The Calabar Bean and its Alkaloids: From Magic, via Miracle, to Memory 9402411909, 9789402411904

Citation preview

Brian Robinson

The Calabar Bean and Its Alkaloids From Magic, via Miracle, to Memory

The Calabar Bean and Its Alkaloids

Brian Robinson

The Calabar Bean and Its Alkaloids From Magic, via Miracle, to Memory

Brian Robinson MSc, PhD, DSc, FRSC Eyam, Derbyshire, United Kingdom

ISBN 978-94-024-1190-4 ISBN 978-94-024-1191-1 https://doi.org/10.1007/978-94-024-1191-1

(eBook)

© Springer Nature B.V. 2023 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature B.V., part of Springer Nature. The registered company address is: Van Godewijckstraat 30, 3311 GX Dordrecht, The Netherlands

My farewell to and with fond memories of Sandra and for the Efik people of Old Calabar

Quotations

Semper aliquid novi Africam adferre (Plinii (the elder), Naturalis Historia, liber VIII) I am afraid that the day is fast approaching when pygmies and bushmen accept Visa and American Express (Neuwinger (1996)) The system of organic chemistry is one of the greatest achievements of the human mind (Sir Frederick Gowland Hopkins as quoted by Sir Robert Robinson (Robinson 1976)) The highest teaching can never be that of him whose chief business is to teach (Roscoe (1874)) If you can look into the seeds of time And say which grain will grow and which will not (William Shakespeare, Macbeth, Act 1, Scene III) Science without religion is lame, religion without science is blind (Einstein (1941)) Nature [she] is but the name For an effect, whose cause is God (Anon, The Buxton Diamonds; or, Grateful Ellen. For the amusement and instruction of children. William Darton, 58, Holborn Hill, London, 1834(?), p. 16) Study to be quiet (I. Thessalonians 4,11 in the New Testament of the Holy Bible)

vii

Dr Brian Robinson (May 1936–February 2023)

The author Brian Robinson has died aged 86, just as this book was completed. He will be remembered as a medicinal chemist who dedicated his career to the understanding of synthetic pathways involved in natural product chemistry and the biosynthetic origins of plant constituents used for medicinal purposes. Brian was born in Hathersage, Derbyshire in May 1936. He received his formal undergraduate and postgraduate education between 1954 and 1960 in the ix

x

Dr Brian Robinson (May 1936–February 2023)

Department of Chemistry at the University of Manchester, during which he obtained the degrees of BSc, MSc and PhD. He subsequently held research fellowships at the Universities of St Andrews and Nottingham, and in 1964 obtained a lectureship in pharmacy at the University of Manchester, where he would spend the rest of his career. In 1970 he was appointed senior lecturer and 10 years later was promoted to Reader in Pharmacy. In 1984 he received a DSc from the university in recognition of his research into the chemistry of indoles and indole alkaloids. During periods spent on leave-of-absence between 1963 and 1986, he held visiting professorships at the University of Illinois, Chicago, USA, and at the University of Sassari, Sardinia, Italy. An active researcher, Brian supervised and mentored many Master’s and PhD students in his lab over the years, and will be remembered by thousands of Pharmacy students as the imposing 60 300 , smartly-dressed, warm-hearted character who introduced them to the wonders of organic chemistry. Those who majored in medicinal chemistry in their final year were treated to a masterclass in heterocyclic chemistry, and knew that an exam essential would be understanding the pivotal [3,3]sigmatropic rearrangement in the Fischer indole synthesis! Brian was a prolific author in the scientific arena. In addition to well over 100 contributions to the literature, he also has several patents to his name, and published a number of books, in particular his seminal treatise The Fischer Indole Synthesis (1983). He also published The History of Pharmaceutical Education in Manchester (1986). In his retirement Brian worked on this book, which would sadly be his last. However – like all of Brian’s books – it is a characteristic blend of meticulous research into the science and historical context, leavened with wry observations. Outside of work, Brian possessed a deep interest in a variety of historical topics of both local and national importance. He was a significant authority on the Royal Maundy, publishing two books on its history: The Royal Maundy (1977) and Silver Pennies & Linen Towels (1992). Other books that reveal the breadth of his interests include Birchinlee: The Workmen's Village of the Derwent Valley Water Board (1983); Walls Across the Valley: Building of the Howden and Derwent Dams (1993); Seven Blunders of the Peak: Some Derbyshire Legends Reassessed (Editor) (1994); and Howden and Derwent: The Building of the Upper Dams of the Derwent Valley Water Board (2004). Brian will be buried in St Michael and All Angels Church at Hathersage where both parents are buried and where plaques commemorate the passing of his brother and sister-in-law Roger and Barbara Robinson. Paul Robinson James Gavin Brian Cox

Preface and Acknowledgements

l-Physostigmine (eserine), the [apparent (however, see Sect. 8.2)] major alkaloidal component of the dried ripe seeds of Physostigma venenosum (Calabar beans), has been for me a generous and stimulating companion throughout my professional scientific life. It appears by both name and structure on the first page of my first published work (Robinson 1958) and as the primary subject matter of what might be my ultimate scientific research paper (Robinson 2002). Following my tenure, from October 1960 to September 1961, as a Wellcome Foundation postdoctoral research fellow in the Department of Pharmacology and Therapeutics at the University of St Andrews – during which the project was conceived – a further postdoctoral research fellowship, from October 1961 to January 1964 at the University of Nottingham, afforded me the opportunity to direct one of my main research endeavours toward an investigation of other alkaloids, the isolation of which from Calabar beans – despite their total alkaloidal content being quite low (circa 0.15–0.3%)1 – had already been reported. However, because of the small quantities of material that had been available, published data on these was usually confined to melting point and either empirical or molecular formulae (Henry 1949; Sumpter and Miller 1954a), but it had been observed that one of them, namely, physovenine, “produces a powerful myotic effect on the pupil of the eye” (Salway 1911). To obviate the expense, if not the impossibility, of my acquisition of large quantities of Calabar beans, Burroughs Wellcome and Co generously afforded me a plentiful supply of the mixed base salicylate residues that remained after their then commercial isolation of l-physostigmine (as its salicylate) from this botanical source. From these residues, I liberated the corresponding mixture of the free bases which, upon partial crystallisation followed by column chromatography, led to an organic chemical Elysium – with crystals, crystals everywhere! Thus, one of the most exciting and formative few hours (Robinson 1964a) of my fifty or so years of scientific research ensued. At this juncture were laid the foundations of what was to become an extremely fertile area of study, which I subsequently developed either by myself or under my supervision and direction of some of the members of the xi

xii

Preface and Acknowledgements

extremely able research group that I ultimately assembled and had the privilege and pleasure to lead within the Victoria University of Manchester. The results from these investigations were not only published but were also presented by me in seminars at various venues throughout Europe, North America and the UK. However, an apposite and, for me, emotive venue for such an exposition was the University of Calabar in Nigeria where, from 10th to 13th April 1988, on what was to be my final (as yet?) of several visits over a period of some twenty-five years to tropical equatorial West Africa, and at the invitation of the West African Society for Pharmacology (Societe Ouest-Africaine de Pharmacologie), I made presentations – to their 17th Annual Scientific Conference concerning “Alkaloids in Medicine” – summarising the results from my by then effected research on the chemistry and pharmacology of the alkaloids of the Calabar bean (Robinson 1988). During this stay, I also visited the nearby tropical rain forest to inspect a massive specimen of the P. venenosum vine that had been located – at the request of the University of Calabar – by nearby villagers. This was, indeed, a rare opportunity since it would appear that “In Calabar nowadays it is difficult to find anyone willing to show the Physostigma plant to an interested party” (Neuwinger 1996), a circumstance that had been previously exemplified when a visitor to Calabar in 1963 “had a good deal of difficulty finding anyone willing to show him the plant from which the bean is obtained” (Holmstedt 1972). The mystique surrounding the Calabar bean and its botanical parent had been earlier manifest by the statements (Barger 1936) that “Before the seeds became an object of commerce they were regarded by the natives with some mystery and were only reluctantly parted with to Europeans” and that “In Old Calabar it was moreover customary to destroy all plants whenever found, except a few which were reserved for judicial purposes”. Unlike many medicinally useful alkaloids which either are or were derived from plants that either still are or were used in folkloric medical practice, l-physostigmine’s botanical source – the Calabar bean – was used to effect folkloric ritualistic executions. Indeed, as noted some years ago (Neuwinger 1996), “Physostigma venenosum seeds belong to one of the most notorious trial by ordeal poisons in West Africa, although the area of use is relatively small. It is found mainly in southeast Nigeria (Calabar Province, the Ibibio, Efik, Ekoi and Aro tribes) and neighbouring Cameroon (Ossidinge District, the Ekoi, Keaka and Boki tribes)”.2 Moreover, it has been claimed (Neuwinger 1996) that “Although one expects a medicinal plant to be as little toxic as possible, experience shows that poisonous plants are the most important source of medicines in Africa” and that “Physostigma is a prime example of the transformation of an extensively-used deadly African poison into a beneficial healing medicine and was an important tool in experimental pharmacology”. In point of fact, if early investigation of the Calabar bean and its use in the ghastly poison ordeal (Macinnis 2004) by the Efiks of Old Calabar (Chapter 1) had not been undertaken, it may consequently have retarded our understanding of some of the fundamental mechanisms occurring in physiology, pharmacology and biochemistry. Undoubtedly, the bean that for centuries was responsible for bringing so much evil

Preface and Acknowledgements

xiii

and death to the Efiks has, through its major alkaloidal component, l-physostigmine, played a fundamental role in the elucidation of the mechanism of neurohumoral transmission at the molecular level (Sect. 10.3); together with curare and atropine, has been the focus of pioneering investigations into pharmacological antagonism [(Henry 1949; Holmstedt 1972; Pal 1900) and (Sects. 10.5 and 10.13) (see footnote 17 of Chap. 1), respectively]; and has also provided a useful treatment – either actual or potential (by providing a template and thereby acting as a “lead compound”) – for a variety of neurological disorders associated with irregularities in cholinergic transmission in which augmentation of cholinergic activity has proved to be beneficial. This has thereby fulfilled the predictions that “We are entitled to infer that the important physiological actions of the kernel of Physostigma may be employed with the greatest advantage in the treatment of disease” (Fraser 1863) and that “The ordeal bean will prove a most valuable addition to the Pharmacopoeia” (Harley 1863). In regard to its above-mentioned use as a pharmacological “tool”, it has been asserted (Flippen-Anderson et al. 2002) that “physostigmine was the first active agent to be isolated as a natural product and used in medicine”. Moreover, it has been the genesis – over the past century or thereabouts – of a sustained international research effort, the progress of which can be traced through a series of reviews (Barger 1936; Bentley 1957; Brossi 1990, 1992; Brossi et al. 1996; Cordell 1981; Coxworth 1965; Dalziel 1948; Henry 1924, 1949; Holmstedt 1972; Hutchinson and Dalziel 1958; Irvine 1961; Karczmar 1970; Longmore and Robinson 1973; Marion 1952; Muhtadi and El-Hawary 1989; Neuwinger 1996; Robinson 1963a, 1964b, 1968, 1971, 1983, 1988, 2002; Saxton 1960; Schneck et al. 1989; Silver 1974b; Sumpter and Miller 1954a; Takano and Ogasawara 1989; Taylor 1966; Triggle et al. 1998; Witkop 1998; Zhao et al. 2004), which was effected without resort to polemics until just over two decades ago (Brossi 1989; Pomponi 1989). However, at this juncture, the situation was superbly defused by the conclusion that “It is hoped that this work will ultimately provide improved drugs for the treatment of geriatric disorders and will continue in many places. After all, patients do not really care whether a drug is developed in Switzerland, Italy or America” (Brossi 1989) – or in the UK? The story of the Calabar bean and its alkaloidal components composes a classical scientific journey throughout some one-and-a-half centuries. Travelling from anthropological, ethnobotanical and ethnopharmacological fieldwork, via a considerable input of innovative organic chemistry and synthesis  including structural elucidation involving the early classical approach of degradation into products of known structures with subsequent retrosynthetic interpretation  to state-of-the-art studies in structural elucidation and in experimental therapeutics, the voyage has reached its current destination where, for example, l-physostigmine is acting as a “lead compound” to afford products that are showing promise (Brossi et al. 1996; Giacobini 2000b; Greig et al. 1995a, 2005a; Iversen et al. 1991; Klein 2007; Marta et al. 1988; Muñoz-Ruiz et al. 2005; Pomponi et al. 1990, 1992; Thal 1991; Yu et al. 2003) for the treatment of that most prevalent form of dementia, namely, of the Alzheimer’s type (Alzheimer’s disease) (Sect. 10.7.2), and have applications (Aarsland et al. 2004) in Parkinson’s disease and dementia with Lewy bodies. It thus offers to the scientific and medical researcher or historian a fascinating tale which not only

xiv

Preface and Acknowledgements

thereby endorses a previous claim (Pomponi et al. 1992) that “The early pioneering investigations on the biochemical and pharmacological action of Phy represent a fascinating piece of medical history” but which is, moreover, still ongoing at the forefront of chemical and medical discovery. As one who has participated in and contributed to this journey for about five decades, it is now my privilege to have this opportunity to present its history (so far, although apparently to the contrary, since lphysostigmine can be considered as the prototype of the insecticidal carbamates (Sect. 10.8) and consequent upon “the curious information (Heyndrickx 1960) that modern pesticides have been used by African witch doctors3 as test poisons” it has been concluded that “The circle of history of the Calabar bean can thereby be considered as closed” (Holmstedt 1972)). Finally, I offer my most sincere gratitude to the staff at the John Rylands University Library of Manchester – especially the young ladies of its Document Supply Unit and the staff of its Information Technology Service Desks – and to John Blunden-Ellis, the faculty team librarian for chemistry in the Joule Library at the University of Manchester, who have all for so long and so willingly put at my disposal their considerable expertise and facilities; to the now late Drs Arnold Brossi and Bernhard Witkop, both of the National Institutes of Health in Bethesda, Maryland, USA, for much stimulating help and encouragement through correspondence; to Miss Panichakorn Jaiyong and Dr Waleed A Zalloum for their expertise in effecting the data input relating to the chemical structures; and to Mrs Sandra Baldwin who was – as she also has with many of my other published works  so skilfully undertaking the long, daunting and arduous task of typing my manuscript and producing it in the formats requested by my publisher but with whom, because of her sudden and untimely death, I shall be unable to share in the ultimate publication of this treatise. Consequent upon Sandra’s grievous passing, I offer my appreciation and wholehearted thanks to Miss Angela Dermody who, after being introduced to me by Mrs Helen Kreissl, has by her skill and hard work completed the monumental task of preparing the data input from my manuscript. Last but, of course, not least, I am mindful of and most sincerely grateful for the help given to me by my publisher, Springer, and in particular that of Drs Jacco Flipsen and Meran Owen and Mrs Ineke Ravesloot for their encouragement and guidance of my study toward its target and for looking after me and my book so well. This became increasingly difficult with the onset of the pandemic, at which juncture Miss Angela Dermody applied her initiative by making available for publication our book’s manuscript, which she had very carefully checked subsequent to my leaving it during my immediate pre-Christmas visit to meet her. Consequently, we were joined by my nephew Paul, to whose late mother and father I had already dedicated footnote 6 of Chap. 10 and to which Paul was to contribute several facets of the crypt-church of St Mary in Lastingham and furthermore to make possible Sects. 8.2 and 8.3 of Chap. 8. It was a pleasure to welcome him to our “team,” to which he has made such a significant contribution. Special thanks to Professor Brian Cox and Dr. James Gavin whose specialist knowledge ensured that this book made it to publication.

Preface and Acknowledgements

xv

Notes 1. From investigations at the Wellcome Chemical Research Laboratories in London, it was noted (Salway 1911, 1912a) that “A representative sample of the Calabar beans . . ., when assayed by the method of the United States Pharmacopoeia, yielded 0.091 per cent of alkaloid” (Salway 1911), but that “The method referred to, however, was found to give results which were much too low, since the amount of physostigmine isolated when working on the large scale, . . ., was equivalent to 0.179 per cent of the material employed” (Salway 1911). Indeed, the total alkaloidal content in the Calabar bean was later (Mutadi and El-Hawary 1989) found to vary between 0.15% and 0.3% and it has also been noted that “The low result by the above method of assay has been ascertained to be due to the fact that three extractions with ether (as required by the Pharmacopoeia) are quite insufficient to remove the alkaloid completely from a solution which has been rendered alkaline with sodium hydrogen carbonate” (Salway 1911). It was found that the alkaloid is more easily extracted from liquids rendered alkaline with sodium carbonate than when sodium hydrogen carbonate is used (Salway 1912a) and the following assay process was therefore suggested:-

Twenty grams of the drug finely powdered are agitated with 200 c.c. of ether, 10 c.c. of 10% aqueous sodium carbonate solution added, and the mixture well shaken at intervals during four hours. One hundred c.c. of the clear ethereal solution are run into a separator, and N/10-sulphuric acid added until the liquid is acid. The separator is well shaken, the acid layer drawn off, and the treatment repeated twice, using 10 c.c. of N/10-acid each time. The acid liquids are combined, made alkaline with 10% sodium carbonate solution, and shaken with ten successive portions (20 c.c. each time) of ether. The combined ethereal extracts are washed with 5 c.c. of distilled water. The solvent is then distilled off, the residue dissolved in 5 c.c. N/10-sulphuric acid, and the excess of acid titrated with N/50-alkali, using iodeosin as indicator (Salway 1912a). 2. Like the Efiks, their Nigerian neighbours to their west, the Ibo (Igbo) have adopted the Physostigma ordeal (Neuwinger 1996) and it has also been reported (Neuwinger 1996) that “trial by ordeal with Physostigma appears to be known in the south of the CAR [Central African Republic] too, from the Cameroon border to the Ubangui River”. 3. “The term ‘witch-doctor’ has been elaborated” (Watt 1956) as being “loosely used with a vague connotation, indicating a type of practitioner who is not really regarded as being a doctor but at the same time is credited with having occult powers. In fact this appellation is a misnomer for there are distinct groups (a) the medicine-man, or herbalist (b) the diviner and (c) the magician. The medicine-man or herbalist is usually a male, hereditary doctor and a dispenser of medicines which are characterised by physical effects, although many of his medicines are ashed and thus rendered inactive. The diviner, on the other hand, is often a woman and operates on a background which assumes that all ills, misfortunes and deaths are due to some active external agency. Although plants, parts of animals and minerals are used by the diviner, he in addition always uses some aid to divination, most commonly the divination dice, “bones” or “delosse”. These are often made of bone but sometimes wood. In addition, the divining bowl is used by the Venda and by the Yao of Nyasaland. The function of the magician is to produce rain, to protect against lightening, hail, flood and other natural calamities, to protect against witchcraft and to vend charms to ward off any evil, including illness”.

Eyam, Derbyshire, UK September 2015

Brian Robinson

Chronology

1846

1855

1861 1863 1863

First European account of the use of the Calabar bean in a native judicial ordeal procedure Investigation of the toxicology of the Calabar bean in the human (by selfexperimentation). Its use suggested for humanely effecting capital punishment Physostigma venenosum botanically described Investigation of the physiological actions and therapeutical uses of the Calabar bean Potential clinical use of the Calabar bean (miosis, antagonised by atropine) recognised

1863–1864

Reports of accidental poisoning with the Calabar bean

1864

Isolation of l-physostigmine (in an amorphous state) l-Physostigmine crystallised and named eserin [eserine] l-Physostigmine used to treat glaucoma by lowering intraocular pressure

1865 1876 1878 1893 1893 1904 1906

Use of the Calabar bean in the administration of trial by ordeal made illegal Isolation of l-eseramine Chemical investigation of l-physostigmine begins Chemical transmission in the autonomic nervous system conceived Steroids isolated from the Calabar bean

WF Daniell

R Christison

JH Balfour TR Fraser TR Fraser A von Graefe DMCLA Robertson TR Fraser J Cameron and JH Evans J Jobst and O Hesse A Vée L Laqueur A Weber (see Sect. 10.5)

A Ehrenberg A Ehrenberg A Petit and M Polonovsky [sic] TR Elliott A Windaus and A Hauth (continued)

xvii

xviii 1911

1915 1921–1935

Chronology Isolation of l-physovenine. Non-alkaloidal components of the Calabar bean investigated Isolation of l-geneserine Discovery of Ach as neurohumoral transmitter

1925

Structure for l-physostigmine proposed

1925 1926

Structure for l-geneserine proposed Recognition of l-physostigmine as an inhibitor of AchE

1934

l-Physostigmine employed in the treatment of Myasthenia gravis (the “miracle” at St Alfege’s) l-Physostigmine synthesised AchE isolated and purified l-Physostigmine found to afford prophylactic protection against intoxication by organophosphates (including “nerve gases”) Pest control with carbamate insecticides Isolation of l-N(8)-norphysostigmine Inhibition of AchE by transcarbamylation

1935 1938 1946

1952 1958 1960 1964 1964 1964–1966

Structure of l-physovenine elucidated Structure of l-eseramine elucidated Structure of l-N(8)-norphysostigmine elucidated

1965 1965 1966 1968

dl-Physovenine synthesised dl-Eseramine synthesised l-Physovenine synthesised Anti-AchE activities of the alkaloids of the Calabar bean investigated in vitro. Proposal of a new pharmacophore for lphysostigmine Structure of l-geneserine revised Absolute configuration of l-physostigmine established

1969 1969

AH Salway

M Polonovski and C Nitzberg HH Dale et al. O Loewi et al. (Dale and Loewi jointly awarded the Nobel Prize in Medicine and Physiology in 1936) G Barger, R Robinson and E Stedman M Polonovski and M Polonovski M Polonovski and M Polonovski O Loewi and E Navratil W Feldberg (see Sect. 10.3) H Fühner (in 1918) (see Sect. 10.3) M Walker

PL Julian and J Pikl D Nachmansohn and E Lederer GB Koelle R Koster

H Gysin et al. J Maier IB Wilson, MA Hatch and S Ginsburg B Robinson B Robinson and G Spiteller RB Longmore and B Robinson G Spiteller and M SpitellerFriedmann RB Longmore and B Robinson B Robinson RB Longmore and B Robinson B Robinson and JB Robinson

C Hootelé RK Hill and GR Newkome RB Longmore and B Robinson (continued)

Chronology 1969

1970 1970 1978 1983 1984–1998

xix Absolute configurations of l-physovenine, l-eseramine, l-N(8)-norphysostigmine and l-geneserine established Anti-AchE activities of d-physostigmine and d-physovenine investigated in vitro Stereochemistry of l-geneserine verified Antinociceptive activity of l-eseroline Treatment of paraplegic anejaculation with l-physostigmine Antiamnesic activity of l-physostigmine and analogues in dementia of the Alzheimer’s type

1986 1987–1989

dl-Geneserine synthesised Antinociceptive activity of:- (i) l-7Bromoeseroline (ii) 6-Hydroxyechibolines

1988

d-Physostigmine found to afford prophylactic protection against intoxication by organophosphates (including “nerve gases”) l-Eseramine synthesised Structure of l-geneserine further revised l-N(8)-Norphysostigmine synthesised Cytotoxic activity of degradation products of l-physostigmine Phenserine, in animal models of Alzheimer’s disease, found to reduce the production of β-amyloid precursor protein, the source of the disease’s toxin β-amyloid

1988 1989 1990 1996 1996–2006

2005

Phenserine subjected to clinical trial for the treatment of dementia of the Alzheimer’s type

RB Longmore and B Robinson

FJ Dale and B Robinson FG Riddell et al. B Robinson and D Moorcroft A and R Bartolini and A Galli et al. P-A Chapelle et al. A Brossi et al. EJ Glamkowski et al. M Pomponi et al. K Fukumoto et al. EJ Glamkowski et al. B Robinson et al. EX Albuquerque et al.

A Brossi et al. A Brossi et al. S Takano et al. GA Cordell et al. Greig et al. Haroutunian Shaw et al. Utsuki et al. Yu et al.

Contents

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

2

l-Physostigmine (Eserine) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 Natural Occurrence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Structure Elucidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Syntheses of:- . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 By Early Approaches, Including the First to be Successful . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2 l-Physostigmine and the 3a–alkyl-1,2,3,3a,8,8a– hexahydropyrrolo[2,3-b]Indole Ring System . . . . . . 2.3.3 d-Physostigmine . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Absolute Configuration, Together with that of the Other Structurally-Established Alkaloids of the Calabar Bean . . . . . 2.5 Biogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Ultraviolet Absorption Spectrum and Reaction in an Acidic Medium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7 Mass Spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8 Detection, Assay and Instability . . . . . . . . . . . . . . . . . . . . . . 2.8.1 Qualitative and Quantitative Analysis . . . . . . . . . . . 2.8.2 Rubreserine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8.3 Eserine Blue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.8.4 Eserine Brown . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . .

25 25 29 36

.

36

. .

48 60

. .

63 71

. . . . . . .

77 80 82 82 84 86 88

l-Physovenine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 Isolation and Structure Elucidation . . . . . . . . . . . . . . . . . . . . . 3.2 Synthesis of the . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.1 Racemate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 l- and d-Enantiomers . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 3a-alkyl-3,3a,8,8a-tetrahydro-2H-furo[2,3-b]indole Ring System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Biogenesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

123 123 127 127 131

3

132 135 xxi

xxii

Contents

4

l-Eseramine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Isolation and Structure Elucidation . . . . . . . . . . . . . . . . . . . . 4.2 Synthesis of the Racemate and the l-enantiomer . . . . . . . . . . 4.2.1 Racemate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2.2 l-enantiomer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

l-N(8)-Norphysostigmine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 5.1 Isolation and Structure Elucidation . . . . . . . . . . . . . . . . . . . . . 145 5.2 Synthesis of the l-enantiomer . . . . . . . . . . . . . . . . . . . . . . . . . 147

6

l-Geneserine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 6.1 Isolation and Structure Elucidation . . . . . . . . . . . . . . . . . . . . . 149 6.2 Synthesis of the l-enantiomer and the Racemate . . . . . . . . . . . 154

7

1

8

Other Alkaloids That Have Been Isolated, or Allegedly So, from the Calabar Bean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1 Calabarine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Eseridine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Isophysostigmine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Calabatine and Calabacine . . . . . . . . . . . . . . . . . . . . . . . . . . 8.5 Investigations Still to Be Effected . . . . . . . . . . . . . . . . . . . . .

. . . . .

141 141 142 142 143

H-, 13C- and 15N–Nuclear Magnetic Resonance Spectra of the Alkaloids of the Calabar Bean . . . . . . . . . . . . . . . . . . . . . . . . 155 . . . . . .

159 159 159 161 162 163

9

Non–Alkaloidal Components of the Calabar Bean . . . . . . . . . . . . . 165

10

Biological Activities of the Alkaloids of the Calabar Bean . . . . . . . 10.1 AntiAchE Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.1.1 AchE – Its Function in Neurohumoral Transmission, Structure and Inhibition . . . . . . . . . . . . . . . . . . . . . 10.2 Pharmacology of l-physostigmine . . . . . . . . . . . . . . . . . . . . . 10.3 Role of l-physostigmine in the Discovery of the Mechanism of Neurohumoral Transmission . . . . . . . . . . . . . . . . . . . . . . 10.4 AntiAchE Activities of the Minor Alkaloids of the Calabar Bean – The l-physostigmine Pharmacophore . . . . . . . . . . . . . 10.5 Clinical use of l-physostigmine in Ophthalmology – Miotic Activity and the Reduction of Intraocular Pressure in Glaucoma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6 Use of l-physostigmine in the Treatment of: - . . . . . . . . . . . . 10.6.1 Myasthenia Gravis . . . . . . . . . . . . . . . . . . . . . . . . . 10.6.2 Paraplegic Anejaculation . . . . . . . . . . . . . . . . . . . . .

. 169 . 169 . 169 . 180 . 181 . 186

. . . .

192 194 194 196

Contents

10.7

10.8 10.9

10.10 10.11 10.12 10.13

xxiii

Use of l-physostigmine in the:- . . . . . . . . . . . . . . . . . . . . . . . . 10.7.1 Prophylactic Protection Against Intoxication by Organophosphates (Including “Nerve Gases”) . . . . . . 10.7.2 Enhancement of Cognition and Memory (Including Antiamnesic Activity in Dementia of the Alzheimer’s Type – Alzheimer’s Disease) . . . . . . . . . . . . . . . . . . . Bactericidal and Insecticidal Activities of l-Physostigmine . . . . Prophylactic Protection with d-physostigmine Against Intoxication by Organophosphates (Including “Nerve Gases”) . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Clinical Uses of l-physostigmine . . . . . . . . . . . . . . . . . . Antinociceptive Activity of l-eseroline and Its Synthetic Analogues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cytotoxicities of Rubreserine and another Structurally-Related Degradation Product of l-Physostigmine . . . . . . . . . . . . . . . . . Poisonings (Either Accidental or Malicious) with the Calabar Bean . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

198 198

199 204

205 207 208 213 214

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265

Abbreviations

Å Ach AchE Bch BchE bp cd chE cm.(cm) cps (d) DAT DMSO ee Et ft. g h in. ir J kg m (m) L M m/e Me mg mm. mm

Angstrom unit(s) Acetylcholine Acetylcholinesterase Butyrylcholine Butyrylcholinesterase Boiling point Circular dichroism Cholinesterase Centimetres Cycles per second (in nmr) Doublet (in nmr) Dementia of the Alzheimer’s type Dimethyl sulphoxide Enantiomeric excess Ethyl Feet Gram(s) Hour(s) Inches Infrared Coupling constant (in nmr) Kilogram(s) Minute(s) Multiplet (in nmr) Molecular ion (in mass spectrometry) Mass to charge ratio (in mass spectrometry) Methyl Milligram(s) Millimetres Millimetre(s) (of mercury re bp) xxv

xxvi

mμ μ mp nm nmr nPr nOe ord Ph ppm (s) (t) tBu uv

Abbreviations

Millimicron(s) (¼ nm and nm) Microns Melting point Nanometre(s) Nuclear magnetic resonance Normal propyl Nuclear Overhauser effect Optical rotatory dispersion Phenyl Parts per million (in nmr) Singlet (in nmr) Triplet (in nmr) Tertiary butyl Ultraviolet

Chapter 1

Introduction

Beginning with their traders, whose appetites had been whetted by Henry the Navigator, and through the efforts of Fernăo Gomes, a Lisbon merchant, whose ships eventually reached the island of Fernando Po and ultimately crossed the equator – and during the course of which discovered an area in which gold was evident in such abundance that it was named the Gold Coast (Holmstedt 1972) [this territory, along with that of the former Ashanti to its immediate north (Fig. 1.1) (Rattray 1923) are now embraced in present-day Ghana], the colonisation and commercial exploitation of the nations along the West African coastal region were initiated during the late fifteenth century by the Portuguese who, within a few decades, were ousted in turn by the Dutch, English, French, and some lesser European nations (Holmstedt 1972). It followed the well-established pattern of infiltration by Christian missionaries who destabilised, eroded, and ultimately destroyed the religious and social infrastructures of the indigenous communities and thereby rendered them vulnerable to subsequent military and commercial exploitation. Thus, this consecutive application of bible and bullet created the so-called Slave Coast with the beginning of the trade in slaves which, by the eighteenth century, had developed into a massive traffic dominated by English companies. These slaves were regarded by European nations primarily as a source of labour for the plantations which they were establishing in the West Indies and other tropical parts of the then so-called New World. They were secured by European traders by purchase from African merchants who, in turn, obtained the majority of them from the interior countryside (Holmstedt 1972). This was a despicable business that was perpetrated through the connivance of both African and European and which, until the late 1830s, when it was replaced by more lucrative, if not extensive, forms of commerce in which cotton and other British manufactured goods were received in exchange for exports of native produce (Daniell 1846), also enjoyed the tacit consent of the Christian missionaries. Notwithstanding this, such missionaries, with the support of the social and business establishments within the colonial – usually British – authority, had the audacity to begin to suppress various local customs which, in their opinion, they found were © Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_1

1

2

1 Introduction

Fig. 1.1 The former territories of Ashanti and Gold Coast [from (Rattray 1923)] which are now embraced in present-day Ghana. The author’s first visit to tropical equatorial West Africa resulted from his tenure – on leave-of-absence from the University of Nottingham – during 1963 from January to September, inclusively, of a Visiting Senior Lectureship at the Kwame Nkrumah University at Kumasi (formerly Coomassie of the Ashanti territory) in Ghana

1 Introduction

3

particularly repugnant to the [slave-trading] Europeans. One of these customs, but one which was an important judicial procedure when perjury was so common that it was recognised as being unsound to permit mere human testimony to condemn an individual, and one which therefore at first carried with it the sanction of the clergy as well as of the civil power (Dragstedt 1945), and probably the most mysterious and gruesome in a manifestly cruel, superstitious and violent society,1 was ésere, the poison ordeal1,2,3,4,5,6 that was practised by the Efik people of Old Calabar (see footnote 2 in Preface and Acknowledgements) who lived on the Cross River in what is now the Calabar Province in southeastern Nigeria7,8,9,10,11 (Holmstedt 1972). Every major Efik settlement had a senior obong, or chief – many of whom carried the title of duke or king (Holmstedt 1972) (with which they had been nonsensically endowed by the colonising Europeans) – who always held high rank (usually as head of its highest grade in the settlement) in the Egbo, or Leopard, Society, the most important male association in the Efik community. Under the aegis of the chief and the town’s elders, Egbo disseminated and enforced laws, judged important cases and constituted the actual executive government of the Efik, enforcing where necessary its laws by capital punishment and to most of the Efik people it represented a mysterious overlordship (Holmstedt 1972). It is therefore not surprising that society in Old Calabar suffered from many superstitions, there being many Efik divinations, ordeals and omens, and an intense fear of witchcraft1 which Harvard’s Clyde Kluckhohn, one of the most notable anthropologists of the present century, so superbly defined (Kluckhohn 1967) as “the influencing of events by supernatural techniques that are socially disapproved”1,2,11. For example, in order to decide upon their guilt or innocence, those accused of crime – primarily of causing death by witchcraft – were made1,2,11 to drink a poisonous mixture usually prepared by pounding and macerating in water, which thereby acquired a milky colour (Daniell 1846), several of the poisonous beans which take their name from this part of West Africa where they grow, namely “Calabar beans”. These are also called “Ésere nuts”, “beans of Etu Ésere” or “chop nuts” (Dragstedt 1945) and are the seeds of the ripe fruit of an aquatic leguminous large vine which is perennial (probably only after some years producing fruit – but, in common with that of many other tropical plants, this ripens at all seasons of the year, affording the most abundant crop in the rainy season from June to September) (Fraser 1863) and is described as a “runner, climbing on the bushes and trees in its neighbourhood” (Fraser 1863) and which, with feeder-stems some 2–3 in. thick, reaches heights of over 50 ft. and is now known as Physostigma venenosum (Figs. 1.2, 1.3 and 1.4) (vide infra). It would appear that the natives themselves did not seem to have much faith in the bean as an ordeal, rather looking upon a summons to undergo the test as a sentence of death, and, if in their power, making their escape and going into exile (Balfour 1861). Nevertheless, remarkably large numbers of free-born men and women accused of witchcraft underwent the ésere ordeal – indeed, it has been noted (Neuwinger 1996) that “The small population of the Uwet in inland Calabar wiped themselves out by this method”, an estimate that in 1840 among a population of some 100,000 with whom the custom prevailed about 100 to 120 were killed annually in this fashion (Dragstedt 1945) being made by William Freeman Daniell12,

4

1 Introduction

Fig. 1.2 Physostigma venenosum. Fig. 1. Branch with pinnately trifoliolate leaves and nodosoracemose infloresence, showing also entire flowers [purple-coloured (Irvine 1961)], persistent calyx and young pod. Fig. 2. Vexillum separated. Fig. 3. Alae. Fig. 4. Carina. Fig. 5. Diadelphous stamens. Fig. 6. Upper part of style, bearded, and with cucullate stigma. Fig. 7. Upper part of bearded style with stigmatic hood laid open. Fig. 8. Calyx and young legume. Figs. 6, 7, 8, magnified [from (Balfour 1861) 19
24 cm]

1 Introduction

5

Fig. 1.3 Fig. 1. Young pod of Physostigma venenosum, with three ovules. Fig. 2. Full grown pods of Do. Fig. 3. Seed or Ordeal Bean seen literally. Fig. 4. The same showing the sulcate and extended hilum on the convex edge. All the figures natural size [from (Balfour 1861) 18
23 cm]

6

1 Introduction

Fig. 1.4 Physostigma venenosum Balf. A, seed, showing hilum. B, pistil. C, zigzag flower-stalk. D, tendril [from (Irvine 1961) 12
19 cm]

1 Introduction

7

who, it has been claimed, was the first European13 to record and bring to scientific attention this judicial use of the Calabar bean – in a paper read by him to the Ethnological Society on 28th January 1846, the relevant sections of which are reported as follows (Daniell 1846):The government of the Old Callebar towns is a monarchical despotism, rather mild in its general character, although sometimes severe and absolute in its details. The king and chief inhabitants ordinarily constitute a court of justice, in which all country disputes are adjusted, and to which every prisoner suspected of capital offences is brought, to undergo examination and judgement. If found guilty, they are usually forced to swallow a deadly potion, made from the poisonous seeds of an aquatic leguminous plant, which rapidly destroys life. This poison is obtained by pounding the seeds and macerating them in water, which acquires a white milky colour. The condemned person, after swallowing a certain portion of the liquid, is ordered to walk about until its effects become palpable. If, however, after the lapse of a definite period, the accused should be so fortunate as to throw the poison from off the stomach, he is considered as innocent, and allowed to depart unmolested. In native parlance this ordeal is designated as “chopping nut.” Decapitation is also practised, but not so much amongst criminals as the former process, being more employed for the immolation of the victims at the funeral obsequies of some great personage. Drowning is sometimes resorted to as a substitute for the first means of destroying life. . . Many cruel and superstitious ceremonies occur upon the death of any influential personage, whether male or female. They mourn for some weeks, which is indicated by their binding a black silk handkerchief across the forehead, and neither washing their body nor changing their clothes; being therefore literally in sackcloth and ashes during the allotted period. Two or three days elapse after the inhumation of the body, when several guns and muskets are fired off, and a proportionate quantity of slaves decapitated to accompany the deceased into the next world. Wives, friends, and confidential servants alike share the same fate, if the departed individual be a man of consequence. Upon the death of Duke Ephraim, one of the former kings of Old Callebar, some hundreds of men, women, and children, were immolated to his manes – decapitation, burial alive, and the administration of the poison-nut, being the methods resorted to for terminating their existence.

Daniell’s above account (Daniell 1846), and as it is quoted by Barger (1936), would appear to confuse the use of the Calabar bean as a means of effecting capital punishment (the accused has already been “found guilty”) and as a trial by ordeal (if the accused vomits the poison and survives, “he is considered as innocent”). However, no such confusion is inherent in a much later account of the judicial use of the Calabar bean that was written by Donald C Simmons, an anthropologist who conducted ethnological research in the area during 1951 and 1953. He states that (Simmons 1956):The Efik [see footnote 2 in Preface and Acknowledgements] believe the esere or Calabar bean (Physostigma venenosum) possesses the power of destroying witchcraft. An individual accused of witchcraft usually demanded his right to undergo the Calabar bean ordeal in order to establish his innocence. The suspect ate eight Calabar beans, and then drank an infusion of several ground Calabar beans and water [interestingly, it has been reported (Neuwinger 1996) that one accused individual survived the unhesitating eating of 24 beans and that a maximum of 35 beans were likewise without a fatal consequence, it being suggested (Neuwinger 1996) that a large number of beans resulted in an increase in emetic activity and thereby in protection against absorption of the beans’ toxic component by the gastrointestinal tract (vide infra)]. If the suspect possessed witchcraft, his mouth shook and mucus came from his nose, but if innocent of witchcraft, he lifted his right hand and then

8

1 Introduction regurgitated. If the poison continued to affect the suspect after he established his innocence, the suspect received an infusion of the excrement of a person of the same sex mixed with water previously used to wash the external genitalia of a female. If the ordeal revealed guilt, the suspect died without aid, and the corpse was thrown into the forest.

If the accused died of the ordeal, which they usually did, this was taken as being in consequence of their guilt. However, in the cases in which they survived, ingestion of the beans had been followed by vomiting. In such circumstances, the accuser would then be compelled to submit to the deadly éseré. Thus, accusations of witchcraft and other crimes were not made lightly and a salutary check was thereby placed upon treachery or private enmity (Fraser 1863), unlike the accusations of witchcraft that were rampant in England just a few centuries ago and where the accuser acted with impunity (Catlow 1976). Occasionally, violent purging that was not accompanied by death occurred, in which instances guilt was again considered proven and the subject was sold into slavery (Robinson 1988a). The purging activity of the beans was also utilised to cleanse bodies that had been allegedly infected by witchcraft, such a use having been recently attested to (Udofia 1988) when some 50 years previously the attester – who was introduced to the author in 1988 – and three of his friends, all four of whom were then in their early teens, had each been given orally one quarter of 40 beans prepared as above – all four recipients were purged and survived (Robinson 1988a). Two explanations have been advanced for the rationale underlying this use of Calabar beans in detecting witches, One is that the traditional judge – the fetish-man who administered the poison – in some cases being convinced that the accused was not guilty, may have therefore either used seeds which very closely resembled Calabar beans but which he knew to be less toxic, or used Calabar beans that had been pre-boiled (in water) which he knew would render them less toxic (Balfour 1861; Neuwinger 1996), it now being known that the beans’ main toxic component is l-physostigmine and that this undergoes decomposition, via hydrolysis of its carbamyloxy group – as would be effected by boiling in an aqueous medium – to afford the much less toxic product, l-eseroline (Sects. 2.8 and 10.1.1) (Coxworth 1965). The other explanation (Ault 2008; Neuwinger 1996; Silver 1974a; Sofowora 1984; Stenberg et al. 1977; Wolff 1960), is that ingestion of Calabar beans is followed by vomiting (vide supra) that is caused by an emetic, the nature of which is as yet unknown, in the seed hull. A guilty person, afraid of the ordeal, might drink the potion and eat hesitantly, thereby resulting in a gradual and steady absorption of the toxic l-physostigmine thus leading to death. On the other hand, the accused but innocent person, having nothing to fear and convinced that they would not die, might not show such hesitancy and the resultant gastric irritation caused the onset of the vomiting which may have been rapid enough to offer protection against absorption of fatal amounts of the toxic alkaloid. Thus, along with this latter, “the simultaneous occurrence of an emetic in the bean, saved the lives of several prisoners and declared them innocent” (Stenberg et al. 1977) (see also Neuwinger 1996 and Sect. 10.13) and therefore “the belief could have had a physiological basis” (Silver 1974a). Moreover, the éseré (a word that would appear to stand for both the poison ordeal and the bean) was also used to effect other ritual killings and, in addition, it also

1 Introduction

9

became increasingly used as either a political or a religious weapon by removing dangerous rivals, weakening powerful houses and settling old scores, and was sometimes resorted to as a means of gratifying private revenge – in common with, for example, the seventeenth century witchcraft trials by ordeal in England (Catlow 1976). In addition, the officials charged with administering the poison may in some cases either have shown favouritism (vide supra) or if, from any cause, they desired the death of their victim, a club might have been employed to compensate for either the slowness or the failure of the action of the poison (Fraser 1863; Neuwinger 1996). It is, therefore, perhaps not surprising – and probably in light of the trials by ordeal for witchcraft that until a few decades earlier had been rife in England but which were by then discredited (Catlow 1976) – that the British colonial authorities sought to outlaw the practice of the éseré ordeal. Thus, in 1878, the then British consul, Hopkins, visited Calabar and, on the instruction of Lord Salisbury – the then Foreign Secretary, drew up an agreement of 15 articles with the kings and chieftains of Duke Town in Calabar to put an end to the many murderous customs which were then practised in the country of which the following two concerned the use of the Calabar bean (Holmstedt 1972; Witkop 1998):Article three: Any person administering the esere bean, whether the person taking it dies or not, shall be considered guilty of murder, and shall suffer death. Article four: Every person taking the esere bean wilfully, either for the purpose of committing suicide, or for the purpose of attempting to prove their innocence of any crimes of which they may have been accused, shall be considered guilty of attempted murder, and shall be fined as heavily as the circumstances will permit, and shall be banished from the country.

However, the efforts by the former British colonial government to eliminate the use of the Calabar bean as an ordeal poison – or even its possession, and attempts to destroy and prohibit the cultivation of its parent plant were attended with difficulties, since the plant could, and often still does, grow wild in the tropical rain forest’s more distant and vegetatively dense tracts, its habitat, since it thrives best on swampy river banks, being the sides and edges of streams (Fraser 1863; Robinson 1988a). From here, considerable quantities of the ripe seed pods, each containing either two or three beans (Figs. 1.3 and 1.4), float down the local streams and rivers into which they have fallen as they have ripened in the remote upstream areas (Fraser 1863; Robinson 1988a). Thus the beans can be easily gathered either from the banks of these water-courses, or by local fishermen, and ultimately find their way to the local markets (Fraser 1863; Robinson 1988a) where, despite Nigerian law – which not only currently forbids either the possession or the use of the Calabar bean (Witkop 1998) but also punishes even it’s possession by a fine and imprisonment (Neuwinger 1996) – they are sold to the local population (indeed, the author had no difficulty in 1988 in locating some for sale in the market at Calabar) – in which belief in witchcraft at all levels of society widely abounds – for use as a means of warding off evil (Simmons 1956). Thus, the bean is often kept under the pillow to ensure a dreamless and safe night’s sleep and to keep at bay the night’s witches, worn on the body as a means of protection against evil spirits, and kept in purses to protect their contents (Robinson 1988a). Furthermore, it has been claimed (Neuwinger 1996)

10

1 Introduction

that “Although strictly forbidden the trials were still being held occasionally in secret in the Seventies” and that “belief in the power of the éseré beans is still alive. Some of the inhabitants of Calabar keep one bean today in their money pouches; they are said to stop witches from making the money disappear”. Moreover, the tradition of the Efiks has refused to be thwarted since they have a second ordeal poison, called mbiam.10 However, long before the éseré ordeal had been officially eliminated, the story of the Calabar bean had entered a scientific phase which began and continued for some time during the mid-nineteenth century (Balfour 1861; Christison 1855; Fraser 1863; Gaddum 1962; Holmstedt 1972) in the School of Medicine at the University of Edinburgh.14,15 This was an understandable location when it is realised that many of that university’s former students, who had a taste for natural science and with whom the staff retained close contact, were employed in the Church of Scotland’s mission which was very active amongst the Efiks in Calabar – whence it spread to other locations (Balfour 1861; Holmstedt 1972).13 The plant that produces the Calabar bean has been fully described botanically and pharmacognostically (Berg and Schmidt 1894; Hutchinson and Dalziel 1958; Irvine 1961) [quoted as ref. 3, 1 and 2, respectively, in (Robinson 1964b)], with initial brief descriptions being given in The Missionary Record13 which began its publication in 1846 (Holmstedt 1972). However, the first botanical description of the whole plant as Physostigma venenosum was made (Figs. 1.2, 1.3 and 1.4) by John Hutton Balfour (1861),16 Professor of Medicine and Botany in the University of Edinburgh and a man of strong Christian persuasions (Holmstedt 1972) who obtained specimens from the many friends and ex-pupils that he had in the Scottish missionary services then active in Calabar. Thus it was that, in 1859, the Rev. WC Thom son, a Prestebyterian missionary, procured flowering specimens for him (Barger 1936) and, after examining their leafy twigs, flowers and fruit at Edinburgh, Balfour “acknowledged them to be a hitherto unknown plant and established the new genus Physostigma [from the Greek meaning “an inflated or bladder-like stigma” (Robinson 1988a)] of the Leguminosae-Papilionaceae with its only known species, venenosum [from the Latin meaning “full of poison” (Robinson 1988a)]” (Neuwinger 1996). Balfour’s description was also reported16 by TR Fraser,17 one of his colleagues in the School of Medicine at the University of Edinburgh,14 and the vine has also been delineated as follows:1. “A climber woody at base, up to 20ft. high or more; branches twining, glabrous; leaves trifoliolate, 6
4 in., 3-nerved at base, central leaflet broadly ovate, laterals obliquely ovate and unequal-sided, base rounded, acuminate; flowers (Aug.) purple, shell-shaped, in axillary drooping racemes, axis stout and zigzag, with swollen nodes; fruit narrowed at ends, 6
2 in., yellowish brown, glabrous, edges raised, veins prominent; seeds few, ellipsoid, over 1 in. long, one side nearly straight, the other grooved by depressed hilum” (Fig. 1.4) (Irvine 1961). 2. “Climber; branches glabrous; leaves 3-foliolate, the central one broadly ovate, the lateral obliquely ovate, rounded at base, abruptly acuminate, up to 16 cm. long and 10 cm. broad, prominently 3-nerved at base, glabrous; stipels semi-lunar;

Notes

11

racemes axillary, drooping, the axis stout and zigzag with swollen nodes; pedicels curved, up to 5 mm. long; calyx broadly cupular, shortly lobed; standard enclosing the wings and keel, much curved round towards the calyx; fruits straight, narrowed at both ends, up to 15 cm. long and 4 cm. broad, glabrous; seeds few, 3 cm. long, ellipsoid, with one side nearly straight, the other side grooved by the depressed hilum” (Hutchinson and Dalziel 1958). 3. “High-climbing large liana or shrubby climber with woody base and cylindrical dark green glabrous branches twining high, trifoliolate leaves and pink or purple flowers curved like shells. Leaves alternate, 3-foliolate, the central one broadly ovate, the lateral obliquely ovate, rounded or weakly cordate at base, each having a struma, which serves as a short petiole, and two small thick acute stipels, abruptly acuminate, prominently 3-nerved at base, 10–16
7–10 cm, petiole about 8 cm long with two triangular stipules on the base. Flowers in axillary pendulous many-flowered racemes, rachis of each raceme robust and zigzag-shaped with swollen fleshy nodes; curved shell-like, veined with pale pink, having a purplish tinge; 10 stamens, stigma blunt, covered by a remarkable ventricular sac or hood, pedicels curved, up to 5 mm long, calyx broadly campanulate, with short, subdeltoid, indistinct teeth. Fruit straight, oblong, narrowed at both ends, with an apicular curved point, up to 15
5 cm, yellow-brown to dark-brown. Seeds two or three, separated from each other by a woolly substance; 2.5
1.8 cm, irregular reniform, with one shorter margin concave to nearly straight, the other longer, convex; this margin with a groove containing the greyish remains of the funiculus; the two surfaces flattened, black-brown, shining, white inside, extremely hard” (Neuwinger 1996).

Notes 1. However, before levelling condemnation, we should remember that witchcraft is one of the most ancient superstitions and that the persecution of the so-called witches has been a permanent feature of the history of mankind which only in recent times became a full-scale heresy hunt, initially in mainland Europe and then in England. Consequently, barbarities were perpetrated in England just a few centuries ago as a result of religious dogma and bigotry, when executions either by being burned alive whilst tied to a stake or by being hanged by the neck until semi-conscious and then, whilst still alive, being disembowelled, dismembered and, finally, beheaded. 2. Relating to this, Robert Christison (1855) wrote “We have no right, however, to express any astonishment at this folly of the benighted pagan African, when we reflect how short a time has gone by since witchcraft was generally believed in throughout civilised Christian Europe; and when the only way of meeting a charge, no less easy to make than difficult to repel, was by undergoing an ordeal of some kind (see footnote 3), quite as preposterous as that by swallowing a deadly poison” 3. The subject of earlier review (Dragstedt 1945), a trial by ordeal (from the Anglo-Saxon “ordaal” – “or” ¼ primitive, “daal” ¼ judgement) is an extremely ancient form of judgement that may generally be considered to have been practiced in every country and by every race from early times and, to some extent, to persist even today. It was used, for example, to

12

4.

5.

6.

7.

8. 9.

1 Introduction ascertain either the guilt or innocence of an accused individual, to detect the guilty member of a group of subjects, and to settle disputes between contesting individuals. For the most part it was a method in which a personage was required to either perform some act, or undergo some test, which would be hurtful to them unless they were protected by divine assistance. Indeed, one of the earliest accounts of such a trial is to be found in the Book of Numbers, chapter 5, verses 18–31) (BC 1490–1451) in the Old Testament of the Holy Bible, where for the detection of conjugal unfaithfulness, a woman was to be commanded “to drink the bitter water that causeth the curse”, whereupon her innocence was indicated be she “free from this bitter water that causeth the curse”. On the other hand, she is told that if she “hast gone aside, to another instead of thy husband, and if thou be defiled, and some man have lain with thee beside thine husband”, then “this water that causeth the curse shall go into thy bowels, to make thy belly to swell, and thy thigh to rot” – the god of love and forgiveness clearly awaited the arrival of the New Testament! Ordeals may be classified under two headings, namely those in which application is external and those in which it is internal. The former type includes the well-known trials by fire (see footnote 4), water (see footnote 5), balance (see footnote 6), rice eating (see footnote 7) and combat (see footnote 8) and the modern day lie-detection (see footnote 9), whereas the latter type comprises perhaps the most curious and interesting forms and encompasses a large selection of vegetable ordeal potions. These include various barks, roots, trees, nuts and beans, and belonging to which in different parts of Africa are a number of potent poisons (see footnote 10), amongst which is the Calabar bean. In which the accused, with a minimal damage to themselves in order to establish their innocence, had to walk through seven circles of fire, and also sometimes to carry a red-hot iron ball in their hands – also too were the ordeals of boiling water or oil into which the accused was required to plunge their hand or arm in order to retrieve either a ring or stone. It is likely that justice may have been directed in all these ordeals by application of information regarding prophylactic treatments that were known to, for example, the fakirs (Dragstedt 1945). In which the accused was required to remain submerged in cold water until either a relative or friend could retrieve an arrow that had been fired from a bow. Some validity for this ordeal might lay within the probability that emotional disturbance in the guilty might make them less able to hold their breath for a long period of time (Dragstedt 1945). An ordeal that has it’s origins in India, where, after appropriate ceremonies, and when no wind was blowing, the accused was placed on a scale and accurately counterbalanced, after which the accusation written on a piece of paper was bound to their head – a subsequent increase in the accused weight was taken to indicate their guilt, whereas a loss of weight was a verification of innocence, although as used later in Europe, the implications were reversed and the accused who lost weight was taken as being guilty – it is clear that either sweating, urination or defecation may have played decisive roles in determining the fate of an anxious accused subject (Dragstedt 1945). This was used when it was necessary to determine the guilty from amongst a number of suspects, all of whom were given some dry rice to chew and then, after a predetermined period of time it was spat out onto a banyan leaf. Rice that was moist and softened with saliva indicated innocence but if the rice when spat out was dry or tinged with blood, the suspect was judged to be guilty. In a later related Anglo-Saxon ordeal, the guilt or innocence of an accused was established by their ability to swallow a prescribed amount of consecrated bread or cheese within a prescribed period of time. The validity of the results from these ordeals may be found in the now generally recognised influence of the emotions on the functions of the alimentary canal and upon the motility of the gastrointestinal tract (Dragsterdt 1945). In which the innocent may at least have had a psychological advantage (Dragstedt 1945). Having only a quasi judicial status, in this modern version of the ancient practice of ordeal, an accused is subjected to nothing more harmful that a barrage of questions, but the delicacy with which the emotional turmoil produced by telling a lie is recorded provides the same sort of results, for the same sort of reasons, as some of the ordeals of a bygone age (Dragstedt 1945).

Notes

13

10. The ordeal poison mbiam is, according to Simmons (1956), “a magical liquid which possesses the power of killing anyone who swears a false oath in its name” and “since Nigerian law has declared the Calabar bean ordeal illegal, mbiam is now given to individuals suspected of witchcraft. The suspect denies the accusation of witchcraft and climaxes his denial by drinking the magical liquid; if the accused swore falsely, the magical liquid will soon kill him.” Although its use as an ordeal poison is time-honoured, since it has been reported (Anon 1986) that in the pre-colonial social organisation of the Efik “The oath-taking ‘Mbiam’ was a common practice for ascertaining innocence or guilt in the region” and the caption to a related illustration reads:“Efik Chiefs swearing ‘Mbiam’ at the ending of the 19th century” (Anon 1986). It would appear that “it [mbiam] has never been investigated chemically” and “Nothing is known about its source or possible toxicological action” (Holmstedt 1972). “The Ibo use other trial poisons too: Erythrophleum suaveolens, Strychnos densiflora, Erythrina senegalensis, Tabernaemontana spp. and Solanum verbascifolium” but whereas “Physostigma is always used alone, the others are mixed together and the decoction is dispensed” (Neuwinger 1996).A number of other potent poisons, including the mbundu root (Imbando) and the tangena nut {Tanghinia veneniflua [venenifera Poir (Watt and BreyerBrandwijk 1962)]} have been likewise used in other areas of the African continent (Dragstedt 1945) and amongst plants which have been used for trial by ordeal in Southern and Eastern Africa, namely Entada abyssinica Steud. (the mode of administration of which is unique in that the juice of the bark and of the cambium is introduced underneath the eyelid, a procedure worthy of further investigation), Erythrophleum guineense G. Don. and other Erythrophleum sp. [the diterpenoid alkaloid cassaine is present in sassay bark – E. guineense – which was also formerly used as an ordeal poison in West Africa (Dalma 1954; Humber and Taylor 1955; Taylor 1963)], Parkia filicoidea Welw. Ex Oliv., Strophanthus courmontii Sacl., S. eminii Asch. & Pax, S. grandiflorus Gilg., S. holosericeus K. Sch. & Gilg., S. kombé Oliv., S. mirabilis Gilg., S. nicholsonii Holm. and S. petersianus Klotzsch, have been specifically referred to - and others have been listed (Watt and Breyer-Brandwijk 1962). Thonga witch-doctors (see footnote 3 in Preface and Acknowledgements) prepare a drink for trial by ordeal from the powdered root of the tree, known by the European names as Molana, Olifantsoor, Appelblaar and Mbandu, and the leaf of Datura stramonium, the related ceremony being known as “drinking the mpondo” (Watt 1956). 11. Witchcraft was a crime so held in abhorrence that rather than lie under suspicion, an accused sometimes would, in order to establish their innocence, insist on being subjected to the ordeal. Thus, for example, it is reported (Balfour 1861) that “a woman who was accused of injuring her child by witchcraft, came in from a distance, strong in innocence, and demanded to have the ordeal administered. She ate twenty-four beans and did not die. Next day, another woman, encouraged by her escape, underwent the ordeal, and she ate twenty-two beans, and died. There was no vomiting in either case. The difference of effect might be owing either to an actual difference in the beans administered in the two cases, or to their mode of preparation” (Chap. 1). 12. William Freeman Daniell (1818–1865), (Rodin 1947) A Liverpudlian by birth, Daniell was awarded the diploma of MRCS in 1841, consequent upon which he joined the British Army’s medical service and spent some 20 years in service on the then pernicious West African coast. It was early in this period that he first observed the use of the Calabar bean in Efik judicial procedure, and the first known record (Chap. 1) (see footnote 13) of its application appears in a paper (Daniell 1846) that he presented to the Ethnological Society of Edinburgh – once again associating the Scottish capital city with the Calabar bean. Daniell returned from West Africa to England in 1863 and later became part of the “Expeditionary Force” to China – being present at the taking of Peking – and proceeded to West India a short time afterward. On his next homecoming his health was broken and he died in 1865. Whilst he was in foreign service he had collected and described many native plants and his publications are concerned with the plants, shrubs and trees, and their fruits, that were

14

1 Introduction

indigenous to those countries that he visited. One of these works relates to the Katemfé, or miraculous fruit of the Sudan and which, as a compliment to him, the botanist who first studied it, was later named Phrynium Danielli, Benn. 13. However, it was reported in 1863 (Harley) that “ Fifteen or twenty years ago Messrs Waddell, Young, Baillie and Taylor, missionaries of the United Presbyterian Church of Scotland, on the West Coast of Africa, gave in the [The] Missionary Record a description of the ordeal bean, and detailed the effect they had seen it produce on the natives of that portion of Africa where they were stationed. These gentlemen also sent to this country, several specimens of the bean.” 14. The School of Medicine at the University of Edinburgh and its role in early studies on the Calabar bean and its alkaloids (Gaddum 1962; Holmstedt 1972) It was about 1670 that the School of Medicine was inaugurated in Edinburgh by a group of physicians who established a college for medical training and laid out a herb garden. The latter was essential since at that time most therapeutic agents were obtained from plants and the only way to obtain reliable medicines was to grow the plants. Consequently, in 1676, a professor of botany was appointed to teach his subject to the medical students, an arrangement that was to continue for sixty-two years. In 1738, Charles Alston (1683–1761) was appointed as professor of both botany and materia medica, the latter then being a major and important subject since in those early days the preparation of medicines was the physicians’ responsibility. Following Alston’s death, John Hope (1725–1786), who introduced the Linnaean system for the classification of plants, succeeded to the joint chair but, in 1768, the two subjects went their different ways and although the medical men of Edinburgh, amongst whom was John Hutton Balfour (1808–1884) (see footnote 16), continued to be good botanists, an independent department of materia medica was created, with Francis Home (1719–1813) as it’s first professor and who gave lectures for 30 years in Edinburgh on materia medica. The next occupant of this chair was Francis Home’s son, James (1760–1844) who, although he did little research, was a fine teacher, with several of his pupils making names for themselves in later life. Thus, for example, Richard Bright and Thomas Addison received the honour of having named after them a form of nephritis known as Bright’s disease and an adrenal deficiency called Addison’s disease, respectively. The next professor of materia medica was Andrew Duncan (junior) (1773–1832) who, in 1832, was succeeded by Robert Christison (1797–1882). Although he was then only 35 years old, Christison had already held the newly-established (and first of its kind in Great Britain) chair of medical jurisprudence in Edinburgh for 10 years, during which time he had been involved in fundamental work on the detection of crime. Thus, for example, when he gave evidence at the trial of Burke and Hare – who made money by selling corpses, which they obtained by murder, to the professor of anatomy – it depended partly on bruises, and Christison carried out experiments to show that these could not have been caused after death. Robert Christison was born and educated in Edinburgh, where he graduated from the faculty of medicine in 1819. Subsequently he went to Paris for further studies, especially in chemistry, during which time he came into contact with the renowned toxicologist MJB Orfila who inspired the young Scot to specialise in toxicology. His usual method was to inject large doses of poisons into animals and then observe the symptoms and the appearance of the organs after death. He investigated the toxic effects of arsenic, and its detection in the corpse, oxalic acid, laburnum, cyanides, alcohol, lead, opium and hemlock, with the paper that he published in 1836 on the poisonous effects of this last substance and its active component, coniine – which had then recently been isolated – being one of the first pharmacological publications concerning the actions of an alkaloid. In his later life, Christison became interested in cocaine and, when he was 78 years old he carried out experiments upon himself. He must certainly have been an energetic old man since this involved walking fifteen miles at four miles per hour and climbing 2,900 feet up a mountain – he found that the cocaine removed extreme fatigue and made him temporarily less hungry and thirsty, but eventually his appetite returned. However, the best known contribution to toxicology made by this most eminent of the subject’s practitioners of his time involved an earlier experiment on himself with the Calabar bean –

Notes

15

two of which he had obtained from Rev Hope Waddell, a missionary in Old Calabar (see footnote 13) and “subsequently, from him and from a mercantile friend from Liverpool, who annually sends a trading vessel to the Gold Coast, and who kindly interested its captain and its surgeon in the cause” he “obtained successively three small parcels of the seeds” (Christison 1855) – and of which he gave the following account (Christison 1855):It appears that many persons think it an easy task to investigate experimentally the physiology of poisoning. But they are assuredly mistaken. A long apprenticeship must be passed before any one can observe with accuracy the phenomena of the action of poisons. These cautions are prefatory to the remark, that it is a matter of great nicety to apprehend the deceptively simple manifestations of the action of the ordeal-bean on the lower animals. Scarcely do signs of uneasiness appear after a fatal dose has been given, when the animal becomes in quick succession languid, prostrate, flaccid, immovable; respiration, now faint, speedily ceases; and death is complete. It may thus appear to die insensible and comatose. But that is not the case. So long as the power of expression remains, amidst the swiftly advancing languor, signs of sensation may be elicited. Or we might infer from the phenomena that it dies of paralysis of the voluntary and respiratory muscles. But this too is in all probability not the fact. For, on dissection immediately after respiration ceases, the heart is found in a state of paralysis; and it is evident that a quickly increasing paralysis of the heart not only explains the mode of death, but might likewise account for the antecedent muscular weakness and flaccidity. These two effects were well exemplified in the first experiment I tried, when twentyone grains of fine powder, made into an emulsion with two drachms of water, were secured in a cavity in the subcutaneous cellular tissue of the flank of a rabbit. For three minutes there was no appreciable change. But the animal then evidently became weaker, especially in the hind legs. Its feebleness quickly increased, and was attended with slight irregular twitches of the muscles of the trunk and extremities, and occasional twitching of the head backwards. But sensation remained; for the animal struggled a little when held up by the ears, and resisted attempts to shove it from behind. In four minutes, when put upon the side, it lay in that position; which the rabbit always vehemently resists so long as it is able. The trunk and extremities immediately afterwards became quite flaccid. Respiration ceased in five minutes certainly; probably indeed sooner; but the precise time could not be fixed, owing to the continuance of slight muscle twitches. The chest being immediately opened, the heart was seen pulsating slowly, feebly, and inefficiently for ten minutes; and when its cavities where then perforated, the left side gave out a much brighter blood that the right, showing that the circulation, owing to paralysis of the heart, had not been maintained after respiration had ceased. The muscles of voluntary motion contracted at this time vigorously under the stimulus of galvanism, and continued to do so twenty-five minutes after death. The same remarkable properties are possessed by the alcoholic extract of the seeds. When two grains and a third of this extract, obtained from one hundred grains of powdered seeds, were introduced into the cellular tissue of a rabbit in the same way as before, at the end of two minutes, without any previous indication, the animal suddenly became weak, fell on its side, struggled a little with its feet, and ceased to breathe in one minute more. On the chest being immediately laid open, the same phenomena were observed as in the last experiment. It is evident that this poison is one of great intensity of action upon the lower animals; but I have not endeavoured to ascertain exactly its degree of energy. I may mention, however, that on making trial of the exhausted powder from which the extract used in the preceding experiment was prepared, although no effect could be detected in the course of an hour, in ninety minutes the animal was observed to become suddenly weak, and it died in a few minutes more exactly like the others. This result, which appeared unintelligible at first, was afterwards satisfactorily traced to the residual farina not

16

1 Introduction having been carefully enough washed clear of the second spirituous decoction; so that a little of the poisonous ingredient was inadvertently allowed to remain before the farina was dried. The quantity must have been very small. The only other fact I have to mention relative to the action of the seed on the lower animals, is one observed incidentally by Mr Macnab. As the seed vegetates, the two fleshy cotyledons or sarcolobes rise partially above ground. In this state one of the seeds growing in the Botanical garden stove-house was attacked by two slugs, one on each cotyledon. Mr Macnab observing that one of them had begun to swell about the head, he removed it for further observation, and in twenty-four hours it was found dead. Having ascertained the mode of death from the action of the ordeal-bean, I did not consider it advisable to study further the details of its action by means of experiments on animals, because I had been fully informed as to this in a more precise manner by an experiment made with the bean in my own person. I shall conclude this notice with an account of what I experienced; and I trust the details will not appear needlessly minute, as they seem to me to establish an action of a very singular kind in the case of this poison, and one of which we might discover other instances among known poisons, had we equally precise opportunities of determining the true phenomenon. Having some doubts whether I had obtained the true ordeal-poison, as it tasted so like an eatable leguminous seed (see footnote 10 in Chap. 10), I ate one evening the eighth part of a seed, or six grains, about an hour after a very scanty supper. During an hour that I passed in bed reading, I could observe no effects whatsoever, and next morning I could still observe none. I am now satisfied, however, that a certain pleasant feeling of slight numbness in the limbs, like that which precedes the sleep caused by opium or morphia, and which I remarked when awake for a minute twice or thrice during the night, must have been owing to the poison. On getting up in the morning I carefully chewed and swallowed twice as much, viz.., the fourth of a seed, which originally weighed forty-eight grains. A slight giddiness, which occurred in fifteen minutes, was ascribed to the force of the imagination; and I proceeded to take a warm shower bath; which process, with the subsequent scrubbing, might take up five or six minutes more. The giddiness was then very decided, and was attended with the peculiar indescribable torpidity over the whole frame which attends the action of opium and Indian hemp in medicinal doses. Being now quite satisfied that I had got hold of a very energetic poison, I took immediate means for getting quit of it, by swallowing the shaving water I had just been using, by which the stomach was effectually emptied. Nevertheless I presently became so giddy, weak and faint, that I was glad to lie down supine in bed. The faintness continuing great, but without any uneasy feeling, I rung for my son, told him distinctly my state, the cause, and my remedy – that I had no feeling of alarm, but that for his satisfaction he had better send for a medical friend. Dr Simpson [James Young Simpson, the professor of midwifery and the discoverer of the anaesthetic effects of chloroform (Gaddum 1962)], who was the nearest, reached me in a few minutes, within forty minutes after I ate the seed, and found me very prostrate and pale, the heart and pulse extremely feeble and tumultuously irregular; my condition altogether very like that induced by profuse flooding after delivery; but my mental faculties quite entire, and my only sensation that of extreme faintness, not, however, unpleasant. Dr Simpson judged it right to proceed at once for Dr Douglas Maclagan as a toxicological authority, and returned with him in a very few minutes. In his absence, feeling sick, I tried to raise myself on my elbow to vomit, but failed. I made a second more vigorous effort, but scarcely moved. At once it struck me – “this is not debility but volition is inoperative”. In a third effort I was more nearly successful; and in the fourth, a resolute exercise of the will, I did succeed. But I could not vomit. The abdominal muscles acted too feebly; nor were they much aided by a voluntary effort to make them act. I then gave up the attempt, and fell back, comforting myself with the

Notes

17 reflection that vomiting was unnecessary, as the stomach had been thoroughly cleared. At the same time the sickness ceased, and it never returned. There were now slight twitches across the pectoral muscles. I also felt a sluggishness of articulation, and, to avoid any show of this, made a strong effort of the will to speak slowly and firmly, through fear of alarming my son, who was alone with me. Dr Maclagan, on his arrival, thought my state very like the effects of an overdose of aconite. Like Dr Simpson, he found the pulse and action of the heart very feeble, frequent, and most irregular, the countenance very pale, the prostration great, the mental faculties unimpaired, unless perhaps it might be that I felt no alarm where my friends saw some reason for it. I had, in fact, no uneasy feeling of any kind, no pain, no numbness, no prickling, not even any sense of suffering from the great faintness of the heart’s action; and as for alarm, though conscious I had got more than I had counted on. I could also calculate, that, if six grains had no effect, twelve could not be deadly, when the stomach had been so well cleared out. Presently my limbs became chill, with a vague feeling of discomfort. But warmth to the feet relieved this, and a sinapism over the whole abdomen was peculiarly grateful when it began to act. Soon afterwards the pulse improved in volume, but not in regularity. I was now able to turn in bed; and happening to get upon the left side, my attention was, for the first time, directed to the extremely tumultuous action of the heart, which compelled me to turn again on the back, to escape the strange sensation. Two hours after the poison was swallowed, I became drowsy, and slept for two hours more; but the mind was so active all the while, that I was not conscious of having been asleep. On awaking, the tumultuous action of the heart continued. In an hour more, however, I took a cup of strong coffee; after which I speedily felt an undefinable change within me, and on examining the condition of the heart, I found it had become perfectly and permanently regular. For the rest of the forenoon I felt too weak to care to leave my bed; and on getting up, after a tolerable dinner I was so giddy as to be glad to betake myself to the sofa for the evening. Next morning, after a sound sleep, I was quite well.

Christison concluded his paper by noting that one “peculiarity in the action of the ordeal – bean which struck” him “forcibly while labouring under it” was that “there was no bodily uneasiness. . .but simply a sense of sinking vitality, with clearness of mind, and without any sensation deserving in the slightest degree to be called physical distress”. He therefore suggested its use for effecting capital punishment since “Philosophers have thought it not unworthy of inquiry, how in criminal executions death may be completed without physical suffering to the criminal. Governments have even consulted science on the subject. But science has not yet satisfactorily solved the question. Meanwhile, I suspect it has been accidentally solved by the negroes of Old Calabar. At least, so far as the effects of their state-poison on myself went . . .Death by simple fainting, without any preparatory painful process, is evidently what a humane execution should aim at producing. And all this, I apprehend, will be affected by the Calabar ordeal-bean.” (Christison 1855). Christison retired in 1877, at which juncture Thomas Richard Fraser (see footnote 17), “supported by a large number of famous people, including Sydney Ringer” (Gaddum 1962) (see footnote 6 in Chap. 10), was elected to the chair of materia medica, which he occupied for 41 years – he was also professor of clinical medicine in the Edinburgh Royal Infirmary that was built soon after his appointment. When Fraser retired, his department was divided between laboratory science and clinical medicine with Arthur Cushny – then at the University College London, where he held the chair in the Department of Pharmacology where the “lectureship” had been “rendered illustrious by the name of Sydney Ringer” (Gaddum 1962) (see footnote 6 in Chap. 10) and, as a graduate of the University

18

1 Introduction

of Aberdeen who was a young man with Schmiedeberg at Strassburg, was thus the first appointee to the chair who had trained outside of Edinburgh – and JC Meakins becoming the professors of pharmacology [to which subject Cushny was to make an important contribution through his work on the biological activities of optical isomers (Cushny 1926)] and therapeutics, respectively. At the same time, a new department of chemistry in relation to medicine was established, with George Barger as its professor, and with fundamental studies being effected on, for example, the structure of thiamine by Alexander Robertus Todd (see footnote 15). However, Edgar Stedman, initially working with Barger, carried on the tradition that important work on physostigmine was undertaken. Indeed, it was Stedman and Barger who first published (Stedman and Barger 1925) the alkaloid’s chemical structure [with the help of an acknowledged suggestion (Robinson 1925) to them to this effect from Robert Robinson (see footnote 5 in Chap. 2) – then at the University of Manchester (Sect. 2.2)] and, furthermore, showed that it competed with Ach for an enzyme in blood, for which they proposed the name cholinesterase – Stedman also prepared the first synthetic anticholinesterases (Sect. 10.1.1), work in which he was ultimately joined by his wife Ellen (Holmstedt 1972) (Stedman and Stedman 1929, 1931a, b, 1932) – see also (Stedman et al. 1932, 1933), “in one of the first applications of the principle of antimetabolites” (Gaddum 1962)

15. Alexander Robertus Todd (1907 – ) (Melius 1993) In 1938, when he was only 30 years old, Todd resigned from the Department of Medical Chemistry at the University of Edinburgh (see footnote 14) and accepted the invitation to become the Director of the Chemistry Laboratories of the University of Manchester, then regarded as the “first class waiting room” for the Universities of Cambridge and Oxford, and Todd would have been well aware of the opportunities thereby offered. In 1944 he resigned from the chair in Manchester – where he had laid the foundations of nucleotide chemistry that later played a crucial role (Watson 1999; Watson and Crick 1953) in the Nobel Prize-winning determination of the structure of DNA (Franklin and Gosling 1953; Watson and Crick 1953; Wilkins et al. 1953) – to become Professor of Chemistry at the University of Cambridge. 16. John Hutton Balfour (1808 – 1884) (Holmstedt 1972) Biographical data on Balfour is scarce. Of Scottish ancestry and with a number of eminent scientists amongst his ancestors, he was born in Edinburgh and studied at the Universities of St Andrews and Edinburgh. Initially, he had prepared for ordination in the Church of Scotland but gave this up in order to become a botanist and a biologist. After obtaining a medical education on the continent, he began medical practice in 1834 in Edinburgh and in that city 2 years later he was to the fore in establishing the Botanical Society. He began, in 1840, to present extraacademical lectures on botany in Edinburgh and in 1841 he gave up the practice of medicine when he became professor of botany at the University of Glasgow. In 1845 he moved to a similar position at the University of Edinburgh (see footnote 14) and in that year he was also serving as the keeper of the Royal Botanic Gardens and as the Queen's botanist for Scotland. Known to his students as “Old Woody Fibre”, his interests lay mainly in teaching and in the writing of textbooks. Consequently, his original work is somewhat limited - the best known is his “Description of the Plant which produces the Ordeal Bean of Calabar” (Balfour 1861). It is only natural that, with his background, Balfour should have had close contact with Old Calabar's missionaries amongst whom were indeed some of his former pupils (Holmstedt 1972). Indeed, Balfour himself wrote “It is pleasing to observe that all the missionaries at Old Calabar have a taste for natural science. They have already contributed many valuable zoological and botanical species. May they long be spared to carry on their noble evangelizing efforts and their natural history pursuits” (Balfour 1861). The Scottish missionaries reported their findings, including their observations relating to the Calabar bean, in The Missionary Record which began publication in 1846 (Holmstedt 1972), and Calabar beans that Rev Hope Waddell himself brought to Scotland (see footnote 13) were germinated in the botanic garden where Balfour was the keeper in Edinburgh and also in Professor Syme’s garden at Milbank

Notes

19

(Balfour 1861). However, although the plants grew vigorously and produced twining stems and leaves, they never flowered – though 2 years old, but Balfour (1861) observed that their “twigs and foliage were quite identical with native specimens which” he had then “lately received from Africa” and, indeed, from his observations in Edinburgh, and from accounts and specimens sent for Calabar, he was able to give for the plant the first botanical description (Balfour 1861) from which, some 2 years later, his colleague at the School of Medicine in Edinburgh, Thomas Fraser (see footnote 17) was to extract and publish the following details (Fraser 1863):Natur. order – Leguminosae. Sub-order – Papilionaceae. Tribe –Euphaseolae. Genus and Species – Physostigma venenosum. – It has generic characters closely resembling those of Mucua and Phaseolus, but is separated from the former by the characters of the flower and pod, and from the latter by its seed. It has accordingly been placed by Professor Balfour in a separate genus, Physostigma, and is itself the only know species venenosum. Generic Charact. – Root, spreading with numerous fibrils, often having small succulent tubers attached. Inflorescence, axillary; on pendulous multifloral racemes; rachis of each raceme zig-zag and knotty. Calyx, campanulate, four cleft at apex, the upper division being notched and its segments ciliated. Corolla, papilionaceous; veined with a pale pink, having a purplish tinge, and curved in a crescentic manner. Stamens, ten, diadelphous. Pistil, more than one. Stigma, blunt, covered by a remarkable ventricular sac or hood, which extends along the upper part of the convexity of the style, having a resemblance to an “admiral’s hat set in a jaunty manner”. Legume, dark brown and straight, when mature, about seven inches in length, elliptico-oblong, with an apicular curved point, and with outer and inner integument easily separable. Seeds two or three, separated from each other by a woolly substance. [See (Balfour 1861)] Characters of the Seed or Bean The part of the plant of interest on account of any known properties is the seed or bean. Synonymes. – Eséré nut; the bean of the Etu esere; chop nut; the ordeal bean of Calabar. Form. – Irregular reniform, having the appearance of a somewhat flattened fusiform body bent on one of its edges. It has two margins, a shorter or concave and a longer or convex, and two flattened surfaces. Extending along the convex margin is a sulcus, having a minute aperture near one of its extremities. Colour. – As obtained from Calabar, the beans have a grey colour, and are incrusted with earthy matter. This is readily removed by washing and a somewhat shining integument is exposed of various shades of brown, ranging from a light coffee to an almost perfect black. Sulcus. – On the convex edge is the furrow or sulcus already alluded to, with elevated edges, which have, externally, a reddish back hue, while, within, the sulcus is generally brown with a shade of yellow. It extends unequally towards either extremity; at the more extended portion, it runs along a part of the extremity of the bean, and terminates in a narrow furrow; at the shorter end, it has a more rounded termination, and is pierced by a foramen. The bottom of the sulcus is of a grey or reddish black colour, and has two parallel markings extending down its centre. Dimensions. – The average length is 1 and 1-16th of an inch, and varies from 1 inch to 1 and 8-16ths. The average breadth is 12-16ths of an inch, and varies from 10-16ths to 14-16ths. These measurements are the extremes in each direction, and the sides slope from the greatest breadth to the comparatively narrow extremities. The average thickness or breadth from one flattened side to the other is 8-16ths, the maximum 11-16ths, and the minimum 6-16ths of an inch. Weight. – The specific gravity is .946, therefore less than that of spring-water, and we can thus understand how the beans should be conveyed down the rivers to the seacoast. A very few, however, sink in distilled water; out of 300, I have found eighteen such, or

20

1 Introduction 6 per cent. The bean weighs, on an average, 63.263 grains. The greatest weight met with was 94, and the least 25 grains. When the covering was removed from the bean, the embryo was found to weigh, on average 46.2grs., varying from 21 to 73 grs., and the removed spermoderm 16.73grs., varying from 13 to 19 grains. The external tegument is of great hardness and toughness. It is with difficulty cut with a knife, and considerable force is required to break it in a mortar. Its internal surface is of a bluish grey colour. So slight is the absorbing power of this covering, that the bean may be exposed to the action of cold water for a long time without undergoing any change. A bean was carefully measured and weighed, and placed in a covered vessel containing water, in which it was left for 4 months. When examined, no change had been produced in the dimensions or weight of the bean. When exposed to the action of water of a temperature of 212 , or to the action of steam for a few hours, the bean swells by imbibing a quantity of fluid, which may be found in the central cavity, and the spermoderm becomes soft, and can be cut readily into sections with a knife. If the heat be prolonged, the spermoderm cracks and fissures, and the colour of the kernel is changed from a yellowish white to a brownish hue. A fractured portion has a distinct odour of cocoa. When a transverse section of the spermoderm is examined microscopically, the following structures are shown:-

16.1. At the bottom of the sulcated hilum are two bodies separated in the median line, forming the floor of the sulcus, and extending a considerable way up its edges. They are about the 1-157th of an inch in perpendicular thickness at the centre, and taper towards the ends. They consist of an aggregation of rods, each extending through the whole depth of the structure, and terminating at both ends in thickened extremities. 16.2. Outer layer of the spermoderm. – It is very similar in structure to the above, and forms the external envelope of the spermoderm. It extends over the entire surface of the bean, except at one end of the floor of the sulcus, where the opening of the foramen occurs. It is between the 1-138th and 1-117th of an inch in perpendicular thickness, and consists of a number of rods placed side by side, each rod being about the 1-2250th of an inch in thickness, and terminated by broad extremities. 16.3. Internal to this is the middle and principal layer of the spermoderm. It varies in thickness at different places from the 1-8th to the 1-130th of an inch, having its smallest measurement at the narrow convex edge of the bean, and it’s greatest at the hilum,. Its structure is cellular, consisting of stellate cells, having six or eight branches which communicate with branches from the neighbouring cells. The cells diminish in size and in the length of their branches as they approach the exterior, and here they appear to form a separate membrane. This is in reality a compressed collection of cells, which extends over the exterior of the middle coat of the spermoderm to a fibro-vascular body to be presently described, and from this to the internal surface of this middle layer. 16.4. Inner Coat of the Spermoderm. – It consists of dark ligneous tissue forming a continuous layer immediately below the former, and varies from the 1-470th to the 1-20th of an inch in thickness.

Notes

21

16.5. A Fibro-vascular structure embedded in the middle layer, immediately below and extending along the whole length of the sulcus. In a transverse section it is seen to possess an elongated ovoid form, and from its lighter colour it is apparent to the naked eye. It has a perpendicular diameter of about the 1-65th of an inch, and a transverse diameter of about the 1-120th. In a transverse section, it appears to contain a number of irregularly oval cells, with transverse markings, having their long axes in the same direction as that of the containing structure. 16.6. Kernel or Embryo.- It consists of two large concavo-convex cotyledons, of a creamy white colour, and easily broken in a mortar or scraped with a knife. In a transverse section these are seen to be in close contact externally with the spermoderm, and internally, to be quite separated from each other, except at the margins of the bean. A large cavity is thus left in the centre, which communicates with the external atmosphere by means of a minute foramen so small and so narrow as to give no opportunity for the escape of the contained air when the bean is immersed in water. Projecting into this cavity, immediately below the broader extremity of the hilum, may be found the plumule with its two radicles. Microscopically, the kernel consists of a cellular texture, with cavities of hexagonal, and often of a very irregular one-sided form. These cells vary from the 1-650th to the 1-140th of an inch in diameter. They contain from one to six starch corpuscles, which are readily detached by washing, and give the usual reactions with iodine, bromine, and boiling water. The general form of these starch granules is an elongated oval, frequently approaching an irregular reniform shape or rounded parallelogram. They have usually a regular margin, but this is often indented. The surface has, in the majority of cases, a central line in the long axis, surrounded by concentric rings, but frequently the central line is superseded by a dark space containing amorphous granules. Occasionally, dark radial lines extend, from the central line or space, more or less completely to the circumference, presenting an appearance similar to radiated cracks in a transparent sphere. The average length of the starch granule is 1-440th of an inch, varying from the 1-700th to the 1-300th. The average breadth was found to be 1-625th of an inch, and varied from the 1-920th to the 1-400th. The bean has been always received remarkably free from all disease, - only one form of abnormality, scarcely deserving the name of disease, having been found. This occurs between the spermoderm and kernel, and affects the inner surface of the former and the outer of the latter. It consists, on the kernel, of a circular, somewhat dark, indented space, varying in diameter from the 1-6th to the 1-8th of an inch, and having a central irregular depression from which a number of faint lines diverge towards the circumference. The kernel is found much softer than usual within this area. On the inner surface of the spermoderm a corresponding space is found, of a brownish colour, and distinct from the usual bluish grey.

22

1 Introduction

This appearance is probably due to the attack of an insect, although our research has never succeeded in finding one 17. Thomas Richard Fraser (1841–1920) (Gaddum 1962; Holmstedt 1972; JTC 1921) Born in India, Fraser obtained his early education at private schools prior to reading medicine at the University of Edinburgh, where he came under the influence of the medical faculty’s virtual galaxy of talent which included Robert Christison (see footnote 14), a pharmacologist before his time who lived and worked before the analytical methods of his subject had been satisfactorily established. Thus, although his description of his own symptoms, both subjective and objective, occasioned by the Calabar bean is extraordinarily exact and instructive (see footnote 14), it was not to be at his hands that the elucidation of these effects was to be accomplished, but by the subsequent work of TR Fraser who, working as one of Christison’s students, subjected the poison which had caused the effects to extensive pharmacological research and who, when he was 21 years old, discovered that an extract of the bean acted on many different organs, such as pupils, the central nervous system, the heart, the glands, the voluntary muscles, and the intestines, and ultimately – in 1862 – presented to the University of Edinburgh a MD thesis (Fraser 1862) – that in common with many such works was awarded a gold medal, which appear to flow copiously in the medical disciplines – entitled “On the Characters, Actions and Therapeutical Uses of the Ordeal Bean of Calabar”. Christison was not slow in recognising promise in his former student and, in the year subsequent to the appearance of the thesis, appointed him as his Assistant in the Department of Materia Medica at the University of Edinburgh, it being during the tenure of this position that Fraser made further advances in investigating the action of drugs, especially the Calabar bean. Working largely with the extract containing its alkaloidal component, physostigmine, in his early work he came to the following conclusions (Fraser 1863) :The two varieties of symptoms following the administration of the bean may be harmonized in the same way. It exerts a special influence on the spinal cord; when this is limited in extent and energy, the only marked effect is paralysis, and death is caused by the extension of this paralysis to the muscles of respiration, causing death by asphyxia. When, on the other hand, this spinal action is more extensive and energetic, the heart is affected, its contractions cease, and death occurs by syncope. We may therefore conclude, that the kernel or embryo of the bean of Physostigma venenosum has the following actions:-

17.1. It acts on the spinal cord by destroying its power of conducting impressions. 17.2. This destruction may result in two well-marked and distinct effects, – (a) In muscular paralysis, extending gradually to the respiratory apparatus and producing death by asphyxia. (b) In a rapid paralysis of the heart, probably due to the extension of this action to the sympathetic system thus causing death by syncope. 17.3. A difference in dose accompanies this difference in effect. 17.4. This action does not extend to the brain proper pari passu with the action on the spinal cord. The functions of the brain may, however, be influenced secondarily. 17.5. It also produces paralysis of muscular fibre, striped and unstriped. 17.6. It acts as an excitant of the secretory system, increasing more especially the action of the alimentary mucous glands.

Notes

23

17.7. Topical effects follow the local application of various preparations; these are – destruction of the contractility of muscular fibre, when applied to the muscles, and contraction of the pupil, when applied to the eyeball. These above observations would appear to invalidate the claim (Neuwinger 1996) that “Schweder (1889) clearly worked out for the first time all physiological effects of physostigma” – once again workers in Edinburgh had been the pioneers in early studies upon Physostigma venenosum and its toxic bean × and, furthermore, as early as 1863, TR Fraser alerted his friend and colleague, Douglas Argyll Robertson,to the use of the Calabar bean in ophthalmic surgery (Robertson 1863; Witkop 1998) (Sect. 10.5). There was a further development from TR Fraser’s research in that the symptoms produced by the Calabar bean and the cause of their occurrence suggested the probability that atropine, which antagonises the action of physostigmine on the eye, might also act as an antidote with regard to the more serious effects that the former causes. The antagonism between physostigmine and atropine, itself a new concept at that time, was quantitatively investigated by TR Fraser by injecting both alkaloids subcutaneously into rabbits. It was found that after a given dose of physostigmine there is a range of doses of atropine that will save life. If too little is given, the rabbit dies of physostigmine poisoning and, if too much is given it dies of atropine poisoning. As the dose of physostigmine is increased, this range become less and, if the dose of physostigmine is more than 3.5 lethal doses, life cannot be saved by any dose of atropine. TR Fraser thus became the discoverer of “The antagonism between the actions of active substances” (Witkop 1998) and expressed the results of these studies graphically, with atropine doses on the horizontal scale and those of physostigmine on a vertical scale, and drew a line which separated combinations of doses that led to death from those that led to recovery. About half a century was to elapse before Dr S Loewe reintroduced graphs such as these and called them isobols (Holmstedt 1972). TR Fraser’s studies showed that small doses of atropine saved life by opposing the lethal action of physostigmine, but evidence was also obtained which showed that small doses of physostigmine causes death by increasing the lethal action of atropine, an affect that TR Fraser had expected and which his experiments were designed to detect. He pointed out that, since both atropine and physostigmine possess a number of separate actions, it is not unreasonable to anticipate that several of them are not mutually antagonistic – in some cases the two drugs might be expected to work together. It was in 1872 that Fraser presented these results to the Royal College of Physicians of Edinburgh in one of two lectures – which was also published (Fraser 1872), and prior to this he effected investigations which established that the lethal effect of l-physostigmine can be averted by atropine (Fraser 1870). Shortly afterwards TR Fraser was to leave academic life to accept appointment as Medical Officer of Health for Mid-Cheshire. However, some 3 years later he returned to Edinburgh, initially as assistant and subsequently as physician and one of the professors of clinical medicine and then, in 1877, he was also elected to the chair of materia medica that had become vacant upon the retirement of Robert Christison (see footnote 14) and, despite his then clinical responsibilities, he still managed to pursue pharmacological research and remain active in the laboratory. He studied the actions of an arrow poison which, like the Calabar bean – to which he was never to return, came from Africa. He isolated the active principle – strophanthin – and introduced it into clinical use as a substitute for digitalis. He also effected an intensive study of snake venoms and produced immunity, not only by repeated injections but also by oral administration.

24

1 Introduction

On leave-of-absence from his duties in Edinburgh, Fraser was to return to the country of his birth as a member and the president of the Commission that was appointed by the Governor- General, with the approval of H.M. Secretary of State for India in consequence of a severe outbreak of a plague in that country in 1898. In addition to his service upon this Commission, which issued a voluminous report in 1901, Fraser was, at various times, called upon to occupy other important positions. Thus, for example, in 1881 he acted as the president of the Section of Materia Medica and Pharmacology at the International Medical Congress in London, he occupied the Presidential Chair of the College of Physicians of Edinburgh, he was appointed by the Admiralty as a member of the Committee charged with an enquiry into an outbreak of scurvy that occurred in Sir G Napes’ Arctic Expedition, he acted as Consulting Medical Officer to the Prison Commissioners for Scotland, he was appointed Honorary Physician in Scotland to HM the King, and he served his University in various capacities – discharging the onerous duties as Dean of Faculty of Medicine for 20 years. Not surprisingly, he was the recipient of many honours that were bestowed upon him by learned bodies both at home and abroad. Honorary degrees were forthcoming from Dublin (MD), Cambridge (BSc) Aberdeen, Glasgow and Edinburgh (all LLD), in 1877 he was elected a Fellow of the Royal Society and also of the Royal Society of Edinburgh, and he was laureated by the Institute of France, by the Turin Academy of Medicine, and by the College of Physicians of Philadelphia, and in 1902 a Knighthood was conferred upon him. Unfortunately, for many years before the end of his active life he suffered from a bronchial condition that was sometimes incapacitating and, when 70 years old, he had the further misfortune of experiencing a fracture of the femur which, though partial use of the injured limb was eventually recovered, left him with restricted movements, often attended by discomfort. In the end, Sir Thomas Fraser, the pharmacologist, became a legend associated with a frail old man with bronchitis, carried onward by an indomitable spirit. He died at his residence in Edinburgh on 4th January 1920, only 16 months after he had relinquished his Professorship of Materia Medica at that city’s university.

Chapter 2

l-Physostigmine (Eserine)

2.1

Natural Occurrence

In his report (Christison 1855) to the Royal Society of Edinburgh on 5th February 1855 of his pioneering studies upon the toxicology of the Calabar bean by selfexperimentation, Robert Christison (see footnote 14 in Chap. 1) stated “that its active properties may be concentrated in an alcoholic extract, which constitutes 2.7 per-cent of the seed”. Although experimental details are not presented (Christison 1855), this deficiency was rectified some 8 years later by Thomas Fraser (see footnote 17 in Chap. 1), formerly Christison’s assistant at the University of Edinburgh (see footnote 14 in Chap. 1), who published the following method for the preparation of the tincture that he used in connection with his investigation into the therapeutic properties of the Calabar bean (Fraser 1863):Take of the kernel, in the form of fine powder, j1; rectified spirit, ij 2. Place the kernel and one ounce of the spirit in a carefully covered vessel, and allow to remain for forty-eight hours. Pack in a procolator [sic], pour in what spirit may be left in the vessel, and add the remaining ounce of spirit. When this has ceased to escape from the procolator [sic], pass as much more spirit through as may be required to obtain two ounces of a golden yellow tincture. This preparation is so far objectionable, that the kernel is not exhausted by the quantity of spirit used; yet it appears preferable to one obtained by reducing to a certain standard, by distillation, a tincture obtained with a much larger proportion of spirit. I have found five minims of this tincture a good dose with which to commence the administration. This appears to possess the activity of three grains of the kernel, as far as can be judged by the effects produced. The dose may be trebled without pushing the physiological action to any extreme. The kernel can only be exhausted by employing a much larger proportion of spirit. By using twelve ounces of rectified spirit with one ounce of powdered kernel, distilling off about eight ounces, and evaporating the remainder, first to a syrupy consistence, in a vapour-bath, and then by spontaneous evaporation, twenty-one grains of an extract of considerable consistence may be obtained, or a proportion of 4.375 per cent. This extract has a deep brown colour, and a peculiar, sweetish, and disagreeable odour, for which I can find no comparison. Its actions differ only in intensity from those of the kernel and tincture. © Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_2

25

26

2 l-Physostigmine (Eserine)

It is apparent from these experimental data and the now known total alkaloidal content (see footnote 1 in Preface and Acknowledgements) and non-alkaloidal components (Chap. 9) of the Calabar bean that far more than l-physostigmine had been extracted and was thereby present in the tincture. Indeed, Fraser stated that “A preparation of unvarying strength will only be obtained when the active principle of Physostigma is discovered” (Fraser 1863). Consequently, the later statements that “the active principle physostigmine was isolated by Fraser in 1863” (Rodin 1947), that he was “the first to isolate the active principle physostigmine” (Rodin 1947), that “Some time later he obtained a purer, crystalline form of the alkaloid” (Holmstedt 1972, Rodin 1947), that “he purified the alkaloid” (Gaddum 1954), that he “separated an amorphous active principle with the properties of a vegetable alkaloid” (Holmstedt 1972), that “An alkaloid of the seeds in the form of an amorphous substance was first isolated in 1863 by Fraser in Edinburgh” (Neuwinger 1996) and that (though, unfortunately, without reference to a primary literature source) “later he obtained a purer crystalline form of the alkaloid” (Holmstedt 1972) would all appear to be either misleading or erroneous, since Fraser reported no further (Fraser 1863) as to either the actual isolation of the active components from his extract or its nature. In connection with the latter, Christison had already noted (Christison 1855) – surprisingly in view of then current knowledge – that his (Christison’s) “extract does not yield a vegetable alkaloid by the more simple of the ordinary methods of analysis” although, once again, experimental details were omitted (Christison 1855). Notwithstanding these above either somewhat exaggerated or erroneous claims, in 1864 alkaloidal material was first reported as being isolable from and responsible for the poisonous activity of the ripe seeds of Physostigma venenosum and was named physostigmine after this botanical source (Hesse 1867; Jobst and Hesse 1864) [from which, on laboratory (Salway 1911) and industrial (Chemnitius 1927; Schwyzer 1927) scales – an adulterant is also known (Watt and Breyer-Brandwijk 1962) that “has on occasion been substituted in commerce for the seed of Physostigma venenosum” – it was later extracted and derivatised as its salicylate (Muhtadi and El-Hawary 1989) which, it has been claimed (Lauter and Foote 1955) “appears to be especially stable toward oxidation” (see Sect. 8.2)]. However, at the juncture of its first isolation, the allegedly pure alkaloid was only obtained in an amorphous, varnish-like state (Hesse 1867; Jobst and Hesse 1864) although the non-existence of its quoted reference may be considered to throw doubt upon the statement (Salway 1911) that “Jobst and Hesse, in a later investigation (Annalen, 1867, 141, 913), assigned to physostigmine the formula C15H21O2N3 [which was ultimately (Petit and Polonovsky [sic] 1893; Straus 1913) confirmed], but still expressed doubt regarding its crystalline character”. However, be-this-as-it-may, the alkaloid was reported in a crystalline state by others (Vée and Leven 1864; 1865) when it was found to have mp 69
C (Vée and Leven 1865) [which was later (Henry 1949, Petit and Polonovsky [sic] 1893, Robinson 1964a, Salway 1911) corrected to 105–106
C. Although this difference between mps was initially suggested (Straus 1913) as being caused by an absorption of either carbon dioxide or water, the alkaloid ultimately was found to be dimorphous with an isomorph of mp either

2.1 Natural Occurrence

27

86–87
C (Henry 1949, Salway 1911) or 85–86
C (Robinson 1964a) being converted into the more stable (Henry 1949) modification of mp 105–106
C by recrystallisation in the presence of a crystal of the latter (Salway 1911) or upon storage at room temperature for 6 months (Robinson 1964c)]. At the earlier juncture (Vée and Leven 1864, 1865) it was also given the name eserin, after the ésere poison ordeal for which the ripe seeds were used by the Efik people of Old Calabar in Nigeria (vide supra) (Chap. 1) and also later accepted as being after the native name for its plant source (Chap. 1). However, since it has chronological priority, l-physostigmine is the name by which the alkaloid is now officially recognised although both names are still used to designate the base [however, see foot-note to (Vée and Leven 1865)]. It has also been referred to as Synapton [Merck Index 2001(h)] – probably because of its role in the discovery of how nerve impulses are transmitted across nerve-nerve and neuromuscular synapses (Sect. 10.3), as Cogmine (Merck Index 1996) – probably because of its role in the enhancement of cognition and memory in DAT (Sect. 10.7.2) and as Physostol (Merck Index 1989). Although the statement (Fraser 1863) that “No part of the physostigma venenosum is known to possess active properties except the seed or bean” may suggest that the natural occurrence of l-physostigmine is confined accordingly, sources other than the ripe seeds of this plant have also been reported for l-physostigmine. Thus, the alkaloid also occurs in P. cylindrospermum Holmes (seeds) (Boit 1961), P. mesoponticum (seeds) (Tanganyika Rep Govt Chem 1952), Vicia calabarica (Orekhov 1955) and Mucuna urens Medic and M. cylindrosperma Welw. Ex Baker (Marques 1957) – all five of which belong to the family Leguminosae – and in various Dioclea species (Freise 1936a) [all of which, being climbers and bearing fruit in long pods (Freise 1936a) are legumes], notably D. microcarpa Huber (Freise 1936a, b) but also allegedly in D. bicolor (Freise 1936b), D. Lasiocarpa, (Freise 1936b), D. reflexa (Freise 1936b) and D. violacea (Freise 1936b). It has also allegedly, but without unequivocal characterisation, been reported in the fresh fruit of Hippomane mancinella (family Euphorbiaceae) (Lauter and Foote 1955). This, known as the manchineel tree, is native to central America and the West Indies and ranks among the most famous of poisonous plants in tropical America (Adolf and Hecker 1975) and in a later investigation (Adolf and Hecker 1975) of its toxic components it is perhaps significant that l-physostigmine was not reported. l-Physostigmine has also been obtained from microbial sources, namely Streptomyces pseudogriseolus (Iwasa et al. 1981), S. pseudogriseolus FERM-P 4084 (Iwasa et al. 1979) and subsp. iriomotensis subsp. nov. (Iwasa et al. 1981, Murao and Hayashi 1986) and S. sp. AH-4 (Murao and Hayashi 1986), observations that could, if necessary, open the way to its production by fermentation (Robinson 2002). This would be a useful alternative to its extraction from the ripe seeds of P. venenosum from which, although it is the major alkaloidal component (Sect. 1 and Chap. 8), it nevertheless appears to be present in only somewhat meagre amounts, being extractable when working on a large scale [for example, using 122.7 kg of Calabar beans (Salway 1911)] only to the equivalent of 0.179% of the material employed (Salway 1911) (see also see footnote 1 in Preface and Acknowledgements). However, it has been suggested that “it is possible that the content

28

2 l-Physostigmine (Eserine)

of the alkaloid might [or could] be increased by selective cultivation” (Dalziel 1948; Irvine 1961) although it was recognised that cultivation in [what was then known as] the British territories was not permitted (Dalziel 1948, Irvine 1961). Indeed, the ripe seeds of P. venenosum may be of limited availability since the parent plant, though distributed from Sierra Leone (Dalziel 1948; Irvine 1961) [the main source of the seeds as formerly used in pharmacy (Irvine 1961) though in 1890 they were collected and sold in Ghana for this purpose (Irvine 1961) and in 1948 they were on sale in the Enugu medicine market (Irvine 1961)] [and perhaps French Guinea (Dalziel 1948)] to Cameroon and B. Congo (Dalziel 1948; Irvine 1961) and to Zaire (Neuwinger 1996) – and also introduced into India and Brazil [Merck Index 2001(g)] – would seem to be rather rare and, in spite of being so widespread,3 of erratic local occurrence (Dalziel 1948) since the former British colonial government prohibited its cultivation and exterminated as far as was possible existing species in its efforts to eliminate the use of the Calabar bean as an ordeal poison. Notwithstanding, the ready availability of l-physostigmine in good yields by facile synthetic procedures, particularly that of Julian and Pikl with its recent improvements (Sect. 3.1)16, 17, 20, 21, 22, 24 would appear to render extraction of the alkaloid – should it still be required – from natural sources, be they either botanical or microbial, redundant. It has been stated (Murao and Hayashi 1986) that [besides l-physostigmine and l-N(8)-norphysostigmine] “there has been only one report of a plant-derived alkaloid being obtained from a microorganism. Namely maytanacine, which had been isolated from Putterlickia verrucosa (Celastraceae), was obtained from fermented broth of Nocardia sp.” Although maytanacine (1) is, indeed, a metabolite both of the plant O Cl MeO

Me O N

O

C Me O

Me

Me O Me

MeO

OH

N H

O

1 P. verrucosa (Kupchan et al. 1975) [but not of the plant Colubrina texensis (Wani et al. 1973) as may have been implied (Higashide et al. 1977)] and also of a fermentation broth of Nocardia sp. No.C-15003(N-1) (Higashide et al. 1977), and does contain two nitrogen atoms, neither of these are basic. Therefore, the compound is not alkaloidal.4 This thereby invalidates the above statement (Murao and Hayashi 1986) and thus leaves l-physostigmine (vide supra) – and l-N(8)norphysostigmine (Chap. 5) which has also been isolated from S. sp. AH-4 (Murao and Hayashi 1986), as the first, and as yet only, examples – since they are plant-derived bases – of alkaloids that have also been obtained from a microbial source.

2.2 Structure Elucidation

2.2

29

Structure Elucidation

In a review (Barger 1936) of the pioneering research on l-physostigmine which included references to the studies that led to the elucidation of the chemical structure of the alkaloid, it was stated that this latter problem was “chiefly investigated by Max Polonovski and his collaborators [the predominant of whom was his brother, Michel, who joined him in the later years of the endeavour], further by Ehrenberg, Salway, Straus, Barger and Stedman”. However, perhaps surprisingly and especially in view of the review’s author, it was not pointed out that the latter duo was working in the Department of Medical Chemistry at the University of Edinburgh (see footnote 14 in Chap. 1), thereby continuing the association between this university and research into the Calabar bean and its major component alkaloid, and, furthermore, overlooked Robert Robinson’s5 significant contributions to the resolution of this problem. Many blind alleys and false trails, a comprehensive elaboration of which would be of interest but which have so far received only cursory attention as parts of reviews (Henry 1949; Marion 1952) [however, see also (Sumpter and Miller 1954a)] were followed in the quest for the solution to this problem which was rendered particularly difficult by the absence of structural analogues in other natural products of then known composition. Indeed, l-physostigmine stood out as the only known [Spande 1979(c), Wieland et al. 1934] naturally-occurring derivative of 5-hydroxyindole until bufotenine – isolated from the parotid glands of the toad Bufo vulgaris – was shown (Wieland et al. 1934) to be 3-(2-dimethylaminoethyl)-5hydroxyindole (2), a postulation verified by synthesis in the following year (Hoshino (CH2)2NMe2

HO N H

2 and Shimodaira 1935, 1936, Hoshino et al. 1935). However, encouraged by the valuable therapeutic properties that it was found to possess (Chap. 10) and following the work by the various above-mentioned investigators which has been the subject of review on several occasions (Bentley 1957; Brossi et al. 1996; Cordell 1981; Coxworth 1965; Henry 1949, Marion 1952; Robinson 1964b; Saxton 1960; Harley-Mason and Jackson 1954a; Taylor 1966), the alkaloid’s structure was eventually established (Robinson 1925; Stedman and Barger 1925) as 3 (R1¼MeNHCO, R2¼Me, X¼NMe), with the positional numbering of the tricyclic 1,2,3,3a,8,8a– hexahydropyrrolo[2,3-b]indole system being adopted later. Me

4

R1O

3b 5

A

6

7a 7

3a

N8 R2

3

3

C

B 8a

X1

2

30

2 l-Physostigmine (Eserine)

Early investigations of l-physostigmine found (Petit and Polonovsky [sic] 1893) its optical rotation “en solution alcoolique” to be “a(D) ¼ 74
,5”{and since established as either [α]D  75.8
(CHCl3) and 120
(C6H6) (Henry 1949; 25

Robinson 1964a) or ½α17 D -76 (CHCl3) and ½αD – 120 (C6H6) [Merck Index 2001(h), Muhtadi and El-Hawary 1989]}, confirmed that it had the composition C15H21N3O2 (Petit and Polonovsky [sic] 1893; Straus 1913), contained three N-methyl groups6 and behaved as a monoacidic tertiary base (Ehrenberg 1893; Petit and Polonovsky [sic] 1893). In addition, the action of either “Alkali” (Ehrenberg 1893) or sodium hydroxide (Salway 1912b) on the alkaloid in the absence of air was found to cause the elimination of methylamine and carbon dioxide and afforded a new base, C13H18ON2, that was named l-eseroline (Ehrenberg 1893, Salway 1912b)7 with these changes being represented (Salway 1912b – see also Ehrenberg 1893) by the equation:-. ðC13 H16 ONÞNH  CO  NHMe þ H2 O ¼ C13 H16 ON  NH2 þ CO2 þ NH2 Me in which l-eseroline was thus regarded (Ehrenberg 1893) as an amino derivative of a structure having a molecular formula C13H16ON. However, equally with this, it could be assumed that the alkaloid is a urethane of the formula NHMeCOOC13H17N2, in which case l-eseroline would be an alcoholic compound C13H17N2OH (Salway 1912b). In fact l-eseroline was ultimately found to be a phenolic monoacidic tertiary base (Polonovski 1916) that contains two N-methyl groups (Straus, 1913, 1914 – see, however, Herzig and Lieb 1918a, b) although an earlier determination, using the methods of Herzig and Meyer, had led to the conclusion that it only “contains one methyl-imide group, NMe” (Salway 1912b)6 and, moreover, the unusual presence of a urethane – (or carbamate) – grouping (or, more specifically, an N-methylcarbamyloxy group) was further confirmed by the observation that heating the alkaloid with sodium ethylate in the absence of air afforded l-eseroline and methyl urethane (Polonovski 1915), its oxidation with potassium permanganate gave methylisocyanate (Me-N¼C¼O) (Polonovski 1915) and pyrolysis at its melting-point caused the evolution of methylisocyanate (Polonovski 1915) “which [it was later found] can be trapped with butanol as butyl N-methylcarbamate” (Brossi et al. 1996; Yu et al. 1987). That the carbamate group in the alkaloid is linked to the phenolic hydroxyl group that is liberated during the formation of l-eseroline is evident since the latter, upon reaction in dry ethereal solution with methylisocyanate in benzene in the presence of a trace of sodium was reconverted into l-physostigmine (Polonovski and Nitzberg 1916) – it was also similarly converted in benzene solution into phenserine (Polonovski 1916) [a compound which some eight decades later (Brossi et al. 1996; Greig et al. 1995a, 2005a; Klein 2007) was to be “a drug candidate of potential use for treating Alzheimer’s disease” (Sect. 10.7.2)] by treatment with phenylisocyanate in ethereal solution in the presence of a trace of sodium (Polonovski 1916). Thus the problem of the alkaloid’s structure was reduced to elucidating that of l-eseroline. The presence of an indole-derived nucleus in l-physostigmine was indicated by its distillation from admixture with zinc dust (Petit and Polonovsky [sic] 1893) in a current of hydrogen which afforded a combined yield of between 5% and 10% of

2.2 Structure Elucidation

31

indolic products claimed to consist of 2-methylindole together with a small quantity (15%) of 1-methylindole (Salway 1912b), although neither of these products were isolated in the pure state and properly characterised (Salway 1912b). A degradative reaction that was of major importance in determining the structure of l-eseroline, and hence that of l-physostigmine, involved heating esermathele methiodide at 200
C in an atmosphere of carbon dioxide. This, as might now be expected,8 led to the facile fragmentation of the rigid tricyclic ring system and to the loss of the elements of C3H7N.MeI to afford physostigmol (C10H11NO), a product that exhibited the typical colour reactions of an indole and contained an N-methyl group and a phenolic hydroxyl function and was thereby suggested as being a hydroxy-1,3-dimethylindole (Straus 1914). Ultimately it was shown to be 5-hydroxy-1,3-dimethylindole (4, R1¼R3¼R4¼H, R2¼Me) by the synthesis of CH2R4

R1O N

R3

R2

4

its ethyl ether 4 (R1¼Et, R2¼Me, R3¼R4¼H) via the pyrolytic decarboxylation (at 250
C) of 2-carboxy-3-carboxymethyl-5-ethoxyindole (4, R1¼Et, R2¼Me, R3¼R4¼COOH) that was synthesised by Fischer indolisation – in warm 50% acetic acid – of the arylhydrazone 5 (R1¼Et, R2¼Me, R3¼R4¼COOH) that was formed in (CH2)2R4

R1O

C N N

R3

R2

5 situ by condensation of 4-ethoxyphenyl-Nα-methylhydrazine with α-ketoglutaric acid (Stedman 1924). Another subsequent synthesis (Keimatsu and Sagasawa 1928) began by reacting diazotised 4-ethoxyaniline with ethyl-α-ethylacetoacetate in ethanolic sodium hydroxide solution to afford 5 (R1¼Et, R2¼R4¼H, R3¼COOEt) that was then subjected to Fischer indolisation in 10% ethanolic sulphuric acid boiling under reflux to yield 4 (R1¼Et, R2¼R4¼H, R3¼COOEt). This, upon sequential saponification and pyrolytic decarboxylation (at 200
C), then gave 4 (R1¼Et, R2-R4¼H) which afforded physostigmol ethyl ether (4, R1¼Et, R2¼Me, R3¼R4¼H) after N1-methylation using sodium and iodomethane. A direct synthesis of physostigmol methyl ether (4, R1¼R2¼Me, R3¼R4¼H) has also been effected (Späth and Brunner 1925) by Fischer indolisation – by heating with anhydrous zinc chloride – of 5 (R1-R2¼Me, R3¼R4¼H), the 4-methyloxyphenyl-Nα-methylhydrazone of propionaldehyde which, surprisingly, was not the carbonyl moiety of choice in the thereby earlier (Stedman 1924) (vide supra) but indirect synthesis of physostigmol ethyl ether.

32

2 l-Physostigmine (Eserine)

The position of the hydroxyl group in l-eseroline, and hence that of the N-methylcarbamyloxy group in l-physostigmine, was established (vide supra) by the structure of physostigmol ethyl ether as (4, R1¼Et, R2¼Me, R3¼R4¼H). This latter product was also formed directly – and significantly in high (66%) yield – by sublimation from a pyrolysis at from 180
C to 220
C under high vacuum (Stedman 1924)8 of l-eseroline ethyl ether (eserethole) methiodide that is prepared by treatment of l-physostigmine with ethanolic sodium ethoxide and ethyl ptoluenesulphonate (Polonovski 1915) followed by quaternisation of the product with iodomethane (Polonovski 1915). The high yield of product resulting from this thermal decomposition and the ease with which physostigmol (4, R1¼R3¼R4¼H, R2¼Me) was formed from l-eseroline methiodide led to the suggestion (Stedman and Barger 1925) that the 3-methyl group of physostigmol is also present as a methyl group in l-eseroline, and thereby in l-physostigmine. Thus it appeared justifiable to assume that the grouping 6 is present in l-eseroline (Stedman Me

HO N

Me

6 and Barger 1925) and it then remained to determine “the manner in which the remainder of the molecule (C3H8N) is linked to this grouping” (Stedman and Barger 1925). That this fragment must be joined to the indoline moiety 6 as –(CH2)2-N (Me)- was apparent since it was known that both l-eseroline and l-physostigmine contain two tertiary basic N-methyl groups (Straus, 1913, 1914 – see, however, Herzig and Lieb 1918a, b).6 These above criteria invalidated as possibilities for l-eseroline structure 7 that had H HO N

H Me

N Me

7 been proposed by Straus (1913), a somewhat bizarre “hypothetical formula which would serve as a basis for some synthetical experiments” proposed by Salway (1913) and several further bizarre structures (Polonovski and Polonovski 1923a, b, c) together with – based mainly on a study of the reduction and the exhaustive methylation of eserethole and, whilst correctly suggesting the presence of a Ph-N-C-N system, is incompatable with a “potential biogenesis” from tryptophan (vide infra) – structure 8 (Polonovski and Polonovski 1924a, b, c, d, e, 1925a, b). However, they are in

2.2 Structure Elucidation

33

H HO N N H Me Me

8 accord with structure 3 (R1¼MeNHCO, R2¼Me, X¼NMe) for l-physostigmine9 that was suggested by Prof Robert Robinson, FRS [then at the University of Manchester5] (Robinson 1925) to and published by Edgar Stedman and George Barger (see footnote 14 in Chap. 1) from the Department of Medical Chemistry at the University of Edinburgh (Stedman and Barger 1925), and also – albeit somewhat tentatively – by the Polonovski brothers (Polonovski and Polonovski 1925a, b). The former group also offered further support for this structure when they recognised that in connection with “the possible mechanism of the phytochemical synthesis of the ring system present in physostigmine” “Perkin and Robinson (1919, 115, 944) [Perkin jun and Robinson 1919] have shown how harmine may be elaborated in the plant from tryptophan.” And that “If the assumption is made that the methylation of an indole nucleus may proceed in the plant in the manner in which it is known to take place in the laboratory, a relation between physostigmine and this amino-acid at once becomes evident. By decarboxylation and methylation, followed by a ring closure, the ring system of physostigmine would be readily formed.” (Robinson 1925, Stedman and Barger 1925) (see also Sect. 5). This very simple example of “potential biogenesis”, considerations of which were to find wide use in the pioneering investigations (Robinson 1955) into alkaloid structural elucidations, provided no support whatsoever for the structures for l-physostigmine that would have been derived from the other proposals (Polonovski and Polonovski 1923a, b, c, 1924a, b, c, d, e, 1925a, b; Salway 1913; Straus 1913) for the structure of l-eseroline. Structure 3 (R1¼MeNHCO, R2¼Me, X¼NMe) for l-physostigmine was also recognised (Stedman and Barger 1925) as being compatible with the observation that reduction [with zinc and hydrochloric acid (Polonovski 1918, Polonovski and Polonovski 1924e – see also 1924a) or by catalytic hydrogenation (Stedman and Barger 1925)] of eserethole (l-eseroline ethyl ether) effected cleavage of a heterocyclic ring with the addition of two hydrogen atoms and the formation of a secondary base for which structure 9 (R1¼EtO, R2¼R3¼Me) was suggested (Stedman and Barger 1925). R2

R1 N

Me

9

NHR3

34

2 l-Physostigmine (Eserine)

Final unequivocal verification of structure 3 (R1¼MeNHCO, R2¼Me, X¼NMe) as that of l-physostigmine was obtained (Robinson and Suginome 1932a) via the formation of eseroline methyl ether (esermethole) (3, R1¼R2¼Me, X¼NMe), the methiodide of which (10) was sequentially subjected to ring opening to afford 11 (R1¼MeO, R2¼H, R3¼OH, R4¼NMe2) – formed in situ by treatment with aqueous potassium hydroxide – and oxidation with an aqueous solution of potassium ferricyanide boiling under reflux to yield the indolin-2-one 11 (R1¼MeO,LR2+R3¼O, R4¼NMe2), the methiodide of which, 11 (R1¼MeO, R2+R3¼O, R4¼ NMe3 I), was obtained by the following synthesis, that was based upon “promising lines 10 developed” by HS Boyd-Barrett at University College, London, working in collaboration with Robert Robinson then at the University of Oxford but formerly at University College5). Thus, starting from the 4-methoxyphenylhydrazone 12 (R¼MeO) that was formed in situ – from 4-methoxyphenylhydrazine and 2-oxo-5-phenoxypentane – and subjected to Fischer indolisation in alcoholic sulphuric acid boiling under reflux afforded, as would now be expected (Robinson 1982), as the sole product compound 13 (R¼MeO). This was then (King and Robinson 1932a) distinguished from the possible alternative product, 14 (R¼MeO) by its failure to react with Erhlich’s reagent (4-dimethylaminobenzaldehyde in aqueous-alcoholic hydrochloric acid) in the cold – thereby eliminating the presence of a free either 2- or 3-position in the indole nucleus, and heated with iodomethane in methanol under pressure to afford the 3H–indolium salt 15 (R¼MeO). This, upon shaking with a mixture of aqueous sodium hydroxide and ether for 8 h gave the corresponding 2-methyleneindoline 11 (R1¼MeO, R2+R3¼CH2, R4¼OPh) which was oxidised in acetone at 0
by the addition of finely powdered potassium permanganate in small amounts over 9 h to the indolin-2-one 11 (R1¼MeO, R2+R3¼O, R4¼OPh).11 The ether linkage of the phenoxyethyl group of this was then cleaved by heating at 150
C for 12 h with an excess of fuming hydrobromic acid (d 1.7) in a Geissler flask – conditions that also concomitantly effected O–demethylation – to yield 11 (R1¼HO, R2+R3¼O, R4¼Br) which was sequentially O–remethylated by reaction with dimethylsulphate in aqueous sodium hydroxide to afford 11 (R1¼MeO, R2+R3¼O, R4¼Br) and then heated at 150
C for 4 to 6 h in a solution of dimethylamine in 95% methanol in a sealed tube to give 11 (R1¼MeO, R2+R3¼O, R4¼NMe2). The methiodide of this product was resolved via fractional crystallisation of the d-bromocamphorsulphonate, and the l-base, as its picrate, proved to be identical with the picrate of the final product resulting from the above degradation of l-physostigmine (King and Robinson 1932a), observations that led to the conclusion – although the actual synthesis of l-physostigmine was ultimately to elude the efforts of Robert Robinson and his research group at the University of Oxford (Sect. 3.1) and12,13,14 – that “Thus the main features of the constitution of this alkaloid have been established by a synthetical method” (King and Robinson 1932a)15.

2.2 Structure Elucidation

3 (R1= MeNHCO, R2= Me, X= NMe)

35 Me

MeO

3 (R1= Me, R2= Me, X= NMe)

I N N Me Me Me 10

Me

R1

11 ( R1=MeO, R2+R3= O, R4= NMe2)

(CH2)2R4 N

R3

R2 Me 11 ( R1= MeO, R2+R3=O, R4= NMe3I )

11 ( R1= MeO, R2= H, R3 =OH, R4= NMe2)

R (CH2)2OPh R

N H

CH2 C N N Me H 12

(CH2)3OPh

14 (CH2)2OPh

R N H

Me

13

Me

R

(CH2)2OPh

11 (R1= MeO, R2+R3= CH2, R4 =OPh) N 11 (R1= MeO, R2+R3= O, R4= OPh)

Me 15

Me

I

11 (R1=HO, R2+R3= O, R4= Br) 11 (R1= MeO, R2+R3= O, R4= Br) 11 (R1=MeO, R2+R3= O, R4=NMe2)

L

In addition, Hofmann degradation of 11 (R1¼EtO, R2+R3¼O, R4¼ NMe3 I), which was likewise (vide supra) obtained via 3 (R1¼Et, R2¼Me, X¼NMe), afforded trimethylamine and an unsaturated compound which was hydrogenated (one mole of hydrogen) to afford a product “which is considered to be 5-ethoxy-1,3dimethyl-3-ethyl-2-indolinone (X)” (Stedman and Barger 1925). It was later added

36

2 l-Physostigmine (Eserine)

that “As an additional confirmation, the synthesis of the substance represented by X [11 (R1¼EtO, R2+R3¼O, R4¼H)] has been undertaken and it is hoped that it will be possible to communicate the results shortly” (Stedman and Barger 1925). Unfortunately, such a communication does not appear to have been forthcoming – the requisite optical resolution may have been problematic – although, in view of the well-established structure of the starting material 11 (R1¼EtO, R2+R3¼O, L 4  R ¼ NMe3 I ), there can be no doubts concerning the nature of the products in this reaction sequence (Polonovski and Polonovski 1925c; Stedman and Barger 1925). Indeed, when this degradative pathway was repeated some 44 years later in connection with one of the investigations into the establishment of the absolute configuration of l-physostigmine, the structures of all the intermediate products were supported by spectroscopic measurements and that of the final product was verified by further degradation and synthesis (Hill and Newkome 1969) (Sect. 2.4). Moreover, within a few years the proposal of structure 3 (R1¼MeNHCO, 2 R ¼Me, X¼NMe) for l-physostigmine (Polonovski and Polonovski 1925a, b; Robinson 1925; Stedman and Barger 1925) was to be further confirmed by total synthesis.

2.3 2.3.1

Syntheses of:By Early Approaches, Including the First to be Successful

Pioneering attempts – of which three were successful – in this area were concomitantly and independently effected by four research groups (one in the United States of America, two in Japan and one in Great Britain) and have been subjected to analysis in review (Robinson 2002). The first successful synthesis of l-physostigmine to be reported resulted from some thorough and extremely focused elegant experimentation (Boggess 1934; Julian et al. 1934; Julian and Pikl 1935a, b, c) effected under the direction of Percy Lavon Julian (Witkop 1998) in the Minshall Laboratory [constructed by the Spring of 1902 from a gift of $25,000 received from DW Minshall in 1901 and demolished in 1973 (Anon 1999)] of DePauw University, Greencastle, Indiana and the aim of which, in tandem with the alkaloid’s structural verification, was to develop a facile and economical synthetic route to l-physostigmine in excellent yields and in any desired quantity (Julian and Pikl 1935a, b, c).

2.3 Syntheses of:-

37

Preliminary exploratory investigations included a synthesis, using the method of Stollé [by reaction of 2-bromopropanoyl bromide 16 with N-methylaniline (17, R1¼H, R2¼Me) to afford the anilide 18 (R1¼H) which was then cyclised by heating with sublimed aluminium chloride16], of 1,3-dimethylindolin-2-one (19, R1¼R2¼H). This, via the in situ formation of the intermediate anion 20 (R1¼H) by reaction with sodium ethoxide in dry ethanol, was then subjected to 3-methylation by addition of iodomethane to the ethanolic solution boiling under reflux to give compound 19 (R1¼H, R2¼Me) (Boggess 1934; Julian et al. 1934). The synthetic objective of the l-physostigmine tricyclic ring system was then addressed (Boggess 1934; Julian et al. 1934) by substitution of iodomethane by chlorocyanomethane in the 3-alkylation stage of the indolin-2-one 19 (R1¼R2¼H). This afforded, in 90% yield, compound 19 (R1¼H, R2¼CH2CN), reduction of which with sodium in boiling (presumably under reflux) dry ethanol17 effected the simultaneous reduction of both the indolin-2-onic carbonyl group and the cyano group to give the intermediate 21 (R1¼H), acidification of which then forming the intermediate 22 (R1¼H) from which, after intramolecular nucleophilic attack on its imino cationic moiety, dl-desoxynoreseroline (23, R1-R3¼H, R4¼Me) was isolated upon basification (Boggess 1934; Julian et al. 1934).17 Methylation with iodomethane in dry ether then afforded dl-desoxyeseroline (23, R1¼R2¼H, R3¼R4¼Me) (Boggess 1934; Julian et al. 1934). However, in this latter approach, “the reduction of a nitrile to an amine with sodium and alcohol does not result generally in good yields” (Julian and Pikl 1935a) (probably because of competitive alkaline hydrolysis of the cyano group to the corresponding carboxylate as a consequence of the difficulty in excluding moisture from the reaction mixture)17 and it was therefore replaced, prior to the reductive cyclisation, by hydrogenation over Adams palladium catalyst, to afford 19 [R1¼H, R2¼(CH2)2NH2] (Julian and Pikl 1935a). This could also be prepared from 1,3-dimethylindolin-2-one (19, R1¼R2¼H) by treatment of the sodium salt in dry ethanol with 2-bromoethylphthalimide to yield 19 [R1¼H, R2¼(CH2)2-phthalimido] (Julian and Pikl 1935a), followed by hydrazinolysis (Julian and Pikl 1935a). The isolation of this amine served not only the purpose of improving the yield of the ultimate cyclised product but also permitted the obviation of another difficulty, namely the N-methylation of dldesoxynoreseroline (23, R1–R3¼H, R4¼Me) to dl-desoxyeseroline (23, R1¼R2¼H, R3¼R4¼Me). Thus, reaction of the amine 19 [R1¼H, R2¼(CH2)2NH2] with benzaldehyde at room temperature readily yielded the benzylidene derivative 19 [R1¼H, R2¼(CH2)2 N¼CHPh] which was subsequently almost quantitatively quaternised by heating with iodomethane in a sealed tube at 100
C, followed by hydrolysis to afford 19 [R1¼H, R2¼(CH2)2NHMe] (Julian and Pikl 1935a). This was reductively cyclised using sodium in “absolute alcohol” (anhydrous ethanol)17 to yield dl-desoxyeseroline (23, R1¼R2¼H, R3¼R4¼Me) (Julian and Pikl 1935a).

38

2 l-Physostigmine (Eserine)

However, since it had been already established that by catalytic hydrogenation (Stedman and Barger 1925), as well as by reduction with zinc and hydrochloric acid (Polonovski 1918 –see also Polonovski and Polonovski 1924e – see also 1924a), eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) is smoothly converted into compound 9 (R1¼EtO, R2¼R3¼Me),18 it was considered possible that nascent hydrogen in an alkaline medium might effect a similar reduction and that, as a result,

2.3 Syntheses of:-

39

the final products resulting from these above syntheses might be 9 (R1¼H, R2¼Me, R3¼H and Me, respectively) (Julian and Pikl 1935a). However, the latter possible structures, which “could not be excluded satisfactorily solely on the basis of analytical figures” (Julian and Pikl 1935a), were duly eliminated when it was shown that the final products from these syntheses had properties characteristic of secondary and tertiary amines, respectively (Julian and Pikl 1935a), and, like eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me), both were reductively cleaved, catalytically, to yield 2,3-dihydrotryptamines – in their cases 9 (R1¼H, R2¼Me, R3¼H and Me, respectively) (Julian and Pikl 1935a) – and other parallel chemical reactivities were noted (Julian and Pikl 1935a). The scene, therefore, was set for a synthesis of l-physostigmine by the following route (Julian and Pikl 1935b, c):-. This started from phenacetin (17, R1¼EtO, R2¼COMe)19 which was N-methylated by sequential reaction in xylene with sodium powder and dimethyl sulphate and then N-deacetylated by warming in aqueous alcoholic potassium hydroxide to afford N-methyl-p-phenetidine (17, R1¼EtO, R2¼Me) (Julian and Pikl 1935b) which was then reacted with 2-bromopropanoyl bromide (16) to yield the anilide (18, R1¼EtO) (Julian and Pikl 1935b). This, upon heating with an excess of sublimed aluminium chloride,16 underwent cyclisation and cleavage of the ethoxy bond to afford, in excellent yield, 1,3-dimethyl-5-hydroxyindolin-2-one (19, R1¼HO, R2¼H) (Julian and Pikl 1935b). After re-ethylation of the 5-hydroxyl group by treatment in 5% potassium hydroxide solution with diethylsulphate, the resulting product 19 (R1¼EtO, R2¼H) was 3-alkylated using chlorocyanomethane to give 19 (R1¼EtO, R2¼CH2CN).20 Hydrogenation over Adams palladium catalyst then yielded 19 [R1¼EtO, R2¼(CH2)2NH2] (Julian and Pikl 1935b). Condensation of this with benzaldehyde – to give 19 [R1¼EtO, R2¼(CH2)2 N¼CHPh] – followed by quaternisation with iodomethane and then hydrolysis effected conversion into 19 [R1¼EtO, R2¼(CH2)2NHMe] (Julian and Pikl 1935b). This could also be prepared by heating with methanolic methylamine at 100
C the product 19 [R1¼EtO, R2¼(CH2)2Br] that resulted from the treatment of 1,3-dimethyl-5-ethoxyindolin-2one (19, R1¼EtO, R2¼H), as its sodium salt, with 1,2-dibromoethane (Julian and Pikl 1935a, c), a reaction sequence that further simplified the synthesis (Julian and Pikl 1935c) and is “of interest for the cheap production of d,l-eserethole in any desired quantity” (Julian and Pikl 1935c) {these results could possibly be further manifest by an even more direct synthesis of 19 [R1¼EtO, R2¼(CH2)2NHMe] by reaction of the sodium salt of 19 (R1¼EtO, R2¼H) with N-methylaziridine}. Reductive cyclisation of compound 19 [R1¼EtO, R2¼(CH2)2NHMe] using sodium in “absolute alcohol”17 then afforded dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) to complete a synthetic sequence “thus rendering d,l-eserethole a readily accessible substance, since the materials employed are inexpensive and the reactions require relatively little time” and whose “yields at every stage are excellent” (Julian and Pikl 1935b). However, attempts to optically resolve21 the dl-eserethole using either dcamphorsulphonic acid or d-tartaric acid were unsuccessful (Julian and Pikl 1935c). Nevertheless, by the successive reaction with these acids, the amine 19 [R1¼EtO, R2¼(CH2)2NHMe] was resolved, in excellent yields, into its d- and lenantiomers (Julian and Pikl 1935c), the latter of which underwent reductive

40

2 l-Physostigmine (Eserine)

cyclisation to yield l-eserethole (Julian and Pikl 1935c).17 This, by boiling under reflux its solution in petroleum ether (bp 70–77
C) in which anhydrous aluminium chloride was suspended,22 was smoothly converted into l-eseroline (23, R1¼HO, R2¼H, R3¼R4¼Me) (Julian and Pikl 1935c).19 Since this product had already been converted into l-physostigmine (3, R1¼MeNHCO, R2¼Me, X¼NMe) (Polonovski and Nitzberg 1916) by reaction of its anhydrous ethereal solution containing a trace of sodium with methylisocyanate in benzene, a first total synthesis of the alkaloid had therefore been achieved.23,24,25 Synthetic studies, apparently initiated concomitantly with but not reported until twenty-three years later – from the Pharmaceutical Institute in the Medical Faculty of the University of Tokyo (Sugasawa and Murayama 1958a – see also 1958b) – than those of Julian and Pikl, independently followed the latter’s route but with an interesting variation on the approach of Stollé to the necessary indolin-2-onic intermediate (Sugasawa and Murayama 1958a, b). Thus, 2-bromo-2-methyl-4phthalimidobutanoyl chloride (24) was reacted with N-methyl-p-anisidine (17, R1¼MeO, R2¼Me) to afford the anisidide 25 which was cyclised by fusion with a (5:1 w/w) mixture of aluminium chloride with sodium chloride.16 These Me phthalimido-(CH2)2 C COCl Br

24

MeO

Br

Me (CH2)2-phthalimido C C

N

O

Me

25 conditions also effected O-demethylation to yield the indolin-2-one 19 [R1¼HO, R2¼(CH2)2-phthalimido] that was then sequentially O-methylated with diazomethane and subjected to hydrazinolysis to give 19 [R1¼MeO, R2¼(CH2)2NH2]. This, following a route analogous to that reported by Julian and Pikl, namely, N-methylation via quaternisation with iodomethane and then hydrolysis of the benzylidine derivative followed by reductive cyclisation using sodium in ethanol, gave dl-esermethole (23, R1¼MeO, R2¼H, R3¼R4¼Me) (Sugasawa and Murayama 1958a). Almost at the same time as Julian and Pikl, and with the same synthetic objective, Hoshino and Kobayashi, with their occasional co-workers, also in Japan developed a scheme in which tryptamines, with both N-atoms – at most – secondary (as in 26), were subjected to 3-methylation via sequential formation, by reaction with either ethyl- or

2.3 Syntheses of:-

41

methylmagnesium iodide, of their adducts 27 followed by reaction with iodomethane in anhydrous either ether, anisole or benzene boiling under reflux. “Work-up” by hydrolysis with acid then afforded the 3H–indolium cation 28, the imonium function of MgI

H 1

1

(CH2)2N

R

R

R3 N H

(CH2)2N

R2

N

R3

R2

MgI

26

27

23 (R1= R4 = H)

H

Me

R1

(CH2)2N N

2

R3

R

H

28 which, as in Julian and Pikl’s synthesis, readily undergoes intramolecular nucleophilic attack,26 with subsequent basification yielding the corresponding 1,2,3,3a,8,8a– hexahydro-3a-methylpyrrolo[2,3-b]-indoles, including for example, compounds 23 (R1–R4¼H) (Hoshino 1932a, b; Hoshino and Tamura 1932), (R1¼R3¼R4¼H, R2¼Me) (Hoshino 1932a, b, Hoshino and Tamura 1932), (R1¼R2¼R4¼H, R3¼Me) (Hoshino and Kobayashi 1935), (R1¼MeO, R2–R4¼H) (Hoshino et al. 1934), (R1¼EtO, R2–R4¼H) (Hoshino et al. 1934)27 and (R1¼EtO, R2¼R4¼H, R3¼Me) (Hoshino et al. 1935; Hoshino and Kobayashi 1935). In an extension to this synthetic approach, 3-methylindole, 2,3-dimethylindole and 5-ethoxy-3-methylindole were reacted sequentially with ethylmagnesium iodide in ether and then with 1,2-dibromoethane to yield the 3H–indoles 29 (R1¼R2¼H, Me

R1

(CH2)2R3 N

R2

29 R3¼Br; R1¼H, R2¼Me, R3¼Br and R1¼EtO, R2¼H, R3¼Br, respectively). These, upon heating with ethanolic ammonia at 100–105
C for 16 h, were converted into 23 (R1–R4¼H; R1¼R3¼R4¼H, R2¼Me and R1¼EtO, R2–R4¼H, respectively), with compound 29 (R1¼EtO, R2¼H, R3¼Br) being similarly converted into 23 (R1¼EtO, R2¼R4¼H, R3¼Me) by heating with “alcoholic” methylamine at 100–105
C for 15 h (Kobayashi 1939).

42

2 l-Physostigmine (Eserine)

In order to advance these above studies toward the synthetic objective of lphysostigmine, a series of investigative N-methylations of their final products were effected. As a result of these it was established that dl-isonoreserethole (23, R1¼EtO, R2¼R4¼H, R3¼Me) (Hoshino et al. 1934; Hoshino and Kobayashi 1935), as its hydrochloride, when heated with iodomethane in methanol yields dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) (Kobayashi 1938) which, via the formation of its salt with either l- or d-tartaric acid, could – contrary to the earlier observation (Julian and Pikl 1935c) (vide supra) – be optically resolved (Hoshino and Kobayashi 1936; Kobayashi 1938)21. Then, application of already-published procedures (Julian and Pikl 1935c) involving O-de-ethylation of the l- and d-enantiomers and the racemate by heating with anhydrous aluminium chloride suspended in petroleum ether22 gave l-, d- and dl-eserolines (Kobayashi 1938) which, upon treatment in anhydrous ether with a trace of sodium followed by methylisocyanate in benzene (Polonovski and Nitzberg 1916) resulted in a second successful total synthesis of lphysostigmine, along with those of its enantiomer and its racemate, respectively (Kobayashi 1938). However, the salicylates of the enantiomers are quoted (Kobayashi 1938) as being dextrorotatory and laevorotatory, respectively, which is contrary to now well-established data (British Pharmacopoeia 1968), and thus it has been concluded (Dale and Robinson 1970) that this work “leaves much to be desired”. Another N-methylation of particular significance (vide infra) involved that by various procedures of dl-dinoreserethole (23, R1¼EtO, R2–R4¼H), reactions from which a base was isolated. This was assigned (Hoshino and Kobayashi 1934a – see also 1934b) a molecular formula C16H24N2O and three N-methyl groups, was ultimately to be designated “methyl-eserethole”, and was given the structure (30) Me

EtO N

N

Me Me Me

30 (Hoshino and Kobayashi 1934a). However, it was later appreciated that this is a “highly improbable constitution” (King and Robinson 1935; King et al. 1934) – and indeed, that it cannot be correct since it “violates fundamental theory” (Jackson 1954; King and Robinson 1935), although it has been regurgitated without comment (Julian and Pikl 1935b). Moreover, concomitantly with further methylations – mostly effected with iodomethane under effectively either neutral or alkaline conditions – of dldinoreserethole (23, R1¼EtO, R2–R4¼H) and dl-isonoreserethole (23, R1¼EtO, R2¼R4¼H, R3¼Me) to give “methyl-eserethole” (Kobayashi 1938), its molecular formula was revised to C15H22N2O and it was found to contain only two N-methyl groups (Kobayashi 1938), conclusions that were later confirmed12 by Robert Robinson’s group in Oxford. Also of considerable significance was the reaction of dl-isonoreserethole methiodide (31) with cold dilute alkali to afford an almost

2.3 Syntheses of:-

43

Me

EtO

29 (R1= EtO, R2= H, R3=NMe2) N I H Me Me H N

31 quantitative yield of “methyl-eserethole”, an observation from which it follows that “methyl-eserethole” has structure 29 (R1¼EtO, R2¼H. R3¼NMe2) (Kobayashi 1938).12 This postulation was ultimately verified (Kobayashi 1939) by synthesis involving the sequential reaction of 5-ethoxy-3-methylindole with ethylmagnesium iodide in anhydrous ether and then with 1,2-dibromoethane to yield the 3H–indole 29 (R1¼EtO, R2¼H, R3¼Br) which, upon heating with alcoholic dimethylamine afforded “methyl-eserethole” (Kobayashi 1939). The fourth group whose synthetic sights were focused upon l-physostigmine during the 1930s was that directed by he who was to be crowned “the king of alkaloid chemistry” (Brossi et al. 1996), Robert Robinson5 – initially at the University of Manchester (Robinson and Suginome 1932a, b) and later at the Dyson Perrins Laboratory at the University of Oxford – and the work of which was reported in a short series of closely inter-related, and consequently somewhat fragmented and reiterative, papers, the clarity of which thus occasionally leaves much to be desired. Nevertheless, they do constitute the first published approach toward l-physostigmine’s synthesis although, unfortunately, without ultimate success12 – the statement “As apparent from the first synthesis of physostigmine (1) by Robinson and co-workers in the early 1930s” (Takano and Ogasawara 1989) is clearly erroneous. So where, why and how did these synthetic efforts break down? Sometimes this group either missed or failed to reach its synthetic destination and, indeed, failed to realise in one instance that it had arrived! This situation, which has been subjected to clarification (Robinson 2002), arose in connection with a reaction sequence that involved the intermediacy of the unequivocally-synthesised compound 19 [R1¼MeO, R2¼(CH2)2Br] (Robinson and Suginome 1932a). When this “was heated under pressure with an excess of saturated alcoholic ammonia, a small amount of a crystalline solid was produced, but it proved to be a secondary base formed from two molecules of the indolin-2-one bromide and one of ammonia (Robinson and Suginome 1932b). However, upon reaction with potassium phthalimide, 19 [R1¼MeO, R2¼(CH2)2Br] afforded 19 [R1¼MeO, R2¼(CH2)2-phthalimido] (King and Robinson 1932b) which upon hydrazinolysis gave 19 [R1¼MeO, R2¼(CH2)2NH2] (King and Robinson 1932b). The pyrrolo ring was then formed by reaction with phosphorus pentoxide in dry xylene boiling under reflux to afford the cyclic amidine (32) (King and Robinson 1932b). This was then reduced by hydrogenation in the presence of a platinum catalyst in ethyl acetate (no more precise conditions were stated) to give a

44

2 l-Physostigmine (Eserine)

product that was thought to be 23 (R1¼MeO, R2¼R3¼H, R4¼Me) (characterised as its picrolonate) (King and Robinson 1932b) which was then N-methylated to yield what was claimed (King et al. 1934) to be dl-esermethole.28 However, it would appear that the conditions employed for the reduction of the cyclic amidine 32 did not preclude the reductive cleavage of the tetrahydropyrrolo ring C in the now-believed to be the intermediately-formed compound 23 (R1¼MeO, R2¼R3¼H, R4¼Me)18 and there is every reason to suppose that the reduction as effected by Robert Robinson’s group gave “not a tricyclic compound” (Kolosov et al. 1953) and thereby ultimately not dl-esermethole (Kolosov et al. 1953) but continued right though to the 2,3-dihydrotryptamine 9 (R1¼MeO, R2¼R3¼H) (Coxworth 1965; Jackson 1954; Kolosov et al. 1953). In his failure to appreciate this possibility, Robert Robinson was clearly less prudent than had been Julian and Pikl (1935a) (vide supra). Indeed, it was later recognised (Jackson 1954) that it was this shortcoming that led the Oxford group astray in the identification of the product resulting from their reduction of compound 19 [R1¼MeO, R2¼(CH2)2NH2] – which they synthesised as described above (King and Robinson 1932b) – using sodium in boiling (presumably under reflux) isoamyl alcohol (attempted “electrolytic reduction” afforded only recovered starting material) (King and Robinson 1932b). By analogy with the subsequent work of Julian and Pikl (vide supra), this reaction could only have given dl-noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me).29 but after considering this structure as “the more remote alternative” it was “rejected” and the contemporary suggestion (King and Robinson 1932b) was that, since the product was, by comparison of their picrolonates, different from that produced by the reduction of the cyclic amidine 32 – which was then thought to have yielded 23 (R1¼MeO, R2¼R3¼H, R4¼Me) – it had afforded 9 (R1¼MeO, R2¼Me, R3¼H) (King and Robinson 1932b).12 It would thus appear “that the actual structures” (Jackson 1954) of the products resulting from the reductions of the cyclic amidine 32 and the indolin-2-one 19 [R1¼MeO, R2¼(CH2)2NH2] “are in fact the reverse of those originally assigned by King and Robinson” (Jackson 1954).

1 2 19 [R = MeO, R = (CH2)2Br]

19 [R1= MeO, R2= (CH2)2-phthalimido]

Me

MeO

19 [R1 MeO, R2= (CH2)2NH2] N Me 32

[23 (R1= MeO, R2= R3= H, R4= Me)]

N

x 1 2 3 9 (R = MeO, R = R = H)

1 2 3 4 23 (R = MeO, R = R = H, R = Me)

2.3 Syntheses of:-

45

In addition to the above studies, Robert Robinson and his team also developed a further potential – but again unsuccessful – route toward a synthesis of l-physostigmine, now starting from acetaldoloxime (33). This was reduced by aluminium amalgam in water into 3-hydroxybutylamine (34) which, when heated with phthalic anhydride, gave 35 (R¼OH) (Robinson and Suginome 1932b). Reaction of this with hydrogen bromide in ethanol afforded 35 (R¼Br) which upon treatment with ethyl acetoacetate (36) in the presence of ethanolic potassium ethoxide led to 37 (Robinson and Suginome 1932b). This, when coupled with 4-ethoxybenzenediazonium chloride in alkaline solution, lost an acetyl group and gave the arylhydrazone 38 (Robinson and Suginome 1932b). Fischer cyclisation (Robinson 1982) of this by reaction in boiling ethanolic hydrogen chloride under reflux then yielded the 3H–indole 39 (R¼COOEt) (Robinson and Suginome 1932b). This, upon hydrolysis with ethanolic potassium hydroxide, afforded 40 which was decarboxylated and dehydrated by boiling in xylene under reflux to afford 39 (R¼H) (King et al. 1933a; Robinson and Suginome 1932b). However, because of competing side-reactions (Robinson and Suginome 1932b), the yield from this latter reaction was very low (6.6–8.5%) (Robinson and Suginome 1932b), but this shortcoming was obviated by the direct preparation of 39 (R¼H) via the Fischer indolisation (Robinson 1982) of a mixture of the aldehyde 41, that was synthesised using standard procedures (King et al. 1933b, 1934), with 4-ethoxyphenylhydrazine (42) in saturated “alcoholic” hydrogen chloride (King et al. 1933b). The protecting phthaloyl group was then removed from the methosulphate of compound 39 (R¼H), using hydrazine hydrate in boiling “alcohol” (under reflux?) followed by acidification with hydrochloric acid, to afford, via the intermediacy of 43 (R¼H, X¼MeSO4) and after basification, dl-noreserethole (23, R1¼EtO, R2¼R3¼H, R4¼Me) (Robinson and Suginome 1932b) [dl-noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me) was also similarly prepared (King et al. 1934)]. The methylation of the former product using methyl p-toluenesulphonate was claimed by Robert Robinson and his co-workers (King et al. 1933a, 1934, Robinson and Suginome 1932b) to give dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) and related to this work it has been stated that “dl-Noreserethole is methylated to dl-eserethole by the action of methyl p-toluenesulphonate” (Marion 1952). However, this was not so12 and it was at this final stage that this synthetic effort broke down to afford as product “methyl-eserethole” (29, R1¼EtO, R2¼H, R3¼NMe2).12 So once again the Oxford group’s endeavours had been thwarted by a cleavage of the tetrahydropyrrolo ring of the 1,2,3,3a,8,8a–hexahydro-3a-methylpyrrolo[2,3-b]indole moiety.

46

2 l-Physostigmine (Eserine)

Although the formation of “methyl-eserethole” from the methylation of either dldinoreserethole (23, R1¼EtO, R2–R4¼H) (Hoshino and Kobayashi 1934a, b;

2.3 Syntheses of:-

47

Kobayashi 1938) or dl-isonoreserethole (23, R1¼EtO, R2¼R4¼H, R3¼Me) (Kobayashi 1938) (vide supra) is without problem since it clearly occurs via a simple dimethylation of the N1-atom (Jackson 1954), its formation from Robert Robinson’s methylation of dl-noreserethole (23, R1¼EtO, R2¼R3¼H, R4¼Me) (King et al. 1933a, 1934; King and Robinson 1935), involving as it does the loss of a methyl group from N(8), is far from clear but has been subjected to considerable speculation (Ault 2008; Jackson 1954 – and also referred to in Coxworth 1965; Robinson 2002)30 and experimentation (Jackson 1954). Although it was thought to be unlikely, the possibility that the methyl p-toluenesulphonate used by the Oxford group in their alkylation of dl-noreserethole (23, R1¼EtO, R2¼R3¼H, R4¼Me) (King et al. 1933a, 1934; King and Robinson 1935) first afforded dl-eserethole ptoluenesulphonate (44, R¼H) which then rearranged into “methyl-eserethole” (29, R1¼EtO, R2¼H, R3¼NMe2)30 was then investigated, and apparently eliminated, when compound 44 (R¼H), formed by treating dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) with one equivalent of p-toluenesulphonic acid, after being “treated under exactly the same conditions” as were employed for the methylation of dl-noreserethole (King et al. 1933a) (see also King et al. 1934, King and Robinson 1935) led to an almost quantitative recovery of unchanged dl-eserethole as the only product (Jackson 1954).30 However, another remaining possibility is that the methyl p-toluenesulphonate effects a dimethylation of N1, the much more basic nitrogen atom, to give compound 44 (R¼Me) which upon basification would afford dl-eserethole methine (racemic 45, R1¼OH, R2¼H) that, as in the work of Robert Robinson’s group,

would be readily extracted into ether and could, not inconceivably, upon distillation simply lose methanol to afford “methyl-eserethole” (29, R1¼EtO, R2¼H, R3¼NMe2) (Jackson 1954), the crucial point being not only the Oxford group’s product distillation but also the latter’s conditions. In an attempt to verify this theory, eserethole methine (45, R1¼OH, R2¼H) was prepared from l-physostigmine as already reported (Polonovski and Polonovski 1924a, e) and subjected to distillation at various temperatures (Jackson 1954). At temperatures of below 120
C and pressures

48

2 l-Physostigmine (Eserine)

of the order of 0.1 mm or less it distilled unchanged whereas at higher temperatures decomposition slowly occurred and, by fractional crystallisation of their picrates, the distillate afforded the picrate of dehydroeserethole methine (45, R1+R2¼O), prepared by alkaline ferricyanide oxidation of eseretholemethine (45, R1¼OH, R2¼H) (Stedman and Barger 1925), and the methopicrate of eserethole – resulting from unchanged eserethole methine (Jackson 1954). Since it appears that the former product may have resulted from aerial oxidation of the eserethole methine, this latter was therefore heated at 170
C for 3 h under an atmosphere of nitrogen and then distilled at 0.1 mm, but the only product was impure eserethole methine, isolated as eserethole methopicrate from the distillate (Jackson 1954). In view of the failure to detect from any of the above experiments any “methyl-eserethole” it was therefore concluded that its formation “on methylation of noreserethole thus still remains something of a mystery [see, however30], though the author’s explanation is regarded as at least a partial solution of the problem” (Jackson 1954). However, further experimental investigation is required in order to clarify this situation (Robinson 2002).30

2.3.2

l-Physostigmine and the 3a–alkyl-1,2,3,3a,8,8a– hexahydropyrrolo[2,3-b]Indole Ring System

Subsequent to the above early synthetic approaches and almost certainly stimulated by an expanding catalogue of useful therapeutic and other applications over recent years (Chap. 10), it is perhaps not surprising that l-physostigmine and its component 1,2,3,3a,8,8a–hexahydro-3a-methylpyrrolo[2,3-b]indole ring system has continued to attract considerable synthetic attention which has been the subject of review on several occasions (Kutney 1977; Lee and Wong 1991; Muhtadi and El-Hawary 1989; Robinson 1971, 1982, 2002; Shishido and Fukumoto 1988; Shishido et al. 1986b; Takano and Ogasawara 1989). Following the successful outcome in 1935 of the endeavours of Julian and Pikl and of Hoshino and Kobayashi (Sect. 3.1), the alkaloid’s synthetic challenge remained dormant for some two decades until the appearance of an entirely different and unique – in that it involved the concomitant formation of the eseroline ring system and its 5-hydroxy substituent – synthetic approach which is based upon a variation on the Nenitzescu synthesis [Spande 1979(b)] of 5-hydroxyindoles. From the University of Cambridge, this evolved from the observation (Harley-Mason and Jackson 1954a; Jackson 1954) that oxidation with ferricyanide of the hydroquinone 46 leads (presumably via the intermediacy of the para-quinone 47) to a good yield of 5-hydroxytryptamine (serotonin) (48). Moreover, in order to determine the effect of N-methylation upon the formation of 5-hydroxyindoles using this approach, 2-(2,5-dihydroxyphenyl)-N-methylethylamine (49) was synthesised and oxidised with potassium ferricyanide to afford (presumably now via the intermediates 50, 51 and 52) 5-hydroxy-1-methylindole (53) (Harley-Mason and Jackson 1954a; Jackson 1954).

2.3 Syntheses of:-

49 (CH2)2NH2

(CH2)2NH2 HO

CH

O

CH

CH2NH2

(CH2)2NH2

HO CH2NH2

OH

46

O

N

47

H

48

(CH2)2NHMe

HO

OH

49

O

(CH2)2NHMe O

50

O

O N OH Me

N Me

51

52

HO N Me

53

The scene was now set for a synthesis of dl-eseroline by the following route (Harley-Mason and Jackson 1954b; Jackson 1954):-. Ammonium acetate/acetic acid catalysed condensation of 2,5-dimethoxyacetophenone (54) with ethylcyanoacetate (55) afforded the product 56 which, when heated in ethanol with potassium cyanide gave – via Michael addition, saponification of the carbethoxy group and decarboxylation – the substituted succinonitrile 57. This was catalytically hydrogenated to yield the primary diamine 58 (R1¼Me, R2¼H) which via sequential di-Schiff base formation by condensation with benzaldehyde, bisquaternisation by heating with iodomethane in a sealed tube at 100
C, and hydrolysis with dilute hydrochloric acid, gave the secondary diamine 58 (R1¼R2¼Me). Heating this with hydrobromic acid then gave the dihydrobromide of the hydroquinone 58 (R1¼H, R2¼Me) which upon oxidation with potassium ferricyanide in the presence of sodium bicarbonate afforded (presumably via the intermediacy of the para-quinone 59 and then 60) dl-eseroline (23, R1¼OH, R2¼H, R3¼R4¼Me), with the yield from this double ring closure stage

50

2 l-Physostigmine (Eserine) Me

O MeO

C

CN Me

+

OMe

CH2 OMe

CO2Et

54

56

55

Me

Me

R1O

C C(CN)CO2Et

MeO

C CH2CN CN OMe

MeO

O NHR2 NHR2 R1 58 (R1= Me, R2= H)

57

58 (R1= R2= Me)

58 (R1= H, R2= Me)

Me

O N

N

Me

Me H 60

O

Me

O NHMe NHMe 59

23 (R1= OH, R2= H, R3= R4= Me)

being 30% of pure product. This was reacted with ethyl-p-toluenesulphonate in ethanolic sodium ethoxide to yield dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) (HarleyMason and Jackson 1954b; Jackson 1954) that had earlier been resolved and then converted into l-physostigmine (Kobayashi 1938). Although the previous synthesis was conceptually innovative, the next to appear (Robinson 1982; Rosenmund and Sotiriou 1964, 1975; Rosenmund and Sadri 1979) [from the University of Frankfurt (Main) and thereby representing

2.3 Syntheses of:-

51

Me

Me

62 (R= CN)

COOEt

COOEt

R

61

62 (R= Br)

R1

63

N(Ph)

62 R=

62 (R= CHO)

N NH2

N(Ph)

R2

Me CH

R1

CH2COOEt

Me

R1

CH2COOEt

CH N N

N

R2

R2 65

64

Me

Me

R1 O

N

O

OEt

Me 67

Me

R1 N

X

N

O

R2 66

H

R2 68 (X= Y= O) 68 (X= NH and NMe, Y= O) 23 ( R2= H, R3= R4= Me)

OH

Me

R1O Y

C O OEt

R2

N

X

Me

H

69

OH

52

2 l-Physostigmine (Eserine)

Germany’s first contribution to this area of research] resembled in many of its salient features the pioneering syntheses of the early 1930s (Sect. 3.1) in that the Fischer synthesis (Robinson 1982) was employed to construct the indole-derived nucleus and the tetrahydropyrrolo ring C was elaborated via an intramolecular nucleophilic attack. Thus, starting with ethyl crotonate (61), sequential hydrobromination with hydrogen bromide in acetic acid, cyanation by heating with ethanolic aqueous potassium cyanide, reductive amination with N,N΄-diphenylethylenediamine in the presence of a Raney nickel catalyst, and acid-catalysed hydrolysis afforded, via 62 [R¼Br, CN and CH(NPh)2(CH2)2 respectively], the product 62 (R¼CHO) (Rosenmund and Sadri 1979; Rosenmund and Sotiriou 1975). Reaction of this aldehyde with Nα-methylphenylhydrazine (63, R1¼H, R2¼Me) then afforded the Nα-methylphenylhydrazone 64 (R1¼H, R2¼Me) which was then subjected to tin(II) chloride-catalysed Fischer indolisation to yield 65 (R1¼H, R2¼Me) that was isolated as a salt formed with the catalyst and which, upon sequential acidification with hydrochloric acid, removal of the tin(II) ion with hydrogen sulphide and basification to pH 9–10, afforded 66 (R1¼H, R2¼Me). This product, upon warming with methanolic potassium hydroxide, gave 68 (R1¼H, R2¼Me, X¼Y¼O), probably via the intermediate 67 [the mechanism of ring-closure of 66 is analogous to the proposed illustrated mechanism of ring L closure of 69 (R1¼Et, R2¼ NMe3 I, X¼O), by treatment with boiling caustic soda, to form the corresponding optical isomer of 68 (R1¼EtO, R2¼Me, X¼O, Y¼H2) (Longmore 1969; Longmore and Robinson 1966) (Sect. 3.2.2)]. The product 68 (R1¼H, R2¼Me, X¼Y¼O) was then reacted with either ammonia or methanolic methylamine to afford 68 (R1¼H, R2¼Me, X¼NH or NMe, respectively, Y¼O) and the latter of these products was reduced by lithium aluminium hydride in boiling tetrahydrofuran under reflux to give dl-desoxyeseroline (23, R1¼R2¼H, R3¼R4¼Me) in good overall yield (Rosenmund and Sotiriou 1964, 1975). This synthetic route was then (Rosenmund and Sadri 1979) applied via the Fischer indolisation of the arylhydrazones 64 (R1¼H, R2¼CH2Ph and CH2CH¼CH2; R1¼MeO and EtO, R2¼Me) to afford products with the corresponding structures 68 (X¼Y¼O), some of which upon sequential reaction with various primary amines followed by reduction with lithium aluminium hydride produced the correspondingly substituted eserine systems (Rosenmund and Sadri 1979), of particular significance being dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) – which had earlier been synthesised by Julian and Pikl (1935b), resolved (Hoshino and Kobayashi 1936; Kobayashi 1938) and then converted into l-physostigmine (Julian and Pikl 1935c; Polonovski and Nitzberg 1916) – of which it thereby constituted yet another formal synthesis with a significant and practical yield (Robinson 1982; Takano and Ogasawara 1989).

2.3 Syntheses of:-

53

54

2 l-Physostigmine (Eserine)

A treatise (Robinson 1982) on the reaction included specific references to the Fischer indole synthesis in synthetic approaches that have been made to the 1,2,3,3a,8,8a– hexahydropyrrolo[2,3-b]indole system. Although some of these have already been referred to in this chapter, other applications have also appeared. Foremost amongst these is an innovative Russian contribution that has resulted from the work of Grandberg’s research group which conceived and developed a new synthesis of tryptamines and homotryptamines using a modification of the Fischer indole synthesis. These studies, which have been the subject of extensive reviews (Grandberg 1974, 1983; Robinson 1982), have established that, in the absence of catalysts, tryptamines 78 (R1¼R3¼H, n¼1) can be obtained directly by reacting arylhydrazines 70 (R1¼H) with 4-chloro-, 4-bromo-, 4-iodo- and 4-tosylbutyraldehyde (71, R3¼R4¼H, R5¼Cl, Br, I and tosyl, respectively, n¼1) – with optimum yields in these and the related reactions described below being obtained using the chloro compounds – in boiling aqueous alcoholic (usually either ethanolic or methanolic) solution under reflux. In analogous reactions under similar conditions, compounds 71 (R3¼Me, R4¼H, n¼1 and R3¼Ph, R4¼H, n¼1) afforded the 2-methyl- and 2-phenyltryptamines 78 (R1¼H, R3¼Me and Ph, respectively, n¼1), other 2-alkyl- and 2-aryltryptamines could be similarly prepared and arylhydrazines 70 reacted with 71 (R3¼H, Me and alkyl, R4¼H, n¼2) to yield homotryptamines, 2-methylhomotryptamines and 2-alkylhomotryptamines 78 (R3¼H, Me and alkyl, respectively, n¼2). It has been suggested (Grandberg 1974, 1983; Robinson 1982) that these direct syntheses of tryptamines proceed via the intermediates 73–77 by a mechanism related to that of the Fischer indolisation after the occurrence of intramolecular quaternisation at the more basic nitrogen atom, Nβ, in either the arylhydrazone 72 or the enehydrazine 74, as shown. Verification for this scheme was forthcoming from many observations, amongst which was that in which compounds corresponding to 77 can be isolated when R4 6¼ H – and thus the reaction became applicable to the preparation of the physostigmine tricyclic ring system. For example, the arylhydrazines 70 reacted with 71 (R3¼R4¼Me, R5¼halogeno, n¼1) in boiling alcoholic solution under reflux to afford the corresponding 77 (R3¼R4¼Me, n¼1) and under similar conditions with 2-(2-chloroethyl)cyclohexanone [71, R3+R4¼(CH2)4, R5¼Cl, n¼1] to give the echibolines 77 [R3+R4¼(CH2)4, n¼1]31 [an approach which has been applied to the synthesis of 6-hydroxyechibolines with antinociceptive activity, (Robinson et al. 1987, 1988) (Sect. 10.11)] and with 71 (R3¼R4¼Me, R5¼halogeno, n¼2) to yield 77 (R3¼R4¼Me, n¼2). Likewise, compounds 79 (R¼H, Me and Et) were prepared by reacting mixtures of 2-pyridylhydrazine and Nα-methyl- and Nα-ethyl-2pyridylhydrazine, respectively, with 71 (R3¼R4¼Me, R5¼Cl, n¼1) in boiling aqueous ethanol under reflux. Another widely-applicable synthesis of tryptamines, first investigated by Abramovitch together with occasional co-workers (Robinson 1982), proceeded via

2.3 Syntheses of:-

55

R4

R6 R5

R3

R7

+ R2

N2X

N

EtOOC

R1 80

R4 R8

R3

R9

R2

R5 N H

R1

O 81

R7

R6

N

N

R8 R9

O 82 (R5= H)

R4

R6

R3 84 (R10= H) R2 R1

N H

R7

R4

R6

R3

NHR9

R8 R10

R2 R1

84 (R10= COOH)

R7 N

N H

R8 R9

O 83

the Fischer indolisation of piperidin-2,3-dione 3-arylhydrazones (82, R1–R4 ¼ variously Me, CH2Et, CMe3, OMe, OEt, OCH2Ph, SMe, SCH2Ph, F, Cl, Br, MeCO, EtCO, PhCO, 4ClC6H4CO, isonicotinoyl, COOH, NH2SO2, Me2NSO2 and CF3, R5¼H; R6–R9¼variously H and Me) (Robinson 1982) – prepared by Japp-Klingemann reaction between aryldiazonium salts 80 and the 3-ethoxycarbonyl-2-oxopiperidines 81 (Robinson 1982) – into the 1,2,3,4tetrahydro-β-carbolin-1-ones 83 which upon hydrolysis with aqueous alcoholic alkali yielded the corresponding tryptamine-2-carboxylic acids 84 (R10¼COOH), decarboxylation of which, usually by reacting in mineral acid solution boiling under reflux, then affording the tryptamines 84 (R10¼H). Unfortunately, an attempt made (Abramovitch 1958) to extend this procedure to the synthesis of the l-physostigmine tricyclic ring system met with failure when it was found that the phenylhydrazone 82 (R1–R4¼R6–R9¼H, R5¼Me) could not be indolised – the phenylhydrazone remaining unchanged when boiled in 90% acetic acid under reflux, and upon heating either with boron trifluoride etherate in acetic acid or with zinc chloride at 170
C afforded only an isomeric product – probably geometrical – in each case. From this reaction the potential product was 85 (R1¼R3¼H, R2¼Me), and the ethoxy analogue of this, 85 (R1¼EtO, R2¼Me, R3¼H), had been, along with other

R1

R2

R3

N

N

85

O

H

56

2 l-Physostigmine (Eserine)

products, obtained earlier (Robinson and Suginome 1932b) (Sect. 3.1) – during some of the pioneering studies that were directed toward a synthesis of l-physostigmine – by the action of hydrazine, followed by hydrochloric acid, on compound 39 (R¼COOEt). However, apart from its facile reduction to the corresponding indoline (Robinson and Suginome 1932b), perhaps surprisingly, and unfortunately in view of the then later synthetic failure of these particular investigations (vide supra), the potential sequential hydrolysis and decarboxylation into 86 (R¼COOH and H, respectively) and cyclisation into the corresponding EtO

Me

N

NH2 R

86 1,2,3,3a,8,8a–hexahydro-3a-methylpyrrolo[2,3-b]indole was not investigated. It may be of interest to explore these possibilities. Indeed, Abramovitch in his later studies reported that Fischer indolisation of the related phenylhydrazone 82 [R1–R4¼R7–R9¼H, R5+R6¼(CH2)4] (probably the trans isomer) appears to have been successfully effected since, after boiling under reflux in 90% formic acid, a product was obtained that has been tentatively assigned the structure 85 [R1¼H, R2+R3¼(CH2)4]. Although ir and uv spectroscopic data supported this postulation, the microanalytical data, whilst clearly indicating that one of the original nitrogen atoms in the starting phenylhydrazone had been lost, does not correspond and, furthermore, attempted alkaline hydrolysis of the product simply gave rise to an isomer having a sharp mp (Abramovitch and Muchowski 1960). It has been concluded that “Clearly, these reactions are worthy of further investigation” (Robinson 1982) as, indeed, is this potential route to the 1,2,3,3a,8,8a–hexahydro-3a-methylpyrrolo[2,3-b]indole ring system starting with the appropriate 4-ethoxy-Nα-methylphenylhydrazone in which the 4-ethoxy (Robinson 1982) and Nα-methyl (Robinson 1982) substituents should not only encourage Fischer indolisation but would also afford the correct potential substituents for the ultimate synthetic target, namely l-physostigmine. The acetic acid catalysed indolisation of the 4-methoxy-Nαmethylphenylhydrazone 87 (n¼1), which along with several related indolisations have been the subject of review (Robinson 1971, 1982), occurred in both possible directions with respect to the asymmetrical ketonic moiety to afford both 88 and 89 (R1¼MeO, R2¼Me, R3¼H, X¼O, n¼1) (Fritz and Fischer 1964). The lactam carbonyl group of this latter product, suspended in ether boiling under reflux, is reduced by lithium aluminium hydride during 1 h to a methylene unit (Fritz and Stock 1970) – in contrast to its homologue,32 and further sequential transformations of the reduction product 89 (R1¼MeO, R2¼Me, R3¼H, X¼H2, n¼1) afforded the interesting tetracyclic analogue of physostigmine, 89 (R1¼MeNHCOO, R2¼R3¼Me, X¼H2, n¼1) (Fritz and Stock 1970). However, the biological properties of this product do not appear to have been investigated (Fritz and Stock 1970), which is unfortunate since a potential increase in lipophilicity related to the presence of the trimethylene bridge might be of interest (Sect. 10.7.2).

2.3 Syntheses of:-

57

A subsequent attempt directly to synthesise 89 (R1¼MeO, R2¼R3¼H, X¼H2, n¼1) using the Grandberg application of the Fischer indole synthesis (vide infra) followed the report (Grandberg 1983; Robinson 1982; Robinson et al. 1988) that reaction of arylhydrazines with 71 [R3+R4¼(CH2)4, R5¼halogen, n¼1], namely a 2-(2-halogenoethyl)cyclohexanone, leads to the tetracyclic system 89 (R2¼R3¼H, X¼H2, n¼2). However, in an attempt (Robinson and Hawkins 1985) to extend this synthetic route to the preparation of compound 89 (R1–R3¼H, X¼H2, n¼1), an equimolar mixture of 2-(2-chloroethyl)cyclopentanone [71, R3+R4¼(CH2)3, R5¼Cl, n¼1] and phenylhydrazine was boiled under reflux in benzene with azeotropic removal of water to afford the corresponding phenylhydrazone 72 [R1¼R2¼H, R3+R4¼(CH2)3, R5¼Cl, n¼1] but when this, in ethanolic solution under an atmosphere of nitrogen, was boiled under reflux for 18 h, the only isolable product, along with phenylhydrazine hydrochloride (16%), was compound 90 (40%), formed by nucleophilic displacement of the chloride group by the Nα atom. This is probably preferred to that involving the Nβ atom because the latter would involve the formation of the relatively sterically-strained system 91 (Robinson and Hawkins 1985).

58

2 l-Physostigmine (Eserine)

N N Ph

90

Ph N N H

91 It has been suggested (Coxworth 1965) that “A further possible route to the physostigmine ring system [or at least to the pyrrolo[2,3-b]indole ring system] would be ring closure of the appropriate amidrazone” 92 (n¼1) “to the α-aminoindole” 93 (n¼1), a n

N N

N

R1

R2

92

n

N

N

R1

R2

93 reaction which “would then be analogous to the well-known Fischer indole synthesis employing ring closure of phenylhydrazones” (Coxworth 1965). Indeed, it has been claimed that a synthesis of 93 (n¼2) has been accomplished by this method, starting from 92 (n¼2) (Rapoport et al. 1959a, b), “prepared from phenylhydrazine and 2-oxopiperidine in the presence of phosphorus oxychloride” (Rapoport et al. 1959b), “and the procedure has also been found useful in the synthesis of α-aminoindoles of types” 93 (n>2) and 94 (Rapoport and Coxworth 1959) although “attempts to extend R4 N

N

R1

R2

94

R3

2.3 Syntheses of:-

59

this reaction to the synthesis of pyrrolo[2,3-b]indoles, and then to physostigmine, have not yet been successful” (Coxworth 1965). “However, it is unfortunate, though perhaps significant, that experimental details relating to these above studies have not [based upon a search through Chemical Abstracts], been published” (Robinson 1982) although a later claim has been made (Saleha et al. 1978) – that also “is difficult to understand” (Robinson 1982) – that 2-carbethoxyacetanilide “phenylhydrazone” (Saleha et al. 1978) when boiled under reflux in aqueous formic acid [“50:50”(Saleha et al. 1978)] for 5 h undergoes a Fischer indolisation – to afford what is presumably 2-(2-carbethoxyphenylamino)indole in “good” (Saleha et al. 1978) yield “with very little complications due to side reactions or decompositions” (Saleha et al. 1978), with the yield being improved using ethanolic formic acid [“1:1” (Saleha et al. 1978)], boiling under reflux for 2.5 h as “a cyclizing agent” (Saleha et al. 1978). Further investigation of all these above claims (Coxworth 1965; Rapoport and Coxworth 1959; Rapoport et al. 1959a, b; Saleha et al. 1978) would be of considerable interest. Other more recent synthetic approaches to the physostigmine ring system [many of which are enantioselective (Ashimori et al. 1993; Bruncko et al. 1994; ElAzab et al. 2000; Lee and Wong 1991; Marino et al. 1992; Node et al. 1991; Takano et al. 1990a, b, 1991; Takano and Ogasawara 1989)] have been already either discussed in detail {ElAzab et al. 2000, Jackson 1954 [quoted as ref. 19 in (Robinson 1963a)], Shishido et al. 1986b, Takano and Ogasawara 1989; Yu et al. 1994} or comprehensively listed and referenced (Ashimori et al. 1993; Bruncko et al. 1994; Lee and Wong 1991; Rege and Johnson 2003; Robinson 2002; Shishido et al. 1986b; Shishido and Fukumoto 1988; Takano et al. 1991; Yu et al. 1994) and two warrant individual reference, namely (Ashimori et al. 1993) – because it “may have practical potential” (Greig et al. 1995a), and (Takano et al. 1990a) – because of its stereochemical originality. Although many of these are certainly innovative, in several instances (ElAzab et al. 2000; Takano et al. 1990a, 1991) they involve as their “key” step that doyen of indole syntheses, namely that of Emil Hermann (Farber 1970, Lucier 1993) Fischer (Robinson 1982), which has already found wide use in chemical studies upon l-physostigmine (vide supra) and in the synthesis of indole alkaloids and other natural products,33 and they may also be considered to be somewhat esoteric and for the most part of academic interest, an opinion that echoes the earlier observation (Greig et al. 1995a) that “Several elegant enantioselective syntheses of Calabar alkaloids were reported” (Takano and Ogasawara 1989), “but they remain primarily of academic interest”. Consequently, early opinions that Julian and Pikl’s pioneering, classical and elegant approach (Sect. 3.1) is “the most satisfactory process” (Bentley 1957) and “appears to be more facile and affords better yields of products” (Robinson and Moorcroft 1970) – it without doubt is also unique in its synthetic versatility – have since been reinforced (Brossi et al. 1996; Greig et al. 1995a; Robinson 2002; Witkop 1998; Yu et al. 1993), it still being considered as offering the best synthetic route to lphysostigmine and its analogues.24 However, an opinion possibly to the contrary accompanied the publication of the latest useful synthesis of dl-physostigmine (Rege 2002; Rege and Johnson 2003) when it was stated (Rege and Johnson 2003) that “The first total synthesis of (-)-physostigmine [98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1] was achieved by Julian and Pikl in 1935. Since then a significant number of total syntheses of this alkaloid have been reported. Unfortunately, a practical and efficient

60

2 l-Physostigmine (Eserine)

synthesis for this seemingly simple molecule has not been reported to date”. However, any aspect of this final sentence that may allude to the synthetic approach of Julian and Pikl (Sect. 3.1) and its recent improvements16, 17, 20, 21, 22, 24 is clearly contentious and at variance with wide opinion to the contrary (vide supra).24

2.3.3

d-Physostigmine

Of the several early successful syntheses of l-physostigmine (Harley-Mason and Jackson 1954b; Julian and Pikl 1935b, c; Kobayashi 1938), only one, that of Teinosuke Kobayashi (1938), has been extended – via the synthesis of dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) and its resolution using either d- or l-tartaric acids – to use in a synthesis of d-physostigmine. However, in the report of this work, the salicylate of l-physostigmine, although shown to be identical with the corresponding salt of “natürlichen Eserins” (Kobayashi 1938), is quoted as having
“½α24 + 63,53  1,61
” (Kobayashi 1938) whereas it is well-established [for D example (British Pharmacopoeia 1968)] that it is laevorotatory, and the salicylate

of d-physostigmine is quoted as having “½α24 D 63,27  0,52 ” (Kobayashi 1938). Furthermore, since d-, l- and dl- physostigmines were not liberated from the salicylates, physical data for the free bases were not obtained (Kobayashi 1938) and therefore it was concluded that “Kobayashi’s work leaves much to be desired” when this synthetic problem was next addressed (Dale 1969; Dale and Robinson 1970). This next synthesis “is grosso modo identical with that of the natural alkaloid reported by Julian and Pikl [1935c]” (Brossi and Pei 1998). Thus, dl-5-hydroxy-1,3dimethylindolin-2-one (19, R1¼HO, R2¼H) (Longmore and Robinson 1967; Robinson 1965b) was sequentially O-ethylated, 3-cyanomethylated, catalytically hydrogenated and N-monomethylated – via formation of the benzylidene derivative – to afford dl-5-ethoxy-1,3-dimethyl-3-(2-methylaminoethyl)indolin-2-one [19, R1¼EtO, R2¼(CH2)2NHMe] as already described (Julian et al. 1934; Julian and Pikl 1935b, c). Since attempted resolution21 of this product with d-camphor-10sulphonic acid, as described earlier (Julian and Pikl 1935c), was unsuccessful (Dale 1969; Dale and Robinson 1970), it was reduced with sodium in ethanol boiling under reflux5 to yield dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) which was resolved, by the method of Kobayashi (1938) with d-tartaric acid, to give deserethole. This was then subjected to a reaction sequence that had already been used (Julian and Pikl 1935c, Polonovski and Nitzberg 1916) to convert its enantiomer into l-physostigmine, namely O-deethylation by boiling under reflux its solution in light petroleum (bp 60–80
C) in which powdered anhydrous aluminium chloride was suspended (cf Julian and Pikl 1935c),22 with subsequent conversion of the resulting d-eseroline (95, R¼H, X¼NMe) (for the absolute configuration, see Sect. 4) into d-physostigmine (95, R¼MeNHCO, X¼NMe) being effected in dry ethereal solution by reaction with methylisocyanate in the presence of a “speck” of sodium (cf Polonovski and Nitzberg 1916; Robinson 1968). However, nearly two decades later (Yu and Brossi 1988), following the observation that this earlier synthesis (Dale and Robinson 1970) involved “a tedious separation of optical isomers at an intermediate stage” and afforded “optically pure”

2.3 Syntheses of:-

61

d-physostigmine “in low yield only”, these two short-comings were obviated when the approach was partially modified (Yu and Brossi 1988). Thus, dl-N1-noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me) was reacted with S-l-1-phenylethylisocyanate and the resulting mixture of the two diastereomeric ureas was separated chromatographically (silica gel, CH2Cl2/MeOH),21 the faster-moving of which either in the presence of sodium pentoxide (prepared in situ) in boiling pentanol under reflux (Schönenberger and Brossi 1986) or upon treatment with sodium amylate in amyl alcohol (Brossi 1985) was converted into d-N1-noresermethole (95, R¼Me, X¼NH) (Brossi 1985, Schönenberger and Brossi 1986),21 isolated as its fumarate in high yield (Yu and Brossi 1988). This was then reductively N-methylated with formaldehyde and sodium borohydride to give a 62% yield of the fumarate of d-esermethole (95, R¼Me, X¼NMe) which, after O-demethylation with boron tribromide (Yu and Brossi 1988) to afford d-eseroline (95, R¼H, X¼NMe) – which has also been obtained by an alternative synthesis (Brossi and Pei 1998), was reacted in anhydrous Me

RO N

H Me

X

95 ethereal solution with a small piece of sodium and methylisocyanate to afford dphysostigmine (95, R¼MeNHCO, X¼NMe) (Brossi 1985; Brossi and Yu 1988a; Yu and Brossi 1988) {and with octylisocyanate, benzylisocyanate and phenylisocyanate under similar conditions and dimethylcarbamoyl chloride to give 95 [R¼Me (CH2)7NHCO (Brossi and Yu 1988a; Yu and Brossi 1988), PhCH2NHCO (Brossi and Yu 1988a; Yu and Brossi 1988), PhNHCO (Yu and Brossi 1988) and Me2NCO (Yu and Brossi 1988), respectively, X¼NMe]} and similarly were prepared 95 in which X¼NH, N-allyl and NCONHCH2Ph (Brossi and Yu 1988a). Indeed, it has been asserted (Brossi and Pei 1998) that “Unnatural (+)physostigmine and its analogs are best prepared today by a modification of the Julian” [Julian and Pikl] “total synthesis developed at Georgetown University in Washington, DC, during 1992-1994”. Thus, in accord with the objective that “A practical resolution of intermediates into the enantiomers at an early stage” (Pei et al. 1996) was effected, the chromatographic enantioseparation on a preparative scale of the Julian and Pikl nitrile of the O-methyl ether series 19 (R1¼MeO, R2¼CH2CN), using microcrystalline cellulose triacetate as the stationary phase and 96% ethanol as the mobile phase,21 sequentially eluted, in good yields, the l-(3aR)-enantiomer 96 followed by the enantiomer 97. However, a similar procedure failed to give

62

2 l-Physostigmine (Eserine)

Me

MeO

CH2CN N

O

Me

96

Me

MeO

CH2CN O

N Me

97 satisfactory results when using the nitriles 19 [R1¼EtO, PhCH2O, HO and tetrahydropyranyl22 O, R2¼CH2CN] although success attended the use of the ethyl ether that had been prepared by chiral phase-transfer catalysed asymmetric 3-cyanomethylation20 of 1,3-dimethyl-5-ethoxyindolin-2-one (19, R1¼EtO, R2¼H) and which was thereby enriched enantiomerically (70%) in the (3S)-enantiomer (Pei and Brossi 1995). The conversion of the enantiomer 96 into d-esermethole (95, R¼Me, X¼NMe) and thence into d-physostigmine (95, R¼MeNHCO, X¼NMe) was then accomplished by well-established reactions (Brossi and Pei 1998; Pallavicini et al. 1994; Pei et al. 1995a, b, 1996). The naturally-occurring l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) can, of course, be similarly synthesised from compound 97. Me

R 1O N R2

H

X

n

98

Although many of the other enantioselective syntheses that have been developed (Ashimori et al. 1993; Bruncko et al. 1994; ElAzab et al. 2000; Lee and Wong 1990 1991; Marino et al. 1992; Node et al. 1991; Takano et al. 1990a, b, 1991; Takano and Ogasawara 1989) (Sect. 3.2) in connection with the synthesis of l-physostigmine are also applicable to that of its enantiomer, there is no doubt that, in accord with much earlier opinion,24 the most satisfactory process is that based upon the pioneering and elegant approach of Julian and Pikl (Sect. 3.1) which, together with its subsequent improvements,16, 17, 20, 21, 22, 24 is also unique in its synthetic versatility.24

2.4 Absolute Configuration, Together with that of the. . .

2.4

63

Absolute Configuration, Together with that of the Other Structurally-Established Alkaloids of the Calabar Bean

In the tricyclic system of l-physostigmine {[α]D  74.5
(alcohol) (Petit and Polonovsky [sic] 1893), [α]D  75.8
(CHCl3) and 120
(C6H6) (Henry 1949; 25

Robinson 1964a), ½α17 D -76 (CHCl3) and ½αD -120 (C6H6) [Merck Index 2001(h), Muhtadi and El-Hawary 1989]}, the aromatic ring A forms a near-planar system with the adjacent ring B and thereby restricts the B/C ring junction to the thermodynamically more stable cis fusion (Hill and Newkome 1969; Jackson 1954; McFarland et al. 1969; Robinson 1963a; Spande et al. 1968; Taylor 1966; Witkop and Hill 1955). This deduction found early support from the X-ray crystallographic determinations of the structures and absolute configurations of echitamine (Hamilton et al. 1962; Manohar and Ramaseshan 1961), chimonanthine (Grant et al. 1965) and hodgkinsine (Fridrichsons et al. 1967, 1974), all of which9 contain the physostigmine ring system with the B/C rings cis-fused, and it was verified by the detection of an internal nOe between the protons of the 3a–methyl group and the 8a–proton in l-physostigmine (Newkome and Bhacca 1969). Thus, although the ring system contains two asymmetric centres, only one pair of enantiomers is possible [and have, indeed “ever been obtained” (Jackson 1954) during the synthetic construction of this tricyclic ring system when only “the natural isomer was obtained” (Newkome and Bhacca 1969)]. Therefore, the absolute configuration of the alkaloid is represented by either 95 (R¼MeNHCO, X¼NMe) or 98 (R1¼MeNHCO, R2¼Me, X¼NMe, n¼1). In fact, the (3aS,8aR)-configuration of the alkaloid, namely 98 (R1¼MeNHCO, 2 R ¼Me, X¼NMe, n¼1), was only settled 34 years after its total synthesis, which has been referred to as “an unorthodox phenomenon in our time” (Brossi et al. 1996), and 44 years after the establishment of its structure which is even more surprising, especially in view of its by then long-standing use and continuing emergence in clinical and laboratory practice (Chap. 10). Moreover, at this juncture it was to become the focus of attention in two distinctly separate laboratories. Thus, not for the first time, a problem that had remained dormant for decades was ultimately to receive the attention of two research groups working entirely independently; in this instance one in the UK (Longmore 1969; Longmore and Robinson 1969a, b) and the other in the USA (Hill and Newkome 1969; Newkome and Bhacca 1969). This was an international situation reminiscent of the early approaches towards the alkaloid’s synthesis by Robert Robinson and his research group and its actual first synthesis by Julian and Pikl (Sect. 3.1). However, in this later endeavour, both groups shared the honours – at least for l-physostigmine although, unfortunately, even this has not been recognised in at least one instance [Merck Index 2001(h)] – when, from mainly essentially different routes of chemical degradation to two different compounds of known absolute configuration which still contained the original C(3a)-atom of lphysostigmine, they verified this long-outstanding facet of its molecular structure as 98 (R1¼MeNHCO, R2¼Me, X¼NMe, n¼1), namely “(3aS-cis) [or (3aS,8aR) (Muhtadi and El-Hawary 1989)]-1,2,3,3a,8,8a-hexahydro-1,3a,8-trimethylpyrrolo

64

2 l-Physostigmine (Eserine)

[2,3b]indol-5-ol methylcarbamate (ester)” [Merck Index 2001(h)], by establishing – in both cases – the absolute configuration at C(3a). Both approaches, which have already been subjected to review (Robinson 1971), involved – following known procedures – the sequential conversion of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1), by reaction with ethanolic sodium ethoxide and ethyl-p-toluenesulphonate, into eserethole (98, R1¼Et, R2¼Me, X¼NMe, n¼1) which in ethereal solution was reacted with iodomethane L to afford (98, R1¼Et, R2¼Me, X¼ NMe2 I, n¼1) (Hoshino et al. 1934; Polonovski 1915). Treatment of this with 6 N–sodium hydroxide gave eserethole methine [99, R1¼EtO, R2¼H, R3¼OH, R4¼(CH2)2NMe2] (Hill and Newkome 1969; Hoshino et al. 1934; Longmore 1969; Longmore and Robinson 1969a, b; Polonovski and Polonovski 1918). The researchers in the UK (at the Victoria University of Manchester) (Longmore 1969; Longmore and Robinson 1969a, b), who were indebted to the Wellcome Foundation for a gift of 30 g of l-physostigmine sulphate with which to effect their investigations, then reduced this eserethole methine either by catalytic hydrogenation in acidic solution or with sodium borohydride to afford 99 [R1¼EtO, R2¼R3¼H, R4¼(CH2)2NMe2] which, via formation of its quaternary hydroxide, was subjected to Hofmann degradation to give 99 (R1¼EtO, R2¼R3¼H, R4¼CH¼CH2), catalytic hydrogenation of which yielded 99 (R1¼EtO, R2¼R3¼H, R4¼Et). Subsequent to unsatisfactory attempts to degrade its benzenoid ring using chromic acid (Longmore 1969), ozonolysis of this indoline followed by sequential treatment with hydrogen peroxide and hydrolysis with boiling hydrochloric acid oxide gave the moiety outlined in 99 as a β-aminoacid which was isolated as its 2,4-dinitrophenyl Me

R1

R4 R3 R2

N Me

99 derivative 100 by reaction with 2,4-dinitrofluorobenzene in the presence of excess sodium hydrogen carbonate, followed by column chromatography on silica gel using ether-chloroform (1:10v/v) as eluant. The absolute configuration of 100 (and hence that Me HOOC

Et

HN NO2

NO2

100

2.4 Absolute Configuration, Together with that of the. . .

65

of l-physostigmine) was determined by the synthesis of its enantiomer from l-3-ethyl3-methoxycarbonyl-3-methylpropionic acid (101, R1¼Me, R2¼COOH), whose absolute configuration was known to be as shown. This acid was converted, in boiling acetic anhydride under reflux, into its anhydride 102 which upon reaction with gaseous ammonia in anhydrous ether afforded the ether-soluble amide 101 (R1¼Me, 1 R2¼CONHL 2) and the ether-insoluble/water soluble ammonium salt 101 (R ¼Me, 2  R ¼COO NH4). After separation, the salt could be recycled and the amide was Me R1OOC

Et

R2

101

Me MeOOC O

Et

C O

2

102 treated with sodium hypobromite (Hofmann reaction), conditions which also effected hydrolysis of the ester group, to give the β-amino acid 101 (R1¼H, R2¼NH2) that was isolated, via reaction with 2,4-dinitrofluorobenzene in the presence of excess sodium hydrogen carbonate (vide supra), as its 2,4-dinitrophenyl derivative 101 [R1¼H, R2¼2,4-(di-NO2)C6H3NH] [cd maxima at 199 nm (Δε¼+1.0)], namely enantiomeric with 100 [cd maxima at 202.5 nm (Δε ¼ 1.4)]. This was further confirmed by the observation that the mp of a 1:1 (w/w) mixture of the samples was the same as that of the racemic modification of 100 which was prepared from dl-3ethyl-3-methoxycarbonyl-3-methylpropionic acid by an identical route to that employed in the above “optically active” synthesis. The pharmacological implications of this determination of the absolute configuration of l-physostigmine which “has thus further defined the stereochemical requirements at, or in the vicinity of, the AchE receptor-site [see Sects. 10.1.1 and 10.4] and may perhaps act as a guide in the future design of molecules with fixed stereochemistry for studies concerned with AchE inhibition” have been discussed (Longmore 1969).

66

2 l-Physostigmine (Eserine)

The investigations in the USA were ultimately published from Princeton University (Hill and Newkome 1969) and involved further sequential reactions of eserethole methine [99, R1¼EtO, R2¼H, R3¼OH, R4¼(CH2)2NMe2], initially by known procedures. Thus, oxidation with ferricyanide yielded 99 [R1¼EtO, R2+R3¼O, R4¼(CH2)2NMe2], the methiodide of which was subjected to Hofmann degradation to give 99 (R1¼EtO, R2+R3¼O, R4¼CH¼CH2)– along with recovered starting material 99 [R1¼EtO, R2+R3¼O, R4¼(CH2)2NMe2] and the alcohol 99 [R1¼EtO, R2+R3¼O, R4¼(CH2)2OH] (Hill and Newkome 1969), hydrogenation of the vinylic product then affording the indolin-2-one 99 (R1¼EtO, R2+R3¼O, R4¼Et) (Hill and Newkome 1969; Polonovski and Polonovski 1925c; Stedman and Barger 1925). This product was then O-deethylated by boiling under reflux in petroleum ether (bp 60–70
C) in which anhydrous aluminium chloride was suspended22 and the resulting phenol 99 (R1¼HO, R2+R3¼O, R4¼Et), together with 5-chloro-1-phenyltetrazole and anhydrous potassium carbonate, were boiled under reflux in acetone to afford the 1-phenyl-5-tetrazolyl ether 99 (R1¼1-phenyl-5-tetrazolylO, R2¼R3¼O, R4¼Et), hydrogenolysis of which in the presence of a 10% Pd-C catalyst yielded 99 (R1¼H, R2¼R3¼O, R4¼Et) (Hill and Newkome 1969). After several unsuccessful synthetic approaches, the absolute configuration of this last product, and hence that of l-physostigmine, was established (Hill and Newkome 1969) by the following synthesis of its enantiomer from l-2-methyl-2-phenylbutyric acid, the absolute configuration of which was known to be as shown in 103 (n¼0). This acid was subjected to Arndt-Eistert homologation to give 103 (n¼1), the acid chloride of which underwent Friedel-Crafts cyclisation to afford the indanone 104 (X¼O). Beckmann rearrangement of the oxime 104 (X¼NOH) of this ketone gave the dihydrocarbostyril 105 (R¼H) which was N-methylated to yield 105 (R¼Me) and then treated with methyllithium to afford 106. Ozonisation of the 2,3-double bond in this 1,4-dihydroquinoline followed by oxidative hydrolysis of the ozonide gave a mixture of the indolin-2-one 108, namely enantiomeric 99 (R1¼H, R2+R3¼O, R4¼Et), together with 107 – hydrolysis of 107 afforded further quantities of 108. A more direct synthesis of the indolin-2-one 108 was subsequently effected (Hill and Newkome 1969) from 103 (n¼0) by its conversion to the hydroxamic acid 109 which was cyclised to 108 by heating with polyphosphoric acid with subsequent N-methylation.

2.4 Absolute Configuration, Together with that of the. . .

Et

67 Et

Me (CH2)n

Me

103 ( n= 1)

COOH X 104 (X= O)

103 (n= 0)

Et

Me

105 (R= Me)

104 (X= NOH) O

N R 105 (R= H)

Et

Me

Et

Me

Et

COOH N

Me

Me 106

Et

Me

N COMe

N

Me 107

Me

O

108

Me CONHOH 109

Further details of the absolute configuration of l-physostigmine have arisen from the failure to detect an internal nOe between the N(1)-methyl group and the 8a– proton which indicated either that there is a completely trans relationship between these two features or that the former is undergoing rapid inversion under the experimental conditions, and from the detection of an internal nOe between the protons of the N(8)-methyl group and the 8a–proton [irradiation of the signal of the former causing a 15% increase in the integrated intensity of the signal of the latter], which was claimed to indicate that this methyl group is cis to the 8a–proton and is not undergoing rapid nitrogen inversion (Newkome and Bhacca 1969). However, the only conclusion that can be validly drawn from this latter observation is that the N(8)-methyl group and the 8a–proton are predominantly in a cis relationship at the unspecified (Newkome and Bhacca 1969) temperature of the solution under investigation [cf (Robinson and Moorcroft 1970)]. Clarification of this situation was forthcoming from the observation (Boardman and Robinson 1973) that irradiation

68

2 l-Physostigmine (Eserine)

of the signal of the N(8)-methyl protons caused in each case a 10% increase in the integrated intensity of the 8a–proton signal when examining the 1H–nmr spectra of a solution of l-physostigmine in CDCl3 at temperatures of 61
, 51
, 34
, 24
, 8
, +10
and +33
C, thereby supporting the theory that in this solution the N(8)methyl group is in a rigid conformation which is cis to the 8a–proton in the molecule (Boardman and Robinson 1973). Of the later investigators, none of whom made any reference to earlier crystallographic studies (Lovel 1953) on the alkaloid, and who sought fit to confirm the absolute configuration of l-physostigmine by X-ray crystallographic analysis [Brossi 1985 (see also Flippen-Anderson et al. 2002), Pauling and Petcher 1973, Petcher and Pauling 1973], two (Petcher and Pauling 1973) were surprisingly apparently also unaware – either through ignorance or for convenience – of the degradative studies and conclusions which had been published four years earlier (Hill and Newkome 1969; Longmore and Robinson 1969a, b; Newkome and Bhacca 1969) (vide supra) and other aspects of their publication (Petcher and Pauling 1973) also left much to be desired and has consequently been subjected to adverse criticism (Longmore and Robinson 1973). The absolute configurations of the other four alkaloids of currently known structures that have been isolated from the Calabar bean, namely l-physovenine (Chap. 3), l-eseramine (Chap. 4), l-N(8)-norphysostigmine (Chap. 5) and lgeneserine (Chap. 6), followed from comparison of their ord spectra with that of lphysostigmine (Dowley et al. 1966; Longmore 1969; Longmore and Robinson 1969a, b). All five alkaloids had closely similar spectra (Fig. 2.1), comprising a negative Cotton effect for the absorption band at circa 300 nm with a second, stronger negative Cotton effect, centred at circa 250 nm – further details of the spectra are presented in Table 2.1. For l-eseramine and l-physovenine, the Cotton effect at longer wavelength was observed only as an inflection on the steeply falling background dispersion curve, although the close similarity of all five spectra was obvious and clearly established. Only for l-N(8)-norphysostigmine were both extrema of the second Cotton effect observed (these arose from the absorption band at circa 250 nm) but in the other cases the negative extrema at circa 265 nm were clear. The observed Cotton effects, which showed a correspondence of sign for corresponding transitions throughout the series, showed that all five of the alkaloids have the same absolute configurations. This conclusion could also be reached independently since, by reactions that could not have caused optical inversion at C-3a, l-physostigmine has been converted into l-physovenine (Longmore 1969; Longmore and Robinson 1966) (Sect. 3.2.2), l-N(8)-norphysostigmine (Takano et al. 1990a) (Sect. 5.2) and l-geneserine (Bacchi et al. 1994; Nakagawa et al. 1975; Polonovski 1917; Yu et al. 1989, 2002) (Sect. 6.1), l-geneserine has been converted into l-physostigmine (Polonovski and Nitzberg 1915a, b; Yu et al. 1989) (Sect. 6.1), and l-eseramine and l-physostigmine have both been synthesised from the same asymmetric intermediate 98 (R1¼MeNHCO, R2¼Me, X¼NH, n¼1) (Yu et al. 1988a) (Sect. 4.2.2). The absolute configurations of the four minor alkaloids are therefore identical with that of l-physostigmine about their B/C ring junctions, with l-physovenine,

2.4 Absolute Configuration, Together with that of the. . . (λ) 250nm

42

40

69

312.5nm

38

36

34

32

400nm

30

28

25 (νin KK;1KK=1000 cm–1)

[M] × 10–3 (degrees)

–4 –8 –12 –16 –20

–24

Fig. 2.1. Optical rotatory dispersion spectra of alkaloids of the Calabar bean. –––––––– , l-physostigmine (rotations x4); . . .. . .. . . .., l-N(8)-norphysostigmine; •-•-•-•-, l-eseramine; ○-○-○-○-, l-geneserine and – – – –, l-physovenine (From Longmore and Robinson 1969b)

l-eseramine and l-N(8)-norphysostigmine thus having structures 98 (R1¼MeNHCO, R2¼Me, X¼O, n¼1) (Longmore 1969; Longmore and Robinson 1969b) – namely “(3aS-cis) [or 3aS,8aR]- 3,3a,8,8a-tetrahydro-3a, 8-dimethyl-2H-furo[2,3-b]indol-5-ol methylcarbamate (ester)” [(Merck Index 2001(i)], 98 (R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1) (Longmore 1969; Longmore and Robinson 1969b) and 98 (R1¼MeNHCO, R2¼H, X¼NMe, n¼1) (Longmore 1969; Longmore and Robinson 1969b) – namely (3aS,8aR)-1,2,3,3a,8,8a–hexahydro-1,3a–dimethylpyrrolo[2,3-b] indol-5-ol methylcarbamate (ester), respectively, and l-geneserine having structure 110 (R¼MeNHCO) – namely “(4aS-cis) [or 4aS,9aR]-2,3,4,4a,9,9a–hexahydro2,4a,9-trimethyl-1,2-oxazino[6,5-b]indol-6-ol methylcarbamate ester” (Merck Index 2001(b)], with the stereochemistry of the B/C ring junction in 110, which molecular Me

RO

4a 9a

N9

H Me

110

O

N

Me

1st Trough λnm [M]
344 520 350 400 355 7600 e e 333 6400b

1st Peak λnm [M]
336 480 328 100 332 4100 e e 333 6400b

2nd Trough λnm [M]
313 1560 303 1100b 313 14000b 318 17,200 313 11,300

2nd Peak λnm [M]
294 940 280 1900b 294 15800b 288 5700 297 10,700

3rd Trough λnm [M]
272 2180 258 3700 270 28000d 263 21,000 270 21,300

Crossover λnm [M]
250 0 252 0c – – – – 255 0

a

Notes. Rotations for physostigmine have been multiplied by 4. b Inflection. c Further data; 242 nm, [M] ¼ +3650
(peak); 230 nm, [M] ¼ 0
; 222 nm, [M] ¼ 1500
(lowest point observed). d Lowest wavelength of observation. e The spectrum for geneserine does not show a peak, trough or inflection in this region, though a marked change of slope occurs – typical values are 357 nm, [M] ¼ 7600
; 335 nm, [M] ¼ 13,000
. f An inflection occurs in this spectrum at 279 nm, [M] ¼ 15,000
(see Fig. 2.1).

Alkaloid l-Physostigminea l-N(8)-Norphysostigmine l-Physovenine l-Geneserine l-Eseraminef

Table 2.1. The salient features of the ord-spectra of alkaloids of the Calabar bean [from (Longmore and Robinson 1969b)].

70 2 l-Physostigmine (Eserine)

2.5 Biogenesis

71

models indicate could be either cis or trans-fused, being established in favour of the former by the detection of an internal n0e between the protons of the C(4a) methyl group and the 9a–proton (Robinson and Moorcroft 1970), a conclusion further supported by a less direct nmr spectroscopic analysis (Riddell et al. 1970). Furthermore, the detection of an internal n0e between the 9a–proton and the N(9)–methyl group’s protons also established that the N(1)-methyl group and the C(2)-proton of the indoline system are predominantly in a cis relationship at the room temperature of 20
C (Robinson and Moorcroft 1970), similar to the situation inherent in l-physostigmine (vide supra). Moreover, in the spectrum of l-geneserine, irradiation at the frequency of the singlet at 7.47τ (caused by the N-methyl group of the tetrahydro-1,2oxazine ring) had no effect upon the intensity of the 9a–proton signal, indicating that these two are probably not close to each other (Robinson and Moorcroft 1970). In view of the latest (Yu et al. 1989) revision to the structure of l-geneserine (Sect. 6.1), its absolute configuration should probably now be expanded to include along with 110 (R¼MeNHCO) a contribution from 98 [R1¼MeNHCO, R2¼Me, X¼N(! O)Me, n¼1].

2.5

Biogenesis

Although an O-carbamyl group was later shown to be present in the antibiotics novobiocin (111) [Kominck 1967, Merck Index 2001(d)] and streptothricins F, E, D, C, B, A and X (112, n¼1–7, respectively) [Merck Index 2001(n)]{streptolin was Me Me MeO

O

O

O

O Me

H2NCOO

Me

O

Me

N OH H OH

111

H H N

HO H2NCOO

OH

O

N H

NH

N H H OH

OH HN O

O

NH2

112

N H

H

n

72

2 l-Physostigmine (Eserine)

formerly the name used for streptothricins either E or D [Merck Index 2001(n)]} and in the biological carbamyl group donor carbamyl phosphate (113) (Lehninger 1970; Mahler and Cordes 1966) (vide infra), as was an N-carbamyl group in the biosynthetic intermediates citrulline (114) (Lehninger 1970) and N-carbamyl aspartic acid (115) (Lehninger 1970; Mahler and Cordes 1966), l-physostigmine O H2NCOO P O O

113

H2NCONH(CH2)3 CH COOH 114

NH2

H2NCONH CH CH2COOH COOH

115 (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) is the first naturally-occurring compound that was shown to contain a carbamyl group – in this instance N-methylated. Indeed, not only in the higher plant alkaloids but, furthermore, in natural products in general, the presence of the N-methylcarbamyl group [for which carbamyl phosphate (113) (Lehninger 1970; Mahler and Cordes 1966) is the likely biosynthetic source] is, as yet, a unique feature in the structures of l-physostigmine and of the four minor alkaloids, namely, l-physovenine, l-eseramine, l-N(8)-norphysostigmine and lgeneserine (Chapters 3, 4, 5 and 6, respectively), associated with it in the Calabar bean and which are currently the only ones of established structures (Chap. 8). However, the presence of an N,N-dimethylcarbamyl group has also been inferred (Sect. 8.4) in two of the other alkaloids, namely calabatine and calabacine, as yet of unknown structure, reported (Döpke 1963) as being isolated from this source. When, based upon a duly acknowledged suggestion by “Professor R. Robinson, F.R.S.” (then at the University of Manchester),5 structure 3 (R1¼MeNHCO, R2¼Me, X¼NMe) for physostigmine was first proposed (Robinson 1925) and published by Stedman and Barger (1925) (Sect. 2) from the Department of Medical Chemistry at the University of Edinburgh (see footnote 14 in Chap. 3), it was accompanied by the supportive statement that “If the assumption is made that the [3] methylation of an indole nucleus may proceed in the plant in the manner in which it is known to take place in the laboratory, a relation between physostigmine and this amino-acid [tryptophan (116, R1–R3¼H, R4¼COOH, R5¼NH2)] at once becomes evident. By decarboxylation and methylation, followed by a ring

2.5 Biogenesis

73

R1

R4 R3 N

R5

R2

116 closure, the ring system of physostigmine would be readily formed”. These principles of this biogenetic postulation were reiterated by Robert Robinson 30 years later in his classical pioneering study upon the biogenesis of natural products (Robinson 1955) in which he states “The chief alkaloidal constituent of the Calabar bean (Physostigma venenosum Balf.) is of great interest on account of the extra methyl group in position 3. Physostigmine has the structure” [3, (R1¼MeNHCO, R2¼Me, X¼NMe)] “and is evidently related to 5-hydroxytryptamine” [(116, R1¼HO, R2–R4¼H, R5¼NH2)]. “N-methylation, C-methylation, and formation of the urethane grouping applied to” [116, (R1¼HO, R2–R4¼H, R5¼NH2)] “could yield” [117], MeNHCOO

Me N HN

117

Me

“which with one further N-methylation could afford” [3, (R1¼MeNHCO, R2¼Me, X¼NMe)], and appends “In accordance with modern biochemical findings the methyl groups could be transferred from a suitable –SMe reagent, and, in the laboratory, analogies for the β-C-methylation are not wanting”. Although with the minor further elaboration that methionine is one of several donors (Lehninger 1970) probably responsible for the C-methyl group – and presumably also possibly the source of the other methyl groups, and that tryptophan arises from anthranilate via indole (Parry 1992; Wenkert 1959), Robert Robinson’s above biogenetic predictions for l-physostigmine remain currently acceptable [see, for examples (Cordell 1981; Muhtadi and El-Hawary 1989; Rahman and Basha 1983(a), Wenkert 1959)]. Indeed, a similar scheme has been postulated (Robinson 1964a for the biogenesis of l-physovenine (Sect. 3.3) with the proviso that transformation of the aminoethyl side-chain to a hydroxyethyl side-chain [probably by dehydrogenation (2H) of the amine to the imine, followed by hydrolysis with subsequent reduction (+2H) to the alcohol] (Udenfriend 1956) occurs before the formation of the tetrahydrofuro ring C. However, an alternative prognostication (Julian and Pikl 1935c), whilst retaining 5-hydroxytryptophan (116, R1¼HO, R2¼R3¼H, R4¼COOH, R5¼NH2) as the biosynthetic source, and in view of Kotake’s observation 5 years previously that tryptophan was converted into kynurenine via its oxidation [“which might occur through certain oxidases” (Julian et al. 1935)] into “oxytryptophan” [presumably (118, R1–R3¼H, R4¼COOH)], suggested that l-physostigmine “might be built up

74

2 l-Physostigmine (Eserine)

by the plant” (Julian and Pikl 1935c) via a similar oxidation, followed by decarboxylation and methylation to afford 118 (R1–R3¼Me, R4¼H) which then undergoes R1 O HN

N

R4

R3

R2

118 reductive cyclisation to afford 23 (R1¼R2¼H, R3¼R4¼Me) (Julian and Pikl 1935c) and, presumably via the appropriate hydroxylation of the aromatic ring and carbamylation at some stage, ultimately l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1). This postulation, involving the C(3)-methylation of the indolin-2-one 118 (R1¼H) followed by reductive cyclisation, finds laboratory analogies in what then was to become the pioneering and elegant synthesis of the alkaloid by Julian and Pikl (Sect. 3.1).24 Indeed, with regards to this, Julian and Pikl (1935c) have stated that “we believe that the route we have taken presents in its essential stages the phytochemical mechanism for the production of this substance”. The observation – from earlier investigations – “that in animal [however, see also (Udenfriend et al. 1956)] metabolism tryptophan is oxidised to 5-hydroxytryptophan which then undergoes decarboxylation to 5-hydroxytryptamine (serotonin)” – observations that were subsequently (Udenfriend et al. 1956) confirmed and extended to the formation (from DL-tryptophan) of dehydrobufotenine (206)6 in the toad Bufo Marinus and 5-hydroxyindole-acetic acid in a patient with malignant carcinoid and likewise from L-tryptophan in the intact dog – raised the question (Witkop and Hill 1955) “whether a similar sequence of reactions may not be operative in the plant for the synthesis of 5-hydroxyindole alkaloids”. This led to the return, as a possible biogenetic precursor of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1), of tryptophan (116, R1–R3¼H, R4¼COOH, R5¼NH2) which in this instance it was hypothesised (Witkop and Hill 1955) might undergo, via formation of 5-hydroxytryptophan (116, R1¼HO, R2¼R3¼H, R4¼COOH, R5¼NH2), C(3)-hydroxymethylation [in accord with the earlier (King and Robinson 1932a) reflections “that the vegetable methylating agent is generally considered to be formaldehyde or an equivalent and the C-alkylation of indoles has not, so far as we are aware, been accomplished by such means. Nevertheless, formaldehyde can be used to methylate amines, and the hetero-enoid systems (N-C¼C) contain carbon atoms which are analogous in quality of reactivity to the nitrogen atom of amines; there is every probability, therefore, that the C-methylation of indoles by means of formaldehyde could be experimentally realised”] to afford 119. This, upon internal CH2OH

HO

COOH

N HN H

H

119

2.5 Biogenesis

75

cyclisation as shown, would afford the eserine ring system-containing amino acid 120 (R1¼HO, R2¼R3¼H, R4¼COOH) which, after hydrogenolysis of the CH2OH

R1 N H

N

R4 R3

R2

120 hydroxymethyl group, methylation of the two secondary nitrogen atoms and decarboxylation, would afford eseroline and, ultimately, l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) – where thought necessary, analogies for the reactions were noted (Witkop and Hill 1955), including that for the ultimate C(3)methylation step involving the then recently reported C-methylation of dihydroberberine with formaldehyde (Bersch 1950). Moreover, attempts were subsequently made (Dolby and Furukawa 1963) to simulate in vitro this postulated biogenesis by treating several 3-indolylacetamides 116 (R1¼R2¼H, R3+R4¼O, R5¼NHCH2Ph and NHcyclohexyl, respectively) with formaldehyde under various conditions, the amides rather than the tryptamines being used in order to decrease the basicity of the amine nitrogen atom and thereby hopefully diminish the formation of 1,2,3,4-tetrahydro-β-carbolines. However, extensive polymerisation occurred during these reactions and none of the required products 120 (R1¼H, R2¼CH2Ph and cyclohexyl, respectively, R3+R4¼O) were isolable from these reactions (Dolby and Furukawa 1963), although, using the Vilsmeier reaction, electrophilic attack apparently occurred at the 3-position in 116 (R1¼R2¼H, R3+R4¼O, R5¼NHCH2Ph and NHcyclohexyl, respectively) using N,N-dimethylacetamide with phosphorous oxychloride followed by the usual alkaline hydrolysis, but the resulting intermediate 3H–indolium cations underwent rearrangements to afford the quinolines 121 (R¼CH2Ph and cyclohexyl, respectively) as the only recognisable products (Dolby and Furukawa 1963). R

OH Me

N

N

121 Partially as a prerequisite to a biosynthetic study on the alkaloids of the Calabar bean, the 13C–nmr spectra of l-physostigmine (Chap. 2), l-physovenine (Chap. 3), leseramine (Chap. 4) and l-N(8)-norphysostigmine (Chap. 5) were examined and their full assignment effected (Crooks et al. 1976) (Chap. 7). However, further studies are awaited into the biosynthesis of these alkaloids involving the feeding of appropriately-labelled potential precursors to an apposite biosystem – perhaps an l-physostigmine-producing microorganism (Iwasa et al. 1979, 1981, Murao and

76

2 l-Physostigmine (Eserine)

Hayashi 1986), although this, not being a plant, is therefore not strictly “alkaloid”-producing.4 Nevertheless, meanwhile it has been verified that tryptophan (116, R1–R3¼H, 4 R ¼COOH,R5¼NH2) is an alkaloidal biosynthetic precursor of the 1,2,3,3a,8,8a–hexahydropyrrolo[2,3-b]indole ring system. Thus, following the recognition of this tricyclic ring system in l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1), its presence has also been detected in a diverse range of natural products,9 amongst which are the dimeric alkaloids chimonanthine (122, R1¼R2¼H), meso-chimonanthine,9 calycanthidine (122, R1¼H, R2¼Me)

*

Me R2 H N N

N R1

N H

*

Me

122 and folicanthine (122, R1¼R2¼Me)9 and which have all been isolated from plants in the botanical order of Calycanthaceae, a small family comprised only of the genera Calycanthus and Chimonanthus (Cordell 1981) – and the enantiomer of chimonanthine, d-chimonanthine, which has been isolated from the skin of Phyllobates terribilis (the Columbian poison-dart frog) (Hino and Nakagawa 1988; Tokuyama and Daly 1983).9 It is still generally accepted (Fridrichsons et al. 1974; Gorman et al. 1971; Hall et al. 1967; Hesse 1981; Kutney 1977; Rahman and Basha 1983b – see also c, Robinson 1963a; Scott et al. 1964; Taylor 1966; Woodward et al. 1960) that these alkaloids are biosynthesised via a ββ’-oxidative dimerisation of two tryptamine units (123), as originally proposed (Robinson 1955; Robinson and Teuber 1954), to afford the intermediate 124 which then undergoes

two intramolecular ring closures as shown, with in vitro syntheses of chimonanthine (Hendrickson et al. 1962; Kutney 1977; Scott et al. 1964) and dl- and mesochimonanthine (Hall et al. 1967; Hendrickson et al. 1962; Kutney 1977; Taylor 1966) apparently substantiating this hypothesis. Initial tracer work (Schütte and Maier 1965) involved feeding DL-[2-14C] tryptophan to Calycanthus floridus

2.6 Ultraviolet Absorption Spectrum and Reaction in an Acidic Medium

77

but the results leave much to be desired since, although incorporation of the radioactive label into chimonanthine, calycanthidine and folicanthine was claimed, all three alkaloids lacked positive identification and necessary degradative work was not effected. However, these deficiencies were rectified the following year (O’Donovan and Keogh 1966) when, following a similar feeding, radioactive folicanthine (122, R1¼R2¼Me) was isolated and shown to be labelled only on the starred carbon atoms by the following degradation. Treatment of the radioactive alkaloid with concentrated hydrochloric acid at room temperature for 4 days afforded 116 (R1¼R3¼R4¼H, R2¼Me, R5¼NHMe) which, via formation of the L methiodide 116 (R1¼R3¼R4¼H, R2¼Me, R5¼ NMe3 I), was subjected to Hofmann degradation to yield radioactive 1-methyl-3-vinylindole. This, upon reaction with osmium tetroxide gave an osmate ester which was decomposed by alkaline sodium sulphite followed by cleavage of the resultant glycol with potassium periodate to give radioactive formaldehyde (isolated as its dimedone derivative) and 3-formyl-1-methylindole, The inactivity of this last product and the concordance between the specific activities of the formaldehyde dimedone derivative, the interL mediate degradation product 116 (R1¼R3¼R4¼H, R2¼Me, R5¼ NMe3 I) and the radioactive folicanthine verify that the latter compound was only labelled on the starred carbon atoms as shown in 122. This result clearly established the involvement of two molecules of tryptophan in the biosynthesis of folicanthine (O’Donovan and Keogh 1966).

2.6

Ultraviolet Absorption Spectrum and Reaction in an Acidic Medium

Early preliminary studies (Brustier 1926) concerning uv light absorption by alkaloids reported that “L’ésérine pure cristallisée présente en solution alcoolique un spectre de trios bandes: I, s’étend de λ ¼ 2254,8 à λ ¼ 2327,5; II, va de λ ¼ 2348,3 à λ ¼ 2664,7 et, dans sa partie moyenne de λ ¼ 2625,8 à λ ¼ 2382,2; III, va de λ ¼ 2783,8 à λ ¼ 3370,9 et, dans partie moyenne, de λ ¼ 2895,1 à λ ¼ 3306,5; la couche présente trios points de rebroussement: le 1er entre λ ¼ 2246,7 et λ ¼ 2254,8; le 2e entre λ ¼ 2343,5 et λ ¼ 2338,1; le 3e à λ ¼ 2767,5” and conclude (Brustier 1926) that this spectrum resembles those of indole, tryptophan and skatole which, by thus suggesting the presence of an indolic chromophore in the alkaloid, was at variance with the structural studies that were by then approaching fruition (Sect. 2.2). A later investigation (Coyne and Paterson 1961) reported that in the uv spectra of “solutions containing physostigmine and eserethole . . .. . . .. one peak only was observed at 245mμ and one at 300-310mμ” but, whilst recognising that “A change of pH of the solutions did not have any effect on separation of the maxima”, failed to perceive that the indoline-type spectrum of physostigmine (and also of eserethole) (vide infra) was retained in dilute acidic solution though with a hypsochromic shift of about 10 nm for the whole spectrum relative to that in neutral solution (Fig. 2.2) (Hodson and Smith 1957) and when it was also recognised that “In acid solution the absorbing species [for l-physostigmine] is a cation” 125 “in which the formal positive charge on N(b) [N(1)] has rendered N(a) [N(8)] virtually non-basic: the

78

2 l-Physostigmine (Eserine) 2.

Log e

1.

3.

4.0

4.0

4.0

3.5

3.5

3.5 c

a 3.0

3.0

b

d

3.0

e f

2.0 200

250

300

350

200

250 300 Wavelength (mm)

200

250

300

350

Fig. 2.2. Uv absorption spectra of 1. physostigmine in (a) EtOH and (b) EtOH-HCl, 2. deoxydihydroajmaline (127) in (c) EtOH and (d) EtOH-HCl, and 3. 1-methyl-3(2-methylaminoethyl)indoline (9, R1¼R2¼H, R3¼Me) in (e) EtOH and (f) EtOH-HCl [from (Hodson and Smith 1957)].

N(a)-electron pair is thus still able to resonate with the benzene ring, with retention of indoline-type absorption. The hypsochromic shift must be a result of the closeness of the positive charge on N(b) to the mesomeric system” (Hodson and Smith 1957). It was later shown (Robinson 1958) that dl-desoxybisnoreseroline (23, R1–R4¼H) (Hoshino 1932a, b; Hoshino and Tamura 1932), dl-iso-desoxynoreseroline (23, R1¼R2¼R4¼H, R3¼Me) (Hoshino and Kobayashi 1935), dl-desoxynoreseroline (23, R1–R3¼H, R4¼Me) (Julian and Pikl 1935a) and dl-desoxyeseroline (23, R1¼R2¼H, R3¼R4¼Me) (Julian and Pikl 1935a) also exhibited this type of spectroscopic behaviour (Table 2.2). Moreover, further decreases in pH eventually caused cleavage of the C(8a)-N(1) bond as shown in 125 – which “began in approximately 1M-hydrochloric acid and was essentially complete in 5M-acid” (Jackson and Smith 1964) – to afford L the corresponding 3H–indolium cation 126 (R1¼Me, R2¼ NH2Me) (Ahmed and Robinson 1967; Jackson and Smith 1964), a species which has been suggested (Robinson and Robinson 1968) as that responsible for the antiAchE activity of l-physostigmine L (Sect. 10.4). Evidence for this cleavage of ring C to give 126 (R1¼Me, R2¼ NH2Me) was forthcoming from uv spectroscopic data which indicated, for “physostigmine”, “esermethole” and “deoxynoreseroline” the change from the indoline (in 95% EtOH) to the 3H–indolium cation (in 6 M–ethanolic HCl) chromophore, and from 1H–nmr spectroscopic investigations of the same three bases dissolved in strongly acidic media, namely in trifluoroacetic acid – and with the latter base also in 4 M–hydrochloric acid – which indicated in particular the low-field resonance of the C(2)-proton of the 3H–indolium cation nucleus (Jackson and Smith 1964).

2.6 Ultraviolet Absorption Spectrum and Reaction in an Acidic Medium

79

Table 2.2. Uv absorption spectra of some analogues of l-physostigminea in neutral and dilute acidicb media [from (Robinson 1958)].

Compound dl-Desoxybisnoreseroline (23, R1–R4¼H) dl-iso-desoxynoreseroline (23, R1¼R2¼R4¼H, R3¼Me) dl-desoxynoreserolinea (23, R1–R3¼H, R4¼Me)

λmax. in nm (ε) in EtOH 299 (2670) 244 (8660) 299 (2850) 244 (9520)

Hypsochromic shift (nm) on acidificationb (ε of new λmax.) 7 (2790) 6 (9250) 7 (2610) 6 (9750)

λmin. in nm (ε) in EtOH 268–9 (590) 221–2 (2870) 268 (490) 221–2 (2900)

Hypsochromic shift (nm) on acidificationb (ε of new λmin.) 9 (590) 5 (3210) 8–9 (290) 5 (3220)

306–7 (2960) 253 (11,440) 303–4 (3010) 251 (11,500)

9 (2750) 8 (12,010) 9 (2490) 8 (11,100)

277–8 (870) 224–5 (2530) 275 (800) 224–5 (2780)

11–12 (410) 3 (2480) 10 (430) 4 (2550)

dl-desoxyeseroline (23, R1¼R2¼H, R3¼R4¼Me) Notes. a The uv spectroscopic data for “deoxynoreseroline” has also been reported (Jackson and Smith 1964) as λmax. (nm) (log ε) 302 (3.44) and 252 (4.02) in 95% EtOH, λmax. 298 (3.40) and 246 (3.97) in 0.01 M–ethanolic HCl. Similar hypsochromic shifts have also been noted in the uv spectra of “esermethole” and “physostigmine” (Jackson and Smith 1964) and of other alkaloids – and various of their transformation products – containing a Ph-N-C-N system (Jackson and Smith 1964; Joule and Smith 1962). b Effected by the addition of one drop of concentrated hydrochloric acid to the cell (1 cm  1 cm  3 cm).

When the protonated more basic nitrogen atom was removed from the indoline nitrogen atom by two carbon atoms as, for example, in deoxydihydroajmaline (127) OH N N H Me H H

127

H Me Et H

(Hodson and Smith 1957; Joule 1983), its protonation still prevented the protonation of the latter so the compound was a monoacidic base, the indoline nitrogen atom having become virtually non-basic as a result of the proximity of the positive charge in the cation and, consequently, the indoline-type uv absorption was retained in acidic solution but the positive charge was sufficiently far removed so as not to affect the resonance of the mesomeric system of the indoline nucleus and there was therefore no appreciable shift in the absorption bands (Fig. 2.2.3) (Hodson and Smith 1957) – ε1hexahydrofluorocurine (128) (Bickel et al. 1955; Hesse et al. 1964) behaved in a

80

2 l-Physostigmine (Eserine)

OH OH H N Me Me

N

128

Et

similar manner (Bickel et al. 1955; Hodson and Smith 1957). To complete the scenario, the indoline 9 (R1¼R2¼H, R3¼Me) was a diacidic base in dilute aqueous acid since protonation of the indoline nitrogen atom is now not prevented by the positively charged side-chain nitrogen atom that is now four carbon atoms distant. Thus, its indoline-type uv absorption in neutral solution is replaced upon acidification by the benzenoid absorption of the anilinium cation (Fig. 2.2.3) (Hodson and Smith 1957). These studies have been further elaborated upon34. Furthermore, the effect of structural factors upon the ring-chain tautomerism in a wide range of compounds containing the physostigmine ring system has been investigated using uv and 1H–nmr spectroscopy (Grandberg et al. 1970).

2.7

Mass Spectrum

The fragmentation of physostigmine (3, R1¼MeNHCO, R2¼Me, X¼NMe) upon electron impact has been investigated, interpreted and reviewed by several groups (Bild and Hesse 1967; Budzikiewicz et al. 1964; Clayton and Reed 1963; Hino and Yamada 1963; Muhtadi and El-Hawary 1989; Robinson 1968; Spiteller and SpitellerFriedmann 1963, 1964). The decomposition pathways together with the resulting spectrum were predominantly controlled by the results from the formation L of relatively stable aromatic ions. Below the molecular ion (m/e 275), the small M -1 peak at m/e 274 was attributed to the loss of the hydrogen atom from the tertiary C-8a to afford 129 in which the positive charge could be delocalised between the two nitrogen atoms as shown. The base peak (m/e 218) of the spectrum was assigned to the radical cation of eseroline (130, X¼NMe) and could arise from the expulsion of the elements of methyl isocyanate as schematically illustrated although still to be established (Budzikiewicz et al. 1964) is the source of the hydrogen atom gained by 130 (X¼NMe). This lost its angular methyl group to give 131 (R¼Me) (small peak at m/e 203), from which loss of a further methyl radical gave rise to another small peak at m/e 188. An alternative mode of fission in 130 (X¼NMe), involved cleavage of the C(8a)N(1) bond to afford 132 which, by homolysis of the original C(2)-C(3) bond led to 133 (medium peak at m/e 175). This underwent ring enlargement to 134 which lost a hydrogen atom to yield the very stable quinolinium ion 135 (R1¼R2¼Me) (intense peak at m/e 174). The intermediate 132 also underwent fission of the original C(3)C(3a) bond as indicated leading to the indolic radical ion 136 (R¼Me) (strong peak at m/e 161) from which loss of a hydrogen atom, to afford the intermediate 137, followed by ring-enlargement as shown gave the quinolinium ion 135 (R1¼H, R2¼Me) (intense peak at m/e 160).

2.7 Mass Spectrum

81

H Me N

C O

Me

O

H Me

N

Me

O

N C O

N

H Me Me

3 (R1= MeNHCO, R2= Me, X= NMe)

N

N

Me

Me

129

-MeN=C=O Me

HO

HO N

N X H Me

R

130 (X= NMe)

N

H

Me 131

H Me

HO

CH2

HO

N

N

N

Me

Me

R

132

136

H Me

HO

CH2

CH

HO

N

N

Me

Me 137

133

R1

Me HO

H

HO

N Me 134

135 (R1= R2= Me)

N R2 1

135 (R = H, R2= Me)

82

2 l-Physostigmine (Eserine)

Alternatively, 137 could be derived by cyclic collapse of 131 as shown (Budzikiewitz et al. 1964). The necessary isotopic labelling experiments have yet to be effected to establish which of these two possible origins of 137 is correct (Budzikiewitz et al. 1964), although the sequence 131 to 137 is strongly favoured (Budzikiewitz et al. 1964) since ions analogous to 137 were found in the mass spectra of the calycanthaceous alkaloids chimonanthine (122, R1¼R2¼H), calycanthidine (122, R1¼H, R2¼Me) and folicanthine (122, R1¼R2¼Me) (Budzikiewitz et al. 1964, Clayton et al. 1962; Hino and Yamada 1963), in all of which the angular C(3a)-methyl group is absent (Budzikiewicz et al. 1964). The mass spectra of related synthetic 1,2,3,3a,8,8a–hexahydro-3a-methylpyrrolo [2,3-b]indoles (Budzikiewitz et al. 1964, Clayton and Reed 1963; Pande et al. 1979; Zhigulev et al. 1972) and 3,3a,8,8a–tetrahydro-3a-methyl-2H-furo[2,3-b]indoles (Clayton and Reed 1963) and some of their derivatives (Budzikiewicz et al. 1964; Clayton and Reed 1963; Pande et al. 1979) and homologues (Zhigulev et al. 1972) have also been investigated and studies in these areas by II Grandberg and his colleagues have been alluded to in review (Grandberg 1974). The chemical ionisation (CI) mass spectrum of physostigmine has been measured using methane as reactant gas and discussed with particular reference to its conventional electron impact mass spectrum (Fales et al. 1970). In addition, the fast atom bombardment (FAB) mass spectrum of this alkaloid and some of its carbamate analogues showed (Rubino and Zecca 1991) a partially different fragmentation process with respect to that of the corresponding CI fragmentation and, interestingly, strongly suggests (Rubino and Zecca 1991) that the observed protonated molecular species in the sample undergoing FAB are more likely to be represented by the corresponding ring C-opened 3H–indolium cations, for example 126 (R1¼Me, R2¼NHMe) (Sect. 6), which have been conjectured (Dale and Robinson 1970; Robinson and Robinson 1968) (Sect. 10.4) as being the pharmacophores of this group of alkaloidal drugs.

2.8 2.8.1

Detection, Assay and Instability Qualitative and Quantitative Analysis

Crystalline salts of l-physostigmine which have been referred to in review (Henry 1949) include the aurichloride (see also Saxton 1960) (mp 163–165
C), benzoate (mp 115–116
C), hydrobromide (mp 224–226
C), mercuric iodide derivative (mp 170
C), picrate (mp 114
C), platinichloride (mp 180
C), salicylate (mp 186–187
C) [mp 185–187
C; about 184
C (Muhtadi and El-Hawary 1989)] and sulphate (deliquescent) (mp 145
C) [mp 140
C; about 143
C (Muhtadi and

2.8 Detection, Assay and Instability

83

El-Hawary 1989)]. The microcrystals, formed by treatment of the alkaloid with various reagents – which include gold chloride, lead iodide, Mayer’s reagent, picric acid, potassium permanganate and Wagner’s reagent – have been illustrated in review (Muhtadi and El-Hawary 1989). Studies describing the behaviour of lphysostigmine upon paper, thin-layer, column, gas and high performance liquid chromatography have been reviewed (Muhtadi and El-Hawary 1989; Robinson 1964b), as have methods for its detection and quantitative determination (Marion 1952; Muhtadi and El-Hawary 1989; Robinson 1964b; Saxton 1960). The alkaloid has also been assayed pharmacologically [by use of it’s antiAchE activity (Sect. 10.2)] (Marion 1952; Vincent and Maugein 1942a, b). However, because of the non-specific nature of the methods available, a reliable quantitative estimation of l-physostigmine in the presence of its decomposition products was not possible until it was found that, in this situation, the alkaloid can be assayed by its reaction with sodium nitrite in acidic solution to afford quantitatively a yellow nitroso-compound – of unknown structure! – that is then extracted with chloroform and estimated by the measurement of its absorbance at 417 nm (Fletcher 1968; Fletcher and Davies 1968) – this work has been included in a review of methods, including spectrophotometric, of analysis for l-physostigmine (Muhtadi and El-Hawary 1989). At the earlier juncture it was also found (Fletcher and Davies 1968) that the rate of decomposition of aqueous solutions of physostigmine sulphate exposed to γ-radiation is independent of the pH of the solutions, whereas the rate of pyrolytic decomposition of such solutions is at a minimum between pH 2.2 and 3.0. The presence of sodium metabisulphite retards the rate of decomposition of the solutions exposed to γ-radiation but has no effect upon the decomposition rate caused by pyrolysis. For solutions containing physostigmine that are sterilised by heat, a solution of pH 3.0, which is too low for ophthalmic preparations unless the solution is buffered with a low-capacity system, is required for maximum heat stability. To overcome these difficulties it is suggested that such solutions could be sterilised by exposure to γ-radiation, under which conditions the rate of decomposition is independent of the pH of the solution. The initial formulation of eyedrops would contain the required excess of physostigmine to take into account its partial decomposition during irradiation. However, this technique assumes (Fletcher and Davies 1968), without any experimental verification, that the radiation decomposition products of physostigmine which, apart from deoxyeseroline (23, R1¼R2¼H, R3¼R4¼Me) that is formed in 10% yield by “a 24 h irradiation of a 3.9  10-3M 2-propanol solution of [physostigmine] with the 300nm low pressure lamps of the Rayonet reactor” (Travecedo and Stenberg 1970), have yet to be structurally identified, are devoid of antiAchE activity. The extraction of l-physostigmine from biological samples and its qualitative and quantitative analysis have been recently the subjects of review (Zhao et al. 2004).

84

2.8.2

2 l-Physostigmine (Eserine)

Rubreserine

Only two of the salts (vide supra) of l-physostigmine are used in medicine, namely the salicylate {known by the generic names Physostigmine Salicylate, Eserine Salicylate and Physostol Salicylate (Muhtadi and El-Hawary 1989), and by the trade names Antilirium (Merck 2001(h); Muhtadi and El-Hawary 1989 (injection); Taylor 1996 (injection); Triggle et al. 1998); Isopto (Moroi and Lichter 1996; Muhtadi and El-Hawary 1989 (ophthalmic solution, with chlorobutanol and sodium sulphite); Taylor 1996 (ophthalmic solution); Triggle et al. 1998) and Synapton (Forest Laboratories, St Louis, MO (slow release formulation) (Luo et al. 2005b; Luo et al. 2006))} and the sulphate [available in an ophthalmic ointment (Taylor 1996)]. However, since the alkaloid’s introduction into clinical practice, a considerable pharmaceutical problem has been posed by the instability of its salts in aqueous solution and thereby its limited efficacy probably related to its relatively short half-life. Thus, the salicylate, despite the claim (Lauter and Foote 1955) that it “appears to be especially stable toward oxidation” and the sulphate, like the alkaloid, are colourless but on standing – especially in alkaline solution – exposed to light and air all three soon begin to assume a pink colouration that ultimately darkens to red. Indeed, since its first isolation (Hesse 1867, Jobst and Hesse 1864) it has been known that the alkaloid in its crystalline form and in solution – especially alkaline – becomes red coloured, this occurring slowly in sodium bicarbonate solution, and rapidly in sodium carbonate solution and immediately in ammonium hydroxide solution. The substance responsible for this colouration was ultimately isolated and named “rubreserine” by Duquesnel, found (Ehrenberg 1893) to have the empirical [later (Salway 1912b) confirmed as the molecular] formula C13H16N2O2, and to be formed via the intermediacy of the colourless phenolic l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) – that is formed by hydrolysis of the N-methylcarbamyloxy group (with the loss of methylamine and carbon dioxide) from l-physostigmine – and which then absorbs two atoms of oxygen to form rubreserine (Salway 1912b) [interestingly, it has been reported (Salway 1912b) that “ On account of the ease with which it oxidises, eseroline possesses strong reducing properties, silver nitrate, gold chloride, and platinic chloride being reduced by it to the metallic state”]. The formation of rubreserine was confirmed (Massart et al. 1939) from the enzymic oxidation of l-eseroline in the presence of indophenoloxidase (cytochrome oxydase) and which, in accordance with the other known oxidation products formed by this enzyme from phenols, led to the suggestion – although lacking other experimental evidence – that rubreserine might be an ortho-quinone (Massart et al. 1939) but its structural formulation as such, namely as 138, did not appear until some 4 years later (Ellis 1943) and was independently confirmed – by reaction of rubreserine with ophenylenediamine [see also (Coyne and Paterson 1961)] to give rubreserine phenazine (Auterhoff and Hamacher 1967) – but also modified (Coyne and Paterson 1961) to the zwitterion 139 which, it was stated (Coyne and Paterson 1961), was “similar to that proposed for adrenochrome”, the red pigment obtained by oxidation of adrenaline (Harley-Mason 1948). However, adrenochrome had been shown

2.8 Detection, Assay and Instability

85

Me

O O

N

N H Me Me

138

Me

O O

N

N H Me Me

139 (Harley-Mason 1948) to have the resonance hybrid structure (140 Ð 141 in which H

O O

OH N Me

140

H

O O

OH N Me

141 the zwitterionic mesomeric structure 141 makes the major contribution) and, from a comparison of the uv, ir and 1H–nmr spectra of adrenochrome with those of rubreserine [for the latter product, 13C–nmr data and some revisions in the 1H–nmr assignments were subsequently published (Poobrasert et al. 1996)] – reported concomitantly with the most recent experimental details for the preparation of rubreserine from l-physostigmine – the structure of rubreserine was shown to be (138 Ð 139, namely a resonance hybrid of the ortho-quinone 138 with its zwitterionic mesomer 139 in which the latter makes the major contribution) (Robinson 1965c). This structure was later (Schönenberger et al. 1986b) confirmed – though without clear indication as to which mesomer makes the major contribution to the hybrid – by solid-state X-ray diffraction analysis. It was also later recognised (Daly et al. 1990) that one of the pseudophrynamines – isolated from the Australian frogs of the genus Pseudophryne (of the family Myobatrachidae) – which is red-coloured and tentatively assigned structure 142 (R¼H) (Daly et al. 1990)9 “may be related to rubreserine, the red pigment formed

86

2 l-Physostigmine (Eserine)

by oxidation of physostigmine” (Daly et al. 1990) – and therefore might have its tentative structure 142 (R¼H) extended accordingly, to include a contribution from the corresponding zwitterionic mesomeric component. Me

COOMe

O RO

N

H

N Me

142 In the past, there was a failure to recognise that the colourless initial degradation product of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1), leseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) has little desirable biological activity as an antiAchE and thus, for example, the heating processes that were used in the sterilisation of solutions of l-physostigmine for use in ophthalmology (Sect. 10.5) [see, for example, (Hemsworth and West 1970; Mair and Miller 1984)] – and which may cause hydrolysis of the N-methylcarbamyloxy group – may result in loss of activity, a situation that may also prevail upon long standing of such solutions. It was the appearance of a red colouration caused by the presence of rubreserine that was taken as the indicator of loss of potency of l-physostigmine solution (or of the drug in its solid state), with the retention of an uncoloured solution being a criterion of stability (Coyne and Paterson 1961; Hemsworth and West 1970). Therefore, the use of antioxidants which have been reported as being effective stabilisers [see, for examples (Miller 1954; Naidu and Zaheer 1959; Savoury and Turnbull 1985; Swallow 1951)] is spurious, since these merely prevent the formation of the red-coloured rubreserine but do not prevent the hydrolysis of the N-methylcarbamyloxy group in l-physostigmine (Coyne and Paterson 1961; Hellberg 1949).

2.8.3

Eserine Blue

One of the qualitative identification tests for l-physostigmine involves the development of a blue colouration when it is treated with aqueous ammonia (Petit 1871). The product responsible for this, eserine blue – which appears (Pavolini et al. 1951) to have been confused by at least one author (Salway 1912b) [an error compounded later (Ellis 1943)] with the basic product of empirical (Ellis 1943) formula C17H23N3O2 but of, as yet, unknown structure, and given the same name but obtained from l-physostigmine by its slow oxidation in alcoholic aqueous barium hydroxide with a limited supply of air (Salway 1912b) – was isolated from the reported evaporation of l-physostigmine in amounts of 10-100 mg with concentrated

2.8 Detection, Assay and Instability

87

ammonium hydroxide on a steam bath in an uncovered watchglass (Ellis 1943; Pavolini et al. 1951). It was shown (Pavolini et al. 1951) to have the empirical formula C36H35N5O3 and postulated (Pavolini et al. 1951) to contain a phenoxazone moiety since rubreserine (138 Ð 139 in which the zwitterionic mesomer 139 makes the major contribution) when treated with ammonia also afforded eserine blue. Upon this evidence and from ir spectroscopic data that indicated the presence of the phenoxazone ring, structure 143 [now showing the absolute configuration (Sect. 4)] was subsequently assigned (Auterhoff and Hamacher 1967) to eserine blue. Me

O

Me N

N N H Me Me

O

143

N N H Me Me

When an absolute ethanolic solution of rubreserine (138 Ð 139 in which the zwitterionic mesomer 139 makes the major contribution) was boiled under reflux under ammonia gas, two further blue products were formed. After their chromatographic separation and from consideration of their spectral properties, these were shown to differ from eserine blue and to have structures 144 or 145, although further studies with model compounds to distinguish between these two structures and clarify their mechanism of formation was required (Poobrasert et al. 1997). Me

Me N

N

N

Me Me

O

N

Me

144

Me Me N

N

Me

N

Me N O

145

N

N Me

Another degradation product – of particular interest because of its cytotoxicity (Sect. 10.12) – of l-physostigmine was formed, along with rubreserine (138 Ð 139 in which the zwitterionic mesomer 139 makes the major contribution) from which it was chromatographically separated, when an absolute ethanolic solution of the alkaloid was boiled under reflux with ammonium hydroxide on a steam bath for 2 h. Spectroscopic analysis of the new product showed it to have structure 146 with

88

2 l-Physostigmine (Eserine)

Me O EtNH

N

146

N Me

its “unusual mechanistic formation” being under further investigation (Poobrasert et al. 1996). Chromatography of the product resulting from the treatment of rubreserine (138 Ð 139 in which the zwitterionic mesomer 139 makes the major contribution) with alcoholic ammonia led (Auterhoff and Hamacher 1967) to the isolation, along with eserine blue, of a violet and a yellow compound. The former compound was converted upon standing into eserine blue and it was suggested that it is a “less hydroxylated” derivative of eserine blue (Auterhoff and Hamacher 1967) although its structure awaits further investigation. The yellow compound, which was also produced in reasonable yield by prolonged treatment of l-physostigmine with ammonia in aqueous ethanolic solution, was suggested (Auterhoff and Hamacher 1967) as being the end product of the ammoniacal decomposition of the alkaloid and, although its ir spectrum indicated that it contained a phenazine and not a phenoxazine moiety (Auterhoff and Hamacher 1967), its structure was neither established nor postulated. A green colouration (Pavolini et al. 1951; Salway 1912b) and amorphous powder (Pavolini et al. 1951) have also been produced during the decomposition of l-physostigmine but structural investigations in this area have yet to be effected.

2.8.4

Eserine Brown

It has been long known (Salway 1912b) that amongst the products resulting from the rapid absorption of oxygen by l-physostigmine in the presence of alkalis is eserine brown. “It is the end product of the decomposition of physostigmine, and forms when physostigmine, eseroline, rubreserine, and eserine blue are either heated in substance or in watery solution, or when they are allowed to stand in alkaline solution” (Ellis 1943). Similarly, “When anhydrous rubreserine is heated for a few hours at 115
C, the colour changes from red to brown, and finally becomes quite black without alteration in weight” (Salway 1912b). Eserine Brown has also been “prepared by evaporating to dryness an aqueous eserine solution which had been allowed to stand for some weeks” and then worked-up accordingly (Coyne and Paterson 1961). The product does not melt below 300
C (Coyne and Paterson 1961; Salway 1912b), “and is undoubtedly a polymerised form of rubreserine” (Salway 1912b). This supposition, coupled with the judgments that “the so-called eserine brown is not a homogenous compound” (Salway 1912b) and that “Eserine brown is not a single chemical compound”(Ellis 1943) and observations that they “show the same solubility properties” (Ellis 1943) have led to the conclusion that eserine brown

Notes

89

is probably structurally closely related to the melanin pigments {the structure of melanin has been reviewed [Spande 1979(a)(d)(f)(i)]} obtained from adrenaline and tyrosine (Ellis 1943; Pavolini et al. 1951). Furthermore, “since all appear to pass through an intermediary indole-quinone compound in the process of their formation” (Ellis 1943) adds further support to this hypothesis.

Notes 1. Apothecaries or Troy ounce (Gunn and Carter 1965). 2. Fluid ounces (Gunn and Carter 1965). 3. A further manifestation of the widespread natural occurrence of P. venenosum throughout much of equatorial west and central Africa is apparent from a list (Neuwinger 1996) of the plant’s vernacular names which includes those in Cameroon, Central African Republic, Congo, Gabon, Ghana, Ivory Coast, Liberia, Sierra Leone and Zaire. 4. Toward the beginning of his most interesting and comprehensive deliberation upon “The Nature and Definition of an Alkaloid”, Pelletier (1983) notes that “The term alkaloid was coined in 1819 by the pharmacist W. Meissner and meant simply, alkalilike (Middle English alcaly, from Medieval Latin alcali, from Arabic alqaliy ¼ ashes of saltwort, from qualey, to fry). The first modern definition by Winterstein and Trier described these substances in a broad sense as basic, nitrogen-containing compounds of either plant or animal origin” and for “True alkaloids” another four “additional qualifications” were defined. Pelletier later proceeds further by stating that “Therefore, following the concept that the structure of a compound [such as the presence of a basic nitrogen atom?] determines classification as an alkaloid, one should accept antibiotics of appropriate structure into the alkaloid class. Examples of antibiotics that, on the basis of structure, can be classed as alkaloids are .......... and the macrocyclic maytansinoids such as maytansine” but fails to note that although the structure of this latter compound contains three nitrogen atoms, like the two in the structurally closely-related maytanacine (1) they are all non-basic. He later elaborates upon this hypothesis by stating that “Though the term alkaloid originally meant alkalilike, we have seen that basicity can no longer be regarded as a necessary property of an alkaloid. Such glaring exceptions as colchicine [200], piperine [201],

MeO NHCOMe MeO MeO O OMe

200 O O

N

O

201

90

2 l-Physostigmine (Eserine) amine oxides......., and important quaternary salts such as laurifoline chloride......require that basicity no longer be included in a definition”. However, in this attempt to make the rule fit the exception rather than the exception fit the rule he meets opinions to the contrary from Wildman (1970) and Brossi (1997) who respectively state that “ Colchicine, C22H25NO6, is not an alkaloid in the strictest sense because the nitrogen atom is not basic being part of an acetamido function” [likewise, ricinine (202), isolated from the castor bean, Ricinus communis L. (Rob-

OMe CN N Me

O

202 inson 1988b), is not an alkaloid since one of its component nitrogen atoms is present in a non-basic lactam moiety – with the other being accounted for by a non-basic nitrile group] and “THE TERM “ALKALOID”, ORIGINALLY RESERVED FOR nitrogen-containing substances with basic properties found in plants, has a much broader meaning today and includes such substances occurring in mammals, fish, and mushrooms as well” (whilst, in the absence of opinion to the contrary, still retaining their basic properties), and also from the author who suggests a reappraisal of the alkaloidal status of the carbazole [so-called] alkaloids (Kapil 1971) since their component nitrogen atom is present in a non-basic carbazole moiety, as furthermore, is that in an essentially non-basic indole moiety [although the “weakly basic character of the indole nucleus” has been noted (Hinman and Lang 1960 – see also Robinson and Smith 1960) and was subsequently elaborated upon, in a review of the acidity [!] and basicity of indoles (Remers 1972), by the statement that “ the lone pair of electrons on the indole nitrogen is an integral part of the π-electron system and is not readily available for salt formation. A high concentration of hydrogen ions is therefore necessary to afford protonation of indoles. Such protonantion occurs mainly on C(3) in solution, but salts in which the proton was on nitrogen could be isolated from certain solutions by precipitation.”], of the two isomeric indole derivatives 203 and 204 which have been isolated from the liverwort, Riccardia sinuata (HOOK.)

N H Me

Me

203

N H

Me Me

204

Notes

91

TREV (Benešová et al. 1969) and of other structurally similar naturally-occurring 6-isoprenylindoles (Ishii and Murakami 1975; Nwaji et al. 1972) which, contrary to their correct designation as natural products (Nwaji et al. 1972), have thus been erroneously referred to as alkaloids (Benešová et al. 1969; Ishii and Murakami 1975) – or even bases (Ishii 1981; Ishii and Murakami 1975). Despite the fact that a combination of both of the above structural features in 6-bromo-Nb-formyl-Nb-methyltryptamine would render both the nitrogen atoms non-basic, this component from the marine bryozoan Flustra foliacea (L.) has also been erroneously referred to as an alkaloid (Christophersen 1985b). 5. Robert Robinson, OM, FRS (1885-1975) The synthesis of brazilin [vide infra] would have no industrial value; its biological importance is problematical, but it is worth while to attempt it for the sufficient reason that we have no idea how to accomplish the task. There is a close analogy between organic chemistry in its relation to biochemistry and pure mathematics in its relation to physics. In both disciplines it is in the course of attack of the most difficult problems, without consideration of eventual applications, that new fundamental knowledge is most certainly garnered. [Sir Robert Robinson (Farber 1963)] Biographical (Farber 1963; Saltzman 1993; Sherwood Taylor 1948) and autobiographical (Robinson 1955, 1976) details relating to Robert Robinson have already appeared. He was born on the 13th September 1885 in Bufford, near Chesterfield, Derbyshire, England, and it was in Chesterfield that his father, William Bradbury Robinson, was a highly successful manufacturer of cardboard boxes and surgical dressings and materials. Although his family were primary members of the Congregational Church in Chesterfield, Robert was a pupil at the Fulneck School, Pudsey Greenside, about midway between Bradford and Leeds, that was run by the Moravian Church in Great Britain – the majority of the masters were candidates for the Moravian ministry – and at which school he exhibited a natural predilection towards mathematics and physics and consequently wanted to be a mathematician. However, following his father’s wish, he decided to become a chemist and in 1902, upon passing the examination of the Joint Matriculation Board of the Universities of Manchester, Liverpool and Leeds, he became an undergraduate in the University of Manchester’s department of chemistry, then the paramount centre for both teaching and research in the subject throughout Great Britain. Although initially his chemical studies found him to be a somewhat reluctant student, it was during his second year at Manchester and consequent upon his attendance at lectures upon organic chemistry presented by William Henry Perkin jun, that the young Robinson became imbued with the subject and, after graduation in 1905 as Bachelor of Science with first-class honours in chemistry, he accepted a place in Perkin’s private laboratory. It would appear that it was from Perkin’s earlier research in the alkaloidal area that Robinson developed his interest in these natural products but, in the meantime, his work in his mentor’s laboratory led, in 1909, to his doctoral (DSc) thesis involving a study of the chemistry of brazilin (vide supra) and the related haematoxylin, the colouring matter of brazilwood. He was to prosecute this subject for the next six decades and, in fact, studies of the chemistry of brazilein were the first (Perkin jun and Robinson 1906) and last (Jaeger et al. 1974) [it is interesting that this paper was published from Shell Research Limited, Egham (vide infra)] of what were to be his some 700 contributions to the scientific literature between 1905 and 1974. Robinson’s appointment to a junior position in the Department of Chemistry at Manchester in 1909 coincided with the acceptance of a senior position in that department by Arthur Lapworth who was to have a second major influence on Robinson’s career. His association with Lapworth between 1909 and 1912 led to Robinson’s interest in the theoretical aspects of organic chemistry to which he was no doubt predisposed in view of his early aptitudes to physics and mathematics. Ultimately, this association with Arthur Lapworth was to lead to Robinson’s development of a general theory of organic reaction mechanism, exemplified by the first use of the “curly arrow” to show electron movement (Kermack and Robinson 1922) and the introduction of the term “aromatic sextet” to denote the six-electron system that accounted for the unique properties of benzene (Armit and Robinson 1925).

92

2 l-Physostigmine (Eserine) In 1912 the Department of Chemistry at Manchester was to play a major personal role in Robert Robinson’s life when he married Gertrude Maud Walsh, one of the first research collaborators of one of his departmental colleagues, Dr Chaim Weizmann. Perhaps not surprisingly, this led to a life-long friendship between the Robinson and Weizmann families and, amongst other considerations, ultimately led – in December 1953, following the death of Weizmann at Rehovoth on 9th November 1952 and who had by then become the first president of Israel – to Sir Robert Robinson being invited to deliver the inaugural Weizmann Memorial Lectures which were ultimately to find publication as a unique and pioneering text (Robinson 1955) which still makes fascinating reading. It was during the period of the preparation of this presentation that Lady Gertrude Maud Robinson died suddenly on 1st March 1954 which elicited the following moving tribute from her husband – “She was the mainstay of my life and I owe to her encouragement and active help in every direction any little success that I have been able to achieve” (Robinson 1955). Indeed, early in what was to be a lifelong mutual collaboration, this Robinson duo was to publish (Robinson and Robinson 1918, 1924) (from the University of Liverpool and from the Universities of St Andrews and Manchester, respectively) (vide infra) what – unlike the postulations of others which were found to be inconsistent with subsequent experimental observations – has become, after further extension and interpretation in light of modern electronic theory, the currently accepted mechanism of that doyen of indole syntheses, namely that discovered by Emil Hermann (Lucier 1993) Fischer (Robinson 1982). The three-stage mechanism which they proposed essentially embodies an earlier concept by Brunner (1898b) in which he “had made a comparison between the acid catalysed conversion of 2,2'-diaminobiphenyl into carbazole and the acid catalysed indolization of methyl isopropyl ketone Nα-methylphenylhdrazone” (Robinson 1982) × but which was not acknowledged by the Robinson duo. In addition to giving chemical analogies for all the three stages of their mechanistic proposal, the Robinsons, furthermore, extended their theory to the synthesis of 2,3,4,5-tetraphenylpyrrole from bisbenzyl phenyl ketone azine (Robinson and Robinson 1918). However, unfortunately, as they appear to have been with Brunner’s 1898 paper, they were also apparently unaware of another German publication (Piloty 1910) which had reported the acidcatalysed conversion of bisdiethyl ketone azine into 2,5-diethyl-3,4-dimethylpyrrole. It is obvious from their publications, that the Robinsons had access to the German literature covering these two German publications and their apparent unawareness of them is, therefore, surprising (Robinson 1982) – for the second enigma in Robert Robinson’s scientific studies, see Sect. 3.1 and vide infra. Also in 1912, Robert Robinson was offered, and accepted, his first professorship – at the University of Sydney – but in 1916 he returned to England to the newly-created Heath Harrison Chair of Organic Chemistry at the University of Liverpool. Here, he had the opportunity to renew his collaboration with Perkin, working on various aspects of alkaloid chemistry [see, for example (Perkin jun and Robinson 1919) from the Universities of Oxford and Liverpool], which led to his first success in this area by the synthesis of tropinone (which is structurallyrelated to atropine and cocaine) by reaction between acetone, methylamine and succindialdehyde (Robinson 1917a). This was published from the University of Liverpool, as was a later paper (Robinson 1917b) which concerned his first speculations relating to alkaloid biogenesis. Whilst in Liverpool, Robert Robinson also participated in research directly related to the World War I (the so called “war to end all wars”) effort of rejuvenating the dyestuffs industry and this led, in 1919, to his full-time appointment as Director of Research at British Dyestuffs in Huddersfield. However, owing to the problems in management he soon resigned from this post when the Chair of Chemistry became available at the University of St Andrews in 1920 – in which year he was also elected a Fellow of the Royal Society. Whilst at St Andrews he was able to return to the study of alkaloids and natural colouring matters and also began to seriously formulate his ideas relating to an electronic theory of mechanism of organic reactions based upon the ideas of those including Arthur Lapworth (vide supra). However, his stay at St Andrews was curtailed in 1922 when he was offered, and accepted, the Chair or Organic Chemistry at Manchester subsequent to Lapworth being appointed as Head of the Department.

Notes

93

During this second period at Manchester of some six years, there was fully developed an electronic theory of organic chemistry and, furthermore, many studies consolidating previous investigations on various alkaloids appeared. Included in these were the inaugural papers (Robinson and Suginome 1932a, b) relating to a potential synthesis of l-physostigmine (Sect. 3.1). The scene was now set for Robinson’s final moves – in 1928 to University College, London – from where the results of studies involved in the verification of the structure of lphysostigmine were published (Boyd-Barrett and Robinson 1932) (see footnote 10 – as quoted in Sect. 2) – and ultimately, in 1930, upon the death of Perkin in 1929, to the Waynflete Professorship of Chemistry at the University of Oxford (Magdalen College) where, amongst many other investigations, were to be the continuation of the quest for synthetic l-physostigmine which, although ultimately without success (Sect. 3.1), was to present an enigma that has only recently (Robinson 2002 – see also Ault 2008 and Sect. 3.1) been clarified. Robinson’s first decade at Oxford also saw the completion of work begun earlier on the plant pigments anthoxanins and anthocyanins and the beginning of work upon synthetic approaches to steroids and in 1939 he received a knighthood and was elected as President of the Chemical Society. The outbreak of World War II witnessed the beginning of his work – in association with Howard W Florey and Ernst Chain – on the chemistry of penicillin which continued until the declaration of peace at which juncture he returned to his research on the synthesis of steroids and, being mindful of the results from degradative work already effected by Hermann Leuchs, turned his attention to the establishment of the structures of strychnine and brucine, success in which was one of his major contributions to the study of the alkaloids. Indeed, in the main publication (Woodward et al. 1963) presenting the verification of the structure of the former alkaloid by total synthesis, it is stated with regard to its structural elucidation that “over a period of forty years, one of the great classics of structural organic chemistry was constructed. In that effort, described in more than two hundred and fifty separate communications, Robert Robinson played a brilliant and commanding role, and the extensive beautiful experimental contributions of Hermann Leuchs were of definitive importance.” In 1947 Robert Robinson was awarded the Nobel Prize in Chemistry (Farber 1963; Saltzman 1993) and elected President of the Royal Society. Perhaps it is not surprising that he became one of the major spokespersons for science and in 1949 he received that most prestigious civilian decoration in the UK when he was elected to the Order of Merit. His tenure was extended at the University of Oxford until 1955 but at this juncture he became a director of Shell Chemical Co Ltd when a small laboratory was placed at his disposal and he took a residence at 170 Piccadilly, London W1. As a consultant he visited the research establishments of Shell and continued in this role until he died in 1975. 6. Prior to the advent of 1H-nmr spectroscopy, the quantitative estimation of methoxy and methylimino groups was based upon a technique that involved initially heating with concentrated hydriodic acid. Under these conditions at 100
C, cleavage of methyl ether groups occurred to afford iodomethane (leading to the Zeisel method) (Zeisel 1895), as does likewise that of methylimino groups when the temperature was raised to 150
C (leading to the HerzigMeyer method) (Herzig and Meyer 1897). Thus, methoxy and methylimino groups could be estimated separately when both were present in the same compound and when, in both cases, the liberated iodomethane was sequentially driven off with carbon dioxide, absorbed in an ethanolic solution of silver nitrate and the silver iodide so formed determined gravimetrically. Alternatively, the iodomethane was pyrolysed to afford iodine which was determined volumetrically by sequential dissolution in potassium iodide solution and titration against thiosulphate. Both these approaches have been the subject of thorough reviews (Anon 1945; Kingscott and Knight 1914; Niederl and Niederl 1948). “Great caution has”, however “to be exercised in interpreting the results. Some substances yielded only one-half or one-third of their number of N-alkyl radicals, such as certain alkaloids and complex N-heterocyclic compounds” (Niederl and Niederl 1948 × see also Anon 1945). Thus, using this technique with l-physostigmine initially indicated the presence of only two of the N-methyl groups and it was only when the product remaining after the removal of those two methyl groups was examined further by the technique that the presence of the third N-methyl

94

2 l-Physostigmine (Eserine) group became apparent (Herzig and Lieb 1918a, b). Similary, only one of the two N-methyl groups in eseroline (3, R1¼H, R2¼Me, X¼NMe) was detected (Salway 1912b) by application of the Herzig-Meyer (1897) method, the second only becoming apparent (Straus 1914) by use of the Pregl apparatus – which is probably that illustrated in (Anon 1945). It appears that the accurate application of this technique depends upon the proportion of hydriodic acid used (Herzig and Lieb 1918a, b) and its quality (Niederl and Niederl 1948). This may also account for the detection (Jensen and Chen 1936) of only one N-methyl group (unfortunately experimental details were not presented) in dehydrobufotenine {the principle indolic component isolable from the parotid glands of the South American toad Bufo marinus (Märki et al. 1961) and one of the components (as its O-sulphate, bufothionine) [Spande 1979 (h)] of the skin of Bufo marinus (Honda et al. 1991) and of Bufo bufo bufo (Robinson et al. 1961) and also found in other Bufo sources [Spande 1979(h)] and in a galenical preparation of one of these known in China as Ch’an Su and in Japan as Senso [Robinson 1917b, 1960 × see, however, Jensen and Chen 1936; Spande 1979(h)]} that led to the erroneous postulation of 205

Me N HO N H

205 as its structure (Robinson 1917b, 1960). This, from earlier chemical and uv spectroscopic properties (Witkop 1956) and primarily a detailed study of its 1H-nmr spectrum, was subsequently corrected to 206 [Märki et al. 1961, Robinson et al. 1961 – see also Spande 1979(c)]

Me Me

N

O N H

206 which was ultimately confirmed by the synthesis of dehydrobufotenine (Gannon et al. 1967) and later by X-ray crystallography of bufothionine (Honda et al. 1991). 7. A facile conversion of l-physostigmine into l-eseroline (obtained as its fumarate salt in 98% yield and from which the free base can readily be isolated) in refluxing butan-1-ol in the presence of a catalytic amount of sodium butoxide was later reported (Brossi 1990; Brossi and Yu 1988b – see also 1988a; Yu et al. 1987, 1988b). This was particularly apposite since l-eseroline “ is extremely sensitive to aerial oxidation” and its preparation by either alkaline or acid hydrolysis of l-physostigmine requires “ neutralization and work-up from aqueous solutions, and proved tedious when carried out on a larger scale” (Yu et al. 1987). 8. The decomposition of eserethole methiodide to yield physostigmol ethyl ether is now known to be analogous to the reaction of thebaine methiodide (207) when heated in the presence of acetic anhydride to afford the product 208 – both cases involve the loss of the fragment CH2CH2⨁ NMe2IƟ and result in the formation of an aromatic system (Coxworth 1965).

Notes

95

MeO

I

O

Me

N Me

MeO

207

MeO MeCOO

MeO

208 9. This was the first recognition of the natural occurrence of the Ph-N-C-N system (Robinson 1963a) and the 1,2,3,3a,8,8a–hexahydropyrrolo[2,3-b]indole ring system which have since been detected in other products which have been isolated and characterised from plants, fungi and animals. Such metabolites comprise:9.1. Two of the Calabar bean’s other structurally closely-related alkaloids, namely l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1) and l-N(8)–norphysostigmine (98, R1¼MeNHCO, R2¼H, X¼NMe, n¼1) (Chaps. 4 and 5 respectively. 9.2. The dimeric tryptamine-derived alkaloids chimonanthine (122, R1¼R2¼H) (Grant et al. 1962, 1965; Hendrickson et al. 1962; Hino and Nakagawa 1988; Hodson et al. 1961; Kutney 1977; Robinson 1963a; Scott et al. 1964; Taylor 1966), calicanthidine (122, R1¼H, R2¼Me) (Grant et al. 1962, 1965; Hino and Nakagawa 1988; Saxton et al. 1962; Robinson 1963a; Scott et al. 1964; Taylor 1966) and folicanthine (122, R1¼R2¼Me) (Grant et al. 1962, 1965; Hino 1961a, b; Hino and Nakagawa 1988; Hino and Yamada 1963; Hodson and Smith 1956; 1957; Kutney 1977; Robinson 1963a; Taylor 1966) × all three of which have been isolated from plants in the botanical order of Calycanthaceae, a small family comprised only of the genera Calycanthus and Chimonanthus (Cordell 1981) [chimonanthine (122, R1¼R2¼H) has also been isolated from Palicourea fendleri (Hino and Nakagawa 1988; Nakano and Martin 1976) and P domingensis (Hino and Nawagawa 1988; Ripperger 1982), mesochimonanthine (209) has been found along with chimonanthine (Gorman et al. 1971; Hall

Me H H N N

N H

N H

209

Me

96

2 l-Physostigmine (Eserine) et al. 1967) in Chimonanthus fragrans, Lindle ((Meratia praecox, Rehd and Wils.), the deciduous shrub commonly known as Winter Sweet (Grant et al. 1965), and the enantiomer of chimonanthine, namely d-chimonanthine (210), has been isolated from the skin of

Me H H N N

N N H H Me

210 Phyllobates terribilis (the Columbian poison-dart frog) (Hino and Nakagawa 1988; Tokuyama and Daly 1983). 9.3. The next higher oligomers, namely the trimer hodgkinsine (211) (Attitulah et al. 1966; Fridrichsons et al. 1967, 1974; Gorman et al. 1971; Hino and Nakagawa 1988) and the tetramers quadrigemine-A (212) (as a mixture of diastereoisomers) and quadrigemine-B (213) (Gorman et al. 1971; Hino and Nakagawa 1988; Parry and Smith 1978) [all three of

Me H H N N Me H H N N

N N H H Me

211 Me N

N H

Me N

H N

N H

N Me

N Me

212

H N

Notes

97

H N

N H

N Me

N Me

N H N H

Me N

N Me

213 which have been isolated from Hodgkinsonia frutescens F.Muell, “ a plant in the family Rubiaceae quite unrelated to the Calycanthaceae” (Cordell 1981)], and the alkaloids with the highest molecular weight so far in this series, namely the pentamers psychotridine (214) which was isolated from Psychotria beccarioides Wernh, also of the family Rubiaceae (Hart et al. 1974; Hino and Nakagawa 1988) and isopsychotridine – tentatively assigned structure (215) and isolated from Psychotria forsteriana A.Gray along with psychotridine (214) and

Me N Me N

N H

Me N

H N

N H

N Me

N Me

214

H N

H N

98

2 l-Physostigmine (Eserine)

Me N Me N Me N Me N

H N

N H

N Me

H N

H N

H N

215 quadrigemines-A and -B (212 and 213, respectively) (Hesse 1981; Hino and Nakagawa 1988; Roth et al. 1985). It is interesting that the statement made consequent upon the elaboration of the structure of hodgkinsine (211) (Fridrichsons et al. 1974) that “the possibilities appear to exist for molecules of greater polymeric elaboration involving 4, 5. . .. . . N-methyltryptamine units” (Fridrichsons et al. 1974) thereby predicted the natural occurrence of structures such as 212, 213, 214, and 215. Perhaps higher oligomers, and even polymers – possibly having structural functions × remain to be discovered! However, the parent monomeric tricyclic moiety has been somewhat elusive (see footnote 4 in Chap. 3). 9.4. Some of the monoterpenoid indole alkaloids such as borreverine [Cordell 1983(a), Herbert 1983(c)], cabuamine [Joule 1983(e)], corymine [Joule 1983(cf), Robinson 1963a], echitamine salt [Creasey 1983(c); Herbert 1983(a); Joule 1983(d)] [in this category, this was the first of these alkaloids – although as a quaternary salt, and in the strictest sense, it is therefore not an alkaloid (see footnote 4) – of which the structural elucidation (as 216, “the

MeOOC

OH 6

H C H H

3 N H H N OH Me

Me H

216 disposition of the groups being with respect to the substituted cyclohexane ring C, which is in the boat form with C3 and C6 above the plane of the paper”) was effected – this work has been comprehensively reviewed (Robinson 1963a; see also Hamilton et al. 1962)], erinicine [Joule 1983(g)], erinine [Joule 1983(fg)], eripine [Joule 1983(c)], hunteracine [Herbert 1983(b), Husson 1983], isocorymine [Joule(c)], peceylanine, peceyline [Cordell 1983(c)], pelankine [Cordell 1983(d)], pleiocorine [Cordell 1983 (b)], rhazidine salt (Hino and Nakagawa 1988; Saxton 1983b) and vincorine [Cordell 1983(b), Joule 1983(e)].

Notes

99

9.5. Isolated from Australian frogs of the genus Pseudophryne (of the family Myobatrachidae), a group of pseudophrynamines, namely 217 [R¼CH2OH, COOMe and CHO (tentative structure)] (Daly et al. 1990), 218 (R1¼H, R2¼OH; R1¼H, R2¼MeO; R1¼OH, R2¼MeO and R1¼R2¼MeO) (Daly et al. 1990), 219 (Daly et al. 1990), 142 [R¼H and Me (both tentative structures)] (Daly et al. 1990), 220 (Daly et al. 1990) and 221 (tentative structure)

Me

R

N H

N Me

217

COOMe

Me

R1, R2 N H

N Me

218

Me OH

HO N

N Me

219 O

O Me

Me

Me

O

O

Me

HO R

220

R

R

R= N H

N Me

221

R

100

2 l-Physostigmine (Eserine) (Daly et al. 1990) and, found in the skin of the Australian frog Pseudophryne coriacea, 217 (R¼CH2OH) (Crich et al. 1995; Mitchell and Le Quesne 1990; Spande et al. 1988), 217 (R¼COOMe) (Spande et al. 1988) and 222 ( 220) (Cozzi et al. 1990; Mitchell and Le Quesne 1990; Spande et al. 1988).

O H

H

O Me Me N N H Me H

N N H Me H

222 9.6. A series of 6-bromo-1,2,3,3a,8,8a-hexahydroypyrrolo[2,3-b]indoles which are also usually 3a-prenylated [for example, flustramine A (223) (Carlé and Christophersen 1979, 1980) and

CH2 Me

Me

Br

N

Me

H

N Me

Me

223 flustramine B (169) (Carlé and Christophersen 1979, 1980; Muthusubramanian et al. 1983;

Me

N

Br

N H

Me

Me

Me

Me

169 Hino et al. 1983] and which have been isolated from the marine bryozoan Flustra foliacea (L) and have already been the subject of reviews (Christophersen 1983, 1985a, b, Hino and Nakagawa 1988) and two of which, urochordamine A and B (224, R1¼Et, R2¼225) and

Notes

101

R2

R1

H Br

N N H H Me

224 Me N

N

NH N

N O

Me

225 (224, R1¼225, R2¼Et), respectively, have been isolated from the tunic (outer body) of the adult tunicate Ciona savignyi (a marine animal) and which promote settlement and metamorphosis of the C. savignyi larvae (Tsukamoto et al. 1993). 9.7. From a variety of sources, microbial metabolites (several of which have biological activity which suggests a potential either therapeutic or other utility) which include:9.7.1. Aszonalenin (226, R¼H) which “apparently induced the abnormal second cleavage the sea urchin embryos” and which was isolated, along with LL S490β

Me Me

CH2 H O

N N RO

N H

226 (226, R¼COMe), from Aspergillus zonatus IFO 8817 (Hino and Nakagawa 1988; Kimura et al. 1982), with the latter metabolite having earlier (Ellestad et al. 1973) been isolated from an unidentified Aspergillus species. 9.7.2. Responsible for the serious animal disease of “facial eczema” (Hino and Nakagawa 1988), which in sheep causes extensive liver damage and ultimate death and which is widespread in New Zealand with outbreaks having also occurred in Australia, sporidesmin A (227) (Fridrichsons and Mathieson 1965), along with the structurally

102

2 l-Physostigmine (Eserine)

OH OH H

Cl

O N N S H 2 Me N O Me Me

MeO MeO

227

closely-related sporidesmins B, C, D, E, F and G, has been isolated from the fungus Pithomyces chartarum (Hino and Nakagawa 1988; Sammes 1975). 9.7.3. Brevianamide E (228, R¼CMe2CH¼CH2) (Hino and Nakagawa 1988; Sammes

OH H O

N N R H O H

N

228 1975) which was isolated from the mould Penicillum brevi-compactum Dierckx (University of Manchester Acc 382) (Birch and Russell 1972; Birch and Wright 1970; Christophersen 1983). 9.7.4. Roquefortine (229), a neurotoxic mycotoxin which showed tremorgenic activity in

Me Me

CH2 H

N H

O

N N

O

H H

N

N

229

H

male mice [LD50 (i.p.) 15-20mg/kg, with doses of 50-100mg/kg causing prostration and an atonic posture (Scott et al. 1976)], was isolated from Penicillium roqueforti (Christophersen 1983; Hino and Nakagawa 1988; Scott et al. 1976; Vleggaar and Wessels 1980) –it was also reported (Ohmomo et al. 1975, 1977, 1978) as being isolated (as roquefortine C) from this same source and as a metabolite of P crustosum (Hino and Nakagawa 1988; de Jesus et al. 1983). Reduction of roquefortine C using zinc powder in boiling acetic acid under reflux afforded a mixture of two stereoisomers of dihydroroquefortine C (230) which were separated

Notes

103

Me Me

CH2 H

N H

O

N N

O

H N N H

230

by column chromatography on silica gel. Although the absolute configuration at the newly-formed asymmetric carbon atom remains to be established in both these isomeric products, one of them was found to be identical with the roquefortine D which was also isolated from P roquefortii – interestingly, the other isomer could not be found in this source (Ohmomo et al. 1978). 9.7.5. Okaramine A (231) and okaramine B (232), produced by penicillium simplicissium

Me

Me

H 2C

N OH

H NH

O

N N H

O

Me Me

231

Me Me Me

N OH O

N OH N

N H

OMe

O

Me Me

232 AK-40 (isolated from a soil sample) when cultured with okara (the insoluble residue of whole soybean) (Hayashi et al. 1989). Both the metabolites exhibited insecticidal activity against the 3rd instar larvae of silkworm, with okaramine B being much more active than okaramine A which was as active as physostigmine (Sect. 10.8) (Hayashi et al. 1989).

104

2 l-Physostigmine (Eserine) 9.7.6. With vasodilating activity, although further details of its pharmacological characterisation await publication, amauromine was isolated from the culture broth of Amauroascus sp. No.6237 (Takase et al. 1984) and its structure was determined as (233) from chemical

Me Me

H2C

H

O H H N N

N N H H O H

Me Me

H2C

233 and spectral data (Takase et al. 1985b) and verified as such by total synthesis (Takase et al. 1985a, 1986a, b). It “ is the first natural diketopiperazine composed of two L-tryptophans” (Hino and Nakagawa 1988) and also appears to have been reported as a metabolite of Penicilliun nigricans, but with no stereochemistry indicated, under the name of nigrifortine which was recognised as “the first example of a prenylated diketopiperazine fungal metabolite in the form of an indolic symmetrical dimer” (Laws and Mantle 1985). 9.7.7. Two dimers WIN 64821 (234, R1¼Ph, R2¼H) and WIN 64745 (234, R1¼CHMe2,

R1 R2 H N O

H

O H H N N

C

B

B

C

N N H H O

H

O

N H R2

Ph

234 R2¼H), obtained from the whole culture fermentation broths of Aspergillus sp,SC319 (ATCC 74177) originally isolated from soil, with the former metabolite being found to be “ a competitive antagonist to substance P(SP) at the human NK1 receptor with an inhibitor affinity constant (Ki) of 230 30nM against [125I] SP in human astrocytoma cells” (Barrow et al. 1993). 9.7.8. A group of structurally-related fungal metabolites which together have been subjected to review (Hino and Nakagawa 1988) and include:9.7.8.1. Chaetocin (235, R1¼R2 ¼ OH, R3¼R4¼H, n¼2) (see also Sammes 1975; Weber 1972) isolated from Chaetomium minutum and possesses antibacterial and cytostatic activity.

Notes

105

R1 MeN O

Sn

O H H N N

R3

R4 O N N Sn H NMe H O R2

235 9.7.8.2. The antibiotics verticillines A (235, R1¼R2¼H, R3¼R4¼OH, n¼2) and B (235, R1¼H, R2-R4¼OH, n¼2) which were isolated from Verticillium sp.(strain TM-759) and possess antimicrobial activity against gram-positive bacteria and mycobacteria and are cytotoxic, both with ED50 against HeLa cells of 0.2μ/ml. 9.7.8.3. Produced by Acrostalagmus cinnabarinus var. melinacidinus, melinacidins II (235, R1¼R3¼H, R2¼R4¼OH, n¼2), III (235, R1-R3¼OH, R4¼H, n¼2) and IV (235, R1-R4¼OH, n¼2) which inhibit the growth of a variety of grampositive bacteria in vitro. A similar metabolite, chetracin A (235, R1-R4¼OH, n¼4) which “exhibits remarkable cytotoxicity to HeLa cells” has been isolated from Chaetomium abuense Lodha and C. retardatum Carter et Khan. 9.7.8.4. Obtained from Aspergillus flavus (strain MIT-M-25, 26 and 27) and A. Flavus var columnaris, ditryptophenaline (234, R1¼Ph R2¼Me, with optical asymmetry at the B/C ring junction inverted) which possesses neither significantly toxic (LD50 200mg/kg) nor antibiotic properties. 9.7.9. Isolated from a strain of Leptosphaeris sp. originally separated from the marine alga Sargassum tortile, leptosins M and M1 (236, R1¼OH, R2¼H, n¼4 and 2, respectively)

OH O NMe H H H N N O OMe 2 R1 R R1 R2 O N N H Sn NMe N O Me

Me

236 and N and N1 (236, R1¼H, R2¼OH, n¼4 and 3, respectively). Significant cytotoxicity against the murine P388 cell line was exhibited by all four of these metabolites (with the activities of leptosins N and N1 being almost 10x greater than those of leptosins M and M1), as it was against a disease-oriented panel of 39 human cancer cell lines. Interestingly, it appears from further investigations that the modus operandi of leptosin

106

2 l-Physostigmine (Eserine) M is probably different from that of any other currently known oncolytic drugs (Yamada et al. 2002). 9.7.10. Chetomin (237), a toxic metabolite which is strongly cytotoxic towards HeLa cells, is produced by various Chaetomium species, including C. abuens, C. cochliodes,

O

MeN

N

S2

NMe OH

O O N N S2 H H NMe O OH

237 C. globosum, C. retardum and C. tenuissimum, and dethio-tetramethylthiochetomin (238) has been isolated from C. globosum Kinze ex Fries (Hino and Nakagawa 1988, Kikuchi et al. 1982).

MeS

O Me N

N H

SMe OH

N

N Me

O SMe O

N O MeS

N Me OH

238 10. This pioneering synthesis from University College, London (see footnote 5) (Boyd-Barrett and Robinson 1932) began with the Fischer indolisation in warm alcoholic sulphuric acid of the phenylhydrazone 12 (R¼H) that was formed in situ – from phenylhydrazine and 2-oxo-5phenoxypentane – to afford 13 (R¼H) { the possible alterative structure 14 (R¼H) was eliminated since “ the colour reactions [with Erhlich’s reagent] do not indicate the presence of a free position in the pyrrolo nucleus” and “a further argument is that, in the later stages,” 14 (R¼H) “would lead to products with an additional methylene group, and this is not indicated by the analyses”}. Compound 13 (R¼H) was then heated in an autoclave with iodomethane in methanol at 120oC for 8h to yield 15 (R¼H) which upon shaking with a mixture of aqueous sodium hydroxide and ether for 7h gave 11 (R1¼H, R2+R3¼CH2, R4¼OPh) that in cooled acetone was oxidised by the gradual addition of finely powdered potassium permanganate over 9h to afford 11 ((R1¼H, R2+R3¼O, R4¼OPh) (see footnote 11). Heating this under reflux with hydrobromic acid (d 1.7) for 12h yielded 11 (R1¼H, R2+R3¼O, R4¼Br) which upon heating in methanol in a sealed tube at 170-180oC for 10h with aqueous dimethylamine afforded 11 (R1¼H, R2+R3¼O, R4¼NMe2) and likewise with aqueous methylamine afforded 11 (R1¼H, R2+R3¼O, R4¼NH Me). This latter product was also a key intermediate in the Julian and Pikl synthetic approach to the physostigmime ring system (Sect. 3.1) and the above studies therefore represent an

Notes

107

alternative route to its synthesis, albeit “a very complicated route” as stated by the American duo (Julian and Pikl 1935a). Moreover, and concomitantly with the completion of this synthesis and with reference to the latter product it was stated that “We are of the opinion that it should be possible to effect the ring closure of substances of type (VIII) [11 (R1¼H, R2+R3¼O, R4¼NHMe)], but this has not yet been investigated” (Boyd-Barrett and Robinson 1932) × and, it has been claimed (Brossi et al. 1996) “was never completed.” However, later in that same year such an investigation was, in fact, reported by King and Robinson (1932b) although they incorrectly identified the product from their reaction and thereby failed to recognise the success of their efforts (Sect. 3.1) It may, however, be significant that the above opinion by Boyd-Barrett and Robert Robinson (1932) [and a similar one expressed by King and Robert Robinson (1932b)] would have been available to Julian and Pikl either before they embarked upon their synthetic studies toward lphysostigmine or during its early phases, in which this type of ring closure is fundamental. Indeed, the paper by Boyd-Barrett and Robert Robinson (1932) is quoted as a reference by Julian and Pikl (1935a). Consequently, it is possible that Percy Julian could have deliberately engineered the altercation (see footnote 12) that was to become “the very public disagreement between himself and Robert Robinson” (Ault 2008). 11. The colourless 1,3,3-trimethyl-2- methyleneindoline (250, R1+R2¼CH2) upon exposure to air

Me Me R2

N R1 Me

250 and sunlight rapidly began to redden [as had been observed previously (Brunner 1898a, b, Plancher 1898) with this compound] and was, over a period of 10 days, completely autoxidised into 250 (R1+R2¼O) and formaldehyde (Robinson 1962, 1963b). By analogy with these observations, it is now suggested that the enamine 11 (R1¼MeO, R2+R3¼CH2, R4¼OPh) which, as also reported (King and Robinson 1932a), “acquired a purple tint in the air” [see also (Boyd-Barrett and Robinson 1932) for an analogous observation with 11 (R1¼H, R2+R3¼CH2, R4¼OPh) and (Boyd-Barrett 1932; Ficken and Kendall 1961; Robinson 1960) for other examples of related enamines which have also been observed to redden upon aerial exposure and thereby likely undergo similar autoxidation], is likely to be similarily converted into 11 (R1¼MeO, R2+R3¼O, R4¼OPh). This would thereby obviate the use of potassium permanganate to effect this transformation (King and Robinson 1932a) and that of 11 (R1¼H, R2+R3¼CH2, R4¼OPh) into 11 (R1¼H, R2þR3¼O, R4¼OPh) (Boyd-Barrett and Robinson 1932). It has not escaped attention that similar autoxidations of either such enamines or of tertiary indolinols {exemplified by the simple analogue 2-t-butyl-1,3,3-trimethylindolin-2-ol 250 (R1¼tBu, R2¼OH) which also undergoes an “interesting” [Spande 1979(e)] air-oxidation to afford 250 (R1+R2¼O) with loss of the t-butyl substitutuent – indolinols with 2-phenyl and 2-hydrogen substitutents are stable (Robinson 1962, 1963b)} could be involved in the formation (including biosynthetic) of alkaloids containing the indolin-2-one moiety, namely the “oxindole alkaloids”. 12. In the communication announcing their synthesis of dl-eserethole, the validity of which is beyond any doubt, Julian and Pikl (1935b) state that “ In a series of ten beautiful papers Robinson and his co-workers have described syntheses of compounds which they call “d,leserethole” and “d,l-esermethole.” Their “d,l-eserethole” is not the compound (see footnote 13) described in this communication as d,l-eserethole, and the constitution of which can hardly be questioned. We believe that the English authors are in error, that the compound they describe as d,l-eserethole is not the substance, and that we are describing for the first time the real d,leserethole”, and summarise that “The “d,l-eserethole” described by other workers in this field is thought to have another constitution than that ascribed to it.” They reinforce this conclusion by

108

2 l-Physostigmine (Eserine)

the comment that “our product was the real d,l-eserethole and that that of the English chemists must be assigned another constitution” in their communication concerning their complete synthesis of l-physostigmine (Julian and Pikl 1935c). This “very public disagreement between Percy Julian and Robert Robinson, as to the identity of “eserethole” ˮ has been the subject of detailed discussion and analysis in review (Ault 2008; Robinson 2002). It was, however, recognised by Julian and Pikl (1935b) [and also later in review (Julian et al. 1952b)] that Robert Robinson’s synthetic approach was intact until its final stage [although unlikely (Ault 2008) doubt would appear to have been cast (Kobayashi 1938) upon this] when they noted that (Julian and Pikl 1935b) “The English authors depend upon methylation of their d,l-noreserethole (which seems to be identical with our product.....) with methyl ptoluenesulfonate and at times methyl sulfate. This could well lead to a substance whose structure is represented by [30]. A substance [methyl-eserethole] with this formula assigned to it has been obtained by Hoshino and Kobayashi [1934a – see also 1934b] through methylation by various procedures of dinoreserethole [vide supra]. Its melting point appears to be identical with the “d,l-eserethole” of Robinson and his co-workers and likewise the melting points of the two picrates seem identical”. Indeed, such was reported (King and Robinson 1935) as being corroborated “ by a direct comparison of the specimens”, at which juncture the molecular formula of “methyl-eserethole” was also confirmed as C15H22N2O which contained only two N-methyl groups – ultimately it was shown to have structure 29 (R1¼EtO, R2¼H, R3¼NMe2) (vide supra). Nevertheless, the Oxford group tenaciously refused to accept the unequivocal fact that their synthesis had collapsed at this stage by their statement that “In our opinion the base is structurally identical with eserethole, and it may be a stereoisomeride of this base (see footnote 13). We accept the evidence of Hoshino and Kobayashi [1934a-see also 1934b] (loc. cit.) that the substance is not d,l-eserethole, although the facts leading to this conclusion have not fallen within our experience” (King and Robinson 1935). However, Robert Robinson (see footnote 5) – the future (1947) Nobel Laureate in chemistry (Farber 1963; Saltzman 1993) “for his work on the synthesis of natural products, the alkaloids” (Saltzman 1993) and who has been regarded as “the master of synthesis” (Saltzman 1993) and as “ the king of alkaloid chemistry” (Brossi et al. 1996) – and his group at Oxford were never to reach the synthetic objective of l-PHYSOSTIGMINE. Indeed, this was further manifest by their failure to prepare dl-noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me) by heating 5-methoxy-1,3dimethylindole with an excess of ethyleneimine in a sealed tube at 120
C for 3h (King and Robinson 1933) which, perhaps not surprisingly, afforded only unchanged indolic starting material. Despite the above observations and conclusions, and with reference to the work of Robert Robinson and some of his team [in particular (King et al. 1933b, Robinson and Suginome 1932a, b)], it has been stated (Sumpter and Miller 1954a) that “The first synthesis of the complete ring system of physostigmine was achieved by the preparation of dl-noreserethole [which would appear to have been recognised by Julian and Pikl (1935b) – see also (Henry 1949, Julian et al. 1952b), although Robert Robinson himself occasionally appears not to have realised that he had reached this synthetic destination (Robinson 2002)], which could be methylated to give dl-eserethole” [which is certainly not the case (vide supra) (Robinson 2002)]. 13. Robert Robinson and his co-worker FE King sought to assign (King and Robinson 1935) differences (see footnote 14) to variations in molecular geometry at the B/C ring junction of the 1,2,3,3a,8,8a-hexahydropyrrolo[2,3-b]indole system by their statement that “It is relevant to note that Linstead and Meade (1934, 935) have isolated cis-cis- and cis-trans-isomerides of fused dicyclooctanes (two five-membered rings), and we provisionally regard the isomerism of d,l-eserethole-a, mp 38
(synthesised by Julian and Pikl, loc. cit), and our d,l-eserethole-b, mp 80
, as another case of the same kind. The fact that the behaviour of d,l-eserethole-a, and d,leserethole-b towards methyl iodide is different is not surprising, because the stereochemical hypothesis closely concerns the configuration of the nitrogen atoms.” (King and Robinson 1935). However, Robinson and King had failed to realise the effect on the eserethole tricyclic system of the aromatic ring A which, by forming a near-planar system with the ring B, thereby restricts the B/C ring junction to a cis fusion (Jackson 1954), (see also Sect. 4). No such restriction is present, of course, in Linstead and Meade’s fused dicyclooctanes and the conclusion to be drawn, therefore, is that “all compounds containing the tricyclic ring system [as in

Notes

109

eserethole] can exist in one configuration only, i.e. with the two pyrrolo rings cis to each other, so that for each individual compound only one pair of enantiomorphic forms is possible” (Jackson 1954) and that “All the experimental results so far obtained are in agreement with this hypothesis, and for each compound prepared only one pair of enantiomorphs has ever been obtained, never two pairs as would be predicted on the basis of the two asymmetric centres which the molecule contains.” (Jackson 1954) (see also Sect. 4). 14. Earlier, along with M Liguori (King et al. 1934), they had also suggested that mp discrepancies might have resulted from either dimorphism [“It appears that the substance exists in two crystalline modifications” (King et al. 1934)] or even the “alternative hopeful idea of spontaneous resolution” (King et al. 1934) although this latter “did not survive the experimental test” (King et al. 1934). 15. Nevertheless, King and Robert Robinson (1935) later found it necessary to further state, with understandable reason, that “The identity of the synthetic salts with those derived by degradation of the alkaloid established for the first time the existence of the angle methyl group and gave important support to the view of the constitution of eserine which is now generally accepted. As this fact does not seem to have been recognised by other workers in the field, we venture to draw attention to it”. From the literature cited by the Oxford duo, the “other workers” undoubtedly could have included Julian and Pikl (1935c) who asserted that “The most convincing work on the constitution of eserine, that from Barger’s laboratory [quoting (Stedman and Barger 1925), namely some seven years before the paper by King and Robert Robinson (1932a)] left one important gap unfilled, namely, the synthesis of oxindole derivatives secured as degradation products”. Perhaps, too, the statement by the Oxford pair reflected a frustration at the failure of the synthetic approaches by Robert Robinson and his research group to the actual alkaloid (Sect. 3.1), the successful synthesis of which by Julian and Pikl (1935c) they concomitantly acknowledged (King and Robinson 1935). 16. If, instead of using aluminium chloride alone, a melt of a mixture of aluminium chloride with sodium chloride (5:1 w/w) is employed to effect this type of cyclisation the resulting yields of product may be substantially increased (Sugasawa and Murayama 1958a). Alternatively, the Stollé approach to the required indolin-2-onic synthetic intermediate could be replaced by the Brunner synthesis (Robinson 1982). Indeed, this is a method of choice for the preparation of 3-alkylindolin-2-ones – affording yields in the region of up to 80% (Robinson 1982) and, furthermore, its use would circumvent the undesired O-dealkylation (Julian and Pikl 1935b; Sugasawa and Murayama 1958a, b) and thereby obviate the necessity for the O-realkylation stage (Julian and Pikl 1935b, Sugasawa and Murayama 1958a, b). An interesting and useful protection of the potential 5-hydroxyl function in a Julian and Pikl type synthesis (Sect. 3.1) involving the use of the tetrahydro-2-pyranyl moiety 251 – apparently

O

251 affording an acetal, such as the methoxymethyl moiety as used by Kolosov et al. (1953) (see footnote 22) – was developed at the Shanghai Institute of Organic Chemistry (Brossi et al. 1996; Pei et al. 1996; Yu et al. 1994). However, although this latter protecting group affords the best synthesis of the racemic alkaloids, its component chiral centre leads to complications in the synthesis of optically active alkaloids (Brossi et al. 1996; Pei et al. 1996). Perhaps it is of ultimate interest that the Stollé synthesis of 1,3-dimethyl-5-hydroxyindolin-2-one has, in fact, been effected without the need to protect the OH group (Pei et al. 1998; Yu et al. 1994). 17. The reductive cyclisation of, for example 19 [R1¼H, R2¼(CH2)2NH2] into 23 (R1-R3¼H, R4¼Me) using sodium in “alcohol” [a Ladenburg reduction (Hino 1961b, Yamada et al. 1963)] requires large amounts of sodium and “alcohol”. This consequently makes the experimental procedure cumbersome and labour-intensive. In addition, the use of large amounts of sodium is dangerous in large-scale reactions and the neutralisation of the base with acid upon work-up results in a large amount of sodium salts (Brossi et al. 1996). These disadvantages, however, were later ameliorated by the introduction of lithium aluminium hydride [which had already

110

2 l-Physostigmine (Eserine)

been used (Julian and Printy 1949) in stirred ether to effect the reduction of 1-methylindolin-2ones to 1-methylindoles – other early studies upon the lithium aluminium hydride reduction of oxindoles have been covered in review (Robinson 1917b, 1969)], in either ether or ether/ dioxane (Hino 1961b; Hino and Ogawa 1961; Yamada et al. 1963) and in either dioxane or tetrahydrofuran (Brossi et al. 1996; Hendrickson et al. 1962; Hino and Yamada 1963; Sandoz Ltd 1966; Yu and Brossi 1988; Yu et al. 1994) boiling under reflux, as the reducing agent. When proceeding via the formation of the carbamate 19 [R2¼(CH2)2 NHCOOMe], these conditions concomitantly result in an ultimate N-methylation to presumably afford the intermediate 19 [R2¼(CH2)2NHMe] (Pallavicini et al. 1994), and they can also be employed in the direct reductive cyclisation of the Julian and Pikl nitriles 19 (R2¼CH2CN) (Brossi 1990; Sandoz Ltd 1966; Yu and Brossi 1988), thereby obviating the low yields caused by a competing hydrolysis of the nitrile group resulting from the difficulty in excluding moisture from the reaction mixture when using sodium in anhydrous ethanol to effect the reductive cyclisation. However, the sensitivity of lithium aluminium hydride toward air and the necessary use of highly-flammable peroxide-forming ethers is hazardous, particularly in large scale preparations (Pei and Bi 1994; Pei et al. 1996). Those factors have led to the introduction of sodium dihydride-bis(2-methoxyethoxy) aluminate, known commercially as Vitride, in toluene (Pei and Bi 1994; Pei et al. 1996), as a reducing agent that is chemically equivalent to lithium aluminium hydride but is stable to oxygen, highly soluble in a variety of solvents, easy to workup, and safe – especially for large scale reductions (Brossi et al. 1996; Pei and Bi 1994; Pei et al. 1996, 1998). Its use can also lead to the isolation of the corresponding 2-indolinols which, via their methiodides can be reacted with a wide variety of nucleophiles to afford products of considerable interest (Pei et al. 1994) (Sect. 3.2.2) (see footnote 3 in Chap. 3). 18. Contrary to the earlier implication (Stedman and Barger 1925) that these reductions, involving zinc and hydrochloric acid (Polonovski 1918, Polonovski and Polonovski 1924e – see also 1924a) and hydrogenation over platinum in glacial acetic acid (Stedman and Barger 1925) must have involved the direct rupture of the C8a-N1 bond, which “seems rather unlikely as this is not activated, e.g. benzylically or allylically”, it has been suggested that they occur via an acidcatalysed ring-C cleavage of, for example, 23, (R2¼R3¼H, R4¼Me) (Jackson and Smith 1964, see also Troxler 1972) to afford the 3H-indolium cation such as 22 and that it is the immonium function thus formed that is reduced (Jackson and Smith 1964). Verification for this arises from the observation that “no reduction occurs in neutral or alkaline solution” (Jackson and Smith 1964). Thus the 1,2,3,3a,8,8a-hexahydro-3a-methylpyrrolo[2,3-b]indole ring system is unaffected under conditions of hydrogenation over a palladium-on-carbon catalyst in methanol (Atack et al. 1989; Brossi 1990; Brzostowska et al. 1992; Yu et al. 1988a, b) and it also explains why the C(8a)-N1 bond in Julian and Pikl’s product is stable to sodium in boiling dry ethanol under reflux. 19. Since phenacetin (acetylphenetidine) (17, R1¼EtO, R2¼COMe) is one of the so-called “coal tar” analgesics and l-eseroline was some three decades ago (Sect. 10.11) also shown to possess analgesic activity, albeit of an opioid type, curiously it may be considered that this synthesis has, this far, pharmacologically returned whence it commenced. 20. Alkylations such as this can now be effected asymmetrically by phase-transfer catalysis using chiral catalysts. For example, with 19 (R1¼MeO, R2¼H) in the presence of 252 (R¼4-CF3,

H

HO H

H

X

N CH2 R

N

252 X¼Br; R¼3,4-Cl2, X¼Br and Cl) under phase-transfer conditions, ees of the S-enantiomer 253 over the R-enantiomer 254 of 72%, 77% and 78%, respectively, were realised at the Chemical

Notes

111

Me

MeO

CN N Me

O

253 Me MeO CN N Me

O

254 Research Department of Hoechst-Roussel Pharmaceuticals Inc. in Somerville, New Jersey, USA (Lee and Wong 1991 × see also Pei et al. 1996). The separation of the mixture of enantiomers thus enriched in one of the two optical isomers can be effected by spontaneous crystallisation (Brossi 1990, 1994; Brossi et al. 1996; Pei and Brossi 1995), a very interesting observation. By thus focussing upon the first prochiral intermediate – namely the nitrile – thereby effected a remarkable improvement in the Julian and Pikl synthesis (Brossi et al. 1996) by affording at an early stage the desired optically pure enantiomer in high yield which obviated the later wasteful formation of undesired enantiomers which would have to be discarded. However, difficulties have arisen using this approach which were overcome by catalytic reduction of the nitrile to the amine which was resolved via formation and a single crystallisation of the diastereoisomeric dibenzoyl-D-tartarates (Lee and Wong 1991). 21. The resolution of racemic amines via their reaction with an optically active acid to afford a mixture of two diastereoisomeric salts that are then separated by fractional crystallisation and from which the optically active amines are subsequently liberated is a widely-employed classical technique. However, sometimes it is either unsuccessful or proves to be difficult and not very practical [despite the use of novel acidic resolving agents which will continue to be of value (Brossi 1994)], with tedious recrystallisations of the salts being required to yield optically pure materials in consequent low yields. Indeed, such is the case with the resolution of dl-N1-noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me) and dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) which can thus be separated into their optically pure stereoisomers in only low yields (Brossi et al. 1996; Schönenberger and Brossi 1986; Schönenberger et al. 1986a). However, an alternative method is now available by which the optically active amines can be obtained from their racemates in high yield and of excellent optical pure amines. It is similar in principle to the earlier classical method of resolution, involving the preparation of a mixture of two diastereoisomeric derivatives (but in this case ureas, obtained by reacting enantiomeric secondary amines with optically pure isocyanates), their separation (in this instance chromatographically), and finally the liberation of the optically pure amines. This technique is applicable to the resolution of a wide range of racemic secondary amines including, for example (Brossi, 1985, 1990, 1992, 1994; Brossi et al. 1996; Brossi and Yu 1988a b, Kawabuchi et al. 1988; Schönenberger et al. 1986a, b; Yu et al. 1988a) that of dl-N1-noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me) in which a stirred solution of the latter in chloroform at 0
C is treated by the dropwise addition of the isocyanate 255 (R¼O¼C¼N) and, after 2h, the solvent is evaporated to leave a mixture of the diastereoisomeric ureas 256 and 257 (a discussion of the

Me R

H

Ph

255

112

2 l-Physostigmine (Eserine)

Me

MeO A

C B Me N N H Me CONH H Ph

256 Me

MeO A

C B Me N N H Me CO NH H Ph

257 stereochemistry of the B/C ring junction in the 1,2,3,3a,8,8a-hexahydro-3a-methylpyrrolo [2,3-b]indole ring system is presented in (see footnote 13)). These are then separated by column chromatography (Brossi 1985; Kawabuchi et al. 1988; Yu and Brossi 1988) [silica gel/CH2Cl2-MeOH (100:1 to 80:1)] and then decomposed by boiling, under reflux and an atmosphere of nitrogen for 2h (Brossi 1985; Schönenberger and Brossi 1986; Schönenberger et al. 1986a, b; Yu and Brossi 1988), with either 1M sodium pentoxide in “pentanol” or sodium butoxide in “butanol”, or in boiling pentan-l-ol or butan-l-ol alone under reflux (Kawabuchi et al. 1988; Schönenberger et al. 1986ab) to afford the component optical isomers of the dl-N1noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me) of excellent optical purity and in high yields [besides 255 (R¼Me(CH2)4O-CO-NH and Me(CH2)3O-CO-NH, respectively) which can be reconverted into 255 (R ¼ O¼C¼N) by hydrolysis to afford 255 (R¼NH2) and then reaction of this with phosgene] (Schönenberger et al. 1986a, b). Resolution can also be directly effected chromatographically on a preparative scale on chiral stationary phases. This topic has received comprehensive coverage in reviews (Anton et al. 1994; Brossi 1994; Brossi et al. 1996; Francotte 1994; Francotte and Junker-Buchheit 1992). Specific examples involve the resolution of the Julian and Pikl nitrile 19 (R1¼MeO, R2¼CH2CN) using a column of microcrystalline cellulose triacetate as the stationary phase and elution with 96% ethanol to afford good yields of the enantiomers 253 and 254 (Brossi et al. 1996; Pei et al. 1996) and a similar analytical chromatographic separation using cellulosebased chiral stationary phases has likewise been affected (Brossi et al. 1996, Lee TBK and Wong 1990) of a variety of indolin-2-ones implicated in the synthesis of l-physostigmine by the Julian and Pikl approach. Furthermore, in another similar application but using 95% ethanol for elution, dl-physovenine (3, R1¼MeNHCO, R2¼Me, X¼O) was separated into its enantiomers 95 (R¼MeNHCO, X¼O) and 98 (R1¼MeNHCO, R2¼Me, X¼O, n¼1) (Yu et al. 1991). Finally, it is important to note that resolution should be effected as early as possible in the reaction sequence (Pei et al. 1996; Witkop 1998) in order to minimise the effect of the wastage of “an equal amount of the undesired enantioner, which has to be discarded” (Brossi et al. 1996). Indeed, work has “focused on the first prochiral intermediate in the Julian [Julian and Pikl] synthesis, the nitriles [253] and [254], and a great variety of analogues” (Brossi et al. 1996), the synthesis of which can be asymmetrically directed using phase-transfer catalysts (see footnote 20) and which can be separated by chromatography on a microcrystalline cellulose triacetate column (Brossi et al. 1996; Pei et al. 1996) (vide supra). 22. “The yield from this reaction was greatly increased when the reaction mixture was stirred continuously throughout the boiling period” (Dale and Robinson 1970), a similar O-demethylation having also been affected when the reaction mixture was boiled under reflux with stirring under an atmosphere of nitrogen (Hill and Newkome 1969). However, attempted O-demethylation of homoesermethole [racemic (98, R1¼R2¼Me, X¼NMe, n¼2)] to produce the corresponding phenol by heating with “anhydrous aluminium chloride in gasoline

Notes

113

(b.p. 70-77o)” was unsuccessful, as was the same attempted cleavage of the ether bond by heating with aniline hydrochloride and by treatment of either the base or its hydrobromide with either hydrobromic acid or hydrochloric acid and with variation in the concentration of acid, reaction temperature and length of heating – apparently, “demethylation proceeded only under condition which led to the formation of an inseparable mixture of substances” (Kolosov et al. 1953). Perhaps O-demethylation in this case might be possible using boron tribromide in dichloromethane which was later (Luo et al. 1990; Robinson et al. 1987, 1988; Schönenberger et al. 1986a; Shishido et al. 1986b; Yu and Brossi 1988) used to effect such transformations in good yields. However, in the absence of this possibility, Kolosov et al. (1953) introduced the methoxymethyl moiety as a protecting group since the resulting methoxymethylether “was, like all acetals, stable enough in alkaline, and readily decomposed in acidic media” from under which conditions the desired product, homoeseroline [racemic (98, R1¼H, R2¼Me, X¼NMe, n¼2)], could be isolated. Similar properties are inherent in the acetal function of the product when using the tetrahydro-2-pyranyl moiety 251 as the protecting group (see footnote 16) and which can be easily removed with 2N-hydrochloric acid (Brossi et al. 1996; Pei et al. 1996; Yu et al. 1994). 23. Percy Lavon Julian (1899–1975) It is poignant to note that the funding of this research [(Julian and Pikl 1935abc), Sect. 3.1] appears to reflect then current social attitudes prevalent in the USA in that the concluding paragraph of the final paper announcing the development of the synthesis – claimed to be “undoubtedly the most significant chemical research publication to come from De Pauw” (Anon 1999) – reads “In acknowledging a generous grant from the Rosenwald Fund, the senior author respectfully dedicates this finished project to the memory of Julius Rosenwald, who has made possible innumerable cultural contributions on the part of young negroes to his country’s civilization” (Julian and Pikl 1935c). Indeed, when he had the pleasure of meeting with this most delightful man at his Julian Research Institute (vide infra) in Chicago on 22nd June 1972, it became apparent to the author that Dr Julian, as a result of his being an Afro-American [he had been born on 11th April 1899 in Montgomery, Alabama, the second generation out of slavery (Anon 1999; Borman 1993)] had, during the early period of his life, for many years experienced and been the victim of the racial segregation and prejudice that was then practised within the USA, a fact that was later verified during a televised biography of him (The Percy Julian Biography Project 2002). His father, James Sumner Julian, was a railway clerk and that his mother, Elizabeth Lena Julian (née Adams), was a school teacher resulted in the young Julian – in a society where only 41% of black children attended school beyond the age of ten years and more than 50% of the black population was illiterate – staying on in school and doing well (Borman 1993). Another acknowledgement of gratitude, but in this instance appended to each of the four publications (Julian et al. 1934; Julian and Pikl 1935abc) leading to the l-physostigmine synthesis, is directed toward Dr WM Blanchard, Dean of the College of Liberal Arts and Senior Professor of Chemistry at DePauw University. The significance of this appreciation not only lies in the fact that it was Blanchard who first imbuded Julian with his predilection for chemistry whilst he was an undergraduate at DePauw – where, in the Autumn of 1916, aged 17, he had been accepted as a sub-freshman (at which juncture his family, in order to enable them to give him their full support, moved with him to Greencastle, Indiana), and from where, in 1920, he graduated BA as valedictorian (ie at the top of his class), despite, in addition to his regular university courses, his having to take remedial classes at a nearby high school – but later created the opportunity for him, after he had left Howard University under a cloud of university politics that were probably racially motivated (vide infra), to do the research at DePauw that led to his celebrated first synthesis of l-physostigmine (Sect. 3.1) (see footnotes 24 and 25). Julian’s career received help from two North American philanthropic organisations, namely the General Education Board (that was financed by the Rockefeller family) and the Julius Rosenwald Fund (Rosenwald was the chairman of the Sears Roebuck catalog sales company), both organisations which played significant roles between 1900 and 1940 in improving the educational opportunities for Afro-Americans by supporting and stimulating the construction of large numbers of appropriate schools across the South of the USA. Although these benefits were too late to help Percy Julian, both these organisations also made selective awards in support of up-and-coming black scientists and amongst whom he was to be included. However,

114

2 l-Physostigmine (Eserine) in the meantime after his graduation from DePauw (vide supra), and because of the lack of future employment opportunities, he spent two years as instructor in chemistry at Fisk University, in Nashville, Tennessee and in 1922, with the help from Blanchard at DePauw (vide supra), under the auspices of an Austin Fellowship he commenced graduate studies at Harvard University from where, after one year, he graduated MA and finished, yet again, at the top of his class but on the campus of which race became a major issue and once again raised its ugly head, with his application for a teaching assistantship being repeatedly rejected lest Southern students were offended by having a “Negro” teacher (Borman 1993). Consequently, Julian, supporting himself with minor Fellowship positions and odd jobs – such as stoking furnaces – outside the university, remained as a researcher at Harvard until 1926 when he moved to West Virginia State College for Negros as professor of chemistry and then, two years later, to the predominately black Howard University in Washington DC as associate professor of chemistry and where he quickly rose to full professor and the head of the chemistry department. It was from here in 1929 that he at last threw off the yolk of racism and, frustrated in his efforts to pursue doctoral studies in the USA, moved as a Rockefeller Foundation Scholar, with the associated financial support from the General Education Board (vide supra), to the University of Vienna where, under the direction of Ernst Späth, he resumed graduate studies and began his doctoral research in the area of the chemistry of natural products, specifically on the indole alkaloids of Corydalis cava (Späth and Julian 1931), a plant that grows in the Vienna Wood. Subsequent upon ultimate graduation as PhD (Vienna) in 1931 as another part of his doctoral research [see, for example (Julian 1931)] he returned to Howard University as a Post-doctoral Research Fellow financed by the Julius Rosenwald Fund and accompanied by Josef Pikl, a friend from his days in Vienna (Witkop 1998). After two years at Howard, internal politics forced them to leave, at which juncture, once again through the efforts of William Blanchard (vide supra), Julian, accompanied by his Julius Rosenwald Post-doctoral Research Fellowship, returned to DePauw and it was here in the Minshall Laboratory in 1935, in collaboration with Josef Pikl, who was then working as an assistant in chemistry (Anon 1999), that he completed the research that resulted in the first total synthesis of l-physostigmine (Sect. 3.1) and (see footnotes 24 and 25) and that established his reputation, at the age of thirtysix years, as a world-renowned organic chemist and an internationally recognised published researcher. However, despite these accomplishments and the recommendations and support of Dean WM Blanchard (vide supra) and DePauw President G Bromley Oxnam, the university’s Board of Trustees rejected Julian’s application for a faculty teaching position – sadly DePauw was not ready for a black professor – and the University of Minnesota likewise rejected him. Thus frustrated in his attempts to secure an academic position, Julian turned his attention to industry and was offered a research position at the Institute of Paper Chemistry at Appleton, Wisconsin. However, the Institute’s officials became aware of an Appleton statute that forbade “housing of a Negro overnight” and, consequently, “Julian’s career at the Institute ended before it began” (Borman 1993). However, a door was soon to be opened which was to lead to a long-term and very successful industrial career and came about as a result of an attempt by Julian – whilst still working at DePauw University (Anon 1999, Witkop 1998) – to isolate geneserine (Chap. 6) from Calabar beans which ultimately led to his isolation, from the component oil of this botanical source, of the hydrate of stigmasterol (175, R¼H) (Chap. 9). Since the most common source of this sterol was soy bean oil, Julian sought five gallons of this oil from The Glidden Company in Chicago. It was at this juncture that a white vice-president at Glidden’s, WJ O’Brien, being aware of Julian’s excellent scientific reputation and his rebuff at Appleton (vide supra), offered him the position as Assistant Director of Research at Glidden’s Soya Product Division, an offer that Julian gladly accepted in 1936. Thus began a remarkable industrial career in the vibrant city of Chicago where he was to spend the rest of his life and which has been superbly reviewed by one of Julian’s friends – over a time span of almost sixty years which went back to the first synthesis of l-physostigmine (Sect. 3.1), Bernhard Witkop (1980, 1998). The numerous patents that emanated over eighteen years from his soy bean protein research work at The Glidden Company, Soya Product Division, and which led after the first

Notes

115

four years to the Division becoming the company’s most profitable single entity, produced many successful products for Glidden, including a paper coating and a fire-retardant foam which saved many lives from its wide use in Aerofoam by US troops during World War II to fight and extinguish petroleum and oil fires. His first patent (Julian et al. 1940), which resulted from a fortuitous isolation of stigmasterol from soy bean oil, ultimately led to synthetic work [see, for example, (Julian et al. 1949, 1950, 1951, 1956a)] which produced in large quantities and at low cost progesterone [see (Witkop 1998)] and cortisone, following the epochal joint Nobel Prize-winning (in 1950) discovery by Edward C Kendall, Philip S Hench and Tadeus Reichstein (Sourkes 1966) that the latter product reversed the symptoms of rheumatoid arthritis. Eventually, Julian became research director in Glidden’s soya products division, and also concurrently acted as manager of their fine chemicals division and research director of its Durkee Famous Foods division (Borman 1993). In 1953 Glidden’s decided to cease all work on steroids and sold Julian’s patents to Syntex and Pfizer, two of the companies involved in the “Great Steroid Wars of the 1950s”. Consequently, in that same year Julian resigned from The Glidden Company to establish his own laboratory, Julian Laboratories, Inc, in Franklin Park, Illinois which specialised in steroid chemistry and the production of therapeutical steroids from Mexican yams. In 1961 he sold this remarkably successful company to Smith, Kline and French Laboratories for more than 2.3 million US dollars whilst remaining as president on a generous salary – later he repurchased one of the laboratories from Smith, Kline and French and established a non-profit making research organisation in Chicago called the Julian Research Institute. From here, via an arrangement with the graduate center at the University of Illinois, he directed the doctoral research of post-graduate students (Borman 1993). Throughout Percy Julian’s remarkable multifaceted professional career runs an unbroken series of publications relating to various aspects of the chemistry of indoles. This begins – as part of his doctoral research at the University of Vienna (vide supra) – with an application of that doyen of the indole syntheses, namely that discovered by Emil Hermann (Farber 1970; Lucier 1993) Fischer (Robinson 1982), to methyl 3,4-dimethoxyphenylethyl ketone phenylhydrazone to furnish 3-(3,4-dimethoxy)benzyl-2-methylindole (Julian 1931). This was followed in turn by a sole contribution – and, in addition, a “non-indolic paper” (Julian and Passler 1932) – from Howard University (vide supra) (Julian and Pikl 1933) and from DePauw University by the celebrated quartet (Julian et al. 1934, Julian and Pikl 1935a, b, c) which ultimately afforded l-physostigmine (Sect. 3.1) and by (Julian et al. 1935 and Julian and Pikl 1935d). Then a series of papers, several concerning yohimbine (Julian et al. 1948a, b, c; Julian and Magnani 1949), were published jointly from DePauw University (vide supra) and from the Research Laboratories of The Glidden Company, Soya Products Division (vide supra) (Julian et al. 1948a) and from the latter organisation alone (Julian et al. 1945, 1948b, c, 1953, Julian and Magnani 1949; Julian and Printy 1949, 1953) and from Julian Laboratories, Inc (vide supra) (Julian et al. 1956b, c). As well as his work on the synthesis of yohimbine alkaloids, he also effected investigations into the metabolism of the amino acid tryptophan (Anon 1999, Julian et al. 1935, 1956b, c, Witkop 1978, 1983). Also from the Research Laboratories of The Glidden Company, Soya Products Division (vide supra) he co-authored (Julian et al. 1952b) a review on “THE CHEMISTRY OF INDOLES” [which incorporates references to some of his as then unpublished studies (Julian 1933, 1952; Julian and Magnani 1952; Julian and Passler 1952; Julian and Printy 1952; Julian et al. 1952a, c)] and which, alongside that by Sumpter and Miller (1954b), provided not only invaluable contemporary comprehensive converages of the subject but still remain as classical texts of considerable utility and interest. Furthermore, an illustration captioned “Julian discusses ongoing research project at Julian Research Institute [vide supra] in 1968” (Borman 1993) shows to his left a blackboard liberally festooned with indole-related chemical structures – clearly, “once an indole-man, always an indole-man!” Further biographical data on Percy Lavon Julian – grandson of slaves, chemistry teacher, teacher and researcher, researcher and administrator, pioneer in the chemical synthesis of therapeuticals, successful industrial research director and manager, millionaire entrepreneur and researcher, the first Afro-American to achieve a supervisory research position within a major American corporation (vide supra) and to serve as a faculty member of a non-Negro

116

2 l-Physostigmine (Eserine)

college in the USA (vide infra), and civil rights steadfast advocate – who overcame prejudice, blatant discrimination in employment, and racist threats including two attacks by bombers and arsonists (fortunately without injury to either himself or any member of his family) on his home in Chicago where he and his family had been the first Afro-Americans to purchase a home in all-white Oak Park, Illinois – can be found in (Anon 1999; Borman 1993; DeKruif 1946; The Percy Julian Biography Project 2002; Witkop 1980, 1998). Toward the end of what must be described as an eventful and extraordinary life, in 1974 Julian became ill and it was necessary for him to scale back his activities. He died, aged 76 years, on 19th April 1975. From Julian’s research in “pure” and “applied” chemistry emanated, respectively, 51 publications from 1931 to 1969 and 117 patents variously issued in Australia, France, Germany, Great Britain and the USA (Witkop 1998). He received 19 honorary degrees – whilst alive or posthumously, in 1973 he was elected to the National Academy of Sciences, and was paid a rare philatelic tribute when he was honoured, in the US Postal Services black heritage commemorative series, by the release on 30th January 1993 of a new 29 cent stamp for nationwide sale (Borman 1993). He also played an active role in the National Association for the Advancement of Colored People, The Chicago Urban League and the Mental Health Association of Greater Chicago and was a trustee at six colleges and universities, included in which was the Board of Trustees of DePauw University to which he was appointed in 1967 –by a “school that had declined to hire him as a faculty member more than 30 years before” (vide supra) (Borman 1993). At this later juncture, “planning began for a new science building, which would replace the 65-year-old Minshall Laboratory [where the celebrated first synthesis of l-physostigmine had been effected (Sect. 3.1)], and construction commenced in 1968. The Science and Mathematics Center was dedicated in September 1972, with Percy Julian giving the dedication address, ‘Science and the Good Life of Man’ ” and “Following Julian’s death, DePauw University named the Percy L. Julian Science and Mathematics Center in his honor” (Anon 1999). Also at DePauw, an annual lecture series in his honour was, in 1993, in its seventeenth year, and a scholarship established by his family has provided help and support to many of that university’s students (Borman 1993). The author’s memory returns to 22nd June 1972 (vide supra) when, at the Julian Research Institute in Chicago (vide supra), he had the great pleasure and privilege to meet and spend a most delightful day with Dr Julian, a person who he fondly remembers with much respect and great affection. 24. In fact, the soundness and versatility of this synthetic approach were soon – indeed, at its outset – recognised when, included in their earliest publication (Julian et al. 1934) – announcing the conversion of 19 (R1¼R2¼H) into 23 (R1¼R2¼H, R3¼R4¼Me), Julian and Pikl (along with Doyle Boggess) stated that “Since, moreover this nitrile [19 (R1¼H, R2¼CH2CN)] is obtained in 90% yield and [19 (R1¼R2¼H)] is a very cheap substance, the way is opened to a ready synthesis of the drug itself and even of homologs, containing other groups in the 3-position of the indole nucleus” (Julian et al. 1934) and confidently predicted that “We hope soon to communicate the complete synthesis of the drug itself by this same procedure” (Julian et al. 1934). This goal was ultimately referred to (Hino 1961b) as a “brilliant synthesis of physostigmine” and, furthermore, as recently as 1998 it has been stated (Witkop 1998) that “Although there have been more than half a dozen subsequent syntheses of physostigmine [vide infra], Julian’s classical approach, with some recent modifications and improvements (see footnote 16, 17, 20, 21, 22), is still the best route to the synthetic alkaloid....”, an opinion with which the author concurs (see footnote 25). The recent improvements (Yu et al. 1994) (see footnote 16, 17, 20, 21, 22) to the Julian and Pikl synthesis were mostly developed at the National Institutes of Health in Bethesda, Maryland, USA, at the Shanghai Institute of Organic Chemistry, China, at Georgetown University in Washington DC, USA and at the Chemical Research Department of Hoechst-Roussel Pharmaceuticals, Inc in Somerville, New Jersey, USA, and all of which have been the subject of review (Brossi et al. 1996, Greig et al. 1995a). Indeed, using these improvements, the Julian

Notes

117

and Pikl approach to the total synthesis of the ring system of l-(3a,S–cis)- physostigmine and its optical isomer has been achieved on a large and technical scale (Pei et al. 1996, Brossi 1992), as has that of phenserine (98, R1¼PhNHCO, R2¼Me, X¼NMe, n¼1) (Sect. 10.7.2) (Brossi et al. 1996). Furthermore, applications of the synthesis to that of 3a–aryl substituted 1,2,3,3a,8,8ahexahydropyrrolo[2,3–b]indoles proceed via the 3-cyanomethylation of the appropriate 3-arylindolin-2-ones as has been exemplified using 3-phenylindolin-2-one [interestingly in the presence of sodium hydride instead of sodium ethoxide as used by Julian and Pikl (1935a, b, c) – the alklation of indolin-2-one at the 3-position has also been well-investigated by others (Hino 1961a)] (Sandoz Ltd 1966) and 5-methoxy-l-methyl-3-phenylindolin-2-one which was prepared via a Stollé synthesis starting from reaction of 4-methylaminophenol (17, R1¼OH, R2¼Me) [an interesting and useful circumvention of the need to protect the hydroxyl function, as also utilised in another instance (Yu et al. 1994)] with 2-bromo-2-phenylacetyl chloride) (Pei et al. 1998). 25. The synthesis’s most recent accolade was commemorated by its designation, conferred by the American Chemical Society, as a National Historic Chemical Landmark (Anon 1999). A plaque marking this event was presented to DePauw University on 23rd April 1999, during the university’s celebration of the centennial anniversary of Percy Julian’s birth, and reads (Anon 1999):-

NATIONAL HISTORIC CHEMICAL LANDMARK SYNTHESIS OF PHYSOSTIGMINE DePauw University Greencastle, Indiana 1935 In 1935, in Minshall Laboratory, DePauw alumnus Percy L. Julian (1899–1975) first synthesised the drug physostigmine, previously only available from its natural source, the Calabar bean. His pioneering research led to the process that made physostigmine readily available for the treatment of glaucoma. It was the first of Julian’s lifetime of achievements in the chemical synthesis of commercially important natural products. American Chemical Society April 23, 1999 26. Similar in mechanistic principle to this ring closure and the related earlier studies is the reaction between Koshland’s protein reagent, 2-hydroxy-5-nitrobenzyl bromide (a reagent which reacts specifically with tryptophan residues in a protein chain) and tryptophan which was shown (McFarland et al. 1969), almost exclusively on the basis of 1H-nmr spectroscopic data, to afford a mixture of two isomeric products, 258 and 259 – although no structure was proposed for a minor reaction product. It is obvious that the formation of ring C in 258 and 259 results from the

HO CH2 C N N CO2 H H H H

258

NO2

118

2 l-Physostigmine (Eserine)

HO CH2 C

NO2

N CO N 2 H H H H

259 electrophilic attack of the 2-hydroxy–5–nitrobenzyl cation at the 3-position of the indole nucleus, followed by ring closure involving nucleophilic addition of the nitrogen atom of the 3-substitent at the C-2 atom of the 3H-indolium cation so formed (Robinson 1917b 1971). However, it would appear that the amino group involved in this cyclisation does not need to be free since N-acetyltryptamine and N-acetyl-L-tryptophan methyl ester also afford similar structurally-related products (Spande et al. 1968). That from the former reaction was acetylated to yield a triacetyl derivative for which structure 260 (R¼H), rather than the ring-C opened 3Hindole 261 (R¼H) was preferred (Spande et al. 1968) although structural confirmation by X-ray

OCOMe CH2 C NO2 N N R CH2 COMe OCOMe O2N

260

O2N OCOMe CH2 N MeOC

R N CH2 OCOMe

O2N

261 crystallography is awaited (Robinson 1971; Spande et al. 1968). Three products were isolated from the latter reaction followed by acetylation. The major product, after separation by chromatography, appears to have structure 260 (R¼COOMe) from the comparison of its uv – and 1H-nmr – spectral properties with these of 260 (R¼H) (Spande et al. 1968) – the

Notes

119

structure of the minor reaction products remain to be determined (Robinson 1971; Spande et al. 1968). A mechanism similar to that involved in the formation of 258 and 259 (vide supra) must be operational in the formation of compound 262 from the reaction between N-acetyltryptamine

Me CH2

CH C Me

N H

N COMe

262 and 3,3-dimethylallyl bromide (Casnati et al. 1969), but now with the initial stage involving an electrophilic attack by a 3,3-dimethylallyl moiety (Casnati et al. 1969; Robinson 1971). These above reactions, together with other mechanically-related transformations, have already been subjected to review (Fontana and Toniolo 1976; Hino and Nakagawa 1988). The above cyclisations leading to 258 and 259 are analogous to those that have been long established as following the 3-alkylation of suitably activated tryptamines (Sect. 3.1). Furthermore, similar ring closure also follows electrophilic addition at the indolic 3-position of tryptamines and tryptophans by the diazonium salt derived from 2-methoxy-4-nitroaniline (“Fast Red B”) (Fontana and Toniolo 1976), a proton (see footnote 4 in Chap. 3), positive halogen [such as Cl⨁ derived from t-butyl hypochlorite (see footnote 4 in Chap. 3), Br⨁ derived from N-bromosuccinimide (see footnote 4 in Chap. 3) and I⨁ derived from iodine azide (Ikeda et al. 1979)], and of singlet and positive oxygen (vide infra). These latter pair provide a synthetic route to the 1,2,3,3a,8,8a-hexahydro -3a-hydroxypyrrolo[2,3-b] indole system – that occurs naturally in brevianamide E (228, R¼CMe2CH¼CH2) (Birch and Russell 1972; Birch and Wright 1970; Hino and Nakagawa 1988; Nakagawa et al. 1975; Sammes, 1975), flustraminols A and B (Christophersen 1985b); hunteracine [Herbert 1983(b); Husson 1983; Nakagawa et al. 1975], okaramine (232) (Hayashi et al. 1989 and sporidesmin A 227) (Fridrichsons and Mathieson 1965; Nakagawa et al. 1975) – which has been effected by photooxidation of tryptamines, and to the corresponding 2-carboxy- and 2-carbomethoxy analogues by either peroxyacetic acid oxidation [in water using one mole of preformed oxidant at 0–5 oC for 24h (Savige 1975)] or photoxidation of either tryptophans or their methyl esters (Fontana and Toniolo 1976; George and Bhat 1979; Saito et al. 1977), respectively. In two further interesting examples of these transformations, compound 263, formed (Hino and Nakagawa 1988) via a Meisenheimer rearrangement and which is structurally analogous to lgeneserine (Sect. 6.1), was isolated as one of the products resulting from the Rose Bengal sensitised photoxidation of N1-methyltryptamine (George and Bhat 1979; Nakagawa et al. 1975; Saito et al. 1977) and the photooxidation of tryptophol afforded a near quantitative yield of 264 (R¼OOH) which is structurally analogous to l-physovenine (Sect. 3.1) and which readily decomposed under a variety of conditions to afford a mixture of products including 264 (R¼OH) (George and Bhat 1979; Saito et al. 1977).

HO N H

O

263

N Me

120

2 l-Physostigmine (Eserine)

R N H

O

264 27. In view of the results obtained by Hoshino et al. (1934), the statement (Taylor 1966) – unfortunately made without reference to a primary source – that “Experimentally, however, treatment of 5-ethoxytryptamine with ethyl magnesium iodide followed by reflux with methyl iodide gave instead the methyl derivative of the ring open tautomer” [namely 29 (R1¼EtO, R2¼H, R3¼NMe2)] would appear to be erroneous. Indeed, it has probably emanated from confusion with the formation of this product, namely “methyl-eserethole”, by methylation of compounds already containing the 5-ethoxy-1,2,3,3a,8,8a-hexahydro-3a-methylpyrrolo[2,3-b] indole system, such as 23 (R1¼EtO, R2-R4¼H, R2¼R4¼H, R3¼Me and R2¼R3¼H, R4¼Me) (see footnote 12). 28. This claim was later shown to be erroneous since the product gave a picrate, small yellow prisms from “alcohol”, mp 180-181
C (King et al. 1934) which is clearly different from the picrate – orange crystals from “alcohol”, mp 150
C (Kolosov et al. 1953) [a later observation reported orange-yellow dice from ethanol, mp 162-163
C (Sugasawa and Murayama 1958a, b)] – of dl-esermethole (23, R1¼MeO, R2¼H, R3¼R4¼Me) that was independently and unequivocally synthesised using Julian and Pikl’s approach involving as the final step the reductive cyclisation of 19 [R1¼MeO, R2¼(CH2)2NHMe] using sodium in boiling either “absolute butyl alcohol” (Kolosov et al. 1953) or “dehyd. ethanol” (Sugasawa and Murayama 1958a, b). 29. This product gave a picrolonate as light brown, tiny, crystalline aggregates from “alcohol”, mp 220
C (decomp) (King and Robinson 1932b) while the picrolonate of dl-noresermethole (33, R1¼MeO, R2¼R3¼H, R4¼Me) that was to be unequivocally synthesised by an application of Hoshino and Kobayashi’s route (vide supra) has mp 221
C (Hoshino et al. 1934). Similarly it afforded a deep red picrate as long rhombic plates from methanol, mp 159
C (King and Robinson 1932b) which would appear to be identical with the picrate [reddish-orange rhomboidal prisms from “alcohol”, mp 162-163
(King et al. 1934) and dark red prisms from ethanol, mp 163-164
C (Robinson 1965a, b)] of dl-noresermethole that was later, respectively, obtained by the Oxford group (King et al. 1934) (vide supra) and subsequently as an intermediate in a synthesis of dl-eseramine (Sect. 4.2.1) via a Julian and Pikl type approach involving the reductive cyclisation of 19 (R1¼MeO, R2¼CH2CN) using sodium in dry ethanol boiling under reflux (Robinson 1917b, 1965a, b). 30. In view of the results from investigations referred to elsewhere (Sect. 6) and (see footnote 18), it might be expected that the intermediate 44 (R¼H) would undergo opening of its tetrahydropyrrolo ring C to afford 43 (R¼Me, X¼p-Tosyl). The latter is of a special interest in that it represents a monoquaternary derivative of a diamine in which it is the least basic nitrogen atom that is quaternised and therefore it might be expected to undergo methyl group transfer from its least basic to a more basicLnitrogen atom. Such a transfer could be intramolecular to afford 29 (R1¼EtO, R2¼H, R3¼ NHMe2 p-Tosyl), basification of which would give “methyl-eserethole” (29, R1¼EtO, R2¼H, R3¼NMe2). Alternatively, since the reaction was effected (King et al. 1933a) by mixing solutions of equimolar quantities of dl-noreserethole (23, R1¼EtO, R2¼R3¼H, R4¼Me) (in dry benzene) and methyl p- toluenesulphonate (in ethyl acetate), the methyl migration might be intermolecular, namely from 43 (R¼Me, X¼p-Tosyl) to 23 (R1¼EtO, R2¼R3¼H, R4¼Me) – since both these species would have co-existed in the reaction medium – to afford 29 (R1¼EtO, R2¼H, R3¼NHMe) and 44 (R¼H) –and this cycle could then continue – and the former product could undergo quaternisation with methyl pL toluenesulphonate to afford 29 (R1¼EtO, R2¼H, R3¼ NHMe2 p-Tosyl) which upon basification would yield “methyl-eserethole”.

Notes

121

In addition, under these reaction conditions it is likely that dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) [produced via the intermolecular quaternisation of 23 (R1¼EtO, R2¼R3¼H, R4¼Me (vide supra)] and methyl p-toluenesulphonate would have co-existed in the reaction medium and could have thereby reacted to yield 44 (R¼Me), opening of the tetrahydropyrrolo ring-C with subsequent intermolecular transfer of the methyl group from the quaternary nitrogen atom of the 3H-indolium cation thus formed to the more basic nitrogen atom of another molecule of dl-eserethole (23, R1¼EtO, R2¼H, R3¼R4¼Me) thus giving “methyl-eserethole” (29, R1¼EtO, R2¼H, R3¼NMe2) and 44 (R¼Me) – and this other cycle could, once again, then continue. It is now suggested that since both these above sequences involve the presence of methyl ptoluenesulphonate, they could explain the lack of formation of “methyl-eserethole” and the almost quantitative recovery of unchanged dl-eserethole as the only product resulting from compound 44 (R¼H) in the investigations by Jackson (1954) which are, unfortunately, not accompanied by relevant experimental details that are far from unequivocal from the quoted literature. Perhaps a clarification of this problem may be sought by an investigation of the product(s) resulting from a repeat of the study by Jackson (1954) (vide supra) in which dleserethole methyl p-toluenesulphonate (44, R¼H) is “ treated under exactly the same conditions” as were employed for the methylation of dl-noreserethole (King et al. 1933a) (see also King et al. 1934; King and Robinson 1935) but now in the presence of either ethyl ptoluenesulphonate or 14C-methyl p-toluenesulphonate . 31. The trivial name echiboline has been derived (Fritz and Fischer 1964) from the observation that this tetracyclic ring system is present in the alkaloid echitamine (216) (Robinson 1963a; see also Hamilton et al. 1962) (see footnote 9) and its positional numbering has been defined to be as shown in 89 (R1-R3¼H2, X¼H2, n¼2) (Fritz and Fischer 1964). A facile synthesis of this parent compound has been effected (Fritz and Fischer 1964; Robinson 1971) by cyclisation of 4a-(2-aminoethyl)-1,2,3,4,4a,9a-hexahydro-4aH-carbazole, formed by treatment with hydrazoic acid of 4a-(2-carboxyethyl)-1,2,3,4,4a,9a-hexahydro-4aH-carbazole which was one of the two products obtained by Fischer indolisation of the phenylhydrazone of 71 [R3+R4¼(CH2)4, R5¼COOH, n¼1]. This reaction sequence as might that starting with the Fischer indolisation of 87 (n¼2) (Sect. 3.2), also be applicable to the synthesis of 6-hydroxyechibolines with antinociceptive activity (Sect. 10.11). Another synthetic route to such latter products could be attempted – using an approach based upon that developed by Hoshino and Kobayashi, with their occasional co-workers (Sect. 3.1) – by reacting 1,2,3,4tetrahydro-6-methoxycarbazole with either ethyl or methyl magnesium iodide followed by N-methylaziridine –a nucleophile suggested (Onaka 1971) for then future use in this type of reaction – with subsequent hydrolytic “work-up”. This should afford 89 (R1¼MeO, R2¼H, R3¼Me, X¼H2, n¼2) and ultimately lead to 6-hydroxyechibolines in which the N(9)-substituent could be other than a methyl group. 32. It is interesting that whereas reduction of 6-methoxy-9-methyl-11-oxoechiboline (see footnote 31) (89, R1¼MeO, R2¼Me, R3¼H, X¼O, n¼2) using lithium aluminium hydride in boiling tetrahydrofuran under reflux afforded 6-methoxy-9-methylechiboline (89, R1¼MeO, R2¼Me, R3¼H, X¼H2 n¼2), an attempt to effect a similar reduction of a suspension of 89 (R1¼MeO, R2¼Me, R3¼H, X¼O, n¼2) in boiling diethyl ether under reflux, in which it is only very slightly soluble, was unsuccessful and when, under these conditions, a Soxhlet apparatus was employed to effect dissolution, an unexpected reductive ring scission occurred along with reduction of the carbonyl group to afford 4a-(2-aminoethyl)-6-methoxy-9-methyl1,2,3,4,4a,9a-hexahydrocarbazole (Cox and Robinson 1988). Since it is well-established (see footnote 18) that reductive cleavage of the Ph-N-C-N system only occurs under acidic conditions, probably via acid-catalysed ring opening to the corresponding 3H-indolium cations (Jackson and Smith 1984), the possible mechanism for the ring opening in this instance was discussed although the reason(s) for the relationship between the nature of the reaction product and the reaction conditions remained unclear (Cox and Robinson 1988).

122

2 l-Physostigmine (Eserine)

33. Indeed, it has been claimed (Kutney 1977) that the Fischer indole synthesis (Robinson 1982) is “the most versatile and widely applied reaction particularly in the syntheses of natural alkaloids”. For examples, it has been employed in the initial stage of the synthesis of strychnine (Kutney 1977; Woodward et al. 1954, 1963), as early stages in those of harmalan and tryptophan (Julian et al. 1952b), in various intermediary stages of those of dl-dasycarpidone, dl-3-epi-dasycarpidone, dl-uleine and dl-3-epi-uleine (Jackson et al. 1969; Kutney 1977), ellipticine (Hewlins et al. 1984; Sainsbury 1977; Stillwell 1964), harmaline (Sumpter and Miller 1954b) and the β-carboline alkaloids (Kutney 1977), in the antepenultimate stage of syntheses of dl-aspidospermine (Ban et al. 1965; Stork and Dolfini 1963 – see also Ban et al. 1969; Inoue and Ban 1970; Kutney 1977; Robinson 1982; Stevens 1977), in the penultimate stage of the synthesis of dl-quebrachamine (Stork and Dolfini 1963 – see also Ban et al. 1969; Kutney 1977; Stevens 1977) and as the final stages in two (Ikezaki et al. 1969; Sallay 1967 – see also Augustine and Pierson 19690; Kutney 1977) syntheses of dl-ibogamine, in that of cinchonamine (Ch’ang-pai et al. 1958, 1960 – see also Robinson 1982) [which, furthermore, had earlier (Witkop 1950) been converted into quinamine using dilute peracetic acid] and the ring-system of yohimbine (Clemo and Swan 1946; Julian and Magnani 1949; Kutney 1977). Furthermore, its applications in Borsche-type syntheses (Robinson 1982), which have also been claimed [Joule 1983(h)] as affording the best method of preparing the ring A-substituted dimethylcarbazoles for use in the Saxton route to the synthesis of ellipticine and its analogues, has led to the carbazole (so-called) (see footnote 4) alkaloids (Kapil 19710; Kutney 1977), in (so-called) abnormal “advanced” Fischer indolisations (Ishii 1980, 1981) to the naturallyoccurring 6-isoprenylindoles (Ishii and Murakami 1975 – see also Ishii 1981), and have also been involved in the syntheses of 5-hydroxytryptamine, 5-hydroxytryptophan and 3-indolylacetic acid, together with many of their analogues and homologues (Julian et al. 1952a, b, c; Robinson 1982). It has also be utilised in some of the synthetic approaches that have been made to the 3a-alkyl-3,3a,8,8a-tetrahydro-2H-furo[2,3-b]indole ring system that is present in physovenine (Sect. 3.2.3) (Robinson 1982). 34. The difference in behaviour of this system in l-physovenine by comparison with that reported (Bardsley 1961) for compound 264 (R¼Me) (Hoshino and Shimodaira 1935) which shows indoline absorption in neutral ethanol, benzenoid absorption in dilute acid and a mixture of benzenoid and 3H-indolium cation absorption in concentrated acid is probably because of the absence in this latter compound of the MeNHCOO group that is present at the 5-position of lphysovenine [cf (Joule and Smith 1962)] (Robinson 1964a). Moreover, and interestingly, it has been concluded (Jackson and Smith 1964) that “It is clear that formation of these indolium salts has been, and will continue to be, a useful diagnostic test for the presence of the Ph-N-C-N and Ph-N-C-O systems in newly discovered indole alkaloids. The two systems can be distinguished from each other by the fact that only the former can undergo quaternisation in dilute acids, and hence give the characteristic small hypsochromic shift first observed by [GF] Smith”. Indeed, such examinations played a major role not only in the elucidation of the structure of lphysovenine as 3 (R1¼MeNHCO, R2¼Me, X¼O) (Robinson 1964a) (Sect. 3.1) but also in establishing the structure of the product resulting from the treatment of 69 (R1¼Et, R2¼⨁NMe3 IƟ, X¼O) with boiling caustic soda as 98 (R1¼Et, R2¼Me, X¼O, n¼1), the uv spectrum of which in ethanol indicated the indoline nucleus, in dilute ethanolic hydrochloric acid was characteristic of a mixture of indoline and 3H-indolium cation chromophores, and in concentrated hydrochloric acid had pure 3H-indolium cation absorption (Longmore 1969; Longmore and Robinson 1966) (Sect. 3.2.2).

Chapter 3

l-Physovenine

3.1

Isolation and Structure Elucidation

Unlike the investigations into the structure of l-physostigmine which followed the then approach involving degradation into products of known structure with subsequent retrosynthetic interpretation (Sect. 2.2), the structures of the other alkaloids of the ripe seeds of Physostigma venenosum, including that of l-physovenine, were, consequent upon the availability of the appropriate instrumentation, established using spectrophotometric and spectrometric techniques with the use of l-physostigmine as a model reference compound. The first isolation of l-physovenine was reported from the Wellcome Chemical Research Laboratories and resulted from the fractional crystallisation of a basic extract of Calabar beans, when the alkaloid was found to have mp 123
C and empirical formula C14H18N2O3 (Salway 1911) and it was noted that when it “is treated with barium hydroxide, there is an immediate precipitation of barium carbonate, and the solution rapidly assumes a deep red colour” (Salway 1911). These latter observations in which the “behaviour of physovenine is very similar to that of physostigmine, since the latter under the influence of alkalis (Ehrenberg, loc. cit.) loses carbon dioxide and methylamine, and becomes converted successively into eseroline and a red colouring matter, rubreserine” (Sect. 2.8.2) led to the suggestion (Salway 1911) that it “appears probable from the properties and composition of physovenine that the latter is an intermediate product in the conversion of physostigmine [(3, R1¼MeNHCO, R2¼Me, X¼NMe)] into eseroline [(3, R1¼H, R2¼Me, X¼NMe)], as represented by the following equations: “

C15 H21 O2 N3 þ H2 O ¼ C14 H18 O3 N2 þ CH3  NH2 Physostigmine Physovenine C14 H18 O3 N2 ¼ C13 H18 ON2 þ CO2 Eseroline” ðSalway 1911Þ:

© Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_3

123

124

3 l-Physovenine

Indeed, the isolation of physovenine from the Calabar bean might be regarded as that of an artefact since it has been allegedly reported amongst the decomposition products from stored solutions of physostigmine (Paraak 1962). However, this claim may be open to doubt since from examination of the primary literature source reporting this study (Parrák 1962a, b) – rather than merely its abstract (Paraak 1962) – it would appear that the identification of the decomposition product as being physovenine seems to be far from unequivocal. It would certainly be of interest to repeat the appropriate sections of this investigation using authentic lphysovenine as a reference substance for identification purposes. Furthermore, it is perhaps significant that no reference is made to physovenine in the abstracts of either related studies seemingly by the same investigator (Parrak 1962; Parrak and Radejova 1962; Parrak et al. 1961) or of their apparent extension by others (Vincent et al. 1962). Surprisingly, a two dimensional paper chromatographic examination of an alkaloidal extract from Calabar beans which claimed (Sinha 1959) to have separated, detected and estimated physostigmine, geneserine, eseramine and – even – eseridine (Sect. 8.2), failed to detect the presence of physovenine. However, a second isolation of this latter alkaloid was effected by employing column chromatographic separation of the basic residues remaining from the then industrial extraction of l-physostigmine (the major alkaloid) from Calabar beans, at which juncture (Robinson 1964a) the previously-determined (Salway 1911) empirical formula was confirmed and shown, initially by Rast molecular weight determination and later by mass spectral molecular ion measurement, to be the molecular formula, the alkaloid was found to

have ½α22:5  92 (EtOH) and, mainly from a comparison of its ir, uv and 1H–nmr D spectra with those of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1),  together with biogenetic considerations (see Sects. 2.5 and 3.3) a structural proposal of 3 (R1¼MeNHCO, R2¼Me, X¼O) was advanced for it (Robinson 1964a). Strong, sharp absorption bands at 1751 and 3473 cm1 in the ir spectrum (in CCl4) of l-physovenine indicated the presence of the MeNHCOO group (C¼O and N-H stretching, respectively) [cf the similar bands in the corresponding spectrum (in CCl4) of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) at 1752 and 3474 cm1]. Since no other absorption occurred between 3600 and 3100 and between 1850 and 1625 cm1, the third oxygen atom in l-physovenine was likely to be contained in an ether-type linkage (Robinson 1964a). Furthermore, the complete ir spectra [which, in CHCl3 solution and using samples provided by the author, were later also published elsewhere (Neuss 1966)] of l-physovenine and lphysostigmine exhibited many features in common, thereby suggesting that the two alkaloids were structurally closely related (Robinson 1964a). The uv spectra (in EtOH) of l-physovenine [λmax 252 and 310 nm (ε ¼ 13,200 and 3300, respectively)] and l-physostigmine [λmax 253 and 310.5 nm (ε ¼ 12,100 and 2800, respectively)] [these two spectra, using samples provided by the author, were later also published elsewhere (Neuss 1966)] were almost superimposable, from which it could be deduced from earlier studies (Chalmers et al. 1958) that the former, like the latter alkaloid, had the MeNHCOO group substituted at the 5-position of an indoline nucleus, this

3.1 Isolation and Structure Elucidation

125

positioning of this substituent being further supported by the ir spectrum which had weak absorption bands at 1183, 1103, 1075 and 1063 cm1 indicative of a 1,2,4-trisubstituted benzene ring (Robinson 1964a). However, unlike the uv spectrum of l-physostigmine which in dilute acid retained indoline-type absorption but, relative to the spectrum in neutral solution, underwent a hypsochromic shift of approximately 10 nm owing to the protonation of N(1) in the Ph-N-C-N system (Sect. 2.6), the uv spectrum of l-physovenine in 1.5 N-ethanolic hydrochloric acid showed no such shift but had absorption characteristic of a combination of 3H-indolium cation and indoline chromophores, the former being indicative of the formation of the 3H-indolium cation 126 (R1¼Me, R2¼OH), L 1 2 similar to that, namely 126 (R ¼Me, R ¼ NH2Me), observed for l-physostigmine in strong acid (Sect. 2.6). This shows that l-physovenine does not contain a Ph-N-C-N system (Robinson 1964a). However, in strong acid, the uv spectrum of l-physovenine indicated that a total cleavage of the ring C occurred to afford 126 (R1¼Me, R2¼OH) as the only absorbing specie (Robinson 1964a) and thereby the presence of the Ph-N-C-O system in the alkaloid (see footnote 34 of Chap. 2). The 1H-nmr spectra of the two alkaloids exhibited the following significant differences and similarities:1. No signal was present in that of l-physovenine corresponding to the 3-proton singlet at 7.36τ in the spectrum of l-physostigmine and caused by the protons of the N(1)-CH3 group. The absence of such a group in l-physovenine was also apparent from the ir spectra, that of l-physostigmine having a medium absorption band at 2791 cm1 that was absent in the ir spectrum of l-physovenine and which is characteristic of the C-H stretching of an aliphatic N-CH3 group (Robinson 1964a). 2. The 1-proton singlet at 4.83τ in the spectrum of l-physovenine and caused by the C(8a)-proton is 0.94τ downfield relative to the corresponding 1-proton singlet in the spectrum of l-physostigmine, a shift expected of a proton attached to a carbon atom flanked by an oxygen and a nitrogen atom instead of by two nitrogen atoms (Robinson 1964a). It is of related interest that the corresponding protons (H*) in compounds 147 (R¼H and MeO) and in dihydrosterigmatocystin (148) gave rise R

R

H O

O H*

147

H O

OH O

O

OMe

148

O H*

126

3 l-Physovenine

to doublets centred at 3.90τ (J ¼ 6cps) (Knight 1965), 3.75τ (J ¼ 6cps) (Knight et al. 1966) and at 3.59τ (J ¼ 6cps) (Knight et al. 1965), respectively, these further downfield shifts being expected since these protons are now on carbon atoms flanked by two oxygen atoms. 3. In both l-physostigmine and l-physovenine, the C(2)- and C(3)-methylene protons form a ABXY system because of the rigidity of ring C (Robinson 1964a). However, whereas the 2-proton multiplets, attributed to the C(3)-methylene protons were in both cases at approximately the same τ values (between 8.09 and 7.81τ for l-physostigmine and between 8.04 and 7.74τ for l-physovenine), the 2-proton multiplet (between 6.74 and 5.76τ) caused by the C(2)-methylene protons in l-physovenine (which are attached to a carbon atom flanked by a carbon and an oxygen atom) was significantly downfield with respect to the multiplet (between 7.30 and 7.01τ) arising from the corresponding methylene protons in l-physostigmine (which are attached to a carbon atom that is flanked by a carbon and a nitrogen atom) (Robinson 1964a). 4. The remaining signals in the spectrum of l-physostigmine1 were 3-proton singlets at 8.53τ and 7.00τ (with corresponding singlets at 8.51τ and 7.02τ, respectively, in the spectrum of l-physovenine) which were attributed to the C-Me and aromatic N-Me protons, respectively, a doublet (J ¼ 4.8cps) at 7.16τ and 7.09τ superimposed upon the multiplet between 7.30τ and 7.01τ, with a 5-proton total intensity of the series of bands within this range which were attributed to the urethane N-Me protons and the C(2)-methylene protons, respectively, [in the spectrum of l-physovenine, the protons of the urethane N-Me group gave rise to a 3-proton doublet (J ¼ 4.8cps) at 7.17τ and 7.05τ] – in the spectra of both lphysostigmine and l-physovenine, the urethane N-H proton gave rise to a 1-proton broad band between 4.71τ and 4.08τ and between 4.91τ 4.33τ, respectively. In both alkaloids the three aromatic protons form a ABX system which in the spectrum of l-physostigmine gave rise to a multiplet between 3.19τ and 3.01τ of a 2-proton intensity and a multiplet between 3.65τ and 3.49τ of a 1 proton intensity, with corresponding multiplets between 3.22τ and 3.03τ and between 3.73τ and 3.58τ in the spectrum of l-physovenine. Needless-to-say, these comparative 1H–nmr data lent even further support (Robinson 1964a) to the postulation of structure 3 (R1¼MeNHCO, R2¼Me, X¼O) for l-physovenine as, indeed, did subsequent (Longmore 1966a; Longmore and Robinson 1967; Reed 1964; Robinson 1968) examinations of it by mass spectrometry. The mass spectrum indicated (Longmore 1966a; Longmore and Robinson 1967; Reed 1964; Robinson 1968) a very easy loss of 57 mass units – caused by the loss of the elements of methyl isocyanate, from the molecular ion at m/e 262 [cf (Sect. 2.7) for a possible mechanism] – to give the base peak 130 (X¼O) at m/e 205. Further key fragments were also obtained at m/e 174, 160 and 147, confirming the presence of the ions 135 (R1¼R2¼Me), 135 (R1¼H, R2¼Me) and 149, respectively, with similar fragments having also been obtained from l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) (Sect. 2.7), 3,3a,8,8a-tetrahydro-3amethyl-2H-furo[2,3,b]indole (150, R1¼R2¼H, R3¼Me) (Clayton and Reed

3.2 Synthesis of the

127

HO N Me

149

R3

N

R2 O

R1

150 1963), and other structurally-related compounds (Clayton and Reed 1963) [see also (Budzikiewicz et al. 1964)] (Sect. 2.7).

3.2 3.2.1

Synthesis of the Racemate

Verification of structure 3 (R1¼MeNHCO, R2¼Me, X¼O) for physovenine that was postulated from its spectral examination as described in the previous section was soon forthcoming by synthesis (Longmore 1966a; Longmore and Robinson 1965, 1967; Robinson 1968). Following the strategy of Hoshino and Kobayashi [in which tryptamines, with both N- atoms at most secondary (as in 26), were subjected to 3-methylation via sequential formation, by reaction with either ethyl– or methylmagnesium iodide, of their adducts 27 with subsequent reaction with iodomethane in anhydrous ether, anisole or benzene boiling under reflux followed by hydrolytic “work-up” with acid – to afford the 3H-indolium cation 28 which upon basification yielded the corresponding 1,2,3,3a,8,8a-hexahydro-3a-methylpyrrolo [2,3-b]indole (Sect. 2.3.1)], tryptophol (151, R1¼R2¼H) in anisole upon sequential R1 N H

151

R2

OH

128

3 l-Physovenine

reaction with ethylmagnesium iodide and iodomethane with subsequent hydrolytic “work-up” afforded 3,3a,8,8a-tetrahydro-3a-methyl-2H-furo[2,3-b]indole (152, R4

R1 N

R3

O

R2

152 R1–R3¼H, R4¼Me) (Hoshino 1934;Hoshino and Shimodaira 1935), and likewise 2-methyltryptophol (151, R1¼H, R2¼Me) in ether afforded the product 152 (R1¼R2¼H, R3¼R4¼Me) (Nakazaki 1959). Consequent upon these observations, an initial synthetic approach toward dl-physovenine involved as its initial stage the reaction of 5-benzyloxytryptophol (151, R1¼PhCH2O, R2¼H) with methylmagnesium iodide in ether followed by iodomethane in anisole and then hydrolytic “work-up. However, none of the required product 152 (R1¼PhCH2O, R2¼R3¼H, R4¼Me) could be isolated from this reaction, although it was attempted many times in different solvents – namely ether, benzene, dioxane and anisole (Longmore 1966a). In only one case was a crystalline product (other than starting material) isolated for which, on the basis of spectroscopic analysis, structure 152 (R1¼PhCH2O, R2¼R3¼H, R4¼OH) was postulated (Longmore 1966a; Longmore and Robinson 1967). The formation of this product has been explained (Longmore 1966a, Longmore and Robinson 1967) by an aerial oxidation of 151 (R1¼PhCH2O, R2¼H) to give via the peroxide intermediate 153 (R¼OOH), the 3-hydroxy-3HR

PhCH2O N

O H

153 indole 153 (R¼OH), followed by nucleophilic attack of the side-chain hydroxy group on the 3H-indole C¼N group as shown, a sequence that finds an analogy (Longmore 1966a; Longmore and Robinson 1967) in the conversion of cinchonamine into quinamine using dilute peracetic acid (Witkop 1950). In an astute variation of this synthetic approach (Onaka 1971), 5-methoxy-3methylindole (154) was reacted with methylmagnesium iodide in ether followed by Me

MeO N H

154

3.2 Synthesis of the

129

an excess of ethylene oxide in ether at room temperature with stirring. After isolation, the basic reaction product was subjected to column chromatography on alumina to afford a 13% yield of 3,3a,8,8a-tetrahydro-5-methoxy-3a-methyl-2Hfuro[2,3-b]indole (152, R1¼MeO, R2¼R3¼H, R4¼Me). This furnished, upon methylation with iodomethane in the presence of sodium hydride in dimethylformamide, 152 (R1¼MeO, R2¼R4¼Me, R3¼H) which, by O-demethylation followed by O-Nmethyl carbamylation, had already been converted into dl-physovenine (Longmore 1966a; Longmore and Robinson 1965, 1967). This earlier synthesis (Longmore 1966a; Longmore and Robinson 1965, 1967) which had involved a longer reaction sequence and afforded much lower overall yields than that of Onaka (1971) had involved an extention to the 3,3a,8,8atetrahydro-3a-methyl-2H-furo[2,3-b]indole system of Julian and Pikl’s elegant and pioneering synthetic approach to the l-physostigmine ring system (Sect. 2.3.1) and which thereby increased even further the versatility (see footnote 9 of Chap. 2) of this latter synthetic pathway. Thus, 3-alkylation of 5-methoxy-1,3-dimethylindolin2-one (19, R1¼MeO, R2¼H) with a 5-molar excess of ethylene oxide in the presence of sodium ethoxide in dry ethanol afforded 3-(2-hydroxyethyl)-5-methoxy-1,3dimethylindolin-2-one [19, R1¼MeO, R2¼(CH2)2OH],2 along with unchanged starting material and its oxidation product 19 (R1¼MeO, R2¼OH) (Longmore 1966a; Longmore and Robinson 1965, 1967) [when only a 1.5 molar excess of ethylene oxide was used, the only products were 19 (R1¼MeO, R2¼OH) and 5,50 -dimethoxy-1,3,10 ,30 -tetramethylleucoisoindigo (155) (both formed by aerial Me

MeO N

OO

Me

Me

OMe

N Me

155 oxidation of the indolin-2-one anion), together with unchanged starting material (Longmore 1966a; Longmore and Robinson 1967). Reductive cyclisation of 19 [R1¼MeO, R2¼(CH2)2OH] with sodium in 95% ethanol boiling under reflux, followed by acidification and then basification gave 3,3a,8,8a-tetrahydro5-methoxy-3a,8-dimethyl-2H-furo[2,3-b]indole (3, R1¼R2¼Me, X¼O). This, upon O-demethylation by boiling under reflux its solution in petroleum ether (bp 80–100
C) in which anhydrous aluminium chloride was suspended (see footnote 8 of Chap. 2) gave 3 (R1¼H, R2¼Me, X¼O) which, upon reaction in dry ethereal solution, in the presence of a trace of sodium, with methylisocyanate was converted into dl-physovenine (3, R1¼MeNHCO, R2¼Me, X¼O) which had uv, ir and mass spectra and RF value [0.44 on a thin layer of alumina-calcium sulphate (20:1 w/w) using ether as the mobile phase and a solution of iodine in chloroform as developer] identical with the alkaloid, the structure of which, previously proposed from its spectral properties, had therefore acquired verification by synthesis (Longmore 1966a; Longmore and Robinson 1965, 1967).

130

3 l-Physovenine

The next two to three decades witnessed the appearance of two further syntheses of dl-physovenine (Luo et al. 1990; Yu et al. 1994) which, although of the same overall strategy as the earlier approach (Longmore 1966a; Longmore and Robinson 1965, 1967) (vide supra), involved several interesting innovations. Thus, in the first to appear (Luo et al. 1990) [for subsequent application of this synthetic strategy, see (Luo et al. 2007 and refs quoted therein)], 5-methoxy-1,3dimethylindolin-2-one (19, R1¼MeO, R2¼H) was C-alkylated with methyl bromoacetate to afford 19 (R1¼MeO, R2¼CH2COOMe) which upon reduction with lithium aluminium hydride (see footnote 20 of Chap. 2) in ether boiling under reflux gave 156 but in stirred tetrahydrofuran at O
C gave the desired product Me

MeO

OH

N Me

156 3 (R1¼R2¼Me, X¼O). This was O-demethylated with boron tribromide in dichloromethane and the resulting phenol 3 (R1¼H, R2¼Me, X¼O) was converted into dl-physovenine (3, R1¼MeNHCO, R2¼Me, X¼O) by reaction in dry ether with methylisocyanate in the presence of a speck of sodium. The other approach (Yu et al. 1994) began with an interesting and procedurally useful circumvention of the need to protect the hydroxyl function – see also (Pei et al. 1998) and (see footnote 20 of Chap. 2) – in which compound 17 (R1¼HO, R2¼Me), after liberation from “Metol”, was, without protection of the phenolic function, reacted, in tetrahydrofuran solution cooled in an ice-bath, with 2-bromopropanoyl bromide (16) added dropwise, with subsequent boiling of the reaction mixture under reflux, to yield the anilide 18 (R1¼HO) which underwent a Stollé cyclisation to afford a 97% yield of 5-hydroxy-1,3-dimethylindolin-2-one (19, R1¼HO, R2¼H). This, in the presence of p-toluenesulphonic acid monohydrate in tetrahydrofuran, was reacted with 3,4-dihydro-2H-pyran to afford 1,3-dimethyl-5tetrahydropyranyloxyindolin-2-one [19, R1¼tetrahydro-2-pyranylO (see footnote 5 and 45 of Chap. 2), R2¼H] which underwent, in tetrahydrofuran and 85% yield, tetrabutylamminium iodide-catalysed 3-alkylation with ethylene oxide to yield 19 [R1¼tetrahydro-2-pyranylO (see footnote 5 and 45 of Chap. 2), R2¼(CH2)2OH]. This, upon reductive cyclisation with lithium aluminium hydride followed by treatment with 2N-hydrochloric acid gave 3 (R1¼H, R2¼Me, X¼O) which was converted into dl-physovenine (3, R1¼MeNHCO, R2¼Me, X¼O) by reaction with methylisocyanate in dry ether in the presence of a trace of sodium. Further syntheses of the racemic alkaloid have been developed. One of the earliest (Shishido et al. 1986a), involving 14 steps, afforded an overall high yield of 40%, others have already been subjected to review (Takano and Ogasawara 1989) and other contributions have also appeared (Shishido and Fukumoto 1988; Shishido et al. 1990).

3.2 Synthesis of the

3.2.2

131

l- and d-Enantiomers

Contrary to a significantly later (Takano et al. 1991) claim, the first synthesis of lphysovenine to be reported (Longmore 1969; Longmore and Robinson 1966) – and apparently used later in another laboratory (Brossi 1990) – started with l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) which was converted, sequentially via l-eserethole L (98, R1¼Et, R2¼Me, X¼NMe, n¼1) and its methiodide (98, 1 2 R ¼Et, R ¼Me, X¼ NMe2 I, n¼1) into the methine 69 (R1¼Et, R2¼NMe2, X¼O). This, upon quaternisation with iodomethane, afforded the methiodide 69 Me

R1O

R2

N

X

Me

H OH

69 L

(R1¼Et, R2¼ NMe3 I, X¼O) which upon reaction in caustic soda boiling under reflux evolved trimethylamine and gave 5-ethoxy-3,3a,8,8a-tetrahydro-3a,8dimethyl-2H-furo[2,3-b]indole (98, R1¼Et, R2¼Me, X¼O, n¼1)3, a product whose structure was not elucidated when this reaction sequence was initially carried out (Polonovski and Polonovski 1918) but when the evolved trimethylamine was characterised as its picrate. The tricyclic product 98 (R1¼Et, R2¼Me, X¼O, n¼1) – the structure of which was verified by comparison (Longmore 1969; Longmore and Robinson 1966) of its uv (see footnote 33 of Chap. 2) and 1H–nmr spectra with those of l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1) and 3,3a,8,8a– tetrahydro-5-methoxy-3a,8-dimethyl-2H-furo[2,3-b]indole (3, R1¼R2¼Me, X¼O) (Sect. 2.1)– was O-deethylated by boiling under reflux its solution in petroleum ether (bp 80–100
C) in which anhydrous aluminium chloride was suspended (see footnote 8 of Chap. 2) and the resulting phenol 98 (R1¼H, R2¼Me, X¼O, n¼1), without either purification or characterisation, was converted into l-physovenine Me

R1O

N

H

R2

98

X

n

132

3 l-Physovenine

(98, R1¼MeNHCO, R2¼Me, X¼O, n¼1) by reaction of its solution in a mixture of dry benzene and dry tetrahydrofuran, containing a trace of sodium with an excess of methylisocyanate. Since the synthesis of l-eserethole (98, R1¼Et, R2¼Me, X¼NMe, n¼1) had already been achieved (Harley-Mason and Jackson 1954b; Hoshino and Kobayashi 1935, 1936; Jackson 1954; Julian and Pikl 1935b, c; Kobayashi 1938, 1939) (Sects. 2.3.1 and 2.3.2), this work (Longmore 1969; Longmore and Robinson 1966) thereby formally represents a total synthesis of l-physovenine, a similar claim also being made later (Node et al. 1991) concomitantly with a synthesis of leserethole (98, R1¼Et, R2¼Me, X¼NMe, n¼1). Furthermore, since optical inversion at C-3a could not have occurred during the above sequence of reactions, it also showed that l-physovenine has the same absolute configuration at the B/C ring junction – namely as shown in 98 (R1¼MeNHCO, R2¼Me, X¼O, n¼1) – relative to that of l-physostigmine, thereby confirming conclusions reached from ord spectral measurements (Sect. 2.4). The synthesis also represented a very useful approach to the 3a-alkyl-3,3a,8,8a-tetrahydro-2H-furo-[2,3-b]indole system from the readily-available 3a-alkyl-1,2,3,3a,8,8a-hexahydropyrrolo[2,3-b]indole system (Sect. 2.3.2) since one of the then routes to the former system proceeds in only relatively low overall yield (Longmore 1966a; Longmore and Robinson 1965, 1967) (Sect. 2.1) and the other (Hoshino 1934; Hoshino and Shimodaira, 1935; Nakazaki 1959) had failed in an attempt (Longmore 1966a; Longmore and Robinson 1967) to apply it to the synthesis of 152 (R1¼PhCH2O, R2¼R3¼H, R4¼Me) (Sect. 2.1). With financial support from the Medical Research Council and starting with deserethole – obtained by the Julian and Pikl route (Sect. 2.3.1) – in place of leserethole, a reaction sequence analogous to that described above afforded, as the ultimate product, d-physovenine (Dale 1969; Dale and Robinson 1970). It has also afforded d- and l-physovenines when starting from the corresponding enantiomeric esermetholes (Yu et al. 1991), Other related enantiospecific syntheses have been developed. One of these involved the chemical resolution of 19 (R1¼MeO, R2¼CH2COOH) with brucine in water and the resultant acid of (3S)-absolute configuration was then, using conventional methods, converted into 98 (R1¼R2¼Me, X¼O, n¼1) (Yu et al. 1993). Others (ElAzab et al. 2000; Sunazuka et al. 2005; Takano et al. 1991) have also elaborated enantiocontrolled total syntheses. However, probably the most significant development in this area has been the resolution of dl-physovenine (3, R1¼MeNHCO, R2¼Me, X¼O) by chromatography (see footnote 23 of Chap. 2) on a column of cellulose triacetate (Yu et al. 1991).

3.2.3

3a-alkyl-3,3a,8,8a-tetrahydro-2H-furo[2,3-b]indole Ring System

A treatise (Robinson 1982) on the reaction included specific references to the use of the Fischer indole synthesis – which has also been used in the synthesis of

3.2 Synthesis of the

133

other indole alkaloids [including l-physostigmine and its component ring system (Sect. 2.3.2)] and natural products (see footnote 32 of Chap. 2) – in synthetic approaches that have been made to this ring system. Indolisation of 157 (R¼H and MeO) in alcoholic hydrogen chloride boiling under reflux gave 158 (R1¼R2¼H, R3¼CH2OPh) (King and Robinson 1933) and 158 (CH2)2OPh

Me CH

R

CH N N H

157

R1

Me R3 N

R2

158

(R1¼MeO, R2¼H, R3¼CH2OPh) (King et al. 1933b), respectively. However, attempts to dephenylate the methochloride of the latter product with hydrobromic acid under a variety of conditions were unsuccessful (King et al. 1933b), in view of which this approach to the ring system of physovenine awaits further investigation (Robinson 1982). However, the ring system was obtained directly when arylhydrazines were reacted with 2-acyl-2-alkylbutyrolactones (159, R2¼alkyl) in a mixture of isopropanol and dilute hydrochloric acid boiling under reflux and, via the in situ formation of 160 (R5¼COOH), 160 (R5¼H) and 161, yielded 162, with specific examples starting from 159 (R1¼R2¼Me, R1¼Me R2¼CH2Ph and R1¼H, R2¼Me) (Robinson 1982). Analogous reactions occurred when the lactones 159 were replaced by 163 which react with arylhydrazine hydrochlorides in dimethylformamide boiling under reflux to directly afford 162, with specific examples starting from 163 [R1¼R2¼Me, R1¼Me, R2¼nPr and CH2Ph, R1¼nPr and CH2Ph, R2¼Me and R1+R2¼(CH2)4] (Robinson 1982). These synthetic approaches to the physovenine ring system have also been included in reviews by others [Grandberg 1974; Spande 1979(g)].

134

3 l-Physovenine

O R3

R2 O

R2

R5

OH

160 (R5=H)

C

O

N N

R1

R1

R4

159

160 (R5=COOH)

R2

R2

R3

R3 N R4

R1

O

N

O H R1

R4

162

161

(CH2)2OH

R2

CH C O

R1

163 The physovenine tricyclic ring system was also obtained as the synthetic intermediate 68 (X¼Y¼O) – en route to that of physostigmine – when arylhydrazines 64 were subjected to Fischer indolisation followed by treatment of the immediate organic product with warm methanolic potassium hydroxide (Rosenmund and Sotiriou 1975). A mixture of phenylhydrazine hydrochloride and 3-methyllaevulinic acid when subjected to Fischer indolisation in methanolic sulphuric acid boiling under reflux afforded a mixture of 158 (R1¼H, R2¼Me, R3¼COOH) /(18% yield) together with its methyl ester 158 (R1¼H, R2¼Me, R3¼COOMe) (52% yield), only the former product (62% yield) being obtained when the indolisation was effected in aqueous sulphuric acid boiling under reflux (Rosenstock 1966). Unfortunately, no attempt was made in this study to effect a further cyclisation to form a tetrahydrofuran ring (Robinson 1982) and thereby a physovenine ring system. Another route to the physovenine ring system (or at least to the 2,3-dihydrofuro [2,3-b]indole ring system) involved the formation of 164 “quantitatively” from

N Me

164

O

Notes

135

3-(2-bromoethyl)-1-methylindolin-2-one either “by heating in benzene solution with tetrahydroisoquinoline” or either “by reaction with tetrahydroisoquinoline alone in the cold” or “by treating in alcoholic solution with sodium ethoxide” (Julian et al. 1948a). However, later approaches (Kondo et al. 1950, 1952) to the synthesis of “3,8-dihydro-2H-furo[2,3-b]indole” were unsuccessful. One of the products resulting from the reaction between N-methylaniline and glycolaldehyde has been found to have structure 150 (R1¼Me, R2¼OH, R3¼H) (Turner 1965). This 3,3a,8,8a–tetrahydro-2H-furo[2,3-b]indole is of particular interest since it does not contain a 3a-alkyl substituent (Robinson 1968 (see footnote 15 of Chap. 2).4

3.3

Biogenesis

A scheme, not surprisingly similar to that for l-physostigmine, has been postulated for the biogenesis of l-physovenine (Robinson 1964a) (Sect. 2.5).

Notes 1. These data and their interpretation relating to the 1H-nmr spectrum of l-physostigmine are in accord with those published elsewhere (Longmore 1966a, Robinson 1965c) (see also Sects. 3.1, 4.1, 5.1 and 6.1) and by others (Cohen et al. 1960, Cordell 1981, Grasselli and Ritchey 1975 Jackson and Smith 1964, Muhtadi and El-Hawary 1989, Murao and Hayashi 1986, Newkome and Bhacca 1969, Varian Associates 1962, Woodward et al. 1960) [see also (Grandberg et al. 1970)]. 2. Contrary to this observation, the reaction of indolin-2-one (265, R1¼R2¼H) in ethanolic sodium ethoxide solution with an excess of freshly distilled ethylene oxide failed to yield any

H R1 N

O

R2

265 3-(2-hydroxyethyl)indolin-2-one [265, R1¼(CH2)2OH, R2¼H], the only product, resulting from N- alkylation, being 265 [R1¼H, R2¼(CH2)2OH], the structure of which was verified by heating it with hydrobromic acid to afford 265 [R1¼H, R2¼(CH2)2 Br], hydrogenolysis of which over 5% palladium-charcoal gave the known 1-ethylindolin-2-one (265, R1¼H, R2¼Et) (Wenkert and Blossey 1962). This result therefore failed to substantiate an earlier preliminary report (Miller and Gamson 1955) which claimed – unfortunately without experimental data – to have prepared 3-(2-hydroxyethyl)indolin-2-one [265, R1¼(CH2)2OH, R2¼H] from indolin -2one (265, R1¼R2¼H) and ethylene oxide. 3. It has been suggested (Longmore 1969, Longmore and Robinson 1966) that the product 98 (R1¼Et, R2¼Me, X¼O, n¼1) [the formation of which would appear to be analogous to the transformation – by “ Hofmann degradation” – of isoechitinolide methine methiodide (266) into compound 267 (Birch et al. 1961, Govindachari and Rajappa 1961, Robinson 1963a)] is formed by the elimination of trimethylamine with concomitant cyclisation as shown in 69 (R1¼Et,

136

3 l-Physovenine

I NMe3

CH2OH Me Me

CO O

N OH H

266 CH2OH Me Me

CO N H

O

O

267 L

R2¼ NMe3 IƟ, X¼O). However, this mechanism appears at first sight to be at variance with the later observations that changing the reacting nucleophile in analogous reactions – which thereby increased the synthetic versatility and utility of what had already been recognised (Longmore 1969, Longmore and Robinson 1966) as representing “ a very useful synthetic approach to the 3a-alkyl-3,3a,8,8a-tetrahydro-2H-furo[2,3b]indole system from the readily available 3a-alkyl1,2,3,3a,8,8a-hexahydropyrrolo[2,3b]indole system” – from ƟOH to ƟSH (using sodium mercaptide) (Brossi et al. 1996, He et al. 1992, Pei et al. 1994), MeNH2 (Brossi et al. 1996, Pei et al. 1994), PhCH2NH2 (Brossi et al. 1996, Pei et al. 1994, 1995a, Yu et al. 1999) and Ph (CH2)2 NH2 (Brossi et al. 1996, Pei et al. 1995b) led to the formation of the ring-systems of 98 [R1¼R2¼Me, X¼S, NMe, NCH2Ph and N(CH2)2Ph, respectively, n¼1]. However, since “The ease of replacement by other functional groups is. . . . characteristic of the hydroxyl group in 2– indolinols” [Spande 1979(e)] and “these reactions very likely proceed by an elimination – addition mechanism” [Spande 1979(e)], it isLlikely that these later above reactions proceed via elimination of ƟOH from 69 (R1¼Me, R2¼ NMe3 IƟ, X¼O) to afford 268 which then reacts

Me

MeO

NMe3

I

N Me

268 L

with ƟSH, etc to afford 69 (R1¼Me, R2¼ NMe3 IƟ, X¼S, etc). Then, by analogy with the earlier proposed mechanism (Longmore 1969; Longmore and Robinson 1966), this reacts further with the nucleophile ƟSH, etc, to eliminate trimethylamine and yield 98 (R1¼R2¼Me, X¼S, etc, n¼1). It would perhaps be of interest to investigate whether a compound such as 98 (R1¼R2¼Me, X¼O, n¼1) might react with either sodium mercaptide or methylamine. Other synthetic approaches have appeared to the ring systems 269 in which X¼NMe, Y¼S (An-naka et al. 1994, Yu et al. 1994), X¼S, Y¼NMe (An-naka et al. 1994), X¼O, Y¼NR (Luo

Me X

269

Y

Notes

137

et al. 2007), X¼NMe Y¼O (Luo et al. 2007, Yu et al. 1994), X¼Y¼O (Castellino and Rapoport 1986, Luo et al 2005a, b, 2006, 2007) [“an important structural component of aflatoxins – mould metabolites studied extensively by Buchi, Rapoport, Townsend, and others” (Luo et al 2005a, b) and of dihydrosterigmatocystin (148) investigated by Roberts and his research group (Knight et al. 1965, 1966)] and X¼CH2, Y¼NH (Chen et al. 1992). 4. Thus, for example, there is no evidence from the 1H-nmr spectra of tryptamine and its derivatives for the existence of the corresponding tricyclic tautomers (Cohen et al. 1960, Hino and Nakagawa 1988, Jackson and Smith 1964), although 239 (R¼Me) was one of two early proposals (Hino and Nakagawa 1988, Hodson and Smith 1956) as a tentative structure for folicanthine but it was soon abandoned (Hino and Nakagawa 1988, Hodson and Smith 1957) – ultimately in favour of 122 (R1¼R2¼Me) (see footnote 15 of Chap. 2). Indeed it might have appeared that a structure such as 239 (R¼Me), with only a tertiary C(3a) –atom, may be incapable of an independent existence since an approach to the synthesis of 239 (R¼H) following the method of Julian and Pikl (Sect. 2.3.1) and involving the potential reductive cyclisation of 240

H N N Me R

239 H

NH2

N O Me

240 using sodium in absolute ethanol boiling under reflux led only to the isolation of 1methyltryptamine (241) (Hino and Nakagawa 1988, Sugasawa and Murayama 1958a) which would appear to have been formed as a consequence of the presence of the active hydrogen at the

NH2 N Me

241 3-position of 240 (Sugasawa and Murayama 1958a × see also Hino 1961bb, Hino and Yamada 1963). Nevertheless, it has been postulated (Grandberg 1974, 1983; Robinson 1982) that a tricyclic intermediate 77 (R4¼H, n¼1) is involved in the formation of tryptamines by reacting arylhydrazines (70) with 4-chloro-, bromo- iodo-, or tosylbutyraldehyde (71, R3¼R4¼H, R5¼C1, Br, I or tosyl, respectively, n¼1) in boiling aqueous alcoholic solution under reflux, and compound 242 has been suggested (Baldwin and Tzodikov 1977) as the hypothetical intermediate in the enzymic prenylation of tryptophan ultimately at the 4-position.

138

3 l-Physovenine

H N N COO H H H H

242 However, a 3a-unsubstituted 1,2,3,3a,8,8a-hexahydropyrrolo[2,3-b]indole system was first synthesised (Hino and Nakagawa 1988, Ohno et al. 1968, 1970) starting from the tryptophan derivative 243 (R1¼COOEt, R2¼COMe). This, upon reaction with either t-butyl hypochlorite or N-bromosuccinmide afforded the corresponding intermediate 3-halogeno – 3H-indoles 244

R1 NH R2

N H

H

243 R

COOEt H NH COMe

N

244 (R¼C1 or Br, respectively) that immediately cyclised to give 245 (R1¼COOEt, R2¼COMe, R3¼Cl or Br, respectively, R4¼H). These underwent either spontaneous or base-catalysed dehydrohalogenation to yield 246 (R¼COOEt). Starting from 243 (R1¼COOEt, CONH2

R3 R1 N N H H R4 R2

245

R N H COMe

N H

246 or H, R2¼COMe) likewise afforded 246 (R¼COOEt, CONH2, or H, respectively) but only the product 246 (R¼COOEt) was subjected to slow (over 3 days) hydrogenation over a rhodiumon-alumina catalyst in ethyl acetate to afford the product 247 (Ohno et al. 1968, 1970).

Notes

139

H COOEt N N H H H COMe

247 It is well-established that in strong acid the tricyclic ring system in l-physostigmineL undergoes opening of the tetrahydropyrrolic ring C as shown in 125 to form 126 (R1¼Me, R2¼ NH2Me) (Sect. 2.6) and that the indole ring undergoes protonation at the 3-position (Hino and Nakagawa 1988). However, in acidic media, tryptamines 243 (R1¼H) are initially protonated on the N-atom of the aminoethyl moiety when, in 243, R2¼H or alkyl, and as the pH of the media is further decreased the diprotonated species 248 (R1¼H) and not the cyclic tautomer 245

R1

H

H NH2 R2

N H

248 (R1-R4¼H) is formed (Hino and Nakagawa 1988). “In order to obtain cyclic tautomers of tryptamines it is necessary to reduce the basicity of Nb below that of the indole ring and to retain the nucleophilicity to attack at the 2-position of the indolenine [248]” (Hino and Nakagawa 1988) and “The methoxycarbonyl group was found to be desired substituent for [R2 in 243]” (Hino and Nakagawa 1988). Thus, a 3a-unsubstituted 1,2,3,3a,8,8a-hexahydropyrrolo[2,3-b]indole was isolated when 243 (R1¼R2¼COOMe) in 85% phosphoric at room temperature underwent tautomerism to afford a high yield of 245 (R1¼R2¼COOMe, R3¼R4¼H) along with a trace of 249 (R¼COOMe)

H COOMe N N H H R

H

249 (Bourne et al. 1991, Taniguchi and Hino 1981), with similar changes also being reported (Sammes and Weedon 1979ab, Taniguchi and Hino 1981) when using trifluroacetic acid and 243 (R1¼COOMe, R2¼COMe). These results and similar observations [including that in which “Nb–Methoxycarbonyltryptamine” (243, R1¼H, R2¼COOMe) “in 85% phosphoric acid cyclized to” 245 (R1¼R3¼R4¼H, R2¼COOMe) – no ring junction stereochemistry was indicated “which was detected by NMR but could not be isolated because of its instability” although “ After acetylation” 245 (R1¼R3¼H, R2¼COOMe, R4¼COMe) “ was isolated in 70% yield”], together with the reactions of the 3a-unsubstituted tricyclic ring system, have been comprehensively reviewed (Hino and Nakagawa 1988), it being later reported (Bruncko et al. 1994) that sulphonylation of the above mixture of 245 (R1¼R2¼COOMe, R3¼R4¼H) and 249 (R¼COOMe) led to the isolation of “the stable crystalline sulphonamide derivative” 245 (R1¼R2¼COOMe, R3¼H, R4¼SO2Ph) “in overall yields typically of around 80-85% as a diastereomerically and enantiomerically pure substance.”

Chapter 4

l-Eseramine

4.1

Isolation and Structure Elucidation

An empirical formula C16H25N4O3 and mp 238–240
C were published for this alkaloid in the report of its first isolation from Calabar beans (Ehrenberg 1893). Nearly 20 years were to elapse before it was next isolated (Salway 1911) but, at this juncture, apart from a modification of the mp to 245
C (dec.) it was not subjected to any further investigation. It was next detected during a two dimensional paper chromatographic examination of an alkaloidal extract from Calabar beans which reported (Sinha 1959) that they contained 0.0088% w/w of eseramine. A third isolation was effected (Robinson 1964a – see also Robinson and Spiteller 1965) – along with those of l-physovenine (Chap. 3) and the lower melting isomorph of l-physostigmine (Sect. 2.1) – using column chromatography, following which the mp was determined as either 216–218
C (Kofler block) or 240–242
C (capillary tube) – an interesting example of variation in mp depending upon the
conditions of its determination – and the alkaloid was found to have ½α23 D – 289 (EtOH) (Robinson and Spiteller 1965). Then, from the following evidence, which has already been reviewed (Robinson 1968), structure 3 (R1¼MeNHCO, R2¼Me, X¼NCONHMe) was proposed for it (Robinson and Spiteller 1965). The previously assigned (Ehrenberg 1893) empirical formula of C16H25N4O3, which is theoretically impossible (Robinson 1988a, b), was corrected to the molecular formula C16H22N4O3 by quantitative elemental analysis and a molecular weight determination by mass spectrometry. The alkaloid was found to contain a MeNHCOO group substituted on the 5-position of an indoline nucleus because the strong absorption bands at 2369 and 1738 cm1 (N–H and C¼O stretching, respectively) in its ir spectrum (in CHCl3), its uv spectrum (in EtOH) and the aromatic proton region of its 1H-nmr spectrum (in DMSO – in these early times deuterated solvents were often not readily available) were similar to the corresponding spectroscopic properties of l-physostigmine and l-physovenine (Sect. 3.1). The presence of the second –NH-CO- group in l-eseramine © Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_4

141

142

4 l-Eseramine

was indicated by additional C¼O and N – H stretching bands in its ir spectrum at 1646(strong) cm1 and at 3283(medium) and 3413(medium) cm1, respectively, the two N-H stretching bands being attributed to free and bonded N-H stretching absorptions in solution, since in a Nujol mull this group gives only one N–H stretching band – at 3286(strong) cm1. The 1H-nmr spectrum of l-eseramine also showed the presence of the C(3a)-Me group, which gave rise to a 3-proton singlet at 8.66τ – similar to the protons of the corresponding methyl group in the spectra of l-physostigmine and l-physovenine (Sect. 3.1), and also had a 1-proton singlet at 4.70τ that was assigned to the C(8a)proton, the downfield shift of 1.03τ in this proton’s signal relative to that of the corresponding proton in l-physostigmine being attributed to the deshielding effect of the neighbouring second –NH-CO- group in l-eseramine. Final support for structure 3 (R1¼MeNHCO, R2¼Me, X¼NCONHMe) for leseramine was available from the alkaloid’s mass spectrum which indicated the very easy loss of one and two 57 mass units (caused by the loss of one and then both of the MeNHCO groups – as MeN¼C¼O) from the molecular ion at m/e 318 to give peaks at m/e 261 and 204 (base peak), respectively, and further key fragments were also obtained with m/e 174, 161 and 160 which confirmed the presence of the ions 135 (R1¼R2¼Me), 136 (R¼Me) and 135 (R1¼H, R2¼Me), respectively, (Sect. 2.7).

4.2 4.2.1

Synthesis of the Racemate and the l-enantiomer Racemate

The above structural proposal for l-eseramine was verified by the following synthesis – which is based upon an adaptation of the Julian and Pikl approach (Sect. 2.3.1) to the tricyclic ring system – of its racemate (Robinson 1965a, b) starting from “Metol” (4hydroxy-N-methylaniline hemisulphate). This was sequentially N-acetylated with acetic anhydride, O-benzylated with benzyl chloride in the presence of dry ethanolic sodium ethoxide boiling under reflux and subjected to alkaline hydrolytic cleavage of the N-acetyl group to afford 4-benzyloxy-N-methylaniline (17, R1¼PhCH2O, R2¼Me). This was subjected to the Stollé oxindole synthesis by reaction with 2-bromopropanoyl bromide (16) to afford the corresponding anilide 18 (R1¼PhCH2O) which, upon pyrolysis with anhydrous aluminium chloride [which also effected O-debenzylation (see footnote 16 in Chap. 2)] gave 5-hydroxy-1,3-dimethylindolin-2-one (19, R1¼HO, R2¼H) (see footnote 16 in Chap. 2) [obtained in only 16% yield, although doubtlessly this could now be improved upon by using an aluminium chloride:sodium chloride mixture (5:1 w/w) in place of aluminium chloride to effect the cyclisation step (see footnote 16 in Chap. 2)]. O-Methylation of this product with dimethyl sulphate gave 5methoxy-1,3-dimethylindolin-2-one (19, R1¼MeO, R2¼H) which was then subjected to 3-alkylation with chlorocyanomethane in the presence of sodium ethoxide in dry ethanol

4.2 Synthesis of the Racemate and the l-enantiomer

143

to yield 3-cyanomethyl-5-methoxy-1,3-dimethylindolin-2-one (19, R1¼MeO, R2¼CH2CN). This was reductively cyclised using sodium in dry ethanol boiling under reflux, followed by acidification and then by basification to afford, after purification via its picrate, dark red prisms, mp 163–164
C [lit (King et al. 1934) mp 162–163
C] [obtained in only 5.5% yield, although doubtlessly this could now be improved upon by effecting the reductive cyclisation with either lithium aluminium hydride or Vitride (see footnote 17 of Chap. 2)], dl-N(1)-noresermethole (23, R1¼MeO, R2¼R3¼H, R4¼Me). This, upon O-demethylation with hydrobromic acid boiling under reflux gave dl-N(1) – noreseroline (23, R1¼HO, R2¼R3¼H, R4¼Me) which, upon treatment, in dry benzene and dry tetrahydrofuran solution, with a minute quantity of sodium followed by an excess of methylisocyanate, was set aside in a stoppered flask under nitrogen, with occasional shaking for 4 days, during which time a crystalline deposit (white needles) of dleseramine (3, R1¼MeNHCO, R2¼Me, X¼NCONHMe) gradually formed (Robinson 1965a, b).

4.2.2

l-enantiomer

Prepared by total synthesis, l-N(1)-noresermethole (98, R1¼R2¼Me, X¼NH, n¼1) was reacted with benzyl bromide in cyanomethane in the presence of potassium carbonate to afford 98 (R1¼R2¼Me, X¼NCH2Ph, n¼1) which was sequentially O-demethylated with boron tribromide in dichloromethane at room temperature to give the phenol 98 (R1¼H, R2¼Me, X¼NCH2Ph, n¼1), reacted with methylisocyanate in dry ether in the presence of a trace of sodium at room temperature to afford 98 (R1¼MeNHCO, R2¼Me, X¼NCH2Ph, n¼1) and N-debenzylated by hydrogenolysis in the presence of palladium on carbon to afford 98 (R1¼MeNHCO, R2¼Me, X¼NH, n¼1). This, upon reaction with methylisocyanate in a shaken ethereal solution, deposited a precipitate of l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1) (Yu et al. 1988a. It has been noted that “This synthesis of [98 (R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1)] is much superior to that used to prepare [l-98 (R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1)] by a direct N-methylcarbamoylation of (-)-N1-noreserolin ----- from which (-)-eseramine could only be isolated in 15% yield after tedious chromatography” (Yu et al. 1988a). These observations are contrary to those of the ultimate step in the earlier synthesis (Robinson 1965a, b) of the corresponding racemate as described above. However, since the intermediate 98 (R1¼MeNHCO, R2¼Me, X¼NH, n¼1), was, by reductive N-methylation, also converted into l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) (Yu et al. 1988a) it follows that the two ultimate alkaloidal products from this overall reaction sequence have identical absolute stereochemistry at their B/C ring junction and that the absolute configuration of l-eseramine is therefore as shown in 98 (R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1), thereby confirming conclusions reached earlier from ord spectral measurements (Sect. 2.4).

Chapter 5

l-N(8)-Norphysostigmine

5.1

Isolation and Structure Elucidation

This alkaloid, mp variously reported as 151
C (Longmore 1966b; Robinson 1968), 151.5
C (Murao and Hayashi 1986) and 152–153
C (Takano et al.
1990a) [and erroneously quoted as 157–159
C (Neuss 1966)] and ½α21 D 108.6 (EtOH) (Longmore 1966b; Robinson 1968) and (95% EtOH) (Neuss 1966),
½α28:5 116
(c 0.40, EtOH) (Takano et al. 1990a) and ½α30 D D 64.4 (c 0.45, MeOH) (Murao and Hayashi 1986), was initially isolated, from Calabar beans, in 1958 by Dr. J Maier of CH Boehringer Sohn at Ingelheim am Rhein (Boehringer Sohn 1964; Spiteller and Spiteller-Friedmann 1964) and was later (Murao and Hayashi 1986) obtained from a microbial source, namely Streptomyces sp. AH-4. Its structure was postulated after examination of its mass spectrum (Spiteller and Spiteller-Friedmann 1964), the salient features of which were a molecular ion peak at m/e 261, a base peak at m/e 204 (indicating the very easy loss of the elements of methylisocyanate from the molecular ion), and other significant peaks at m/e 189, 160, 147 and 146 [indicating the presence of the ions 131 (R¼H), 135 (R1¼Me, R2¼H), 136 (R¼H), and 135 (R1¼R2¼H), respecHO N R

H

N Me

131

© Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_5

145

146

5 l-N(8)-Norphysostigmine

R1 HO N

135

(R1=

R2 H, R2 = Me)

H CH2

HO N R

136 tively]. This fragmentation pattern was very similar to that of l-physostigmine (Sect. 2.7) but with all six peaks 14 mass units lower, which clearly indicated the secondary nature of N(8) in the new alkaloid for which structure 98 Me

R1O N R2

H

X

n

98 (R1¼MeNHCO, R2¼H, X¼NMe, n¼1) was suggested (Spiteller and SpitellerFriedmann 1964) although, of course, the position of the MeNHCOO substituent still awaited confirmation. Further evidence for the presence of this group in the alkaloid was forthcoming (Longmore 1966b; Robinson 1968) from its ir spectrum (in Nujol) which showed, along with a band at 1710  10 cm1 (C¼O stretching) (Longmore 1966b), two bands in the N-H stretching region, one at 3331 cm1 (due to the N(8)-H group) and the other at 3200 cm1 (due to the N-H of the methylcarbamyl group) [cf the corresponding band in the ir spectrum of l-physostigmine (in Nujol) at 3198 cm1]. In chloroform, however, the two N-H stretching bands of l-N(8)-norphysostigmine are nearly superimposed and give rise to a single absorption band with a shoulder (Neuss 1966). The uv spectra of l-N(8)-norphysostigmine and of l-physostigmine (Sect. 2.6) in ethanol are almost superimposable from which it can be deduced (Chalmers et al. 1958) that the former, like the latter alkaloid has the MeNHCOO group substituted at the 5-position of an indoline nucleus (Longmore 1966b; Robinson 1968), further evidence for which is forthcoming from H1-nmr spectroscopic data (vide infra). Moreover (Longmore 1966b; Robinson 1968), both uv spectra in dilute acid undergo a hypsochromic shift of about 10 nm, with retention of indoline absorption, relative

5.2 Synthesis of the l-enantiomer

147

to their spectra in ethanol owing to protonation of N(1) in their Ph-N-C-N systems (Sect. 2.6). However, whereas the spectrum of l-physostigmine is completely changed to 3H–indolium cation absorption in 3 N–hydrochloric acid, the weakest acid that will similarly completely change the spectrum of l-N(8)-norphysostigmine is 8 N–hydrochloric acid (Robinson 1965d), these observations supporting the suggestion (Jackson and Smith 1964) that such acid-catalysed ring openings of Ph-N(a)-C-N and Ph-N(a)-C-O systems (Sect. 2.6) (see footnote 34 in Chap. 2) are facilitated by the presence of a methyl substituent on N(a). The 1H–nmr spectra of the two alkaloids are similar except that the spectrum of lN(8)-norphysostigmine shows no signal corresponding to the N(8)-CH3 group protons present in the spectrum of l-physostigmine (Chap. 7) (see footnote 1 in Chap. 3) but does have 1-proton doublets centred at 5.52τ and 5.75τ (J ¼ 6 cps) due to the mutually coupled N(8)-H and C(8a)-H protons, respectively, (Longmore 1966b; Robinson 1968) although this coupling was surprisingly apparently absent in a later report (Murao and Hayashi 1986) comparing these two 1H–nmr spectra in which the corresponding signals are quoted at “δ ¼ 3.40” and “4.31” respectively, and as “(1H, b.r.s)” and “(1H, s)”, respectively. However, this later study does note for both l-physostigmine and l-N(8)-norphysostigmine a significant coupling between the 6-H and 7-H protons which afford signals (all “1H, d, J ¼ 9 Hz”) at “δ ¼ 6.78 and 6.33 and at 6.72 and 6.48”, respectively, (Murao and Hayashi 1986) whereas the 4-H protons both give rise to 1H singlets at “δ ¼ 6.77 and 6.76”, respectively. These data are all in accord with the MeNHCOO group being substituted at the 5-position of the indoline nucleus in l-N(8)-norphysostigmine. The structure of the alkaloid was ultimately identified as 98 (R1¼MeNHCO, R2¼H, X¼NMe, n¼1) by its methylation with iodomethane in alkaline methanol to afford physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) (Murao and Hayashi 1986).

5.2

Synthesis of the l-enantiomer

A synthesis of this minor alkaloid of the Calabar bean has been effected (Takano et al. 1990a) from its major alkaloidal component by a route involving two remarkably-selective reactions. Thus, l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) was oxidised using pyridinium dichromate in dichloromethane to afford a 25% yield of 98 (R1¼MeNHCOO, R2¼CHO, X¼NMe, n¼1) which, upon hydrolysis with diluted hydrochloric acid (10%) at room temperature was converted into a 72% yield of l-N(8)-norphysostigmine (98, R1¼MeNHCO, R2¼H, X¼NMe, n¼1) (Takano et al. 1990a). Furthermore, since optical inversion at C(3a) could not have occurred during the above sequence of reactions, this also showed – as did its N-methylation with iodomethane in alkaline methanol to afford l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) (Murao and Hayashi 1986) (Sect. 5.1) – that l-N(8)norphysostigmine had the same absolute configuration at the B/C ring junction, namely as shown in 98 (R1¼MeNHCO, R2¼H, X¼NMe, n¼1) – relative to that of

148

5 l-N(8)-Norphysostigmine

l-physostigmine, thereby confirming the conclusion reached earlier from ord spectral measurements (Sect. 2.4). An approach to the synthesis of physostigmine by Wijnberg and Speckamp was published in 1978. It has been reported (Brossi 1990) that efforts to prepare “N(8)norphysostigmine” following this route “has not yet [in 1990] afforded the desired alkaloid” but that “This investigation is continuing despite an elegant total synthesis of (-)-N(8)-norphysostigmine [then] just published by Japanese scientists” as described above (Takano et al. 1990a). Using the method of Hoshino and Kobayashi (1935) (Sect. 2.3.1), 5-methoxy-Nmethyltryptamine (26, R1¼MeO, R2¼H, R3¼Me) had afforded a 45% yield of dl-N(8)noreseroline methyl ether (23, R1¼MeO, R2¼R4¼H, R3¼Me) (Brzostowska) and

H (CH2)2N

R1

R3 N H

R2

26 attempts to convert this into l-N(8)-norphysostigmine had “been initiated” (Brossi 1992). Furthermore, the sequential reaction of 5-methoxy-3-methylindole (154) in ether Me

MeO N H

154

with methylmagnesium iodide and N-methylaziridine (see Onaka 1971) might well lead to a facile synthesis of dl-N(8)-norphysostigmine and the successful application of a similar synthetic strategy to 9-unsubsittuted-6-alkoxy-1,2,3,4-tetrahydrocarbazoles would afford pharmacologically and potentially therapeutically interesting (Sects. 10.7.2 and 10.11) echibolines.

Chapter 6

l-Geneserine

6.1

Isolation and Structure Elucidation

This alleged (see footnote 4 in Chap. 2) alkaloid was extracted, along with lphysostigmine, with cold ether from pulverised Calabar beans which were treated with sodium bicarbonate, in increased yield when using hot sodium carbonate solution and boiling ether, and in maximum yield (0.1%) – and free from l-physostigmine – using ether and botanical material that had previously been either soaked in (or moistened with?) 2% aqueous sodium hydroxide solution (Polonovski and Nitzberg 1915a). It had mp 128–129
C, [α]D  175
(EtOH) and 188
(dilute H2SO4), an empirical formula C 15H21N3O3, and was only weakly basic – allegedly failing to give crystalline salts with mineral acids [Henry 1949; Merck Index 2001(b)] although its hydrochloride has been reported (Bacchi et al. 1994) and it has also afforded a picrate (Henry 1949; Nakagawa et al. 1975; Polonovski and Nitzberg 1915a; Yu et al. 1989), salicylate and methiodide (Henry 1949; Polonovski and Nitzberg 1915a), 2-hydroxybenzoate (Flippen-Anderson et al. 2002) and formate (Yu et al. 1989). Methods for its detection and quantitative determination have been reviewed (Robinson 1964b). Further investigations, which have been included in reviews (Coxworth 1965; Henry 1949; Marion 1952; Robinson 1964b; Sumpter and Miller 1954a), showed that its reactions paralleled in many ways those of l-physostigmine. For example, when heated to 160
C it yielded methylisocyanate, which was also formed by its oxidation with acidic potassium permanganate or chromium trioxide, thus indicating the presence of a N-methylcarbamyl group (Polonovski and Nitzberg 1915a), as it also was by the production of one mole of carbon dioxide and one of methylamine upon heating it with aqueous barium hydroxide (Polonovski and Nitzberg 1915a). When heated with sodium ethoxide in the absence of air it afforded ethyl N-methylcarbamate and geneseroline (C13H11N2O2) – analogous to l-eseroline and which, analogous to the latter, can be converted into l-geneserine and its phenyl analogue by treatment in anhydrous ether, in the © Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_6

149

150

6 l-Geneserine

presence of a trace of sodium, with methylisocyanate and phenylisocyanate, respectively (Polonovski 1916; Polonovski and Nitzberg 1916), and by reaction with bromoethane and sodium ethoxide into its ethyl ether, geneserethole, which was also obtained directly from l-geneserine by the action in ethanol of ethyl p-toluenesulphonate in the presence of sodium ethoxide (Polonovski and Nitzberg 1915a). Moreover, l-geneserine was reduced to l-physostigmine by treatment with zinc powder in either ethanolic acetic acid boiling under reflux for 1 h (Polonovski and Nitzberg 1915a) or in acetic acid stirred at room temperature for ½ h (Yu et al. 1989) – the former conditions also reduce geneseroline to eseroline (Polonovski and Nitzberg 1915a) – and by the action of sulphur dioxide (Polonovski and Nitzberg 1915b). These reductions were also reversed when l-physostigmine was subjected to oxidation either with hydrogen peroxide in either acetone (Polonovski 1917) or acetone and phosphate buffer (Bacchi et al. 1994) or, more recently, with 3-chloroperbenzoic acid in chloroform at room temperature (Nakagawa et al. 1975; Yu et al. 1989, 2002), with the first conditions similarly converting eserethole into geneserethole (Polonovski 1917). From the earlier of these above observations, and since the molecular formula of l-geneserine differs from that of l-physostigmine only by the addition of one oxygen atom, it was concluded (Polonovski 1917, 1918; Polonovski and Nitzberg 1915a; Polonovski and Polonovski 1918, 1925a, b) that l-geneserine was the N(1)-oxide of l-physostigmine and claimed (Polonovski and Polonovski 1924f) that it was the first alkaloid with an aminoxide function [although as such, and in the strictest sense, it is not an alkaloid (see footnote 4 in Chap. 2)] to be isolated from a plant. This structural postulation, namely 3 [R1¼MeNHCO, R2¼Me, X¼N(!O)Me], was accepted for some 40 years (Coxworth 1965; Henry 1949; Marion 1952; Robinson 1963a, 1964a, b; Sumpter and Miller 1954a), with even the observation (Bild and Hesse 1967) that the mass spectrum of l-geneserine does not show M⨁ – 16, M⨁  17 and M⨁  18 peaks characteristic of N-oxides failing to lead to any suggestion that this alkaloid may not be an N-oxide. Ultimately (Hootelé 1969), however, these mass spectral properties, together with the previously-observed (Polonovski 1930) nonhygroscopic character and low water solubility of l-geneserine – which were “in contrast with the polar properties of true N-oxides” (Hootelé 1969) – were realised as being inconsistent with its structure being an N-oxide and it was thereby modified to 165 by analysis of the following 1H–nmr data (Hootelé 1969). On the basis of structure 3 [R1¼MeNHCO, R2¼Me, X¼N(!O)Me] the chemical shift of the protons of the N(1)-methyl group was unexpectedly the same as that for the corresponding protons in the spectrum of l-physostigmine, whereas by analogy with the signal usually associated with the protons of the N(!O)Me group it should have been circa 0.75τ downfield relative to the corresponding signal in the spectrum of l-physostigmine. Furthermore, the N(1)-methyl signal would have been expected to have had a chemical shift roughly comparable with that of the N(1)methyl group in l-physostigmine methiodide although this was not the case, the latter signal being circa 1τ downfield relative to the N(1)-methyl signal of l-geneserine. Another observation inconsistent with structure 3 [R1¼MeNHCO, R2¼Me, X¼N(!O)Me], on both a theoretical basis and by analogy with the observation

6.1 Isolation and Structure Elucidation

151

that the signal caused by the C(2)-H2 protons in N-methylpyrrolidine is moved circa 1τ downfield in its N-oxide, is that the C(2)-H2 protons have approximately the same chemical shifts in both l-geneserine and l-physostigmine. Indeed, the only significant difference between the 1H–nmr spectra of the two alkaloids is the relative downfield shift of circa 0.65τ experienced by the C(9a)-proton in the former. All these above data are consistent with structure 165 (Hootelé 1969), in particular the downfield shift of the C(9a)-proton’s signal as compared with that of the C(8a)-proton in l-physostigmine – expected of the signal of a proton attached to a carbon atom flanked by an oxygen and a nitrogen atom – see, for example, as in l-physovenine (Sect. 3.1) – as opposed to one flanked by two nitrogen atoms. Further confirmation (Hootelé 1969) for the correctness of structure 165 was forthcoming from a comparison of the 1H–nmr spectrum of l-geneserine with that of N-methyltetrahydro-1,2oxazine (166) from which it could be seen that the chemical shifts of the NCH2 and NMe protons in 166 closely resemble those of the corresponding protons of this system in l-geneserine (165). Furthermore, the observation (Hootelé 1969) that the solvent effects on the chemical shifts of the tetrahydro-1,2-oxazine moiety’s NMe protons in l-geneserine parallel those of the NMe protons in 167 also supported (Hootelé 1969) structure 165 for l-geneserine. Me

MeNHCOO

N

H Me

O

N Me

165

O

N Me

166

O

N Me

N

167 An earlier suggestion for the existence of this alkaloid as tautomerides – distinguished by their relative abilities to decolourise methylene blue – has been presented (Polonovski 1939; Polonovski and Desgrez 1938), although this interesting observation, with the exception of only two (Henry 1949; Robinson 1964b) of the several pertinent reviews

152

6 l-Geneserine

(vide supra), remained unrecognised in the literature, including that in which further evidence for this tautomerism was reported when it was shown (Yu et al. 1989) that l-geneserine base (110, R¼MeNHCO) and salt (168, R¼MeNHCO) (vide infra) both showed “two spots of different polarity on Si gel plates” and that upon treatment with Me

RO

X N H OH Me Me N

168

Me

RO

4a 9a

N9

H Me

O

N

Me

110 acids – for example in the formation of salts – l-geneserine (110, R¼MeNHCO) and l-geneseroline (110, R¼H) are converted into and adopt the corresponding N-oxide 168 (R¼MeNHCO and H, respectively) structures which were supported from their 1H-nmr spectral properties (Yu et al. 1989) and confirmed by X-ray crystallographic analysis of geneseroline hydrochloride (168, R¼H, X¼C1(Yu et al. 1989) and of geneserine hydrochloride (168, R¼MeNHCO, X¼C1(Bacchi et al. 1994). It was suggested that 165 arises from the oxidation of l-physostigmine (vide supra) via the formation of the N-oxide 3 [R1¼MeNHCO, R2¼Me, X¼N(!O)Me] as a transient intermediate from which 165 is formed by nucleophilic attack of the N(1)-oxygen atom on C(8a) with consecutive enlargement of ring C (Hootelé 1969). Indeed, this rearrangement of the N-oxide into 165 has been recognised (Nakagawa et al. 1975; Shishido et al. 1986a, 1987; Yu et al. 1989) as being of a Meisenheimertype. The reduction of l-geneserine into l-physostigmine by zinc powder in ethanolic acetic acid (vide supra) was envisaged as proceeding via hydrogenolysis of the N-O bond in 165 (as observed for 167) followed by nucleophilic substitution of the OH group thus formed at C(2) of the indoline nucleus by the p-electron pair of the nitrogen atom of the aminoethyl side-chain (Hootelé 1969). It is, however, either possible (Nakagawa et al. 1975) or, indeed, more likely that, under the acidic conditions of the reduction, the initially produced 2-hydroxyindoline looses hydroxyl ion to afford the 3H–indolium cation and that l-physostigmine is then produced by the nucleophilic attack of the p-electron pair of the nitrogen atom of the aminoethyl side-chain at the C(2)-atom of this 3H–indolium cation nucleus, analogous to the ring closures leading to ring C in the early syntheses of l-physostigmine and its tricyclic ring system (Sect. 2.3.1 and (see footnote 26 in Chap. 2).

6.1 Isolation and Structure Elucidation

153

Moreover, since optical inversion at C(3a) and C(4a), respectively, cannot have occurred during the oxidative/reductive interconversion between l-physostigmine and l-geneserine (vide supra) this, together with the appropriate nOe studies (Robinson and Moorcroft 1970) also shows that l-geneserine has the same relative absolute configuration at the B/C ring junction – namely as shown in 110 (R¼MeNHCO) – as that of l-physostigmine, thereby confirming conclusions reached from ord spectral measurements (Sect. 2.4). The structural relationship existing between l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) and l-geneserine (110, R¼MeNHCO) has also been found (Christophersen 1985a, b; Carlé and Christophersen 1980; Keil et al. 1986) between two of the bases which have been isolated from the marine bryozoan Flustra foliacea, (see footnote 9 in Chap. 2) namely flustramine B (169) and flustrarine B (170). Oxidation of the former with Me

Me

N

Br

Me

H

N Me

Me

169

Me

Me

Br

N

O H

Me

N Me

Me

170 aqueous hydrogen peroxide in acetone afforded an 88.6% yield of the latter (Keil et al. 1986) and, moreover, the possibility that this product is the corresponding N-oxide is eliminated by the absence of M⨁ – 16, M ⨁  17 and M⨁  18 fragments in its mass spectrum (Keil et al. 1986) which does, however, indicate a fragmentation of the molecular ion to give rise to peaks corresponding to M⨁ – C2H5NO and M⨁C2 H6NO, “in close analogy to the findings reported [Bild and Hesse 1967] for geneserine” (Keil et al. 1986). In addition, the 1H–nmr

154

6 l-Geneserine

spectrum accounted for all the protons of the flustramine B (169) skeleton, thereby excluding the presence of a hydroxyl function (Keil et al. 1986). However, in view of the later observations (Yu et al. 1989) relating to l-geneserine (vide supra), it would appear possible that salts of flustrarine B would also contain the moiety corresponding to the protonated N(1)-oxide of 169.

6.2

Synthesis of the l-enantiomer and the Racemate

An essential component of the studies which led to the elucidation of the structure of l-geneserine involved the transformation of l-physostigmine into 98 [R1¼MeNHCO, R2¼Me, X¼N(!O)Me, n¼1] that underwent Meisenheimer-type rearrangement (Nakagawa et al. 1975; Shishido et al. 1986a, 1987; Yu et al. 1989) to afford 110 (R¼MeNHCO) (vide supra). This work, since the total synthesis of l-physostigmine has previously been effected (Sects. 2.3.1 and 2.3.2), thereby formally represents a total synthesis of l-geneserine, a similar claim also being made later (Node et al. 1991) concomitantly with a synthesis of l-eserethole (98, R1¼Et, R2¼Me, X¼NMe, n¼1). A total synthesis of the racemate of geneserine via an unambiguous route has been effected in studies (Shishido et al. 1986a, 1987; Wright et al. 1987) that, along with related investigations, have already been comprehensively reviewed (Takano and Ogasawara 1989).

Chapter 7 1

H-, 13C- and 15N–Nuclear Magnetic Resonance Spectra of the Alkaloids of the Calabar Bean

Comparisons of the 1H-nmr spectra of l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1), l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1), l-N(8)- norphysostigmine (98, R1¼MeNHCO, R2¼H, X¼NMe, n¼1) and l-geneserine [(98, R1¼MeNHCO, R2¼Me, X¼N(! O)Me, n¼1)  later (110, R¼MeNHCO)] with that of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) played a significant role in the elucidation of the structures of the first four alkaloids and details of these spectra of all five of these alkaloids have therefore already been referred to and discussed (Sects. 3.1, 4.1, 5.1 and 6.1, respectively). Partial 1H-nmr data of l-geneserine (110, R¼MeNHCO) have also been tabulated (Yu et al. 1989 – see also Hootelé 1969). Independent investigations (see footnote 1 in Chap. 3) have also been manifest by their recording and interpretation of the 1H-nmr spectra of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) and related studies on analogous compounds by Grandberg and his colleagues have been alluded to in review (Grandberg et al. 1970). As a prerequisite to a biosynthetic study on this group of alkaloids (Sects. 2.5 and 3.3) and as a useful aid in the structural investigation of further alkaloids which may be isolated from this source (Chap. 8), the 13C–nmr spectra of l-physostigmine, l-N(8)-norphysostigmine, l-eseramine and l-physovenine have been examined (Crooks et al. 1976) (Table 7.1). The carbon assignments, which were also later (Stenberg et al. 1977) ultimately shown for all 15C of l-physostigmine to be consistent with those reported in an even later related examination of this alkaloid in “DMSOd6”(Muhtadi and El-Hawary 1989), were fully established and were based upon proton noise decoupled, off-resonance decoupled and proton coupled spectra together with comparison with already-published data (Johnson and Jankowski 1972) and with the 13C–nmr spectrum of the model indoline 171 (R1¼H, R2-R4¼Me) (Ahmed 1966; Ahmed and Robinson 1965) (with numbering analogous to that in the l-physostigmine ring system 3). The correlations are generally very close but useful diagnostic shifts are observed (Crooks et al. 1976) in the series, particularly at C(2) where a downfield shift of 14.1 ppm for l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1) and an upfield shift of 7.5 ppm for © Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_7

155

67.3

–

–

–

45.7

52.5

37.0

c

2 53.2

1-Me 36.9

–

41.6

38.5

40.7

3 40.7

42.7

52.3

50.4

53.7

3a 52.6

12.0 22.7

24.6

26.9

26.9

3a-Me 27.2

140.0

135.2

135.1

137.8

3b 137.4

115.5

116.5

116.0

116.5

4 116.1

149.2

147.9

147.4

146.9

5 149.3

120.0

120.8

120.7

120.5

6 120.4

107.4

105.5

105.8

109.0

7 106.5

143.8

143.0

142.8

144.0

7a 143.3

34.1

31.2

33.6

-

8-Me 38.4

72.9d

104.7

89.2

90.3

8a 98.1

27.4

27.7

26.9

27.9

Meb NH 27.5

156.3

156.3

155.6

156.3

COb 156.3

7

From (Crooks et al. 1976) Notes. aWith the exception of l-eseramine that was examined in DMSO solution, all other spectra were recorded in CHCl3 (Crooks et al. 1976) b Compare MeNH-COOEt (Me-27.4; CO-157.8 ppm) (Johnson and Jankowski 1972) c CO-157.7, Me-23.3 d Me-25.5

Assignment (ppm) Compounda l-Physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) l-N(8)Norphysostigmine (98, R1¼MeNHCO, R2¼H, X¼NMe, n¼1) l-Eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1) l-Physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1 171 (R1¼H, R2–R4¼Me)

Table 7.1 The 13C-nmr spectral assignmentsa for alkaloids of the Calabar bean

156 1

H-, 13C- and 15N–Nuclear Magnetic Resonance Spectra of. . .

7

H-, 13C- and 15N–Nuclear Magnetic Resonance Spectra of. . .

1

R1O

4 5

Me 3b

A

6 7

157

7a

3a

3

B C N 8 8a X1 R2

2

3

MeNR1COO

R4

4 5

3b 3a

6

7a 7

N

8a 8

R3 R2

Me

171 l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1) are observed relative to the C(2) value in l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) and, similarly, C(8a) showed an upfield shift of 7.8 and 8.9 ppm in l-N(8)-norphysostigmine (98, R1¼MeNHCO, R2¼H, X¼NMe, n¼1) and l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1), respectively, and a downfield shift of 6.6 ppm in l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1). The N(8)-methyl group also showed an upfield shift of 7.2 and 4.8 ppm in l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1) and l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1), respectively (Crooks et al. 1976). These above data, corroborated – with the exception of the assignments “for four shift groupings” – by those later reported (Stenberg et al. 1977) for the 13C–nmr spectrum of l-physostigmine have been included in reviews (Shamma and Hindenlang 1979) [see also (Cordell 1981)] and also have already been used in the determination of the structure of dihydroflustramine (172), one “of a series of bromoalkaloids belonging to the physostigmine group” and present in “the bryozoan Flustra foliaceae (L) from Scandanavian waters” (Wright 1984) (see footnote 9 in Chap. 2), and of flustrarine B (170), another related metabolite (Keil et al. 1986). Me

Me

Br

N

O H

Me

Me

170

N Me

158

7

H-, 13C- and 15N–Nuclear Magnetic Resonance Spectra of. . .

1

CH2 Me

Me

Br

N

H

H

N Me

172 It is considered (Stenberg et al. 1977) that “spectroscopically, physostigmine’s N-nmr is notable because it contains nitrogens in three different environments”. Indeed, it is not surprising “that the proton decoupled 15N–nmr spectrum – adjusted to 15NH4Cl as an external reference set at 0.0 ppm and with positive values indicating downfield shifts (Stenberg et al. 1977) – of the alkaloid has three resonances”. These are a doublet centred at 44.8 ppm (attributed to the N-atom of the carbamyl group on the basis of its coupling to the attached proton) (Stenberg et al. 1977), a singlet at 44.1 ppm (attributed to N(8) because this is conjugated with the aromatic ring) (Stenberg et al. 1977) and a poorly resolved multiplet at 30.1 ppm (attributed to N(1) because of its long-range coupling with the C(2) and C(8a) protons) (Stenberg et al. 1977). Unfortunately, these 15N-nmr studies were not extended to other of the Calabar bean’s alkaloids, especially since from these by then already published initial 13C– nmr studies (Crooks et al. 1976) it should have been apparent that samples of these were available, albeit on the other side of the Atlantic Ocean.

15

Chapter 8

Other Alkaloids That Have Been Isolated, or Allegedly So, from the Calabar Bean

8.1

Calabarine

The presence of this, as a second alkaloid, in the Calabar bean, was indicated in 1876 although the substance was of a very indefinite character and was mainly distinguished by its tetanus-like effects – thereby in its physiological action resembling strychnine more nearly than physostigmine (Zhao et al. 2004)
on the living organism (Harnack and Witkowski 1876). Moreover, subsequent investigation (Ehrenberg 1893) led to the conclusion that calabarine was, in all probability, an artefact produced by decomposition of the bean’s alkaloids either during or after their extraction, an opinion that was later endorsed (Holmstedt 1972; Robinson 1964b; Salway 1911; Zhao et al. 2004).

8.2

Eseridine

This crystalline base, mp 132  C and thereby differing from l-physostigmine – by comparison with which it was also much less poisonous, was obtained from the Calabar bean by Boehringer and Söhne (1888) who also reported that “Dieses Alkaloid steht dem Physostigmin sehr nahe, es geht schon beim Erhitzen mit verdünnten Säuren in dasselbe über. Aus diesem Grunde ist darauf zu achten, dass beim Auflösen des Eseridins in verdünnten Säuren eine Erwärmung thunlichst vermieden wird”. It was apparently the subject of further study by Schweder (1889) and it was also subsequently stated (Salway 1911) that “Eseridine has been further examined by Eber (Pharm. Zeit. 1892,37,483), who assigned to it the formula C15H23O3N3, and it is thus seen to differ from l-physostigmine only in the elements of one molecule of water”, although neither this literature source nor its abstract (Eber 1888a) [the former of which was misquoted – its year of publication being 1888 – by Salway (1911) (vide supra) and likewise by Henry (1924, 1949) but later © Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_8

159

8 Other Alkaloids That Have Been Isolated. . .

160

correctly quoted by others (Coxworth 1965; Marion 1952)] made any reference to this formula. However, it was quoted (Eber 1888b) (as C15H23N3O3) with relation to “Eseridinum purum” and likewise in the first [Eber 1888c(a)] of a series of four communications [Eber 1888c(a)(b)(c)(d)] concerning the alkaloid’s pharmacology, but in neither instance [Eber 1888b, c(a)] without any reference to relevant supportive experimental analytical data. This omission is particularly unfortunate since they may have helped to differentiate between the molecular formula C15H23N3O3 (calculated: C, 61.41; H, 7.90; N, 14.32) and that for l-geneserine (Chap 6), C15H21N3O3 (calculated: C,61.84; H, 7.27; N, 14.42), consequent upon the assumption that the former is incorrect when it was later (Merck 1926) suggested that eseridine was probably identical with l-geneserine, mp 128–129  C. This proposal also appears to have found acceptance by others [British Pharmacopoeia 1998; Heathcote 1932; Janot and Chaigneau 1947; Kutney 1977; Merck Index 2001(b)] although it appears to be invalidated by a report (Sinha 1959) which claimed that eseridine and geneserine were separated by two-dimentional paper chromatography of the total alkaloid content of the Calabar bean. However, the validity of this communication is questionable since, although the use of authentic specimens of the individual alkaloids as reference compounds on the chromatograms was claimed, no mention was made as to their source, and a separation of eseridine from geneserine was reported even though the Rf values of the two alkaloids were quoted as being 0.29 and 0.28, respectively. Be-that-as-it-may, the observation (Boehringer and Söhne 1888) that eseridine was converted into l-physostigmine upon heating with dilute mineral acid (vide supra) might not be expected if it was identical with l-geneserine. Indeed, it may perhaps be pertinent that when “geneserine was dissolved in 7M hydrochloric acid” it “was recovered in quantitative yield after basification (Na2CO3) and extraction with chloroform” (Riddell et al. 1970). Furthermore, in view of this acid-catalysed transformation [which would explain the failure by Salway (1911) to find any evidence for the presence of eseridine in the Calabar bean since during his extraction procedure the total bases were extracted with successive portions of 5% sulphuric acid] and the previously-determined formula of C15H23N3O3 (vide supra), structure 173 has been postulated (Robinson 1964b) for eseridine, its acid-catalysed conversion into l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) proceeding, as shown in 173, via the formation of 174 by a mechanism analogous to that which occurred in the formation of the 1,2,3,3a,8,8a–hexahydro-3a-methylpyrrolo[2,3-b] indole system during the synthesis of l-physostigmine by Julian and Pikl Me

MeNHCOO N

H Me

173

NHMe OH H

8.3 Isophysostigmine

161

Me

MeNHCOO

NMe N

H

Me

174 (1935a, b, c) (Sect. 2.3.1). Indeed, it appears feasible that eseridine could be the Calabar bean’s actual major alkaloidal component which, by exposure to conditions of low pH either during (see footnote 1 in Preface and Acknowledgements) or after [for example, in the formation of salts such as the sulphate and salicylate (Sect. 2.8.2)] extraction, is converted into l-physostigmine which could, thereby, like calabarine (vide supra) be an artefact!

8.3

Isophysostigmine

This alkaloid has been described (Ogiu 1904a, b) as a constituent of the Calabar bean, its name arising because the molecular formula assigned to it – although, as with eseridine (vide supra), relevant supportive experimental analytical data were not presented (Ogiu 1904a, b) – was the same as that of l-physostigmine (Ogiu 1904a, b). Apart from the results of some related pharmacological investigations, the only data relating to it is that it was either only slightly soluble or insoluble (unlike l-physostigmine which is very soluble) in ether and afforded a sulphate, mp 202  C (Ogiu 1904a, b) [which differs significantly from that of l-physostigmine sulphate – variously quoted as 140–142  C (Ogiu 1904a, b), 145  C (Henry 1949) and either 140  C or 143  C (Muhtadi and El-Hawary 1989)]. Unfortunately (vide infra) – and surprisingly – the mp of the free alkaloid was not reported (1904a, b). Moreover, other examinations of Calabar beans (Robinson 1964a; Salway 1911; Sinha 1959) have not confirmed its occurrence and the earlier observations by Ogiu (1904a, b) are not referred to in either of subsequent publications when it was reported (Polonovski and Nitzberg 1916; Polonovski 1916) that heating – in the absence of sodium! – 0.5 g of l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) in 10 ml of benzene with 0.22 g of methylisocyanate in 10 ml of benezene in a sealed tube at 100  C for 4 h afforded a product that was isomeric with l-physostigmine (C15H21N3O2) and named “isoeserine”. It had mp 195–196  C, [α]D-236 (95% alcohol) and gave a picrate, mp 170  C, but not a methiodide. Unfortunately (vide infra) a sulphate was not documented, and neither was it when the above reaction (Polonovski and Nitzberg 1916) was repeated but in the absence of the final “forcing conditions”, to again afford “isoeserine”, mp 197–198  C (Kobayashi 1938) and it was noted that the product “sont insolubles dans l’eau, tres peu solubles dans la benzine froide et [perhaps significantly] dans l’ether”. When treated with barium hydroxide it yielded barium carbonate and methylamine, but much more slowly than in the case of

8 Other Alkaloids That Have Been Isolated. . .

162

l’physostigmine and it appears to be a substituted urea derivative of l’physostigmine, with a phenolic hydroxyl group (Polonovski and Nitzburg 1916). Although yet again no reference to the studies by Ogiu (1904a, b) was made when an interesting analysis of 1H- and 13C-nmr data led to (Rosenmund et al. 1989) the elucidation of the structure and conformation of “isoeserine”. However, at this juncture and either what might appear to be from an unawareness of the studies on isophysostigmine (Ogiu 1904a, b) (vide supra) or from a prudent unacceptance of the alkaloid’s identity [because of the absence of its mp (vide supra)] with “isoeserine” [because of the absence of the latter’s sulphate (vide supra)], the opportunity was not taken to postulate for isophysostigmine a structure such as, for example, 3 (R1¼H, R2¼Me, X¼Me N C¼ON Me) containing the substituted R1O

4 5

Me 3b

A

6 7

7a

3a

B C N 8 8a X1 R2

3 2

3

urea derivative suggested by Polonovski and Nitzberg (1916) (vide supra) and for which, by comparison with and in consideration of that of eseramine (Chap. 4), there is a biogenetic rationale.

8.4

Calabatine and Calabacine

These two new alkaloids were isolated (Döpke 1963) from the Calabar bean by means of alumina column chromatography and counter-current distribution in the system Mc Ilvaine buffer/chloroform.  Calabatine, mp 119  C and ½α24 D – 98 (CHCl3), was reported to have a formula (presumably empirical) C17H25N3O2, afforded a picrate, mp 128  C (from methanol) and a salicylate, mp 211  C (from methanol) and had absorption bands in its ir spectrum (CHCl3) at 5.78μ and 6.10 μ (Döpke 1963).  Calabacine, mp 138  C and ½a24 D – 198 (CHCl3), was reported to have a formula (presumably empirical) C17H25N3O3, afforded a picrate, mp 215  C (from methanol) and a salicylate, mp 138  C (from methanol) and had absorption bands in its ir spectrum (CHCl3) at 5.76 μ and 6.1 μ (Döpke 1963). Since it would appear that sufficient quantities of these two alkaloids were available for but were wasted upon the preparation of their picrates and salicylates, it is particularly unfortunate that, if necessary through a cooperative effort with one of the many laboratories in which the requisite instrumentation was then available, their uv, 1H- and 13C–nmr and mass spectra were not investigated, and their structural elucidation therefore awaits further study. In the meantime, it is tempting to speculate

8.5 Investigations Still to Be Effected

163

that because of the C¼O stretching absorption at 5.78 μ (1730 cm
1) and 5.76 μ (1736 cm
1), respectively (Robinson 1968) and the apparent lack (Döpke 1963) of any absorption indicating N-H stretching, and by analogy with the C¼O stretching absorption of the N-methycarbamyl groups in the ir spectra (CHCl3) of l-physostigmine and l-physovenine at 1751 cm
1 and 1752 cm
1, respectively (Sect. 3.1), both these new alkaloids may contain a carbamyl group which is N,N-dialkylated (most likely N,N-dimethylated) and, furthermore, because of their relative formulae, the structural relationship between calabatine and calabacine may resemble that which exits between l-physostigmine and l-geneserine, respectively (Sect. 6.1).

8.5

Investigations Still to Be Effected

The minor (and other?) alkaloidal components of the Calabar bean are clearly worthy of further investigation. To this effect it would be of interest not only to repeat the extraction procedure of Döpke (1963) (vide supra) but also to subject a total base extract [preferably one which has not been subjected to conditions involving a low pH in order, thereby, to permit the conservation of acid-labile alkaloids such as eseridine (vide supra)] of the ripe seeds to a detailed high pressure liquid and/or gas chromatographic-mass spectral investigation. Perhaps, too, the application of these techniques should be extended to other parts of the parent vine, despite the claim by Fraser (1863) that “no part of the Physostigma venenosum is known to possess active properties except the seed or bean” [indeed, the plant was “being grown in Freetown, Sierra Leone in order that the alkaloidal content (if any) of the leaves, root bark and stem-bark may also be examined” (Robinson 1964b)], and to other natural sources, both botanical and microbial [the latter which may not thereby be regarded (see footnote 4, Chap. 2) as being alkaloidal related], in which the presence of l-physostigmine has already been established (Sect. 2.1).

Chapter 9

Non–Alkaloidal Components of the Calabar Bean

In the report of his pioneering studies at the University of Edinburgh on the toxicology of the Calabar bean by self-experimentation, Robert Christison (1855) (see footnote 14 in Chap. 1) states that “the seed, like others of its natural order, contains much inert starch and legumin, and 1.3 per-cent of fixed oil, also probably inert”, and some one century later it was also stated – although unfortunately, without reference to any experimental data whatsoever, that “the seeds contain about 48% of starch, 23% of albuminoids” (Dalziel 1948 – see also Irvine 1961), “and small amounts of oil, sugar [presumably glucose (vide infra)], mucilage etc” (Dalziel 1948). Homogenous non-alkaloidal metabolites were first isolated from Calabar beans by Windaus and Hauth (1906, 1907) who began – using the method of Hesse (1878) – by preparing from this botanical source “phytosterol”, which was shown to be a mixture of stigmasterol, C30H480 (20%) and sitosterol, C27H460 (80%) which were separated via fractional crystallisation of the acetyl derivatives of their tetraand di-bromo addition products, respectively, and are now known to have structures 175 (R¼H)[Merck Index 2001(m)] and 176 (R¼H)[Merck Index 2001(k), Me Me Me Me H

Me H

Me

H H

RO

175

© Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_9

165

166

9 Non–Alkaloidal Components of the Calabar Bean

Me Me Me Me H

Me H

Me

H H

RO 176

respectively. These reports (Windaus and Hauth 1906, 1907) of this isolation and separation were later independently confirmed (Salway 1911). Yet a serendipitous further related isolation of stigmasterol (175, R¼H) – now as its hydrate – has been reported (Julian 1936). This was obtained as “glistening small crystals” when, in attempts to isolate geneserine (Chap. 6) from Calabar beans their component oil was first extracted, washed sequentially with dilute acid and water and, when “wet” was “set aside” for “some weeks”. The beans were also found (Salway 1911) to contain a new dihydric alcohol, calabarol, C23H34O2(OH)2 which afforded a dibenzoyl derivative, and trifolianol, C21H34O2(OH)2 which had first been isolated (Power and Salway 1910) from red clover flowers – it would appear that the structures of neither calabarol nor trifolianol have, as yet, been established. At the later juncture (Salway 1911), the beans were also shown to contain the glycerides [presumably] of palmitic, stearic, behenic, oleic and linoleic acids {177, n¼14[Merck Index 2001(f)], n¼16[Merck Index 2001(l)], n¼20[Merck Index 2001(a)], 178 [Merck Index 2001(e)] and 179 [Merck Index 2001(c)], respectively} and “A sugar yielding d-phenylglucosazone (m.p. 205 )”. Me(CH2)nCOOH

177

(CH2)6COOH

Me(CH2)6

178

(CH2)6COOH

Me(CH2)3

179

9 Non–Alkaloidal Components of the Calabar Bean

167

In view of the widespread occurrence of the N-methylcarbamyl group throughout the alkaloids in the Calabar bean, its presence, where possible, in the non-alkaloidal components might not be unexpected, to afford, for examples 175 (R¼MeNHCO), 176 (R¼MeNHCO) and calabarol and trifolianol in which either one or both of the hydroxyl groups is/are N-methylcarbamylated. Perhaps the loss of this group in the above instances (Salway 1911; Windaus and Hauth 1906, 1907) was the result of some of the conditions employed during the isolation procedure, as exemplified by “. . . . the fatty oil was next hydrolysed by heating with an excess of potassium hydroxide in the presence of alcohol, when [very significantly] a considerable quantity of ammoniacal vapours was evolved. After this treatment the greater part of alcohol was removed, water added, and the alkaline liquid repeatedly extracted with ether” (Salway 1911). Perhaps a further investigation of the non-alkaloidal, as with the alkaloidal (Sect. 8.5), components of the Calabar bean might be of interest.

Chapter 10

Biological Activities of the Alkaloids of the Calabar Bean

10.1

AntiAchE Activity

10.1.1 AchE – Its Function in Neurohumoral Transmission, Structure and Inhibition Ach (180) is the cationic neurotransmitter that in the central and peripheral nervous MeCOO(CH2)2NMe3

180 systems effects the transmission of action potentials across nerve-nerve and neuromuscular synapses. In response to an action potential it is released from the presynaptic nerve and then diffuses across the synapse ultimately to bind to the Ach L receptor which serves, amongst other functions, as an ion gate for the entry of K into either the postsynaptic nerve process or the muscle or gland cell, the series of events that then follows ultimately resulting in the triggering of an action potential in the postsynaptic cell (Brand 1960; Gearien 1970; Greig et al. 1995a; Quinn 1987; Rosenberry 1975). To ensure that an orderly flow of impulses across the synapses occur, the chemical mediator, Ach (180), must be deactivated after its reaction with the receptor site and subsequent dissociation. The termination of this impulse transmission at cholinergic synapses is the principal biological role of AchE (an Ach hydrolase). This enzyme is localised in the synaptic cleft of cholinergic neurones where it is bound by a network of collagen and glucosaminoglycans to the post-synaptic terminal (Witkop 1998) and it springs into action when Ach (180) is released from the presynaptic nerve and effects the rapid hydrolysis of the neurotransmitter to choline (181) and acetate, thus terminating the Ach receptor mediated ion gating. The

© Springer Nature B.V. 2023 B. Robinson, The Calabar Bean and its Alkaloids, https://doi.org/10.1007/978-94-024-1191-1_10

169

170

10

Biological Activities of the Alkaloids of the Calabar Bean

HO(CH2)2NMe3

181 enzyme has “high stability” – “It can be kept in the refrigerator for years without loss of activity” and “freezing does not affect it” (Nachmansohn and Wilson 1951). It also has an amazing catalytic power, with a turnover number which is amongst the highest reported – namely approximately 20 million per minute (Nachmansohn and Wilson 1951; Rothenberg and Nachmansohn 1947) – for enzyme catalysis, either a single catalytic site being able to hydrolyse 104 molecules per second (Greig et al. 1995a; Quinn 1987; Rosenberry 1975) or “1 mg of enzyme is able to split about 75,000mg of acetylcholine per hour” (Nachmansohn and Wilson 1951). To a similar effect are the statements that “AChE is one of the most efficient enzymes known and has the capacity to hydrolyze 6
105 Ach molecules per molecule of enzyme per minute: this indicates a turnover time of 150 micro-seconds” (Brown and Taylor 1996; Taylor 1996), that “one enzyme molecule seems to be able to split one molecule of acetylcholine in 3 to 4 millionths of a second” (Nachmansohn and Wilson 1951) and that “The catalytic activity of this enzyme is remarkable and is expressed by its very high turnover number of 25,000, which means that AchE cleaves a molecule of acetylcholine in 40 microseconds, so that synapses can transmit 1000 impulses per second, provided the membrane recovers its normal polarization (resting potential) within fractions of a milli-second” (Witkop 1998). AchE is also a “promiscuous catalyst” with a broad substrate specificity (Quinn 1987), including l-physostigmine when this acts as its inhibitor (Fig. 10.1) (vide infra). Following Ach hydrolysis, the choline (181) is then actively transported into the presynaptic cell for subsequent conversion into the chemical mediator (Soreq and Zakut 1993) by its choline acetyltransferase-catalysed acetylation with acetyl coenzyme A. Thus, antiAchEs, by their inhibition of AchE, bring about an accumulation of Ach at the cholinergic receptor sites and are thereby potentially capable of eliciting the physiological effects of excessive stimulation of the cholinergic receptors throughout the central and peripheral nervous systems. Consequently, in view of the widespread distribution of cholinergic neurones, it is not surprising that antiAchEs have received extensive investigation as therapeutic agents associated with the treatments of cholinergic disorders, in which augmentation of cholinergic activity has proved to be beneficial, and application as toxic agents, namely as insecticides (Sect. 10.8) and – such is the nature of the human creature – as insidious warfare “nerve gases”.1 Consequently, as might be expected, the structure and mode of action of AchE, which was first isolated and purified in 1938 (Nachmansohn and Lederer 1939; Nachmansohn and Rothenberg 1945; Nachmansohn and Wilson 1951; Rothenberg and Nachmansohn 1947), has attracted considerable attention and the published results from these investigations have been the subject of reviews (Coxworth 1965; Lewis 1964; Nogrady 1985, (def); Quinn 1987; Silman and Sussman 2000; Silver

10.1

AntiAchE Activity

– O MeNH H O

O C

N ES

C H N O

171 Me

HO Me

O +

N NHMe Me

MeNH – O H

O C

N

N

O C +

N N H Me Me

O

ES

AS

AS

N-Methylcarbamylated Enzyme

Tetrahedral Intermediate

H2O O – O

O C

MeNH C H + OH

H N

N

O

ES

MeNH2 CO2

AS

Regenerated Enzyme

Fig. 10.1 Schematic representation of the inhibition of AchE by the ring-C opened tautomer of lphysostigmine (Sect. 10.4) (ES, esteratic site; AS, anionic site). Transfer of the N-methylcarbamyl group from the inhibitor leads to the formation of a N-methylcarbamylated enzyme (Wilson et al. 1960, 1961) which is relatively unstable [with a half-life of about 38m (interestingly, those of the carbamyl and dimethylcarbamyl enzyme are about 2m and about 27m, respectively) (Wilson 1963)] and rapidly hydrolyses in water (Main and Hastings 1966, Triggle et al. 1998, Wilson et al. 1960, 1961) [see also (Dale and Robinson 1970; Robinson and Robinson 1968)]

1974a; Soreq et al. 1992; Soreq and Zakut 1993; Taylor 1996; Taylor and Radić 1994; Wilson 1960, 1963). Early kinetic studies (Nachmansohn and Wilson 1951) indicated that the active site of the enzyme contained two subsites, namely the “anionic” locus and the “esterase” (an “esteratic”) locus, which binds – by coulombic attraction  the quaternary nitrogen atom of Ach and corresponds to the cholinebinding pocket and the catalytic system, respectively (Fig. 10.2) (Harel et al. 1996; Sussman et al. 1991; Wilson 1960). AchE is an α/β protein (Ollis et al. 1992; Sussman and Silman 1992) with an ellipsoidal shape measuring 45
60
65 Å (Sussman and Silman 1992). Its enzymic biological properties are essentially totally embraced within what, for Torpedo Californica AchE, has been shown to be the 537 members of its polypeptide backbone of which the amino acid sequence and secondary structure have been established (Sussman et al. 1991). Thus it presents its Ach-binding pocket as an extraordinary aspect of its structure known as the “activesite gorge”, a deep narrow gorge some 20 Å in depth that penetrates halfway into the enzyme and widens out close to its base (Sussman and Silman 1992) where the catalytic machinery (Fig. 10.3) (vide infra) is located. A substantial portion (circa 40%) of the surface of the gorge is lined with fourteen aromatic residues, namely

172

10

Biological Activities of the Alkaloids of the Calabar Bean

ES –

H N

O C

O

N H Me

O O

C

Me CH2

CH2

+

N

O

Me

Me AS

Fig. 10.2 The interaction of Ach with AchE (ES, esteratic site; AS, anionic site)

O

O C Glutamate moiety 327

O

O –

N

H

C

H

O

N

O CH2 CH2 NMe3

Serinyl moiety 198

O C O H N

C

O CH2 CH2 NMe3

N H

O

Tetrahedral Intermediate O

–

C

Me O H C O N H

Me

O H N

Histidyl moiety 438

O +

+

Me

–

C

O H

O

O +

–

Me N

N

H

C + HO CH2 CH2 NMe3

O O

H

Acetylated Enzyme

Tetrahedral Intermediate O C O

–

–

H N

N

H

O

O

+ Me C

+

+

H

O Regenerated Enzyme

Fig. 10.3 The involvement of the catalytic triad in the hydrolytic process at the esteratic subsite (locus) of human AchE – the so-called “charge relay system” (Blow et al. 1969; Taylor 1996) – see also [Nogrady 1985 (e)]

10.1

AntiAchE Activity

173

tyrosinyl [tyr (182, R¼OH)] 70, tryptophanyl [trp (183)] 84, trp 114, tyr 121,

R

CH2

NH CH CO

182

NH CH2 CH CO N H

183 tyr 130, trp 233, trp 279, phenylalanyl [phe (182, R¼H)] 288, phe 290, phe 330, phe 331, tyr 334, trp 432 and tyr 442 (Sussman and Silman 1992). Because of this, it is also referred to as the “aromatic gorge” (Harel et al. 1996). Significantly, the gorge contains only a few acidic residues including, at the very top, aspartate [asp (184, n¼1)] 285 and glutamate [glu (184, n¼2)] 273, about halfway down, asp

NH C (CH2)n CH O CO

O

184 72 (hydrogen-bonded to tyr 334), and near the bottom, glu 199 (Sussman and Silman 1992). In keeping with these data, earlier studies using charged and uncharged isosteric substrates and inhibitors had led to the conclusion that “the anionic site is uncharged {contrary to “the previously predicated [(Quinn 1987)] anionic domain” (Soreq et al. 1992)} and lipophilic” and other studies had “supported the presence of hydrophobic aromatic residues within the active site (Sussman and Silman 1992, see also Soreq et al. 1992),. It has also been suggested (Sussman and Silman 1992) that Ach “may also use the aromatic surface of the gorge to slide down to the bottom via a series of low-affinity sites” in which “This aromatic guidance is somewhat similar in concept to the “electrostatic guidance. . .” The role of these aromatic residues would seem to be “substrate recognition by dipole-dipole interaction between the π

174

10

Biological Activities of the Alkaloids of the Calabar Bean

electrons of the aromatic rings and the quaternary nitrogen of the Ach. Recognition is not merely confined to the substrate-binding pocket but seems also to provide an ‘aromatic guidance’ mechanism for rapid translation of the substrate from the surface to the active site” (Sussman et al. 1991; Sussman and Silman 1992), the cationic substrate being also attracted to the surface by the strong electrostatic field that is generated by the enzyme (Gilson et al. 1994). Moreover, there is a strong electrostatic dipole directed towards the bottom of the gorge and whereas this should accelerate the penetration of the positively-charged substrate toward the catalytic site (Ripoll et al. 1993; Tan et al. 1993), it “would impede the clearance of the product choline from the mouth of the active site” and which could therefore be released through a “back door” in a “thin wall near the base of the active site, at residues Met83 [methionyl (185) 83] and Trp84 [tryptophanyl (183) 84]” (Gilson et al. 1994).

NH Me S (CH2)2 CH CO

185 Although “Site directed mutagenesis experiments [Faerman et al. 1996; Kronman et al. 1994], designed to test the existence of such putative ‘back doors’ or ‘side doors’, have not provided support for their existence” (Silman and Sussman 2000), evidence in favour of a “back door” has been forthcoming from a crystallographic study (Bartolucci et al. 1999) of the conjugate of Torpedo Californica AchE with the l-physostigmine analogue 98 (R1¼186, R2¼Me, X¼NMe, n¼1). Nevertheless, it Me O

N (CH2)8 NHCO

Me

186 would appear (Silman and Sussman 2000) that further experimental verification of this hypothesis may still be required. However, the necessity at all for a “back door” may be obviated if, post-hydrolytically and thus possibly invoking a temporary disruption to the “catalytic triad” (vide infra), a new aromatic profile developed in the “aromatic gorge” which would allow, and perhaps even facilitate, the passage and release of the product choline (181). At the bottom of the “active-site gorge” is situated the “esteratic site”, now known as the “catalytic triad” (Sussman et al. 1991, Taylor 1996), namely the three amino acid moieties involved in the hydrolytic process which are, in human AchE,2 the

10.1

AntiAchE Activity

175

glutamate 327, histidyl 438 and serinyl 198 moieties in the enzymic polypeptide chain (Fig. 10.3) (Greig et al. 1995a; Soreq and Zakut 1993) [see also (Soreq et al. 1992)] [but see, for example, (Yu et al. 2010)]. Although these functional groups are separated by many chemical (peptide) bonds, they nevertheless are able to participate in multifunctional enzyme catalysis when they are in stereochemical proximity attainable by the operation of God’s (Robinson 2012) hydrogen bonds (Blow et al. 1969) between histidyl 438 and glutamate 327 and between histidyl 438 and serinyl 198 (Fig. 10.3). This forms what has been described as a “charge relay system” (Blow et al. 1969; Taylor 1996) which conducts electrons from the buried glutamate 327 carboxylate group to the surface where it renders the oxygen atom of the serinyl 198 hydroxyl group strongly nucleophilic, and enough so as to be capable of attacking the carbon atom of the carboxyl group in a substrate, in this case Ach (180) ultimately to release choline (181) and leave the acetylated enzyme which then rapidly reacts with water to afford acetate and regenerate the enzyme (Figs. 10.3 and 10.4). This enzyme, however, is still enigmatic since “The active site of an enzyme with a rapid turnover that operates almost at the rate of diffusion, would tentatively be expected to be easily accessible to the substrate, rather than buried deeply within the enzyme” (Soreq et al. 1992) and “why the active site of a particularly rapid [vide supra] enzyme should be located near the bottom of a deep cavity remains a mystery” (Sussman and Silman 1992). In what has been refered to (Gaddum 1962) as “one of the first applications of the principle of antimetabolites”, the first synthetic anticholinesterases were prepared by – O

H N ES

Me O HC N O

+ NMe3

O

– O

H Me N

C

N

O

HO

C O

+ NMe3

+

ES Acetylated Enzyme

AS

Tetrahedral Intermediate

AS

H2O O

O

– O C

+ H

Me C

H

H N

N

+

– +

O

O

ES AS Regenerated Enzyme Fig. 10.4 Schematic representation of Ach acting as a substrate for AchE (ES, esteratic site; AS, anionic site)

176

10

Biological Activities of the Alkaloids of the Calabar Bean

Edgar Stedman who, together with George Barger (his colleague in the Department of Medical Chemistry at the University of Edinburgh) (see footnote 14 of Chap. 1), and Professor Robert Robinson FRS (then at the University of Manchester) (see footnote 5 of Chap. 2), had been one of the two groups who first published (Robinson 1925; Stedman and Barger 1925) (Sect. 2.2) the structure of physostigmine. By these later related synthetic endeavours, in which he was soon joined by his wife Ellen (Holmstedt 1972) (Stedman and Stedman 1929, 1931a, b, 1932see also Stedman et al. 1932, 1933) he thereby allegedly3 “carried on the tradition that all the best work on physostigmine was done in Edinburgh” (Gaddum 1962) (see footnote 3). Edgar Stedman’s investigations4 in this area, which have been comprehensively reviewed (Barger 1936; Henry 1949, Long 1963; Long and Evans 1967; Marion 1952; Stempel and Aeschlimann 1956), began with the observation that the physiological activity of l-physostigmine disappeared upon its hydrolysis to the phenol, eseroline (3, R1¼H, R2¼Me, X¼NMe), which “led to the suggestion {although “it is now well-known that the presence of a carbamyl group is not essential for activity (Avison 1954; Galli et al. 1982; Long and Evans 1967) but may be a factor contributing to high anticholinesterase activity” (Long and Evans 1967)} that the physiological properties of physostigmine depended upon the fact that the alkaloid was a phenyl ester of methylcarbamic acid (ie a urethane)” (Stedman 1926). Consequently, urethanes were prepared, using methylisocyanate, from simple phenols that also contained a basic group “in order to permit the formation of [presumably water] soluble salts” (Stedman 1926). Since methylisocyanate would also react with primary and secondary amines, the three isomeric dimethylaminophenols were employed as starting materials to afford products that were found to have a physiological activity, namely a miotic activity on the cat’s eyes, quite similar to that of physostigmine itself, with the interesting observation that whereas quaternisation with iodomethane of the ortho- and para- isomeric products decreased activity, an increase resulted from a similar reaction with the meta-isomer (Barger 1936; Stedman 1926; Stempel and Aeschlimann 1956). It is interesting that in the above studies (Stedman 1926, 1929; Stedman and Stedman 1929) it was the meta isomers that showed “marked physostigmine activity wheres members of the corresponding ortho and para series were only slightly active”, it being suggested and experimentally supported that “this anomaly may be the result of activation by the alkyl residue ortho to the amino group attached to the benzene ring in physostigmine” (Stevens and Beutel 1941). Subsequently, Stedman (1929) extended his studies to the synthesis and pharmacological evaluation of the urethanes of isomeric hydroxybenzyldimethylamines which ultimately led to the discovery of the appropriately-named miotine (187), “a detailed pharmacological MeNHCOO

CHMe NMe2

187

10.1

AntiAchE Activity

177

investigation” of which “showed that in addition to the [miotic] effects on the pupil, the other actions, notably that on the intestine, was also similar to those of physostigmine” (Barger 1936). These activities were also independantly observed in the laboratories at ROCHE (Hoffmann-La Roche), whilst seeking a l-physostigmine substitute for use in post-operative intestinal stasis, and the consequent synthesis and examination of a long series of urethanes, (Aeschlimann and Reinert 1931), led to neostigmine (188, R1¼R2¼Me), from which the methysulphate (188, R1¼R2¼Me, R1NCOO R2 NMe3

X

188 X¼MeSO4) was selected and introduced into practical medicine by ROCHE (Hoffmann-La Roche) under the name Prostigmine (Barger 1936; Long and Evans 1967). Another active quaternary urethane is pyridostigmine (189, R1¼R2¼Me) in R1NCOO R2 N X Me

189 which the quaternary grouping has been formed from a heterocyclic N-atom (Avison 1954, Stempel and Aeschlimann 1956). Unlike l-physostigmine, the quaternary salts cannot cross the blood-brain barrier (Birtley et al. 1966; Duvoisin and Katz 1968; Greig 1992; Lee et al. 1975; Muhtadi and El-Hawary 1989; Pardridge 1988; Wright 1976). Consequently, their introduction into medical practice has “greatly diminished the medical use of the relatively toxic Phy [physostigmine] {the subject of some deliberation (Caine 1979) – references to studies on toxicity data relating to lphysostigmine can be found in [Merck Index 2001(h)]} for the treatment of cholinergic disorders” (Brossi 1990) (see also Taylor 1996). The first synthetic antiAchEs incorporating a para-aminophenol moiety in a 5-hydroxyindoline nucleus and thus omitting ring C of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) and “showing physostigmine” activity by intravenous injection into mice, although their activity was not compared with that of l-physostigmine, were 171 (R1-R3¼Me, R4¼H) and its methiodide, with like activity also being inherent in similar derivatives of 1,2,3,4-tetrahydro-6-and 8-hydroxy-l-methylquinolines (Gardner and Stevens 1947). Within some 6 years, other similar 5-hydroxyindoline derivatives, namely 171 (R1–R4¼H; R1-R3¼H,

178

10

Biological Activities of the Alkaloids of the Calabar Bean

R4¼Me and R1¼R3¼Me, R1¼R4¼H), as their hydrochlorides and methiodides were found to exhibit miotic activity in the rabbit eye, with 171 (R1–R3¼H, R4¼Me) hydrochloride, in particular, being found to be a highly-active miotic even at a concentration of 1 in 100,000 but its activity was not compared with that of l-physostigmine (Kolosov and Preobrazhenskii 1953a, b, c), and the syntheses of 171 (R1¼R2¼H, R3¼Me, R4¼Me and Et), have been reported (Kolosov and Preobrazhenskii 1953b). These above studies have previously been reviewed (Ahmed 1966; Dale 1969). MeNR1COO

5

R4

4 3b

3a 6

N

7a

7

8

8a

R3 R2

Me 171

Another 5-hydroxyindoline derivative 171 [R1¼R2¼H, R3+R4¼(CH2)2 NMe (CH2)2] has also been synthesised but has been found to possess only weak antiAchE activity (Kretz et al. 1952). However, subsequent to the syntheses of 171 (R1¼R2¼H, R3¼R4¼Me) (Ahmed 1966; Ahmed and Robinson 1965; Kolosov and Preobrazhenskii 1953b) and 171 (R1¼H, R2-R4¼Me) (Ahmed 1966; Ahmed and Robinson 1965) it was found (Dale 1969; Dale and Robinson 1970; Robinson 1971) that the former’s methiodide was circa 100 times more active as an antiAchE than was l-physostigmine when measured in vitro using erythrocyte AchE. It is interesting that the hydrochloride and methiodide of 190 (n¼2) and the hydrochloride of 190 (n¼3) all exhibit miotic activity in the rabbit eye (Kolosov and Preobrazhenskii 1953c). Me

Me2NCOO

(CH2)nNMe2 N

O

Me

190 It has long been recognised that “almost all functions of the body are controlled by the nervous system” (Nachmansohn and Wilson 1951) and that l-physostigmine possesses certain drawbacks as a therapeutic agent, since it is unstable in both aqueous solutions (as a salt) and as a solid (sect. 2.8) and in doses only slightly greater than the therapeutic it is very toxic (Brossi 1990; Kolosov and Preobrazhenskii 1953a; Taylor 1996). Therefore it is perhaps not surprising that an amazingly large number of compounds have been investigated for antiAchE activity and consequently that “Agents which inhibit cholinesterases must be counted in

10.1

AntiAchE Activity

179

thousands rather than hundreds and, in consequence, they form a vast field of study which impinges on many disciplines. They have a relevance to biochemistry, pharmacology, physiology, toxicology and ecology and during the last 30 years [as of 1974] have engendered a truly enormous literature” (Silver 1974a) with such studies having been the subject matter of many reviews. For example, it has been stated that “In the monograph by Augustinsson about 900 publications are quoted which appeared during the last 15 years” (Nachmansohn and Wilson 1951) and a spate of work over some three decades has been comprehensively alluded to in review (Silver 1974a), a study that incorporates some twenty references to earlier reviews. Other reviews have also appeared (Avison 1954; Brufani and Filocano 2000; Fisher 2002; Gearien 1970; Giacobini 2000a,b, 2004; Koelle 1963; Koelle and Gilman 1949; Long 1963; Long and Evans 1967; Robinson 1968,1971; Soreq and Zakut 1993; Stempel and Aeschlimann 1956; Taylor 1996), as have interesting reports upon a diverse range of other new antiAchEs (Abramson et al. 1989; Coleman et al. 1987; Hersh 1981; Karlsson et al. 1985; Lin et al. 2001; Marquis 1983; Marquis and Lerrick 1982; Roddick 1989; Zhu et al. 1996). “Clinicalyrelevant cholinesterase inhibitors” as of 2006 have been listed (Luo et al. 2006 – see also Muñoz-Ruiz et al. 2005; Taylor 1996). A wide range of alkaloids5 and of organophosphates (see footnote 1) have also been found to exhibit antiAchE activity, the former observations being much as would be expected (see footnote 5) and the latter, since they are not structural analogues of physostigmine, have been omitted from another such compilation (Stempel and Aeschlimann 1956). The reaction between AchE and its substrate Ach involves the initial attachement (Fig. 10.2) of the latter with the enzyme, ultimately affording the acetylated enzyme which then undergoes rapid hydrolysis to regenerate the enzyme (Figs. 10.3 and 10.4). The routes by which antiAchEs interfere with this process, which is in all cases similar to the reaction between the Ach and the enzyme, depends upon the nature of the inhibitor and may involve (Bergmann and Shimoni 1952; Silver 1974a) the blockage of either the anionic site or the anionic and esteratic sites or the esteratic site. Those inhibitors that attach to the anionic site include some of the quaternary compounds and basic compounds [including alkaloids (see footnote 5)] in their protonated form, whose structures can extend in either instance far enough also to mask the esteratic site, thereby offering the enzyme some degree of prophylactic protection against inhibition by organophosphates (see footnote 1). Inhibitors which bind with both sites simultaneously do so by a reaction that is qualitatively similar to that between the enzyme and Ach and are exemplified by l-physostigmine (Fig. 10.1) and Prostigmine (188, R1¼R2¼Me, X¼MeSO4), the inhibition resulting from the relative slowness by which the enzyme is hydrolytically regenerated (Fig. 10.1) – and both of which may thus protect the enzyme against the irreversible combination with organophosphates (Sect. 10.7.1 and see footnote 1), and those where sole reaction is at the esteratic site include the majority of the organophosphates (see footnote 1) which afford phosphorylated enzyme which are extremely resistant to hydrolysis – an interesting type of inhibitor involves organophosphates that also contain quaternary ammonium groups (Avison 1954). The most recently introduced

180

10

Biological Activities of the Alkaloids of the Calabar Bean

inhibitors are of the “dual-binding site” type in which a catalytic site binding moiey is separated by a “linker” of suitable length and nature from a moiety that binds simultaneously with a peripheral site (Muñoz-Ruiz et al. 2005).

10.2

Pharmacology of l-physostigmine

Apart from its use for ritualistic killing (Chap. 1) in the cruel and superstitious Efik society that once prevailed in Southeastern Nigeria (see footnote 12 of Chap. 1), the Calabar bean has also found many non-macabre uses throughout the coastal regions of West Africa. Thus, “in Zaire the Yassi and Congo tribes in the southwest of the country use the seeds against asthma in children: the beans are put into a calabash containing glowing embers, the sick children inhale the vapours once a day” (Neuwinger 1996). In Cameroons [sic], crushed beans mixed with palm oil have been used as a remedy for lice (Dalziel 1948, Irvine 1961) – as they also have been in Calabar and Equat. Africa (Irvine 1961), with the powdered cotyledons sometimes being left on the head for several days (Irvine 1961). Similary in Zaire “the Babwa make a paste against lice from the seed oil” and “the Gbaya of the CAR [Central African Republic] use the seeds or the powdered fruit to kill head lice, the same report comes from Gabon” (Neuwinger 1996). Also in Cameroons [sic] the Bakwiris and the Kpe have used the crushed seeds, along with Spilanthes and other drugs, in the local treatment of articular rheumatism (Santesson 1926) and rheumatism in the joints (Neuwinger 1996), respectively. On Iv. Coast, doses of small pieces of crushed bean in a little water have been used for oedemas (Kerharo and Bouquet 1950), and the Attié have also used crushed seeds for local application in parasitic skin diseases (Ivanoff). The beans have also been used for killing mice (Wilczek) and constitute one of the ingredients of arrow poisons used in hunting in the AbongMbang region in Southeastern Cameroon (Neuwinger 1996), in the forest area south of Banqui (Central African Republic) (Neuwinger 1996), and by the pygmies in the border area of Gabon/Cameroon (Neuwinger 1996) – in the preparation of these poisons, all the ingredients are finely ground, mixed with a little water to a red-brown paste and smeared on the arrow tips (Neuwinger 1996). In 1948, the seeds were on sale in Enugu medicine market (Irvine 1961) and in 1988 the author had no difficulty in locating some for sale in the market at Calabar. These above data indicate that the Calabar bean is, indeed, ethnopharmacologically fairly widespread throughout western and central equatorial Africa. The pharmacology of its major biologically-active alkaloidal component, l-physostigmine, has also been included in reviews (Brossi 1990; Brossi et al. 1996; Coxworth 1965; Fraser 1938a, b, c; Giacobini 2004; Henry 1949; Holmstedt 1972; Koelle and Gilman 1949; Long and Evans 1967; Muhtadi and El-Hawary 1989; Neuwinger 1996; Robinson 1971 1988a, b, 2002; Schneck et al. 1989; Silver 1974b; Takano and Ogasawara 1989; Taylor 1996; Taylor and Radić 1994; Triggle et al. 1998; Zhao et al. 2004 – see also Moroi and Lichter 1996) – of which the most conventional and comprehensive is that by Triggle et al. (1998) and articles

10.3

Role of l-physostigmine in the Discovery of the Mechanism. . .

181

specifically relating to its pharmacokinetics and metabolism have also appeared (Boyer and Somani 1985; Hartvig et al. 1986; Isaksson and Kissinger 1987; Zhao et al. 2004). However, the earlier of these reviews and articles may be regarded as being somewhat out-of-date, although further experimental data in these latter areas will, no doubt, be forthcoming from clinical trials on l-physostigmine and its analogues that may become necessary. Of interest in this area have been the syntheses of l-40 -hydroxyphenserine (98, R1¼4-HOC6H4NHCO, R2¼Me, X¼NMe, n¼1) and l-40 -hydroxyphenyl-N(1)–norphenserine (98, R1¼4HOC6H4NHCO, R2¼Me, X¼NH, n¼1) which have been suggested as potential metabolites of phenserine (Yu et al. 1999) which is under investigation for the treatment of dementia of the Alzheimer’s type (Sect. 10.7.2) – surprisingly, the in vivo hydrolysis of the phenylcarbamyloxy group to ultimately afford l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) [probably a major metabolite of l-physostigmine (Schönenberger et al. 1986b) and the formation of which “may contribute to the toxic effect of physostigmine” (Zhao et al. 2004) and the potentially toxic aniline [cf (Fig. 10.1)] (Sect. 10.7.2) was ignored in this investigation (Yu et al. 1999). Likewise, too, the pharmacological reviews might be deserving of updating, the need for which may be considered to have been, at least, partially satisfied by the subject matter currently presented in this chapter.

10.3

Role of l-physostigmine in the Discovery of the Mechanism of Neurohumoral Transmission

“The alkaloids – notwithstanding the advent of the antibiotics and of many new and valuable synthetic drugs – still constitute an indispensable part of our medicinal arsenal” (Svoboda 1963), a situation that also currently appertains. Those which either still play or have played a role in the cure and alleviation of diseases have been subject to review (Brossi 1997) (see also Creasey 1983(a) (limited to monoterpenoid indole alkaloids), Svoboda 1963; Taylor 1963) and include (Brossi 1997) atropine (mydriatic and antidote in organophosphate poisoning), cocaine (local anaesthetic), codeine (analgesic and antitussive), colchicine (antimitotic and the drug of choice to treat acute gout and familial Mediterranean fever), demecolcine (clinically useful in the treatment of chronic leukaemia), ephedrine (bronchodilator and orally-active sympathomimetic), ergotamine (smooth muscle stimulant, especially of blood vessels and the uterus-stimulating contraction), morphine (analgesic), quinine (antimalarial) (see also Creasey 1983(g)), quinidine (used in the control of cardiac arrhythmias) (see also Creasey 1983(g)), and vinblastine {vincaleukoblastine; either Velban or Velbe (Creasey 1983(f))} and vincristine {leurocristine, Oncovin (Creasey 1983(f))} [the most important alkaloids (employed in 1997) in medical practice because of their use for the treatment of malignant and non-malignant diseases such as leukaemia, breast cancer, and Hodgkin’s disease (Brossi 1997)]. Other examples of clinically useful alkaloids include (Creasey 1983(a)) ajmaline

182

10

Biological Activities of the Alkaloids of the Calabar Bean

(useful in the management of both ventricular and supraventricular arrhythmias), curare alkaloids (neuromuscular blocking agents), ellipticine and derivatives (exhibit cytotoxic activity), ibogamine (marginally cytotoxic and in rats exhibits diuretic activity and marginal hypoglycaemic activity and has a suggested clinical use as an antidepressant) (Creasey 1983(e)), pilocarpine (a miotic used in the treatment of glaucoma) (Taylor 1963) and (Sect. 10.5), and reserpine (used in the treatment of either severe or intractable hypertension) (claimed by Taylor in 1963 that “up to the present time it is the most valuable natural drug to have been discovered”). In addition to being part of the armamentarium in the treatment of disease, some alkaloids – for examples atropine, curare, muscarine and nicotine – have also found use as “experimental tools” in the elucidation of physiological and pharmacological mechanisms (Holmstedt 1972; Robinson 2002). Indeed, in both this and the therapeutic domain, not least l-physostigmine occupies significant positions. Thus, as one of the first examples whereby the mechanism of action of a drug could be defined at the molecular level relatively simply, l-physostigmine was the first alkaloid that was proven to exert its pharmacological activity through the inhibition of an enzyme when it was recognised as a reversible inhibitor, both in vitro and in vivo, of AchE (Holmstedt 1972; Robinson 1964b, 2002). Consequently it became (and still is in many laboratories) the blocking agent of choice for this enzyme and thereby played such a fundamental role in the elucidation of the mechanism of neurohumoral transmission at the molecular level (Holmstedt 1972; Robinson 1964b, 2002) (Sect. 10.1.1), namely how nerve impulses are transmitted across nerve-nerve and neuromuscular synapses by the mediation of the cationic neurotransmitter Ach (180), and is also the reason for its use, or that of its analogues which have been developed to improve the therapeutic index of the somewhat toxic parent alkaloid and which have thereby been the genesis of such a valuable group of therapeuticals for the treatment of a wide variety of neurological disorders associated with irregularities in cholinergic transmission in which augmentation of cholinergic activity has proved to be of value (vide infra). Certainly, the bean that was responsible for so much evil and death in the manifestly cruel, superstitious and violent society that once prevailed – as it had done for centuries – in Calabar (Chap. 1), in our’s has, via its component toxic alkaloid l-physostigmine, ultimately become the source of so much good. It was TR Elliott, then a young George Henry Lewes Student in Physiology [a studentship that has been held earlier by Henry Dale (Dale 1958)] at the University of Cambridge (Elliott 1905a, b), who was the first to brilliantly anticipate and express himself clearly – alluded to by his friend Henry Dale (Feldberg 1969) as a “brilliantly precocious suggestion” (Dale 1958) – upon chemical transmission in the autonomic nervous system, in a short prophetic note (Elliott 1904) presented to the Physiological Society on 21st May 1904.6 Although this note has been referred to in review (Bacq 1975), it is not recognised in either earlier (Feldberg and Fessard 1942) or later (Feldberg 1979) studies, both of which, nevertheless, clearly state that Elliott was “the first to conceive the idea of chemical transmission in the autonomic nervous system” but refer to his Sydney Ringer Memorial Lecture [(Elliott 1914) and (see footnote 6)] as his means of communication. However, at this juncture – when he was then an assistant physician

10.3

Role of l-physostigmine in the Discovery of the Mechanism. . .

183

to the University College Hospital at which Ringer himself had previously made such major contributions [(Elliott 1914) and (see footnote 6)] – it was recognised (Feldberg 1979; Feldberg and Fessard 1942) that Elliott had “even made an attempt to extract the active substance” when he wrote “I have tried in vain to discover an active substance in the muscle plates of striped muscle. And Professor Herring was also disappointed when he examined for this purpose the electrical organs of the skate, which are exaggerated motor plates. The ganglia of other visceral nerves, the auriculo-ventricular bundle of His, the spinal cord itself – all have been extracted and nothing as yet found. But it is hard to forego the belief that such discoveries lie in the lap of the future” (Elliott 1914). Feldberg (1979) has recounted how that over dinner one evening at the Athenaeum, when he and Elliott were the guests of Henry Dale, he “asked Elliott what kind of extraction they had used. As far as he could remember they had simply used water and saline extracts. If only they had used hydrochloric acid and thus prevented the spontaneous and enzymic hydrolysis of acetylcholine they could hardly have missed the high activity in the extracts from the electric organ. What a fantastic near-miss it had been!”A bull’s-eye was, however, to be achieved using l-physostigmine! The two main uses of l-physostigmine as an “experimental tool” have already been delineated (Silver 1974a) when it was noted that “In bioassay it will prevent the cholinesterase of the test-organ (e.g. dorsal muscle of the leech) from hydrolysing any ACh in the solution under test. Similarly, in physiological investigations of cholinergic mechanisms it will preserve endogenous ACh, or ACh and other cholinomimetics which are injected into the animal either locally or systemically. Without such an inhibitor, cholineserase-susceptible drugs would be hydrolysed before they could exert their effects”. The work that initially verified the existence of a chemical transmission across the nervous synapse from nerves to an end organ was effected by Otto Loewi in whose absorbing – and in parts alarming – autobiography (Loewi 1960) is presented the following dramatic account of the commencement of these investigations:As far back as 1903, I discussed with Walter M. Fletcher from Cambridge, England, then an associate in Marburg, the fact that certain drugs mimic the augmentary as well as the inhibitory effects of the stimulation of sympathetic and/or parasympathetic nerves on their effector organs. During this discussion, the idea occurred to me that the terminals of those nerves might contain chemicals, that stimulation might liberate them from the nerve terminals, and that these chemicals might in turn transmit the nervous impulse to their respective effector organs. At that time I did not see a way to prove the correctness of this hunch, and it entirely slipped my conscious memory until it emerged again in 1920. The night before Easter Sunday of that year I awoke, turned on the light, and jotted down a few notes on a tiny slip of thin paper. Then I fell asleep again. It occurred to me at six o’clock in the morning that during the night I had written down something most important, but I was unable to decipher the scrawl. The next night, at three o’clock, the idea returned. It was the design of an experiment to determine whether or not the hypothesis of chemical transmission that I had uttered seventeen years ago was correct. I got up immediately, went to the laboratory, and performed a simple experiment on a frog heart according to the nocturnal design. I have to describe briefly this experiment since its results became the foundation of the theory of chemical transmission of the nervous impulse.

184

10

Biological Activities of the Alkaloids of the Calabar Bean

The hearts of two frogs were isolated, the first with its nerves, the second without. Both hearts were attached to Straub canulas filled with a little Ringer solution. The vagus nerve of the first heart was stimulated for a few minutes. Then the Ringer solution that had been in the first heart during the stimulation of the vagus was transferred to the second heart. It slowed and its beats diminished just as if its vagus had been stimulated. Similarly, when the accelerator nerve was stimulated and the Ringer from this period transferred, the second heart speeded up and its beats increased. These results unequivocally proved that the nerves do not influence the heart directly but liberate from their terminals specific chemical substances which, in their turn, cause the well-known modifications of the function of the heart characteristic of the stimulation of its nerves.

The results from further investigations on this subject were published over the following 6 years in a series of papers (Holmstedt 1972). The first of these (Loewi 1921a), in four pages, and with no references presented the inaugural demonstration for the existence of chemical transmission from nerves to an end organ and has been referred to as “truly one of the most remarkable communications in the history of science” (Holmstedt 1972) and as “one of the most remarkable communications of medical science” (Neuwinger 1996). During the course of the subsequent papers the identity of the chemical transmitter was gradually narrowed – indeed, a choline ester was an early suggestion in the second paper (Loewi 1921b) – until a footnote “Bei der auβerordentlich starken Wirksamkeit des Vagusstoffes könnte sogar daran gedacht werden, daβ er Acetylcholin ist” in the penultimate communication (Loewi and Navratil 1926a) pointed to Ach. Further verification of this was forthcoming from what has been referred to as “The famous eleventh paper in the series” (Holmstedt 1972) in which it was reported (Loewi and Navratil 1926b) that a rise in the concentration of Ach at the synapses, with a resulting increase in the parasympathetic excitation occurs consequent upon the inhibition of an enzyme [later called acetylcholinesterase (AchE)] – primarily by l-physostigmine, this being recognised (Loewi and Navratil 1926b) as the first time that the pharmacological action of an alkaloid had been defined in terms of the inhibition of an enzyme. It might appear that Holmstedt (1972) in his colourful review has credited the above investigations of Loewi as resulting in the first recognition and use of lphysostigmine as an inhibitor of AchE which permitted, by inhibition of its enzymic destruction, the detection of Ach released in a biosystem. However, although he refers to Fühner’s discovery (Fühner 1918) “that soaking the plain muscle of a leech in a very weak solution of eserine (1ppm.) did not alter the stimulant action of pilocarpine or of choline, but increased that of acetylcholine a million-fold” which led to the suggestion “that the effect of eserine was to inhibit hydrolysis of acetylcholine by the tissues”, and notes that “Fühner’s technique was later developed by Feldberg [(Feldberg 1968)] into a bioassay for acetylcholine,” he was apparently unable to give the latter duo a share with Loewi in the honours for the application of this technique (vide infra). As Loewi was to be when he was forced to leave his native city of Graz, Austria for London on 28th September 1938 (Loewi 1960), Feldberg had been also but an

10.3

Role of l-physostigmine in the Discovery of the Mechanism. . .

185

earlier victim of the Nazi authorities as he describes in his reminiscences (Feldberg 1979): One day in 1933, shortly after Hitler came to power, the Director of the Institute of Berlin where I was working sent for me and informed me that I had been dismissed, must leave the Institute at the latest by mid-night that day, and was not allowed to enter it any more. A few weeks later, someone told me that the representative of the Rockefeller Foundation was staying in Berlin and that I should try to see him. I succeeded. He was most sympathetic but said something like this: ‘You must understand, Feldberg, so many famous scientists have been dismissed whom we must help that it would not be fair to raise any hope of finding a position for a young person like you.’ Then, more to comfort me, ‘But at least let me take down your name. One never knows.’ And when I spelt out my name for him, he hesitated, and said, ‘I must have heard about you. Let me see’. Turning back the pages of his diary, he suddenly said, delighted himself: ‘Here it is. I have a message for you from Sir Henry Dale whom I met in London about a fortnight ago. Sir Henry told me, if by chance I should meet Feldberg in Berlin, and if he has been dismissed, tell him I want him to come to London to work with me. So you are all right,’ he said warmly. ‘There is at least one person I needn’t worry about any more’. Having received Dale’s message from the representative of the Rockefeller Foundation, I wrote at once to Sir Henry, immediately received an invitation, and as soon as all necessary formalities were completed, I was literally thrown out by my wife who, with our two small children, stayed behind to get everything ready and packed for what might, and eventually did, become an emigration. I landed at Harwich on 7 July 1933.

Thus it was that, in 1933 Feldberg found himself working with Sir Henry Dale in his laboratory in London, he having known Dale since 1927 when he was a visiting scientist in his department (Feldberg 1979). Feldberg continues in his reminiscences (Feldberg 1979): Between 1933 and 1936, the index volume of the Journal of Physiology for volumes 61 to 100 lists, from Dale’s laboratory, fourteen publications, printed communications or full papers, on acetylcholine in ganglionic and neuromuscular transmission, and another ten publications on transmission by acetylcholine in other tissues: in the suprarenal medulla, stomach wall, sweat glands, salivary glands and central nervous system. I cannot help it, but my name appears as co-author on all twenty-four of them. What is the explanation? And what was my contribution which resulted in this embarrassing fact, that I was an eye-witness in all of them? To make use of a metaphor: perhaps it was that I had brought with me a key that would open the doors. Dale and Gaddum seemed to know what lay behind them, but I had the key. So I was asked to open first this one, then that one and so on, one after the other; and I never refused to do so. And what was the ‘key’ that I had brought with me? It was my method of detecting acetylcholine released during nerve stimulation. It was the use of eserine, its intravenous injection or, in perfusion experiments, its addition to the perfusion fluid in order to inhibit the enzymic destruction of the released acetylcholine and then to detect it in the venous effluent, blood or perfusate, by means of a simple, specific and sensitive test, the eserinized leech muscle preparation. In the historical abstract to my demonstration on acetylcholine release from the perfused sympathetic ganglion, I have told the story of how it came about that this preparation became our routine test for assaying minute amounts of acetylcholine. With this ‘key’ we obtained the first direct experimental evidence for the role of acetylcholine both in ganglionic and neuromuscular transmission. The second evidence followed in quick succession. It, too, was based on the effect of eserine to inhibit the enzymic destruction of acetylcholine. Eserine should affect the response to nerve stimulation, if due to acetylcholine, in a way consistent with persistence of undestroyed

186

10

Biological Activities of the Alkaloids of the Calabar Bean

acetylcholine at the site of release. Eserine had been shown before to affect the reponse to nerve stimulation in this way but it had not been used so persistently for this purpose as we now used it, in order to obtain evidence for the role of acetylcholine in transmission processes. During these experiments it became evident that the effect need not necessarily be a potentiation of the response; eserine could have the opposite effect and produce attenuation due to a paralysing action of excess acetylcholine, for instance in the sympathetic ganglion.7

However, without any doubt, “But for physostigmine’s ability to prevent the destruction of acetylcholine it would have been difficult to prove that the latter is liberated by nerve action under normal conditions” (Dragstedt 1945; Anon 1970) and, whatever may be the merits for sharing the honours between Loewi, Feldberg and Fühner (vide supra){furthermore, it has also been noted (Henry 1949) that “The explanation of the action of physostigmine began with the discovery of the parasympathetic action of acetylcholine by Hunt and Taveau [1910]” and that “Feldberg and Gaddum [1934] proved that acetylcholine was involved as the transmitter of the nervous impulse in the autonomic ganglia”}, it is also without any doubt that the uses of l-physostigmine as described above by these three investigators played an indispensable role in the discovery of the mechanism of neurohumoral transmission involving Ach which ultimately led in 1936 to Dale (Feldburg 1969) and Loewi being jointly awarded the Nobel Prize in Medicine and Physiology [(Dale 1958; Holmstedt 1972; Loewi 1960  for further references to another “able, exhaustive and critical account” of this work, see (Fraser 1938a) – see also Vincent 1938], “For their discoveries relating to the chemical transmission of nerve impulses” [(Sourkes 1966(a)].

10.4

AntiAchE Activities of the Minor Alkaloids of the Calabar Bean – The l-physostigmine Pharmacophore

Since the physiological substrate of AchE, Ach (180) [and many of its structural analogues], and one of the enzyme’s inhibitors, neostigmine (188, R1¼R2¼Me), are both quaternary esters, it was supposed (Goldstein et al. 1974; Nachmansohn and Wilson 1951; Wilson and Bergmann 1950) that they both form an ionic bond by coulombic attraction (Wilson 1960) with the anionic site in the active centre of the enzyme (Fig. 10.2) and that the activity of such an inhibitor should be independent of pH. The broad pH optimum of the enzyme permitted the experimental verification (Wilson 1960; Wilson and Bergmann 1950) of this latter supposition, whereas inhibition by l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) – also a structural analogue of the substrate but a tertiary base – was shown to be greatest on the acidic side of its pKa, when it would be practically totally protonated, and fell toward zero on the alkaline side when almost all of the drug would be present as the free base (Goldstein et al. 1974; Nachmansohn and Wilson 1951; Wilson 1960; Wilson and Bergmann 1950). However, the proposal, suggested independently by two groups (Goldstein 1951; Goldstein et al. 1974; Pomponi et al. 1992) [and

10.4

AntiAchE Activities of the Minor Alkaloids of the Calabar Bean –. . .

187

apparently perpetuated by others (Triggle et al. 1998) when they depicted the tetrahedral intermediate (Fig. 10.1) formed by the interaction of physostigmine with AchE as (191)] that it is the N(8)-protonated species of the l-physostigmine, Me Me

N Me

N O C N

O NHMe

O

191 192 (now showing absolute configuration), that is the pharmacophore is clearly Me

MeNHCOO

N N H Me H Me

192 erroneous since it is well-established (Sect. 2.6) that the ring system of l-physostigmine undergoes protonation at N(1), the aliphatic and therefore the most basic of its nitrogen atoms, to afford 125 which, as a result of comparative studies on the effect of pH on the inhibition of AchE by l-physostigmine and neostigmine bromide (188, R1¼R2¼Me, X¼Br) (Wilson and Bergmann 1950), has previously (Nachmansohn and Wilson 1951, Wilson 1960) been suggested as the pharmacophore. As a result, the distance between the N⨁ and the C¼O group in 125 would be greater than it is in Ach (180) and neostigmine bromide (188, R1¼R2¼Me, X¼Br) – and pyridostigmine (189, R1¼R2¼Me), and interaction at the AchE receptor site by the protonated alkaloidal inhibitor would therefore be impeded. However, this would be obviated if the interatomic parities were restored – as a consequence of ring-C cleavage in l-physostigmine as shown in 125 to afford, as the pharmacophore, the 3H–indolium cation 126 (R1¼Me, R2¼⨁NH2Me) (Robinson and Robinson 1968). In fact, this concept had first been mooted some 2 years earlier along with recognition (Ahmed 1966) that “the distance between N(a) and the carbonyl group in physostigmine happens to be the same as the maximum possible distance between the active centres in acetylcholine” and consideration (Ahmed 1966) of the relative antiAchE activities of l-physostigmine, l-physovenine and l-eseramine (vide infra) (Ainscow et al. 1964) and the relative eases of their ring C cleavages to afford the corresponding 3H– indolium cations, but with the proviso (Ahmed 1966) that “This assumes that the

188

10

Biological Activities of the Alkaloids of the Calabar Bean

receptor sites in the acetylcholinesterase are rigidly fixed a constant distance apart, which may not be the case {see ref. [(Solter 1965)] and references therein}”. It has been shown, on the basis of uv and 1H–nmr spectroscopic studies, that 126 1 (R ¼Me, R2¼⨁NH2Me) is formed from l-physostigmine at low pH (Sect. 2.6). In practice, this conversion into the ring-C opened 3H–indoluim cation 126 (R1¼Me, R2¼⨁NH2Me) begins in approximately 1 M–hydrochloric acid and is essentially complete in 5 M–acid (Jackson and Smith 1964) which, it was recognised, involve pH values “too low to be considered applicable to a biological system” (Robinson and Robinson 1968) although it was suggested that “the possibility that analogous enzyme-catalysed reactions occur at the acetylcholinesterase surface is not precluded”. Evidence in support of this hypothesis appears to be forthcoming (Robinson and Robinson 1968) from a comparison of the in vitro antiAchE activities (using erythrocyte AchE) of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1), l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1),8 l-N(8)norphysostigmine (98, R1¼MeNHCO, R2¼H, X¼NMe, n¼1), l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1) (see footnote 8) and l-geneserine (110, R¼MeNHCO)]9 and (see footnote 7) when the first three of these alkaloids exhibited comparable activities and the last two were inactive (Robinson and Robinson 1968), results in accord with those from earlier investigations (see footnote 8 and 9). The three active alkaloids were competitive inhibitors after 1 m preincubation with the enzyme prior to the addition of the substrate, mixed (namely competitive and non-competitive) inhibitors after a 3 m preincubation period and non-competitive inhibitors after a 10 m preincubation period, observations which may be explained by the formation of a carbamylated enzyme by transfer of a carbamyl group from the inhibitor (Main and Hastings 1966; Triggle et al. 1998; Wilson et al. 1960; Wilson 1963) (Fig. 10.1) (Dale and Robinson 1970; Robinson and Robinson 1968). Whereas protonation of N(1) could account for the high antiAchE activities of l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) and l-N(8)-norphysostigmine (98, R1¼MeNHCO, R2¼H, X¼NMe, n¼1) and the inactivity (under the then experimental conditions) of l-eseramine (98, R1¼MeNHCO, R2¼Me, X¼NCONHMe, n¼1) and l-geneserine (110, R¼MeNHCO), in both of which the atoms corresponding to N(1), namely that which is protonated, in l-physostigmine and l-N(8)-norphysostigmine, are non-basic, it does not account for the high inhibitory activity shown by l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1) in which the basic N(1)-Me group of l-physostigmine is replaced by a relatively non-basic oxygen atom. It has been suggested “that interaction with the enzyme at this point is mediated through a H-bond” (Yu et al. 1991) “rather than via electrostatic attraction” (Pei et al. 1999). However (Robinson and Robinson 1968), the high antiAchE activity of l-physovenine might be the result of its biotransformation into the pharmacophore (126, R1¼Me, R2¼OH), a species that is known (Robinson 1964a) to be produced by lysis of its ring-C (analogous to that described above for l-physostigmine) in 11 N–hydrochloric acid. Since l-N(8)-norphysostigmine in 8 N– hydrochloric acid is likewise converted into (126, R1¼H, R2¼⨁NH2Me) (Longmore

10.4

AntiAchE Activities of the Minor Alkaloids of the Calabar Bean –. . .

189

1966b; Robinson 1965d, 1968) and the two alkaloids devoid of antiAchE activity do not undergo analogous facile ring openings along with formation of the corresponding 3H–indoium cations (Sects. 4.1 and 6.1), it has been postulated (Robinson and Robinson 1968) that the 3H–indolium cations (126, R1¼Me, Me

MeNHCOO

R2

N R1

126 R2¼⨁NH2Me; R1¼Me, R2¼OH; and R1¼H, R2¼⨁NH2Me) are the pharmacophores of the corresponding active three alkaloids, namely l-physostigmine, l-physovenine and l-N(8)-norphysostigmine, respectively, although it was clearly recognised (Robinson and Robinson 1968) that “the pH values at which such cleavages of ring C in l-physostigmine, l-physovenine and l-N(8)norphysostigmine have been effected are too low to be considered applicable to a biological system” but it was suggested (Robinson and Robinson 1968) “that the possibilily that analogous enzyme-catalysed reactions occur at the acetylcholinesterase surface is not precluded”. Both these assertations have been endorsed by the statements, that [following further experimentation, (Dale and Robinson 1970)] “It could be that the difference in activity between the (+)- and the (-)-isomers [of physostigmine and physovenine  in both cases, the d-isomers are far less active antiAchEs than their natural enantiomers (Dale and Robinson 1970)] reflect the preferential enzymic opening of ring C in the (-)-isomers [see (Longmore 1969) in Sect. 2.4], which affords the 3H-indolium cations [126], the biologically active species, at or near the acetylcholinesterase receptor site”, that L [in review (Robinson 1988a, b)] “the 3H-indolium cation [126 (R1¼Me, R2¼ NH2Me)] is the active species of (-)-physostigmine”, that [Neuwinger 1996 – quoting (Dale and Robinson 1970)] “The analogous 3H-indolium cations [126 (R1¼Me, R2¼OH and L 1 2 R ¼H, R ¼ NH2Me)]” were likewise postulated [Longmore and Robinson 1973, Robinson and Robinson 1968] as “being the active forms of physovenine and N-8norphysostigmine,” respectively, and that “Although the pH values at which such cleavages of ring C in these three alkaloids occurs are too low to be considered applicable to a biological system, the possibility that analogous enzyme-catalysed reactions occur at the acetylcholinesterase surface was not precluded” (Robinson and Robinson 1968). Further support for the 3H–indolium cations 126 as the pharmacophores was also forthcoming (Dale and Robinson 1970; Robinson 1971) from the observation that the methiodide of 171 (R1¼R2¼H, R3¼R4¼Me) (Ahmed 1966; Ahmed and Robinson 1965; Kolosov and Preobrazhenskii 1953b) is circa 100 times more active as an antiAchE than is l-physostigmine when measured in vitro using erythrocyte AchE and is independently alluded to (Grandberg et al. 1970) in studies investigating “the reversibility of the process” 98 Ð126 “and the

190

10

Biological Activities of the Alkaloids of the Calabar Bean

effect on it of structural factors, since these could be decisive factors in the physiological activity of compounds of this type” (Grandberg et al. 1970). “l-Physostigmine is a tertiary amine and therefore crosses (Birtley et al. 1966; Lee et al. 1975; Muhtadi and El-Hawary 1989; Wright 1976) the blood-brain barrier (Grieg 1992; Pardridge 1988) and it has been claimed (Pomponi et al. 1992) that “From a structure-activity point of view, the pyrrolodine rings were, actually, the critical pharmacophore aiding in penetration of the blood-brain barrier”, On the other hand, the pharmacophores 126 are quaternary amines and would, therefore – like neostigmine bromide (188, R1¼R2¼Me, X¼Br) and pyridostigmine (189, R1¼R2¼Me) (Muhtadi and El-Hawary 1989; Wright 1976) – be unable to cross the blood-brain barrier, thus giving credence to the suggestion (Robinson and Robinson 1968) that they are formed, as and when required, at the AchE receptor surface. However, opinion against the 3H–indolium cations 126 as the pharmacophores was initially presented when attention was drawn to “the fact that ring-opening [to form the 3H-indolium cations 126]“only occurs at pH 3 or below [Jackson and Smith 1964] which is rather unphysiological” (Brossi 1985) – although this situation had been clearly recognised, and resolved, when the hypothesis was first presented (Robinson and Robinson 1968) (vide supra) – and by the observation (Brossi 1985) that the high analgesic activity of l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) [Sect. 10.11] is not paralleled in the corresponding ring-opened species 193 (which has only low analgetic activity) – although these latter data, when Me

HO

NMe2 N Me

193 apparently taken as invalidating the pharmacophore involved at the AchE receptor, “are abrogated by the qualitative differences between the compared biological activities” (Robinson 1988a, b). Further investigations (Yu et al. 1991) have included repetition and confirmation of the results of the earlier investigation (Robinson and Robinson 1968) – that had initially lent support to the 3H– indolium cation 126 (R1¼Me, R2¼⨁NH2Me) pharmacophore hypothesis (vide supra) – when it showed that l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) and l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1) have “almost identical potencies” in inhibiting in vitro AchE and BchE from human plasma. However, in this instance (Yu et al. 1991) these results were interpreted as showing “that a basic N-atom (N(1)) in the tricyclic molecule need not be present, and that interaction with the enzyme at this point is mediated through a H-bond” (Yu et al. 1991). Further evidence for this conclusion was also forthcoming from the detection of potent antiAchE activity in thiaphysovenine (98, R1¼MeNHCO,

10.4

AntiAchE Activities of the Minor Alkaloids of the Calabar Bean –. . .

191

R2¼Me, X¼S, n¼1) (He et al. 1992), 8-carbaphysostigmine and its analogues which, consequent upon the replacement of the N(8)-Me group in l-physostigmine by a methylene group, contain the ring system 269 (X¼CH2, Y¼NMe) (Chen et al. 1992) and tetrahydrofurobenzofuran analogues of l-physostigmine which contain the ring system 269 (X¼Y¼O) resulting from the replacement of the basic N(1)-Me and N(8)-Me groups in l-physostigmine by non-basic oxygen atoms (Luo et al. 2005a, b; 2006). In these inhibitors, it is likely that the p-electron pair of either the oxygen or sulphur atom is, in common with the amino N-atoms of l-physostigmine and l-physovenine (Luo et al. 2005b) involved in H-bonding with the enzyme via molecular moieties which have been elaborated (Luo et al. 2005b, 2006), thereby validating the earlier conclusion (Yu et al. 1991) that “The high activity of [lphysovenine] against AchE and BchE is in good support of the view that the central effects observed with these compounds are caused by tricyclic molecules and not ring-open indolium species”. However, a less categorical conclusion “suggesting that optimal potency may possibly require a tricyclic, rather than a bicyclic template” had been reached following the observation (Brossi et al. 1986), using AchE from electric eel and acetylthiocholine as the substrate, that the dihydroseco compound 194 (R1¼Me, R2¼H) derived from l-physostigmine (98, R1¼MeNHCO, R2¼Me, Me

R1NHCOO

NR2Me

N Me

194 X¼NMe, n¼1), is “about 25-times less potent” than the parent alkaloid. Furthermore, when later (Pei et al. 1999) in vitro investigations using human erythrocyte AchE and plasma BchE found that the products 194 (R1¼Me, Ph and 4-Me2CC6H4, R2¼H and Me) – unfortunately, in none of these products is the N(1) quaternary – and their parent compounds l-physostigmine, l-phenserine and l-cymserine (98, R1¼MeNHCO, PhNHCO, and 4Me2CHC6H4NHCO, respectively, R2¼Me, X¼NMe, n¼1) all possess potent in vitro inhibitory activity, it was concluded (Pei et al. 1999) that this “provides direct evidence that the tricyclic structure of physostigmine is not an essential component for inhibiting either AchE or BchE”, a decision that is not at variance with the original hypothesis (Robinson and Robinson 1968) (vide supra) that, when it is acting as an antiAchE, l-physostigmine (98, R1¼MeNHCO), R2¼Me, X¼NMe, n¼1) is converted as required, by an enzyme-catalysed reaction at the AchE receptor site into the pharmacophore 126 (R1¼Me, R2¼⨁NH2Me) (Robinson and Robinson 1968). Further experimental support for this hypothesis may reside in the metabolic pool of l-physostigmine, with synthetic products already being available (Yu et al. 1999 – quoted in Sect. 10.2) that “will prove of value in characterising potential ring-C open metabolites in phase I clinical trials on phenserine” (Sect. 10.2).

192

10

Biological Activities of the Alkaloids of the Calabar Bean

However, “A better insight into the molecular interactions of the enzyme and inhibitor can be obtained at the transition-state. Whereas it is not currently possible to produce a crystal structure of an enzyme complexed with its substrate in the transition-state [Figs. 10.3 and 10.4], due to the short transition-state lifetime versus the time required for X-ray data collection, one can exploit the high affinity of transition-state analogues” (Luo et al. 2006), such as these with inhibitors (Fig. 10.1). In view of the apparently conflicting results obtained between the antiAchE activities of compounds 98 (R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) and 194 (R1¼Me, R2¼H) (Brossi et al. 1986; Pei et al. 1999) (vide supra), it might appear that it is possibly the nature of the experimental conditions, including the source of the enzyme, which determines whether or not it is the ring-C opened 3H–indolium cation 126 (R1¼Me, R2¼⨁NH2Me) that is the inhibitory moiety which interacts with the AchE receptor site. Indeed, such a concept seems to underline the statement that “The wide structural diversity of available ChEIs suggests that unlike types of inhibitors bind with AChE in different ways via disparate yet specific interactions between compound and enzyme” [(Luo et al. 2006) and refs 27–29 therein quoted].

10.5

Clinical use of l-physostigmine in Ophthalmology – Miotic Activity and the Reduction of Intraocular Pressure in Glaucoma

This subject has already been reviewed on several occasions (Anon 1970; Holmstedt 1972; Robinson 1988a; b; Rodin 1947; Triggle et al. 1998 – see also Moroi and Lichter 1996), during the course of which it was noted (Rodin 1947) that “It is interesting that [during his self-experimentation with the Calabar bean (see footnote 14 of Chap. 1)], while Christison was under the effect of the drug, he did not notice its action on the eye”. The first therapeutic utility of extracts of the Calabar bean was discovered independently by:-. 1. Douglas Moray Cooper Lamb Argyll Robertson (Robertson 1863), an ophthalmic surgeon at the Edinburgh Royal Infirmary (see footnote 14 of Chap. 1), who, amongst his several publications dealing with eye diseases included the pupillary condition which became generally known as the “Argyll Robertson phenomenon” and, being a sign of neurosyphilis, led to the suggestion by Sir Anderson Citchett “that it was far better to be an Argyll Robertson pupil than to have one” (Holmstedt 1972; Mackay 1909), and who also, as early as 1863, had been alerted to the use of the Calabar bean in ophthalmic surgery (Robertson 1863) by his friend, Thomas Richard Fraser (Witkop 1998), 2. the said Fraser (see footnote 17 of Chap. 1) in papers (Fraser 1863) – based upon his MD thesis submitted in 1862 to the University of Edinburgh (Fraser

10.5

Clinical use of l-physostigmine in Ophthalmology – Miotic. . .

193

1862) – in which he described the action of physostigmine (in the form of a crude extract of the Calabar bean) on his own eye as follows (Fraser 1863):“A small drop of a syrupy extract was placed on the point of a thin probe, and applied to the conjunctiva over the left eye-ball. A copious discharge of tears immediately occurred. In five minutes, the left pupil was a little contracted, and very evidently so in eight minutes, the left being one-half the size of the right. In ten minutes, the left pupil was the one-sixteenth of an inch in diameter. Vision with this eye was imperfect, the visual distance being lessened, but the iris was mobile. A slightly painful sensation was now experienced in the supraorbital region of the left side, and a sensation of heat in the left eyeball. In thirty minutes, no change had occurred in the right pupil; the left was a mere speck. Vision with the left eye was almost lost; there was a little redness, and tenderness on exposure to the light. In one hour and a-half, all disagreeable sensations had gone, the dimness of vision was less marked, but the extreme contraction of the pupil continued. In four hours, the dimness of vision disappeared; but the contraction of the left pupil continued unchanged for twenty-four hours. It gradually diminished after this, but very slowly, as the symptom continued for five days,”

3. Albrecht von Graefe (von Graefe 1863) of Berlin, the then doyen of ophthalmic surgery and medicine. It was probably because of the latter’s seniority and his firm belief in iridectomy that whereas all three of these investigators recognised the extract’s use as a miotic and suggested its use after atropinisation of the eye (indeed in alternation with a mydriatic such as atropine, l-physostigmine proved to be useful for the breaking of adhesions between the iris and the lens or cornea (Laqueur 1876; von Graefe 1863)) they did not realise its potential for use in the treatment of glaucoma. It has been suggested that Ludwig Laqueur was (in 1876) the first (Anon 1970; Neuwinger 1996; Taylor 1996) ophthalmologist – although it appears (Anon 1970) that he shared the honours with Adolph Weber (1876, 1877) – to use l-physostigmine, the major alkaloidal component of the Calabar bean (Chap. 2), therapeutically in the treatment of glaucoma (Holmstedt 1972; Rodin 1947). “He noticed that it lowered intraocular pressure temporarily or permanently” [which is, of course, accompanied by constriction of the eye’s pupil], as the result of increased drainage of intraocular fluid (Cannon 1981) through the canal of Schlemm (Moroi and Lichter 1996) and “From that time on, eserine has been used more and more extensively, and [now] occupies a prominent place in the treatment of glaucoma” (Holmstedt 1972). Although for a time it came to be replaced by certain organophosphorus compounds (see footnote 1), it was found that the use of these may produce cataracts and consequently many ophthalmologists reverted to the use of lphysostigmine (Axelsson 1969; Holmstedt 1972) which thus became wellestablished in ophthalmology (Axelsson 1969). However, in this role a current alternative to the antiAchE l-physostigmine is now the cholinomimetic alkaloid pilocarpine (195) [Anon 1970; Battersby and Openshaw 1953; Hill and Barcza Me

Et O

N O

N

195

194

10

Biological Activities of the Alkaloids of the Calabar Bean

1966; Merck Index 2001(j); Taylor 1963; van Rossum et al. 1960], although “there is no consensus on medical therapy for glaucoma” (Moroi and Lichter 1996) which has also been embraced by the therapeutically-versatile prostaglandins (Camras et al. 1989).

10.6

Use of l-physostigmine in the Treatment of: -

10.6.1 Myasthenia Gravis It has been claimed (Drachman 1994) that “Myasthenia gravis is undoubtedly the most thoroughly understood of all human autoimmune diseases and has served as a model for the elucidation of mechanisms underlying other autoimmune disorders”. Accounts of the development of the study of this disease, “the basic abnormality” of which “is a decrease in the number of acetylcholine receptors at the neuromuscular junction” (Drachman 1994), have already been presented (Drachman 1994; Fisher 2002; Holmstedt 1972; Robinson 1988a, b; Taylor 1996). It is “a disorder that was first described clinically 300 years earlier [than 1972] by the great physiologist Thomas Willis” (Willis 1672) and “is not rare with a prevalence of 50 to 125 cases per million population, or approximately 25,000 affected persons in the United States. The incidence is age- and sex-related, with one peak in the second and third decades affecting mostly women and a peak in the sixth and seventh decades affecting mostly men” (Drachman 1994) [see also (Fisher 2002)]. It has been claimed (Holmstedt 1972; Robinson 1988a, b) than it was initially studied by Guillermo Erb and new observations upon it were presented in 1891 by Goldflam, in recognition of which it has been called Erb-Goldflam’s disease (see Goni 1946). Subsequently, Frederick Jolly, (1844–1904) in 1895 published the results of his investigations into muscular reaction to electrical stimulus and obtained a myasthenia reaction which led to the disease being known as pseudo-paralytic myasthenia gravis, which is actually the name mostly used. Moreover, since it was known that physostigmine has a pharmacological action the opposite of the myasthenia reaction, known as the myostonic action, Jolly suggested, in the presentation of his results in a lecture to the Berlin Medical Society on 5th December 1894 (Jolly 1895a), the use of physostigmine in the treatment of the disease (Jolly 1895a, b). However, no reference was made to Jolly’s concept in the first reported treatment of Myasthenia gravis with physostigmine, namely that by Mary Walker, a graduate in medicine in 1913 from the University of Edinburgh – an establishment that had already been associated (see footnote 14, 16, 17 of Chap. 1)10 with many of the pioneering investigations into the Calabar bean and its major alkaloidal component – but who was then working as a Senior Hospital Medical Officer at St Alfege’s Hospital in London, Her report of the case (Walker 1934) begins “The abnormal fatiguability in myasthenia gravis has been thought to be due to curare-like poisoning of the motor nerve-endings [a suggestion that was apparently made to her by that prominent neurologist Derek Denny-Brown during a visit by him to St Alfege’s

10.6

Use of l-physostigmine in the Treatment of: -

195

(Fisher 2002)] or of the “myoneural junctions” in the affected muscles. It occurred to me recently that it would be worth while to try the effect of physostigmine, a partial antagonist to curare, on a case of myasthenia gravis at present in St. Alfege’s Hospital in the hope that it would counteract the effect of the unknown substance which might be exerting a curare-like effect on the myoneural junctions. I found that hypodermic injections of physostigmine salicylate did have a striking though temporary effect”. Indeed, the treatment was attended with such remarkable success that it has been referred to as “the Miracle of St. Alfege’s” (Viets 1965). However, this was accompanied by the secondary reaction of parasympathetic stimulation and in an attempt to circumvent this, on 16th June 1934 she replaced the injection of physostigmine salicylate by a successful one of Prostigmine (188, R1¼R2¼Me, X¼MeSO4) and reported this second case before the Royal Society of Medicine on 8th February 1935 (Walker 1935) [In 1935 she was also awarded the gold medal of medicine from the University of Edinburgh for her work on myasthenia and in 1955 she became an honorary member of the Medical Advisory Board of the Myasthenia Gravis Foundation (Holmstedt 1972, where is also published some further biographical data on Mary Walker)]. It had, indeed, been some 3 years earlier and a few months after neostigmine (188, R1¼R2¼Me) first became commercially available as a result of the work of Aeschlimann and Reinert (1931) at ROCHE (Hoffmann-La Roche) (Sect. 10.1.1) that “an account was published of its use in a patient with Myasthenia gravis by Lazar Remen [(Remen 1932)], then attached to Kehrer’s clinic at the University of Münster in Westphalia” (Viets 1965), but his short note on this subject remained, until much later (Viets 1965) [see also (Fisher 2002; Holmstedt 1972; Neuwinger 1996)], entirely unrecognised. It certainly appears that Mary Walker was unaware of Remen’s work but such a fate was not to be shared by her own observations (Viets 1965), with able support from her mentor at St Alfege’s, Philip Hamill (Hamill 1935). “This opened the floodgates, using Prostigmine, and confirmatory observations filled the medical literature for months to come. First was Blake Pritchard (1935) at the Hospital for Epilepsy and Paralysis, Maida Vale, with seven cases followed by Laurent (1935) who matched Blake Pritchard at the University College Hospital [the alma mater in earlier years of Sydney Ringer (see footnote 6)] and brought the total cases up to sixteen, without a single failure. Success could no longer be denied and the ‘miracle at St. Alfege’s’ became an integral part of medicine” (Viets 1965), together with the introduction of the use of Prostigmine in a related diagnostic test (Viets 1965; Viets and Schwab 1935). As of some decade ago (Fisher 2002) and despite other therapies, cholinesterase inhibitors still retained an important role in the management of Myasthenia gravis, “and remain crucial to symptomatic treatment” (Yu et al. 2010), and in this, the neostigmine (188, R1¼R2¼Me) of Lazar Remen and Prostigmine (188, R1¼R2¼Me, X¼MeSO4) of Mary Walker et al. (vide supra) had been joined by two further monoquaternaries, namely pyridostigmine (189, R1¼R2¼Me) and edrophonium, and the bisquaternary ambenonium. However, “erratic absorption after oral administration” and “A short duration of pharmacodynamic action, usually 30-120 min and allied with rapid metabolism, has made the routine clinical use of

196

10

Biological Activities of the Alkaloids of the Calabar Bean

these compounds [Prostigmine (188, R1¼R4¼Me, X¼MeSO4) and pyridostigmine bromide (189, R1¼R2¼Me, X¼Br)] troublesome. This has resulted in necessary multiple dosing (10 doses over 24h) as well as the development of slow-release formulations to achieve efficacy” (Yu et al. 2010). An attempt to circumvent this problem followed the observations that l-phenserine, l-tolserine, l-cymserine (98, R1¼PhNHCO, 2MeCHC6H4NHCO and 4-Me2CHC6H4NHCO, respectively, R2¼Me, X¼NMe, n¼1) (Sect. 10.7.2) and l-phenylethylcymserine [98, R1¼4Me2CHC6H4NHCO, R2¼Me, X¼N(CH2)2Ph, n¼1] all “possess far longer durations of in vivo cholinesterase inhibitory action compared to not only their methyl carbamate analogue, l-physostigmine [98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1], but also to neostigmine and pyridostigmine” (Yu et al. 2010). Thus, the phenyl carbamate analogues of Prostigmine (188, R1¼Ph, R2¼H, X¼MeSO4) and of pyridostigmine bromide (189, R1¼Ph, R2¼H, X¼Br), as well as the methyl quaternary analogues of l-physostigmine, l-phenserine, l-tolserine, l-cymserine (98, R1¼MeNHCO, PhNHCO, 2MeC6H4NHCO and 4-Me2CHC6H4NHCO, respectively, R2¼Me, X¼⨁NMe2 BrƟ, n¼1) and l-phenylethylcymserine [98, R1¼4Me2CHC6H4NHCO, R2¼Me, X¼⨁N(Me)(CH2)2Ph BrƟ, n¼1] were synthesised and their pharmacological utility was assessed. It was found that their antichE inhibition against human enzyme ex vivo demonstrated that whereas the first two compounds possessed only marginal activity, the remaining five proved to be potent antichEs and an extended duration of chE inhibition was determined in rodent, “making them of potential interest as long-acting agents for myasthenia gravis and congenital myasthenic syndromes” (Yu et al. 2010).

10.6.2 Paraplegic Anejaculation “Especially young people, almost 80% of them male, are victims of a traumatic cross-lesion. Next to the motoric and sensory deficiencies, their sexual and reproductive live [sic] comes to a sudden end often at a time when the wanted family size hasn’t been reached yet”. (Blockmans and Steeno 1988). “The primary causes of this low fertility rate are ejactulatory dysfunction and poor semen quality” (Leduc et al. 1992 – see also Griffith et al. 1973; Linsenmeyer and Perkash 1991; Tarabulcy 1972; Yarkony 1990) and the various sperm recovery methods which have been described to assist infertile men with spinal cord injury include vibratory stimulation of the penis, electroejaculation, sperm aspiration and pharmacological stimulation (Bindley 1984; Leduc et al. 1992; Linsenmeyer and Perkash 1991; Rawicki and Hill 1991; Rawicki and Lording 1988), all such studies on potency and fertility in the spinal man being consequent upon the greatly increased survival rate during the two decades up to 1971 of patients with spinal cord injuries (Guttmann and Walsh 1971). The pioneering therapeutic use of neostigmine (188, R1¼R2¼Me) that resulted from the work of Aeschlimann and Reinert (1931) at ROCHE (Hoffmann-La Roche) (Sect. 10.1.1) in the treatment of Myasthenia gravis (Remen 1932) has already been

10.6

Use of l-physostigmine in the Treatment of: -

197

referred to (Sect. 10.6.1). The same drug also appeared at the forefront of medical practice when it was introduced for administration via intrathecal injection by Ludwig Guttmann (1949) as a new way to treat paraplegic anejaculation. A similar clinical result was later reported (Guttmann and Walsh 1971; Tarabulcy 1972) by injecting Prostigmine (188, R1¼R2¼Me, X¼MeSO4) intrathecally. However, being quaternary amines, neither neostigmine nor Prostigmine can cross the bloodbrain barrier and must, therefore, be administered in situ by lumbar puncture to make use of their central effects to produce ejaculation (Blockmans and Steeno 1988; Chapelle et al. 1983). The search for a less invasive route to administration (Chapelle et al. 1974) has led to a return to the parent alkaloid in the use of a subcutaneous injection, of a minimal dose of 1 mg in a 70 kg man (Chapelle et al. 1983) of physostigmine – which crosses the blood-brain barrier (Birtley et al. 1996; Greig 1992; Lee et al. 1975; Muhtadi and El Hawary 1989; Pardridge 1988; Wright 1976) and “is said to increase the concentration of acetylcholine in the sympathetic ganglion after the pre-ganglionic fibres have left the spinal cord at the emission center (T10-L2)” (Leduc et al. 1992) although “hasty conclusions should not be drawn as to the neuropharmacology of ejaculation” (Chapelle et al. 1983) – as an alternative to evoke an ejaculation (Blockmans and Steeno 1988; Chapelle et al. 1983; Leduc et al. 1988, 1992; Linsenmeyer and Perkash 1991; Rawicki and Hill 1991; Yarkony 1990). When obtained in this manner, spermatozoa was present in the ejaculate (Chapelle et al. 1983) and artificial insemination of the wife with a sample of the sperm obtained ultimately led to the spontaneous birth of a healthy baby (Blockmans and Steeno 1988). In other instances “There have been 8 pregnancies from 6 couples” (Rawicki and Hill 1991) and insemination was effected in six couples and of three normal births, one resulted after fifteen inseminations (six performed in hospital, nine at home) and “in the case of two other couples, no more than 3 inseminations were performed at the outpatient clinic to achieve pregnancy in spite of the decreased sperm mobility” (Leduc et al. 1992). In another use of subcutaneous physostigmine masturbation ejaculation was effected in seventy-five out of one hundred and thirty-five patients, fifteen of whom fathered children (Chapelle et al. 1988). Thus, the major toxic component of the Calabar bean which was responsiblie for so much death caused by the latter’s use in the ghastly poison ordeal trials by the Efiks of Old Calabar (Chap. 1) has, through therapeutical development and application, played a crucial role in the creation of life, a role that is clearly worthy of further investigation and development.

198

10.7

10

Biological Activities of the Alkaloids of the Calabar Bean

Use of l-physostigmine in the:-

10.7.1 Prophylactic Protection Against Intoxication by Organophosphates (Including “Nerve Gases”) In what prima facie might appear to be a paradoxical phenomenon, an initial non-lethal dose of l-physostigmine (the peripheral autonomic effects of this, thereby enabling the injection of larger doses and offering a further degree of protection, can be prevented by prior administration of a small dose of atropine), a reversible antiAchE, when administered to cats was found to afford protection against several otherwise lethal doses of diisopropyl fluorophosphate (DFP) (270, R1¼R2¼Me2CHO, X¼O, Y¼F), an irreversible antiAchE (see footnote 1) (Koster 1946). In a probable explanation for this phenomenon, it has been proposed that “protection against one anticholinesterase agent by another is suggestive of the antagonistic effects which have been demonstrated between components similar in structure” and that “It is conceivable that the protective action exerted by eserine when injected prior to DFP might result from the reversible combination of eserine with the active groups of the cholinesterase molecules [the hydroxyl group of the serinyl 198 moiety (Sect 10.1.1 and Fig. 10.3) is a prime candidate (see footnote 1)], thereby blocking access to DFP and the subsequent formation of an irreversible complex” (Koelle 1946) (see footnote 1). Since vital organs contain far more AchE than is necessary for normal functioning, the inhibition of a proportion of this enzyme in such organs would prevent complete phosphorylation by the DFP and, because the inhibition is reversible, free enzyme would gradually be regenerated. If, in parallel with this, hydrolytic metabolism and excretion of excess uncombined DFP occurred, the liberated AchE could then resume its physiological function (Berry and Davies 1970; Deshpande et al. 1986; Koelle 1946) – see also (Albuquerque et al. 1985). In accord with this hypothesis, of nineteen other antiAchEs that were similarly evaluated, only two exhibited a marked protection, namely Prostigmine (188, R1¼R2¼Me, X¼MeSO4) and carbamylcholine, the only ones in the group that would have been able to protectively reversibly carbamylate the enzyme (Koelle 1946). In view of this, it is perhaps not surprising that protection of animals (guinea pigs) against poisoning by organoposphates [including Soman (272)] by pretreatment with carbamates (Sect. 10.8) has been observed (Berry and Davies 1970; Gordon et al. 1978) although prima facie to the contrary under similar conditions an effective prophylaxis using bisquaternary pyridinium salts was observed against poisoning by Soman (272) of female mice (Schoene et al. 1976). Ultimately protection against poisoning by Soman (272) has been observed using pyridostigmine (189, R1¼R2¼Me) (Deyi et al. 1981; Dirnhuber et al. 1979; Tuovinen et al. 1999). It is interesting that during the so-called “Operation Desert Storm”, 41,650 American soldiers were provided, for oral self-administration of 30 mg every 8 h “while under the threat of nerve agent attack (for 1 to 7 days)”, with

10.7

Use of l-physostigmine in the:-

199

“pyridostigmine bromide” (pyridostigmine) (189, R1¼R2¼Me, X¼Br). This action was taken since “Intelligence reports indicated that the Iraqi chemical arsenal contained nerve, vesicant and blood agents” (Keeler et al. 1991). However, since as would be expected of a quaternary amine (Greig 1992; Pardridge 1988), “pyridostigmine iodide” (189, R1¼R2¼Me, X¼I) should not cross the blood-brain barrier (Birtley et al. 1966; Taylor 1996) and therefore has very limited access to the central nervous system (Deshpande et al. 1986), it is probably fortunate that the truth behind the above (so-called) “intelligence reports” left much to be desired although such falsehoods were still used by the USA and UK and their conspirators in an attempt to justify their invasion of Iraq – such is the murky nature of politics and its practitioners!

10.7.2 Enhancement of Cognition and Memory (Including Antiamnesic Activity in Dementia of the Alzheimer’s Type – Alzheimer’s Disease) As described by Alois Alzheimer in 1906, “classical signs of Alzheimer’s disease are brain atrophy and neuronal cell loss, deposition of amyloid plaques in the brain extracellular space, formation of intracellular neurofibrillary tangles and, of course, the clinical symptoms of dementia (i.e., loss of attention, memory, and intellectual and cognitive processes)” (Klein 2007). It is currently preferred “not to distinguish between Alzheimer’s disease and senile dementia so that the term dementia of the Alzheimer type (DAT) is often used” (Marchbanks 1982). There is evidence that aluminium is associated with the disease and that a genetic component is present in its occurrence (Marchbanks 1982). It “is the most common form of dementia, accounting for approximately 70% of dementia cases in most industrialized countries” (Greig et al. 2005a). Worldwide, the disease is estimated to affect more than 15 million individuals (Greig et al. 2005a; Luo et al. 2006). In western societies it is “the fourth leading cause of death” (Greig et al. 1995a; Marta et al. 1988), “in the United States alone, there were more than 4.5 million afflicted in 2000 – a number anticipated to increase to over 13 million by 2050” (Greig et al. 2005a). “The mean survival time from diagnosis is approximately 8 years” (Greig et al. 1995a). “Patients....experience increasing deterioration as the disease progresses, resulting in memory impairments, behavioral abnormalities and increasing difficulty in maintaining independence and performing activities of daily living. Eventually, all patients become dependent on the care of others” (Greig et al. 2005a). Thus, “it has become evident that its incidence and the necessity for nursing care have important economic consequences” (Marchbanks 1982). The disease is characterised neurochemically by a reduced activity of the enzyme choline acetyltransferase with a consequent decrease of the Ach concentration in the brain areas associated with memory and learning (Bartus et al. 1982; Marchbanks 1982;Marta et al. 1988 – see also Becker and Giacobini (eds) 1991; Luo et al.

200

10

Biological Activities of the Alkaloids of the Calabar Bean

2006) and it is thereby a further example of a neurological disorder associated with an irregularity in cholinergic transmission characterised by a cholinergic deficit in which augmentation of cholinergic activity may prove to be of value. “The cholinergic forebrain deficit found early in the Alzheimer brain (Marchbanks 1982; Whitehouse et al. 1982), together with the critical role of cholinergic function in memory and learning (Drachman and Leavitt 1974) spurred the use of physostigmine as the first anticholinesterase [a prerequisite for the cholinergic action of such agents being “that sufficient residual cholinergic function is present in patients’ brains” (Klein 2007 – see also Marchbanks 1982)] in the treatment of Alzheimer’s disease” (Flippen-Anderson et al. 2002). Thus, it is perhaps not surprising that, although there have been reports to the contrary (Caltagirone et al. 1982; Jotkowitz 1983), physostigmine has been found to enhance the impaired memory processes in Alzheimer’s disease (Davidson et al. 1986; Davis and Mohs 1982; Davis et al. 1978, 1979, 1983; Greig et al. 1995a; Gustafson et al. 1987; Hartvig et al. 1986; Levin and Peters 1984; Merck Index 2001(h); Mohs et al. 1985; Peters and Levin 1979; Stern et al. 1987; Thal 1991; Thal and Fuld 1983; Thal et al. 1983) – it is also of interest that it improves retentive abilities in both experimental animals and humans (Marchbanks 1982), the long-term memory processes in normal humans [Davis et al. 1978, 1979, 1983; Merck Index 2001(h)] and the memory and cognitive functioning of normal elderly subjects (Drachman and Sahakian 1980) and is useful in the treatment of memory loss following traumatic brain injury (Cardenas et al. 1994). However, physostigmine is somewhat unstable (Sect. 2.8.2), non-selectively inhibits in vitro both human erythrocyte AchE and human plasma BchE (Brossi et al. 1996), has serious, potentially lethal side effects {the subject of some deliberation (Caine 1979) – references to studies on toxicity data relating to l-physostigmine can be found in [Brossi et al. 1996; Merck Index 2001(h)]} and is characterised by a short half-life, short duration of action, poor bioavailability, and a narrow therapeutic window (Brossi et al. 1996; Brufani et al. 1986; Iversen et al. 1991; Marta et al. 1988; Perola et al. 1997 – see also Greig et al. 1995a). Furthermore, the “results of therapeutic trials with physostigmine in the treatment of Alzheimer’s disease (AD) have been inconsistent [see also (Greig et al. 1995a)] and controversy persists concerning safety and efficacy” (Sano et al. 1993), although investigation of these problems led to the suggestion that “oral physostigmine is safe and may improve memory in AD” (Sano et al. 1993). Nevertheless, in order to produce more stable and selective carbamylated antiAchEs, and compounds able to cross the blood brain barrier (Birtley et al. 1996; Greig 1992; Lee et al. 1975; Muhtadi and El-Hawary 1989; Pardridge 1988; Wright 1976) and with improved selectivity for the inhibition of AchE alone, analogues of l-physostigmine have been investigated in which, for example, the lipophilicity has been increased by appropriate replacement of the methyl group of the N-methylcarbamyl moiety. Investigations to this effect have been undertaken at the Catholic and La Sapienza Universities in Rome and at Mediolanum Farmaceutici in Milan, Italy and at the National Institutes of Health in Bethesda, MD, USA (Brossi et al. 1996; Marta et al. 1988; Pomponi et al. 1992; Yu et al. 1988b), and other related investigations have also been effected at Novartis in Basel, Switzerland (Luo et al. 2006) and at Hoechst-Roussel Pharmaceuticals Inc.

10.7

Use of l-physostigmine in the:-

201

in Somerville, NJ, USA (Brossi et al. 1996) [see, for examples (Glamkowski 1989; Glamkowski and Kurys 1989; Glamkowski et al. 1989a; Hamer et al. 1986, 1988) – and ref. 2 quoted in (Lee and Wong 1991)] (for other studies by this group, see Sect. 10.11).Early Italian endeavours to this effect included the synthesis and evaluation of antiAchE activity of 98, (R1¼butyl, t-butyl, phenyl, c-hexyl, heptyl, nonyl, undecyl, and pentadecyl-NHCO, R2¼Me, X¼NMe, n¼1) (Marta et al. 1988, Pomponi et al. 1992) and of which the heptyl analogue, “heptastigmine” 98 [R1¼Me(CH2)6NHCO, R2¼Me, X¼NMe, n¼1) – which was also the subject of related investigations at the National Institutes of Health in the USA (Iijima et al. 1992) –was found (Brufani et al. 1986, 1987; Greig et al. 1995a; Iversen et al. 1991; Pomponi et al. 1990; de Sarno et al. 1989) to be “the drug of choice” (Pomponi 1989) and patents were accordingly secured (Brufani et al. 1985–1986, Pomponi 1989 – see also Brossi 1992). Unfortunately, toxicity has reared its unwelcome head and it has been reported that heptastigmine “produces haematological toxicity” (Brossi et al. 1996), namely it exhibits “reversible hematopoietic toxicity” (Greig et al. 1995a), “causes agranulocytosis with the consequence that its clinical trials have been discontinued” (Witkop 1998) and “was withdrawn from clinical studies because two patients developed aplastic anaemia [the cause of eptastigmine bone marrow toxicity is unknown, but could be related either to the presence of the heptyl chain (physostigmine does not produce this effect) or to haematoxic effects of the eseroline metabolite” (Giacobini 2000b)] and “during the first phase of clinical experimentation, two cases of reversible neutropenia have been observed among the 96 patients treated with this drug” [Mediolanum Farmaceutici, personal communication – quoted in Perola et al. 1997]. “The causes of this phenomenon are not clear. They could be attributed either to the cholinesterase inhibition itself or to the action of some products of metabolic degradation. In both cases they could be related to the presence of the alkyl chain, since physostigmine is free from this effect” (Perola et al. 1997). Indeed, it is of interest that, with regard to the alkylcarbamyl analogues of physostigmine, it has been noted that “By reactivation of carbamylated enzymes, an alkylamine is produced [cf Fig. 10.1], whose biological effect should be accurately tested before these compounds are used in clinical practice” (Pomponi et al. 1992). However, the above-mentioned withdrawal of the lipophilic heptastigmine because of its toxicity – which followed the similar withdrawal (because of its hepatotoxicity) (Greig et al. 1995a; Marx 1987; Muñoz-Ruiz et al. 2005) of the erstwhile promising tacrine (Brossi et al. 1996) from the therapeutic armamentarium for the treatment of Alzheimer’s disease by cholinesterase inhibitors (Brossi et al. 1996; Klein 2007; Pomponi et al. 1990) – did not deter further efforts in this approach using lipophilic and thereby brain-directed antiAchEs for the treatment of Alzheimer’s disease. Thus, phenserine (98, R1¼PhNHCO, R2¼Me, X¼NMe, n¼1), although apparently dismissed by Italian investigators from their statement that “The anticholinesterase action is greatest when R1 is an alkyl group .... only slight activity was observed when R1 was a phenyl group” (Bufani et al. 1986), has been found, by investigators at the National Institutes of Health in the USA, to possess relevant useful activity. Subsequent to these studies, which have been

202

10

Biological Activities of the Alkaloids of the Calabar Bean

subjected to review on several occasions (Brossi 1990; Brossi et al. 1996; Greig et al. 1995a, 2005a; Klein 2007) – as have related studies on tolserine and cymserine (98, R1¼2-MeC6H4NHCO and 4-Me2CHC6H4NHCO, respectively, R2¼Me, X¼NMe, n¼1) (Greig et al. 1995a; Witkop 1998), have been the interesting and important observations that d-, l- and dl-N(1)-norphenserine were not only all active (cf Sect. 10.9) against human erythrocyte AchE and plasma BchE but also against human neuroblastoma cell β-amyloid precursor protein secretion in cell culture (Yu et al. 2003) and that “when phenserine is given to animal models of Alzheimer’s disease, rats with basal forebrain lesions, it reduces the production of β-amyloid precursor protein, which is the source of the Alzheimer’s disease toxin β-amyloid” (Brossi et al. 1996; Haroutunian 1995 – see also refs 4 and 5 as quoted in Yu et al. 1989). These observations have been further elaborated upon in subsequent investigations with phenserine (Shaw et al. 2001; Utsuki et al. 2006) and cymserine (Greig et al. 2005b), and analogues. The earlier of the above interesting and encouraging properties in phenserine have led to it being subjected to clinical trial (Brossi et al. 1996; Greig et al. 1995a, 2005a) – along with that for Posiphen (“the positive isomer of Phenserine”) and “BisNorCymcerine” [Axonyx 2005 (a)(b)(c)] and to patent cover being secured (Brossi et al. 1996; Greig et al. 1995b) – the already published synthesis of phenserine some eight decades earlier by reaction of l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) in benzene solution with phenylisocyanate in ethereal solution in the presence of a trace of sodium (Polonovski 1916) (Sect. 2.2) made it impossible to get a substance claim in this patent but the biological and pharmacological studies effected at the National Institutes of Health made it possible to get use patents covering dementia (Brossi 2004). However, by analogy with the toxicity of heptastigmine which may be caused by its metabolism to afford heptylamine (vide supra), it appears possible that one of the metabolites of phenserine could be aniline (cf Fig. 10.1) which must be borne-in-mind should toxic side-effects become manifest. Perhaps it was with this possibility in mind that the dl-3(a)-phenyl congener of physostigmine (and of phenserine) were synthesised and biologically evaluated but were, unfortunately, “found to lack cholinesterase inhibitory activity”. “Clearly, substitution of the 3a-methyl group with a sterically bulky phenyl group in either physostigmine or phenserine is not well tolerated and is detrimental to inhibitory cholinesterase activity” (Pei et al. 1998). However, other approaches could be taken to increase the lipophilicity of lphysostigmine as a “lead compound” in order to facilitate brain-targeting and thereby make possible treatment for Alzheimer’s disease. Thus:-. 1. Because of its antinociceptive activity (Sect. 10.11), it is apparent that dl-9,12dimethyl-6-hydroxyechiboline (89, R1¼HO, R2¼R3¼Me, X¼H2, n¼2) penetrates the blood-brain barrier and therefore a biological investigation of its O-(N-methylcarbamylated) derivative 89 (R1¼MeNHCOO, R2¼R3¼Me, X¼H2, n¼2) would be of interest, as would that of its lower homologue 89 (R1¼MeNHCOO, R2¼R3¼Me, X¼H2, n¼1) which has already been synthesised (Fritz and Stock 1970) (Sect. 2.3.2) but apparently not studied pharmacologically.

10.7

Use of l-physostigmine in the:-

R1

203 5

6

9a

7 8

4

4a

N

R2

10 1

n3

2

N12 11 X

R3 89 (R1= MeO, R2= Me, R3= H, X= O, n=1)

89 (R1= MeO, R2= Me, R3= H, X= H2, n=1)

89 (R1, MeNHCOO, R2= R3= Me, X= H2, n=1) 2. Another series of homologues of l-physostigmine worthy of evaluation would be 98 (R1¼MeNHCO, R2¼Me, X¼NMe, n2). The synthesis of these could involve in their latter (ring-closure) stage a Julian and Pikl type approach (Sect. 2.3.1), as has already been the case in that of the racemate of 98 (R1¼H, R2¼Me, X¼NMe, n¼2) (Kolosov et al. 1953d) (see also Hino 1961b; Hino and Ogawa 1961; Hino and Yamada 1963; Sugasawa and Murayama 1958b; Yamada et al. 1963). 3. It has been suggested (Robinson 2002) that, with respect to DAT, “Such antiamnesic activity may also be manifest in one of the minor alkaloids of P. venenosum,namely l-physovenine [98, R1¼MeNHCO, R2¼Me, X¼O, n¼1 (Sects. 2.4 and Chap. 3)], that is, like l-physostigmine [98, R1¼MeNHCO, R2¼Me, n¼1 (Sect. 2.4)], a highly active inhibitor of AchE [Robinson and Robinson 1968 (Sect. 10.4 and see footnote 8)] but may be more lipophilic – with a consequent increased ability to cross the blood-brain barrier – than the latter alkaloid”. It is of related interest that when l-physovenine analogues 98 (R2¼EtNHCO, 2-MeC6H4NHCO and 4-Me2CHC6H4NHCO, R2¼Me, X¼O, n¼1) and other synthetic products were shown to have potent antiAchE and antiBchE activities – similar to that of the corresponding physostigmines (Greig et al. 1995a) it was also noted that “These enzymes are validated targets in the treatment of Alzheimer’s disease (AD)” (Luo et al. 2007). However, some physovenine analogues have shown “significant toxicity after s.c. application to mice” (Yu et al. 1991) – see also (Greig et al. 1995a) which may make them unattractive for further investigation as potential therapeuticals. On the other hand, an approach to increasing the armamentarium for the treatment of DAT might be initiated by synthetic and pharmacologic investigations into the braintargeting of cholinomimetic moieties.

204

10.8

10

Biological Activities of the Alkaloids of the Calabar Bean

Bactericidal and Insecticidal Activities of l-Physostigmine

It has been reported (Iwasa et al. 1981) that l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) (Chap. 2) is both insecticidal and bactericidal and (Murao and Hayashi 1986) that, toward silkworm larvae, l-physostigmine (98, R1¼MeNHCO,R2 ¼ Me, X¼NMe, n¼1) (Chap. 2) exhibited strong insecticidal activity whereas l-N(8)-norphysostigmine (98, R1¼MeNHCO,R2¼H, X¼NMe, n¼1) (Chap. 5) was less active, suggesting (Murao and Hayashi 1986) that “the methyl at N-8 plays an important role in the activity”. However, be-this-as-it-may, a review of the carbamate insecticides (Holmstedt 1972) [which have also been reviewed by others (Kuhr and Dorough 1976; Matsumura 1975; Metcalf 1955)] had included the investigations by Thomas Fraser (see footnote 17 of Chap. 1) which resulted from his receiving in 1863 Calabar beans that showed indications of attack by an insect and led to his conclusion that “(1) The caterpillar of [the moth] Deiopeia pulchella feeds on the virulent poison contained in the kernel of the seed of Physostigma venenosum; and that (2) This caterpillar is unaffected by the poisonous principle of the kernel-eserine”. The review also included the observations that “Physostigmine is indeed a poor insecticide, as has been shown by modern testing” although it “can also be considered as the prototype of the carbamates used as insecticides” (Holmstedt 1972). “Before and after World War II, cholinesterase inhibitors of the organophosphorus type [see footnote 1] were developed as insecticides” (Holmstedt 1972) although this use was clearly curtailed by the observation that “OP insecticides alone cause over one million acute injuries and over 20,000 deaths per year worldwide, and sub-acute occupational exposure to such insecticides was shown to increase the risk for farmers to develop leukemias” (Soreq et al. 1992) – see also (Taylor 1996). Work on carbamates as insecticides began with their development as insect repellents and was initiated by Gysin and his colleagues at J.R.Geigy A.-G Basel [see, for example (Gysin 1952a, b, 1954; Gysin and Margot 1954, 1956; Gysin et al. 1954a; b, c)]. One of the compounds investigated was 5,5-dimethyldihydroresorcinol dimethylcarbamate (196) which ultimately became known as Dimetan and, although it possessed only a weak insect repellent activity, when subject to investigation by the biologists Wiesman and Lotmar was found to exhibit promising insecticidal activity. Consequently, a considerable number of analogous carbamates were synthesised and biologically evaluated, as a result of which one compound was found to display systematic insecticidal activity (Holmstedt 1972). The introduction of N-methylcarbamates increased the versatility of the carbamates as is exemplified by Sevin (or Carbaryl) (197) which, because of its low acute and chronic toxicity to mammals and its environmental degradability, is one of the most widely used broad spectrum insecticides. By 1972 it had replaced DDT for a number of uses and was recommended for more than 149 uses on 46 crops (Holmstedt 1972). A further example is Baygon (198) which became a standard for household pest

10.9

Prophylactic Protection with d-physostigmine Against Intoxication. . .

205

O Me OCONMe2

Me

196

OCONHMe

197

OCONHMe OCHMe2

198 control and by 1967 was showing promise as a residual insecticide against adult malaria vectors, having passed all stages of the World Health Organisation programme (Holmstedt 1972). It would appear “that esterase-inhibition is an essential factor of insecticidal action. But it is still not clear why some cholinesterase inhibitors are not insecticides” (Holmstedt 1972). Furthermore, “While it is well established that the acute toxicity of carbamates and organophosphates [see footnote 1] is mediated primarily by a marked increase in acetylcholine concentration subsequent to central and peripheral cholinesterase inhibition, there is good evidence for direct noncholinergic effects of these compounds which may not be mediated by acetylcholine accumulation” (Marquis 1985). Clearly, the mode of action of insecticides leaves much to be discovered (Holmstedt 1972; Marquis 1985).

10.9

Prophylactic Protection with d-physostigmine Against Intoxication by Organophosphates (Including “Nerve Gases”)

It has long been recognised {for examples, [Casy 1970, Cushny 1926 (see footnote 6), Nogrady 1985(a)]} that the activites of many biologically-active compounds that contain chiral entities are usually enantiospecific. Indeed, many instances are

206

10

Biological Activities of the Alkaloids of the Calabar Bean

known [see, for examples (Brossi, 1985, 1994; Brossi and Pei 1998; Casy 1970)] in which the biological activities of optically isomeric compounds exhibit either qualitative or quantitative differences, with Louis Pasteur being the first to make such an observation as long ago as 1858 when he found that yeasts and moulds metabolically differentiate between d- and l-tartarates (Pasteur 1858). Indeed, “Because they interact with optically active, asymmetric biological macromolecules such as proteins, polynucleotides, or glycolipids acting as receptors, it is eminently reasonable that many drugs show stereochemical specificity” [Nogrady 1985(a)]. Thus, when referring to the optical isomers of physostigmine – and presumably of eseroline (Sect. 10.11), the statement, that “the remarkable enantiospecificity in pharmacological activities such as antiacetylcholinesterase and analgesic activities has been recognised” (Takano and Ogasawara 1989) is misleading. Thus, for example, in view of the already-known data (vide supra), there is nothing “remarkable” about the observation (Dale 1969; Dale and Robinson 1970) [including that of asymmetrical symbiosis] – investigated with financial support from the Medical Research Council, that measurements of the antiAchE activities of the alkaloids l-physostigmine and lphysovenine and their respective synthetic enantiomers in vitro showed that the latter compounds were largely inactive in the assay, a finding that was later confirmed for l- and d-physostigmines by Brossi and his co-workers (Atack et al. 1989; Brossi and Yu 1988a, b; Brossi et al. 1986; Takano and Ogasawara 1989). However, what is truly “remarkable” is that a noteable and currently relevant (Sect. 10.7.2) exception to the above generalisation would appear to be observation that all three isomers of l-N1-norphenserine (98, R1¼PhNHCO, R2¼Me, X¼NH, n¼1), namely its enantiomers and its racemate, when investigated in vitro for inhibitory activity against human erythrocyte AchE and plasma BchE as well as against human neuroblastoma cell ß-amyloid precursor protein secretion in cell culture, were all appreciably active (Yu et al. 2003). Perhaps, via opening of their rings C at the enzymic receptor site (Sect. 10.4) with ultimate generation of sp2-hybridisation and thereby loss of asymmetry at their former C(3a)- atoms, all three might be transformed into the same biologically-active metabolite? More in keeping with the above generalisation are a number of enantiomeric pairs of compounds that have been assessed as antiAchEs and the following results of which have already been presented in review [(Long 1963), as quoted in (Dale and Robinson 1970)]. Thus, the l-isomer of isomethadone (Me2NCH2CHMeCPh2COEt) is thirty times more active than the d-isomer, the d-isomers of active l-amino acids are far weaker inhibitors than the corresponding l-isomers and it is the l-isomer of miotine (187) which is the active isomer. Since a wide range of alkaloids have been found to exhibit antiAchE (see footnote 5), it would be of interest to investigate the comparative antiAchE activities of the optical isomers that have been obtained by resolution of the natuarallyoccurring racemic alkaloid, ψ-akuammicine (Edwards and Smith 1960) and the d- and lmitraphyllines that it is claimed (Badger et al. 1950) have been isolated from the bark and leaves, respectively of Mitragyna rubrestipulaceae. Financial support from the Medical Research Council permitted (Dale 1969; Dale and Robinson 1970) investigations inhibiting in vitro erythrocyte AchE which found that the activity of synthetic dphysovenine (95, R¼MeNHCO, X¼O) is much lower than that of the natural alkaloid,

10.10

Other Clinical Uses of l-physostigmine

207

l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1), an observation later confirmed by others (Yu et al. 1991) and analogous to that between the naturally-occuring lphysostigmine (98, R1¼MeNHCO, R2¼Me X¼NMe, n¼1) and its much less active synthetic d-enantiomer 95 (R¼MeNHCO, X¼NMe) (vide supra) and both of which invalidate (Longmore and Robinson 1973) the claim (Petcher and Pauling 1973) that “AChE appears to be relatively insensitive to molecular shape as far as reversible inhibition is concerned”. “During the years of the Cold War with all the anxiety about chemical warfare the search for the effective antidote led to a surprising discovery” (Witkop 1998), namely that, notwithstanding its very weak antiAchE activity, a similar protection to that provided by l-physostigmine (Sect. 10.7.1) against the lethality of the organophosphate (see footnote 1) Sarin (271) also results from pretreatment with its d-enantiomer. However, in this later occurrence the “protective effect almost certainly does not depend at inhibition of acetylcholinesterase but may depend on a direct blockade at the nicotinic-acetylcholine receptor and its ion channel” (Kawabuchi et al. 1988).

10.10

Other Clinical Uses of l-physostigmine

l-Physostigmine is established for the relief of central cholinergic intoxication (Duvoisin and Katz 1968; Grancher and Baldessarini 1975) for anticholinergic agents such as atropine (see footnote 17 of Chap. 1) (Brown and Taylor 1996; Taylor 1996) [and related antimuscarinic drugs (Brown and Taylor 1996; Muhtadi and El-Hawary 1989; Triggle et al. 1998)] and curare (decurarisation in anaesthesiology) (Henry 1949; Holmstedt 1972) and from overdoses of tricyclic antidepressants (Brown and Taylor 1996; Burks et al. 1974; Grancher and Baldessarini 1975; Muhtadi and El-Hawary 1989; Triggle et al. 1998; Wright 1976); antiparkinson drugs (Grancher and Baldessarini 1975), phenothiazines (Brown and Taylor 1996; Muhtadi and El-Hawary 1989), antihistamines (Brown and Taylor 1996; Muhtadi and El-Hawary 1989; Triggle et al. 1998), antipsychotics (Triggle et al. 1998), benzodiazepines (Muhtadi and El-Hawary 1989; Triggle et al. 1998) and asthma powders (stramonium) and sleeping preparations (Muhtadi and El-Hawary 1989). It also offers a treatment for “intestinal atony”and for “urinary retention” (Brown and Taylor 1996; Taylor 1996; Yu et al. 1991 – see Merck Index 1989) and in both animals and humans improves morphine analgesia (Hartvig et al. 1986). Furthermore, using l-physostigmine as a “lead compound”, the work of Aeschlimann and Reinert (1931) at ROCHE (Hoffmann- La Roche) (Sect. 10.1.1) resulted in the commercial availability of Prostigmine (188, R1-R2¼Me, X¼MeSO4) which has been found, especially in addicts, to potentiate the analgesic activity of morphine (Andrews 1942; Himmelsbach et al. 1942; Slaughter et al. 1940, 1943).

208

10.11

10

Biological Activities of the Alkaloids of the Calabar Bean

Antinociceptive Activity of l-eseroline and Its Synthetic Analogues

l-Physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1), in various tests, has been shown to possess antinociceptive activity (Dayton and Garrett 1973; Flodmark and Wramner 1945; Harris et al. 1969; Hendershot and Forsaith 1959; Ireson 1970; Klee and Streaty 1974; Pleuvry and Tobias 1971). The formation of l-eseroline (98, Me

R1 O N R2

H

X

n

98

R1¼H, R2¼Me, X¼NMe, n¼1), “probably a major metabolite of the drug” (Schönenberger et al. 1986b), may be the likely cause of this activity (Schönenberger et al. 1986b), a possibility having found support in the disclosure from the University of Firenze (Florence) in Italy (Bartolini et al. 1978) that l-eseroline, which is the product from the hydrolysis of l-physostigmine (see footnote 7 of Chap. 2), possesses antinociceptive activity that was dependent upon opioid receptor stimulation and comparable in potency with that of morphine. This observation was elaborated upon by the Italian group (Bartolini et al. 1981; Fürst et al. 1982; Galli et al. 1979) and patents secured (Bartolini et al. 1979, 1982), verified and shown – perhaps not surprisingly (Sect. 10.9) – to be enantiospecific by Brossi and his colleagues at the National Institutes of Health in the USA (Schönenberger et al. 1986b) who also provided “a qualitative structure-activity profile for the eseroline group of analgesics”. Furthermore, together with a summary of these studies was presented the intriguing suggestion that “The finding of potent analgesics among hexahydropyrroloindoles will undoubtedly stimulate the development of analgesics and analgesic antagonists in the indole alkaloid series, up to now almost exclusively a domain of isoquinoline alkaloids” (Brossi 1990) – support for this hypothesis is awaited with interest! In the meantime, “it will be interesting to see whether substitution of the nitrogen atoms [in l-eseroline] with allyl or cyclopropylmethyl groups will afford analgesic antagonists” (Brossi 1992). Unfortunately, along with its antinociceptive activity, it would appear that l-eseroline also retains undesirable antiAchE activity (Bartolini et al. 1981; Galli et al. 1982) although reports to the contrary are available (Bartolini et al. 1981; Ellis et al. 1943; Galli et al. 1979; Hemsworth and West 1970). Moreover, as the free base, l-eseroline is unstable and is readily oxidised to afford rubreserine (138 Ð 139 in which the zwitterionic mesomer 139 makes the major contribution) (Sect. 2.8.2) and, although it has been reported that “Its salts, however, are more stable and can be stored as dry powders or in solution in the presence of an antioxidant without noticeable loss of activity: in this study we have used the salicylate”(Galli et al. 1979) and that “some of its salts with acids, like salicylic, fumaric and tartaric, were quite stable in solution and thus more suitable for pharmaceutical use” (Bartolini et al. 1981), others have suggested that the “sensitivity of eserolines to air-oxidation, although suppressed by antioxidants, such as ascorbic acid,

10.11

Antinociceptive Activity of l-eseroline and Its Synthetic Analogues

209

is in our opinion a serious drawback to further development of compounds of this series as pharmaceutically useful drugs” (Schönenberger et al. 1986b). Nevertheless, this obstacle might be obviated if the drugs were formulated in the solid phase as the appropriate salts for oral administration, notwithstanding the possibility of deactivation by in vivo post-administrative oxidation. However, these considerations may be rendered unnecessary by the results (Glamkowski et al. 1989c) (see also Glamkowski et al. 1989a, b, Hamer et al. 1986, 1988) of research eminating from the laboratories of Hoechst-Roussel Pharmaceuticals Inc. in North-Sommerville, NJ, USA. In this instance it was observed (Glamkowski et al. 1989c) that l-7-bromoeseroline (199) {prepared Me

HO

Br

N N H Me Me

199 from l-physostigmine (98, R1¼MeNHCO, R2¼Me, X¼NMe, n¼1) by sequential bromination with N-bromosuccinimide [a similar bromination of dl-physovenine (152, R1¼MeNHCOO, R2¼R4¼Me, R3¼H) to dl-7-bromophysovenine was also R4

R1 N 2

R

R3

O

152 later (Luo et al. 1990) effected] and alkaline hydrolysis, and code-named HP-736} is a highly potent [considerably even more so than l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1)] morphine-like centrally acting analgesic with excellent oral activity and stability and agonist actions that are μ-receptor mediated although it may have antagonist actions at the x- and δ-receptors as well, and superior to morphine in its antinociceptive effects in rodents with significantly reduced side effects and a duration of action in mice about half that of morphine. The results of the proposed (Wright 1984) investigation into “the variety of biological activity” of dihydroflustramine (172) – and, hopefully, other CH2 Me

Me

Br

N

H

H

172

N Me

210

10

Biological Activities of the Alkaloids of the Calabar Bean

structurally-related natural products containing the 7-bromophysostigmine ring system that have been isolated from the marine bryozoan Flustra foliacea (L) (see footnote 9 of Chap. 2) – are awaited with interest. In attempts (Yu et al. 1991) to develope analgesics using l-physovenine (98, R1¼MeNHCO, R2¼Me, X¼O, n¼1) (Chap. 3) as a “lead compound”, dphysovenine, dl-physovenol (3, R1¼H, R2¼Me, X¼O), dl-7-bromophysovenine (Luo et al. 1990) (vide supra) and the l-phenylcarbamate 98(R1¼PhNHCO, R2¼Me, X¼O, n¼1) have been investigated along with the alkaloid, when it was found that “these compounds do not exhibit opiate-like analgesic effects ......but show considerable activity in the Writhing test” with “the natural alkaloid” being “twice as potent” as its enantiomer (cf Sect. 10.9), although, however, “the compounds show significant toxicity after s.c. application to mice, making this group of compounds unattractive for further development [either] as analgesic agents” (Yu et al. 1991) or “as nonsteroidal and nonopiatelike analgesics” (Brossi 1992). Nevertheless, is has been claimed that “the drastic changes in biochemical behaviour seen after substitution of the N(1)-nitrogen atom in [98, R1¼H, R2¼Me, X¼NMe, n¼1] by an oxygen atom invites further study” (Brossi 1992). Well-known and potent analgesics, which have been subjected to review [Casy and Parfitt 1986; Nogrady 1985(c)], include compounds having the molecular skeleton of morphine with the addition of a substituted 2-carbon atom bridge between the 6 and 14 carbon atoms in the morphine structure. Although this modification of the structure still has not taken away the high dependence liability it has given rise to the situation whereby the opium alkaloid thebaine – which is not a narcotic analgesic but a convulsant – undergoes Diels-Alder reactions with alkyl vinyl ketones ultimately to afford, for example, etorphine which is 5000–10,000 times more potent than morphine as a result of its very high lipophilicity and very easy penetration of the blood-brain barrier [a concept appertaining to the development from l-physostigmine of therapeuticals for use in the treatment of DAT (Sect. 10.7.2)] and has in guinea pigs an agonist analgesic potency some 8600 times that of morphine. However, although in humans it is also a potent analgesic, it has a low therapeutic index, causing considerable respiratory depression in primates. Its use has therefore been limited to that of “Immobilon” in veterinary practice. Nevertheless, this concept when applied in another approach to potentially increasing the therapeutic index of l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) as an analgesic received financial support from the Medical Research Council and involved the introduction of a tetramethylene bridge between the C(3a)- and C(8a)- atoms of eseroline’s ring system with a concomitant removal of the C(3a)-methyl group (Cox and Robinson 1988; Rees et al. 1989; Robinson 2002; Robinson et al. 1987, 1988). Initially to this effect (Robinson et al. 1987; 1988), an equimolar mixture of 4-methoxyphenylhydrazine (70, R1¼4-MeO, R2¼H) with 2-(2-chloroethyl) cyclohexanone [71, R3 + R4¼(CH2)4, R5¼Cl, n¼1] in ethanolic solution was boiled under reflux to afford dl-6-methoxyechiboline (89, R1¼MeO, R2¼R3¼H, X¼H2, n¼2) (see footnote 31 of Chap. 2). Bis–N–formylation of this product was then effected, either by heating with a mixture of acetic anhydride and formic acid or by treatment with acetic-formic anhydride in dichloromethane at room

10.11

Antinociceptive Activity of l-eseroline and Its Synthetic Analogues

211

temperature, to yield 89 (R1¼MeO, R2¼R3¼CHO, X¼H2, n¼2) which upon reduction with lithium aluminium hydride in boiling tetrahydrofuran under reflux gave 89 (R1¼MeO, R2¼R3¼Me, X¼H2, n¼2). This was then O-demethylated using boron tribromide in boiling dichloromethane under reflux to yield dl-6-hydroxy-9,12– dimethylechiboline (89, R1¼HO, R2¼R3¼Me, X¼H2, n¼2) which was characterised as its dihydrochloride (Robinson et al. 1987, 1988 – see also Rees et al. 1989). Another synthetic route to this product and to analogues in which the N(12)substituent could be other than a methyl group (Cox and Robinson 1988; Robinson et al. 1987) started when an equimolar mixture of 4-methoxy-Nαmethylphenylhydrazine (70, R1¼4-MeO, R2¼Me) with ethyl cyclohexanone–2R1 N NH2 R2

70 acetate in dry benzene containing a trace of glacial acetic acid was boiled under reflux employing a Dean & Stark separator to give the corresponding arylhydrazone 87 (n¼2) which was isolated as an oil that, without further characterisation was in MeO

CH2COOEt n

N N Me

87

boiling glacial acetic acid solution under reflux, converted into dl-6-methoxy-9methyl-11-oxoechiboline (89, R1¼MeO, R2¼Me, R3¼H, X¼O, n¼2) which underwent reduction with lithium aluminium hydride in boiling tetrahydrofuran solution under reflux [see also (Cox and Robinson 1988) (see footnote 32 of Chap. 2)] to yield 6-methoxy-9-methylechiboline (89, R1¼MeO, R2¼Me, R3¼H, X¼H2, n¼2). This was reacted:1. With acetic-formic anhydride in dichloromethane under anhydrous conditions at  O C with cooling to afford 89 (R1¼MeO, R2¼Me, R3¼CHO, X¼H2, n¼2) which was reduced with lithium aluminium hydride in tetrahydrofuran boiling under reflux to yield dl-9,12–dimethyl–6–methoxyechiboline (89, R1¼MeO, R2¼R3¼Me, X¼H2, n¼2) (Robinson et al. 1987). 2. With a strirred solution of cyclopropylcarboxylic acid anhydride in dichloromethane at room temperature (initially with cooling) to afford 89 (R1¼MeO, R2¼Me, R3¼COcyclopropyl, X¼H2, n¼2) which was reduced with lithium aluminium hydride in boiling tetrahydrofuran under reflux to afford 89 (R1¼MeO, R2¼Me, R3¼CH2cyclopropyl, X¼H2, n¼2). O-Demethylation of this by treatment with boron tribromide in dichloromethane at room temperature yielded 89 (R1¼HO, R2¼Me,

212

10

Biological Activities of the Alkaloids of the Calabar Bean

R3¼CH2cyclopropyl, X¼H2, n¼2) which was characterised as its dihydrobromide (Robinson et al. 1987). 3. As a solution in aqueous dimethylformamide in the presence of potassium carbonate with a solution of allyl bromide in dichloromethane and dimethylformamide at room temperature to yield 89 (R1¼MeO, R2¼Me, R3¼CH2CH¼CH2, X¼H2, n¼2) which was O-demethylated, using boron tribromide in dichloromethane at room temperature, to afford 89 (R1¼HO, R2¼Me, R3¼CH2CH¼CH2, X¼H2, n¼2) which was characterised as its dihydrobromide (Robinson et al. 1987). The in vitro pharmacology of dl-9,12–dimethyl-6-hydroxyechiboline (89, R1¼HO, R ¼R3¼Me, X¼H2, n¼2), which has analgesic activity comparable with that of morphine and was antagonised by naloxone (Robinson 2002; Robinson et al. 1987), has been compared with that of its 12-allyl analogue 89 (R1¼HO, R2¼Me, R3¼CH2CH¼CH2, X¼H2, n¼2), and with l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1), morphine and ketazocine (Rees et al. 1989; Robinson 2002; Robinson et al. 1987). Of the utmost importance is the observation that, in contrast to l-eseroline, dl-9,12–dimethyl–6–hydroxyechiboline lacks antiAchE activity – unfortunately, the 12-allyl analogue was not investigated for such activity. Furthermore, l-eseroline and the two echiboline derivatives 89 (R1¼HO, R2¼Me, R3¼Me and CH2CH¼CH2, X¼H2, n¼2) possess opioid activity on in vitro systems including the guinea pig ileum and the mouse vas deferens. That of dl-9,12–dimethyl–6– hydroxyechiboline (89, R1¼HO, R2¼R3¼Me, X¼H2, n¼2) suggested that is has a receptor profile similar to that of a μ-opioid agonist but it differed from morphine in its susceptibility to naloxone on the mouse vas deferens. The analgesic activity predicted from the above in vitro screen has been confimed (Robinson et al. 1987) using two standard animal tests in mice, namely the Hot Plate Test and the Abdominal Constriction (acetic acid) Test, with the test drug being given by subcutaneous injection. The drug is antagonised by naloxone and the potency relative to morphine is somewhat greater than that predicted from the in vitro tests. It is apparent that these echibolines are worthy of further investigation, involving their resolution [probably chromatographically – either of the free bases using a chiral stationary phase or via the formation, where possible, of diastereoisomeric ureas (see footnote 21 of Chap. 2)], and in vitro pharmacological investigation of the enantiomers thereby obtained. Ultimately their clinical evaluation would clearly be of interest and of potential utility toward the discovery of a broad spectrum, effective and safe analgesic for, hopefully, relief from the extremely debilitating effects of acute and/or prolonged pain. Another series of homologues of l-eseroline (98, R1¼H, R2¼Me, X¼NMe, n¼1) worthy of analgesic investigation would be 98 (R1¼H, R2¼Me, X¼NMe, n2). The synthesis of these could involve in their latter ring-closure stage a Julian and Pikl type approach (Sect. 2.3.1) as has already been the case in that of the racemate of 98 (R1¼H, R2¼Me, X¼NMe, n¼2) (Kolosov et al. 1953) (see also Hino 1961b; Hino and Ogawa 1961; Hino and Yamada 1963; Sugasawa and Murayama 1958b; Yamada et al. 1963). Synthetic approaches to the echiboline ring system (vide supra and see footnote 31 and 32 of Chap. 2) might likewise be adapted for the preparation of homologues. 2

10.12

10.12

Cytotoxicities of Rubreserine and another Structurally-Related. . .

213

Cytotoxicities of Rubreserine and another Structurally-Related Degradation Product of l-Physostigmine

The therapeutic profile relating to l-physostigmine has been recently widened by the interesting detection of cytotoxic activity amongst its degradation products. Thus, although “Compounds [144] and [145] were evaluated for their cytotoxic activity Me

Me

N N

N

O

Me

Me

Me

Me

N

N

N

Me

144

Me

N O

Me

N

N

N Me

145

using a battery of human cancer cell lines [elaborated in (Poobrasert et al. 1996)] and were judged to be inactive” (Poobrasert et al. 1997) and in similar screening rubreserine (138 Ð 139 in which the zwitterionic mesomer 139 makes the major Me

O O

N Me

H

N Me

138 Me

O O

N

N H Me Me

139

214

10

Biological Activities of the Alkaloids of the Calabar Bean

contribution) exhibited only limited very weak activity (Poobrasert et al. 1996), “Compound [146] was very active in all of the human cancer cell lines tested Me

O EtNH

N

146

N Me

(ED50