Circadian Clocks: Methods and Protocols [1st ed.] 9781071603802, 9781071603819

Citation preview

Methods in Molecular Biology 2130

Steven A. Brown Editor

Circadian Clocks Methods and Protocols

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, UK

For further volumes: http://www.springer.com/series/7651

For over 35 years, biological scientists have come to rely on the research protocols and methodologies in the critically acclaimed Methods in Molecular Biology series. The series was the first to introduce the step-by-step protocols approach that has become the standard in all biomedical protocol publishing. Each protocol is provided in readily-reproducible step-bystep fashion, opening with an introductory overview, a list of the materials and reagents needed to complete the experiment, and followed by a detailed procedure that is supported with a helpful notes section offering tips and tricks of the trade as well as troubleshooting advice. These hallmark features were introduced by series editor Dr. John Walker and constitute the key ingredient in each and every volume of the Methods in Molecular Biology series. Tested and trusted, comprehensive and reliable, all protocols from the series are indexed in PubMed.

Circadian Clocks Methods and Protocols

Edited by

Steven A. Brown Chronobiology and Sleep Research Group, Institute of Pharmacology and Toxicology, University of Zurich, Zürich, Switzerland

Editor Steven A. Brown Chronobiology and Sleep Research Group Institute of Pharmacology and Toxicology University of Zurich Zu¨rich, Switzerland

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-0380-2 ISBN 978-1-0716-0381-9 (eBook) https://doi.org/10.1007/978-1-0716-0381-9 © Springer Science+Business Media, LLC, part of Springer Nature 2021 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 Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.

Dedication I would like to thank the members of my laboratory who helped to proofread this book, funding sources that have made our research possible (including the Swiss National Science Foundation, the Velux Foundation, the Human Frontiers Science Program, and the University of Zurich), and especially Jacqueline Vicario, whose formidable organizational skills and tireless efforts put form to chaos.

v

Preface Designing a Circadian Experiment Well before the first circadian clock mutants were discovered in the fruitfly, Drosophila melanogaster [1], biologists were using a wide diversity of model organisms and methods to study the workings of the circadian clock. As the Nobel Prize in Physiology and Medicine 2017 recognized, these daily rhythms play an essential role in nearly all aspects of physiology, and their disruption is likely a causative factor in human disease [2, 3]. Several excellent textbooks have been written recently to describe what is known about the molecular workings of circadian clocks and the physiology that they control [4–8]. Within the context of this book, studying circadian clock function is essentially the art of studying processes across the 24-h day, so we present a compilation of protocols from all branches of biology, describing how to study daily changes at a molecular level in many physiological systems. Although our topic is circadian rhythms, many of the same methods can apply to rhythms of any frequency, including ultradian (shorter than a day) or infradian (longer than a day), even circalunar (about a month) or circannual (about a year). Beyond the methods themselves, experimental design is an important factor, and it is that aspect we discuss here first.

Circadian or Just Rhythmic? A circadian rhythm has been formally defined [9] as a rhythm occurring with a period of approximately a day (hence circadian, from the Latin circa diem) [10] that persists in constant (or “free-running”) environmental conditions, that can be entrained to the local environment [11], and that is temperature-compensated—i.e., that has the same freerunning period at different environmental temperatures [12]. As an inflexible yardstick or litmus test, however, this definition is poorly adapted to modern experimental biology and medicine: entrainment may be compromised or altered, even if a process is normally under circadian control (e.g., human physiology during shiftwork), and temperature compensation is hard to verify in a homeotherm such as mouse or man. Thus, practically, the first step in experimental circadian science is deciding which of these criteria are relevant within a given study. Most simply, many biological processes may vary with time of day, but are directly environmentally driven—e.g., light-dependent gene expression. Both “true” circadian processes fulfilling all aspects of the formal definition, and others simply displaying a near-24h periodicity, are generally termed as “rhythmic” or “cyclic.” In specific contexts, many investigators today make the conscious decision that circadian rhythmicity per se is irrelevant to the scientific question being posed or too laborious to fully verify and will intentionally choose to define their process as cyclic.

vii

viii

Preface

Characterizing Cyclic Processes As the first step, even before one approaches the question of whether a process is circadian, one must ascertain whether it is cyclic or rhythmic at all. Mathematically, this probability is directly related to the frequency of sampling relative to the period length of the oscillation. It is also related to the number of replicates and the experimental noise. At the end of this Foreword, multiple free software platforms are presented for determining probability of rhythmicity. However, even before one arrives at the analysis of rhythmic data, one must collect it, and the design of this collection is far from trivial. The first question to ask is whether sampling frequency is sufficiently dense and length of sampling sufficiently long relative to the period of oscillation that “misfitting” of an improper rhythm is not an issue. When one can answer yes to this question, rhythmic analysis is straightforward. For this reason, an initial analysis is best conducted as an automated highfrequency measurement where possible. These include constant behavioral and physiological analyses (movement, body temperature) or the so-called reporter analyses in which a particular molecular event is reported constantly in real time by genetically engineered means, e.g., interrogation of introduced fluorescent or bioluminescent proteins. These methods are covered extensively for different model organisms in several chapters. When they can be used, initial experimental design is simplified to observation over multiple days in the desired paradigm, and analysis generally consists of fitting the acquired data to a known rhythmic function (e.g., a decaying sine wave), or simple fast Fourier analysis. Using these methods, even very weak or complex rhythms can be identified with statistical certainty. Generic statistical packages in R, Python, and Matlab can be used for this purpose. However, several web-based solutions and bespoke software packages also exist, and most of these are free. We list them at the end of this introduction. For many newly studied rhythmic processes, measurement is often complicated or expensive, so sampling is reduced to a small number of “timepoints.” At their simplest, rhythms might even be analyzed at only two timepoints, suspected peak and suspected nadir. Though requiring the least effort, such an approach is highly problematic because changes in relative values at two timepoints could result either from changes in amplitude or from changes in phase of an underlying cycle (Fig. 1). Nevertheless, many investigators, after verifying that a process is cyclic at higher resolution, will then use only two timepoints for the detailed biological characterization of the process in question if phase-shifting (changes in oscillatory phase across experimental conditions) is not an issue. In general, the question to be asked before embarking upon a two-timepoint design is simple: “am I 100% certain that I have correctly identified the peak and nadir of the analysed process, and are these the only times of interest to me?” Presuming that one has answered yes to this question, statistical analysis is easy: timepoints are directly compared to different experimental conditions, by ANOVA or Student’s t-test, and completely standard methodologies can be used to make power calculations for the number of replicates required. Initially, however, nearly all investigators interested in rhythmic processes will answer in the negative, and now a more complicated design ensues, involving a compromise between effort, expense, and resolution. If determining rhythmicity is the sole criteria of interest, somewhat counterintuitively a greater number of timepoints per oscillation or timepoints collected over a greater number of oscillations is statistically much more powerful than increasing the number of replicates per timepoint. Different strategies have been analyzed in detail elsewhere, and simulation software exists to generate

ix

Expression level

Preface

Control Outcome 1 Outcome 2 ZT0

6

12

18

0

Fig. 1 (a) Relative to Control, is the underlying rhythm in the experimental condition changed in amplitude (Outcome 1) or in phase (Outcome 2)? Impossible know with only two timepoints! Both outcomes show equivalent experimental values (blue and red dots) when only two timepoints are sampled

synthetic datasets to examine the consequences of such choices (CircaInSilico, [13]). For example, a single sample per timepoint at two-hour resolution is far more powerful than triplicates at 6-h resolution, even though 12 samples per 24 h are collected in each case. For this reason, the artificial end-to-end concatenation of replicates or duplication of data to make longer time series is statistically unwise from a circadian perspective. When statistical significance of rhythmicity is the only important question, strengths and weaknesses of different designs have been discussed at length [13]. However, a counterbalancing consideration in many “-omics”-based methods—transcriptomics and epigenetics in particular—is that standard analysis packages require technical triplicates for analysis. Our laboratory, for this reason, typically uses six timepoints per 24 h, in technical triplicate. Essentially, this design sacrifices some power in pure “circadian/not circadian” p-value, but gains statistical power for determining quantitative differences in epigenetic or transcriptomic features across neighboring timepoints. Of course, as in any other -omics-type experiment, some type of correction for multiple testing is required, generally in the form of a Benjamini–Hochberg-type false discovery rate [14], and usually incorporated directly into analysis software packages. From a circadian perspective, with less dense sampling, evaluation methods typically involve more complicated circular statistics. For such analyses, multiple software packages are listed at the end of this introduction. From these, a p-value for probability of oscillation will typically be assigned for each variable analyzed. In the case of -omics-based experiments, this means that one obtains a p-value (or q-value) for each transcript, protein, or metabolite. Thus, one must choose a “cutoff.” To aid in choosing such a threshold, we usually generate a profile of significant genes at different p-values. Typically, there will be a sharp “shoulder” on such curves suggesting a logical approximate cutoff (Fig. 2). Alternatively, one can perform downstream gene ontology analyses at multiple cutoffs: both above and below optimal stringency—which itself is dependent upon inherent experimental noise—statistical significance of ontological terms will generally decrease across all significant categories.

x

Preface

Fig. 2 Q-value (i.e., BH-corrected p-value) histogram for a sample circadian proteomics experiment. Beyond p ¼ 0.05, a sharp drop in additional significant transcripts is observed under control conditions (baseline, BL), and at the same threshold a sharp increase under experimental conditions (SD, sleep deprivation). It is easy to see that p ¼ 0.05 makes a logical choice of significance threshold to differentiate between these two conditions (from Noya, Bru¨ning et al., unpublished data)

Characterizing Circadian Processes Periodic sampling is all that is necessary to characterize a process as rhythmic or cyclic. To call it circadian, however, further steps are required. As we mentioned earlier, the formal definition of circadian rhythmicity contains multiple aspects (such as temperature compensation and entrainment) that are seldom systematically investigated. However, several additional steps are still required in order to call a rhythmic process circadian. Most importantly, the rhythmicity must continue under constant conditions. In experiments involving cultured cells, this requirement has probably already been fulfilled: the incubator itself maintains admirable stability. For animal experiments, generally the subjects are placed into a constantly dark environment for some days after the initial experiment, and periodic sampling is repeated. Animal ethics considerations in some countries may prohibit keeping of experimental animals, even nocturnal rodents, in constant darkness. Because the circadian clock is insensitive to dim red light [15] (1 log) between labeled and nonlabeled cells is selected (background signal should also be taken into account and minimized as much as possible). 6. The 0.83% NH4Cl solution is used to lyse the red blood cells. It is really important to resuspend the cells well in the NH4Cl solution and then to incubate for no more than ~5 min to prevent cell mortality. 7. IL-4 is a home-made reagent from a cell culture of P815-IL 4 cells. It takes some time to culture the cells and prepare a stock of this reagent. Therefore, it is a good idea to plan ahead of time, and to start renewing the stock when there are several 50 ml tubes left. 8. For LPS, GM-CSF, PMA/Ionomycin, Brefeldin A and OVA257-264 peptide, it is better to prepare small aliquots to avoid multiple freeze–thaw cycles (not more than 3–5 cycles for each aliquot). 9. Set the centrifuge to decelerate without brake for the density gradients. Otherwise the gradients might be disrupted by the deceleration.

48

Chloe´ C. Nobis et al.

10. The trypan blue solution is used to distinguish the dead and live cells. It is essential to take into account viability/mortality during the cell counting to adjust to the right number of cells for the experiments. Cells dyed in blue are the dead cells. After putting the 10 μl of trypan blue/cell mix in the hemocytometer, it is recommended to wait 1 min before counting the cells (otherwise they move and are harder to count). 11. Bacterial endotoxin lipopolysaccharide (LPS) stimulates immune cells such as monocytes via their Toll-like receptor (TLR) 4, leading to the release by these cells of cytokines such as IL-1β, TNF-α, or IL-6. The lectin phytohemagglutinin (PHA) activates T cell, leading to the release of cytokines such as IFN-γ or IL-2. The kinetics are different in these two cases, which explains why we recommend using 4 h and 24 h for stimulation with LPS and PHA, respectively. 12. For flow cytometry stainings, 96-round well plates are recommended to ensure that the cells pellet well during centrifugation in the washing steps. Some laboratories rather use conicwell plates; however in this case the cells are more packed at the bottom of the wells, which could possibly lead to increased cell mortality during the staining procedure. 13. Fc Block is used to block unspecific binding of the Fc receptorexpressing cells such as macrophages. Usually, the Fc Block is composed of antibodies against CD16/CD32. 14. “Fluorescence minus one” (FMO) corresponds to a staining including all antibodies except one (e.g., in a staining with four antibodies, and thus with four different fluorochromes, the number of FMO tubes that have to be included in the experiment is four, each with one of the fluorophores removed). This control staining allows controlling for possible nonspecific signal. If signal is observed in a given channel in the absence of the corresponding fluorophore (in the FMO), then one can conclude that this signal is not specific. 15. Either forceps and scissors or a scalpel can be used to remove the tissue that surrounds the bone. 16. Similar to what was described above for the stimulation of human PMBCs, LPS is used to activate dendritic cells. This activation signal leads to the upregulation of various molecules such as costimulatory molecules (e.g., CD86) and Major Histocompatibility Complex (MHC) (H2Kb, class I molecule and IAb, class II molecule). Therefore, it is important to activate the dendritic cells before adding the antigenic peptide to the cell culture, to increase the capacity of the dendritic cells to be loaded with the antigen via MHC-I or -II molecules. 17. According to the article published in 2015 by J. Helft [22] and colleagues, mouse bone marrow-derived dendritic cells present

Circadian Rhythms in the Immune System

49

a variety of dendritic cells with specific cell signatures. In this respect, the loosely adherent cells represent the dendritic cells that have the best capacity to present an antigen both in vitro and in vivo. To collect these cells, one has to aspirate up and down the cells in each well without scrapping off the cells, and then proceed to the Histodenz 14.7% gradient. 18. In general, to assess the dendritic cell activation and loading, antibodies against costimulatory molecule CD86, MHC-I or -II molecules (IAb and H2Kb), the MHC-I–OVA257-264 complex, and the specific marker of dendritic cells CD11c are used. Figure 3 presents an example of staining of BMDCs after differentiation, activation and OVA peptide loading. There should be at least 80% of CD11c+ cells in the culture. CD86 and IAb staining must increase after LPS activation. H2Kb should be positive in all conditions. As for OVA loading, the mean fluorescence intensity for the MHC-I–OVA257-264 complex staining should be at least 4–5 times higher for the OVA-loaded dendritic cells than for the cells without OVA peptide. 19. During a circadian experiment it is important to keep all parameters identical between circadian times (CTs). Thus, in a dendritic cell vaccination protocol over 24 h it is essential to keep the cells alive and with similar characteristics from the first to the last CT. To achieve this, cells should be kept in 25 ml of complete RPMI 1640 10% FBS in a 50 ml tube. The tube should be kept lying on ice on a slow shaker, with the ice changed every 4–6 h. Before each dendritic cell collection, the cells must be counted to test the viability and the same numbers of live cells should be injected. 20. Mice have two tail veins, one on each side of the tail. In general, it is difficult to see the tail veins and even more so under dim red light. To improve the injection, it is thus important to use an indirect source of light (i.e., point the red light toward the wall instead to pointing it toward the mouse). 21. The activation of dendritic cells and the subsequent loading with an antigen can lead to an increased mortality of the dendritic cells when they are kept for a long time after the treatment. To control this phenomenon, it is important to count the cells before each time point, and to check the activation and loading level again after the last time point. This ensures that the immune variation across the experiment is due to a circadian rhythm and not to a problem with the dendritic cells (e.g., a change in viability, numbers, or state over the course of the experiment). 22. The spleen is a secondary lymphoid organ where immune responses occur. In the context of intravenous injection, the

50

Chloe´ C. Nobis et al.

antigen presentation and T cell response occur mostly in the spleen, which filtrates the blood. Moreover, in a context of dendritic cells loaded with the OVA257-264 peptide the peak of the T cell response occurs 7 days postvaccination. 23. The ex vivo restimulation of splenocytes allows for the assessment of the cytokine production by T cells. In the case of flow cytometry analysis, one has to analyze the cytokine production by intracellular staining of the cytokines within the secretory pathway. To avoid the exocytosis of the cytokines during the restimulation, an inhibitor of this process is used, such as Brefeldin A. 24. The cell fixation step is essential to the intracellular staining. Indeed, one has to fix the cells and then permeabilize them to assess the cytokine secretion by the cells. After the cell fixation, the staining and acquisition of the cells by flow cytometry should be completed within 5 days. 25. The washing steps after the cell fixation or the permeabilization steps are important. Some markers could be affected by the formaldehyde and the saponin (e.g., the CD8 marker). 26. H2Kb-OVA257-264 is a tetramer that allows to assess the CD8+ T cell expansion. This molecule recognizes the T cell receptor that is restricted to the MHC-I–OVA257-264 peptide complex.

Acknowledgments The authors thank all members of the Cermakian, Labrecque, and Boivin laboratories for insightful tips and comments. This research has been supported by grants from the Canadian Institutes of Health Research (MOP-119322 to N.C. and N.L., and MOP-102724 to D.B.B. and N.C.). References 1. Duguay D, Cermakian N (2009) The crosstalk between physiology and circadian clock proteins. Chronobiol Int 26(8):1479–1513. https://doi.org/10.3109/ 07420520903497575 2. Dibner C, Schibler U, Albrecht U (2010) The mammalian circadian timing system: organization and coordination of central and peripheral clocks. Annu Rev Physiol 72:517–549. https://doi.org/10.1146/annurev-physiol021909-135821 3. Curtis AM, Bellet MM, Sassone-Corsi P, O’Neill LA (2014) Circadian clock proteins and immunity. Immunity 40(2):178–186.

https://doi.org/10.1016/j.immuni.2014.02. 002 4. Labrecque N, Cermakian N (2015) Circadian clocks in the immune system. J Biol Rhythm 30 (4):277–290. https://doi.org/10.1177/ 0748730415577723 5. Born J, Lange T, Hansen K, Molle M, Fehm HL (1997) Effects of sleep and circadian rhythm on human circulating immune cells. J Immunol 158(9):4454–4464 6. Abo T, Kawate T, Itoh K, Kumagai K (1981) Studies on the bioperiodicity of the immune response. I. Circadian rhythms of human T, B, and K cell traffic in the peripheral blood. J Immunol 126(4):1360–1363

Circadian Rhythms in the Immune System 7. Cuesta M, Boudreau P, Dubeau-Laramee G, Cermakian N, Boivin DB (2016) Simulated night shift disrupts circadian rhythms of immune functions in humans. J Immunol 196 (6):2466–2475. https://doi.org/10.4049/ jimmunol.1502422 8. Keller M, Mazuch J, Abraham U, Eom GD, Herzog ED, Volk HD, Kramer A, Maier B (2009) A circadian clock in macrophages controls inflammatory immune responses. Proc Natl Acad Sci U S A 106(50):21407–21412. https://doi.org/10.1073/pnas.0906361106 9. Gibbs J, Ince L, Matthews L, Mei J, Bell T, Yang N, Saer B, Begley N, Poolman T, Pariollaud M, Farrow S, DeMayo F, Hussell T, Worthen GS, Ray D, Loudon A (2014) An epithelial circadian clock controls pulmonary inflammation and glucocorticoid action. Nat Med 20(8):919–926. https://doi. org/10.1038/nm.3599 10. Nguyen KD, Fentress SJ, Qiu Y, Yun K, Cox JS, Chawla A (2013) Circadian gene Bmal1 regulates diurnal oscillations of Ly6C (hi) inflammatory monocytes. Science 341 (6153):1483–1488. https://doi.org/10. 1126/science.1240636 11. Nakamura Y, Ishimaru K, Tahara Y, Shibata S, Nakao A (2014) Disruption of the suprachiasmatic nucleus blunts a time of day-dependent variation in systemic anaphylactic reaction in mice. J Immunol Res 2014:474217. https:// doi.org/10.1155/2014/474217 12. Nakamura Y, Nakano N, Ishimaru K, Hara M, Ikegami T, Tahara Y, Katoh R, Ogawa H, Okumura K, Shibata S, Nishiyama C, Nakao A (2014) Circadian regulation of allergic reactions by the mast cell clock in mice. J Allergy Clin Immunol 133(2):568–575. https://doi. org/10.1016/j.jaci.2013.07.040 13. Bollinger T, Leutz A, Leliavski A, Skrum L, Kovac J, Bonacina L, Benedict C, Lange T, Westermann J, Oster H, Solbach W (2011) Circadian clocks in mouse and human CD4+ T cells. PLoS One 6(12):e29801. https://doi. org/10.1371/journal.pone.0029801 14. Hemmers S, Rudensky AY (2015) The cellintrinsic circadian clock is dispensable for lymphocyte differentiation and function. Cell Rep 11(9):1339–1349. https://doi.org/10.1016/ j.celrep.2015.04.058

51

15. Nobis CC, Dubeau Larame´e G, Kervezee L, Maurice De Sousa D, Labrecque N, Cermakian N (2019) The circadian clock of CD8 T cells modulates their early response to vaccination and the rhythmicity of related signaling pathways. Proc Natl Acad Sci USA 116(40):20077–20086. https://doi.org/10.1073/pnas.1905080116 16. Levi FA, Canon C, Touitou Y, Sulon J, Mechkouri M, Ponsart ED, Touboul JP, Vannetzel JM, Mowzowicz I, Reinberg A et al (1988) Circadian rhythms in circulating T lymphocyte subtypes and plasma testosterone, total and free cortisol in five healthy men. Clin Exp Immunol 71(2):329–335 17. Dimitrov S, Benedict C, Heutling D, Westermann J, Born J, Lange T (2009) Cortisol and epinephrine control opposing circadian rhythms in T cell subsets. Blood 113 (21):5134–5143. https://doi.org/10.1182/ blood-2008-11-190769 18. Depres-Brummer P, Bourin P, Pages N, Metzger G, Levi F (1997) Persistent T lymphocyte rhythms despite suppressed circadian clock outputs in rats. Am J Phys 273(6 Pt 2): R1891–R1899 19. Fernandes G, Halberg F, Yunis EJ, Good RA (1976) Circadian rhythmic plaque-forming cell response of spleens from mice immunized with SRBC. J Immunol 117(3):962–966 20. Suzuki S, Toyabe S, Moroda T, Tada T, Tsukahara A, Iiai T, Minagawa M, Maruyama S, Hatakeyama K, Endoh K, Abo T (1997) Circadian rhythm of leucocytes and lymphocytes subsets and its possible correlation with the function of the autonomic nervous system. Clin Exp Immunol 110(3):500–508 21. Fortier EE, Rooney J, Dardente H, Hardy MP, Labrecque N, Cermakian N (2011) Circadian variation of the response of T cells to antigen. J Immunol 187(12):6291–6300. https://doi. org/10.4049/jimmunol.1004030 22. Helft J, Bottcher J, Chakravarty P, Zelenay S, Huotari J, Schraml BU, Goubau D, Reis e Sousa C (2015) GM-CSF mouse bone marrow cultures comprise a heterogeneous population of CD11c(+)MHCII(+) macrophages and dendritic cells. Immunity 42 (6):1197–1211. https://doi.org/10.1016/j. immuni.2015.05.018

Chapter 4 Measuring Circadian Rhythms in Human Cells Ngoc-Hien Du and Steven A. Brown Abstract Human cells, especially primary fibroblasts from skin punch biopsy, have emerged over the last decade as powerful, unlimited, and easily accessible resources that bridge the gap between animal models and human subjects in basic as well as clinical research. The cells also retain molecular circadian clocks that reflect subject-specific differences in circadian physiology, and the cellular rhythms can be measured easily in large scale. This is a series of protocols that describes the procedure to measure circadian rhythms in these cells, starting from deriving fibroblasts from skin punch biopsy, to generation of stable cells expressing a circadian reporter, and finally measurement of cellular rhythms in large scale. Key words Human skin punch biopsy, Human primary skin fibroblasts, Stable cell line generation, Circadian rhythm measurement

1

Introduction Studies in humans, especially in the context of disease, are challenging due to the scarceness of relevant human tissues and the difficulty and high cost of recruiting human subjects. Peripheral blood is easily available and is the most widely used human material, but its collection must include tight control for confounding factors such as food, sleep, and stress. In addition, blood cells are terminally differentiated and are therefore less suitable for many molecular manipulations. In contrast, human primary skin fibroblasts are easily obtained from skin punch biopsy, a mostly painless and fairly noninvasive procedure [1]. The primary cells can be kept in culture for up to 20 passages and thus provide ample material for long-term assessments as well as genetic and biochemical manipulations. They have been used successfully to explore the systemic cellular effects of pathogenesis, for example, in neurodegenerative diseases, psychiatric disorders, and diabetic neuropathy [2–4]. Moreover, with the discovery of induced pluripotent stem cells, human fibroblasts have become a primary source to derive human pluripotent stem cells (iPSCs). Therefore, human fibroblasts provide useful

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_4, © Springer Science+Business Media, LLC, part of Springer Nature 2021

53

54

Ngoc-Hien Du and Steven A. Brown

platforms that are complementary to human and animal models that could help to advance understanding of the molecular mechanisms underlying human pathology. Circadian rhythms in biophysiological processes such as sleep, behavior, and metabolism are implicated in human well-being as well as human pathology. It is now recognized that these rhythms are disrupted in a wide range of diseases from neurodegenerative diseases to metabolic disorders [5, 6]. Prospective studies have identified disturbance of circadian behaviors, including decreased activity amplitude and excessive daytime sleepiness, as predictors of impaired cognitive functions and neurodegeneration, respectively [7–10]. As a result, there is an increasing need to understand if the circadian clock plays a causal role in pathogenesis, as well as the molecular mechanisms underlying such linkage. Cultured skin fibroblasts represent valuable resources for such studies, since they express circadian rhythms whose clock properties (period, amplitude) correspond to circadian properties observed in human physiology, including melatonin rhythms, behavioural rhythms, and human chronotype [11–13]. In addition, analysis of fibroblasts cellular circadian traits together with genotypes of corresponding subjects has identified the genetic basis that partially explains individual differences in circadian parameters [14]. Importantly, due to large size effect and precise measurements of molecular traits, such a study is sufficiently powered with hundreds of subjects compared to conventional thousands, if using human behavior and physiology as phenotypes. This type of cellular genetics in circadian context can be extended to explore the relationship between the circadian clock and human diseases. Human cell lines such as U2OS, derived from malignant bone tumor, have been used successfully in RNAi or drug screens to identify regulators of circadian clock function [15–17]. However, such cancerous cell lines lack individual differences present in the general population, possibly hindering their utility in a clinically relevant context. Therefore, characterization of the circadian clock in personalized cell lines such as primary fibroblasts from patients will likely become a necessity in the current era of precision medicine. Overall, the obtention, cultivation, and measurement of primary human fibroblasts are surprisingly easy. We described below a series of protocols: (1) derivation of fibroblasts from human skin punch biopsy, (2) maintenance of fibroblasts in large scale, (3) production of lentivirus carrying a circadian reporter (Bmal1-luciferase), (4) generation of cell lines that stably express circadian reporters, (5) measurement of cellular rhythms in these cells, and (6) a brief description of data analyses.

Measuring Circadian Rhythms in Human Cells

2

55

Materials

2.1 Human Skin Biopsy Processing

1. Collection medium: DMEM (with 4.5 g/l glucose, e.g., D5796 from Sigma) supplemented with 50% FBS (fetal bovine serum), 1% Pen–Strep solution. 2. Digestion medium: Prepare 2 ml digestion medium per biopsy, combining 1.8 ml of DMEM (with 4.5 g/l glucose, e.g., D5796 from Sigma, supplemented with 10% FBS, 1% amphotericin B) with 0.2 ml of Liberase (see below). 3. Culture medium: DMEM (with 4.5 g/l glucose, e.g., D5796 from Sigma) supplemented with 20% FBS, 0.1% gentamycin (Cat. G1397 from Sigma, see Note 1). 4. Liberase stock solution (Cat. 05401119001 from Roche): dissolve 9 mg or 28 WU in 8 ml of DMEM, freeze at
20  C in 0.2 ml aliquots. 5. Amphotericin B (Cat. A2942 from Sigma). 6. Trypsin-EDTA (0.5%) no phenol red (Cat. 154000054 from Gibco). 7. DPBS, no calcium, no magnesium (137 mM NaCl, 2.7 mM KCl, 8 mM Na2HPO4, 1.5 mM KH2PO4). 8. Millicell Cell Culture insert (Cat. PICMORG50 from Millipore).

2.2 Simultaneous Culturing of Hundreds of Lines

1. Advanced culture medium: Advanced DMEM (Cat. 12491023 from Gibco) supplemented with 10% FBS, 1% GlutaMax, and 0.1% gentamycin (see Notes 1 and 2). If cells are confirmed to be mycoplasma-free, gentamycin can be replaced by 1% Pen– Strep solution. 2. TrypLE Express Enzyme (1), no phenol red (Cat. 12604021 from Gibco, see Note 3). 3. DPBS, no calcium, no magnesium. 4. Freezing medium, 2: combine 20 ml of advanced culture medium, 10 ml of DMSO (dimethyl sulfoxide), and 20 ml of FBS.

2.3 Production of Lentiviral Vector by Calcium Phosphate Transfection

1. Culture medium: DMEM (with 4.5 g/l glucose, e.g., D5796 from Sigma) supplemented with 10% FBS, 1% Pen–Strep solution. 2. Trypsin-EDTA (0.5%) no phenol red (Cat. 154000054 from Gibco): dilute 1:10 in PBS as working solution (0.05% final). 3. Buffered H2O: Add 125 μl of 1 M HEPES pH 7.3 in 50 ml of H2O (2.5 mM final). Filter-sterilize through 0.22-μm nitrocellulose filter. Store up to 6 months at 4  C.

56

Ngoc-Hien Du and Steven A. Brown

4. CaCl2 0.5 M: Dissolve 36.75 g CaCl2·2H2O (MW: 147) in 500 ml H2O. Filter-sterilize through 0.22-μm nitrocellulose filter. Store up to 2 years at
80  C in 50-ml aliquots. Once thawed, the solution can be kept at 4  C for several weeks without significant change in the transfection efficiency. 5. HEPES-buffered saline (HeBS), 2: Dissolve the following reagents in 800 ml H2O: 16.36 g NaCl (MW: 58.44; 0.28 M final), 11.9 g HEPES (MW: 238.3; 0.05 M final), 0.213 g anhydrous Na2HPO4 (MW: 142; 1.5 mM final). Adjust pH to 7.00 with 10 M NaOH. Proper pH is critical. Add H2O to 1000 ml and make final pH adjustment to pH 7.00 (see Note 4). Filter-sterilize through 0.22-μm nitrocellulose filter. Store up to 2 years at
80  C in 50-ml aliquots. 6. Plasmids: pMD2G (encoding the VSV G envelope protein, #12259 from Addgene), psPAX2 (encoding the HIV-1 Rev protein, #12260 from Addgene), pWPT-GFP (#12255 from Addgene), and pLV6-Bmal-luc (#68833 from Addgene). 2.4 Generation of Stable Cell Lines Expressing Lentiviral Vector

2.5 Measurement of Cellular Circadian Rhythms

1. Blasticidin (e.g., Cat. R21001 from Gibco) 10 mg/ml stock solution: dissolve 10 mg in 10 ml cell culture grade water, sterilize using 0.22 μm filter. Store in aliquots at
20  C. 2. Protamine sulfate 8 mg/ml stock solution: dissolve, for example, 80 mg in 80 ml water, sterilize using 0.22 μm filter. Store in aliquots at
20  C. 1. Synchronization medium: Advanced culture medium supplemented with 100 nM dexamethasone (dilute 1:10,000 from 1 mM stock, see below). 2. Counting medium: DMEM (high glucose, no glutamine, no phenol red, e.g., Cat. 31053028 from Gibco) supplemented with 10% FBS, 1% GlutaMax, 1% Sodium pyruvate, 0.1% gentamycin, and 0.1 mM luciferin (dilute 1:200 from 20 mM stock, see below). 3. Dexamethasone 1 mM stock solution: dissolve 5 mg (MW: 392.46) in 12.7 ml EtOH. Store in aliquots at
20  C. 4. Luciferin 20 mM stock solution: Dissolve 70 mg of luciferin in 6 ml of PBS pH 12.0 then vortex well. Add 4 ml of PBS pH 2.0 then adjust pH to 7.5 with few μl (about 30 μl) of NaOH 10 M. Add PBS pH 7.5 to a final volume of 12 ml. Sterilize using 0.22 μm filter. Store in aliquots at
80  C. If preferred, potassium and sodium salts of luciferin (water soluble) are also commercially available. 5. Black 24-well plate (e.g., #41082 from Berthold). 6. Real-time luminometer.

Measuring Circadian Rhythms in Human Cells

2.6

3

Data Analysis

57

Rhythmic data can be analyzed using a commercial software, LumiCycle Analysis, from Actimetrics [18]. There are also freely available tools, which are listed in Chap. 1 of this book.

Methods

3.1 Human Skin Biopsy Processing

1. Prior to taking biopsy, fill collection tubes (e.g., 1.5 ml Eppendorf tube) with collection medium (e.g., 1.5 ml), and keep on ice (see Note 5). 2. Biopsies are standard 2 mm round dermal punches, taken from upper arm or buttocks by professional dermatologist (Fig. 1). It is advisable to take 2 biopsies per subject if possible. Use a pair of clean tweezers to move and immerse the biopsies in prechilled collection medium immediately (Fig. 1, see Note 6). 3. Use a pair of tweezers to remove the biopsy from the collection tube and place each punch in a separate 4-cm culture dish. Be careful not to completely close the tweezers and crush the sample. 4. Incubate the biopsy in 2 ml prewarmed digestion medium for 4–14 h at 37  C/5–7% CO2. Do not digest longer (see Note 7). Figure 2 shows examples of a biopsy before and after digestion. 5. Using a pipet tip whose end has been cut to allow passage of larger fragments to remove the digestion medium to a falcon tube containing 10 ml warm PBS. Rinse plate with a bit more PBS to obtain all fragments. 6. Spin samples for 5 min at 1000 rpm (200  g). Remove supernatant. 7. Resuspend pellet in 0.2 ml culture medium.

Fig. 1 Skin biopsy punch. (Left picture) The skin around the biopsy is stretched perpendicular to the lines of least skin tension. (Middle picture) The biopsy punch instrument is pressed again the skin and rotated downward. (Right picture) The skin biopsy is collected in a 1.5-ml Eppendorf tube

58

Ngoc-Hien Du and Steven A. Brown

Fig. 2 Examples of a biopsy before and after digestion. (a and c) Before digestion, the biopsy shows sharp and clear edges. (b and d) Four hours after digestion, the edges are dissolved

8. Place all in the center of a fresh 4 cm dish. Overlay fragments with a Millicell Cell Culture insert (Fig. 4). Prior to use, “feet” should be cut from the Millicell Cell Culture insert so that the insert presses tissue fragments against the bottom of the plate (Fig. 3). 9. Add 1.5 ml of the same medium to the interior of the insert, 0.5 ml to the exterior. Spread medium by gently inclining plate in each direction (Fig. 4). 10. Incubate plates for about 2 weeks at 37  C/5% CO2 (see Note 8). Fig. 5a, b show examples of fibroblasts outgrowing from a cultured skin biopsy. 11. Change medium every 3–4 days. Change medium by aspirating medium from the side of the Millicell Cell Culture insert, then from the top of the insert, and adding new medium in reverse order, to avoid that the insert floats away from the biopsy (see Note 9). 12. One week after biopsy, amphotericin B can be omitted from growth medium.

Measuring Circadian Rhythms in Human Cells

59

Fig. 3 Preparation of a Millicell Cell Culture insert. (Left panel) A foot of a Millicell Cell Culture insert is indicated by a white arrow. Each insert has 3 feet. (Right panel) One of the feet is removed

Fig. 4 Culturing skin biopsy with a Millicell Cell Culture insert. (a) The skin biopsy is placed in the middle of a 35-mm cell culture dish. (b and c) The skin biopsy is overlaid with a Millicell Cell Culture insert. (d) The skin biopsy after 4 days in culture with the Millicell Cell Culture insert

60

Ngoc-Hien Du and Steven A. Brown

Fig. 5 Examples of fibroblasts outgrew from the skin biopsy. (a and b) Four days after in culture, fibroblasts are seen outgrowing from the skin biopsy. (c and d) Examples of typical fibroblasts’ morphology

13. Cells are ready to be harvested when the combined volume of their foci is about 1/2 the volume of the plate to which they will be amplified (see Note 10). 14. Trypsinize and replate cells as for normal cells. (a) Aspirate medium and rinse plates with warm PBS. (b) Cover with 0.4 ml of warm Trypsin–EDTA 0.5% solution. (c) Incubate for 3 min at 37  C, or until cells detach. (d) Resuspend in 2 ml of culture medium. (e) Dilute into 10 ml of culture medium in a 15-ml falcon tube, spin for 5 min at 1000 rpm (200  g). Remove supernatant. Resuspend cells in 1 ml culture medium (see Note 11). (f) Place cells into a well of a 12-well plate. (g) Incubate cells in incubator until they reach confluence (4–7 days). 15. Amplify cells each time they reach confluence, splitting 1:2 or 1:3, until they reach the equivalent of three 4-cm dishes (see Notes 12 and 13). Fig. 5c, d show examples of typical fibroblasts’ morphology.

Measuring Circadian Rhythms in Human Cells

61

16. Freeze cells in culture medium + 10% DMSO, as normally. Use at least one 4-cm or one 6-cm dish per vial. 17. In addition to human subjects, we have tested the described method in migratory birds, spiny mice, reindeers, bats, and bears [19]. Culturing of avian cell lines requires coating the cell culture dish with final 0.2% gelatin to enhance cell attachment. 3.2 Simultaneous Culturing of Hundreds of Lines 3.2.1 Thaw Cells Without Centrifugation

1. Thaw a cryotube containing 1 ml of frozen cells (e.g., from a well of a 6-well plate) in water bath, 37  C, remove from the water bath when there is still a small piece of frozen part in side (see Note 14). 2. Wipe water from outside of the tube; use 70% ethanol to wipe around the cap to disinfect. 3. Mix 1 ml of thawed cells with 8 ml of culture medium. 4. Distribute 3 ml per well of a 6-well plate. This gives some cells to work with and some to freeze without increasing too much passage number. 5. Change medium next day. This step is important to remove DMSO in the freezing medium.

3.2.2 Freeze Cells Without Centrifugation from 6-Well Plate

1. Wash cells with 2 ml/well of PBS. 2. Wash cells with 1 ml/well of TrypLE. 3. Add 50 μl/well of TrypLE, tap the plate to distribute TrypLE over the well. 4. Incubate at 37  C for 5–10 min (see Note 15). 5. Add 500 μl of culture medium to each well. 6. Add 500 μl of 2 freezing medium to each well, mix well and put in a cryotube (see Note 16). 7. Put the tubes in a freezing box or a Styrofoam box at
80  C, overnight. 8. Next day transfer tubes to liquid nitrogen tank.

3.3 Production of Lentiviral Vector by Calcium Phosphate Transfection

Note that other transfection methods are available and work fine. We use calcium phosphate transfection because it is cheaper compared to other transfection reagents, and it works well in HEK293T cells. This protocol is adapted from a protocol by the Trono lab [20]. 1. Maintain 293T cells in culture medium in 10-cm tissue culture dishes. Split at a ratio of 1:4 to 1:10 using diluted trypsin, three times per week (e.g., every Monday, Wednesday, and Friday, see Note 17).

62

Ngoc-Hien Du and Steven A. Brown

2. The day before transfection, seed cells at 1–3  106 cells per dish, such that the cells will be 40–50% confluent on the day of transfection, that is, next day (see Note 18). 3. Dissolve all DNA in TE buffer, pH 8.0. Adjust the DNA concentration of all plasmids to 1 mg/ml using the same buffer. 4. Make plasmid mix for each 10-cm plate to be transfected in a 1.5 ml eppendorf tube: (a) 3 μg of pMD2G. (b) 8 μg of psPAX2. (c) 10 μg of transfer vector (pWPT-GFP or pLV6-Bmal-luc, see Note 19). 5. Add up to 250 μl of sterile buffered water, mix well by pipetting. 6. Add 500 μl of 2 HeBS to the 1.5 ml tube prepared in step 5, mix well by pipetting. 7. For each plate to be transfected, put 250 μl of 0.5 M CaCl2 in a 15-ml sterile conical tube. 8. To each 15-ml tube prepared in step 7, slowly transfer, dropwise, the 750 μl of mixture prepared in step 6, while vigorously vortexing. 9. Incubate at room temperature for 20–30 min. A fine white translucent precipitate should form. 10. Add the 1 ml of precipitates from step 9 dropwise to the cells prepared in step 2. Mix by gentle swirling until the medium has recovered a uniformly red color. 11. Early the next morning, aspirate the medium, wash with 10 ml of warm PBS, and gently add 10 ml of fresh warm culture medium. Incubate 24 h (see Note 20). 12. Collect the culture medium (now containing viral particles), store at 4  C. Add 10-ml of fresh warm culture medium and incubate for another 24 h. 13. Collect the culture medium, pool with the one from step 12. Centrifuge for 5 min at 500  g, 4  C, to pellet detached cells and debris. 14. Filter through a 0.45-μm syringe filter (see Note 21). 15. Concentrate supernatant by ultracentrifugation at 24,000 rpm for 2 h in SW-28 rotor. Gently discard the supernatant by inversion. Let the tube dry inverted for 5 min. 16. Resuspend the pellet (usually invisible) with desired medium in desired volume by pipetting up and down 50 times (see Note 22). 17. Store virus in aliquots at
80  C.

Measuring Circadian Rhythms in Human Cells

3.4 Generation of Stable Cell Line Expressing Lentiviral Vector

63

We use a customized bioluminometer to measure circadian rhythms of bioluminescence for prolonged periods in cells plated on 24-well plate. Scale accordingly for different culture dish sizes. There are commercially available devices from Actimetrics that record from samples in 35-mm Petri dishes or 24-well plates [18]. 1. The day before infection, split confluent cells 1:2 or 1:3 on a well of a 12-well plate so that the cells will be 50% confluent on the day of infection. 2. On the day of infection, remove medium and replace with same medium supplemented with 8 μg/ml final concentration of protamine sulfate. Add 20–50 μl of 10 concentrated viral particles to medium (see Note 23). 3. After 6–24 h change to fresh medium. If cells look stressed, infection can be done early in the morning, with medium change at the end of the day. 4. After 3 days, change to medium containing final 10 μg/ml of blasticidin antibiotic if using pLV6-Bmal-luc (see Note 24). 5. The next day or when cells become confluent, split the cells 1:3 to black 24-well plate, continue with selection. If desired, cells can be amplified for freezing (see Note 25). 6. When cells reach confluence, conduct measurement.

3.5 Measurement of Cellular Circadian Rhythms

1. Change medium to synchronization medium. Incubate in the incubator for 30 min. 2. Wash cells twice with PBS. 3. Replace medium with 1 ml/well of counting medium. 4. Place cells into real-time luminometer device kept in a cellculture incubator (34.5  C, 5% CO2). 5. Measure for at least 5 days or until rhythms dampen to flatness.

3.6

Data Analysis

In our hands, fibroblasts can show rhythms up to 7 days after synchronization. Signals during the first 24 h after synchronization are usually variable and thus do not reflect the true endogenous cellular rhythms. We therefore conventionally analyze data starting at least 24 h after synchronization. Baseline drift should be removed from raw data before any period calculation, usually by moving average using a window of 24 h around a certain data point, that is, 12 h before and 12 h after that point. If the software LumiCycle Analysis from Actimetrics is used, this is done automatically when choosing their “Base fit” parameter. For period calculation, LumiCycle offers six different methods to find the dominant period length in the data. We conventionally use “Sin fit (damped)” that fits the data to a damped sine wave.

64

Ngoc-Hien Du and Steven A. Brown

Any other software (see description in Chap. 1) that analyses rhythmic data would work fine. The important thing to keep in mind is data are comparable only if they are analyzed using the same methods.

4

Notes 1. Gentamycin prevents replication of mycoplasma and should not be substituted by penicillin/streptomycin. Humans are carriers of mycoplasma, and 1 biopsy in 20 will have it. 2. We found that Advanced DMEM enhance proliferation rate of human skin fibroblasts, as compared to conventional culture medium prepared in Subheading 2.1. 3. We found that TrypLE is better for splitting human skin fibroblasts, especially when processing a lot of samples at the same time. 4. Having the correct pH is extremely important. Below 6.95, the precipitate will not form, while above 7.05, the precipitate will be coarse and the transfection efficiency will be low. It is advisable to prepare a series of tubes with different pH values (e.g., pH 6.90, pH 6.95, pH 7.00, pH 7.05, pH 7.10, and pH 7.15) then perform a pilot transfection to screen for a good one. 5. Tubes should be almost filled to facilitate immersion of biopsy, which adheres easily to dry surfaces. 6. Biopsies in collection medium, on ice, can be kept for several days (4 days tested). They can be shipped on wet ice at this stage, and are still processable after 4 days. They must not be frozen. 7. Concentration of Liberase and time of digestion will vary according to batch of enzyme. Hence, it is advantageous to titrate a large quantity of enzyme at once. The goal is to have, at the end of digestion, whole tissue pieces with dissolved edges, and multiple clumps containing tens to hundreds of cells. If all cells are dispersed, reduce enzyme up to four times or reduce time up to four times or both. If all tissue fragments are whole and sharp-edged, increase accordingly. A good starting point is 4 h of digestion. If in doubt, biopsy can be examined regularly during digestion. 8. After 3–4 days, fibroblasts could be seen growing out from the biopsy. If no cell is seen after a week, it is advised to take another biopsy. 9. After a week, insert can be removed for ease of changing medium. We usually leave it until splitting.

Measuring Circadian Rhythms in Human Cells

65

10. We usually trypsinize the cells when they cover 1/6 to 1/4 the surface of the 4-cm dish, and replate them in a single well of a 12-well plate. Length of culture will vary with the number of viable fibroblast foci. This varies with age of subject (younger ¼ better), and optimization of digestion. 11. Fragments of tissue are now irrelevant: they can come or stay as it happens. The washing step can be omitted from next passaging. 12. Cells can be left at confluence up to 2 days without ill effects. They should not be oversplit. Date and passage number should be noted. Cells will become senescent and useless around passage 20. 13. Cell growth varies a lot among lines. They can be divided roughly into 3 types: fast grower (passaging every 5–7 days), medium (10–14 days), and slow (>14 days). 14. Do not over thaw the cells since this reduces viability. It is advised to thaw not more than 4 vials/time to avoid overthawing. 15. Human skin fibroblasts are quite sticky, it might take 10 min to detach them all. Do not incubate for more than 15 min. 16. It is advised to work with no more than 6 wells at a time for this step since cells do not like to be in DMSO for a long time. Once in cryotube, keep the cells on ice until the whole round is finished. At maximum, four plates of 6-well plates could be done at the same time. 17. Frequent passages and keeping the cells as single cells will ensure high transfection efficiency. Do not grow cells beyond 80% confluence. 18. If cells are more than 50% confluent on the day of transfection, it is advised to prepare another dish, since the transfection efficiency drops quickly when cells are too confluent. If cells are work day sleep duration:

Asking the Clock

81

Sleep debt corrected MSF ðMSFsc Þ ¼ MSF  0:5  ðsleep durationf ree day  sleep durationaverage Þ where the weighted average sleep duration is computed by the following formula: (work day sleep duration  number of work days + free day sleep duration  number of free days)/7. If free day sleep duration < work day sleep duration: MSFsc ¼ MSF We consider MSFsc as an extremely practical and readily assessable marker for circadian phenotyping (see Notes 7–9). 4. In addition to assessing chronotype and sleep duration, the MCTQ also allows to estimate the strain that an individual’s circadian system is exposed to within their daily life. While circadian strain can occur at multiple levels [9, 10, 11], the MCTQ allows for the computation of a measure of strain on the behavioral level. The underlying assumption here is that the closer the sleep timing on workdays is to the unrestricted sleep timing on work-free days, the lower the circadian strain. In analogy to travel jet lag, this difference between sleep timing on workdays and work-free days has been called social jet lag [12]. Social jet lag ¼ |MSF-MSW| (see Note 10).

4

Notes 1. It is the circadian phenotype, the clock’s phase of entrainment that is exposed to evolutionary and environmental pressure; this is likely where selection takes place, not at the level of the free-running period in constant conditions. 2. The precision required for a particular study will determine its study instruments. Approximating strain to the circadian system in shift workers to examine associations with health and safety outcomes can be done using questionnaires [6, 13]. In contrast, if exact timing of medication in relationship to circadian phase is necessary, we recommend using biomarkers of circadian phase, such as dim light melatonin onset, for increased precision (e.g., [14, 15]. 3. Depending on the question, one might either assess a circadian phenotype or a measure of circadian strain. Circadian strain can, for example, be computed by social jet lag [12], phasor analysis [16], mid sleep deviations [17], or chronodisruption [18]. Chronotype per se is a bad proxy for circadian strain. 4. A detailed overview of all MCTQ variables and their computation can be found in the online supplement of [8] and

82

Ce´line Vetter et al.

Roenneberg et al. [19], as well as at www.thewep.org/ documentations/mctq/item/mctq-variables. 5. The MCTQ does not assess sleep fragmentation, quality, or disorders, nor does it query chronicity of sleep problems. Most versions also do not ask about daytime naps. 6. Individuals suffering from sleep disorders might not be able to report sleep-and-wake behavior adequately [20]. In populations with a high prevalence of sleep disorders, we thus recommend an additional, extensive sleep disorder screening in addition to the phenotyping (to allow subgroup analyses or statistical adjustment for those effects) and/or the usage of preference-based circadian phenotype assessments such as the Morningness/Eveningness Questionnaire [21]. 7. A study of more than 2,500 participants in Europe showed that this sleep-timing based measure of circadian phenotype correlated well (r ¼ 0.73) with the score [22] from the preference assessment counterpart of the MCTQ, the Morningness/ Eveningness Questionnaire (MEQ, [21]), one of the most widely used preference measures in circadian phenotyping. A Japanese study with 450 participants reported a similarly high correlation between the MEQ Score and the MSF (r ¼ 0.65) and the MSFsc (r ¼ 0.61) [23]. The external validity of the MSFsc measure with its analog derivations from actigraphy and sleep logs were moderately high, both in non-shift working adults (r ¼ 0.66 and 0.80, respectively [24]; r ¼ 0.55 [7] and among shift workers (r ¼ 0.74 and 0.78, respectively) [25]. A number of studies examined the association between sleeptiming based circadian phenotypes and physiological phase markers. For example, Brown et al. showed that the period length of circadian clock gene expression in human fibroblasts varies systematically with the MCTQ-based phenotype assessment [26]. Nova´kova´ et al. [27] reported a significant association between the dim light melatonin onset (DLMO)— currently the gold standard marker for circadian phase [28, 29] and the MCTQ phenotype. The Japanese MCTQ validation study [23] reported a moderately strong association between DLMO and MSFsc (r ¼ 0.54), and recently, Kantermann et al. [30] reported a correlation of r ¼ 0.68. Similar associations between sleep timing and DLMO have been reported before, but using sleep logs rather than one-time questionnaires [31, 32]. We have observed a high level of test–retest reliability of the MSFsc parameter, in a study of 19 healthy adults (r ¼ 0.86 [7]). 8. Categorizations (such as into early, intermediate, or late types) based on the MCTQ-chronotype parameter MSFsc are arbitrary. MSFsc is a time point and thus, by definition, continuous. If using relative categorization, it is important to keep in mind that results are not necessarily comparable across studies. For

Asking the Clock

83

interpretation, it is thus imperative to report mean MSFsc and its variability per category. 9. We believe that sleep timing reflects the systems current state and will thus only partially reflect trait-like characteristics. In contrast, diurnal preference of behaviors seems to be a more stable trait [33, 34], and might thus be more useful to study genetic underpinnings of variability in circadian phenotypes (see [35–37] for recent examples) than, for example, more variable measures of sleep timing that are influenced by work schedules [5, 38], light exposure [39–42] and season [42–44], age [45, 46], and so on. 10. We usually compute absolute social jet lag, that is, the absolute amount of difference between MSF and MSW. However, the directionality might be informative, especially when examining the differential effects of social jet lag and sleep deprivation on early versus late chronotypes, so one may want to differentiate between negative and positive values. References 1. Roenneberg T, Wirz-Justice A, Merrow M (2003) Life between clocks: daily temporal patterns of human Chronotypes. J Biol Rhythm 18:80–90 2. Levandovski R, Sasso E, Hidalgo MP (2013) Chronotype: a review of the advances, limits and applicability of the main instruments used in the literature to assess human phenotype. Trends Psychiatry Psychother 35:3–11 3. Roenneberg T, Pilz LK, Zerbini G, Winnebeck EC (2019) Chronotype and social jetlag: A (self-) critical review. Biology (Basel) 8(3):pii E54. https://doi.org/10.3390/ biology8030054 4. Smith PA, Brown DF, Di Milia L, Wragg C (1993) The use of the circadian type inventory as a measure of the circadian constructs of vigour and rigidity. Ergonomics 36:169–175 5. Roenneberg T (2012) What is chronotype? Sleep Biol Rhythms 10:75–76 6. Juda M, Vetter C, Roenneberg T (2013) The Munich ChronoType questionnaire for shiftworkers (MCTQShift). J Biol Rhythm 28:130–140 7. Ghotbi N, Pilz LK, Winnebeck EC, Vetter C, Zerbini G, Lenssen D, Frighetto G, Salamanca M, Costa R, Montagnese S, Roenneberg T (2019) The μMCTQ: an ultra-short version of the Munich ChronoType Questionnaire. J Biol Rhythms 748730419886986. https:// doi.org/10.1177/0748730419886986

8. Roenneberg T, Allebrandt K, Merrow M, Vetter C (2012) Social jetlag and obesity. Curr Biol 22:939–943 9. Qian J, Scheer FAJL (2016) Circadian system and glucose metabolism: implications for physiology and disease. Trends Endocrinol Metab 27:282–293 10. Roenneberg T, Merrow M (2016) The circadian clock and human health. Curr Biol 26: R432–R443 11. Vetter C (2018) Circadian disruption: what do we actually mean? Eur J Neurosci. https://doi. org/10.1111/ejn.14255. 12. Wittmann M, Dinich J, Merrow M, Roenneberg T (2006) Social jetlag: misalignment of biological and social time. Chronobiol Int 23:497–509 13. Barton J, Spelten E, Totterdell P, Smith L, Folkard S, Costa G (1995) The standard Shiftwork index: a battery of questionnaires for assessing shiftwork-related problems. Work Stress 9:4–30 14. Wittenbrink N, Ananthasubramaniam B, Mu¨nch M, et al (2018) High-accuracy determination of internal circadian time from a single blood sample. J Clin Invest 128(9):3826– 3839. https://doi.org/10.1172/JCI120874 15. Laing EE, Mo¨ller-Levet CS, Poh N, Santhi N, Archer SN, Dijk DJ (2017) Blood transcriptome based biomarkers for human circadian phase. Elife 6:e20214. https://doi.org/10. 7554/eLife.20214

84

Ce´line Vetter et al.

16. Rea MS, Bierman A, Figueiro MG, Bullough JD (2008) A new approach to understanding the impact of circadian disruption on human health. J Circadian Rhythms 6:7 17. Fischer D, Vetter C, Roenneberg T (2016) A novel method to visualise and quantify circadian misalignment. Sci Rep 6:38601 18. Erren TC, Reiter RJ (2009) Defining chronodisruption. J Pineal Res 46:245–247 19. Roenneberg T, Keller LK, Fischer D, Matera JL, Vetter C, Winnebeck EC (2015) Chapter twelve—human activity and rest in situ. In: Amita S (ed) Methods in enzymology. Academic Press, Cambridge, Massachusetts, pp 257–283 20. Suh S, Ryu H, Kim S, Choi S, Joo EY (2017) Using Mid-sleep time to determine chronotype in young adults with insomnia-related symptoms. Sleep Med Res 8(2):107–111 21. Horne JA, Østberg O (1976) A self-assessment questionnaire to determine morningnesseveningness in human circadian rhythms. Int J Chronobiol 4:97–110 22. Zavada A, Gordijn M, Beersma D, Daan S, Roenneberg T (2005) Comparison of the Munich Chronotype questionnaire with the ¨ stberg’s Morningness-Eveningness Horne-O score. Chronobiol Int 22:267–278 23. Kitamura S, Hida A, Aritake S, Higuchi S, Enomoto M, Kato M, Vetter C, Roenneberg T, Mishima K (2014) Validity of the Japanese version of the Munich ChronoType questionnaire. Chronobiol Int 31:845–850 24. Peres I, Vetter C, Blautzik J, Reiser M, Po¨ppel E, Meindl T, Till R, Gutyrchik E (2011) Chronotype predicts activity patterns in the neural underpinnings of the motor system during the day. Chronobiol Int 28:883–889 25. Vetter C, Fischer D, Matera JL, Roenneberg T (2015) Aligning work and circadian time in shift workers improves sleep and reduces circadian disruption. Curr Biol 25:907–911 26. Brown SA, Kunz D, Dumas A, Westermark PO, Vanselow K, Tilmann-Wahnschaffe A, Herzel H, Kramer A (2008) Molecular insights into human daily behavior. Proc Natl Acad Sci 105:1602–1607 27. Nova´kova´ M, Sla´dek M, Sumova´ A (2013) Human Chronotype is determined in bodily cells under real-life conditions. Chronobiol Int 30:607–617 28. Arendt J (1998) Melatonin and the pineal gland: influence on mammalian seasonal and circadian physiology. Rev Reprod 3:13–22

29. Arendt J (2006) Melatonin and human rhythms. Chronobiol Int 23:21–37 30. Kantermann T, Sung H, Burgess HJ (2015) Comparing the morningness-eveningness questionnaire and Munich ChronoType questionnaire to the dim light melatonin onset. J Biol Rhythm 30:449–453 31. Martin SK, Eastman CI (2002) Sleep logs of young adults with self-selected sleep times predict dim light melatonin onset. Chronobiol Int 19:695–707 32. Burgess HJ, Savic N, Sletten T, Roach G, Gilbert SS, Dawson D (2003) The relationship between the dim light melatonin onset and sleep on a regular schedule in young healthy adults. Behav Sleep Med 1:102–114 33. Barclay NL, Rowe R, O’Leary R, Bream D, Gregory AM (2016) Longitudinal stability of genetic and environmental influences on the association between diurnal preference and sleep quality in young adult twins and siblings. J Biol Rhythm 31:375–386 34. Broms U, Pitk€aniemi J, B€ackmand H, Heikkil€a K, Koskenvuo M, Peltonen M, Sarna S, Vartiainen E, Kaprio J, Partonen T (2014) Long-term consistency of diurnal-type preferences among men. Chronobiol Int 31:182–188 35. Jones SE, Tyrrell J, Wood AR, Beaumont RN, Ruth KS, Tuke MA, Yaghootkar H, Hu Y, Teder-Laving M, Hayward C et al (2016) Genome-wide association analyses in 128,266 individuals identifies new Morningness and sleep duration loci. PLoS Genet 12:e1006125 36. Hu Y, Shmygelska A, Tran D, Eriksson N, Tung JY, Hinds DA (2016) GWAS of 89,283 individuals identifies genetic variants associated with self-reporting of being a morning person. Nat Commun 7:10448 37. Lane JM, Vlasac I, Anderson SG, Kyle SD, Dixon WG, Bechtold DA, Gill S, Little MA, Luik A, Loudon A et al (2016) Genome-wide association analysis identifies novel loci for chronotype in 100,420 individuals from the UK biobank. Nat Commun 7:10889 38. A˚kerstedt T (2003) Shift work and disturbed sleep/wakefulness. Occup Med (Lond) 53:89–94 39. Roenneberg T, Kumar CJ, Merrow M (2007) The human circadian clock entrains to sun time. Curr Biol 17:R44–R45 40. Vetter C, Juda M, Lang D, Wojtysiak A, Roenneberg T (2011) Blue-enriched office light competes with natural light as a zeitgeber. Scand J Work Environ Health 37:437–445 41. Wright KP, McHill AW, Birks BR, Griffin BR, Rusterholz T, Chinoy ED (2013) Entrainment

Asking the Clock of the human circadian clock to the natural light-dark cycle. Curr Biol 23:1554–1558 42. Stothard ER, McHill AW, Depner CM, Birks BR, Moehlman TM, Ritchie HK, Guzzetti JR, Chinoy ED, LeBourgeois MK, Axelsson J et al (2017) circadian entrainment to the natural light-dark cycle across seasons and the weekend. Curr Biol 27(4):508–513 43. Kohsaka M, Fukuda N, Honma K, Honma S, Morita N (1992) Seasonality in human sleep. Experientia 48:231–233 44. Allebrandt KV, Teder-Laving M, Kantermann T, Peters A, Campbell H, Rudan I, Wilson JF, Metspalu A, Roenneberg

85

T (2014) Chronotype and sleep duration: the influence of season of assessment. Chronobiol Int 31:731–740 45. Crowley SJ, Van Reen E, LeBourgeois MK, Acebo C, Tarokh L, Seifer R, Barker DH, Carskadon MA (2014) A longitudinal assessment of sleep timing, circadian phase, and phase angle of entrainment across human adolescence. PLoS One 9:e112199 46. Roenneberg T, Kuehnle T, Pramstaller PP, Ricken J, Havel M, Guth A, Merrow M (2004) A marker for the end of adolescence. Curr Biol 14:R1038–R1039

Chapter 7 Simple Kinetic Models in Molecular Chronobiology J. Patrick Pett, Pa˚l O. Westermark, and Hanspeter Herzel Abstract Circadian rhythms are constituted by a complex dynamical system with intertwined feedback loops, molecular switches, and self-sustained oscillations. Mathematical modeling supports understanding available heterogeneous kinetic data, highlights basic mechanisms, and can guide experimental research. Here, we introduce the basic steps from a biological question to simple models providing insight into generegulatory mechanisms. We illustrate the general approach by three examples: modeling decay processes, clock-controlled genes, and self-sustained oscillations. Key words Mathematical model, Differential equation, Gene regulation, Feedback, Oscillation, Circadian clock

1

Introduction There has been a long tradition for mathematical modeling in chronobiology [1–3]. In a sense, cartoons of mechanisms already constitute models. Mathematical modeling is one particular strong discipline transcending sciences and involving precise computations of causal relations. We attempt to exemplify when and how modeling may be useful and worthwhile in molecular chronobiology of the cell. As a starting point we list ten steps to construct and analyze kinetic models associated with time-resolved data. More extensive introductions to mathematical modeling can be found in the excellent textbooks of Kaplan and Glass [4], Segel [5], Murray [6], Goldbeter [7], Cornish-Bowden [8], Heinrich and Schuster [9], and Ingalls [10]. We apply modeling concepts to the molecular biology of circadian rhythm generation. In the discussion we list some successful simple kinetic models in chronobiology. We suggest to keep the following ten points in mind if a data-based model is developed and studied:

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_7, © Springer Science+Business Media, LLC, part of Springer Nature 2021

87

88

J. Patrick Pett et al.

1. What biological question is addressed, and how may quantitative modeling help? Often, a model can be used to check the self-consistency of proposed mechanisms: It can point to necessary and sufficient conditions for the observed dynamics. If multiple mechanisms are debated, model selection techniques can help to identify the most plausible explanation. Model predictions can then aid the design of informative follow-up experiments. 2. What is a minimal set of dynamic variables? Models are simplifications, and the task is to focus on the heart of the matter (which depends on the biological question addressed), while ignoring phenomena of lesser relevance. Consequently, a small number of dynamical variables (e.g., mRNA and protein levels, metabolites) should be selected that vary in the time span of interest and that are accessible to measurements. 3. What parameters control the dynamics? Quantitative models contain time-varying variables for which causal relationships are defined. But they also contain parameters, which are usually constants with known or unknown values (although time-varying parameters are possible as well). To ensure reproducibility, experimental conditions should be controlled carefully: Corresponding parameters such as temperature, ATP levels, etc. should be approximately constant in the course of the experiment. Controlled parameter variations in experiments can be compared to parameter variations in the mathematical model. 4. Quantify uncertainties, measurement errors, missing components. Models in molecular biology typically represent only subsystems. Other parts are either out of the scope or more difficult to quantify experimentally. Processes that are not directly part of the model can, however, be included as slow parameter variations or random noise. Here, time-scale separation might be useful. 5. Formulate simple kinetic equations. Many kinetic processes might be described by linear terms. An example is degradation, which often occurs with a rate roughly proportional to the abundance of a given molecular species. Simple nonlinearities arise via mass-action kinetics or enzymatic reactions (e.g., the Michaelis–Menten equation). Note that the choice of the kinetic terms might be different in models of gene regulation [11], signaling [12], and metabolism [8]. A sound approach is to use equations that have parameters inferable from the data at hand (e.g., fold changes), while still being based on physicochemical first principles [13, 14]. 6. Solve equations for different initial conditions.

Simple Kinetic Models

89

If the number of variables is small and if equations are mostly linear, analytical solutions are possible as described in textbooks mentioned above. More often, numerical solutions are necessary. In both cases, convenient software packages such as Matlab, Mathematica, and XPPAUT [15] are available, all three having add-on libraries that support export into standardized formats such as SBML [16]. 7. Study the sensitivity of the model by parameter variations. The robustness of the outcome of simulations should be tested by varying parameters of the model. In many cases models can be simplified by non-dimensionalization [17]. In particular, if only relative measurements are available, normalizations of concentrations can reduce the number of parameters. 8. Are there qualitative changes due to parameter variations (molecular switches, bifurcations)? The outcome of model simulations might depend sensitively on parameter values. For example, cell cycle check points or cell fate decisions can be described as kinetic switches. Such switches can be represented in models by ultrasensitivity or bistability. Parameter variations can also lead to an onset of self-sustained oscillations (“Hopf bifurcation”). 9. Compare simulations critically with available data. First of all, models should represent the basic dynamical features such as bistability or oscillations. In some cases, enough data are available to fit model simulations directly to measured data. Here goodness-of-fit methods such as likelihood-ratio tests or Akaike criteria can be applied for parameter fitting and model selection. However, if this is not possible, semiquantitative models that capture basic dynamical features can contribute crucial insights (see, e.g., [18]). 10. Formulate the biological insights and the limitations of the models. A common aphorism claims: “All models are wrong, but some are useful.” Models represent only cartoons of complex biomedical systems pointing to the most essential features. Thus, it is important to understand which questions are within the scope of a particular model at hand.

2

Modeling Degradation Processes mRNA expression is controlled by production and degradation of transcripts. By blocking transcription via actinomycin D or by mRNA labeling, one can monitor decay processes. Figure 1a shows a hypothetical decay curve of a concentration x(t). Here x(t) represents the amount of mRNA of a gene of interest. Even though

90

J. Patrick Pett et al.

5

4.4 4.2 4.0 0

1

2

5

6

4

5

6

30 20

80 70 60

4

40

D

50

3

time [h]

0

10

100

3.8

6

simple model extended model

90

3.6

log(% remaining mRNA) 4

% remaining mRNA

3

time [h]

30

% remaining mRNA

2

40

C

1

3.4

90 80 70 60 50

% remaining mRNA

40 30 0

4.6

B 100

A

0

1

2

3

4

5

6

0

1

time [h]

2

3

time [h]

Fig. 1 A simple model of mRNA degradation. (a) Simulated data with added noise. (b) Parameter estimation by linear regression on the log-scale. (c) Comparison of the simple model with an extended model incorporating a production rate. (d) The extended model leads to a non-zero steady state, while the simple model results in decay until no mRNA is abundant anymore

degradation processes are highly complex and might be modulated by RNA structure formation, miRNAs, and RNA binding proteins, in many cases rate limiting steps may lead to an exponential decay described as a first-order reaction (absolute degradation rate proportional to mRNA abundance). The mathematical translation is this differential equation: dx ¼ d  x dt

ð1Þ

Simple Kinetic Models

91

a linear first-order differential equation with the degradation rate d with unit 1/h. Using separation of variables or Laplace transformation, one directly obtains the well-known solution: xðtÞ ¼ xð0Þ  e dt

ð2Þ

Plotting the logarithm of x(t) as illustrated in Fig. 1b, the unknown parameter d can be estimated using linear regression. In our example the regression leads to the estimate d^ ¼ 0:14=h corresponding to a half-life t1/2 ¼ 4.95 h. The model with just a single parameter seems to fit the data reasonably well. However, the values for later times are systematically above the regression line. Thus a decay to a plateau C might be considered as an alternative model: xðtÞ ¼ xð0Þ  e dt þ C

ð3Þ

Such a decay to a constant level can be described by a model with basal production p and degradation: dx ¼pdx dt

ð4Þ

Asymptotically, the solution approaches a steady state xss ¼ p/d by exponential relaxation with rate d. Thus, the equilibrium is governed by the ratio of production rate p and decay rate constant d. The time constant of relaxation is given by the rate constant d. This underlines the essential role of degradation kinetics. The regression curve of the extended model fits the data very well (compare Fig. 1c) leading to estimates d^ ¼ 0:6= h and C^ ¼ 39:9. Note that the corresponding estimated half-life of t 1=2 ¼ 1:15 h is considerably smaller compared to fitting a purely exponential decay to zero. This leads to the topic of model selection. Obviously both models represent the data reasonably well. The first model has just a single parameter d, whereas the second model has two parameters p and d. Thus it is not surprising that the larger model fits the data better. Here, criteria are needed that take the number of model parameters and the fitting error into account. For example, the Akaike criterion or Likelihood-ratio tests based on Chi-square values can be applied for model selection [19]. Furthermore, a simple experiment can be suggested to compare both models: If the initial condition x(0) is below the steady state p/d, we expect an increase of x(t), inconsistent with a pure exponential decay (see Fig. 1d). Note that in many large-scale studies, production and degradation models are exploited to quantify mRNA and protein half-lives [20–22]. Using mathematical models, production rates have been successfully computed from steady state levels and from measured decay rates [21].

92

3

J. Patrick Pett et al.

Modeling Clock-Controlled Gene Expression Typically, thousands of mRNA abundances oscillate in a given mammalian tissue [23]. In many cases, transcription factors such as BMAL1 activate transcription with a circadian period of about 24 h. This assumption leads to a generalization of the model discussed above:    dx 2π t dx ð5Þ ¼ p  1 þ A  sin τ dt Here, the production rate p is modulated with an amplitude A  1 and a period τ of about 24 h, an example of a time-varying parameter (see Point 3 in the introduction). Using, for example, the “variations of constants” method or the Fourier transform, the differential equation can be solved analytically (see, e.g., Korencˇicˇ et al. [24], Supplement S3.3 or [25]). After some transient time the solution approaches   2π xðtÞ ¼ C þ a  sin ϕ ð6Þ τ i.e., the mRNA level oscillates periodically around its mean C ¼ p/d. Figure 2a shows simulated target gene expression profiles. Interestingly, the phase shift ϕ and the relative amplitude a/C depend strongly on the half-life t 1=2 ¼ ln ð2Þ=d (see Fig. 2b, c). For short-lived gene transcripts the mRNA is almost in phase with the transcriptional modulation and the relative amplitude is large. Long-lived gene transcripts have delays approaching 6 h but small amplitudes as lifetimes increase. Such dependencies have indeed been found for many clock-controlled genes (Martelot et al. [26], Figure 6C). These insights highlight that some model predictions are quite general: They transcend, and thus do not require, explicit fits to data. In this case, the model with rhythmic transcription implies that transcripts of clock-controlled genes have phase shifts up to 6 h and small relative amplitudes for long-lived mRNAs. Significant deviations from these limits would suggest additional regulatory mechanisms such as combinatorial regulation by multiple transcription factors or post-transcriptional regulation such as periodically modulated mRNA decay rates. Cases like this have been indeed analyzed recently [14, 24, 25]. It turns out that combinatorial transcriptional regulation and rhythmic degradation can lead to almost any phase. In general, an intuitive “phase-vector model” describes synergistic effects of multiple periodic regulations well [14, 25, 27]. Multiple rhythmic regulatory processes are also relevant for glucose homeostasis [28], cell cycle regulation [29], and flux regulation in metabolic networks [30]. Interestingly, out-of-phase

93

0.5

1.0

1.5

d = 0.15 / t 1/2 = 4.6 h d = 0.8 / t 1/2 = 0.9 h

0.0

mRNA abundance

A

2.0

Simple Kinetic Models

0

12

24

36

48

72

60

time [h]

C

0.8 0.6 0.4

rel. amplitude

4 3 2 1 0

phase delay f [h]

5

1.0

B

0

2

4

6

8

10

0

half−life t 1/2

2

4

6

8

10

half−life t1/2

Fig. 2 Model of clock-controlled gene expression. (a) Oscillations in mRNA abundance generated under two different parameter choices for the degradation rate (relative amplitudes plotted). (b) Phase delay ϕ depending on the half-life. (c) Relative amplitude depending on the half-life. In (b) and (c) the two parameter choices shown in (a) are also indicated by blue lines

activators or in-phase activators and inhibitors can also generate 12-h harmonics from 24-h rhythms [14]. Such 12-h oscillations have been described in high-resolution expression profiles [31]. This mechanism of harmonics generation by nonlinearities could explain the origin of tidal rhythms from a circadian clock [14] and other ultradian rhythms [32, 33].

4

Modeling Self-sustained Oscillations (Limit Cycles) The oscillations described in the section above were driven by periodic transcription given a-priori. However, modeling has also helped understanding the basic mechanisms of rhythm generation by the core clocks of various organisms. In a pioneering study,

94

J. Patrick Pett et al.

Hardin et al. [34] found that the protein product of the Drosophila Period gene represses its own expression. Such a negative feedback loop can be described by the classical Goodwin oscillator [35]. We discuss a simplified version where production parameters have been removed via non-dimensionalization ([36], Supplement S5 Appendix): dx 1  dx  x ¼ dt 1 þ ð2zÞh

ð7Þ

dy ¼ x  dy  y dt

ð8Þ

dz ¼ y  dz  z dt

ð9Þ

Here, x represents a clock gene transcript level such as Period2 mRNA abundance, y represents the corresponding cytoplasmic protein, and z stands for a nuclear inhibitor complex. The only nonlinear term in the first equation models a switch-like inhibition of transcription. For h ¼ 10 (which we will assume hereafter) and z ¼ 1/4, repression can be neglected since (1/2)10 is just 1/1024. For z ¼ 1, however, the inhibition is almost perfect since 210 ¼ 1024. This clearly illustrates the self-evident oversimplification of the model. We do not discuss here other core clock genes, phosphorylations, complex formations, epigenetic regulations, or mRNA localization. Still some insight can be gained from the Goodwin model: For vanishing inhibition the equations imply a steady state xss ¼ 1/dx, yss ¼ 1/(dx  dy), and zss ¼ 1/(dx  dy  dz). The approach to this steady state is governed solely by the degradation rates. If transcription starts from the initial condition x ¼ y ¼ z ¼ 0, the variables x, y, and z subsequently approach their steady states as long as z is small and inhibition thus can be neglected. When z exceeds the threshold z ¼ 1/2, transcription is suddenly inhibited and x, y, and z decay one after the other exponentially (due to the linear degradation assumption) according to their degradation rate constants. Then, the cycle starts again. Although this basic oscillator design can be understood without resorting to mathematics, we list some insights that could only be gained by explicit computations below. The oscillations resulting from our version of the Goodwin model are shown in Fig. 3a for relatively large degradation rates corresponding to short half-lives. This implies that the delay between the phases of x and z is relatively small and, consequently, the period of the oscillations is small. By slowing down degradation (compare Fig. 3b) the period increases. This finding illustrates a general principle of negative feedback oscillators: there is a direct

Simple Kinetic Models

95

1.5 0.5

1.0

z y x

0.0

Normalized abundance

A

0

12

24

36

48

60

72

time [h]

1

2

3

4

z y x

0

Normalized abundance

5

B

0

12

24

36

48

60

72

time [h]

Fig. 3 Simulation of the Goodwin oscillator. After a transient, stable limit cycle oscillations are reached. (a) Short-period oscillations for a high degradation rate of d ¼ 0.8 and corresponding half-life of t1/2 ¼ 0.87. (b) Long-period oscillations for a low degradation rate of d ¼ 0.15 and corresponding half-life of t1/2 ¼ 4.6

relation between delays, half-lives, and oscillator period. Under quite generic assumptions one can show rigorously that the delay is between 1/4 and 1/2 of the period (compare Korencic et al. [24], Supplement S2). This implies that for circadian periods a delay of at least 6 h is needed. In Drosophila this delay is achieved by a delayed nuclear import of PER and TIM proteins [37, 38]. In mammals and Neurospora, however, the mechanisms of such long and tightly controlled delays are still debated [18, 39]. Certainly, multiple phosphorylations, complex formations, nuclear translocation, and epigenetic regulations are involved. Despite its simplicity, the Goodwin model illustrates some crucial ingredients for the generation of self-sustained oscillations in core-clock models: (1) negative feedbacks are necessary, (2) switch-like nonlinearities play an essential role, (3) degradation times contribute to the delay, and (4) delays of at least 6 h are needed to achieve 24-h rhythms.

96

J. Patrick Pett et al.

In general, parameters have to be tuned to achieve selfsustained limit cycle oscillations. If degradation is too fast or if the non-linearity, in our Goodwin model version quantified by the Hill coefficient h ¼ 10, is too weak, we get only damped oscillations or no oscillations at all. This highlights another lesson from simple models that could not be reached without explicit mathematical reasoning. Gene-regulatory feedbacks in cells might be able to generate self-sustained oscillations under certain circumstances, but minor changes of parameters (metabolic state, coupling strength etc.) can lead to damped oscillations. Thus the same network can generate multiple dynamic outputs. Such features are conveniently studied using “bifurcation diagrams” as described in textbooks [4, 7, 10]. There are many useful refinements of the basic Goodwin oscillator. It has been shown, for example, that complementary positive feedback loops enhance the capabilities of rhythm generation [40– 43]. Moreover, not all feedback loops are explicitly visible—sequestration (formation of inactive protein complexes) can also form implicit feedback loops. A prominent example is the sequestration of KaiA molecules in the cyanobacterial clock [44].

5

Applications of Minimal Kinetic Models In the final section we discuss some selected applications of minimal models in chronobiology. Even though current models are not quantitative in all details, they can provide valuable insights into regulatory mechanisms of the circadian clock.

5.1 Temperature Compensation via Balancing Reactions

Circadian clocks can be entrained by temperatures since temperature pulses induce phase shifts [45–48]. Nevertheless, the period is almost constant at different environmental temperatures despite temperature dependencies of individual biochemical reactions. Using the Goodwin model and Arrhenius equations, Ruoff and Rensing have shown in 1996 that balances of antagonistic processes can achieve the experimentally observed temperature compensation [49]. More recently, it has been shown that even single phosphorylation and dephosphorylation cycles allow temperature compensation if the Arrhenius-type temperature dependence of the reaction rate is compensated by an opposite exponential dependence of enzyme concentration (see Hatakeyama and Kaneko [50] for details). This mechanism has been illustrated using models of the cyanobacterial clock [50].

5.2 Phosphorylations Control Period via Stability and Nuclear Translocation

Already in early models of the circadian clock, phosphorylation was considered [51, 52]. Motivated by studies related to Familial Advanced Sleep Phase Syndrome (FASPS) [53] and kinase inhibitors [54], the control of circadian periods via multiple phosphorylation has been modeled successfully [18, 55]. It has been found

Simple Kinetic Models

97

that enhanced degradation can generate shorter periods, whereas phosphorylation-assisted nuclear retention can explain long phenotypes [18]. A recent model of a phosphoswitch can even explain circadian temperature compensation [56]. 5.3 Sequestration as Hidden Feedback

6

Modeling reveals that oscillations require negative feedbacks. Furthermore, positive feedbacks can induce switch-like behavior [57]. Sometimes, however, the involved feedback loops are not directly visible in biochemical reaction schemes that are typically drawn to illustrate mechanisms. If enzymes are involved in several processes in a reaction network, sequestration of enzymes can also serve as hidden feedback. Examples are sequestration of XIAP in apoptosis [58] and sequestration of MEK in ERK activation [59] leading to switch-like dynamics. Recently, the sequestration of kinase and phosphatase has been suggested as a mechanism to generate post-translational oscillations [60]. The circadian clock of cyanobacteria can be reconstituted in vitro by mixing KaiA, KaiB, and KaiC proteins with ATP [61]. This leads to temperature-compensated 24 h rhythms of KaiC hexamer phosphorylations [61]. Here the phosphorylation reactions are assisted by KaiA. The relatively slow subsequent phosphorylation and dephosphorylation reaction form the core cycle of the clock, but where is the negative feedback required for an oscillator? How are different KaiC hexamers synchronized? These questions have been addressed by mathematical modeling. It turns out that sequestration of KaiA within complexes of hyperphosphorylated KaiC and KaiB plays an essential role [44, 62, 63]. This mechanism is more clearly illustrated by a minimal model of just six differential equations [64].

Concluding Remarks To summarize, some of the lessons learned from studying minimal models in molecular chronobiology are: 1. Expression levels are controlled by the balance of production and degradation. 2. Delays are largely determined by half-lives of the involved regulators. 3. Long-lived transcripts of clock-controlled genes exhibit small amplitudes and can be delayed by up to 6 h. 4. Combinatorial regulations including rhythmic degradation allow almost any phase and amplitude of clock outputs. 5. Nonlinear combinatorial regulation can generate harmonics (8 and 12 h rhythms).

98

J. Patrick Pett et al.

6. The generation of self-sustained oscillations requires delayed negative feedback loops. 7. Switch-like behavior due to cooperativity and positive feedback loops assist rhythm generation. 8. Temperature compensation can be achieved by balancing reactions, even on the level of single phosphorylation–dephosphorylation cycles. 9. Via protein stability and nuclear translocation phosphorylations can control delays and periods effectively. 10. Sequestration of regulators can provide hidden feedbacks allowing switches and self-sustained oscillations. References 1. Wever R (1965) A mathematical model for circadian rhythms. Circadian Clocks 47: 47–63 2. Winfree AT (1970) Integrated view of resetting a circadian clock. J Theor Biol 28 (3): 327–374 3. Kronauer RE, Czeisler CA, Pilato SF, MooreEde MC, Weitzman ED (1982) Mathematical model of the human circadian system with two interacting oscillators. Am J Physiol Regul Integr Comp Physiol 242 (1): R3–R17 4. Kaplan D, Glass L (2012) Understanding nonlinear dynamics. Springer Science & Business Media, New York 5. Segel LA (1984) Modeling dynamic phenomena in molecular and cellular biology. Cambridge University Press, Cambridge 6. Murray JD (2002) Mathematical biology I: an introduction. Interdisciplinary applied mathematics, vol 17. Springer, New York 7. Goldbeter A (1997) Biochemical oscillations and cellular rhythms. Cambridge University Press, Cambridge 8. Cornish-Bowden A, Ca´rdenas ML (2013) Control of metabolic processes, vol 190. Springer Science & Business Media, New York 9. Heinrich R, Schuster S (2012) The regulation of cellular systems. Springer Science & Business Media, New York 10. Ingalls BP (2013) Mathematical modeling in systems biology: an introduction. MIT Press, Cambridge 11. Bintu L, Buchler NE, Garcia HG, Gerland U, Hwa T, Kondev J, Phillips R (2005) Transcriptional regulation by the numbers: models. Curr Opin Genet Dev 15 (2): 116–124 12. Chen WW, Niepel M, Sorger PK (2010) Classic and contemporary approaches to modeling biochemical reactions. Genes Dev 24 (17): 1861–1875

13. Westermark PO, Kotaleski JH, Bjo¨rklund A, Grill V, Lansner A (2007) A mathematical model of the mitochondrial NADH shuttles and anaplerosis in the pancreatic β-cell. Am J Physiol Endocrinol Metab 292 (2): E373–E393 14. Westermark PO, Herzel H (2013) Mechanism for 12 hr rhythm generation by the circadian clock. Cell Rep 3 (4): 1228–1238 15. Ermentrout B (2002) Simulating, analyzing, and animating dynamical systems: a guide to XPPAUT for researchers and students, vol 14. SIAM, Philadelphia 16. Hucka M, Finney ABBJ, Bornstein BJ, Keating SM, Shapiro BE, Matthews J, Kovitz BL, Schilstra MJ, Funahashi A, Doyle JC, Kitano H (2004) Evolving a lingua franca and associated software infrastructure for computational systems biology: the systems biology markup language (SBML) project. Syst Biol 1 (1): 41–53 17. Buckingham E (1914) On physically similar systems; illustrations of the use of dimensional equations. Phys Rev 4 (4): 345 18. Vanselow K, Vanselow JT, Westermark PO, Reischl S, Maier B, Korte T, Herrmann A, Herzel H, Schlosser A, Kramer A (2006) Differential effects of PER2 phosphorylation: molecular basis for the human familial advanced sleep phase syndrome (FASPS). Genes Dev 20 (19): 2660–2672 19. Maiwald T, Timmer J (2008) Dynamical modeling and multi-experiment fitting with PottersWheel. Bioinformatics 24 (18): 2037–2043 20. Friedel CC, Do¨lken L, Ruzsics Z, Koszinowski UH, Zimmer R (2009) Conserved principles of mammalian transcriptional regulation revealed by RNA half-life. Nucleic Acids Res 37 (17): e115

Simple Kinetic Models 21. Schwanh€ausser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, Chen W, Selbach M (2011) Global quantification of mammalian gene expression control. Nature 473 (7347): 337–342 22. Eden E, Geva-Zatorsky N, Issaeva I, Cohen A, Dekel E, Danon T, Cohen L, Mayo A, Alon U (2011) Proteome half-life dynamics in living human cells. Science 331 (6018): 764–768 23. Zhang R, Lahens NF, Ballance HI, Hughes ME, Hogenesch JB (2014) A circadian gene expression atlas in mammals: implications for biology and medicine. Proc Natl Acad Sci 111 (45): 16219–16224 24. Korencˇicˇ A, Bordyugov G, Rozman D, Golicˇnik M, Herzel H et al (2012) The interplay of cis-regulatory elements rules circadian rhythms in mouse liver. PLoS One 7 (11): e46835 25. Lu¨ck S, Thurley K, Thaben PF, Westermark PO (2014) Rhythmic degradation explains and unifies circadian transcriptome and proteome data. Cell Rep 9 (2): 741–751 26. Le Martelot G, Canella D, Symul L, Migliavacca E, Gilardi F, Liechti R, Martin O, Harshman K, Delorenzi M, Desvergne B, Herr W, Deplancke B, Schibler U, Rougemont J, Guex N, Hernandez N, Naef F, the CycliX consortium (2012) Genome-wide RNA polymerase ii profiles and RNA accumulation reveal kinetics of transcription and associated epigenetic changes during diurnal cycles. PLoS Biol 10 (11): e1001442 27. Ukai-Tadenuma M, Yamada RG, Xu H, Ripperger JA, Liu AC, Ueda HR (2011) Delay in feedback repression by Cryptochrome 1 is required for circadian clock function. Cell 144 (2): 268–281 28. Lamia KA, Storch K-F, Weitz CJ (2008) Physiological significance of a peripheral tissue circadian clock. Proc Natl Acad Sci 105 (39): 15172–15177 29. Eser P, Demel C, Maier KC, Schwalb B, Pirkl N, Martin DE, Cramer P, Tresch A (2014) Periodic mRNA synthesis and degradation co-operate during cell cycle gene expression. Mol Syst Biol 10 (1): 717 30. Thurley K, Herbst C, Wesener F, Koller B, Wallach T, Maier B, Kramer A, Westermark PO (2017) Principles for circadian orchestration of metabolic pathways. Proc Natl Acad Sci USA 114(7):1572–1577. https://doi.org/10. 1073/pnas.1613103114 31. Hughes ME, DiTacchio L, Hayes KR, Vollmers C, Pulivarthy S, Baggs JE, Panda S, Hogenesch JB (2009) Harmonics of circadian gene transcription in mammals. PLoS Genet 5 (4): e1000442

99

32. Locke JCW, Kozma-Bogna´r L, Gould PD, Fehe´r B, Kevei E, Nagy F, Turner MS, Hall A, Millar AJ (2006) Experimental validation of a predicted feedback loop in the multi-oscillator clock of Arabidopsis thaliana. Mol Syst Biol 2 (1): 59 33. Blum ID, Zhu L, Moquin L, Kokoeva MV, Gratton A, Giros B, Storch K-F (2014) A highly tunable dopaminergic oscillator generates ultradian rhythms of behavioral arousal. eLife 3: e05105 34. Hardin PE, Hall JC, Rosbash M (1990) Feedback of the Drosophila Period gene product on circadian cycling of its messenger RNA levels. Nature 343 (6258): 536–540 35. Goodwin BC (1965) Oscillatory behavior in enzymatic control processes. Adv Enzyme Regul 3: 425–437 36. Pett JP, Korencˇicˇ A, Wesener F, Kramer A, Herzel H (2016) Feedback loops of the mammalian circadian clock constitute repressilator. PLoS Comput Biol 12 (12): 1–15 37. Zeng H, Qian Z, Myers MP, Rosbash M (1996) A light-entrainment mechanism for the Drosophila circadian clock. Nature 380 (6570): 129–135 38. Myers MP, Wager-Smith K, RothenfluhHilfiker A, Young MW (1996) Light-induced degradation of TIMELESS and entrainment of the Drosophila circadian clock. Science 271 (5256): 1736 39. Gallego M, Virshup DM (2007) Posttranslational modifications regulate the ticking of the circadian clock. Nat Rev Mol Cell Biol 8 (2): 139–148 40. Meinhardt H (1982) Models of biological pattern formation. Academic, London 41. Tsai TY-C, Choi YS, Ma W, Pomerening JR, Tang C, Ferrell JE Jr (2008) Robust, tunable biological oscillations from interlinked positive and negative feedback loops. Science 321: 126–129 42. Stricker J, Cookson S, Bennett MR, Mather WH, Tsimring LS, Hasty J (2008) A fast, robust and tunable synthetic gene oscillator. Nature 456: 516–519 43. Ananthasubramaniam B, Herzel H (2014) Positive feedback promotes oscillations in negative feedback loops. PLoS One 9: e104761 44. Clodong S, Du¨hring U, Kronk L, Wilde A, Axmann I, Herzel H, Kollmann M (2007) Functioning and robustness of a bacterial circadian clock. Mol Syst Biol 3 (1): 90 45. Zimmerman WF, Pittendrigh CS, Pavlidis T (1968) Temperature compensation of the circadian oscillation in Drosophila pseudoobscura and its entrainment by temperature cycles. J Insect Physiol 14 (5): 669–684

100

J. Patrick Pett et al.

46. Rensing L, Ruoff P (2002) Temperature effect on entrainment, phase shifting, and amplitude of circadian clocks and its molecular bases. Chronobiol Int 19 (5): 807–864 47. Abraham U, Granada AE, Westermark PO, Heine M, Kramer A, Herzel H (2010) Coupling governs entrainment range of circadian clocks. Mol Syst Biol 6 (1): 438 48. Buhr ED, Yoo S-H, Takahashi JS (2010) Temperature as a universal resetting cue for mammalian circadian oscillators. Science 330 (6002): 379–385 49. Ruoff P, Rensing L (1996) The temperaturecompensated Goodwin model simulates many circadian clock properties. J Theor Biol 179 (4): 275–285 50. Hatakeyama TS, Kaneko K (2012) Generic temperature compensation of biological clocks by autonomous regulation of catalyst concentration. Proc Natl Acad Sci 109 (21): 8109–8114 51. Goldbeter A (1995) A model for circadian oscillations in the Drosophila PERIOD protein (PER). Proc R Soc Lond B Biol Sci 261 (1362): 319–324 52. Tyson JJ, Hong CI, Thron CD , Novak B (1999) A simple model of circadian rhythms based on dimerization and proteolysis of PER and TIM. Biophys J 77 (5): 2411–2417 53. Toh KL, Jones CR, He Y, Eide EJ, Hinz WA, Virshup DM, Pta´cˇek LJ, Fu Y-H (2001) An hPer2 phosphorylation site mutation in familial advanced sleep phase syndrome. Science 291 (5506): 1040–1043 54. Lowrey PL, Shimomura K, Antoch MP, Yamazaki S, Zemenides PD, Ralph MR, Menaker M, Takahashi JS (2000) Positional syntenic cloning and functional characterization of the mammalian circadian mutation tau. Science 288 (5465): 483–491 55. Gallego M, Eide EJ, Woolf MF, Virshup DM, Forger DB (2006) An opposite role for tau in

circadian rhythms revealed by mathematical modeling. Proc Natl Acad Sci 103 (28): 10618–10623 56. Zhou M, Kim JK, Eng GWL, Forger DB, Virshup DM (2015) A Period2 phosphoswitch regulates and temperature compensates circadian period. Mol Cell 60 (1): 77–88 57. Ferrell JE (1996) Tripping the switch fantastic: how a protein kinase cascade can convert graded inputs into switch-like outputs. Trends Biochem Sci 21 (12): 460–466 58. Legewie S, Blu¨thgen N, Herzel H (2006) Mathematical modeling identifies inhibitors of apoptosis as mediators of positive feedback and bistability. PLoS Comput Biol 2 (9): e120 59. Legewie S, Schoeberl B, Blu¨thgen N, Herzel H (2007) Competing docking interactions can bring about bistability in the MAPK cascade. Biophys J 93 (7): 2279–2288 60. Jolley CC, Ode KL, Ueda HR (2012) A design principle for a posttranslational biochemical oscillator. Cell Rep 2 (4): 938–950 61. Nakajima M, Imai K, Ito H, Nishiwaki T, Murayama Y, Iwasaki H, Oyama T, Kondo T (2005) Reconstitution of circadian oscillation of cyanobacterial KaiC phosphorylation in vitro. Science 308 (5720): 414–415 62. Qin X, Byrne M, Xu Y, Mori T, Johnson CH (2010) Coupling of a core post-translational pacemaker to a slave transcription/translation feedback loop in a circadian system. PLoS Biol 8 (6): e1000394 63. Brettschneider C, Rose RJ, Hertel S, Axmann IM, Heck AJR, Kollmann M (2010) A sequestration feedback determines dynamics and temperature entrainment of the KaiABC circadian clock. Mol Syst Biol 6 (1): 389 64. Axmann IM, Legewie S, Herzel H (2007) A minimal circadian clock model. Genome Inform 18: 54–64

Part II Genome-Wide Analyses in Circadian Biology

Chapter 8 Searching Novel Clock Genes Using RNAi-Based Screening Bert Maier, Stephan Lorenzen, Anna-Marie Finger, Hanspeter Herzel, and Achim Kramer Abstract RNA interference (RNAi) allows for the selective downregulation of gene expression by neutralizing targeted mRNA molecules and has frequently been used in high-throughput screening endeavors. Here, we describe a protocol for the highly parallel RNAi-mediated downregulation of gene expression in order to search for components involved in circadian rhythm generation. We use lentiviral gene transfer to deliver shRNA expressing plasmids into circadian reporter cells ensuring for efficient and stable knockdown. Circadian rhythms are monitored using live-cell bioluminescence recording of synchronized reporter cells over several days. In addition, we present a new software tool (ChronoStar) for efficient, parallel time-series analysis to extract rhythm parameters such as period, phase, amplitude, and damping. Key words Circadian rhythms, RNA interference, Screen, Luciferase reporter, Time-series analysis, U-2 OS cells, ChronoStar, Live-cell imaging

1

Introduction In mammals, Clock was the first clock gene to be identified. Pioneering studies in Drosophila by Seymour Benzer and his colleagues [1] inspired Joe Takahashi’s laboratory to perform a forward genetic screen (from phenotype to gene), where the progeny of mice treated with the mutagen N-ethyl-N-nitrosourea (ENU) were screened for circadian alterations in behavior [2]. Clock was identified as semidominant mutation that led to long circadian periods and eventually to arrhythmicity in wheel-running activity. While forward genetics allows for unbiased screening, reverse genetics (from gene to phenotype) studies what phenotypes arise as a result of mutation or knockout of specific candidate genes. For example, period and cryptochrome genes have been knocked out (and thus identified as clock components) because of the function of their orthologous genes in Drosophila and Arabidopsis [3–8]. The discovery of RNA interference (RNAi) and the development of rationally designed RNAi libraries at the beginning of this

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_8, © Springer Science+Business Media, LLC, part of Springer Nature 2021

103

104

Bert Maier et al.

century [9] allowed for genome-wide reverse genetic screens, that is, the systematic and parallel silencing of many thousand genes. In mammals, circadian oscillators are present in virtually every cell and even exist in some immortalized cell lines with very similar molecular architecture [10]. Such cells can be equipped with suitable circadian reporters and exploited for large-scale (chemical or genetic) screening endeavors [11–13]. This has several advantages compared to behavioral outputs: (1) the sensitivity for detecting circadian phenotypes is likely to be higher, since loss-of-function genetics results in more severe circadian phenotypes in uncoupled cells [14]; (2) clock genes essential for development or masked by developmental compensation may be uncovered; (3) a high throughput with relative low costs can be achieved; (4) human cell lines can be used, which is more relevant for human health. Here, we describe a detailed protocol for large-scale RNAibased screening for novel clock genes in human U-2 OS osteosarcoma cells—an established cellular circadian clock model. U-2 OS cells were equipped with luciferase-based circadian reporters and circadian phenotypes were measured using live cell bioluminescence imaging for several days (Fig. 1). For long-lasting gene knockdown, we use micro-RNA-based RNAi constructs from a second-generation library of the Hannon and Elledge laboratories [15–18], which are lentivirally delivered and thus stably integrated into the genome. Transfecting siRNA is also possible and has been applied in a genome-wide screen [13].

Fig. 1 Graphical workflow of RNAi-based screen for novel clock genes. The 15-day protocol described here affords 6 days with partial hands-on time including free weekends

RNAi Screening for Novel Clock Genes

105

Fig. 2 Screenshot of ChronoStar time-series analysis software. Samples of a 96-well bioluminescence recording have been loaded, of which well A5 has been selected for display (data table). In the Raw data panel counts per time point (black) and respective nonlinear trend curve (blue) were plotted. Fit data panel shows detrended data (black) and fitted cosine function (blue). The Control panel comprises all adjustable parameters of the analysis

In addition, we introduce a software tool (ChronoStar) for a highly parallel analysis of bioluminescence time-series (Fig. 2). ChronoStar is compatible with various different input formats and automatically fits cosine wave functions to detrended bioluminescence data to calculate rhythm parameters such as period, phase, amplitude, damping as well as the quality of the fit.

106

2

Bert Maier et al.

Materials Prepare all reagents and media under aseptic conditions under laminar flow. Follow all waste disposal and biosafety regulations when disposing of/decontaminating materials. All media and buffers can be stored at 4
C for up to 6 months. Fetal bovine serum, luciferin, protamine sulfate, dexamethasone, and antibiotics should be stored at 20
C. Catalogue numbers were added for convenience. While we recommend using the indicated products, which we tested during protocol establishment, similar products might work equally well.

2.1

Cell Lines

1. HEK293T (ATCC® CRL-3216™). 2. U-2 OS Bmal1:Luc reporter cells [12]. Alternative circadian reporter cells are also possible (e.g., see [13]).

2.2

Reagents

1. Culture medium: (a) 500 mL High Glucose DMEM (Gibco® #11965092). (b) 50 mL FBS (Gibco® #10270-106). (c) 12.5 mL 1 M HEPES, pH 7.3. (d) 5 mL 10.000 U/mL Penicillin/Streptomycin (Gibco® #15140122). 2. Reporter medium: (a) 500 mL High Glucose DMEM, phenol red-free (Gibco® #21063029). (b) 50 mL FBS (Gibco® #10270-106). (c) 5 mL 10.000 U/mL Penicillin/Streptomycin (Gibco® #15140122). (d) 500 μL 10 mg/mL Puromycin (Sigma-Aldrich™ #P9620). (e) Immediately before use add luciferin (PJK #102111) to a final concentration of 250 μM (see Note 1). 3. Opti-MEM (Gibco® #31985). 4. psPAX plasmid (Addgene #12260). 5. pMD2G plasmid (Addgene #12259). 6. GIPZ lentiviral shRNA constructs (Dharmacon™ GIPZ™) (see Note 2). 7. Lipofectamine2000® (ThermoFisher Scientific #11668019). 8. Dexamethasone #D4902).

solubilized

in

EtOH

(Sigma-Aldrich™

RNAi Screening for Novel Clock Genes

107

9. Protamine sulfate (Sigma-Aldrich™ #53597254). 10. Puromycin (Sigma-Aldrich™, #P9620). 2.3

Equipment

1. Bioluminometer (e.g., TopCount™, LumiStar™, Orion II™) (see Note 3). 2. Incubator for bioluminometer (e.g., Galaxy® 170R, New Brunswick). 3. ALPS 50™ V-Manual Heat Sealer (Abgene™). 4. Clear 96-well tissue cultured treated plates (BD Falcon™ #353072). 5. White 96-well tissue culture treated plates (NUNC #136101). 6. 96-well PCR V-bottom microtiter plate (Costar #3363). 7. MultiScreen® HTS filter plates (Millipore #MSFBN6B50). 8. 96-well plate adhesive tape (Roth #EN83.1). 9. Diamond Clear Seal (Thermo Scientific™ #AB0812). 10. Optional: Fluorescence plate reader (e.g., Infinite F200Pro™, TECAN™). 11. Optional: Liquidator™ 96 Manual Pipetting System (Mettler Toledo).

2.4 Software ChronoStar

3

ChronoStar is a program for analysis of time-series data describing damped oscillations overlaid by a nonlinear trend. A typical but not exclusive application is the analysis of biological time-series with damped circadian oscillations (e.g., bioluminescence data). The software runs with a graphical user interface with visualization and real-time adjustment of time-series data and fitting parameters. The program is freely available for download for academic institutions at www.achim-kramer-lab.de/downloads.html.

Methods

3.1 RNAi Screen: General Considerations

Here, we describe a screening protocol in 96-well plate format; depending on the experimenter’s needs, it might be scaled accordingly. Selection of genes for the screening procedure might be unbiased or instructed by experimenter’s background. For the arrayed screen design described here, we recommend including at least 2–3 different RNAi constructs each targeting a gene of interest (see Note 2). Similarly, three to five nontargeting controls (also called nonsilencing shRNA) should be included per 96-well plate to enhance statistical power. Plasmid DNA of shRNA constructs can be prepared by standard mini-preparation from E. coli glycerol stock cultures and should be adjusted to a final concentration of 0.14 μg/μL. To array the shRNA constructs, transfer 20 μL/well of RNAi plasmids to a 96-well PCR V-bottom microtiter plate and

108

Bert Maier et al.

store at 20
C. In addition, psPAX packaging and pMD2G envelope plasmids should be prepared by standard midi-preparation, and their concentration should be adjusted to 1 μg/μL for convenience. 3.2 Day 1: Seeding of Virus-Producing Cells

1. Wash HEK293T cells 1 with PBS and trypsinize for 5–10 min under standard cell culture conditions to detach cells. 2. Harvest cells in fresh culture medium and count in hemocytometer. Adjust concentration of the cellular suspension to 3  105 cells/mL (see Note 4). 3. Seed 100 μL/well (3  104 cells/well) of the cellular suspension into the clear flat well 96-well tissue culture treated plates. Incubate cells under standard cell culture conditions overnight (see Note 5).

3.3 Day 2: Transfection of Virus-Producing Cells

1. Prepare two 15 mL falcon tubes with 2.75 mL Opti-MEM each and per 96-well plate of HEK293T cells. Add 11 μg psPAX packaging and 6.6 μg pMD2G envelope plasmid to tube A. Add 55 μL Lipofectamine 2000® to tube B. Incubate both tubes for ~5 min at room temperature. 2. To prepare the transfection solution use a multichannel pipette to transfer 25 μL/well of solution A into a 96-well PCR V-bottom microtiter plate. Again using a multichannel pipette add 1 μL/well of your arrayed shRNA plasmids to the 96-well PCR V-bottom microtiter plate containing solution A. Use new pipette tips for each column/row of the RNAi screen library to prevent cross-contamination of plasmid DNA. Add 25 μL/well of solution B to the same 96-well V-bottom microtiter plate containing solution A and shRNA plasmids. Again, use new pipette tips for each column/row of the PCR V-bottom microtiter plate (see Note 6). 3. Seal the PCR V-bottom microtiter plate containing the transfection mix with adhesive tape. Vortex, and spin down with 500  g for ~30 s. Incubate transfection mix for 20–40 min at room temperature. 4. Add 50 μL/well transfection mix to HEK293T cells from day 1 without changing or discarding the culture medium (150 μL/well final volume). Incubate cells under standard cell culture conditions overnight.

3.4 Day 3: Replacement of Viral Medium

1. Use a multichannel pipette to aspirate and discard viruscontaining medium from transfected HEK293T cells. Add 150 μL/well fresh culture medium. Use new pipette tips for each column/row of the 96-well plate to avoid crosscontamination of viral supernatant. 2. Incubate cells under standard cell culture conditions overnight (see Note 7).

RNAi Screening for Novel Clock Genes

3.5 Day 4: A. Harvest of Viral Supernatant and Transduction of Reporter Cells

109

1. Place a MultiScreen® HTS filter plate on top of a white 96-well cell culture treated plate. 2. With a multichannel pipette transfer supernatant of transfected HEK293T cells (~150 μL/well) to the MultiScreen® HTS filter plate on top of the white 96-well tissue culture treated plate. Use new pipette tips for each column/row of the 96-well plate containing transfected HEK293T cells in order to avoid cross-contamination of the viral supernatant (see Notes 7 and 8). 3. Filter supernatant through the MultiScreen® HTS filter plate into the white 96-well tissue culture treated plate by spinning both plates on top of each other at 3000  g for 1 min at 4
C. Keep white 96-well tissue culture treated plate with viral supernatant at 4
C until use.

3.6 Day 4: B. Seeding of Reporter Cells

1. Wash U-2 OS Bmal1:Luc reporter cells 1x with PBS and trypsinize for 5–10 min under standard cell culture conditions to detach (see Note 9). 2. Harvest cells in fresh culture medium and count in hemocytometer. Adjust concentration of the cellular suspension to 4  105 cells/mL. Add protamine sulfate to a final concentration of 32 μg/mL. 3. Seed 50 μL/well (2  104 cells/well) of the cellular suspension + protamine sulfate into white 96-well tissue culture treated plates containing the viral supernatant (seeNote 10).

3.7 Day 5: Antibiotic Selection of Transduced Reporter Cells

1. Using a multichannel pipette aspirate and discard supernatant from transduced U-2 OS reporter cells. Use new pipette tips for each column/row of the white 96-well plate in order to prevent cross-contamination with virus. 2. Prepare 150 μL/well fresh culture medium with 10 μg/mL puromycin. Substitute transduced U-2 OS Bmal1:Luc reporter cells with the antibiotic selection medium. Incubate cells for at least 24 h under standard cell culture conditions (see Note 11).

3.8 Day 6: Synchronization and Long-Term Monitoring (See Note 11)

1. Prepare 50 μL/well fresh culture medium with 4 μM dexamethasone. 2. Using a multichannel pipette add 50 μL/well of the dexamethasone solution to white 96-well tissue culture treated plates containing transduced U-2 OS Bmal1:Luc reporter cells. The same pipette tips can be used for all wells. Incubate cells under standard cell culture conditions for 20–30 min to synchronize circadian oscillations.

110

Bert Maier et al.

3. Using a multichannel pipette wash U-2 OS reporter cells 2 with 100 μL/well PBS. The same pipette tips can be used for all wells. 4. Prepare 150 μL reporter medium per well. Transfer reporter medium (including luciferin) to Bmal1:Luc U-2 OS reporter cells. Seal white 96-well tissue culture treated plate with Diamond Seal using the ALPS 50™ V-Manual Heat Sealer. Seal twice at 165
C for 2–3 s each, rotating the plate 180
in the horizontal plane after the first round. Do not place transparent lid back onto the sealed plate. 5. Place sealed white 96-well plate containing transduced U-2 OS Bmal1:Luc reporter cells into bioluminometer placed in a cell culture incubator. Monitor oscillations for at least 4–6 days (see Note 3). 3.9 Time-Series Analysis with ChronoStar Software

1. Open ChronoStar software. 2. Open bioluminescence time-series data in ChronoStar. File ! Open ! select file(s). Data should be formatted in ∗.cs3 format (see Note 12). 3. Select single bioluminescence time-series. 4. Visually inspect data- and fit-quality by scrolling through single traces using up/down arrow keys. The Raw data window displays selected time-series in black and corresponding trends (! trend estimation) in blue. Trend-eliminated time-series (black) and sine wave fits (blue) are displayed in the Fit data window (see Note 13). 5. Select all time-series by clicking at upper-left corner of the table or press ctrl+A. 6. Adjust data analysis settings via the Control panel. Borders— left/right crop your dataset to exclude nonrelevant parts (e.g., initial artifacts or final signal drops due to cell death). Space— data will be analyzed either in linear (Abs) or in logarithmic (Rel) mode. We recommend using the logarithmic mode for screen data evaluation. Detrend—raw data should be detrended before sine-fitting (Average 24) unless already performed by other means (none). AmpRef—set time for amplitude parameter extraction to 24 h to avoid extrapolation. 7. Save and export data analysis. File ! Export Selection ! select path and provide file name. Data will be saved in ∗.xml file format for further processing in spreadsheets, such as Microsoft Excel. You will find a comprehensive table of parameter estimates from analyzed time-series within the Param sheet (see Note 14).

RNAi Screening for Novel Clock Genes

4

111

Notes 1. Some papers report sensitivity of D-luciferin to light, oxygen, and moisture as powder and in solution. Thus, luciferin should be stored in aliquots at 20
C in lighttight boxes and be added to reporter medium only directly before use. Moreover, D-luciferin can be purchased from a large number of suppliers. However, purity in respect to D-luciferin–L-luciferin ratio strongly affects quantum yield and therefore signal intensity [19]. Hence, if using a different luciferin from the one indicated above, initial comparative tests of free samples provided by suppliers are recommended. 2. In our laboratory we use GIPZ lentiviral shRNA constructs because of the following features: (1) they show generally high transfection and transduction efficiencies; (2) shRNA is coexpressed with tGFP; (3) they express a fast-acting selection marker (puromycin); (4) they usually result in efficient knockdown of the gene of interest. However, the RNAi screen described here is not limited to these specific lentiviral shRNA constructs, and the protocol may be adjusted for other transfection-based shRNA or even siRNA libraries. Nevertheless, for every RNAi library of choice, careful consideration should be given to issues such as transfection/transduction efficiency, knockdown efficiency, knockdown stability (i.e., half-life), selectable and fluorescent markers (e.g., to control for success of transfection/transduction). 3. Our screen setup consists of a NXT TopCount with 12 photomultiplier tubes (PMTs) stalled in a custom-made incubator with heating and cooling units. However, other devices might be used as well. Real-time measurements should be performed at ambient CO2 and humidity levels, as both factors will influence corrosion of electronic components at normal cell culture conditions. We recommend a final sampling interval of less than 30 min to obtain high resolution especially important for phase information. The minimal sampling interval duration is critically influenced by the number of PMTs for parallel sampling, integration time (time of PMT on one sample/interval), PMT sensitivity/noise level, signal intensity of the samples, and number of plates monitored in parallel. 4. To save time during cell plating, seeding density can be estimated roughly. For HEK293T cells, one 175 cm2 flask will contain ~6  107 cells, which suffices for seeding in seven 96-well plates with 100 μL/well. For U-2 OS Bmal1:Luc reporter cells, we recommend using one 175 cm2 flask of cells for seeding in 3–4 96-well plates with 50 μL/well.

112

Bert Maier et al.

5. To optimize lentivirus production by HEK293T cells, culture time of virus producing cells should not exceed 2 months. If cells are detaching too easily during transfection/medium exchange, seeding density can be reduced to 2  104 cells/ well to prevent cell culture overgrowth. In general, HEK293T cells should be handled with care because cells may stick loosely to plastic surface of clear 96-well tissue culture plates. 6. For high-throughput approaches (number of 96-well reporter cell plates >4) we recommend to use the Liquidator™ 96 Manual Pipetting System instead of a simple multichannel pipette and columnwise pipetting for ease of handling and time saving. 7. Even though the used lentiviruses are replication-deficient, all transfection samples, transfected cells, and lentiviral preparations have to be handled with care and under adherence to biosafety regulations. Gloves should be worn the entire time, and decontamination should be performed with 70% ethanol, bleach, or Decosept®. 8. The same MultiScreen® HTS filter plates can be used for filtration of 96-well replicate plates as long as well to well cross-contamination with viral supernatant can be excluded. If necessary, the filter plates can also be reused for different RNAi screens if they are washed thoroughly between experiments, ensuring decontamination of filter membranes. 9. Double RNAi knockdown screens can be conducted by extension of the protocol described here. In order to maintain the one-week experimental design we recommend transducing U-2 OS Bmal1:Luc reporter cells with previously prepared lentivirus encoding shRNA against your first target gene by standard cell transduction in a larger format (e.g., T-75 culture flask) and on day 2 of the screen protocol. The next day, reporter cells are then selected for expression of the RNAi constructs with the respective antibiotic. We recommend using RNAi constructs with puromycin resistance as this antibiotic allows for very fast and reliable selection. Following, on day 3 of the screen protocol, transduced and selected U-2 OS Bmal1:Luc reporter cells are seeded and transduced with the second RNAi construct in a 96-well format and as described. For the second transduction, RNAi constructs with a different antibiotic resistance should be used, if possible, in order to allow for secondary selection of doubly transduced reporter cells on day 5. However, based on our experience we assume that also single selection is sufficient to achieve good knockdown of both target genes if the concentration of the lentivirus is high enough during transduction. 10. In order to semiquantitatively control for transfection and transduction efficiency GFP expression can be measured in a fluorescence microplate reader since lentiviral RNAi plasmids coexpress

RNAi Screening for Novel Clock Genes

113

tGFP. Fluorescent emission of HEK293T cells can be measured on day 4 after transfer of viral supernatant, while GFP expression by transduced U-2 OS Bmal1:Luc reporter cells can be measured during the second PBS washing step on the last day of the protocol. Fluorescence intensity should lie significantly over background (empty clear/white 96-well tissue culture treated plates) to be indicative of sufficient shRNA expression. 11. Antibiotic selection of transduced reporter cells may be done for more than 24 h. For example, to fit our protocol in a weekly schedule, we usually start with day 1 on Monday and select transduced reporter cells with puromycin over the weekend. Thus, synchronization and start of bioluminescence monitoring may also be performed on day 8 (Monday of the second week). 12. ChronoStar (∗cs.3) input format should be the following: Tab delimited text file (ASCII code); First row: column headers; First column: time in days; Columns 2n: bioluminescent counts; Number format: rational, positive elements, point as decimal separator. 13. ChronoStar views can be customized. Individual panels can be freely dragged to convenient places on your desktop. Similarly, panels may be resized and differently arranged by snapdragging them to other places within the ChronoStar application. Columns of the data table may be rearranged or individually selected in menu ! Settings ! Customize table. 14. ChronoStar output files provide raw data, detrended data, trends as well as fits of your individual time-series in respective data sheets and may be used to prepare time-series graphs.

Acknowledgments We thank Maike Mette-Thaben and Daniel Lohse for excellent technical assistance. Work in AK’s laboratory is supported by the Deutsche Forschungsgemeinschaft (SFB740/D2 and TRR186/ A17). References 1. Konopka RJ, Benzer S (1971) Clock mutants of Drosophila melanogaster. Proc Natl Acad Sci U S A 68:2112–2116 2. Vitaterna MH, King DP, Chang AM et al (1994) Mutagenesis and mapping of a mouse gene, clock, essential for circadian behavior. Science 264:719–725 3. Bae K, Jin X, Maywood ES et al (2001) Differential functions of mPer1, mPer2, and mPer3

in the SCN circadian clock. Neuron 30:525–536 4. Cermakian N, Monaco L, Pando MP et al (2001) Altered behavioral rhythms and clock gene expression in mice with a targeted mutation in the Period1 gene. EMBO J 20:3967–3974 5. Shearman LP, Jin X, Lee C et al (2000) Targeted disruption of the mPer3 gene: subtle

114

Bert Maier et al.

effects on circadian clock function. Mol Cell Biol 20:6269–6275 6. van der Horst GT, Muijtjens M, Kobayashi K et al (1999) Mammalian Cry1 and Cry2 are essential for maintenance of circadian rhythms. Nature 398:627–630 7. Zheng B, Larkin DW, Albrecht U et al (1999) The mPer2 gene encodes a functional component of the mammalian circadian clock. Nature 400:169–173 8. Zheng B, Albrecht U, Kaasik K et al (2001) Nonredundant roles of the mPer1 and mPer2 genes in the mammalian circadian clock. Cell 105:683–694 9. Downward J (2004) Use of RNA interference libraries to investigate oncogenic signalling in mammalian cells. Oncogene 23:8376–8383 10. Balsalobre A, Damiola F, Schibler U (1998) A serum shock induces circadian gene expression in mammalian tissue culture cells. Cell 93:929–937 11. Wallach T, Kramer A (2015) Chemical chronobiology: toward drugs manipulating time. FEBS Lett 589:1530–1538 12. Maier B, Wendt S, Vanselow JT et al (2009) A large-scale functional RNAi screen reveals a role for CK2 in the mammalian circadian clock. Genes Dev 23:708–718

13. Zhang EE, Liu AC, Hirota T et al (2009) A genome-wide RNAi screen for modifiers of the circadian clock in human cells. Cell 139:199–210 14. Liu AC, Welsh DK, Ko CH et al (2007) Intercellular coupling confers robustness against mutations in the SCN circadian clock network. Cell 129:605–616 15. Silva JM, Li MZ, Chang K et al (2005) Secondgeneration shRNA libraries covering the mouse and human genomes. Nat Genet 37:1281–1288 16. Stegmeier F, Hu G, Rickles RJ et al (2005) A lentiviral microRNA-based system for singlecopy polymerase II-regulated RNA interference in mammalian cells. Proc Natl Acad Sci U S A 102:13212–13217 17. Stegmeier F, Rape M, Draviam VM et al (2007) Anaphase initiation is regulated by antagonistic ubiquitination and deubiquitination activities. Nature 446:876–881 18. Draviam VM, Stegmeier F, Nalepa G et al (2007) A functional genomic screen identifies a role for TAO1 kinase in spindle-checkpoint signalling. Nat Cell Biol 9:556–564 19. Ando Y, Niwa K, Yamada N et al (2008) Firefly bioluminescence quantum yield and colour change by pH-sensitive green emission. Nat Photonics 2:44–47

Chapter 9 Chromatin Immunoprecipitation (ChIP) from Mouse Liver Nuclei Sara S. Fonseca Costa and Ju¨rgen A. Ripperger Abstract Understanding the binding of regulatory proteins to their cognate genomic sites is an important step in deciphering transcriptional networks such as the circadian oscillator. Chromatin immunoprecipitation (ChIP) enables the detection and temporal analysis of such binding events in vivo. Here, we describe the individual steps from the generation of formaldehyde-cross-linked chromatin from mouse liver nuclei, fragmentation thereof, immunoprecipitation, reversal of cross-links, fragment cleanup, and detection of binding sites by real-time PCR. Depending on the quality of the employed antibody, a clear enrichment signal over the background is expected with a resolution of about 500–800 base pairs around the selected primer–probe pair. Key words DNA binding, Formaldehyde, Reversible cross-link, Sonication, Antibody, Immunoprecipitation, DNA purification, Real-time PCR

1

Introduction Formaldehyde was originally used to map the positions of nucleosomes in the DNA virus SV40 [1]. Subsequently, the method was more and more refined to allow the detection of DNA binding proteins to chromatin [2, 3]. The first attempts of the method were quite cumbersome because after the cross-link with formaldehyde chromatin was isolated as the fraction between free DNA and free protein. This was achieved by centrifugation through CsCl gradients. Later on, the method was more simplified [4, 5] and applied to the analysis of, for example circadian time series [6–8]. Finally, the basic method combined with ultradeep sequencing approaches allowed for the detection of binding sites within the chromatin with close to nucleotide resolution [9–11]. At a first glance, the method is quite simple. It is possible to cross-link a DNA-binding protein to DNA in its close proximity using formaldehyde. The chromatin is fragmented usually by sonication in the presence of sodium dodecyl sulfate (SDS) to increase

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_9, © Springer Science+Business Media, LLC, part of Springer Nature 2021

115

116

Sara S. Fonseca Costa and Ju¨rgen A. Ripperger

the viscosity of the solution. Protein–DNA complexes are then immunoprecipitated with a specific antibody. Due to the chemical nature of the cross-links, these can be reversed easily by lowering the pH at an elevated temperature, and then the liberated DNA fragments are detected. However, the devil is in the detail and this is why we elaborated a method of chromatin isolation from mouse liver nuclei [7]. Briefly, cross-links are established immediately during the homogenization process to avoid the detachment of the regulatory proteins from DNA due to their intrinsic off-rate of binding [12]. In a further step, pure nuclei essentially devoid of cytoplasmic contamination are obtained by centrifugation through 2.05 M sucrose cushions. Both steps improve the signal-to-noise ratio. However, in cases where to detect proteins not intimately bound to DNA but part of complexes associated with chromatin, it is better to use an additional cross-linker with a longer spacer between the functional groups, such as disuccinimidyl glutarate (DSG) [9].

2

Materials Prepare all solutions with at least double-distilled water or equivalent. All solutions were sterilized by vacuum-assisted filtration through 0.22 mm polyethersulfone (PES) membranes, which for the sucrose solutions takes quite a while. Sucrose solutions can be stored at 4
C for at least 2 months; the other solutions are stable at RT for at least 6 months, unless otherwise indicated.

2.1 Purification of Nuclei

1. Set of surgical instruments suitable for work with mice. 2. PBS: phosphate-buffered saline. 14.7 mM KH2PO4, pH 7.4, 26.8 mM KCl, 137 mM NaCl, and 79.7 mM Na2PO4, usually prepared as 10 concentrate and then diluted just before use. 3. 60 ml glass beaker kept on ice. 4. 15 ml conical centrifuge tubes, any model will do as long as there is a volume graduation on the side. 5. PYREX tissue grinder, 15 ml with Teflon pestle. The pestle must be attached to an engine, for example the Eurostar power-b basic overhead stirrer from IKA Laboratory Equipment (see Note 1). 6. 37% formaldehyde: use high-quality product concentrated 37% (v/v) and stabilized with methanol. 7. Set of GILSON pipettes, or equivalent, with filter tips. 8. 2.4 M sucrose STOP buffer. 2.4 M sucrose, 166.7 mM glycine, 10 mM HEPES·KOH, pH 7.6, 15 mM KCl, 0.5 mM spermidine, 0.15 mM spermine, 1 mM DTT, 0.5 protease inhibitor cocktail (e.g., Roche cOmplete ULTRA tablets, EDTA-free).

Chromatin Immunoprecipitation (ChIP) from Mouse Liver Nuclei

117

Bringing this amount of sucrose into solution takes a while but do not speed up the process by heating. Just add small quantities of sucrose at a time and let it dissolve. 9. 2.05 M sucrose cushion buffer: 2.05 M sucrose, 10% glycerol, 125 mM glycine, 10 mM HEPES·NaOH, pH 7.6, 15 mM KCl, 0.5 mM spermidine, 0.15 mM spermine, 1 mM DTT, 0.5 protease inhibitor cocktail (e.g., Roche cOmplete ULTRA tablets, EDTA-free). Bringing this amount of sucrose into solution takes a while, but do not speed up the process by heating. 10. Thinwall polyallomer (PA) ultracentrifugation tubes, 13.2 ml. 11. Swinging bucket rotor Thermo Fisher Scientific Sorvall TH-641 and compatible ultracentrifuge (see Note 2). 12. Soft tissue paper. 13. NDB: Nuclear dialysis buffer. 20 mM HEPES·NaOH, pH 7.6, 100 mM KCl, 0.2 mM EDTA, 20% glycerol, 2 mM DTT, 1 protease inhibitor cocktail (e.g., Roche cOmplete ULTRA tablets, EDTA-free). Even with the inhibitor the solution can be used for 1 month if kept at 4
C. 14. 1.5 ml Eppendorf centrifuge tubes and centrifuge, any model will do. 15. 80
C freezer (25
C will do for the short term). 2.2 Sonication and DNA Fragment Analysis

1. Nuclei resuspension buffer: 22 mM Tris–HCl, pH 7.5, 165 mM NaCl, 2.2 mM EDTA. 2. Small vortex instrument (see Note 3). 3. 10% SDS: dissolve 10 g of sodium dodecyl sulfate in 80 ml of water, fill to 100 ml. 4. BRANSON SLPe sonicator with microtip, or equivalent (see Note 4). 5. Protection for the ears. 6. Oven to be heated up to 68
C. 7. Qiagen MinElute PCR purification kit (# 28004). For the purification of the DNA fragments, we found this kit quite useful. The SDS in the chromatin reversal buffer tends to precipitate in most buffers but the binding buffer PB (Qiagen). 8. 1 TAE: Tris acetate EDTA buffer. 40 mM Tris Base, 0.57% (v/v) glacial acetic acid, 1 mM EDTA, usually prepared as 50 stock and diluted before use. 9. 6 loading dye for gel electrophoresis: 100 mM EDTA, pH 8, 80% glycerol, 0.01% (w/v) xylene cyanol, 0.01% (w/v) bromophenol blue.

118

Sara S. Fonseca Costa and Ju¨rgen A. Ripperger

10. 1.5% agarose gel: prepare by boiling 1.5 g of agarose in 1 TAE buffer and cool down to roughly 55–60
C. Add 0.01% SYBR Safe dye and cast into any kind of electrophoresis device gel holder. After appropriate solidification, place into the electrophoresis device, fill with 1 TAE buffer, and run the gel according to the specifications of the device. 11. Gel documentation system with UV screen, any system will do. 2.3 Chromatin Immunoprecipitation

1. Antibody diluted in TSE I (see Note 5). 2. Rotating wheel, any system will do as long as it does not turn too fast. 3. Protein A-agarose (e.g., Roche #11134515001) diluted in TSE I (see Note 6). 4. Chromatin dilution buffer: 20 mM Tris–HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1.1% Triton X-100. 5. TSE I: 20 mM Tris–HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% Triton X-100. 6. TSE II: 20 mM Tris–HCl, pH 7.5, 500 mM NaCl, 2 mM EDTA, 0.1% SDS, 1% Triton X-100. 7. TSEIII: 20 mM Tris–HCl, pH 7.5, 250 mM LiCl, 2 mM EDTA, 0.5% Na deoxycholate, 0.5% Nonidet NP40 substitute. 8. TSE IV: 20 mM Tris–HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA. 9. Cross-link reversal buffer: 20 mM Tris–HCl, pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% SDS. 10. 1 ml syringes with 27G needle (see Note 7).

2.4 Detection of DNA Fragments

1. Real-time PCR machine, any model will do. 2. Master mix for real-time PCR, any will do. 3. Probe and primer pairs for real-time PCR (see Note 8).

3

Methods

3.1 Preparation of Formaldehyde-Fixed Nuclei

1. Anesthetize and kill the mouse according to your animal experimentation guidelines. Open the body cavity, remove the liver with a pair of scissors and place the organ immediately into 30 ml of ice-cold PBS. 2. Transfer the liver into a 15 ml conical centrifuge tube, wash once with PBS, adjust to a total volume of 3 ml with the same buffer and transfer everything to the 15 ml PYREX tissue grinder.

Chromatin Immunoprecipitation (ChIP) from Mouse Liver Nuclei

119

3. Add 81 μl of 37% formaldehyde (i.e., 1% final concentration) to initiate the cross-linking, start homogenization with a Teflon pestle at 2000 rpm (four times up and down) and keep for 5 min at room temperature (see Note 9). 4. Transfer the homogenate into a 15 ml conical centrifuge tube containing 8 ml of 2.4 M sucrose STOP buffer to stop the cross-linking reaction, mix well and keep on ice. 5. Carefully lay over 2 ml of a 2.05 M sucrose cushion in a 13.2 ml ultracentrifuge centrifugation tube, and centrifuge balanced tubes for 60 min at 100,000  g and 4
C using a Sorvall TH-641 swinging bucket rotor (see Note 2). 6. Discard all of the tissue and fat layer on top and all of the supernatant including the cushion, and dry the sides of the tubes with a soft tissue paper. 7. Resuspend the nuclear pellet in 500 μl of NDB using a widebore tip, transfer to a 1.5 ml Eppendorf centrifugation tube, spin down for 30 s in an Eppendorf centrifuge (2500  g, RT) and discard the supernatant. 8. Resuspend the nuclei in 200 μl of NDB and store at 80
C (see Note 10). 3.2 Preparation of Sheared Chromatin Fragments

1. Thaw the nuclei, spin down for 30 s in an Eppendorf centrifuge (2500  g, RT) and remove the supernatant. 2. Add 450 μl of nuclei resuspension buffer, resuspend well, spin down for 30 s in an Eppendorf centrifuge (2500  g, RT) and discard the supernatant. 3. Resuspend in another 450 μl of nuclei resuspension buffer and—while vortexing the Eppendorf centrifuge tube with the nuclei on a vortex mixer—slowly add 50 μl of a 10% SDS solution (see Note 3). 4. Set the sonicator to level 50% and sonicate 5–20 times for 10 s. In between, keep the samples on ice to prevent over-heating (see Note 11). 5. Spin down the insoluble matter for 2 min at 16,000  g in an Eppendorf centrifuge. Combine 400 μl of soluble chromatin in the supernatant with 3.6 ml of chromatin dilution buffer and mix well. 6. Transfer another 40 μl of the soluble chromatin into a 1.5 ml Eppendorf centrifugation tube and place in an oven at 68
C for at least 2 h (this is used to analyze the fragment size in the next step). 7. Meanwhile, spin down the diluted chromatin for an additional 5 min at 16,000  g, transfer the supernatant to three 1.5 ml Eppendorf centrifuge tubes and store at 80
C (see Note 12).

120

Sara S. Fonseca Costa and Ju¨rgen A. Ripperger

3.3 Analysis of DNA Fragment Size

1. Add 160 μl of binding buffer PB (Qiagen) to the 40 μl sample of temperature-reversed chromatin, mix well, and load on a spin column, which is placed into a 2 ml collection tube (see Note 13). 2. Spin for 30 s in an Eppendorf centrifuge (10,000  g, RT) and discard the flow-through. 3. Wash once with 300 μl of binding buffer PB (Qiagen), twice with washing buffer PE (Qiagen) and spin each time for 30 s in an Eppendorf centrifuge (10,000  g, RT). 4. Spin for 2 min in an Eppendorf centrifuge (10,000  g, RT) to dry the column. 5. Place the spin column into a fresh 1.5 ml Eppendorf centrifuge tube and add 40 μl of elution buffer EB (Qiagen). 6. Wait 1 min before to spin for 30 s in an Eppendorf centrifuge (10,000  g, RT). 7. Mix 10 μl of the eluate with 2 μl of 6 loading dye and run along side a 100 bp DNA size marker on a 1.5% agarose gel in 1 TAE buffer with 0.01% v/v SYBR Safe. 8. Make a photo of the gel to determine the average fragment size (see Note 14).

3.4 Chromatin Immunoprecipitation

1. Thaw the chromatin and spin for 30 min in an Eppendorf centrifuge (16,000  g, RT) to preclear from insoluble matter. 2. Combine 400 μl of the supernatant with 50 μl of diluted antibody in TSE I in a 1.5 ml Eppendorf centrifuge tube (see Note 5). 3. Add another 4 μl of the supernatant to 46 μl of cross-link reversal buffer in a 1.5 ml Eppendorf centrifuge tube, which represents the input control and is directly placed at 65
C to reverse the cross-links. 4. Incubate the immunoprecipitation reactions for 1 h at RT on a rotating wheel, then spin down for 30 s in an Eppendorf centrifuge (16,000  g, RT). 5. Add 50 μl of prepared protein A-agarose slurry in TSE I. Be careful to mix well the slurry because the beads settle down very rapidly (see Note 6). 6. Incubate the immunoprecipitation reactions with the protein A-agarose for one more hour at RT on the wheel, and then spin down for 30 s in an Eppendorf centrifuge (16,000  g, RT). 7. Remove the supernatant with a 1 ml syringe equipped with a 27G needle, add 500 μl of TSE I buffer and spin down for 30 s in an Eppendorf centrifuge (16,000  g, RT).

Chromatin Immunoprecipitation (ChIP) from Mouse Liver Nuclei

121

8. Remove the supernatant with a 1 ml syringe equipped with a 27G needle, add 500 μl of TSE II buffer and spin down for 30 s in an Eppendorf centrifuge (16,000  g, RT). 9. Remove the supernatant with a 1 ml syringe equipped with a 27G needle, add 500 μl of TSE III buffer and spin down for 30 s in an Eppendorf centrifuge (16,000  g, RT). 10. Remove the supernatant with a 1 ml syringe equipped with a 27G needle, add 500 μl of TSE IV buffer and spin down for 30 s in an Eppendorf centrifuge (16,000  g, RT). 11. Add 50 μl of cross-link reversal buffer, vortex briefly and place into the oven overnight at 65
C to reverse the cross-links. 3.5 Quantification of the DNA Fragments by Real-Time PCR

1. Spin down the reactions for 2 min in an Eppendorf centrifuge (16,000  g, RT). 2. Transfer 40 μl of supernatant from the input or immunoprecipitation reaction to new 1.5 ml Eppendorf centrifuge tubes with 160 μl of binding buffer PB (Qiagen) and mix well. 3. Load the entire 200 μl into a spin column and spin down for 30 s in an Eppendorf centrifuge (10,000  g, RT). 4. Discard the flow-through, wash with 300 μl of binding buffer PB, twice with 300 μl of washing buffer PE (Qiagen) and spin each time for 30 s in an Eppendorf centrifuge (10,000  g, RT). 5. Discard the flow-through and dry the columns by centrifugation for 3 min at 10,000  g, RT. 6. Add 68 μl elution buffer EB (Qiagen) to each reaction and incubate for 1 min at RT. 7. Spin for 30 s in an Eppendorf centrifuge (10,000  g, RT). 8. The eluate contains the purified DNA fragments and can be stored at 20
C (see Note 15). 9. Perform real-time PCR using 5 μl of eluate from the input and ChIP reaction (see Note 8). 10. Calculate the percent of immunoprecipitated material using the following formula in EXCEL: ¼POWER(2,[CT value from input][CT value from ChIP])/factor  100%, where factor ¼ volume of ChIP DNA/volume of input DNA, that is, 100 (see Note 16) and plot a corresponding graph (Fig. 1).

4

Notes 1. We have seen many different devices in labs all over the globe and all of them work. Nevertheless, it is a good idea to perform some tests in advance to optimize the speed and number of

122

Sara S. Fonseca Costa and Ju¨rgen A. Ripperger

Fig. 1 Time course of Rev-Erbα binding to its own gene. Shown is the binding of Rev-Erbα to the promoter region of its gene (4 h intervals, when Zeitgeber time 0 and 12 represent lights on and off, respectively). We used chromatin from wildtype animals (white bars) and Rev-Erbα-deficient mice as control (black bars). Note that binding of Rev-Erbα is not detectable in Rev-Erbα-deficient mice demonstrating specificity of the anti-Rev-Erbα antibody and providing an estimate for background binding (adapted from ref. 8)

strokes necessary to obtain the maximum number of nuclei. Also, wear a pair of suitable protection gloves and goggles in case of failure of the grinder or the spilling out of organ pieces. 2. Again, we have been working with a multitude of different types of ultracentrifuges as long as the rotor is of the swinging bucket type and could reach 100,000  g. It is easy to upscale the process to a larger rotor such as an SW28 rotor from Beckman. However, it is difficult to use a smaller rotor unless the sample is divided into multiple tubes. The reason being the homogenization of the liver requires a certain volume. If it is too small, then the liver homogenate sometimes turns into a nearly solid, unusable mass. 3. Do not vortex too harsh to avoid the formation of bubbles. If you hit such a bubble with the sonicator, you obtain a lot of foam and bad chromatin. The chromatin turns—after contact with the SDS—into a rubbery mass, which sometimes becomes even kind of solid. But this is not a problem the sonicator cannot fix. 4. Again, we have been using a variety of sonicators with microtips, and every time the conditions for the optimal fragment size have to be established. Always wear ear protections! An average fragment size of 500–800 bp is optimal (normally achieved after 5 sonication cycles) and for ChIP sequencing experiments, 150 bp (normally achieved after 20 sonication cycles). The smaller the fragment size, the better the resolution

Chromatin Immunoprecipitation (ChIP) from Mouse Liver Nuclei

123

of the experiment and the lower the background. However, we also lose the signal, which can be a problem for the detection of some proteins. 5. The antibody is the most critical component of the experiment. Even “ChIP grade” antibodies have to be carefully characterized. We generally perform the experiment with probes specific for the suspected region of binding and an unrelated genomic site to determine the signal-to-noise ratio. Also, chromatin itself is not very soluble and tends to precipitate, which can be detected as a peak in the control genomic site. If available, a knock-out mouse for the protein of interest would provide the best specificity check (Fig. 1). For ChIP sequencing experiments essentially use all of the material and upgrade the reaction and wash steps accordingly. 6. The protein A-agarose (50 μl per reaction) is washed a couple of times with water, followed by TSE I. We normally prepare a stock of washed beads and carefully transfer the same volume to all of the reactions using a wide-bore tip. In case of too much background, the beads may be blocked with bovine serum albumin (BSA). Paramagnetic beads also work, but in our hands, they have more background binding. 7. The syringe may also be attached to a small aspiration system but do not to aspirate the beads, nor cross-contaminate the samples. 8. Probe and primer pairs for real-time PCR should be designed as close as possible to the suspected binding sites. Any detection system for real-time PCR will work to obtain a CT value, that is, the number of cycles necessary to reach a certain threshold. Do not forget to choose another genomic region as control. 9. This is a somewhat critical step. The cross-linking reaction is temperature-dependent, for example the signal obtained for a cross-link at 4
C is only about 1–2% of the signal obtained at 37
C. A good compromise is RT, which is sometimes difficult to keep constant over the course of a circadian experiment. 10. Under these conditions, the formaldehyde cross-linked nuclei are stable for years. 11. We keep the Eppendorf centrifuge tube in a metal rack placed on ice and shift its position by one hole after each sonication cycle. If the samples become too cold, then the SDS will precipitate. We found it as a good compromise to process six tubes at the same time, which represents a good circadian time series. In this way, the samples stay cold, but the SDS does not precipitate. However, do not contaminate your samples with material from other samples.

124

Sara S. Fonseca Costa and Ju¨rgen A. Ripperger

12. Also this material is quite stable. Alternatively, the diluted chromatin may be kept at 4
C for up to 1 month. 13. Follow the instructions that come with the kit. 14. The average fragment size should be 500–800 bp if not looking for higher resolution. All samples should have a similar distribution. 15. The eluted DNA fragments are quite stable at 20
C. The amount should be sufficient to analyze three genomic regions and one control region in triplicate each. After upscaling, there should be sufficient material to prepare a library for ChIP sequencing (in this case a pilot experiment should yield at least 0.3% of precipitation from the input with low background in the control region). 16. In our hands, background binding is in the range of 0.02–0.05% of input (see Fig. 1, Rev-Erbα-deficient control mice). Antibodies against specific DNA-binding proteins precipitate in the range of 0.1–1% of the input and histones and their modifications in the range of 0.3–10% of the input. Any higher efficiency may be due to over-cross-linking of the chromatin or calculation errors.

Acknowledgments This work was supported by the Canton of Fribourg, Switzerland, by the Swiss National Science Foundation grant 31003A_152792, and the SystemsX grant 51PHD0_157318. References 1. Solomon MJ, Varshavsky A (1985) Formaldehyde-mediated DNA-protein crosslinking: a probe for in vivo chromatin structures. Proc Natl Acad Sci U S A 82:6470–6474 2. Orlando V, Paro R (1993) Mapping Polycomb-repressed domains in the bithorax complex using in vivo formaldehyde crosslinked chromatin. Cell 75:1187–1198 3. Solomon MJ, Larsen PL, Varshavsky A (1988) Mapping protein-DNA interactions in vivo with formaldehyde: evidence that histone H4 is retained on a highly transcribed gene. Cell 53:937–947 4. Strutt H, Cavalli G, Paro R (1997) Co-localization of Polycomb protein and GAGA factor on regulatory elements responsible for the maintenance of homeotic gene expression. EMBO J 16:3621–3632 5. Strutt H, Paro R (1999) Mapping DNA target sites of chromatin proteins in vivo by

formaldehyde crosslinking. Methods Mol Biol 119:455–467 6. Lee C, Etchegaray JP, Cagampang FR, Loudon AS, Reppert SM (2001) Posttranslational mechanisms regulate the mammalian circadian clock. Cell 107:855–867 7. Ripperger JA, Schibler U (2006) Rhythmic CLOCK-BMAL1 binding to multiple E-box motifs drives circadian Dbp transcription and chromatin transitions. Nat Genet 38:369–374 8. Stratmann M, Stadler F, Tamanini F, van der Horst GT, Ripperger JA (2010) Flexible phase adjustment of circadian albumin D site-binding protein (DBP) gene expression by CRYPTOCHROME1. Genes Dev 24:1317–1328 9. Cho H, Zhao X, Hatori M, Yu RT, Barish GD et al (2012) Regulation of circadian behaviour and metabolism by REV-ERB-alpha and REV-ERB-beta. Nature 485:123–127

Chromatin Immunoprecipitation (ChIP) from Mouse Liver Nuclei 10. Koike N, Yoo SH, Huang HC, Kumar V, Lee C et al (2012) Transcriptional architecture and chromatin landscape of the core circadian clock in mammals. Science 338:349–354 11. Rey G, Cesbron F, Rougemont J, Reinke H, Brunner M et al (2011) Genome-wide and phase-specific DNA-binding rhythms of

125

BMAL1 control circadian output functions in mouse liver. PLoS Biol 9:e1000595 12. Stratmann M, Suter DM, Molina N, Naef F, Schibler U (2012) Circadian Dbp transcription relies on highly dynamic BMAL1-CLOCK interaction with E boxes and requires the proteasome. Mol Cell 48:277–287

Chapter 10 Using GRO-Seq to Measure Circadian Transcription and Discover Circadian Enhancers Bin Fang, Dongyin Guan, and Mitchell A. Lazar Abstract Circadian gene transcription transmits timing information and drives cyclic physiological processes across various tissues. Recent studies indicate that oscillating enhancer activity is a major driving force of rhythmic gene transcription. Functional circadian enhancers can be identified in an unbiased manner by correlation with the rhythms of nearby gene transcription. Global run-on sequencing (GRO-seq) measures nascent transcription of both pre-mRNAs and enhancer RNAs (eRNAs) at a genome-wide level, making it a unique tool for unraveling complex gene regulation mechanisms in vivo. Here, we describe a comprehensive protocol, ranging from wet lab to in silico analysis, for detecting and quantifying circadian transcription of genes and eRNAs. Moreover, using gene-eRNA correlation, we detail the steps necessary to identify functional enhancers and transcription factors (TFs) that control circadian gene expression in vivo. While we use mouse liver as an example, this protocol is applicable for multiple tissues. Key words Circadian, Enhancer RNA, Global run-on sequencing, Transcription factor, Nascent RNA

1

Introduction The mammalian circadian clock is composed of a transcriptional– translational feedback loop that is controlled by TFs [1]. A variety of genetic studies have implicated the BMAL1/CLOCK heterodimer as the main positive transcriptional regulator, which drives the transcription of PER/CRY and Rev-erbα/β that feedback and repress BMAL1/CLOCK activity and Bmal1 (aka Arntl) gene expression [2]. This mutual regulation results in the rhythmic expression of core clock TFs that oscillate with a period of approximately 24 h. In addition to regulating each other, core clock TFs control other genes, referred to as clock output genes, which are also expressed in a circadian manner. Although BMAL1/CLOCK was initially viewed as the main controller of clock output genes, it

Bin Fang and Dongyin Guan contributed equally to this work. Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_10, © Springer Science+Business Media, LLC, part of Springer Nature 2021

127

128

Bin Fang et al.

is now clear that all of the core clock TFs contribute to the rhythmic transcription of many genes [3, 4]. TFs bind DNA in a sequence-specific manner at promoters and at enhancers to regulate gene transcription. Although tremendous effort has been devoted to understanding the precise mechanism controlling gene transcription, major questions remain concerning the location of functional enhancers and the identify of TFs that regulate a gene of interest. In earlier studies, DNA sequences recognized by specific TFs were usually examined by in vitro binding methods such as electrophoretic mobility shift assays (EMSA), and functionality of putative binding sequences was then tested using reporter assays [5]. With the help of protein binding microarrays, tens of thousands of DNA oligomers have also been simultaneously tested providing a thorough understanding of TF binding capacity and plasticity [6, 7]. However, in vitro binding dynamics of TFs do not necessarily recapitulate their in vivo specificities in part due to the degenerative nature of binding motifs. Moreover, the search for enhancer (and repressor) elements was often limited to within a few kilo base pairs (kb) of transcription starting site (TSS) with these early methods. Today, in the new era of high-throughput DNA sequencing, our capabilities of detecting genome-wide TF binding sites have greatly improved. For example, based on recent chromatin immunoprecipitation followed by deep sequencing (ChIP-seq) studies, it is now appreciated that TFs normally bind to thousands of locations across the genome [8]. It is less clear whether a binding site is functional, and accordingly, several strategies have evolved to help sift out the functional binding sites that harbor regulatory activities. Popular strategies often correlate TF binding data with other epigenomic markers of enhancers such as specific histone modifications as determined by ChIP, or chromatin accessibility as determined by DNase hypersensitivity followed by deep sequencing (DHS-seq) or with transposase-accessible chromatin with highthroughput sequencing (ATAC-seq) assays [9–11]. Importantly, these epigenomic marks can be used to discover de novo enhancers without the need to focus on a specific TF [12, 13]. Other emerging techniques are helping to map active enhancers to their target genes. For example, genome scale chromatin looping as measured by 4C, 5C, Hi-C, and ChIA-PET techniques [13–16] has revealed complex topological associations between genes and enhancers. In addition to looping, identification of concordant eRNA and nearby gene expression patterns has been informative [6, 17, 18]. Recently discovered short noncoding RNAs transcribed at enhancers (eRNAs) can serve as markers enhancer activities [19, 20], which can be quantified along with nascent gene transcripts by global run-on followed by deep sequencing (GRO-seq) [21]. Here we provide a detailed protocol, from wet lab procedures to bioinformatics analysis, for performing GRO-seq

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

129

and mapping genome-wide active enhancers in mouse livers. GRO-seq is particularly appealing for the study of circadian transcription, as it provides an unbiased and detailed map of functional genomic sites targeted by core clock TFs. It further yields mechanistic insights that may help explain why many circadian genes exhibit expression phases that are distinct from the activity phases of core clock TFs.

2 2.1

Materials Nuclei Extraction

2.1.1 Reagents

1. Experimental animals: C57Bl/6 mice are purchased from Jackson Laboratories and housed under standard 12 h light:12 h dark cycles, with lights on at 7 am and lights off at 7 pm. All animal procedures are performed following an approved protocol from the University of Pennsylvania Perelman School of Medicine Institutional Animal Care and Use Committee. 2. 1
swelling buffer: 10 mM Tris–HCl pH 7.5, 2 mM MgCl2, 3 mM CaCl2, 2 U/mL SUPERase-In. Mix together 5 mL of 1 M Tris–HCl pH 7.5, 1 mL of 1 M MgCl2, 1.5 mL of 1 M CaCl2, and 492.5 mL of ultrapure water. Add SUPERase-In before use. 3. Lysis buffer: 1
swelling buffer, 10% glycerol, 1% Igepal, 2 U/ mL SUPERase-In. Add 10 mL of glycerol, 1 mL of Igepal to 89 mL of 1
swelling buffer. Add SUPERase-In before use. 4. Resuspension buffer: 1
swelling buffer, 10% glycerol, 2 U/ mL SUPERase-In. Add 5 mL of glycerol to 45 mL of 1
swelling buffer. Add SUPERase-In before use. 5. Freezing buffer: 40% glycerol, 5 mM MgCl2, 0.1 mM EDTA, 50 mM Tris–HCl pH 8.0. Mix together 27.5 mL of ultrapure water, 20 mL of glycerol, 250 μL of 1 M MgCl2, 10 μL of 0.5 M EDTA, and 2.5 mL of Tris–HCl pH 8.0. Add SUPERase-In to solution before use. 6. SUPERase-In RNase inhibitor, ThermoFisher Scientific, AM2696. 7. 100 μm cell strainer, Corning, 431752. 8. Dounce tissue homogenizer set, working volume 15 mL.

2.2 Nuclear Run-On (NRO) 2.2.1 Reagents

1. 2
Nuclear Run-On reaction buffer: 10 mM Tris–HCl pH 8, 5 mM MgCl2, 300 mM KCl, 1 mM DTT, 500 μM ATP, 500 μM GTP, 500 μM Br-UTP 2 μM CTP, 200 U/mL SUPERase-In, 1.0% Sarkosyl. For each sample, 100 μL of freshly prepared buffer is needed. 2. TRIzol LS reagent, Thermo Fisher Scientific, 10296010.

130

Bin Fang et al.

3. Chloroform. 4. GlycoBlue, Thermo Fisher Scientific, AM9515. 2.3 NRO-RNA Purification 2.3.1 Buffers ( See Note 1)

1. Binding buffer: 0.25
SSPE, 0.05% Tween, 37.5 mM NaCl, 1 mM EDTA. Mix together 175 μL of 20
SSPE, 70 μL of 10% Tween, 105 μL of 5 M NaCl, 28 μL of 0.5 M EDTA, and 13.622 mL of nuclease-free water. 2. Blocking buffer: ~1
binding buffer, 0.1% Polyvinylpyrrolidone (PVP), 0.1% BSA. Add 20 μL of 10% Polyvinylpyrrolidone (PVP) and 40 μL of 50 mg/mL BSA to 1.94 mL of 1
binding buffer. 3. Low salt wash buffer: 0.2
SSPE, 0.05% Tween, 1 mM EDTA. Add 100 μL of 20
SSPE, 50 μL of 10% Tween, and 20 μL of 0.5 M EDTA to 9.83 mL of nuclease-free water. 4. High salt wash: 0.2
SSPE, 137.5 mM NaCl, 0.05% Tween, 1 mM EDTA. Add 100 μL of 20
SSPE, 275 μL of 5 M NaCl, 50 μL of 10% Tween, and 20 μL of 0.5 M EDTA to 9.53 mL of nuclease-free water. 5. Elution buffer: 50 mM Tris–HCl pH 7.5, 150 mM NaCl, 0.1% SDS, 1 mM EDTA, 20 mM DTT. Mix together 250 μL of 1 M Tris–HCl pH 7.5, 150 μL of 5 M NaCl, 50 μL of 10% SDS, 10 μL of 0.5 M EDTA, and 4.44 mL of nuclease-free water. Add fresh 100 μL of 1 M DTT to the solution before use.

2.3.2 Reagents

1. E. coli Poly(A) Polymerase, New England Biolabs, M0276S. 2. UltraPure Glycogen, Thermo Fisher Scientific, 10814010. 3. TURBO DNA-free Kit, Thermo Fisher Scientific, AM1907. 4. Micro Bio-Spin P-30 Gel Columns, Bio-Rad, 732-6250. 5. RNA Fragmentation Reagents, Thermo Fisher Scientific, AM8740. 6. T4 Polynucleotide Kinase, New England Biolabs, Y904L. 7. Anti-BrU agarose sc-32323 AC.

beads,

Santa

Cruz

Biotechnology,

8. SSPE (20
), Thermo Fisher Scientific, AM9767. 2.4 GRO-Seq Library Preparation

1. Low Range ssRNA Ladder, New England Biolabs, N0364S.

2.4.1 Reagents

3. TrackIt 25 bp DNA Ladder, Thermo Fisher Scientific, 10488022.

2. APE 1, New England Biolabs, M0282S.

4. Novex TBE-Urea Gels, 10%, 10 well, Thermo Fisher Scientific, EC6875BOX. 5. Novex TBE Gels, 10%, 12 well, Thermo Fisher Scientific, EC62752BOX.

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

131

6. CircLigaseTM ssDNA Ligas, Epicentre, CL4111K. 7. Novex TBE Running Buffer (5
), Thermo Fisher Scientific, LC6675. 8. ChIP DNA Clean & Concentrator-Capped column, Zymo Research, D5205. 9. DNA LoBind Microcentrifuge Tube, Eppendorf, 30108.051. 10. Exonuclease I (E. coli), New England Biolabs, M0293L. 11. Novex TBE-Urea Sample Buffer (2
), Thermo Fisher Scientific, LC6876. 12. Novex Hi-Density TBE Sample Buffer (5
), Thermo Fisher Scientific, LC6678. 13. SuperScript III Reverse Transcriptase, Thermo Fisher Scientific, 18080-051. 14. Deoxynucleotide (dNTP) Solution Mix, New England Biolabs, N0447S. 15. SYBR Gold Nucleic Acid Gel Stain, Thermo Fisher Scientific, S11494. 16. Phusion Hot Start II DNA Polymerase, Thermo Fisher Scientific, F-549L. 17. Adapter for GRO-seq. oNTI223: 50 /5Phos/GA TCG TCG GAC TGT AGA ACT CT/idSp/CAA GCA GAA GAC GGC ATA CGA TTT TTT TTT TTT TTT TTT TT VN-30 where/ 5Phos/ means 50 Phosphorylation, idSp means 0 0 1 ,2 -Dideoxyribose (dSpacer), and VN means degenerate nucleotides. 18. GRO-seq PCR primers. oNTI200: 50 -CAAGCAGAAGACGGCATA-30 , oNTI201: 50 -AATGATACGGCGACCACCGACAGGTTCAGAGTTCTACAGTCCGACG-30 . 19. Sequencing primers. oNTI202: 50 -GACAGGTTCAGAGTTCTACAGTCCGACGATC-30 . 2.5 Computational Tools and Software

The following tools are used in our study of circadian enhancers [22]. Note that newer versions might be available in some cases. Alternative programs may also be applied, but they are not listed here. 1. FastQC 0.11.2, a software with graphic user interface for examining the quantity of raw reads generated by highthroughput DNA sequencing (http://www.bioinformatics. babraham.ac.uk/projects/fastqc/). 2. Bowtie 1.1.1, a software program for aligning short DNA fragments to reference genomes (http://bowtie-bio. sourceforge.net/index.shtml) [23].

132

Bin Fang et al.

3. SAMtools 1.1, software programs for handling sequence alignment files in .bam format (http://samtools.sourceforge.net/) [18]. 4. Bedtools 2.19, a package of tools for various analysis on data types with genomic coordinates (http://bedtools.readthedocs. org/en/latest/) [24]. 5. Homer v4.6, a computational tool package for processing and analyzing high-throughput DNA sequencing data (http:// homer.salk.edu/homer/) [25]. 6. JTK_cycle, an R package for detecting rhythmic changes of genomic data (https://github.com/mfcovington/jtkcycle) [26]. 7. IGV 2.3, a graphical user interface for visualizing DNA sequencing data (http://software.broadinstitute.org/soft ware/igv/home) [27]. 8. R 3.1.0, a software programming environment for statistical computing and graphics (www.r-project.org/). 9. Perl v5.16, a scripting programming language widely used in bioinformatics studies (https://www.perl.org).

3

Methods The following protocol is optimized for GRO-seq in mouse liver and would require adjustment for other tissue types (see Note 2). The total time of this protocol is 7 days (Fig. 1).

3.1

Nuclei Extraction

3.1.1 Steps

1. Harvest mouse livers every 3 h throughout the 24-h circadian cycle. Liver samples can be either processed fresh or snapfrozen in liquid nitrogen for later usage. 2. Wash the liver with cold swelling buffer. 3. Dounce homogenize the liver ten times with piston A in 15 mL of swelling buffer and incubate at 4  C for 20 min. 4. Dounce homogenize another 20 times with piston B and add 15 mL of swelling buffer. Use the cell strainer (100 μm) to filter the cells and centrifuge at 400
g at 4  C for 10 min. 5. Remove the supernatant and resuspend the pellet in 10 mL of resuspension buffer. Slowly vortex while adding drop by drop 10 mL of lysis buffer and incubate at 4  C for 5 min. Add 30 mL of lysis buffer and centrifuge at 600
g at 4  C for 5 min. 6. Remove the supernatant and wash the pellet with 25 mL of lysis buffer. Centrifuge at 600
g at 4  C for 5 min. Then, remove the supernatant and resuspend the pellet in 10 mL of freezing buffer and count the nuclei (see Note 3).

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . . Day 1

133

Circadian tissue collection: 1 day

*

Nuclei extraction : 3h Day 2

Nuclear Run-On reaction: 1h Precipitation: overnight Dnase treatment: 1h Hydrolysis of RNA: 0.5h

Day 3

PNK treatment: 2.5h NRO-RNA immunoprecipitation: 2h Precipitation: overnight PolyA-tailing reaction: 1h

Day 4

Reverse transcription: 1h cDNA purification: 4h Precipitation: overnight

*

Circularization: 2h Day 5-6

Relinearization: 2h PCR amplification: 1h

*

Library DNA purification: 8h

*

Day 7

Product size check: 2h

*

qPCR quantification: 3h

Fig. 1 Flowchart of wet bench protocol for GRO reaction and deep-seq library preparation. ∗ Refers to a safe stopping point in the protocol

7. Centrifuge at 900
g at 4  C for 6 min. Remove the supernatant and resuspend the nuclei to reach the concentration of 107 nuclei/100 μL of freezing buffer. Aliquot 100 μL of nuclear extract into each tube. A total of 400 μL (4 individual tubes) is needed for each library. Snap-freeze the additional nuclei and store at 80  C for future usage. 3.2 Nuclear Run-On Reaction 3.2.1 Steps

1. Mix 100 μL of nuclear extract with 100 μL of nuclear run on reaction buffer by pipetting 15 times up and down with end-cut 200 μL pipette tips. Then, incubate at 30  C for 5 min. 2. Add 450 μL of TRIzol, and incubate at room temperature (RT) for 5 min. Then, add 120 μL of chloroform to the TRIzol solution containing nuclei and shake the mixture by hand for 15 s. Incubate at RT for 2–3 min and centrifuge at 12,000
g at 4  C for 15 min.

134

Bin Fang et al.

3. Take the colorless upper layer containing RNA, transfer to a 2 mL tube, add NaCl (up to 300 mM), 2 μL glycol-blue, and 2.5 volumes of cold 100% ethanol. Incubate at 20  C overnight. 4. Centrifuge at 14,000
g at 4  C for 5 min. Remove supernatant and wash the pellet with 75% ethanol, vortex and centrifuge at 7500
g at 4  C for 5 min. Pool the 4 pellets in 1 tube (started with 4 tubes per condition). Remove supernatant, spin and remove the rest of supernatant. Air-dry the pellet. 5. Suspend RNA into 22 μL of DNase, RNase-free water containing SUPERAse-In (1 U/μL). Incubate the pellet at 60  C for 10 min. 3.3 NRO-RNA Immunoprecipitation 3.3.1 DNase Treatment Steps

1. To eliminate DNA contamination, 2.8 μL of 10
DNase buffer and 4 μL of DNase (from TURBO DNA-free Kit) are added to 20 μL of RNA samples and incubated at 37  C for 30 min. 2. Add 4 μL of DNase Inactivation Reagent and incubate at RT for 5 min with occasional mixing. 3. Centrifuge at 10,000
g for 2 min and transfer RNA to a fresh tube.

3.3.2 Hydrolysis of RNA

1. Divide the RNA into two tubes (12 μL each) and add 8 μL of H2O along with 2 μL of fragmentation reagent.

Steps

2. Incubate at 70  C for 13 min. Add 2 μL of stop solution (from RNA Fragmentation Reagents) and put on ice. Combine the 2 tubes and purify the RNA using Micro Bio-Spin P-30 Gel Columns.

3.3.3 Treatment of NRORNA with PNK

1. Divide the RNA into two tubes (add water up to 25 μL to each tube) and add 1 μL of SUPERase-In, 3 μL of 10
PNK buffer, and 1.5 μL of PNK to each tube.

Steps

2. Incubate at 37  C for 1 h, and then add 1 μL of PNK and incubate for another 1 h. 3. Stop the reaction by adding EDTA (final 10 mM). Inactivate the enzyme by incubating at 75  C for 5 min and put samples on ice for at least 2 min. Combine the 2 tubes and bring the volume to 100 μL with binding buffer.

3.3.4 Binding to AntiBrUTP Beads Steps

1. Equilibrate beads by washing two times in 500 μL of binding buffer for 5 min with rotation. Spin down beads at 1000
g for 2 min. Place on ice for 1 min before removing supernatant (see Note 4). 2. Block beads with 5 volumes of blocking buffer (50 μL of beads in original suspension per sample) at RT for 1 h.

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

135

3. Wash beads twice in 500 μL binding buffer for 5 min with rotation. Spin down beads at 1000
g for 2 min. Place on ice for 1 min before removing supernatant. 4. Resuspend beads in 500 μL of binding buffer. Bring the denatured RNA volume to 100 μL with binding buffer and mix with beads. 5. Incubate RNA with beads at RT for 60 min. 6. Wash beads twice in 500 μL of binding buffer for 5 min with rotation. Spin down beads at 1000
g for 2 min. Place on ice for 1 min before removing supernatant. 7. Wash beads twice in 500 μL of low salt buffer for 5 min with rotation. Spin down beads at 1000
g for 2 min. Place on ice for 1 min before removing supernatant. 8. Wash beads in 500 μL of high salt buffer for 5 min with rotation (do not exceed 5 min). Spin down beads at 1000
g for 2 min. Place on ice for 1 min before removing supernatant. 9. Wash beads twice in 500 μL of TET buffer (1
TE with 0.05% tween) for 5 min with rotation. Spin down beads at 1000
g for 2 min. Place on ice for 1 min before removing supernatant. 10. Elute NRO-RNA from beads using 100 μL of elution buffer for 10 min with rotation (see Note 5). Spin down beads at 1000
g for 2 min. Place on ice for 1 min before saving the supernatant. Repeat this step for three additional times to collect 400 μL of eluate. 11. Add 2 μL of glycogen and bring NaCl to the concentration of 300 mM by adding 12 μL of 5 M NaCl (already some NaCl in elution buffer). 12. Add 2.5 volumes of 100% Ethanol and incubate at 20  C overnight. 13. Centrifuge the tubes at maximum speed at 4  C for 25 min. Carefully discard the supernatant without disturbing the pellet. 14. Wash cDNA pellet with 1 mL of 75% ethanol. Centrifuge the tube at maximum speed at 4  C for 5 min. Pool the reactions by suspending the pellet in 5 μL of ultrapure water (SUPERase-In 1 U/μL and 0.05% Tween). 15. Add one negative control sample (5 μL of water). Purified RNA is denatured by incubating at 65  C for 3 min and subsequently placed in ice.

136

Bin Fang et al.

3.4 cDNA Library Preparation 3.4.1 PolyA-Tailing Reaction

3.4.2 Reverse Transcription

Set up the following reaction and incubate at 37  C for 30 min. Components

Volume (μL)

Denatured RNA (10–25 pmol)

5

Poly(A)-polymerase buffer
10 (contains 10 mM MgCl2)

0.7

2.5
diluted ATP (stock ATP 10 mM)

0.25

E. coli poly(A)-polymerase (3.75 U)

0.7

Total volume

6.5

Add 0.9 μL of dNTP mix (10 mM each) and 0.9 μL of oNTI223 (10 pmol/μL) primer to poly(A)-RNA. Heat the solution to 75  C for 3 min and chilled on ice. Then add the following reagents and incubate at 48  C for 40 min. Components

Volume (μL)

10
reverse transcription buffer

1.7

25 mM MgCl2

3

0.1 M DTT

1.7

SUPERase-In (20 U/μL)

0.5

SuperScript III reverse transcriptase (200 U/μL)

1

3.4.3 Exonuclease Treatment to Remove Extra Primer

1. Add 3 μL of exonuclease I and incubate at 37  C for 15 min. Then inactivate enzyme activity and eliminate RNA by adding 2 μL of 1 M NaOH and incubating at 98  C for 20 min. The reaction is then neutralized with 1 μL of 2 M HCl.

Steps

2. Prepare 10% TBE-urea gel (10-well) for gel-extraction. All the samples are denatured at 70  C for 3 min and then put on ice before loading. Gel is run for 15 min at 180 V before loading. Then, load samples along with Low Range ssRNA Ladder and run for 60–80 min at 170 V. Poststain the gel by adding 25 mL of 1
TBE buffer with 2.5 μL of SYBR Gold nucleic acid gel stain at RT for 5 min. Extract fragments between 105 and 400 nt in size (before the smear of 80 nt marker starts). 3. Puncture the bottom of a sterile, nuclease-free, 0.5 mL tube five times with a 21 G needle. 4. Put the extracted gel slice into the punctured tube and place the tube into a 2 mL sterile round-bottom nuclease-free tube. 5. Let the gel slice pass through small holes into the 2 mL tube by spinning at 14,000
g for 3 min. 6. Add 500 μL of 1
Elution buffer (1
TE with 0.1% Tween) to gel debris. Rotate at RT for 4 h.

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

137

7. Transfer eluate and gel debris to Spin X filter. Spin at 14,000
g for 2 min. 8. Precipitate cDNA by adding 1 μL of glycogen, 30 μL of 5 M NaCl, and 1 mL of 100% ethanol. Incubate at 20  C overnight. 9. Centrifuge the samples at maximum speed at 4  C for 30 min. Carefully discard the supernatant without disturbing the pellet. Wash cDNA pellet with 1 mL of 75% ethanol. Centrifuge the tube at maximum speed at 4  C for 5 min and dry the pellet. 3.4.4 Circularization

1. Suspend the pellet into circularization reaction.

Steps

Components

Volume (μL)

H2O (0.05% Tween)

7.5

10
reaction buffer

1

1 mM ATP (final 0.05 mM)

0.5

50 mM MnCl2 (final 2.5 mM)

0.5

CircLigase (100 U/μL)

0.4

3.4.5 Relinearization

2. Circularization is performed at 60  C for 1 h, and the reaction is heat-inactivated at 80  C for 10 min. 1. Circular single-stranded DNA is relinearized by adding 3.3 μL of relinearization mix (4
mix containing 100 mM KCl and 2 mM DTT) followed by 1 μL of APE1. 2. The reaction is incubated for 45 min. Add 1 μL of APE1 to the solution and incubated for another 45 min at 37  C. 3. Inactivate APE1 by incubating at 65  C for 20 min. Dilute the samples up to 15 μL with water.

3.4.6 PCR Amplification Steps

1. Purify and concentrate the reactions using ChIP DNA clean & Concentrator Kit. Elute with 15 μL EB-buffer or water and then set up the following amplification PCR reaction. Components

Volume (μL)

cDNA

5

oNTI 200 (10 μM)

2.5

oNTI 201 (10 μM)

2.5

5
HF buffer

10

10 mM dNTPs

1

Betaine

5

DNA polymerase

0.5

H2O

23.5

138

Bin Fang et al.

Set up the following PCR program. Step

Temperature ( C)

Time

1

98

30 s

2

98

10 s

3

57

30 s

4

72

15 s

5

72

5s

6

4

Indefinite

2. Load reaction products onto 8–10% polyacrylamide TBE gel and run at 120–150 V for 2 h along with TrackIt 25 bp DNA Ladder. Poststain the gel by adding 25 mL of 1
TBE buffer with 2.5 μL of SYBR-gold nucleic acid gel stain at RT for 5 min. Excise fragments between 150 and 300 bp (see Note 6). 3. Puncture the bottom of a sterile, nuclease-free, 0.5 mL tube five times with a 21 G needle. 4. Add gel slice into the punctured tube and place the tube into a 2 mL sterile, round-bottom nuclease-free tube. 5. Pass the gel into the 2 mL tube by spinning at 14,000
g for 2 min. 6. Add 400 μL 1
Elution buffer (1
TE with 0.1% Tween and150 mM NaCl) to gel debris and rotate for 3 h at RT. 7. Transfer eluate and gel debris to Spin X filter and spin at 14,000
g for 2 min. 8. Purify above eluate using ChIP DNA clean & Concentrator Kit and then elute with 10–20 μL of EB-buffer. 9. Quantify the concentration of library (target concentration 50–100 ng/μL) using a NanoDrop spectrophotometer. Further evaluate the size and concentration of the libraries using High Sensitivity DNA analysis kit and Agilent BioAnalyzer (Fig. 2). The exact concentration of library (50–200 nM) is determined by KAPA Library Quantification Kits. 10. Sequencing was subsequently performed on Illumina HiSeq using a single end sequence read length of 50 bp.

Fig. 2 Ideal Agilent Bioanalyzer High Sensitive DNA profile of GRO-seq library

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

139

3.5 Quality Control of GRO-Seq Data

Multiple steps of processing are required before GRO-seq reads can be aligned to the reference genome. A detailed protocol from preprocessing to sequence alignment has been documented in a previous chapter [28]. Since each step filters out certain reads, it is common that 40–50% of the total GRO-seq reads are excluded before the last step. Thus, the minimum GRO-seq reads required for consistent measurement of GRO-seq transcription needs to be determined. A downsampling approach as illustrated in Fig. 3a can be applied. Transcription levels of eRNAs at selected bona fide enhancers were repetitively calculated using decreasing numbers of reads that are randomly chosen from the total reads. Consistency between datasets was then determined using Pearson’s correlation. In this example, 30 million of uniquely mapped reads are required for consistent measurement of eRNA transcription (r > 0.99). This number can be lower (~20 million reads) for measuring gene body transcription. Since the quality and robustness of GRO-seq data are largely dependent on biological samples and library preparation, it is recommended that the minimum number of reads be determined for each experimental setup.

3.6 Quantification of Gene Transcription

GRO-seq detects nascent transcription of pre-mRNA; thus, its reads are mapped to both exonic and intronic regions with unique strandness. Gene transcription is then measured using reads mapped to the sense strand of the gene. It is known that GRO-seq signals gradually decay toward the 30 end of transcripts [29], likely due to gradually decreased Pol2 activity and stability in the in vitro nuclear run-on system (Fig. 3b). In a different scenario, paused Pol2 at the TSS may result in accumulated GRO-seq signals at the immediate downstream of the TSS, forming a TSS peak that is much higher than the rest of the genic region (Fig. 3b). To minimize the negative effects of these factors in the accurate measurement of gene transcription, the average GRO-seq signal (in rpkm, reads per kb per million total reads) within a 10 kb window (from +2 kb to +12 kb relative to TSS) of each Refseq transcript is used to estimate the real-time transcription of each gene. Genes with lengths between 5 kb and 12 kb are measured using smaller windows from +2 kb to the transcription end site (TES) while the entire gene body is used for genes shorter than 5 kb. Eventually, GRO-seq read counts will be normalized to 1 kb window size for all genes. GRO-seq read counts at selected genomic loci can be obtained by running public software programs such as bedtools [24] and HTseq [30]. The following is an example of a Homer command line for counting GRO-seq reads in the gene body: $ annotatepeaks 10kbwindows.bed mm9 –size given –strand + -d zt10Tags/ >10kbwindows.tagCount.txt 10kbwindows.bed is an input file containing genomic coordinates (in .bed format) of 10 kb windows selected for each gene.

140

Bin Fang et al.

A

10 0

5

5 million, r=0.83 15 million, r=0.96 30 million, r=0.99

-5

Down sampled reads

eRNA transcription (log2 rpktm)

-5

0

5

10

Total reads (60 million)

B

GRO-seq gene transcription Pol2 pausing Signal decay

Anti-sense transcription TSS

Fig. 3 Sample GRO-seq results. (a) Correlation between eRNA transcription (RPKTM, read per kb per 10 million reads) calculated using 60 million GRO-seq reads (x-axis) and randomly downsampled reads (y-axis). Pearson correlations for three sampling datasets are indicated in different colors. (b) Example of GRO-seq gene transcription. As indicated by the dash line, transcription signals of gene body (blue) gradually decreases toward the 30 end of the gene. A high signal peak at the TSS is the indication of paused Pol2. Antisense transcription at TSS (in red) is detected on the antisense strand

zt10Tags/ is a Homer tag directory containing all the aligned GRO-seq reads. “-strand +” tells the program to count the sense strand reads only. Detailed instructions for installing and running Homer can be found at http://homer.salk.edu/homer/. The program output contains GRO-seq read count in each 10 kb window (one for each gene) normalized to ten million of total reads. These numbers are further converted to rpkm by normalizing window size to 1 kb. 3.7 Identification of Circadian Genes

GRO-seq datasets from multiple time points are combined in order to identify oscillating gene expression patterns. Taking our previous study for example [22], 8 GRO-seq datasets were collected from samples taken every 3 h throughout the 24-h cycle. Previous studies

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

141

indicated that about ten thousand genes are actively transcribed in mouse liver at a given time [29, 31], based on which an expression threshold of 0.5 rpkm was selected, and inactive genes whose maximum transcription at all time points was lower than the threshold were excluded from downstream analysis. To make oscillation amplitudes comparable across different genes, the absolute transcription level (in rpkm) was converted to relative transcription which is defined as log2 fold change to the average transcription of the day (Fig. 4). Cycling transcription of active genes can be detected by an R package JTK_cycle, which uses nonparametric algorithm to predict circadian expression patterns [26]. This program takes as input a matrix containing time-course transcriptions and outputs predicted cycling phase, amplitude, period, and P-value for each gene (Fig. 4). Genes with P-value < 0.05, amplitude >0.5, and period between 21 and 27 h were considered as oscillating genes. Note that JTK_cycle takes the first data point as time 0, thus the real oscillation phase needs to be adjusted with the offset of the actual starting time (Fig. 4). Sensitivity of JTK_cycle largely relies on the number of time points in the dataset, thus duplication or concatenation of datasets might be necessary for thorough discovery of cycling targets [29]. To eliminate false positive targets, one can apply multiple prediction programs, such as COSOPT, CirWave, and ARSER, and select commonly detected hits [32]. 3.8 Identification of Circadian Enhancers

Several studies have shown that eRNAs are epigenomic markers of active enhancers that can be identified using eRNA transcripts as the single source of input [20, 22]. However, few examples discuss the quantitative characterization of dynamic eRNAs. In this section, protocols for eRNA identification and quantification are described.

3.8.1 De Novo Identification of Intergenic eRNAs

In a previously described protocol, eRNAs were identified based on intergenic RNA transcriptions captured by groHMM [28]. Here we introduce an alternative approach using Homer peak calling algorithm [25]. Unlike gene transcription, eRNA transcripts are generally short, likely due to unstable elongation and early termination. This results in accumulation of GRO-seq signals at the 50 end of eRNA transcripts, forming high peaks resembling those observed in ChIP-seq experiments. These peaks can be readily detected by Homer peak finding function with customized parameters, such as lowered enrichment ratio over local background (-L and -localSize). Note that GRO-seq reads are stranded; thus, peaks on different strands must be called separately. The following is an example of a Homer command line for peak finding: $ findPeaks groseqTagsDir –style factor -o auto –L 3 –localSize 5000

142

Bin Fang et al.

A GRO-seq transcription - Absolute ID Gene_1 Gene_2 Gene_3

ZT1 84.6 27.03 6.01

ZT4 81.3 21.9 4.3

ZT7 50.1 18.7 3.39

ZT10 54.35 19.47 2.76

ZT13 44.7 18.67 1.8

ZT16 61.23 14.45 2.1

ZT19 74.53 15.81 2.26

ZT22 67.05 20.49 2.62

ZT19 0.20 -0.31 -0.48

ZT22 0.05 0.07 -0.27

B GRO-seq transcription - Relative ID Gene_1 Gene_2 Gene_3

ZT1 0.39 0.47 0.93

ZT4 0.33 0.16 0.45

ZT7 -0.37 -0.06 0.10

ZT10 -0.25 -0.01 -0.19

ZT13 -0.53 -0.07 -0.81

ZT16 -0.08 -0.44 -0.59

ZT10 ZT13 ZT16 ZT19 ZT22

C JTK_cycle output ID

ZT1

ZT4

ZT7

Gene_1 Gene_2 Gene_3

0.39 0.47 0.93

0.33 0.16 0.45

-0.37 -0.25 -0.53 -0.08 -0.06 -0.01 -0.07 -0.44 0.10 -0.19 -0.81 -0.59

ADJ Pvalue 0.20 0.05 0.002024 -0.31 0.07 0.002024 -0.48 -0.27 0.000244

PER AMP Phase 24.00 0.52 24.00 0.41 24.00 0.91

0.00 0.00 0.00

ADJ phase 1.00 1.00 1.00

Fig. 4 Identification of circadian gene transcription using JTK_cycle. (a) Time course of transcription (in rpktm) of three genes. (b) Input of JTK_cycle, showing relative transcription of each time point, which is derived from the log2 fold change to the mean of all time points. (c) Example of JTK_cycle output. Predicted oscillating phases were adjusted according to the time of the first data point

Since RNA transcript starts from 50 end on both stands, the negative strand peak in a bidirectional eRNA must be on the left side of the positive strand peak (Fig. 5a). Once all peaks are identified on two strands, bidirectional eRNAs can be identified by searching all negative strand peaks having a positive strand peak to the right. Appropriate distance between the paired peaks in a bidirectional eRNA locus can be estimated using a previously described method [33]. Briefly, all possible peak pairs formed by selected high-confidence peaks are generated using brute force algorithm. Distances between paired peaks in real eRNAs should follow normal distribution which can be distinguished from randomly paired peaks as illustrated in Fig. 5b. The boundary of distinct distributions is select as the distance cutoff for eRNA peak pairs. Note that the peak calling approach described here trades off stringency for extensive discovery of eRNAs, one can exclude potential false positive loci, especially some unidirectional eRNAs, by applying further filtering criteria such as histone modifications and chromatin states [22, 28]. 3.8.2 De Novo Identification of Intragenic eRNAs

eRNA transcripts originating in the gene body can be masked by gene body transcription, which likely explains why eRNAs were ignored in most previous studies. Here we introduce a protocol for identifying intragenic eRNAs: (1) Separate GRO-seq reads on + and  strand. (2) Call GRO-seq peaks on each strand using Homer

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

143

A intrargenic bidirectional

intergenic bidirectional

log2 number of pairs

B

C

Average signal profile of bidirectional intergenic eRNAs

Population 1 Population 2

Distance cutoff

log2 distance between peak pairs

2kb

Fig. 5 Identification and quantification of eRNA transcription. A. Examples of intragenic bidirectional eRNA is illustrated in the upper panel. Gene body transcription is on the negative strand and shown in red. The lower panel shows an example of intergenic bidirectional eRNA. (b) Distance distribution of GRO-seq peak pairs. Dash lines indicate two distribution populations and the intersection defines the distance cutoff for real bidirectional eRNAs. (c) Average GRO-seq signal suggests a 2 kb window for quantifying bidirectional eRNAs. GRO-seq signal on  strand is indicated by blue and red, respectively

peak finding function. (3) Use peaks on the antisense strand of genes as anchors. Bidirectional eRNAs can be identified if a paired peak can be found on the sense strand and unpaired antisense peaks are grouped into unidirectional eRNAs. (4) High-confidence eRNA loci can be determined if additional enhancer markers are available. Note that antisense transcription at TSS is a widespread phenomenon; thus, TSS regions are normally excluded for eRNA analysis [22, 28].

144

Bin Fang et al.

3.8.3 Identification of Circadian eRNAs

Sizes of eRNAs vary from site to site. Defining a window size for quantifying eRNA is a challenging open question. We use the average signal profile of eRNA transcripts at preselected bona fide enhancers to estimate the size of eRNAs (Fig. 5c) and transcription level is calculated using the following steps: 1. Intergenic bidirectional eRNAs are quantified by counting GRO-seq reads on both strands in a 2 kb window. 2. To eliminate influence from gene transcription, intragenic bidirectional eRNAs are quantified using reads on the antisense strand only, and a factor of 2 is then multiplied to estimate the total read count. 3. Unidirectional eRNAs are measured using 1 kb windows surrounding the peak summit and only reads mapped to the same strand are counted. 4. Finally, read counts for all types of eRNAs are normalized to rpkm. Once eRNA transcripts are quantified, circadian eRNAs can be detected using the same method as described in Subheading 3.7.

3.9 Identification of TFs Controlling Circadian Genes

Since eRNA transcription is correlated with their target genes [20, 22], functional enhancers of circadian genes should produce eRNAs oscillating in phase with target genes. Therefore, target genes can be predicted by “nearest neighboring in-phase gene” approach. As illustrated in Fig. 6a, compared with out-of-phase and noncircadian eRNAs, in-phase eRNAs within 100 kb of a circadian gene are more likely to be functional enhancers of that gene. When multiple in-phase circadian genes exist near an eRNA site, the enhancer will be assigned to the nearest gene. Motif mining protocol has been described previously [28]; however, eRNA loci are usually large (1–2 kb in size) making motif mining a challenge. Here, we introduce a way to precisely select central eRNA regions for motif mining. Real TF binding data such as ChIP-seq data can be used to estimate the relative position of a TF binding site within an eRNA locus. As illustrated in Fig. 6b, average signal profiles of bidirectional eRNAs and ChIP-seq peaks are centered at the same location, thus suggesting the TF binds in the center of bidirectional eRNA loci. In contrast, ChIP-seq signal is located about 200 bp upstream of unidirectional eRNAs. Thus, optimal motif mining windows for unidirectional eRNAs should be shifted accordingly. Once the window center is determined, a 100–200 bp window surrounding the center can be selected for mining TF motifs.

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . .

145

A Oscillation phase

Gene A

Gene B eRNA

B

eRNA

Bidirectional eRNAs

Unidirectional eRNAs

TF

TF

Gro-seq ChIP-seq

–1000 –500

Gro-seq ChIP-seq

0

500 1000

–1000 –500

0

500 1000

Fig. 6 Identification of TFs controlling circadian eRNAs and genes. (a) Functional enhancers of circadian genes can be found using correlated/in-phase transcriptions. Note that enhancers may not be assigned to the nearest genes. (b) Schematic models of TF-mediated eRNA transcription, above the overlay of ChIPseq signal and GRO-seq signals surrounding the centers of identified eRNAs. TF binds about 200 bp upstream of unidirectional eRNAs (indicated by the red dash line)

4

Summary and Conclusions Understanding regulated gene transcription is critical to advancing our modern understanding of biology. Within a given organism, the same genome is transcribed differently as a function of cell type, and within each cell type this is further modified by intrinsic and extrinsic signals [34]. The circadian transcriptome is particularly complex, since it derives from a clock that is present in all tissues and yet each tissue has different sets of circadian genes, operating at different circadian phases. We have demonstrated the utility of GRO-seq to identify the gene-specific functions of core clock TFs as well as other circadian TFs that drive alternatively phased transcription in the mouse liver. Here, we have explained the experimental basis of this technique, both in sample acquisition and analysis, to allow for it to be reproduced in other labs. We urge the circadian field to utilize this protocol in other tissues and species in the hope of collectively solving the puzzle of integrative transcriptional mechanisms underlying circadian gene expression.

146

5

Bin Fang et al.

Notes 1. DTT, Tween, PVP, and BSA are dissolved in ultrapure water. Add 2 μL of RNase inhibitor/10 mL of the above buffers. All buffers except for the elution buffer should be kept on ice. 2. GRO-seq also works in other tissues, such as adipocyte tissue [35] and muscle [36]. 3. Liver from one mouse typically yields 100–200 million nuclei. 4. These beads settle loosely at the bottom, so do not disturb them when moving or opening the tubes. Also, during washing, allow ~50 μL of the supernatant to remain at the bottom of the tube so as to not disrupt the beads. It is helpful to hold the tube up to a light to make sure no obvious residual beads that are floating. 5. To increase the elution efficiency, it is recommended to preheat the elution buffer to 42  C. 6. The size of product is above 50 bp, as that is the read length sequenced (single-end 50 bp).

Acknowledgments We thank Romeo Papazyan for careful reading of the manuscript. Work on circadian rhythms and GRO-seq in the Lazar lab is funded by NIH grants DK45586 and DK43806, as well as by the JPB Foundation. References 1. Bass J, Lazar MA (2016) Circadian time signatures of fitness and disease. Science 354 (6315):994–999. https://doi.org/10.1126/ science.aah4965 2. Zhang EE, Kay SA (2010) Clocks not winding down: unravelling circadian networks. Nat Rev Mol Cell Biol 11(11):764–776. https://doi. org/10.1038/nrm2995 3. Feng D, Liu T, Sun Z, Bugge A, Mullican SE, Alenghat T, Liu XS, Lazar MA (2011) A circadian rhythm orchestrated by histone deacetylase 3 controls hepatic lipid metabolism. Science 331(6022):1315–1319. https://doi. org/10.1126/science 4. Feng D, Lazar MA (2012) Clocks, metabolism, and the epigenome. Mol Cell 47 (2):158–167. https://doi.org/10.1016/j. molcel.2012.06.026 5. Geertz M, Maerkl SJ (2010) Experimental strategies for studying transcription factor-DNA

binding specificities. Brief Funct Genomics 9 (5-6):362–373. https://doi.org/10.1093/ bfgp/elq023 6. Berger MF, Philippakis AA, Qureshi AM, He FS, Estep PW 3rd, Bulyk ML (2006) Compact, universal DNA microarrays to comprehensively determine transcription-factor binding site specificities. Nat Biotechnol 24 (11):1429–1435. https://doi.org/10.1038/ nbt1246 7. Fang B, Mane-Padros D, Bolotin E, Jiang T, Sladek FM (2012) Identification of a binding motif specific to HNF4 by comparative analysis of multiple nuclear receptors. Nucleic Acids Res 40(12):5343–5356. https://doi.org/10. 1093/nar/gks190 8. Slattery M, Zhou T, Yang L, Dantas Machado AC, Gordan R, Rohs R (2014) Absence of a simple code: how transcription factors read the

Using GRO-Seq to Measure Circadian Transcription and Discover Circadian. . . genome. Trends Biochem Sci 39(9):381–399. https://doi.org/10.1016/j.tibs.2014.07.002 9. Buenrostro JD, Wu B, Chang HY, Greenleaf WJ (2015) ATAC-seq: a method for assaying chromatin accessibility genome-wide. Curr Protoc Mol Biol 109(29):21–29. https://doi. org/10.1002/0471142727.mb2129s109 10. Valouev A, Johnson DS, Sundquist A, Medina C, Anton E, Batzoglou S, Myers RM, Sidow A (2008) Genome-wide analysis of transcription factor binding sites based on ChIPSeq data. Nat Methods 5(9):829–834. https:// doi.org/10.1038/nmeth.1246 11. Boyle AP, Davis S, Shulha HP, Meltzer P, Margulies EH, Weng Z, Furey TS, Crawford GE (2008) High- resolution mapping and characterization of open chromatin across the genome. Cell 132(2):311–322. https://doi. org/10.1016/j.cell.2007.12.014 12. Calo E, Wysocka J (2013) Modification of enhancer chromatin: what, how, and why? Mol Cell 49(5):825–837. https://doi.org/ 10.1016/j.molcel.2013.01.038 13. Hon GC, Hawkins RD, Ren B (2009) Predictive chromatin signatures in the mammalian genome. Hum Mol Genet 18(R2): R195–R201. https://doi.org/10.1093/ hmg/ddp409 14. Dostie J, Richmond TA, Arnaout RA, Selzer RR, Lee WL, Honan TA, Rubio ED, Krumm A, Lamb J, Nusbaum C, Green RD, Dekker J (2006) Chromosome conformation capture carbon copy (5C): a massively parallel solution for mapping interactions between genomic elements. Genome Res 16 (10):1299–1309. https://doi.org/10.1101/ gr.5571506 15. Lieberman-Aiden E, van Berkum NL, Williams L, Imakaev M, Ragoczy T, Telling A, Amit I, Lajoie BR, Sabo PJ, Dorschner MO, Sandstrom R, Bernstein B, Bender MA, Groudine M, Gnirke A, Stamatoyannopoulos J, Mirny LA, Lander ES, Dekker J (2009) Comprehensive mapping of long-range interactions reveals folding principles of the human genome. Science 326 (5950):289–293. https://doi.org/10.1126/ science.1181369 16. Hakim O, Misteli T (2012) SnapShot: Chromosome confirmation capture. Cell 148 (5):1068. https://doi.org/10.1016/j.cell. 2012.02.019 17. Step SE, Lim HW, Marinis JM, Prokesch A, Steger DJ, You SH, Won KJ, Lazar MA (2014) Anti-diabetic rosiglitazone remodels the adipocyte transcriptome by redistributing transcription to PPARgammadriven

147

enhancers. Genes Dev 28(9):1018–1028. https://doi.org/10.1101/gad.237628.114 18. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R, Genome Project Data Processing S (2009) The sequence alignment/map format and SAMtools. Bioinformatics 25 (16):2078–2079. https://doi.org/10.1093/ bioinformatics/btp352 19. Kim TK, Hemberg M, Gray JM, Costa AM, Bear DM, Wu J, Harmin DA, Laptewicz M, Barbara-Haley K, Kuersten S, MarkenscoffPapadimitriou E, Kuhl D, Bito H, Worley PF, Kreiman G, Greenberg ME (2010) Widespread transcription at neuronal activity-regulated enhancers. Nature 465(7295):182–187. https://doi.org/10.1038/nature09033 20. Hah N, Murakami S, Nagari A, Danko CG, Kraus WL (2013) Enhancer transcripts mark active estrogen receptor binding sites. Genome Res 23(8):1210–1223. https://doi.org/10. 1101/gr.152306.112 21. Li W, Notani D, Rosenfeld MG (2016) Enhancers as non-coding RNA transcription units: recent insights and future perspectives. Nat Rev Genet 17(4):207–223. https://doi.org/ 10.1038/nrg.2016.4 22. Fang B, Everett LJ, Jager J, Briggs E, Armour SM, Feng D, Roy A, Gerhart-Hines Z, Sun Z, Lazar MA (2014) Circadian enhancers coordinate multiple phases of rhythmic gene transcription in vivo. Cell 159(5):1140–1152. https://doi.org/10.1016/j.cell.2014.10.022 23. Langmead B, Trapnell C, Pop M, Salzberg SL (2009) Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol 10(3):R25. https:// doi.org/10.1186/gb-2009-10-3-r25 24. Quinlan AR, Hall IM (2010) BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26(6):841–842. https://doi.org/10.1093/bioinformatics/ btq033 25. Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, Cheng JX, Murre C, Singh H, Glass CK (2010) Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell 38 (4):576–589. https://doi.org/10.1016/j. molcel.2010.05.004 26. Hughes ME, Hogenesch JB, Kornacker K (2010) JTK_CYCLE: an efficient nonparametric algorithm for detecting rhythmic components in genome-scale data sets. J Biol Rhythm 25(5):372–380. https://doi.org/10. 1177/0748730410379711

148

Bin Fang et al.

27. Robinson JT, Thorvaldsdottir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP (2011) Integrative genomics viewer. Nat Biotechnol 29(1):24–26. https://doi.org/10. 1038/nbt.1754 28. Nagari A, Murakami S, Malladi VS, Kraus WL (2017) Computational approaches for mining GRO-Seq data to identify and characterize active enhancers. Methods Mol Biol 1468:121–138. https://doi.org/10.1007/ 978-1-4939-4035-6_10 29. Menet JS, Rodriguez J, Abruzzi KC, Rosbash M (2012) Nascent-Seq reveals novel features of mouse circadian transcriptional regulation. Elife 1:e00011. https://doi.org/10.7554/ eLife.00011 30. Anders S, Pyl PT, Huber W (2015) HTSeq--a python framework to work with highthroughput sequencing data. Bioinformatics 31(2):166–169. https://doi.org/10.1093/ bioinformatics/btu638 31. Vollmers C, Schmitz RJ, Nathanson J, Yeo G, Ecker JR, Panda S (2012) Circadian oscillations of protein- coding and regulatory RNAs in a highly dynamic mammalian liver epigenome. Cell Metab 16(6):833–845. https://doi. org/10.1016/j.cmet.2012.11.004 32. Koike N, Yoo SH, Huang HC, Kumar V, Lee C, Kim TK, Takahashi JS (2012) Transcriptional architecture and chromatin landscape of the core circadian clock in mammals. Science 338(6105):349–354. https://doi. org/10.1126/science.1226339

33. Fullwood MJ, Liu MH, Pan YF, Liu J, Xu H, Mohamed YB, Orlov YL, Velkov S, Ho A, Mei PH, Chew EG, Huang PY, Welboren WJ, Han Y, Ooi HS, Ariyaratne PN, Vega VB, Luo Y, Tan PY, Choy PY, Wansa KD, Zhao B, Lim KS, Leow SC, Yow JS, Joseph R, Li H, Desai KV, Thomsen JS, Lee YK, Karuturi RK, Herve T, Bourque G, Stunnenberg HG, Ruan X, Cacheux-Rataboul V, Sung WK, Liu ET, Wei CL, Cheung E, Ruan Y (2009) An oestrogen-receptor-alpha-bound human chromatin interactome. Nature 462(7269):58–64. https://doi.org/10.1038/nature08497 34. Heinz S, Romanoski CE, Benner C, Glass CK (2015) The selection and function of cell typespecific enhancers. Nat Rev Mol Cell Biol 16 (3):144–154. https://doi.org/10.1038/ nrm3949 35. Jager J, Wang F, Fang B, Lim HW, Peed LC, Steger DJ, Won KJ, Kharitonenkov A, Adams AC, Lazar MA (2016) The nuclear receptor rev-erbalpha regulates adipose tissue-specific FGF21 signaling. J Biol Chem 291 (20):10867–10875. https://doi.org/10. 1074/jbc.M116.719120 36. Hong S, Zhou W, Fang B, Lu W, Loro E, Damle M, Ding G, Jager J, Zhang S, Zhang Y, Feng D, Chu Q, Dill BD, Molina H, Khurana TS, Rabinowitz JD, Lazar MA, Sun Z (2017) Dissociation of muscle insulin sensitivity from exercise endurance in mice by HDAC3 depletion. Nat Med 23 (2):223–234. https://doi.org/10.1038/nm. 4245

Chapter 11 Circadian Metabolomics from Breath Steven A. Brown and Pablo Sinues Abstract Metabolites like melatonin are essential in determining circadian phase. In the recent years, comprehensive metabolome analyses have unveiled entire panels of small biomolecules fluctuating in a circadian fashion, thus enabling a more precise determination of inner time and understanding of how circadian clock operates at the molecular level. Emerging analytical techniques allowing for the determination of exhaled metabolites in breath show promise to gain further insights noninvasively and in vivo into circadian metabolism. Key words Metabolomics, Breath analysis, Real-time analysis, Secondary electrospray ionization

1

Introduction Metabolites are downstream end products resulting from the interaction between genes and environmental factors. The dynamic response of changes of metabolite levels in body fluids to environmental factors is very fast as compared to genes and proteins. This makes them ideal molecular indicators to track and gain insights into the mechanisms of action of the circadian clock. As a matter of fact, small molecules like melatonin and cortisol have been widely used as molecular beacons indicating circadian phase [1]. In the recent years, a number of global metabolomics initiatives have extended the analysis of the circadian clock at the metabolic level. As a result, a comprehensive overview of panels of metabolites modulated by the circadian clock (i.e., fluctuating in a 24 h cycle) has been unveiled, thus gaining insights into metabolic pathways activated during different circadian phases [2–5]. Such global metabolomics studies typically rely on the analysis of body fluids (e.g., plasma or urine) using well-established analytical platforms like liquid chromatography–mass spectrometry (LC-MS) and nuclear magnetic resonance (NMR). Because the goal is to identify rhyth-

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_11, © Springer Science+Business Media, LLC, part of Springer Nature 2021

149

150

Steven A. Brown and Pablo Sinues

mic metabolites fluctuating during 24 h cycles, body fluid samples are collected at regular intervals (e.g., every 2 h) to capture with sufficient time resolution such changes in metabolites levels. During the last decade, less conventional analytical approaches have arisen. The goal has been to achieve the detection of volatile and semivolatile molecules in exhaled breath, yet in real time [6, 7]. One such technique is secondary electrospray ionizationmass spectrometry (SESI-MS) [8]. This technique sacrifices selectivity and sensitivity as compared to LC-MS or NMR, but, on the other hand, it represents a convenient way to probe metabolite levels in a completely noninvasive way from a nearly unlimited body fluid. This makes real-time breath analysis an especially attractive technique because it allows for probing metabolite levels with virtually limitless time resolution. SESI-MS has been exploited by us and several authors to study volatile metabolites in a wide range of applications [9–12]. In a proof-of-principle study, we have shown the feasibility of the concept by using real-time breath metabolomics to track species fluctuating in a circadian fashion [13]. Breath samples were conveniently acquired hourly during a 24-h circadian cycle.

2

Materials

2.1 Breath Analysis Interface

1. A commercially available disposable round mouthpiece featuring a one-way valve for exhalation-only testing. 2. The breath sampling tube connected to the SESI ion source (See Notes 1, 2) can be of Teflon or stainless steel (recommended inner diameter, I.D. 4–6 mm). The tube should be heated to at least 130  C to prevent vapor adsorption onto the tube walls (See Note 3). 3. A digital manometer in the range 0–30 mbar.

2.2 Secondary Electrospray Ionization

1. Ultrapure water (18 MΩ-cm at 25  C). 2. Formic acid. 3. Liquid reservoir (1–5 mL). 4. Noncoated silica capillary I.D. 20 μm. Nanospray emitters are commercially available (SilicaTip™ Emitters). 5. Platinum wire. 6. Electrometer.

2.3 Mass Spectrometry

Ionization of vapor species in SESI takes place at atmospheric pressure. For this reason, it is recommended to apply SESI-MS using any commercially available atmospheric pressure ionization

Circadian Metabolomics from Breath

151

mass spectrometer (API-MS). Further modifications of the ionization stage are recommended to maximize sensitivity. 1. For untargeted metabolomics a high resolution and high mass accuracy instrument is preferred because it allows covering a wide range of species and often enables unambiguous determination of molecular formulae. A quadrupole time-of-flight (Q-TOF) with resolutions in the order of 20,000, or even higher resolution (i.e., >70,000) mass analyzers based on Orbitrap technology are preferred. 2. For targeted analysis—where a set of known metabolites are meant to be detected—a triple quadrupole mass spectrometer operating in multiple reaction-monitoring mode is preferred. This enables high sensitivity and high selectivity toward the species of interest. 3. The standard electrospray ion source delivered together with commercial mass spectrometers have been optimized to efficiently ionize analytes from the liquid phase—typically delivered by an LC system upstream. However, they tend to perform poorly to ionize volatile species. For this reason, an embodiment where an efficient ionization of the vapors take place is recommended. Such embodiments can be machined in-house when a mechanical workshop is available. Alternatively, commercial plug-and-play SESI sources can also be purchased for preexisting mass spectrometers (e.g., FIT, Spain).

3 3.1

Methods SESI Ion Source

1. An in-house electrospray source coupled to a commercial quadrupole time-of-flight mass spectrometer (Sciex, Canada) is shown in Fig. 1. It shows all the necessary elements to perform SESI-MS. Further details of the embodiment can be seen in these patents [14, 15] (See Notes 1, 2). 2. Set the voltage to 4 kV when operating in positive ion mode or to 3.2 kV in negative mode. 3. Pressurize the liquid reservoir with compressed air or nitrogen to 1 bar. For a capillary of 50 cm long, 20 μm I.D. and water (0.2% formic acid) as SESI solvent (See Note 5), it provides a flow rate in the range of 50 nL/min (Poiseuille flow formula). 4. After a couple of minutes, the liquid emerges at the emitter tip to form the electropray (See Note 4). Adjust the voltage as necessary to operate in the so-called cone jet mode. If the SESI chambers features windows the formation of the electropray can be monitored by visual inspection by using a magnifier and

152

Steven A. Brown and Pablo Sinues

Fig. 1 Lab-built SESI ion source for the real-time analysis of vapors. The SESI chamber is hooked on a commercial mass spectrometer. Water (0.1–0.2% formic acid) in a reservoir is pushed at about 10 nL/min through a silica capillary (50 cm long; 20 μm I.D.). At the same time, the water is electrified at around 4 kV, thus forming an electrospray with currents in the range of 100 nA (read in this setup on the multimeter). Formation of a stable spray can additionally be inspected visually with a magnifier. Neutral vapors enter the SESI chamber, whereby some of the species are ionized and mass analyzed in real time. Excess gas is discharged to the room through the exhaust port. The interlock is necessary for the mass spectrometer to allow operating an external ion source

Fig. 2 (a) Picture of the sharpened tip of the silica tube (o.d. 360 μm, i.d. 20 μm). The o.d. of the tip was approximately 50 μm. (b) Detail of the silica needle-grounded electrode. Observe the cone-jet of water anchored at the needle exit. (c) Steady cone-jet of deionized water in air (reprinted with permission from [16])

a torch. Figure 2 shows an electropray of water operating the cone-jet mode (adapted from [16]). Typical current should be in the range of 100 nA as measured by an electrometer.

Circadian Metabolomics from Breath

3.2 Mass Spectrometry 3.2.1 Positive Ion Mode

153

1. In full MS mode, scan in the range 50–500 Da with an accumulation time of 1 s. In especially sensitive setups, species with molecular weights approaching 1000 Da can be detected, and in this case the scanning mass range should therefore be extended [17]. 2. When the electrospray is on, the mass spectrum will show typical background contaminants in the entire mass range. Some of these chemical interferences are well known and are useful as lock-masses to ensure a high mass accuracy (within 3 ppm) across hours of experiments. Typical background signals covering the mass range of interest include acetone, phthalates, and polysiloxanes [18]. 3. Most of the species will be detected in the protonated form [M + H]+; however, ammonium and sodium adducts are also common in positive ion mode. 4. Typical metabolites detected in positive ion mode include N-containing compounds (e.g., urea and indole), aldehydes, and even amino acids [19, 20].

3.2.2 Negative Ion Mode

1. In full MS mode, scan in the range 50–400 Da with an accumulation time of 1 s. 2. When the electrospray is on, the mass spectrum will show a typical series of fatty acids that can be used as calibrants [18]. 3. In negative ion mode most species are detected in the deprotonated form [M H] ; however, formic acid adducts are also observable.

3.3 Breath Maneuvers

1. Acquire the background spectrum for at least 30 s, ensuring that the total ion current measured by the mass analyzer remains stable. 2. Provide a full exhalation through the mouthpiece ensuring that the overpressure delivered is in the range of 10 mbar (as read by the breathing subject in the digital manometer). The corresponding optimal flow rate will vary depending on the particular geometry of the flow restriction but will typically be in the range of 0.2–1 L/min. 3. As the subject breathes, the TIC should increase steeply to eventually reach a plateau. The time to reach the plateau will depend on the exhalation flow rate but will typically be in the order of 10–15 s. 4. Once the full exhalation is provided, the TIC levels should decrease steeply to the previous background levels.

154

Steven A. Brown and Pablo Sinues

Fig. 3 (a) Example of one compound rising during the six replicate exhalations of one participant. Note that in less than 3 min, six repeatable breathprints are recorded. Average signal at the plateau of the replicate exhalations was computed for all three participants, and if the intensity at all the time points was lower than 4000 counts, the feature was not considered. (b) Final data set showcasing inter- and intraindividual differences of an exhaled compound (tentatively identified as indole; m/z 118.0675) during 24 h for three different participants. Note the tightly clustered replicate breath for the same clock time. Red shading signifies 95% confidence interval around the mean. The separation between hourly measurements has been arbitrarily set to ease visualization

5. Repeat the process. Figure 3a shows six replicate exhalations while monitoring the channel m/z 118.0675, corresponding to indole. Figure 3b shows indole levels measured hourly for three subjects during 24 h (adapted from [13]).

4

Notes 1. Should an in-house system be developed, care must be taken to avoid any electrical shock, as the electrospray solvent is electrified at high voltage. A commercially available electrospray source is depicted in Fig. 4. 2. Seek advice from the MS vendor on the details of the interlock in order to bypass the instrument to use an in-house made SESI source. 3. The sampling tube should be heated to prevent adsorption of exhaled metabolites. To prevent burns, the heating element should be insulated. 4. The emitters clog frequently. Should the electrospray stop working, flush and sonicate the emitter. 5. Prepare fresh water–formic acid solutions at least once per month and keep them in the fridge.

Circadian Metabolomics from Breath

155

Fig. 4 Commercially available SESI source for real-time breath analysis

References 1. Sharma M, Palacios-Bois J, Schwartz G, Iskandar H, Thakur M, Quirion R, Nair NPV (1989) Circadian rhythms of melatonin and cortisol in aging. Biol Psychiatry 25 (3):305–319. https://doi.org/10.1016/ 0006-3223(89)90178-9 2. Dallmann R, Viola AU, Tarokh L, Cajochen C, Brown SA (2012) The human circadian metabolome. Proc Natl Acad Sci U S A 109 (7):2625–2629. https://doi.org/10.1073/ pnas.1114410109 3. Davies SK, Ang JE, Revell VL, Holmes B, Mann A, Robertson FP, Cui N, Middleton B, Ackermann K, Kayser M, Thumser AE, Raynaud FI, Skene DJ (2014) Effect of sleep deprivation on the human metabolome. Proc Natl Acad Sci U S A 111(29):10761–10766 4. Kasukawa T, Sugimoto M, Hida A, Minami Y, Mori M, Honma S, Honma K, Mishima K, Soga T, Ueda HR (2012) Human blood metabolite timetable indicates internal body time. Proc Natl Acad Sci U S A 109 (37):15036–15041. https://doi.org/10. 1073/pnas.1207768109 5. Minami Y, Kasukawa T, Kakazu Y, Iigo M, Sugimoto M, Ikeda S, Yasui A, van der Horst GTJ, Soga T, Ueda HR (2009) Measurement of internal body time by blood metabolomics. Proc Natl Acad Sci U S A 106(24):9890–9895. https://doi.org/10.1073/pnas.0900617106 6. Jens H et al (2009) On-line breath analysis with PTR-TOF. J Breath Res 3(2):027004

7. Sˇpaneˇl P, Smith D (2013) Chapter 4 - Recent SIFT-MS Studies of Volatile Compounds in Physiology, Medicine and Cell Biology. In: Amann A, Smith D (eds) Volatile Biomarkers. Elsevier, Boston, p 48–76 8. Martinez-Lozano Sinues P, Zenobi R, Kohler M (2013) Analysis of the exhalome: a diagnostic tool of the future. Chest 144(3):746–749. https://doi.org/10.1378/chest.13-1106 9. Zhu JJ, Bean HD, Jimenez-Diaz J, Hill JE (2013) Secondary electrospray ionizationmass spectrometry (SESI-MS) breathprinting of multiple bacterial lung pathogens, a mouse model study. J Appl Physiol 114 (11):1544–1549. https://doi.org/10.1152/ japplphysiol.00099.2013 10. Bean HD, Zhu J, Hill JE (2011) Characterizing bacterial volatiles using secondary electrospray ionization mass spectrometry (SESI-MS). J Vis Exp 52:e2664. https://doi.org/10.3791/2664 11. Dillon LA, Stone VN, Croasdell LA, Fielden PR, Goddard NJ, Paul Thomas CL (2010) Optimisation of secondary electrospray ionisation (SESI) for the trace determination of gas-phase volatile organic compounds. Analyst 135(2):306–314 12. Zhu J, Bean HD, Kuo YM, Hill JE (2010) Fast detection of volatile organic compounds from bacterial cultures by secondary electrospray ionization-mass spectrometry. J Clin Microbiol 48(12):4426–4431. https://doi.org/10. 1128/JCM.00392-10

156

Steven A. Brown and Pablo Sinues

13. Martinez-Lozano Sinues P, Tarokh L, Li X, Kohler M, Brown SA, Zenobi R, Dallmann R (2014) Circadian variation of the human metabolome captured by real-time breath analysis. PLoS One 9(12):e114422. https://doi.org/ 10.1371/journal.pone.0114422 14. Martinez-Lozano Sinues P, Fernandez de la Mora J (2015) Method to analyze and classify persons and organisms based on odor patterns from released vapors; US Patent No: US9121844 B1 15. Martinez-Lozano Sinues P, Fernandez de la Mora J (2010) Method for detecting volatile species of high molecular weight; US Patent No: US 20100264304 A1 16. Lo´pez-Herrera J, Barrero A, Boucard A, Loscertales I, Ma´rquez M (2004) An experimental study of the electrospraying of water in air at atmospheric pressure. J Am Soc Mass Spectrom 15(2):253–259. https://doi.org/ 10.1016/j.jasms.2003.10.018 17. Gaugg MT, Garcia Gomez D, Barrios Collado C, Vidal de Miguel G, Kohler M, Zenobi R, Martinez-Lozano Sinues P (2016) Expanding metabolite coverage of real-time breath analysis by coupling a universal

secondary electrospray ionization source and high resolution mass spectrometry—a pilot study on tobacco smokers. J Breath Res 10 (1):016010 18. Keller BO, Sui J, Young AB, Whittal RM (2008) Interferences and contaminants encountered in modern mass spectrometry. Anal Chim Acta 627(1):71–81. https://doi. org/10.1016/j.aca.2008.04.043 19. Garcı´a-Go´mez D, Gaisl T, Bregy L, Cremonesi A, Sinues PM-L, Kohler M, Zenobi R (2016) Real-time quantification of amino acids in the Exhalome by secondary electrospray ionization–mass spectrometry: a proofof-principle study. Clin Chem 62 (9):1230–1237. https://doi.org/10.1373/ clinchem.2016.256909 20. Garcı´a-Go´mez D, Martı´nez-Lozano Sinues P, Barrios-Collado C, Vidal-De-Miguel G, Gaugg M, Zenobi R (2015) Identification of 2-alkenals, 4-hydroxy-2-alkenals, and 4-hydroxy-2,6-alkadienals in exhaled breath condensate by UHPLC-HRMS and in breath by real-time HRMS. Anal Chem 87 (5):3087–3093. https://doi.org/10.1021/ ac504796p

Chapter 12 A Lipidomics View of Circadian Biology Rona Aviram, Chunyan Wang, Xianlin Han, and Gad Asher Abstract Lipidomics approaches provide quantitative characterization of hundreds of lipid species from biological samples. Recent studies highlight the value of these methods in studying circadian biology, and their potential goes far beyond studying lipid metabolism per se. For example, lipidomics analyses of subcellular compartments can be used to determine daily rhythmicity of different organelles and their intracellular dynamics. In this chapter we describe in detail the procedure for around the clock shotgun lipidomics, from sample preparation to bioinformatics analyses. Sample preparation includes biochemical fractionation of nuclei and mitochondria from mouse liver harvested throughout the day. Lipid content is determined and quantified, in unbiased manner and with wide coverage, using multidimensional mass spectrometry shotgun lipidomics (MDMS-SL). Circadian parameters are then determined with nonparametric statistical tests. Overall, the approach described herein is applicable for various animal models, tissues, and organelles, and is expected to yield new insight on various aspects of circadian biology and lipid metabolism. Key words Shotgun lipidomics, Mass spectrometry, Circadian, Mouse, Liver, Nuclei, Mitochondria, Lipid metabolism

1

Introduction Lipids play essential roles in the structure/function of biological membranes of cells and organelles, and thus measurably define their identity [1]. Therefore, comprehension of lipid composition is essential in understanding the related functionality. Alongside increasing evidence tying circadian clocks with lipid homeostasis [2–4], technological advances evinced lipidomics to be a powerful tool that can be used as a reliable “time-teller,” and readout for circadian biology [5–7]. Specifically, lipidomics analyses of subcellular compartments throughout the day can be used to determine spatial and temporal dynamics in lipid composition and uncover daily rhythmicity in intracellular organelles.

Rona Aviram and Chunyan Wang contributed equally to this work. Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_12, © Springer Science+Business Media, LLC, part of Springer Nature 2021

157

158

Rona Aviram et al.

We commence with a detailed description of the protocol used for biochemical isolation of nuclei and mitochondria from mouse tissue around the clock. These intracellular organelles were selected in view of their critical role in circadian biology and metabolism. The sample isolation procedure is based on multiple centrifugation steps, performed in different sucrose-based media. Subsequently, samples are homogenized and spiked with appropriate internal standards. Lipids are extracted and analyzed by multidimensional mass spectrometry shotgun lipidomics (MDMS-SL). This technique provides quantitative measurements of >95% of the lipid mass, covering hundreds to thousands of lipid molecular species [8–10]. Depending on the nature of the biological sample, the technology described herein is expected to yield quantitative measurements of glycerophospholipids (e.g., cardiolipin, phosphatidic acid, phosphatidylcholine, phosphatidylethanolamine, phosphatidylglycerol, phosphatidylinositol, phosphatidylserine), lysoglycerophospholipids (e.g., lysocardiolipin, lysophosphatidylcholine, lysophosphatidylethanolamine), sphingolipids (e.g., ceramide and sphingomyelin) and glycerolipids such as triacylglycerols. These molecules are characterized according to class and biochemical structure: acyl-chain length and linkage type (if relevant), and degree of saturation (i.e., number of double bonds). Finally, the circadian analysis of the daily lipid profiles is performed using statistical tests such as the one offered in JTK_CYCLE algorithm [11]. Additional commonly used statistical analyses can be applied to dissect the lipid composition and dynamics of the different organelles. The method described herein provides the lipid landscape of nuclei and mitochondria isolated from mouse liver throughout the day, and therefore can be used to monitor their “internal” time. This approach can be applied on different animal models, tissues, and organelles and thereby unravel various novel facets of circadian biology and lipid metabolism.

2

Materials

2.1 Sample Preparation

∗ All instruments and reagents are to be cold during use (4
C). Since storing in cold may cause rust, preferably cool only prior to use.

2.1.1 Nuclei Isolation from Mouse Liver

Supplement all buffers with 0.5 mM DTT, 100 μM NaF, 100 μM Na2VO4, 0.5 μM PMSF, and 1:500 Protease Inhibitor cocktail (we use CALBIOCHEM). 1. Homogenization buffer A, 0.3 M sucrose: 10 mM Hepes pH 7.6, 15 mM KCl, 2 mM EDTA, 0.15 mM spermine,

Circadian Lipidomics

159

0.5 mM spermidine, 10% glycerol (prepare ~4 mL per liver sample).1 2. Homogenization buffer B, 2.2 M sucrose: 10 mM Hepes pH 7.6, 15 mM KCl, 2 mM EDTA, 0.15 mM spermine, 0.5 mM spermidine, 10% glycerol (prepare 25 mL per liver sample: 7 for homogenization, 18 for homogenate dilution).2 3. Cushion buffer 2.05 M sucrose: 10 mM Hepes pH 7.6, 15 mM KCl, 2 mM EDTA, 0.15 mM spermine, 0.5 mM spermidine, 10% glycerol (prepare ~11 mL per liver sample).3 4. Isotonic buffer: 10 mM Hepes pH 7.6, 100 mM KCl, 0.15 mM spermine, 0.5 mM spermidine, 10% glycerol. Store in 4
C for up to 6 months. 5. Cold phosphate-buffered saline (PBS). 6. Swinging-bucket ultracentrifuge (4
C), we use Beckman, SW28 rotor. 7. Beckman ultracentrifuge polycarbonate4 tubes, suitable for the above rotor (40 mL volume). 8. Electric homogenizer (we use IKA 3720001 T-18 Ultra-Turrax Digital Homogenizer). 9. Dounce homogenizer: consisting of a glass tube with a tightfitting Teflon pestle (we use KONTES glass 23, 15 mL capacity at least). 10. Vacuum pump. 11. Syringe (of any kind, suitable to wash a tube as described in Subheading 3.2.1). 12. Spatula. 13. Microtubes (1.5 mL). 14. Delicate paper wipes (e.g., Kimwipes). 2.1.2 Mitochondria Isolation from Mouse Liver

1

1. Mitochondrial Isolation Buffer (MIB): 70 mM sucrose, 200 mM mannitol, 10 mM HEPES, 1 mM EDTA, pH 7.5.5 Buffer can be stored in 4
C for up to 6 months (visually check for contaminations before use, filtration is recommended). 2. MIB supplemented with 0.2% bovine serum albumin (BSA) fatty-acid free.6

Melt sucrose on a slightly warm plate, prior to addition of other ingredients. Prepare buffer in advance, at least 1 day prior to fractionation. This ensures hardening which will sustain the biological sample. This buffer can be stored in 4
C for up to 1 month. Titer to pH with either KOH or NaOH. 2 See footnote 1. 3 See footnote 1. 4 Polycarbonate is important for reuse of tubes, from our experience other materials might damage when reusing. 5 Titer with KOH to improve extraction. 6 Its purpose is to remove fat tissue from the sample.

160

Rona Aviram et al.

3. Dounce homogenizer, consisting of a glass tube with a tightfitting Teflon pestle (we use KONTES glass 23), use 15 mL capacity at least. 4. Refrigerated centrifuge (4
C) with SS34 rotor. 5. Round bottom tubes suitable for the above rotor. 6. Glass rod. 2.2 Shotgun Lipidomics

Quantity of biological samples for analysis: >0.2 mg protein of nuclei or >0.4 mg protein of mitochondria.7 1. Nano-ESI source device (TriVersa NanoMate, Advion Bioscience Ltd., Ithaca, NY, USA). 2. Chipsoft 8.3.1 software (TriVersa NanoMate, Advion Bioscience Ltd., Ithaca, NY, USA). 3. Mass spectrometer (Thermo TSQ Quantiva™ Triple Quadrupole Mass Spectrometer, San Jose, CA, USA). 4. Xcalibur™ software (Thermo Fisher Scientific, San Jose, CA, USA). 5. Kontes Microtube Pellet Pestle Rods with Motor (we use Daigger Scientific Inc., Vernon Hills, IL, USA). 6. 6 mL and 10 mL reusable culture tube with PTFE lined cap. 7. 5.7500 disposable borosilicate glass Pasteur pipets (we use Thermo Fisher Scientific). 8. Vortex mixer. 9. Tabletop centrifuge. 10. 24 position N-EVAP nitrogen evaporator (we use Organomation Associates, Inc., Berlin, MA). 11. Chemical resistance 96-well microplates. 12. Calibrated micropipettes. 13. Chloroform. 14. Methanol. 15. Methyl-tert-butyl ether (MTBE). 16. Millipore deionized water. 17. Hydrochloric acid. 18. Glacial acetic acid. 19. Trimethylsilyl diazomethane solution 2.0 M in hexanes. 20. Lithium chloride. 21. Isopropanol. 22. Lithium hydroxide.

7

Else, could be performed on 106 cells, 100 μL body fluids

Circadian Lipidomics

161

23. BCA protein assay kit (we use Fisher Scientific). 24. Lipid internal standards8: (a) 1,2-Dimyristoleoyl-sn-glycero-3-phosphocholine (di14:1 PC). (b) 1,2-Dipalmitoleoyl-sn-glycero-3-phosphoethanolamine (di16:1 PE). (c) 1,2-Dipentadecanoyl-sn-glycero-3-phosphoglycerol (sodium salt) (di15:0 PG). (d) 1,2-Dimyristoyl-sn-glycero-3-phospho-L-serine (sodium salt) (di14:0 PS). (e) 1,2-Dimyristoyl-sn-glycero-3-phosphate (sodium salt) (di14:0 PA). (f) 1,10 ,2,20 -Tetramyristoyl cardiolipin (T14:0 CL). (g) 7,7,8,8-d4-Palmitic acid (d4-16:0 NEFA) (Cambridge Isotope Laboratories, Andover, MA, USA). (h) N-Lauroyl sphingomyelin (N12:0 SM). (i) N-Heptadecanoyl ceramide (N17:0 Cer). (j) 1-Heptadecanoyl-2-hydroxy-sn-glycero-3-phosphocholine (17:0 lysoPC). (k) 1-Myristoyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (17:0 lysoPE). (l) 1,2,3,4-13C4-Palmitoyl-L-carnitine hydrochloride (13C416:0 CN) (Sigma-Aldrich, St. Louis, MO, USA). (m) Triheptadecenoin (T17:1 TAG) (Nu Chek, Inc. Elysian, MN).

3 3.1

Methods Mice Handling

3.2 Liver-Organelle Isolation

8

Use three-month-old C57BL male mice. Mice are housed under routine light schedule of 12-h light/dark regimen for at least 2 weeks to allow accommodation. Mice are sacrificed by cervical dislocation or carbon dioxide exposure at the desired light regimen, namely, under light/dark or dark/dark, according to the experimental aims. Livers should be harvested immediately upon sacrifice.9 Perform all stages on ice see Fig. 1.

We purchase all of the lipid internal standards from Avanti Polar Lipids, Inc., Alabaster, AL, USA except if otherwise noted. 9 In case of liver analysis without fractionation rapidly freeze the tissue in liquid nitrogen and keep frozen until lipidomics.

162

Rona Aviram et al.

Fig. 1 General scheme of nuclear and mitochondrial fraction isolation from mouse liver. The isolation procedure is based on multiple centrifugation steps, performed in different sucrose-based media. Nuclei isolation can be done with a liver which has been frozen in 80
C, while for mitochondria only fresh liver is to be used. For whole liver analysis without fractionation rapidly freeze the tissue in liquid nitrogen and keep frozen until lipidomics 3.2.1 Nuclei Isolation from Mouse Liver10

1. Rinse liver in cold PBS. 2. Place liver sample in glass homogenizer tube, complete volume to 5 mL with 0.3 M sucrose Homogenization buffer A (i.e., add ~4 mL). 3. Add 7 mL of 2.2 M sucrose Homogenization buffer B. 4. Homogenize liver using tight fitting pestle at 1150–1200 rpm (according to IKA motor setup). Run through the sample with three vertical motions. Avoid air bubbles. 5. Following homogenization, add 18 mL of 2.2 M Homogenization buffer B. 6. Shake vigorously by hand. 7. Prepare Beckman ultracentrifugation tube with 10.5 mL Cushion buffer 2.05 M sucrose. 8. Softly place the sample on top of the Cushion.11 9. Centrifuge for 1 h, at 48,490  g, at 4
C.

Nuclei isolation can be done with a liver which has been frozen in 80
C. This is a crucial step: materials must be very cold to avoid turbulence. Transfer the sample in slope, gently on the wall of the tube. The two phases must not mix, this would destroy the separation. If mixture occurs cease immediately and resume with the rest of the sample on top of a new Cushion. The mixed samples cannot be used.

10 11

Circadian Lipidomics

163

10. Remove the upper fat layer (using spatula or other instrument of choice). 11. Remove the red fraction with vacuum.12,13 12. Keep tube inverted to drain the remaining liquids. While inverted, wash the walls of the tubes with DDW (it is easy done using a syringe14) and dry the water with delicate paper. 13. Resuspend pellet in ~0.2 mL Isotonic buffer and transfer to Microtubes tubes (1.5 mL).15 14. Freeze samples in 80
C and store for lipidomics analyses. 3.2.2 Mitochondria Isolation from Mouse Liver16

1. Rinse liver in cold Mitochondrial Isolation Buffer (MIB) several times to remove blood. 2. Mince liver using razor or scissors in 3 mL MIB supplemented with 0.2% BSA, and transfer the sample to the glass tube. 3. Use Teflon Dounce to homogenize the sample, stroke up and down two times.17 4. Add 15 mL MIB and transfer18 homogenate to centrifuge tubes. 5. Centrifuge at 600  g, for 10 min, at 4
C. 6. Following centrifugation use a pipet to extract the middle fraction.19 Move this desired fraction to a clean centrifuge tube. 7. Centrifuge at 7000  g, for 15 min, at 4
C. 8. Remove supernatant and resuspend pellet in 10 mL MIB. 9. Centrifuge at 7000  g, for 15 min, at 4
C.20 10. Remove sup and resuspend pellet in 100 μL MIB. 11. Freeze samples in 80
C, and store for lipidomics analyses.

3.3

Lipid Extraction

Perform stages 1 and 2 on ice. 1. Homogenize sample21 in 500 μL cold PBS for 1 min using the pellet pestle rods with cordless motor at 2000–3000 RPM.

12

This contains crude cytosolic fraction, which, if desired, can be used for western blot. In that case, collect with a pipette and store at 80
C until use. 13 Vacuum far from the pellet, tilt tube to distance the liquid from the pellet. 14 Avoid scratching the tube for further use 15 For maximal yield first resuspend in half the volume, transfer it, and thoroughly collect the remaining pellet with the rest. 16 Only fresh liver is to be used, mitochondria do not fractionate well from frozen liver. 17 Manual homogenization; Avoid creating vacuum force and/or air bubbles! 18 For maximal yield first resuspend in half the volume, transfer it, and thoroughly collect the remaining pellet with the rest. 19 Avoid the fat layer on top and the lower layer at the bottom. Aim for 10–13 mL to be on the safe side. 20 Steps 8 and 9 are an optional wash, we perform it. 21 Equivalent to 10–20 mg wet weight of tissue or 1  106 cells.

164

Rona Aviram et al.

2. Perform protein assay with 5–10 μL aliquot of homogenate for each sample. 3. Prepare a mixture of internal standards: di14:1 PC, di16:1 PE, di15:0 PG, di14:0 PS, di14:0 PA, T14:0 CL, d4-16:0 NEFA, N12:0 SM, N17:0 Cer, 14:0 lysoPE, 13C4-16:0 carnitine, 17:0 lysoPC, and T17:1 TAG.22 4. Transfer a precise volume of homogenate of each sample into a disposable culture borosilicate glass tube (10 mL). Add the premixture of internal standards, based on the protein content of the transferred homogenate sample.23 5. Prepare extraction solvent using: chloroform/methanol (1/1, v/v) (solvent A), chloroform (solvent B), and 10 and 50 mM lithium chloride solutions. 6. For extraction add 4 mL solvent A to the 10 mL glass tube from step 3, and 2 mL 50 mM LiCl. Cap the tubes and vortex for 20 s. 7. Centrifuge at 4000  g, for 10 min, at 4
C. 8. Collect the bottom layer to a new borosilicate glass tube,24 and add 2 mL solvent B to the residual top layer. Cap the tubes and vortex them for 20 s. 9. Centrifuge at 4000  g, for 10 min, at 4
C. 10. Collect the bottom layer and combine it with that collected in step 7. 11. Evaporate the combined chloroform layers under a nitrogen stream. 12. Resuspend individual residue in step 10 with 4 mL of solvent A, and add 2 mL of 10 mM LiCl. Cap the tubes and vortex them for 20 s. 13. Repeat steps 8–12. 14. Resuspend individual lipid extract residue from step 12 with solvent A in a volume of 200 μL/mg protein in original sample. 15. Store the samples at 20
C until MS analysis.

22

The stock solution of each single internal standard is prepared in chloroform/methanol (1:1, v/v) with a concentration approximately 1 mg/mL. The amount of each single lipid species in the premixture is based on the abundance of the corresponding lipid class in the samples. The molecular species of internal standards are selected because they represent 20,000 distinct phosphopeptides in the 64 analyzed samples originated from 4 biological replicates collected every 3 h across two consecutive days [13]. Quantification was achieved for almost half of the phosphopeptides from which more than 25% showed circadian oscillations of abundance. Similarly and more recently, using this workflow we have characterized, in an unprecedented depth, the daily dynamics of the proteome and phosphoproteome in whole mouse forebrain as well as isolated

Quantitative Proteomics of the Circadian Clock

187

synapses from rested and sleep deprived mice [17, 18]. Consequently, the use of this method enabled us to reveal a dominant and widespread role for phosphorylation in circadian control of physiology, which was not appreciated at this scale previously.

2

Materials The EasyPhos method can be performed with cell pellets or homogenized tissue as starting material.

2.1 Cell and Tissue Lysis

1. Lysis buffer: 100 mM Tris-HCl, pH 8.5, 4% sodium dodecyl sulfate (SDS), 10 mM Tris(2-carboxyethyl)phosphine (TCEP), and 40 mM 2-chloroacetamide (CAA) (see Note 1). 2. Ultra-Turrax® (IKA) blender. 3. Bioruptor® (Diagenode).

2.2

TFE Digestion

1. TFE digestion buffer: 100 mM ammonium bicarbonate (ABC), 10% 2,2,2-Trifluoroethanol (TFE). 2. Trypsin and LysC.

2.3 Phosphopeptide Enrichment

1. Titansphere™ Phos-TiO (TiO2) beads 10 μM (GL Sciences). 2. Deep-well plate (DWP) 96/2 ml Protein LoBind (Eppendorf 0030504305) and silicon microplate sealing mats (Eppendorf 951030147) or 2 ml tubes (see Note 2). 3. Loading buffer: 80% acetonitrile (ACN), 6% trifluoroacetic acid (TFA). 4. Wash buffer: 60% ACN, 1% TFA. 5. Transfer buffer: 80% ACN, 0.5% acetic acid. 6. C8 (3 M Empore) StageTips (2 layers) prepared as described [19]. 7. TiO2 elution buffer: 40% ACN, 3.75% NH4OH (stock as 25%, HPLC grade). 8. Styrene divinylbenzene—reversed-phase sulfonated (SDB-RPS) (3 M Empore) StageTips (2 layers) prepared as described [19]. 9. SDB-RPS elution buffer: 80% ACN, 1.25% NH4OH (25%, HPLC grade). 10. LC-MS loading buffer: 2% ACN and 0.1% TFA.

188

3

Franziska Bru¨ning et al.

Methods

3.1 Cell and Tissue Lysis

1. Solubilize cell pellets or homogenized tissue (see Notes 3 and 4) with lysis buffer (see Note 5). 2. Immediately boil samples for 5 min at 95  C to rapidly inactivate proteases and phosphatases, and cool samples on ice for 5 min. 3. Sonicate homogenates in a Bioruptor® using the high-power setting at 4  C for 15 min (15 cycles of 30 s) or until a homogeneous suspension is formed. 4. Estimate the protein concentration using reducing-reagent compatible BCA or tryptophan assay [20] (see Note 6). 5. Precipitate 1–2 mg of protein lysate with 80% acetone as follows: (a) Add 4 volumes of

20  C 100% acetone.

(b) Precipitate protein overnight at

20  C.

(c) Collect precipitated protein by centrifugation at 1500  g for 10 min at 4  C. (d) Wash pellets 1–2 times with ice-cold 80% acetone, centrifuge after each wash at 1500  g for 10 min at 4  C (see Note 7). (e) After washing, leave the protein pellets inverted at RT to dry (see Note 8). 3.2

TFE Digestion

1. Add 500 μl TFE digestion buffer to protein pellets and resuspend by sonication in a Bioruptor® (maximum power at 4  C for 5 min) or with a tip-probe sonicator (see Notes 9 and 10). 2. Add 1:100 (enzyme–protein) LysC followed by 1:100 (enzyme–protein) Trypsin, and incubate at 37  C overnight with rapid agitation (2000 rpm). If desired, a small volume of digested peptides can be taken at this point to measure the proteome of the same sample prior to phosphopeptide enrichment (see Note 11). 3. Add to the peptide solution the following in the following order (final volume 1.6 ml): (a) 150 μl of 3.2 M KCl. (b) 55 μl of 150 mM KH2PO4. (c) 800 μl of 100% ACN. (d) 95 μl of 100% TFA.

Quantitative Proteomics of the Circadian Clock

3.3 Phosphopeptide Enrichment

1. Incubate peptides at RT for 5 min with shaking Clear the peptides by centrifugation at RT (16,000  g for 2 ml tubes, or 3000  g for transfer clear supernatant into a clean 2 ml tube Note 12).

189

(1600 rpm). for 30 min DWPs), and or DWP (see

2. Meanwhile, prepare TiO2 beads: (a) Weighed TiO2 beads at a ratio of 10:1 beads/protein. (b) Resuspend beads in loading buffer using 100 μl per sample. (c) Sonicate TiO2 bead suspension in a sonicating water bath for 2 min to ensure a homogeneous suspension is achieved. 3. Add TiO2 beads (see Note 13). Incubate peptides and beads at 40  C for 5 min with rapid agitation (2000 rpm). 4. Centrifuge at 2000  g for 1 min at RT to pellet TiO2 beads. 5. Aspirate supernatant, being careful to avoiding aspirating the beads (see Note 14). 6. Resuspend beads in 1 ml wash buffer and transfer to a clean 2 ml tube or DWP (see Note 15). 7. Repeat wash steps a further four more times, with 1 ml of wash buffer. 8. After the final wash, resuspend pelleted beads in 100 μl transfer buffer, and transfer them on top of C8 StageTips. Centrifuge at 500  g for 3–5 min at RT or until no liquid is visible in the StageTip (discard the flow-through). 9. Elute phosphopeptides from the TiO2 beads on the StageTips by adding 60 μl TiO2 elution buffer and centrifuge at 500  g for 3 min at 4  C. Collect the eluates containing the phosphopeptides into PCR tubes. 10. Concentrate immediately samples in a centrifugal evaporator for 15 min at 45  C or until about 20 μl of volume remains (it is important not to dry the peptides at this stage). Add 10 μl of 10% TFA to each sample to acidify the peptides before desalting in SDB-RPS StageTips. 11. Equilibrate SDB-RPS StageTips with the following solutions (in this order): (a) 100 μl of 100% ACN. (b) 100 μl of 30% MeOH in 0.2% TFA. (c) 100 μl of 0.2% TFA. 12. Load samples onto the equilibrated SDB-RPS StageTips (see Note 16).

190

Franziska Bru¨ning et al.

13. Wash StageTips twice with 100 μl 0.2% TFA and centrifuge at 500  g for 3 min at RT. 14. Elute with 60 μl SDB-RPS elution buffer and immediately concentrate samples in a centrifugal evaporator at 45  C for about 35 min, or until a final volume of 10% of the volume in a centrifugal evaporator. After this SDB-RPS loading buffer is added to the peptide solution to a final volume of 50 μl. Peptides can be stored at 20  C until being desalted in a SDB-RPPS StageTip (see step 12 of Subheading 3.3). Before desalting, make sure to acidify the peptides with TFA to a final concentration of 1%. 12. It is important that peptides are completely cleared of particulates prior to phosphopeptide enrichment. If this cannot be achieved by centrifugation, peptide solution can also be filtered using a filter plate (e.g., Millipore part number MDRPNP410). 13. TiO2 beads settle very quickly, therefore it is important to ensure that the beads are evenly dispersed before pipetting. This can be achieved by mixing the bead solution with a vortexer between each pipetting step. 14. The supernatant contains unbound nonphosphorylated peptides.

192

Franziska Bru¨ning et al.

15. In order to facilitate transferring all beads to a new plate, the washing can be performed in two steps, using 500 μl wash buffer each time. 16. This step can be done also with the proteome samples collected prior to the enrichment of the phosphopeptides. 17. Peptides can be resuspended in LC-MS loading buffer with the aid of a sonicating water bath. 18. Resuspended peptides can be stored at 20  C for months prior to the measurement. Thaw, sonicate briefly in a water bath, and briefly spin down peptides prior to injection. 19. Label-free intensities can be first normalized, for example by median centering (subtracting the median of all log2 intensities in each sample).

Acknowledgments We thank Matthias Mann and group members for their support. References 1. Panda S et al (2002) Coordinated transcription of key pathways in the mouse by the circadian clock. Cell 109(3):307–320 2. Storch KF et al (2002) Extensive and divergent circadian gene expression in liver and heart. Nature 417(6884):78–83 3. Zhang R et al (2014) A circadian gene expression atlas in mammals: implications for biology and medicine. Proc Natl Acad Sci U S A 111 (45):16219–16224 4. Masri S et al (2013) Circadian acetylome reveals regulation of mitochondrial metabolic pathways. Proc Natl Acad Sci U S A 110 (9):3339–3344 5. Reddy AB et al (2006) Circadian orchestration of the hepatic proteome. Curr Biol 16 (11):1107–1115 6. Humphrey SJ, Azimifar SB, Mann M (2015) High-throughput phosphoproteomics reveals in vivo insulin signaling dynamics. Nat Biotechnol 33(9):990–995 7. Sharma K et al (2014) Ultradeep human phosphoproteome reveals a distinct regulatory nature of Tyr and Ser/Thr-based signaling. Cell Rep 8(5):1583–1594 8. Junger MA, Aebersold R (2014) Mass spectrometry-driven phosphoproteomics: patterning the systems biology mosaic. Wiley Interdiscip Rev Dev Biol 3(1):83–112

9. Larance M, Lamond AI (2015) Multidimensional proteomics for cell biology. Nat Rev Mol Cell Biol 16(5):269–280 10. Altelaar AF, Munoz J, Heck AJ (2013) Nextgeneration proteomics: towards an integrative view of proteome dynamics. Nat Rev Genet 14 (1):35–48 11. Mauvoisin D et al (2014) Circadian clockdependent and -independent rhythmic proteomes implement distinct diurnal functions in mouse liver. Proc Natl Acad Sci U S A 111 (1):167–172 12. Robles MS, Cox J, Mann M (2014) In-vivo quantitative proteomics reveals a key contribution of post-transcriptional mechanisms to the circadian regulation of liver metabolism. PLoS Genet 10(1):e1004047 13. Robles MS, Humphrey SJ, Mann M (2017) Phosphorylation is a central mechanism for circadian control of metabolism and physiology. Cell Metab 25(1):118–127 14. Wang J et al (2017) Nuclear proteomics uncovers diurnal regulatory landscapes in mouse liver. Cell Metab 25(1):102–117 15. Humphrey SJ, Azimifar SB, Mann M (2015) High-throughput phosphoproteomics reveals in vivo insulin signaling dynamics. Nat Biotechnol 33(9):990–U142 16. Humphrey SJ, Karayel O, James DE, Mann M (2018) High-throughput and high-sensitivity

Quantitative Proteomics of the Circadian Clock phosphoproteomics with the EasyPhos platform. Nat Protoc 13(9):1897–1916 17. Noya SB, Colameo D, Bru¨ning F, Spinnler A, Mircsof D, Opitz L, Mann M, Tyagarajan SK, Robles MS, Brown SA (2019) The forebrain synaptic transcriptome is organized by clocks but its proteome is driven by sleep. Science 366 (6462):eaav2642 18. Bru¨ning F, Noya SB, Bange T, Koutsouli S, Rudolph JD, Tyagarajan SK, Cox J, Mann M, Brown SA, Robles MS (2019) Sleep-wake cycles drive daily dynamics of synaptic phosphorylation. Science 366(6462):eaav3617

193

19. Rappsilber J, Ishihama Y, Mann M (2003) Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics. Anal Chem 75(3):663–670 20. Wisniewski JR (2013) Proteomic sample preparation from formalin fixed and paraffin embedded tissue. J Vis Exp (79) 21. Tyanova S et al (2016) The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat Methods 13 (9):731–740

Chapter 15 Circadian Phosphorylation of CLOCK and BMAL1 Hikari Yoshitane and Yoshitaka Fukada Abstract Daily rhythms of behaviors and physiologies are driven by transcriptional–translational negative feedback loops of clock genes and encoded clock proteins (Bass and Takahashi Science 330:1349–1354, 2010; Brown et al. Dev Cell 22:477–487, 2012). Posttranslational modifications of clock proteins, including protein phosphorylation, play an essential role for normal oscillation of the circadian clock through regulation of their activities, stabilities, interactions, and intracellular localization (Gallego and Virshup Nat Rev Mol Cell Biol 8:139–148, 2007; Hirano et al. Nat Struct Mol Biol 23:1053–1060, 2016). In this chapter, we describe detailed methods for quantitative analysis of phosphorylation levels of clock proteins, particularly focusing on circadian phosphorylation of CLOCK, BMAL1, and their complex (Yoshitane et al. Mol Cell Biol 29:3675–3686, 2009). Key words Circadian rhythms, Clock proteins, Protein phosphorylation, Immunoprecipitation, Immunoblotting

1

Introduction In mammalian circadian clockwork, a couple of positive regulators, CLOCK and BMAL1, rhythmically bind to E-box DNA sequences and activate transcription of a series of genes including Per and Cry genes. Translated PER and CRY proteins associate with each other and enter into nuclei to inhibit E-box-dependent transcriptional activation [1, 2], thus serving as negative regulators of the circadian transcription. In human PER2, Ser662Gly mutation at its phosphorylation site was found as a mutation responsible for Familial advanced sleep-phase syndrome (FASPS) [3], and genetic analysis identified the other mutation of kinase CKIδ gene in a family with FASPS [4]. CKI-mediated phosphorylation of PER protein controls its subcellular localization and protein stability (reviewed in [5, 6]). Meanwhile, we identified Ser265 and Ser557 as phosphorylation sites of CRY2 by mass spectrometry analysis [7]. An antibody specific to the Ser557-phosphorylated form of CRY2 revealed circadian change of Ser557 phosphorylation in the mouse liver [8] and SCN [9]. The priming phosphorylation of CRY2 at Ser557 by

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_15, © Springer Science+Business Media, LLC, part of Springer Nature 2021

195

196

Hikari Yoshitane and Yoshitaka Fukada

DYRK1A allows for subsequent phosphorylation at Ser553 catalyzed by GSK3β, and the two-step phosphorylation leads to its rhythmic degradation by proteasome [8, 10, 11]. Circadian phosphorylation is observed also in the positive factors, CLOCK and BMAL1, in the mouse liver [12]. CLOCK phosphorylation was stimulated by a negative regulator CIPC [12], which was first identified to bind to CLOCK but not to CLOCKΔ19 protein [13] encoded by mutant Clock [14–16]. In the Clock mutant liver, CLOCKΔ19 is poorly phosphorylated, and CLOCKΔ19 protein is accumulated twofold higher than that of wild-type CLOCK in the liver [12]. A mass spectrometry analysis identified phosphorylation of CLOCK at Ser38, Ser42, and Ser427 [12]. Ser38 and Ser42 are located in the basic region of bHLH DNA-binding domain of CLOCK [17], and the dual luciferase reporter assay showed strong inhibition of the CLOCK-BMAL1mediated transactivation by mutations at these two phosphorylation sites [12]. The circadian phosphorylation of CLOCL might have a role of supporting circadian binding of CLOCK-BMAL1 complex to E-box sequences [18, 19]. Circadian regulation of Ebox-dependent transcription was recently shown to expand downstream outputs such as epitranscriptome mediated by E-box-regulated A-to-I RNA editing enzyme ADAR2 [20]. In order to analyze protein phosphorylation at amino acid levels, mass spectrometry and anti-phospho antibody are respectively strong strategies for identifying new phosphorylation sites and quantitative analysis of the target phosphorylation sites. Phosphorylation at some of the sites causes band shifts in SDS-PAGE gel, and, in some cases, protein purification by immunoprecipitation clarifies the protein band shifts.

2

Materials

2.1 Preparation of Nuclear Proteins of Mouse Liver

1. Mouse: 8–12-week-old male mice (C57BL/6J or BALB/c, Tokyo Laboratory Animals Science Co., Ltd). 2. Phosphate Buffered Saline (PBS): 10 mM sodium phosphate (pH 7.4), 140 mM NaCl, and 1 mM MgCl2. 3. Buffer A: 10 mM HEPES-NaOH (pH 7.8), 10 mM KCl, 0.1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM dithiothreitol (DTT), 1 mM phenylmethylsulfonyl fluoride (PMSF), 4 μg/ml aprotinin, 4 μg/ml leupeptin, 50 mM NaF, and 1 mM Na3VO4. Add DTT solution (1 M stock, store at
20  C), aprotinin and leupeptin solution (4 mg/ml each stock, store at
20  C), NaF powder, and Na3VO4 powder to Buffer A (
) containing 10 mM HEPES-NaOH (pH 7.8), 10 mM KCl, 0.1 mM EDTA (filtered with a 0.22 μm-pore-size filter) in each experimental day, and add PMSF solution

Studying Circadian Phosphorylation

197

(200 mM stock in ethanol, store at
20  C) immediately before use, because it is unstable in aqueous solutions. 4. Buffer C: 20 mM HEPES-NaOH (pH 7.8), 400 mM NaCl, 1 mM EDTA, 5 mM MgCl2, 2% (v/v) glycerol, 1 mM DTT, 1 mM PMSF, 4 μg/ml aprotinin, 4 μg/ml leupeptin, 50 mM NaF, and 1 mM Na3VO4. Add DTT solution, aprotinin and leupeptin solution, NaF powder, and Na3VO4 powder to Buffer C (
) containing 20 mM HEPES-NaOH (pH 7.8), 400 mM NaCl, 1 mM EDTA, 5 mM MgCl2, 2% (v/v) glycerol (filtered with a 0.22 μm-pore-size filter) in each experimental day, and add PMSF solution immediately before use. 2.2 Immunoblot Analysis

1. 5 sodium dodecyl sulfate (SDS) sampling buffer: 50 mM Tris–HCl (pH 6.8), 10 mM EDTA, 30% (v/v) glycerol, 250 mM DTT, 10% (w/v) SDS, 0.1% (w/v) Coomassie Brilliant Blue R-250, store at
20  C (see Note 1). 2. SDS-polyacrylamide gel electrophoresis (PAGE) gel: SDS-PAGE gel is prepared by using Mini-PROTEAN System (Bio-Rad) according to the manufacturer’s protocol. Separating gel: 375 mM Tris–HCl (pH 8.8), 6–10% acrylamide (see Note 2), 0.1% (w/v) SDS (10% stock in water, store at room temperature), 0.1% (w/v) ammonium persulfate (APS), 0.1% (v/v) N,N,N0 ,N0 -tetramethyl-ethylenediamine (TEMED). Mix all of the solutions except APS and TEMED. Add APS solution (10% stock, store at 4  C for up to a month) and TEMED, and immediately fill up the space made by the two glass plates with 1.5 mm-thickness. Layer water-saturated 2-butanol carefully on the top of the acrylamide layer. Leave it at room temperature for at least 15 min to polymerize. Check the residual acrylamide in the plastic tube to confirm that the polymerization reaction completed. Remove the residual solution by inverting the glass plate, and wash the gel face twice with water. Stacking gel: 125 mM Tris–HCl (pH 6.8), 5% acrylamide, 0.1% SDS, 0.1% APS, 0.1% TEMED. Mix all of the solutions except APS and TEMED. Add APS solution and TEMED, and immediately fill it to the top of the gel. Immediately insert comb. Leave at room temperature for at least 15 min to complete polymerization. 3. SDS-PAGE Buffer (SPB): 25 mM Tris, 192 mM glycine, 0.1% SDS. 4. Western Blotting Buffer 2 (WBB2): 25 mM Tris, 192 mM glycine, 10% (v/v) methanol. 5. Tris-based saline (TBS): 50 mM Tris–HCl (pH 7.4), 200 mM NaCl, 1 mM MgCl2. 6. Tween-containing Tris-based saline (T-TBS): 50 mM Tris– HCl (pH 7.4), 200 mM NaCl, 1 mM MgCl2, 0.05% (v/v) Tween 20.

198

Hikari Yoshitane and Yoshitaka Fukada

2.3 Immunoprecipitation for Analysis of CLOCK-BMAL1 Complex

1. Buffer D1: 20 mM HEPES-NaOH (pH 7.8), 5.5 mM NaCl, 1 mM EDTA, 6.5% (v/v) glycerol, 1.5% (v/v) Triton X-100, 1 mM DTT, 1 mM PMSF, 4 μg/ml aprotinin, 4 μg/ml leupeptin, 50 mM NaF, and 1 mM Na3VO4 (see Note 3). Add DTT solution, aprotinin and leupeptin solution, NaF powder, and Na3VO4 powder to Buffer D1(
) containing 20 mM HEPES-NaOH (pH 7.8), 5.5 mM NaCl, 1 mM EDTA, 6.5% (v/v) glycerol, 1.5% (v/v) Triton X-100 (filtered with a 0.22 μm-pore-size filter) in each experimental day, and add PMSF solution immediately before use. 2. Immunoprecipitation Buffer 2 (IPB2): 20 mM HEPES-NaOH (pH 7.8), 137 mM NaCl, 1 mM EDTA, 5% (v/v) glycerol, 1% (v/v) TritonX-100, 1 mM DTT, 2 mM PMSF, 4 μg/ml aprotinin, 4 μg/ml leupeptin, 50 mM NaF, and 1 mM Na3VO4. Add DTT solution, aprotinin and leupeptin solution, NaF powder, and Na3VO4 powder to IPB2 containing 20 mM HEPES-NaOH (pH 7.8), 137 mM NaCl, 1 mM EDTA, 5% (v/v) glycerol, 1% (v/v) TritonX-100 (filtered with a 0.22 μmpore-size filter) in each experimental day, and add PMSF solution immediately before use. 3. Protein G-Sepharose: Protein G-Sepharose 4 Fast Flow (GE Healthcare). Prior to use, wash Protein G-Sepharose (store in 20% EtOH at 4  C) three times with ice-cold water, and equilibrate it twice with ice-cold IPB2 (see Note 4). After the final wash, remove the buffer to make 50% slurry (beads/ liquid ¼ 1:1).

3

Methods

3.1 Preparation of Nuclear Proteins of Mouse Liver

1. Keep the mice in 12-h light:12-h dark cycles at least 2 weeks for their entrainment (see Notes 5 and 6). 2. Euthanize the mice by cervical dislocation and remove the eye balls under dim red light in the second day of constant dark conditions. Isolate the mouse liver under a room light and wash twice with 2 ml of ice-cold PBS. 3. Homogenize the tissue in a glass-on-glass Dounce homogenizer with ice-cold Buffer A (9 ml for 1 g wet weight of tissues). 4. Centrifuge the tissue lysate at 700  g at 4  C for 5 min. Centrifuge the supernatant again, and the final supernatant is used as a “cytoplasmic extract.” 5. Resuspend the two precipitates with ice-cold Buffer A, and centrifuge again. Resuspend the precipitate with ice-cold Buffer C (2 ml for 1 g wet weight of tissues), and rotate it at 4  C for 30–60 min.

Studying Circadian Phosphorylation

199

6. Centrifuge the suspension twice (for 30 min and then 10 min) at 21,600  g at 4  C, and the final supernatant is used as a “nuclear extract” (see Note 7). 3.2 Immunoblot Analysis

1. Add 5 SDS sampling buffer to the nuclear and cytoplasmic extracts, and boil them at 98  C for 5 min. Flash the boiled sample and leave the sample at room temperature until loading onto the gel (see Note 8). 2. Load the boiled samples on a polyacrylamide gel. In parallel, load prestained protein maker in the both ends of lanes. Electrophorese at 60 mA in SPB until the dye front reached the bottom of the gel. 3. Rinse the gel with WBB for 5–10 min, and electrotransfer to a PVDF membrane at 300 mM for 1 h in WBB by using miniPROTEAN 3 System (see Note 9). 4. Rinse the membrane with TBS, and cut out excess parts according to the signals of prestained protein maker (see Note 10). Block the membrane with a blocking buffer (1% skim milk in TBS) at 37  C for 1 h. 5. Incubate the membrane with primary antibody diluted in blocking buffer at 37  C for 1 h (see Notes 11 and 12). Wash with blocking buffer three times (for 2, 5, and 10 min). 6. Incubate the membrane with horseradish peroxidase (HRP)linked anti-IgG secondary antibody diluted in blocking buffer at 37  C for 1 h. Wash it with T-TBS three times (for 2, 5, and 10 min) and once with TBS for 5 min. 7. Detect the signal with the ECL plus kit and LAS 4000 system (GE Healthcare) or X-ray film. A sample result from such an experiment can be seen in Fig. 1.

3.3 Immunoprecipitation for Analysis of CLOCK-BMAL1 Complex

1. Add 100 μl of Buffer D1 to 50 μl of the nuclear extracts. Add 40 μl of 50% slurry of Protein G-Sepharose to the lysate, and rotate the mixed slurry for 30 min at 4  C. 2. Centrifuge at 384  g for 1 min at 4  C. The supernatant is used as a “precleared input.” 3. Add 1 μg of anti-CLOCK antibody (CLNT1) to the precleared input, and rotate for 1 h at 4  C (see Note 13). 4. Add 40 μl of 50% slurry of Protein G-Sepharose and rotate for 1 h at 4  C. 5. Centrifuge at 384  g for 1 min at 4  C. Wash the precipitate with 1 ml of ice-cold IPB2. Add 5 SDS sampling buffer to the precipitate and boil for 5 min at 98  C. 6. CLOCK and BMAL1 are immunoblotted as described in Subheading 3.2 (see Note 14). A sample result from such an experiment can be seen in Fig. 2.

200

Hikari Yoshitane and Yoshitaka Fukada

ZT6 ZT18 Nuc Cyt Nuc Cyt anti-CLOCK

anti-ATF2 anti-Raf1

Fig. 1 Immunoblot analysis of nuclear and cytoplasmic proteins in the mouse liver. The nuclear (Nuc) and cytoplasmic proteins (Cyt) were prepared from the mouse liver at ZT6 and ZT18, and were subjected to immunoblot analysis with anti-CLOCK monoclonal antibody CLSP3 (top). ATF2 (middle) and Raf1 (bottom) serve as nuclear and cytoplasmic markers, respectively. Marginal changes of phosphorylation levels of CLOCK protein were detected as band shifts in the gel (modified from ref. 12)

Fig. 2 Circadian phosphorylation of CLOCK and BMAL1 in the mouse liver. (a, b) Nuclear proteins prepared at the indicated time points were immunoprecipitated (IP) with CLNT1 anti-CLOCK monoclonal antibody and were subjected to immunoblot analysis (WB) with CLSP3 anti-CLOCK monoclonal antibody and B1BH2 anti-BMAL1 monoclonal antibody. Apparent changes of phosphorylation levels of CLOCK and BMAL1 proteins were detected as band shifts in the gel (modified from ref. 12). ZT: Zeitgeber time, a biological time in the 12-h light:12h dark cycles in which ZT0 and ZT12 correspond to the light-on and the light-off time, respectively. CT: Circadian time, biological time under the constant dark condition. (c) Dbp mRNA rhythms determined by qRT-PCR analysis showed circadian transcription driven by circadian clock (modified from ref. 12)

Studying Circadian Phosphorylation

4

201

Notes 1. SDS in buffers precipitates at 4  C. Therefore, the buffer needs to be warmed prior to use. 2. Wear gloves because acrylamide (30% stock, store in the dark at 4  C, acrylamide/N,N-methylenebisacrylamide ¼ 37.5:1) is neurotoxic. 3. Buffer D1 is a dilution buffer for immunoprecipitation using samples in Buffer C (1 Buffer C + 2 Buffer D1 ¼ 3 IPB2). 4. Gently invert or tap the Protein G-Sepharose, because it is fragile. For wash, centrifugation speed should be less than 384  g. 5. The room temperature is controlled in 23  1  C, and the animals have free access to commercial chow (CLEA Japan, Inc.) and tap water. 6. Divide the mice into two groups, and keep the first group in LD cycles (e.g., 7:00 light on/19:00 light off) and the second group in the opposite DL cycles (e.g., 7:00 light off/19:00 light on). The reverse entrainment procedures will prevent you from working through the night for the circadian sampling. 7. After the first centrifugation in Buffer C, freeze the supernatant with liquid nitrogen and store the aliquots at
80  C. Just before use, thaw the sample quickly and centrifuge them at 21,600  g at 4  C for 10 min. 8. SDS in buffer precipitates at 4  C. Therefore, do not keep them at 4  C after boiling. When the boiled sample is stored at
20  C, boil again at 95  C for 2 min prior to electrophoresis. 9. To keep WBB at 4  C during the electrotransfer, gently mix the buffer by a small stir bar on ice. 10. To detect CLOCK and BAML1 in the same membrane, cut the membrane according to the signals of prestained protein maker. 11. Condition for the reaction with a primary antibody should be optimized for each antibody. Conventional conditions are; at 37  C for 1 h, at 25  C for 2 h, or 4  C for 12 h. 12. For the reaction with a primary antibody, put the membrane on the plastic dish and add the primary antibody solution to the membrane so that the surface tension is not broken. 13. In this step, the rotation is not essential (leave at 4  C). The antibody reaction time should be optimized for each antibody. Conventional conditions are; at 4  C for 1 h, 2 h, or 12 h. To minimize noise signals in control sample (e.g., IP with control

202

Hikari Yoshitane and Yoshitaka Fukada

IgG), centrifuge at 21,600  g at 4  C for 10 min after the reaction with an antibody, and add Protein G-Sepharose to the supernatant. 14. 1 μg/ml of anti-CLOCK (CLSP3) and 1 μg/ml of antiBMAL1 (B1BH2) in blocking buffer.

Acknowledgments This work was partially supported by Grants-in-Aid for Scientific Research from MEXT, Japan (to H.Y. and Y.F.). References 1. Bass J, Takahashi JS (2010) Circadian integration of metabolism and energetics. Science 330:1349–1354 2. Brown SA, Kowalska E, Dallmann R (2012) (Re)inventing the circadian feedback loop. Dev Cell 22:477–487 3. Toh K, Jones C, He Y, Eide E, Hinz W, Virshup D, Pta´cek L, Fu Y (2001) An hPer2 phosphorylation site mutation in familial advanced sleep phase syndrome. Science 291:1040–1043 4. Xu Y, Padiath Q, Shapiro R, Jones C, Wu S, Saigoh N, Saigoh K, Pta´cek L, Fu Y (2005) Functional consequences of a CKIdelta mutation causing familial advanced sleep phase syndrome. Nature 434:640–644 5. Gallego M, Virshup DM (2007) Posttranslational modifications regulate the ticking of the circadian clock. Nat Rev Mol Cell Biol 8:139–148 6. Hirano A, Fu YH, Pta´cˇek LJ (2016) The intricate dance of post-translational modifications in the rhythm of life. Nat Struct Mol Biol 23:1053–1060 7. Sanada K, Harada Y, Sakai M, Todo T, Fukada Y (2004) Serine phosphorylation of mCRY1 and mCRY2 by mitogen-activated protein kinase. Genes Cells 9:697–708 8. Harada Y, Sakai M, Kurabayashi N, Hirota T, Fukada Y (2005) Ser-557-phosphorylated mCRY2 is degraded upon synergistic phosphorylation by glycogen synthase kinase-3 beta. J Biol Chem 280:31714–31721 9. Kurabayashi N, Hirota T, Harada Y, Sakai M, Fukada Y (2006) Phosphorylation of mCRY2 at Ser557 in the hypothalamic suprachiasmatic nucleus of the mouse. Chronobiol Int 23:129–134

10. Kurabayashi N, Hirota T, Sakai M, Sanada K, Fukada Y (2010) DYRK1A and GSK-3{beta}: a dual kinase mechanism directing proteasomal degradation of CRY2 for circadian timekeeping. Mol Cell Biol 30(7):1757–1768 11. Hirano A, Kurabayashi N, Nakagawa T, Shioi G, Todo T, Hirota T, Fukada Y (2014) In vivo role of phosphorylation of cryptochrome 2 in the mouse circadian clock. Mol Cell Biol 34:4464–4473 12. Yoshitane H, Takao T, Satomi Y, Du N, Okano T, Fukada Y (2009) Roles of CLOCK phosphorylation in suppression of E-boxdependent transcription. Mol Cell Biol 29:3675–3686 13. Zhao W, Malinin N, Yang F, Staknis D, Gekakis N, Maier B, Reischl S, Kramer A, Weitz C (2007) CIPC is a mammalian circadian clock protein without invertebrate homologues. Nat Cell Biol 9:268–275 14. Vitaterna M, King D, Chang A, Kornhauser J, Lowrey P, McDonald J, Dove W, Pinto L, Turek F, Takahashi J (1994) Mutagenesis and mapping of a mouse gene, clock, essential for circadian behavior. Science 264:719–725 15. King D, Zhao Y, Sangoram A, Wilsbacher L, Tanaka M, Antoch M, Steeves T, Vitaterna M, Kornhauser J, Lowrey P et al (1997) Positional cloning of the mouse circadian clock gene. Cell 89:641–653 16. Antoch M, Song E, Chang A, Vitaterna M, Zhao Y, Wilsbacher L, Sangoram A, King D, Pinto L, Takahashi J (1997) Functional identification of the mouse circadian clock gene by transgenic BAC rescue. Cell 89:655–667 17. Huang N, Chelliah Y, Shan Y, Taylor CA, Yoo SH, Partch C, Green CB, Zhang H, Takahashi JS (2012) Crystal structure of the

Studying Circadian Phosphorylation heterodimeric CLOCK:BMAL1 transcriptional activator complex. Science 337:189–194 18. Koike N, Yoo SH, Huang HC, Kumar V, Lee C, Kim TK, Takahashi JS (2012) Transcriptional architecture and chromatin landscape of the core circadian clock in mammals. Science 338(6105):349–354 19. Yoshitane H, Ozaki H, Terajima H, Du NH, Suzuki Y, Fujimori T, Kosaka N, Shimba S, Sugano S, Takagi T, Iwasaki W, Fukada Y

203

(2014) CLOCK-controlled polyphonic regulation of circadian rhythms through canonical and noncanonical E-boxes. Mol Cell Biol 34:1776–1787 20. Terajima H, Yoshitane H, Ozaki H, Suzuki Y, Shimba S, Kuroda S, Iwasaki W, Fukada Y (2017) ADARB1 catalyzes circadian A-to-I editing and regulates RNA rhythm. Nat Genet 49:146–151

Part III Imaging and Manipulating Brain Clocks

Chapter 16 Fluorescence Live Imaging of Drosophila Circadian Pacemaker Neurons Virginie Sabado and Emi Nagoshi Abstract Live imaging of the molecular clockwork within the circadian pacemaker neurons offers the unique possibility to study complex interactions between the molecular clock and neuronal communication within individual neurons and throughout the entire circadian circuitry. Here we describe how to establish brain explants and dissociated neuron culture from Drosophila larvae, guidelines for time-lapse fluorescence microscopy, and the method of image analysis. This approach enables the long-term monitoring of fluorescence signals of circadian reporters at single-cell resolution and can be also applicable to analyze real-time expression of other fluorescent probes in Drosophila neurons. Key words Drosophila whole brain explant culture, Drosophila dissociated neuronal culture, Timelapse fluorescence microscopy

1

Introduction Much progress has been made in the field of imaging live neurons. Observation of live neurons helps understanding dynamics of a plethora of neuronal processes, which range in time from milliseconds to days. When studying the mechanisms underlying circadian behavioral control, long-term observations of molecular and neuronal processes lasting hours to days are required. Circadian behavior is robust, quantifiable, and universal to most animals. In adult Drosophila, approximately 150 clock neurons compose the circuitry controlling circadian behavior. Because of its numerical simplicity compared to the mammalian circadian pacemaker circuit in the suprachiasmatic nuclei [1] and the availability of powerful genetic tools, the Drosophila circadian circuit offers a unique opportunity to delineate the intricate interactions between molecular and neural mechanisms underlying the operation of circadian circuitry at the single cell level. Fly clock neurons are clustered into seven subgroups: two clusters of ventral lateral neurons (small-ventral lateral neurons, s-LNvs; large-ventral lateral

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_16, © Springer Science+Business Media, LLC, part of Springer Nature 2021

207

208

Virginie Sabado and Emi Nagoshi

neurons, l-LNvs), dorsal lateral neurons (LNds), lateral posterior neurons (LPNs), and three clusters of dorsal neurons (DN1s, DN2s, and DN3s). Only three clusters of clock neurons (LNvs, DN1s, and DN2s) constitute the larval circadian circuitry, which is sufficient to regulate circadian light and temperature response behavior [2–4]. Larval LNvs give rise to adult s-LNvs, the master pacemakers that coordinate behavioral output by synchronizing the circuitry [5]. Only a handful of studies have performed bioluminescence or fluorescence imaging of fly clock neurons over a long period of time [6–12]. Although these studies successfully monitored molecular clockwork or neuronal activity in the brain, long-term circadian live imaging of dissociated fly clock neurons had not been performed. In addition, a better understanding of the interplay between network interactions and the molecular clockwork requires a system that enables the manipulation of network communication while monitoring the dynamics of the molecular clock. Here, we describe new methods for long-term fluorescence time-lapse imaging of Drosophila larval brain explants and cultured dissociated clock neurons. These methods were successfully used for imaging the newly developed fluorescent circadian reporters: one to monitor CLK/CYC transcriptional activity rhythms, and the other to follow PER protein levels and localization. The CLK/CYC transcriptional activity reporter, termed 3x69-VNP, expresses short-lived yellow fluorescent protein fused to a nuclear localization signal (VENUS-NLS-PEST, VNP) [13] under the control of three tandem repeats of the 69-bp E-box-containing enhancer of per. The PER protein reporter per-TdT, which was modified from the previously established BG-luc reporter [14], expresses the N-terminal two-thirds of PER fused with tandem TOMATO (TdT) red fluorescent protein under the control of per regulatory sequences and the 30 UTR [15, 16]. The protocols in this chapter describe how to establish and maintain healthy brain and primary neuron cultures for several days, and the guidelines for circadian fluorescent reporter timelapse imaging and image analysis. The use of a confocal microscope equipped with a resonant scanner enables a high-speed image acquisition at single-cell resolution and minimizes phototoxicity. This experimental system allows not only the addition of compounds to the culture media during live image acquisition but also dual imaging of circadian reporters and neuronal activity reporters (e.g., genetically encoded calcium indicators) and thus would yield a further understanding of the neuromolecular interactions important for the functioning of the circadian circuitry. More broadly, this flexible system is adaptable to time-lapse imaging of any fluorescent probes in fly neurons, providing powerful tools for research in Drosophila neurobiology.

Fluorescence Live Imaging of Drosophila Circadian Pacemaker Neurons

2

209

Materials All reagents should be prepared under a cell culture hood.

2.1 Preparation of Brain and Primary Neuronal Cultures

1. Stereomicroscope with light source (Nikon SMZ745) for screening Drosophila larvae and dissecting. 2. Small spatula and sterile petri dishes (35 mm and 10 mm of diameter) to collect larvae. 3. Fine forceps, size 5 (Dumont). 4. Dissection plate that can be chilled by ice water: a hollow metal plate connected to a small aquarium pump to circulate chilled water. 5. Glass depression well plate that can be filled to at least 500 μl. 6. SM active medium protocol modified from [17, 18]. Mix together all the reagents in Table 1. Filter-sterilize with a 500 ml bottle top filter unit (0.22 μm, Millipore). Keep the medium in dark for 3 days at 25  C. Add insulin (2 μg/ml) and Bis-Tris (5 mM). At this stage, the pH is ~5.7 and has to be adjusted to 6.8–6.9 by adding ~8 drops of 10 N NaOH. Filtersterilize with a 0.22 μm 500 ml bottle top filter unit. Aliquot to 5 ml, flash-freeze in LN2, and store at 80  C for up to ~ 6 months. Single use. 7. Dissecting Saline modified from O’Dowd’s protocol [19]. Mix together all the reagents in Table 2. Filter-sterilize with a 0.45 μm filter (see Note 1). Keep at 4  C. 8. Antibiotics: 100 penicillin/streptomycin (P/S), cell culture grade, store 500 μl aliquots in 20  C. Once thawed the aliquot can be kept 1 month at 4  C, provided it is only opened under laminar flow.

2.2 Ex Vivo third Instar Larval Brain Culture

1. Glass bottom dishes (MatTek 35 mm-diameter, 20 mm Microwell, No. 1.5 cover glass (see Note 2). 2. Teflon® FEP film (Dupont) cut to fit the inside of a glass bottom dish. Under laminar flow, rinse the discs in 70% EtOH and wash with autoclaved dH2O. To sterilize, turn on the UV lamp and keep under laminar flow for about 30 min, keep in a sterile petri dish at room temperature. Single use. 3. Autoclaved silicone grease. 4. Sterile 1.5 ml-Eppendorf tubes. 5. Sterile tips: P1000, P200, P20, P2. 6. Fibrinogen from bovine plasma (Millipore) can be prealiquoted to 10 mg and kept at 4  C. Single use.

210

Virginie Sabado and Emi Nagoshi

Table 1 Modified active Schneider’s Medium (SM active) Reagents

Concentration

KH2PO4

4.18 mM

CaCl2

1.05 mM

MgSO4.7H2O

0.7 mM

NaCl

116 mM

NaHCO3

0.7 mg/ml

Glucose

2 mg/ml

Trehalose

2 mg/ml

α-Ketoglutaric acid

0.35 mg/ml

Fumaric acid

6 μg/ml

Malic acid

0.6 mg/ml

Succinic acid

6 μg/ml

Yeast extract

2 mg/ml

NON heat-inactivated FBS

20%

Autoclaved dH2O All solutions have to be sterilized

Table 2 Dissecting Saline solution (1) Reagents

Concentration (mM)

HEPES-KOH, pH 7.4

9.9

NaCl

137

KCL

5.4

NaH2PO4

0.17

KH2PO4

0.22

Glucose

3.3

Sucrose

43.8

Autoclaved dH2O All solutions have to be sterilized

7. Thrombin from bovine plasma (Sigma, 1500 NIH unit/mg dissolve in SM active medium, prepare 10 μl aliquots, flashfreeze in LN2 and store at 80  C), keep at room temperature (RT) while dissecting larvae. Single use.

Fluorescence Live Imaging of Drosophila Circadian Pacemaker Neurons

2.3 Dissociated Neuron Culture from third Instar Larval Brain

211

1. Glass bottom dishes (MatTek 35 mm-diameter, 10 mm Microwell, No. 1.5 cover glass (see Note 2). 2. Sterile petri dish (90 mm of diameter). 3. Sterile 1.5 ml Eppendorf tubes. 4. P1000 (12 cm in length) and P200 (6 cm in length) ART® selfsealing barrier pipette tips (see Note 3). 5. FireboyTM Bunsen burner under a cell culture hood. 6. Concanavalin A (ConA), Canavalia ensiformis (Calbiochem). Dissolve ConA to 5 mg/ml in 150 mM NaCl (or directly in Dissecting Saline solution), add drops of 0.1 N NaOH until it dissolves (see Note 4). Then dilute the solution to 1 mg/ml in autoclaved dH2O (or Dissecting Saline solution). Do not vortex. Pipet gently if necessary. Filter-sterilize through 0.45 μm (not smaller) syringe filter. Aliquot and keep at 20  C. 7. ConA-coated glass bottom dishes (10 mm Microwell) (see Note 5). Thaw 1 mg/ml ConA solution under a culture hood, spot 70–100 μl of ConA on the cover glass of the glass bottom dish and make sure the glass is completely covered. Incubate at 37  C for 1 h with the lid closed. Then, place the dishes under a culture hood with the lid open so that ConA will be air-dried completely (it is fine to pipet most of ConA and leave a thin layer to dry out). Once the dishes are dried, rinse once with autoclaved dH2O and air-dry completely again under a culture hood and UV lamp to sterilize further (see Note 6). Keep the ConA-coated dishes at 4  C and use within 1 month. Single use. 8. L-cysteine-activated papain (Worthington). Dissolve to 50 unit/ml in Dissecting Saline solution; incubate at 37  C for 10 min to activate. Make 50 μl aliquots, flash-freeze in LN2, and store at 80  C. Single use. 9. Autoclaved dH2O.

3

Methods

3.1 Collection of Larvae

1. To collect entrained third instar (L3) nonwandering larvae on the day of dissection, incubate the vials for about 4–4.5 days after egg laying in 12 h:12 h Light/Dark (LD) cycles at 25  C (see Note 7). Just prior to dissection, clean the tools for dissection with 70% EtOH and then rinse in deionized water. 2. Fill half a bucket with ice (large enough to accommodate the aquarium pump), add water until the ice floats. Place the metal dissection plate connected to the aquarium pump under a stereomicroscope and switch on the pump to start circulating chilled water through the plate. 3. Thaw 5 ml of SM active medium (see Note 8).

212

Virginie Sabado and Emi Nagoshi

4. Prepare the following under a cell culture hood: pour 400 μl Dissecting Saline solution into three glass depression well plates, two to rinse the specimen, one to dissect in larval brain. Keep them on ice or on the chilled dissection plate. 5. Scrape fly food containing larvae with the spatula and place it on a 35 mm petri dish to collect nonwandering L3 larvae. Rinse the larvae twice in chilled Dissecting Saline and keep them in a depression well containing Dissecting Saline on ice until dissection. 3.2 Ex Vivo Third Instar Larval Brain Culture

The following protocol was modified from [20, 21]. The dissection and embedding of the brains should not take more than 40 min. 1. Under a cell culture hood, prepare 1 ml fresh solution of fibrinogen (10 mg/ml in SM active medium) and incubate at 37  C for at least 30 min to a maximum of 2 h (see Note 9). Single use. (a) Pour 400 μl SM active medium in a glass depression well to collect the dissected brains. Keep on ice. (b) Prepare one aliquot of 1 ml SM active medium with 1 P/S antibiotics, keep at 25  C. (c) Thaw an aliquot of thrombin at RT. 2. Apply autoclaved silicone grease on the plastic part around the cover glass of a glass bottom dish. 3. Bring the reagents and dishes back to the lab where you can dissect. 4. Dissect out the larval brain in chilled Dissecting Saline solution on the cold plate using the inside-out method [22] (see Note 10). Transfer the brains still attached to the mouth hooks (though the brain detached from the mouth hooks can be maintained in healthy culture) to the well containing chilled SM active medium on ice (collect 4–5 brains). 5. Pipet one brain in prewarmed 3.5 μl SM active medium and add it on a sterile petri dish, mix in another 3.5 μl of SM active medium and add 3.5 μl of warmed fibrinogen solution by gently pipetting up and down. Apply about 2 μl of the brain containing solution on the coverslip of the glass bottom dish and spread around the drop in a small circle. Then add 0.8 μl of thrombin and very quickly spread around with the tip, as it gets very sticky. 6. After about 40 s (the polymerization of fibrinogen should be visible), pipet one brain in a very small volume of fibrinogen solution and quickly add it on top of the fibrinogen matrix. You may use two pairs of forceps to stretch the forming matrix with the tips of the forceps and fold it around the brain to flatten and position the brain as close as possible to the coverslip (see Note 11). At this point, the fiber is sticky, so patience and

Fluorescence Live Imaging of Drosophila Circadian Pacemaker Neurons

213

micromovements are key. Add 20 μl of warmed SM active medium with 1 P/S antibiotics on top of the embedded brains to stop the action of thrombin. Repeat the procedure for the remaining brains. 7. At the end, add ~600 μl warmed SM active medium with 1 P/S antibiotics. Put the Teflon membrane on top and stick it to the silicone grease (see Note 12). Place the glass bottom petri dish in a 10 mm petri dish with wet paper and place at 25  C in LD cycles until the start of live imaging (see Note 13). 8. On the day of imaging, make sure that the integrity of the brain surface and neuronal cell membrane is intact. Expressing membrane-bound florescent marker in the neuron of interest is useful for this purpose. 3.3 Dissociated Neuron Culture from Third Instar Larval Brain

1. Take out an aliquot of 50 μl papain from 80  C and reactivate at 37  C for 10 min, and keep at RT while dissecting. 2. Under a cell culture hood, prepare two 1.5 ml-Eppendorf tubes, one containing 250 μl SM active medium and another one with 1 P/S antibiotics in 500 μl SM active medium, keep on ice. 3. Bring the reagents back to the lab to dissect. 4. Dissect out the larvae in chilled Dissecting Saline solution on the cold plate, as quickly as possible by tearing apart the larva to expose the brain. Dissect out the brain and remove unnecessary tissues such as the ventral ganglion or ring gland to eliminate non-clock PER+ cells. As soon as one brain is dissected, immediately transfer it to the 250 μl SM active medium set on ice. Collect about 20 brains per dish (see Note 14). 5. All the following procedures should be performed under a culture hood. 6. Spin the tube in a centrifuge in the cell culture room (swing rotor) at 2000 rpm (~425  g) for 1 min, RT. 7. Remove supernatant carefully (see Note 15) and wash the brains by adding 500 μl chilled sterile Dissecting Saline. Spin again at 2000 rpm (~425  g) for 1 min. 8. Remove supernatant carefully. Add 500 μl chilled sterile Dissecting Saline. Leave the tube at RT for 2 min to equilibrate to RT. Spin again at 2000 rpm (~425  g) for 1 min. 9. Remove supernatant carefully. Add 40 μl/ml activated papain (2 μl/brain) and leave at RT for 10 min. Flicker gently the tube from time to time (see Note 16). 10. In the meantime (or in advance), prepare flame-rounded P1000 (medium and small opening) and P200 tips. Pipet in and out autoclaved dH2O to check the aperture (see Note 17).

214

Virginie Sabado and Emi Nagoshi

11. Add SM active medium, at least five times the volume of papain solution (or add directly 500 μl), to stop the digestion. 12. Spin the tube at 2000 rpm (~425  g) for 1 min. 13. Remove supernatant carefully. Add 500 μl SM active medium, spin 2000 rpm (~425  g) , 1 min, RT. Repeat three times. 14. Remove supernatant carefully. Add SM active medium with 1 P/S antibiotics, 10 μl/brain (enough to plate two ConAcoated glass bottom dishes). 15. Triturate carefully with the flame-rounded prewetted tips in SM active medium as follows (see Note 18): P1000 (medium opening), 10 times; P1000 (small opening), 10 times; P200, 10 times. 16. Calculate the number of cells with a hemocytometer or Neubauer chamber. Most of the neurons should still display projections and pieces of brain should be spotted. 17. Plate cell homogenate in ConA-coated glass bottom dishes at 4~5  105 cells/cm2 corresponding to about 70–100 μl/dish (see Note 19). 18. Place the culture in a humid chamber (by simply placing the culture plate in a 10 mm petri dish with wetted paper). Incubate the cells in a 25  C incubator with about 80% humidity for 4–5 h. 19. Close to the end of the first incubation, thaw one aliquot of SM active medium quickly at 37  C, add 1 P/S antibiotics to the medium and incubate it at 25  C for 10 min. 20. Add 2 ml of the 25  C SM active medium with 1 P/S antibiotics to the culture and incubate further for 2 days in LD or constant darkness (DD) depending on the experimental purposes before imaging. 21. On the day of imaging, make sure that most neurons are alive and have developed long projections. 3.4 Live Imaging Setup

To perform fluorescence imaging of live samples, it is necessary to use a fast-speed microscope with confocal quality. Many microscopy systems can be adapted to this end, such as spinning disk confocal microscopes. Here is an example with a Leica TCS SP5 tandem scanner inverted confocal microscope equipped with a resonant scanner. 1. The microscope should be equipped with a chamber to control temperature and humidity. We use the temperature control chamber customized for Leica TCS SP5 confocal microscope (Life Imaging Services, Switzerland) and place a stage-top humidity controller (Life Imaging Services, Switzerland) inside the chamber. The setting should be 25  C and 80% humidity.

Fluorescence Live Imaging of Drosophila Circadian Pacemaker Neurons

215

2. To perform experiments in DD, cover the microscope environmental chamber with a black opaque cloth (see Note 20). 3. Acquire images with 40 or 63 water-immersion objective. Position a Water Immersion Micro Dispenser (dispenser and extended micropump MP6 series, Leica microsystem CMS GmbH) on the objective and program a protocol with the Autoimmersion Objective Controller software to dispense a constant amount of water, permitting long-term live imaging. 4. To reduce phototoxicity to a minimum, use the resonant scanner at 8000 Hz in bidirectional mode scanning 8 line average with 512  512 image resolution (see Note 21). 5. Use Z-section steps of 2 μm  ~40 and 1.7 μm  8 to image ex vivo brain culture and dissociated cultured neurons, respectively. 6. Use the mark and find window to select multiple positions (see Note 22) and set z-stack for each sample (see Note 23). 7. Choose xyzt recording mode to acquire time-lapse movies. 8. Set up time acquisition. For both types of culture, image the sample every 2–3 h for up to 72 h (see Note 24). 3.5 Image Analysis and Representation

3.5.1 Ex Vivo Brain Explant Time-Lapse Movie

FIJI [23] image analysis software can be used to quantify fluorescent circadian reporter expression and data representation. In certain cases, Bitplane Imaris software can be effective for image analysis. 1. Use FIJI to analyze the time-lapse movies of the ex vivo brain cultures. Briefly, construct a SUM z-stack containing the neuron of interest and manually draw the area around each neuron (see Note 25). 2. Measure the mean fluorescence intensity of the defined area. 3. Draw three nearby areas to measure the background fluorescence level. 4. Calculate the corrected total relative intensity of each cell as follows: (AreaneuronMean_Intensityneuron)(AreaneuronBackground) in arbitrary units (https://sciencetechblog.com/2011/05/24/ measuring-cell-fluorescence-using-imagej/).

3.5.2 Dissociated Cultured Clock Neuron Time-Lapse Movie

If available, use Imaris software (Bitplane). Create a 3D surface (region of interest in 3D) of each cell by thresholding fluorescence levels after background subtraction. Extract the intensity SUM of

216

Virginie Sabado and Emi Nagoshi

each 3D surface from the statistical data automatically generated by the program. 3.5.3 Dissociated Cultured Clock Neuron Time-Lapse Movie

4

For presentation purposes only, a  10 iterative deconvolution process can be applied to time-lapse movies using AutoQuant (MediaCybernetics) and Imaris. Similarly, time-lapse movies with xyz drift can be fixed by the 3D correction plugin from FIJI [24].

Notes 1. 10 Dissection Saline stock can be prepared and kept at 4  C. To prepare 1 solution, it is important to filter-sterilize the solution after diluting the 10 stock with autoclaved dH2O. Due to the high percentage of sucrose, Dissecting Saline solution is prone to contamination by bacteria and mold. Keep the aliquots at 4  C, and only open under a culture hood. 2. To gain maximum fluorescence intensity, choose the thickness of the cover glass according to the objective of the confocal microscope. 3. ART® self-sealing barrier pipette tips are used for their relatively long length. Compared to shorter tips, these are less harmful to the cells during trituration. 4. Without salt and NaOH, it is impossible to dissolve ConA to 5 mg/ml. 5. It is better to use dishes with a small glass diameter, as it permits a sufficient neuronal density. For a 10 mm-glass bottom dish, 3.1~4  105 plated cells makes 4~5  105 cells/cm2. 6. It is very important that the dishes are dried completely. 7. Flies can be cultured in the cornmeal–agar plate or in vials, depending on how sensitive the developmental stage is for your experiment. To collect well-synchronized population, change plates twice a day. 8. Thaw quickly SM active medium at 37  C rather than at room temperature, about 10 min. 9. It is preferable to have a 37  C incubator close by when embedding brains as it is important to keep the fibrinogen solution at a constant temperature. 10. The ventral ganglion and imaginal disks should remain intact to prevent tearing apart the brain, which otherwise leads to microcuts that compromise the brain culture. 11. Keep in mind the orientation of the chosen microscope (inverted or upright) when placing the brains. Larval LNvs are more easily detectable when the dorsal brain and ventral ganglion face the coverslip.

Fluorescence Live Imaging of Drosophila Circadian Pacemaker Neurons

217

12. It is possible to incubate the culture without Teflon membrane for a short brain culture (max 12 h) when a drug has to be added along the course of time-lapse imaging. In this case, add up to 2 ml of SM active medium with 1 P/S antibiotics. 13. Larval brain explants can be imaged immediately at the start of the culture or a few hours later. It is possible to culture them further. We were able to observe healthy neurons after 5 days in culture. However, the larval brain is still in development, and this should be taken into account. 14. Even when you need less material, it is better to start with about 20 brains, or else trituration will be difficult. 15. It is possible to see the brains when the lighting is optimal. It is always better to leave some solution instead of pipetting out or drying the brains. 16. When adding the papain do not touch the brain, as it would stick to the tip. Make sure no brain sticks to the wall of the tube. 17. The water should come out of the tip in a jet with a bit of resistance. 18. Do not over triturate; tiny pieces of brains should still be visible. 19. A too dense cell culture indicates overtrituration, which results in less healthy culture. However, the cell density should not be too low, as it is also unhealthy. 20. For circadian live imaging, short wavelengths should be avoided to prevent cryptochrome (CRY) excitation. When designing fluorescent reporters, choose the fluorophores whose excitation wavelengths are above 514 nm to minimize the effect on CRY. 21. Our microscope system possesses high-sensitivity HyD detectors that allow for the reduction of laser power to a minimum. To minimize it further, gain and pinhole can be readjusted. 22. Use the Mark and Find window to set 4–5 positions for ex vivo brain explant and as many as possible for dissociated neurons by selecting the cells that express the reporter of interest (it takes about 90 min to select 60 neurons). It is possible to perform a Tile Scan for long-term imaging but be aware that it will generate a huge volume of data. 23. When using an inverted microscope, set the beginning of the z-stack at the surface of the sample farthest from the objective and the end at the closest. The z-stack should encompass the sample and extra z-sections to avoid losing the region of interest, as the sample can move during live acquisition.

218

Virginie Sabado and Emi Nagoshi

24. For short-term ex vivo brain culture (up to 12 h), it is fine to image every hour. When using sequential mode to avoid cross talk between channels, choose in-between stack scanning mode, which is the fastest. 25. In the brain, clock neurons are often tightly clustered, which prevents the discrimination of single neurons by automated software.

Acknowledgments This work was supported by the JST PRESTO program, the Swiss National Science Foundation (31003A_149893), the European Research Council (ERC-StG-311194), and the Novartis Foundation for Medical-Biomedical Research (13A39). References 1. Hastings MH, Brancaccio M, Maywood ES (2014) Circadian pacemaking in cells and circuits of the suprachiasmatic nucleus. J Neuroendocrinol 26(1):2–10. https://doi.org/ 10.1111/jne.12125 2. Malpel S, Klarsfeld A, Rouyer F (2004) Circadian synchronization and rhythmicity in larval photoperception-defective mutants of Drosophila. J Biol Rhythm 19(1):10–21. https:// doi.org/10.1177/0748730403260621 3. Mazzoni EO, Desplan C, Blau J (2005) Circadian pacemaker neurons transmit and modulate visual information to control a rapid behavioral response. Neuron 45(2):293–300. https://doi.org/10.1016/j.neuron.2004.12. 038 4. Kaneko H, Head LM, Ling J, Tang X, Liu Y, Hardin PE, Emery P, Hamada FN (2012) Circadian rhythm of temperature preference and its neural control in Drosophila. Curr Biol 22 (19):1851–1857. https://doi.org/10.1016/j. cub.2012.08.006 5. Nitabach MN, Taghert PH (2008) Organization of the Drosophila circadian control circuit. Curr Biol 18(2):R84–R93. https://doi.org/ 10.1016/j.cub.2007.11.061 6. Roberts L, Leise TL, Noguchi T, Galschiodt AM, Houl JH, Welsh DK, Holmes TC (2015) Light evokes rapid circadian network oscillator desynchrony followed by gradual phase retuning of synchrony. Curr Biol 25(7):858–867. https://doi.org/10.1016/j.cub.2015.01.056 7. Sehadova H, Glaser FT, Gentile C, Simoni A, Giesecke A, Albert JT, Stanewsky R (2009) Temperature entrainment of Drosophila’s

circadian clock involves the gene nocte and signaling from peripheral sensory tissues to the brain. Neuron 64(2):251–266. https:// doi.org/10.1016/j.neuron.2009.08.026 8. Sellix MT, Currie J, Menaker M, Wijnen H (2010) Fluorescence/luminescence circadian imaging of complex tissues at single-cell resolution. J Biol Rhythm 25(3):228–232. https:// doi.org/10.1177/0748730410368016 9. Guo F, Yu J, Jung HJ, Abruzzi KC, Luo W, Griffith LC, Rosbash M (2016) Circadian neuron feedback controls the Drosophila sleepactivity profile. Nature 536(7616):292–297. https://doi.org/10.1038/nature19097 10. Haynes PR, Christmann BL, Griffith LC (2015) A single pair of neurons links sleep to memory consolidation in Drosophila melanogaster. elife:4. https://doi.org/10.7554/ eLife.03868 11. Liang X, Holy TE, Taghert PH (2016) Synchronous Drosophila circadian pacemakers display nonsynchronous Ca(2)(+) rhythms in vivo. Science 351(6276):976–981. https:// doi.org/10.1126/science.aad3997 12. Mezan S, Feuz JD, Deplancke B, Kadener S (2016) PDF signaling is an integral part of the Drosophila circadian molecular oscillator. Cell Rep 17(3):708–719. https://doi.org/10. 1016/j.celrep.2016.09.048 13. Nagoshi E, Saini C, Bauer C, Laroche T, Naef F, Schibler U (2004) Circadian gene expression in individual fibroblasts: cellautonomous and self-sustained oscillators pass time to daughter cells. Cell 119(5):693–705. https://doi.org/10.1016/j.cell.2004.11.015

Fluorescence Live Imaging of Drosophila Circadian Pacemaker Neurons 14. Stanewsky R, Jamison CF, Plautz JD, Kay SA, Hall JC (1997) Multiple circadian-regulated elements contribute to cycling period gene expression in Drosophila. EMBO J 16 (16):5006–5018. https://doi.org/10.1093/ emboj/16.16.5006 15. Sabado V, Vienne L, Nunes J, Rosbash M, Nagoshi E (2017) Fluorescence circadian imaging reveals a PDF-dependent transcriptional regulation of the Drosophila molecular clock. Sci Rep 7:41560 16. Sabado V, Vienne L, Nagoshi E (2017) Evaluating the autonomy of the Drosophila circadian clock in dissociated neuronal culture. Front Cell Neurosci 11:317. https://doi.org/10. 3389/fncel.2017.00317 17. Kuppers-Munther B, Letzkus JJ, Luer K, Technau G, Schmidt H, Prokop A (2004) A new culturing strategy optimises Drosophila primary cell cultures for structural and functional analyses. Dev Biol 269(2):459–478. https://doi.org/10.1016/j.ydbio.2004.01. 038 18. Schneider I (1964) Differentiation of larval Drosophila eye-antennal discs in vitro. J Exp Zool 156:91–103 19. Jiang SA, Campusano JM, Su H, O’Dowd DK (2005) Drosophila mushroom body Kenyon cells generate spontaneous calcium transients

219

mediated by PLTX-sensitive calcium channels. J Neurophysiol 94(1):491–500. https://doi. org/10.1152/jn.00096.2005 20. Sakiyama SE, Schense JC, Hubbell JA (1999) Incorporation of heparin-binding peptides into fibrin gels enhances neurite extension: an example of designer matrices in tissue engineering. FASEB J 13(15):2214–2224 21. Forer A, Pickett-Heaps JD (1998) Cytochalasin D and latrunculin affect chromosome behaviour during meiosis in crane-fly spermatocytes. Chromosom Res 6 (7):533–549 22. Hafer N, Schedl P (2006) Dissection of larval CNS in Drosophila melanogaster. J Vis Exp 1:85 23. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, Rueden C, Saalfeld S, Schmid B, Tinevez JY, White DJ, Hartenstein V, Eliceiri K, Tomancak P, Cardona A (2012) Fiji: an opensource platform for biological-image analysis. Nat Methods 9(7):676–682. https://doi.org/ 10.1038/nmeth.2019 24. Parslow A, Cardona A, Bryson-Richardson RJ (2014) Sample drift correction following 4D confocal time-lapse imaging. J Vis Exp (86). https://doi.org/10.3791/51086

Chapter 17 Monitoring Electrical Activity in Drosophila Circadian Output Neurons Annika F. Barber and Amita Sehgal Abstract Drosophila melanogaster is a powerful model organism used to study circadian rhythms, historically for elucidating the molecular basis of the clock and, more recently, for allowing for dissection of neural circuits underlying rhythmic behavior. The fly can be used to investigate the neuronal basis of complex behaviors at single-neuron resolution. Patch clamp electrophysiology permits single-neuron recording of resting membrane potential and action potential firing in response to genetic or environmental manipulations or application of drugs and neurotransmitters. Here we describe a protocol for dissecting Drosophila brains for electrophysiology, setting up and using a patch clamp system, and analyzing firing data around the circadian day and in stimulation-response experiments to test for functional neuronal connectivity in circadian circuits. Key words Drosophila, Electrophysiology, Patch clamp, Circadian clock, Pars intercerebralis

1

Introduction Drosophila melanogaster has proven to be a powerful model organism for probing the circadian clock at the molecular, neuronal, and behavioral levels. Availability of a sophisticated genetic toolbox allowed for work in Drosophila to lead the way in the discovery of the core molecular transcription–translation feedback loop of the molecular clock [1–5]. In addition, the circadian system in Drosophila is now well characterized at the neuronal levels. The clock network consists of ~150 interconnected neurons that express the core molecular clock proteins distributed in discrete clusters within the fly brain. The application of whole-cell patch clamp electrophysiology in flies lagged behind molecular and behavioral methods due to the technical difficulty of recording from such small neurons. However, in 2008 the first characterizations of clock cell spontaneous activity showed circadian patterns of excitability [6–8] in small and large ventral lateral neurons. More recently, circadian excitability has

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_17, © Springer Science+Business Media, LLC, part of Springer Nature 2021

221

222

Annika F. Barber and Amita Sehgal

been characterized in DN1 dorsal neurons [9]. Within the clock network, time-of-day information appears to be transmitted by rhythmic changes in RMP and firing rate. These physiological rhythms in turn drive rhythmic release of peptides and neurotransmitters, most notably PDF [10]. However, less is known about how time-of-day information from the Drosophila clock network is transmitted to output cells, which do not express core clock proteins, to drive rhythms of behavior and organismal physiology. Recent work from our lab has shown that at least one group of clock output neurons—insulin-producing cells—displays circadian physiological rhythms. The patch clamp technique allows for single-neuron measurements of resting membrane potential changes and currents at high temporal resolution. Patch clamp uses low-resistance glass micropipettes to form a high resistance seal with the membrane of a cell body. Suction is used to rupture the membrane under the micropipette, allowing electrical access to the intracellular space. Membrane potential (i.e., firing) and current flow can be measured basally at different times of the circadian day, or in response to stimulation by bath-applied drugs. In flies, recordings can be obtained from live flies or acutely dissected brains, to preserve the native neuronal connections (vs. dissociated neurons or slices). Neurons suitable for recording must have large somata and be superficially accessible after removal of the glial sheath. Here we describe a protocol for obtaining basal firing data from circadian output neurons in acutely dissected brains from light- and foodentrained flies. We also describe how this method can be combined with pharmacological stimulation of putatively connected neurons via P2X2 expression [11].

2 2.1

Materials Solutions

1. Internal and external solutions are based on the formulation by Gu and O’Dowd [12]. There is relatively little variation in these solutions between Drosophila electrophysiology labs, though some labs add additional sugars (trehalose and sucrose) to the external solution. There are also optional additives to internal solution described below. 2. Extracellular solution (see Table 1), in mM: 101 NaCl, 1 CaCl2, 4 MgCl2, 3 KCl, 5 glucose, 1.25 NaH2PO4, and 20.7 NaHCO3, with osmolarity 250 mOsm and a pH of 7.2. A 10 stock of external solution omitting NaHCO3 can be stored at 4  C for several weeks and diluted within a week of use.

Drosophila Clock Output Neuron Recordings

223

Table 1 Drosophila patch clamp external solution

Reagent

MW

[Stock] (M)

mL stock for 500 mL 10 g for 500 mL stock solution

[Final] (mM) 101

NaCl

58.44 5

146.10

101

CaCl2

110.98 0.5

27.75

10

1

MgCl2

95.21 1

47.61

20

4

KCl

74.55 3

111.83

5

3

180.16 1

90.08

25

5

–

–

1.25

glucose

NaH2PO4 119.98 0.5

First prepare filtered stock concentrations of each reagent. Prepare 10 external solution stock by adding all components except NaH2P04 and diluting to 500 mL in water. To create final working external solution, combine 50 mL 10 stock with 1.25 mL of NaH2PO4, dilute to 500 mM. Adjust pH to 7.2 and osmolarity to 250 mOsm

Table 2 Drosophila patch clamp internal solution Reagent

MW

[Stock] (M) g for 100 mL stock Amt stock for 100 mL solution [Final] (mM)

K-gluconate 234.20 0.5

11.71

20.4 mL

102

HEPES

238.30 0.1

2.38

8.5 mL

8.5

EGTA

380.35 0.1

3.80

940 μL

0.94

CaCl2

See Table 1

17 μL

0.085

MgCl2

See Table 1

170 μL

1.7

NaCl

See Table 1

340 μL

17

First prepare filtered stock concentrations of each reagent (some stocks from external solution can be used for internal as well). Prepare working solution by combining indicated amounts of stock solutions and diluting to 100 mL. Adjust pH to 7.2 and osmolarity to 235 mOsm

3. Intracellular solution (see Table 2) in mM: 102 K-gluconate, 0.085 CaCl2, 1.7 MgCl2, 17 NaCl, 0.94 EGTA, 8.5 HEPES, with an osmolarity of 235 mOsm and a pH of 7.2. 4. Solution preparation is most easily accomplished using sterile filtered high molarity stocks of each component, which remain stable for months at 4  C (see Tables 1 and 2). Working solutions should be made within a week of use and stored at 4  C when not in use. 5. Optional solution additives. (a) ATP or GTP can be added to internal solution to avoid rundown of channels that require ATP or G-proteins. Nucleotides should be added to internal solution on the day of use, and solutions must be kept on ice (see Note 1).

224

Annika F. Barber and Amita Sehgal

(b) Filling dyes such as Alexa Fluor-conjugated biocytin can be added to internal solution to visualize cells after recording. For visualization of fine neuronal detail, filling with biocytin followed by brain fixation and use of a streptavidin-conjugated dye may be necessary [13]. 2.2

Equipment

1. General lab equipment: (a) Incubator with controlled light cycle and temperature. (b) Drosophila activity monitoring system (DAM, Trikinetics) (optional). (c) Dissecting forceps.

microscope,

dissection

plate,

dissecting

(d) pH meter, (e) Osmometer. (f) Electrode puller (Narishige PC-10 or equivalent). (g) Soldering equipment. 2. Electrophysiology equipment: (a) Upright microscope (Nikon Eclipse FN1 or equivalent) with 40 immersion objective, recording chamber (Siskiyou PC-H perfusion chamber), homemade or commercial harp slice grid to hold brain, and fixed stage with X-Y translation, or moveable stage. Air table and Faraday cage to reduce vibrational and electrical noise. (b) Amplifier and compatible head stage and electrode holder for current and voltage clamp recordings, data acquisition system and acquisition software, BNC cables, computer. From Axon instruments, this would be Multiclamp 700B amplifier with CV-7b head stage, Digidata 1550 acquisition system, and pClamp software (see Note 2). For a basic understanding of key concepts in electrophysiology and practical advice on physical setup and software configuration for patch clap, the Axon Guide [14] is an invaluable resource. Using an analog–digital (A/D) converter requires choosing sampling and filtering frequencies appropriate for the signal being captured in order to avoid aliasing and remove unwanted noise. Refer to the Axon Guide [14] and your A/D converter user manual to configure your sampling and filtering rates. (c) Accessories: l

For electrodes and filling.

l

Microfil needles or equivalent.

l

Borosilicate glass, 1.5 OD  0.86 ID.

Drosophila Clock Output Neuron Recordings l

Syringes.

l

For ground electrode.

225

– Silver wire. – Insulated wire (for making ground). – Heat shrinkable tubing. l

For harp slice grid. – Nylon fiber (for harp slice grid). – Platinum wire (for harp slice grid).

(d) Micromanipulator to target cell (see Note 3). For single-cell patch clamp, only one manipulator is required. For multiple cells, or for targeted drug delivery, a second manipulator is required. (e) Perfusion system tubing and stopcocks to plumb perfusion system (see Note 4). (f) Illumination source and filter set for fluorescence if targeting cells by fluorophore expression. (g) For dark dissection.

3

l

IR camera and video monitor (optional).

l

600 nm LED plate

l

Safelight.

l

600 nm filter for microscope light source.

Methods

3.1 Creating and Entraining Flies

A genetic line that labels the neurons targeted for recording should be used. We use PDF-Gal4, Clk4.1, and DILP2 Gal-4s to identify LNv, DN1p, and insulin-producing neurons, respectively [6, 7, 9, 15]. The Gal-4/UAS system can be used to drive expression of a fluorescent marker in neurons of interest (see Note 5). We often use the UAS-dORK-NC1 marker which has GFP fused to a nonconducting open rectifier potassium channel and provides bright, membrane delimited GFP expression that assists in visualizing the membrane for sealing [6]. Not all neurons are amenable to wholecell recording, as the recording pipette must physically access the cell body of the target neuron. Neurons deep within the brain are challenging to access, as the overlying neurons must be blown or scraped away to access the target cell. This damages the native connections to the target cell, which may alter the electrophysiological properties under investigation, and it can prove challenging to access deeper neurons without clogging the recording electrode.

226

Annika F. Barber and Amita Sehgal

Male or female flies can be used for recording (see Note 6). Flies used for recording are entrained to a light cycle for at least 72 h prior to dissection, so flies are 5–10 days old on the day of recording (see Note 7). For constant-darkness experiments, flies are entrained to a light cycle for at least 72 h prior to transferring them to constant darkness (see Note 8). 3.2

Brain Dissection

1. Anesthetize flies for dissection on ice in glass or plastic vials. 2. Perform dissection in cold external solution. Using fine forceps, first remove the proboscis and associated air sacs, then remove head from thorax. Gently detach head cuticle and eyes from the brain. After the brain is removed clean off all connective tissue, air sacs and trachea. To avoid cleavage of surface ion channels, the glial sheath must be removed by hand rather than being removed enzymatically using a reagent such as papain (see Note 9). 3. For recordings in constant conditions (DD) and between ZT 12 and 0, dissections must be done in red light to avoid activating the clock circuitry [6] (see Note 10). Flies should be placed on ice in the dark (see Note 11). Dissection and recording must be done under red light. We perform dissections on a plate illuminated by 600 nm LEDs to avoid activation of retinal receptors and CRYPTOCHROME [16, 17]. A 610 nm longpass filter (Thorlabs) was placed over the microscope light to allow for cell identification and seal formation in red light. A brief illumination with a wavelength that excites your fluorophore of choice will be needed to confirm cell identification during seal formation. 4. Transfer a single brain to the recording chamber and hold in place with a harp slice grid. The recording chamber selected should be as small as possible to allow for rapid exchange of solutions via perfusion. We use the Siskiyou PC-H chamber which allows for smooth solution flow across a small volume bath. Harp slice grids and hold-downs can be purchased or fabricated in the lab to optimize access to brain regions of interest (Fig. 1). Brain orientation in the grid should be adjusted according to the cells of interest (see Note 12). To fabricate a harp slice grid, begin by bending 0.5 mm platinum wire into ~6  3 mm U-shaped frame. Undyed nylon threads can then be glued across this frame in any spacing and configuration. To achieve even spacing of threads, tape threads to thin strip of cardboard and wind around the cardboard several times. Tape the other end firmly so that threads are taut. Use forceps to adjust threads to approximate the desired spacing on the cardboard. Then slide the platinum frame under the taut threads and glue with superglue. When glue is dry, use a razor to trim threads loose from cardboard.

Drosophila Clock Output Neuron Recordings

227

Fig. 1 Brain holder configurations. (a) Brain holder configured for dorsal neuron and PI recordings—three nylon threads run horizontally across brain. (b) Brain holder configured for central neuron and SOG recordings—two nylon threads run across each optic lobe or edge of central brain

5. Position recording chamber on the microscope stage and allow brains to rest in flowing external solution (can be oxygenated with 95% oxygen, 5% carbon dioxide) for at least 10 min prior to recording. Recordings should be obtained within 1 h of dissection to ensure that they accurately represent physiological firing. External solution should be perfused continuously during seal formation/recording; a multibarrel perfusion system can allow for solution switching during recording. 3.3 Seal Formation and Whole-Cell Access

1. Ensure that silver ground wire and pipette wire are freshly chlorided. Place the ground electrode in the bath (see Note 13). After placing the electrode in the bath, use the pipette offset feature of Clampex to cancel the junction potential (see Note 14). 2. Pull borosilicate glass pipette to desired shape, fill pipette with internal solution using MicroFil needle (World Precision Instruments), and place on electrode holder mounted onto micromanipulator. Pipettes should be pulled on the day of use. The pipette resistance (RP) must be adjusted by cell size/type. For large peptidergic cells such as l-LNvs and PI neurons, RP should be 8–10 MΩ. For smaller neurons, such as DN1 neurons, RP should be 10–12 MΩ. Follow your pipette puller’s instruction manual to adjust tip size and taper. 3. Bring brain region of interest into focus using 40 immersion objective. Identify cells of interest using a fluorescence marker (see Notes 15 and 16). 4. Manipulate recording pipette to the area above the target neuron while monitoring RP (see Note 17). 5. RP can be monitored by giving a 10 mV voltage step which can be observed in the pClamp seal test window—which also provides estimates of other helpful parameters discussed later.

228

Annika F. Barber and Amita Sehgal

Maintain positive pressure in the recording electrode during manipulation to prevent clogs (see Note 18). Position the electrode just above the neuron so that it will make contact near the center of the soma when driven straight down. When positioned correctly with appropriate positive pressure, slight dimpling should be observed on the soma (see Note 19). 6. To form a seal, once the electrode is properly positioned just above the target cell and a slight dimple is seen, release the positive pressure. The cell should spring back to contact the tip of the pipette and begin forming a cell-attached seal. Monitoring RP will allow for assessment of seal formation. Some cells may continue to form a tight seal (>5–10 GΩ) with no additional manipulation. To assist in seal formation, the holding voltage command can be set from 30 to 60 mV and gentle constant negative pressure can be applied by syringe or mouth suction. If a high resistance seal does not form, use a new pipette to try again with a different cell. 7. For whole-cell experiments the membrane under the recording pipette must be ruptured. This can be done with a combination of suction (via mouth or syringe) and the “Buzz” feature of the Multiclamp amplifier, which drives a high-frequency oscillatory current through the recording pipette to facilitate penetration. Break-in can be monitored using a 10 mV voltage step and observing a change in the capacitative transients, though this change is often very small. Break-in can also be observed in current-clamp mode as a drop in resting membrane potential and increase in firing amplitude. If using a dye in the recording electrode, dye filling of the cell will also begin to be observable soon after break-in. 3.4 Assessing Recording Quality

1. It is important to assess the quality of the recording before and during data acquisition. Membrane capacitance (Cm), seal resistance (Rseal), series resistance (Rs), and input resistance (Rin) should be monitored prior to initiating recording and between recordings. 2. Membrane capacitance is a function of cell surface area and can be estimated from the exponential decay in response to a voltage step. Clampex software will estimate Cm in the seal test window. A drop in Cm can indicate a failing seal or a dying cell. 3. Seal resistance is a measure of contact between the cell membrane and the pipette tip. Seal resistance depends on multiple factors including electrode quality, cell health, and the quality of the preparation. Cell-attached seal resistance should exceed 10 GΩ. Whole-cell seal resistance should exceed 1 GΩ. Wholecell seal resistance may deteriorate over long recording periods, so this parameter must be checked in the seal test window between each recording to ensure the seal quality remains high.

Drosophila Clock Output Neuron Recordings

229

4. Series resistance is the resistance between the pipette electrode and the cell. Series resistance will be high in small cells such as Drosophila neurons. Initial series resistance after break-in should be noted in the seal test window, as the series resistance should remain constant throughout recording. An increase in series resistance can indicate that a patch is resealing. 5. Input resistance reflects the membrane resistance (Rm), which is a function of the number of open channels, and the access resistance. Access resistance should remain constant throughout recording, so any change in input resistance reflects changes in Rm. A reduction in Rin indicates that the cell health is failing and Rm is dropping, making the cell leakier. This parameter requires assessment by injecting a small hyperpolarizing current to silence the cell and then injecting 5 pA hyperpolarizing current steps from 5 pA to 45 pA and reading the resulting membrane potential. Rin is determined by measuring the slope of the I-V curve plotted from this protocol. For large peptidergic neurons Rin is typically 2–4 GΩ. 3.5 Acquiring and Analyzing Data

For whole-cell recordings, wait 5 min after break-in to whole-cell configuration before acquiring data, to allow for equilibration of ion concentrations between the cytosol and the intracellular recording solution. If applying a stimulus such as a perfused drug or neurotransmitter, acquire baseline data before applying the stimulus. Stimuli can be applied locally with a picospritzer or can be applied via bath perfusion when the target is sufficiently specific, such as UAS-GAL4 expression of P2X2 channels [11]. Whole-cell seals can remain stable for up to an hour in dissected brains and live flies [18]. Brains not used within 1 h of dissection should be discarded. The success rate for acquiring high quality whole-cell recordings can be frustrating; even experienced electrophysiologists may acquire data from 4 to 8 cells within a day, and only 50–75% of those cells may be suitable for further analysis. Baseline firing frequency and resting membrane potential can be determined from a 1–2 min current clamp recording. The firing frequency in whole-cell and cell attached recording modes should be comparable in high-quality recordings. If it is slightly different, a small depolarizing or hyperpolarizing current can be injected; however, if it is substantially different, the cell should be discarded. Cellular excitability is measured in current clamp mode by injecting 10 s depolarizing current steps in small (e.g., 2–5 pA) increments. Data from this protocol can be plotted as a firing frequency–current (f-I) plot. Voltage-clamp is used to examine specific ionic conductances and their response to time of day or applied neurotransmitters or pharmacological agents. Isolation of specific conductances requires use of specific electrophysiology protocols to selectively activate/ inactivate channel populations and/or the use of toxins or pharmacological agents to selectively block channel populations.

230

4

Annika F. Barber and Amita Sehgal

Notes 1. Addition of these components will affect pH and osmolarity; readjust these before use. 2. The other major manufacturer of electrophysiology systems is HEKA. This protocol is described based on Axon products, but HEKA equipment and software can also perform all functions described here. 3. The Sutter micromanipulator systems are excellent. MP225 is sufficient for one manipulator, the MPC200 adds functionality for multiple manipulators. 4. ValveLink 8.2 is a good gravity driven system; specialized perfusion systems are also available for perfusion of volatile or reactive compounds. 5. To avoid exciting the clock circuitry in dark conditions, illumination with light 12 h [8]; enable the delivery of carefully controlled sensory stimuli; can be used alongside an array of established or novel experimental tools (local or systemic application of pharmacological agents, electrical stimuli, optogenetics); and are easily adapted to study activity in any other brain region [9–13]. With little modification, this method can also easily be adapted for the implantation of electrodes for chronic recordings in freely moving rodents [10].

2

Materials Prepare all solutions used for surgery in sterile in physiological saline. Prepare and store all reagents at room temperature, unless stated otherwise. Waste disposal regulations apply to urethane and PFA.

Acute In Vivo Multielectrode Recordings from the Mouse Suprachiasmatic Nucleus

2.1

Surgery

251

1. Sterile physiological saline (0.9% NaCl in sterile water). 2. Urethane: 20% solution in physiological saline. Store at room temperature and use an opaque container or cover (e.g., aluminum foil) to minimize light exposure (see Note 1). 3. Atropine: 1% solution dissolved in physiological saline. 4. Mineral oil (see Note 2). 5. Hydrogen peroxide. 6. DiI (Cell Tracker CM-DiI; Invitrogen, UK) (see Note 3). 7. Homeothermic heat mat. 8. Stereotaxic frame. 9. Micromanipulators. 10. Electrode (see Note 4).

2.2

Histology

1. Sucrose: 30% solution dissolved in water. 2. Paraformaldehyde (PFA) solution: 4% solution dissolved in PBS (pH 7–8). Weigh 8 g PFA powder (store at 4
C) and add to 200 ml phosphate buffered saline (PBS). Heat solution to 55–60
C while mixing. Once heated, continue mixing at 55–60
C for 10 min. Add 1 M NaOH dropwise until solution clears. Remove from heat and filter. Make up to 400 ml total volume with PBS. The pH of the solution at room temperature should be between 7 and 8, add NaOH or HCl as appropriate. Cool and store at 4
C, for up to 3 months (see Note 5).

3 3.1

Methods Surgery

1. Anesthetize mouse with intraperitoneal injection of urethane (dose: 1.5–1.6 g/kg). 2. Place mouse onto homeothermic heat mat (37
C) to maintain stable body temperature throughout surgery. 3. Once surgical plane anesthesia has been achieved (see Note 6), place mouse into the stereotaxic frame by placing the mouse into the bite bar and inserting the ear bars into the ear canals. 4. Expose the surface of the skull with a scalpel cut so that bregma and lambda are visible (see Note 7). 5. Adjust the position of the bite bars and ear bars, if necessary, to ensure the skull is level within the frame (see Note 8). 6. Using the micromanipulator, measure from Bregma to the point of entry on the skull (see Note 9).

252

Joshua Mouland et al.

7. Mark the point of entry on the skull surface with a cross (see Note 10). 8. Using a micro drill, drill into the skull at the center of the cross ensuring that the hole is large enough for the electrode to pass through. Clear the hole using forceps to remove the dura on the surface of the brain (see Note 11). 9. Before proceeding further, prepare the animal and/or position any apparatus required to produce experimental stimuli (see Note 12). 10. Dip the probe in fluorescent dye (DiI) and connect to the headstage/data acquisition system. 11. If bregma is still clearly visible, remeasure with the electrode to the point of insertion. Failing this, place the electrode on the surface of the brain in the center of the hole, aligned with the markings made previously in step 6. 12. Lower the electrode to the level of the SCN (~5.5 mm) using a fluid-filled micromanipulator (see Note 13). 13. Correct electrode placement can be confirmed by the presence of light-dependent changes in neural activity (see Note 14). 14. Allow probe/neural activity to settle (20+ min) before starting recording to ensure stable recordings. This preparation allows for continuous recordings for over 12 h. However, waveform properties of neurons may drift slightly with long recordings. 15. At the end of the experiment, apply a small amount of sterile saline to facilitate probe removal. Attention is required when moving the electrode as it is very easy to break. Lateral or forceful movement while the probe is imbedded can cause the shanks to shear from the probe. Once removed, the electrode can be gently rinsed in distilled water before returning to storage. If shanks are stuck together or tissue is attached to the probe, additional probe cleaning and maintenance might be required (see Note 15). 3.2

Data Acquisition

Optimal parameters for data acquisition will be dependent on the nature of the experiment being performed and the data acquisition hardware/electrodes being used. The key general principles are as follows: 1. Set an appropriate gain and sample rate for data acquisition. With the low-impedance electrodes used for multielectrode recording, action potentials recorded from SCN neurons are relatively small (7 mm is advised. For more superficial areas, such as the IGL/OPN, a shorter length (5 mm) electrode can be used. (c) Shank thickness—Thinner shanks do less damage during electrode insertion than thicker ones and so are advantageous when recording from smaller structures. Conversely, thinner electrodes can be more susceptible to fracturing following slight impacts. (d) Distance between contact sites—The greater the distance between contact sites, the larger area one can record over. However, densely packed contacts allow for polytrode recordings which can facilitate isolating single units (see Note 17). (e) Size of contacts—Smaller sites (i.e., 177 μm2) are better for detecting neuronal firing from single units. Larger sites (i.e., 703 μm2) are better for local field potential recordings and/or for electrical stimulation studies. (f) Electrode materials—Commercially available silicon probes typically have iridium or iridium oxide contacts (although other electrode materials/coatings may be available). For studies involving stimulation, electrode material (alongside area) will determine how much current can be delivered. Bare Iridium has a relatively low

Acute In Vivo Multielectrode Recordings from the Mouse Suprachiasmatic Nucleus

257

charge-carrying capacity, so if electrical stimulation is required, we suggest using an alternative material and/or coating (e.g., iridium oxide). 5. PFA is a known carcinogen, corrosive to skin and eyes and is harmful if inhaled. Prepare in fume hood, wearing gloves and goggles to protect from exposure to dust particles. Storing powder at 4
C prevents release of vapors. 6. The correct depth of anesthesia can be assessed by examining the corneal (blink) reflex and toe pinch reflex. When these are abolished, surgical plane anesthesia has been achieved. Additional top-ups of urethane can be administered into the other intraperitoneal space or subcutaneously. We recommend using either 20% or 10% urethane to allow greater precision when applying small doses. 7. If bregma is difficult to visualize, a cotton bud can be used to apply hydrogen peroxide solution (30%) to the skull surface— bregma and lambda turn white, making them easier to see. Hydrogen peroxide also acts as an antiseptic, to sterilize the skull surface for surgery. 8. If the skull is level, (1) bregma and lambda should run parallel to the micromanipulator in the rostral–caudal direction), (2) bregma and lambda should be at the same depth, (3) to correct for lateral tilt move 2 mm lateral to bregma—the points at either side of bregma should be at the same depth. 9. The SCN is very close to the midline and therefore is almost directly under the superior sagittal sinus and the azygos pericallosal artery. Approaching the SCN directly from above carries a substantial risk of damaging the overlying vessels, resulting in hard-to-control bleeding. We thus recommend approaching the SCN at an angle with respect to the sagittal axis. Using basic trigonometry (lateral displacement ¼ SCN depth  tanθ), one can calculate the distance from the midline for the entry point if approaching at a given angle. The depth of the SCN can be calculated from the Mouse Atlas (Paxinos and Franklin) at about 5.5 mm. This technique is also very useful if implanting multiple probes. If using a multishank electrode, we recommend directing the long axis of the probe along the sagittal axis, as this helps to increase the number of shanks within the SCN. 10. Connect a marking device to the manipulator (a pencil will suffice; we use a syringe needle). Place the tip of the marker on bregma, raise and then measure the required distance lateral and caudal. Lower the manipulator to draw a dot on the skull surface at the point of entry. Draw a cross centered on this point using a scalpel extending the lines as far as you can.

258

Joshua Mouland et al.

11. It is easier to use a rounded head drill bit (0.5 mm ball diameter) if you are manually drilling as you hold the drill at an angle (similar to holding a pen). When drilling the hole it is advisable to make the hole slightly larger than you need—this saves time if it is necessary to reposition the electrode. Ensure that the arms of the cross remain visible to enable correct alignment of the electrode. 12. While setting up, the craniotomy can be covered with cotton wool in sterile saline to prevent the cortex drying. For experiments of a visual nature, place atropine onto the surface of the eye to dilate the pupil and mineral oil to prevent drying out of the cornea. For diffuse light stimulus, the light apparatus can be positioned in contact with the eyes ensuring that the light stimulus is centered on the pupil. For this stimulus, it is best to remove the whiskers because they can prevent a proper seal and act as light guides. If only one eye is receiving the light stimulus, cover the other eye. 13. Although the mouse atlas measures ~5.5 mm distance to the SCN on the dorsal–ventral plane, there will be some variation between individuals/mouse strains. As such, we recommend using extreme caution when approaching the SCN. Once the electrode is ~500 μm above target, the electrode should be lowered very slowly (~50 μm increments) while visualizing under the dissecting microscope. If possible, we also suggest assessing for the presence of light responsiveness throughout this procedure. Our experience is that robust light responses are not found outside the SCN region. As such with multielectrode probes, it is usually possible to infer the depth within the SCN based on the profile of light evoked firing (which will cover at maximum 350 μm dorsal–ventral). If during this procedure the electrode comes into contact with the base of the brain (e.g., if probe placement is sufficiently off-target that no light responses are detected), a slight flexion in the electrode is usually observed (for silicon probes). Note here that the durability of electrodes to this kind of treatment is variable (depending on thickness, material, and design). Extreme care is required to avoid fracturing the recording probe. 14. SCN neurons are light responsive. If no light responses are detected on the electrode sites then the electrode will require repositioning. Typically, a 2–5 s light pulse to both eyes is sufficient to determine light responsiveness. 15. Electrodes can be reused and if looked after they can last multiple recordings. When using multishank electrodes, there is a tendency for shanks to coalesce when retrieving the electrode from the brain. A short (~5–30 min) immersion of the

Acute In Vivo Multielectrode Recordings from the Mouse Suprachiasmatic Nucleus

259

shanks into a protease enzymatic detergent (e.g., Tergazyme [Alconox]) is usually sufficient to separate the shanks. The electrodes should then be carefully rinsed in water. There are commercial devices that can help maintain and prolong your electrodes by activating sites to adjust impedance. 16. We recommend a sample rate of 40 kHz. Higher sample rates may assist in spike sorting but require more memory to store and to process. 17. Our experience is that a simple ground connection from the data acquisition system to the stereotaxic frame (and therefore the animal) results in a very low noise recording. If electrical noise persists, surrounding the electrophysiology rig with a Faraday cage, and/or grounding any nearby electrical devices to a common location may be required. In the worst-case scenario mains noise can also be removed from recordings by applying an appropriate notch filter (i.e., 50/60 Hz), although identifying and removing the source of noise by appropriate grounding is strongly advised. Some probes also contain a distant, low-impedance reference electrode which can be used to subtract noise/artifacts that might be present at all recording sites (e.g., due to occasional muscular activity/breathing). If the electrode does not have a specific reference electrode, an alternative option (primarily useful for spike recordings) is to apply referencing during offline data processing by subtracting the common median across all electrodes from each data trace. 18. If the data files are too large to be manageable, try reducing the sampling rate to reduce the file size. Alternatively many data acquisition systems can filter and extract spikes online, allowing you to record just the spike waveforms. This dramatically reduces the size of the files but also reduces the flexibility in analysis. 19. It is advisable to preserve the attachment of the optic tract when removing the brain. If the optic tract is ripped, the SCN can also be pulled away from the base of the slice. Additionally, it is much easier to identify the position of the electrodes on the rostrocaudal axis by the size/shape of the optic tract. 20. We recommend coronal sections as it is easier to identify the anatomical boundaries of the SCN. The sections containing the SCN begin shortly after the anterior commissure comes together and ends around the appearance of the hippocampus. When starting out it can be useful to keep several sections either side. In general we recommend 100 μm sections

260

Joshua Mouland et al.

for confirming probe location; this preserves substantial portions of the dye-labeled track even if this is slightly incongruent with the plane of section. These thicker slices are also much easier to handle than thinner sections. 21. While there are automated spike sorting algorithms we primarily recommend manual sorting. One automated system that we have had some success with is KiloSort [16], which works well for dense electrode arrays where a single spike may be detected at more than four adjacent electrodes (and which are very time-consuming to sort manually). Even with these automated methods, manual validation of all isolated single units is still essential. 22. If the electrode contact sites are in close proximity (0.5%) occurring within 1 ms of one another. Two example units that are well isolated are shown in Fig. 2. Cross correlograms can be used to confirm that units are well separated. Unless connected synaptically (or by gap junction), two cells should fire independently from one another. Cross correlograms of well-isolated units are relatively flat (Fig. 2).

Acute In Vivo Multielectrode Recordings from the Mouse Suprachiasmatic Nucleus

261

Fig. 2 Isolation of single units from SCN tetrodes using principal component analysis. Screenshots from Offline Sorter (Plexon, TX, USA) showing two isolated single units from a single tetrode, colored yellow and green, respectively. (a) The waveforms from both units are separated according to principal components and their waveform properties. “Virtual” tetrode waveforms (mean  3 SD) for each unit are shown in (b). (c) Crosscorrelogram of the two units (0.2 ms bins). Note the dip at time 0 indicating that spikes rarely occur at the same time in both units as would be expected if spikes were being inappropriately allocated to both units. (d) Histograms of interspike interval (0.1 ms bins) for each of the units depicted in A and B. Note that in neither case did spikes fall within 1 ms (dashed red line) of one another within the same unit References 1. Rusak B, Robertson HA, Wisden W, Hunt SP (1990) Light pulses that shift rhythms induce gene expression in the suprachiasmatic nucleus. Science 248(4960):1237–1240 2. Gillette MU (1986) The suprachiasmatic nuclei: circadian phase-shifts induced at the time of hypothalamic slice preparation are preserved in vitro. Brain Res 379:176–181 3. Yamazaki S, Numano R, Abe M, Hida A, Takahashi R, Ueda M, Block GD, Sakaki Y, Menaker M, Tei H (2000) Resetting central and peripheral circadian oscillators in transgenic rats. Science 288:682–685 4. Meijer JH, Groos GA, Rusak B (1986) Luminance coding in a circadian pacemaker: the suprachiasmatic nucleus of the rat and the hamster. Brain Res 382:109–118 5. Meijer JH, Rusak B, G€anshirt G (1992) The relation between light-induced discharge in the

suprachiasmatic nucleus and phase shifts of hamster circadian rhythms. Brain Res 598 (1–2):257–263 6. Walmsley L, Brown TM (2015) Eye-specific visual processing in the mouse suprachiasmatic nuclei. J Physiol 5937:1731–1743 7. Walmsley L, Hanna L, Mouland J, Martial F, West A, Smedley AR, Bechtold DA, Webb AR, Lucas RJ, Brown TM (2015) Colour as a signal for entraining the mammalian circadian clock. PLoS Biol 13(4):e1002127 8. Brown TM, Wynne J, Piggins HD, Lucas RJ (2011) Multiple hypothalamic cell populations encoding distinct visual information. J Physiol 589(5):1173–1194 9. Howarth M, Walmsley L, Brown TM (2014) Binocular integration in the mouse lateral geniculate nuclei. Curr Biol 24 (11):1241–1247

262

Joshua Mouland et al.

10. Storchi R, Bedford RA, Martial FP, Allen AE, Wynne J, Montemurro MA, Petersen RS, Lucas RJ, Storchi R, Bedford RA, Martial FP, Allen AE, Wynne J (2017) Modulation of fast narrowband oscillations in the mouse retina and dLGN according to background light intensity report modulation of fast narrowband oscillations in the mouse retina and dLGN according to background light intensity. Neuron 93(2):299–307 11. Allen AE, Brown TM, Lucas RJ (2011) A distinct contribution of short-wavelength-sensitive cones to light-evoked activity in the mouse pretectal olivary nucleus. J Neurosci 31 (46):16833–16843 12. Dasilva M, Storchi R, Davis KE, Grieve KL, Lucas RJ (2016) Melanopsin supports irradiance-driven changes in maintained activity in the superior colliculus of the mouse. Eur J Neurosci 44:2314–2323

13. Sakhi K, Wegner S, Belle MDC, Howarth M, Delagrange P, Brown TM, Piggins HD (2016) Intrinsic and extrinsic cues regulate the daily profile of mouse lateral habenula neuronal activity. J Physiol 22:5025–5045 14. Maggi CA, Meli A (1986) Suitability of urethane anesthesia for physiopharmacological investigations in various systems. Part 1: general considerations. Experientia 42 (2):109–114 15. Nair G, Kim M, Nagaoka T, Olson DE, Thule´ PM, Pardue MT, Duong TQ (2011) Effects of common anesthetics on eye movement and electroretinogram. Doc Ophthalmol 122:163–176 16. Pachitariu M, Steinmetz N, Kadir S, Carandini M, Harris KD (2016) Kilosort: realtime spike-sorting for extracellular electrophysiology with hundreds of channels. bioRxiv

Chapter 20 Perforated Multi-Electrode Array Recording in Hypothalamic Brain Slices Mino D. C. Belle, Beatriz Ban˜o-Otalora, and Hugh D. Piggins Abstract The ability to record ensemble action potential (AP) discharge frequencies from large populations of neurons over extended periods of time in vitro offers clear advantages in neuroscience and circadian biology research. Here, we provide an overview of a step-by-step method to perform multisite extracellular AP activity recordings in suprachiasmatic and mediobasal hypothalamic nuclei brain slices, using a state-of-theart perforated multielectrode array system. Further, we describe in detail a setup architecture which systematically delivers stable, high-quality recordings with excellent anatomical accuracy and consistency. We also provide some procedural, technical, and methodological troubleshooting notes and examples of good quality recordings. Key words Perforated multielectrode array, Electrophysiology, Suprachiasmatic nuclei, Mediobasal hypothalamus, Multiunit activity, Circadian

1

Introduction The development of multielectrode methods to simultaneously record and analyze ensemble activity from circuit-wide neuronal populations represents a powerful advance in neuroscience [1, 2]. This is especially useful in circadian biology studies as these tools can simultaneously probe the unfolding landscape of activity dynamics of large populations of neurons with incredible temporal and spatial resolution [3–9]. However, acute brain slice recording performed with traditional multielectrode array (MEA) systems can suffer from a number of shortcomings, such as poor signal-to-noise ratio (due to poor contact between the brain slice and recording electrodes) and unintended tissue movement over time. These become particularly apparent when measuring electrical output from slices over several hours or at the circadian timescale. From our own experience and through the work of others (e.g., [3, 4, 10]), we have concluded that these and other technical issues are readily resolved by using dual-perfusion perforated MEA (pMEA)

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_20, © Springer Science+Business Media, LLC, part of Springer Nature 2021

263

264

Mino D. C. Belle et al.

systems. For example, to make multisite recordings, the surface of the brain slice must be in continuous contact with the fabric of the array, and in conventional MEAs, this can restrict oxygenation and nutrient supply to the recording surface and compromise slice viability. The pMEAs solve this important issue by allowing artificial cerebrospinal fluid (aCSF) to penetrate deep into the parenchyma, providing ample oxygenation and nutrient supply to the brain slices [11]. Together, these highlight how the pMEA systems are especially powerful for recording electrical activity in brain slices over several hours. In this chapter, we therefore document an overall step-by-step method of how to achieve such recordings, placing tissue survivability and recording stability at the forefront while attaining excellent signal-to-noise ratio.

2

Materials

2.1 Equipment and Background 2.1.1 MEA2100-HS2x60-System Workstation

We describe the setting up and specification for perforated MEA recording using the new state-of-the-art dual-MEA2100-HS2x60 head-stage system from Multi Channel Systems (MCS GmbH, Reutlingen, Germany; see Figs. 1 and 2 for our setup layout; and http://www.multichannelsystems.com/systems/multiwellmea2100-2x60-system). This device integrates data acquisition from 120 MEA recording electrodes (60 electrodes per well/chamber/probe: hereafter appropriately referred to as pMEA probe/ chamber/array) and communicates with the computer via the MEA2100-interface board. In addition, a 2-channel temperature controller (TC02), two heatable perfusion cannulas with temperature sensor (PH01), 2 peristaltic perfusion systems (PPS2), a Constant Vacuum Pump (CVP), and other accessories are needed (detailed below and shown in Figs. 1 and 2). We recommend purchasing all these from MCS, if possible, as they have been designed for the optimal functioning of the MEA system. The methods and procedures we describe below can be easily adapted to use with the single head-stage perforated MEA2100-HS60 or with the older MEA1060 amplifier systems (see MCS website). The dual head-stage system (labeled as 1 and 2 in Figs. 1d and 2d), however, offers the unique advantage for two recordings to be performed in parallel from brain slices collected from two independent animals under appropriate experimental conditions (e.g. knockout vs its congenic wild-type littermate; treated vs control animal). This provides an ideal internal control. For an experienced user, setting up for dual-pMEA recordings, including aCSF and brain slice preparation and acclimatization, takes approximately 3–4 h.

Perforated Multi-Electrode Array Recording

265

Fig. 1 Schematic layout for our pMEA system setup. (a) aCSF reservoirs, (b) drug delivery reservoirs, (c) PPS2 peristaltic pumps, (d) pMEA head-stage unit with dual-recording capability (labeled as 1 and 2). Inset (d1) shows the cross section of a pMEA–perfusion element assembly in situ. (e) CVP pump, (f) MEA2100-interface board, (g) 2-channel temperature controlling unit (TC02), and (h) acquisition computer. Perfusion tubes are color-coded to indicate upper and lower inflow and outflow systems. Blue arrows indicate the direction of perfusion flow (see also Fig. 3j, k). (a1) shows the hollow glass tubes, (a2) the inflow tubing, and (a3) the tubes that supply 95% O2 and 5% CO2 contained in the gas bag to the bottles via the hollow glass tubes. This assembly offsets the vacuum that builds inside the bottles as the aCSF drains. A Tap configuration key is provided with red arrows indicating directions of solution flow when syringe plunger is pulled or pushed for suction and flushing, respectively. (i) shows a diagram of the syringe attached to a 3-way tap, and (ii) Tap Position-A to -C configurations as referred to in the text. Thick green arrows indicate points where earth wires fitted with crocodile clips can be attached, with inset (a4) showing an enlarged diagram of such tubing-earth points. PPS2 2 peristaltic perfusion system, PH01 heatable perfusion cannulas with temperature sensor, TC02 2-channel temperature controller, CVP Constant Vacuum Pump, MPH magnetic perfusion holder. See Fig. 2 for a photograph of the setup with associated labeling 2.1.2 Perforated MEAs (60pMEA100/30iR-Ti-gr or 60pMEA200/30iR-Ti-gr)

The 60pMEA100/30iR-Ti-gr and 60pMEA200/30iR-Ti-gr are pMEAs with 60 titanium nitride electrodes, 30 μm in diameter each, embedded in a perforated polyimide foil (Figs. 4a and 5a). The perforations allow for aCSF exchange from above and below the pMEA probes. When carefully set up, this permits the formation of a negative pressure/suction underneath the brain slice (see below). The electrodes in the 60pMEA100/30iR-Ti-gr MEAs are

266

Mino D. C. Belle et al.

Fig. 2 Photograph of our pMEA system setup. Letters (a–h) identify the components shown schematically in Fig. 1. (i), gas bag, (j) rubber bung with hollow glass tube and tubing assembly, (k) water bath, (l) ramps of stopcocks Luer-lock for drug delivery, (m) heatable cannulas (PH01), (n) syringe–tap assemblies for under perfusion system (see Fig. 1 for Tap-A to -D setup), (o) gooseneck of light source, and (p) binocular microscope

100 μm apart, arranged in a 6  10 layout (ideally suited for small brain structures such as the suprachiasmatic nuclei (SCN)), while in the 60pMEA200/30iR-Ti-gr the electrodes are 200 μm apart arranged in an 8  8 layout (suited for larger brain circuits such as the mediobasal hypothalamus (MBH); see Figs. 4a and 5a). Unless necessary, we recommend to use pMEAs with 6 mm high glass ring (identified by the code -Ti-gr), rather than plastic rings (-Ti-pr), as these are more durable and, with due diligence and care, can be used for at least 40 experiments. From new use and between uses, store pMEAs overnight at 4  C in an enzyme-active powder detergent (such as 1% Terg-a-zyme; Sigma-Aldrich).

Perforated Multi-Electrode Array Recording

267

Solutions

Prepare incubation/cutting and recording aCSF solutions in ultrapure deionized and distilled water (e.g., using the MilliQ water purifying system to attain a sensitivity of 18 MΩ-cm at 25  C; Merck Millipore). When preparing these aCSF solutions, use analytical grade salts (For example, VWR-AnalaR NORMAPUR grades). For accuracy, use calibrated volumetric flasks to prepare and/or dilute all solutions (e.g., Fisherbrand™ clear borosilicate glass with stopper). Make up 10 and 5 stock solutions for recording and incubation/cutting aCSF solutions, respectively. It is advisable to store these solutions at 4  C. Dilute the incubation solution from the 5 stock on the day before recording and keep overnight at 4  C. Prepare the recording solution from the 10 stock on the day of recording.

2.2.1 Cutting Solution

The composition of the cutting/incubation aCSF can slightly vary across laboratories, but ours constitutes in mM: NaCl 95; KCl 1.8; KH2PO4 1.2; CaCl2 0.5; MgSO4 7; NaHCO3 26; glucose 15; sucrose 50; Phenol Red 0.005 mg/L; oxygenated with 95% O2; 5% CO2; pH 7.4. We recommend using such low Na+/Ca2+, high Mg2+ sucrose-based cutting/incubation aCSF while preparing SCN and MBH brain slices as this maintains and protects the tissue from unavoidable anoxia and neurotoxic cell death that occurs during decapitation, dissection, and slicing [12].

2.2.2 Recording Solution

The composition of the recording solution is identical to the incubation/cutting solution, except for the following in mM: NaCl 127; CaCl2 2.4; MgSO4 1.3; sucrose 0. It is advisable that for long-term recordings, antibiotic solution (such as Gentamicin, Sigma G1272 1 mL/L) is added to the aCSF. Set the water bath at 1  C higher than the temperature setting chosen for the heatable perfusion cannulas (PH01, see below), and warm the aCSF solution in the water bath for at least 1 h. This allows the aCSF solution to uniformly reach the desirable temperature before gassing. Gas the aCSF to saturation, which normally takes ~20 min, before adding CaCl2. We use a sealed aCSF reservoir-flow system to store and circulate aCSF (see Note 1 and Figs. 1a and 2). Although the osmolarity of the aCSF solutions can be estimated/calculated using empirical methods, we strongly recommend to measure this directly with an osmometer prior to use. For both solutions, this should be in the 305–315 mOsmol/kg range. Addition of the pH indicator Phenol Red in the aCSF provides an instantaneous visual cue (color change) to any alteration in the gaseous status/content of the aCSF. This is helpful for monitoring aCSF gas saturation both during gassing and recording.

2.2

268

Mino D. C. Belle et al.

Fig. 3 Photograph of the pMEA and Parafilm protection lay-down process, and over perfusion system setup. (a) shows one of the two perfusion elements with (a1) identifying the grove/notch, (a2 and a3) showing inlet and outlet of the under perfusion system, respectively, (b) positioning of the blue rubber O-ring (b1) in the grove/ notch, (c) shows the O-ring snuggly in place in the grove with (c1) demonstrating the water pool referred to as 3c in the text. This water pool becomes continuous within the inlet and outlet tubing (from the reservoir bottle to the CVP pump: see text), (d) depicts the coverslipping action during the pMEA lay-down process, showing the leading water edge (d1), (e) displays the pMEA properly laid down onto the blue rubber O-ring with no air bubbles and overspills, and (f) showing an example of a poorly laid array, indicating a large pocket of air (f1) trapped underneath the pMEA chamber. (e1) shows the electrode contact pads (where the electrodes form

Perforated Multi-Electrode Array Recording

3

269

Methods

3.1 Brain Slice Preparation

Deeply anaesthetize the animal with isoflurane (Baxter, Norfolk, UK; code: FAGG9623) to minimize pain, discomfort and stress, and kill by cervical dislocation followed by decapitation. Using a razor blade, make an incision along the scalp, carefully cut away the top and side parts of the cranium and remove the brain. During this process avoid stretching or applying mechanical strain to the optic nerve since these can damage the optic chiasm and preoptic regions of the hypothalamus, including the SCN. Rest the brain (ventral side up) on an aCSF-soaked tissue paper and prepare a block of tissue containing the whole hypothalamus by grossly cutting away the anterior and posterior aspects of the brain. Glue the caudal side of the tissue block (rostral side facing uppermost) on the stage of a vibroslicer. Once in the cutting chamber, submerge the tissue block in ice-cold incubation/cutting aCSF solution and, with the ventral side of the hypothalamic block facing the blade, carefully cut coronal hypothalamic slices containing the SCN and MBH. As a rule of thumb, we recommend using ceramic blades, as these are sharper than stainless steel ones, producing less damage to the tissue. The section thickness should not be less than 200 μm (250 μm is an ideal thickness, but you may want to cut at 300–400 μm for long-term recordings: see Note 11). Transfer the slices to a holding chamber in which the incubation/cutting aCSF solution is continuously gassed at room temperature. Allow the slices to rest and recover here for at least 30 min. Finally, transfer the slices to a separate holding chamber containing continuously gassed recording aCSF solution, and leave the slices to acclimatize here at room temperature for at least another 30 min for full recovery before transferring to the pMEA chamber. Use these time windows to set up the MEA system.

3.2 MEA Setup: Overview for a One-Off Setup from New

We provide a comprehensive set of diagrams and photographs to show the overall setup layout of our dual-head-stage pMEA system (Figs. 1, 2, and 3), which offers efficient, ergonomic and easy access during slice setup. This way of assembly also enables exceedingly stable and high-quality recordings to be performed from brain slices in a systematic way, allowing for great anatomical accuracy and consistency across experiments. The two head-stages (labeled as 1 and 2 in Figs. 1d and 2d, and built into one unit) and pMEA chambers operate independently, permitting independent aCSF/ drug delivery to the recording tissues and acquisition. It is advisable


Fig. 3 (continued) connection with the head-stage). (g–i) show the Parafilm setting up process with (j and k) showing the complete over perfusion setup without (j) and with (k) Parafilm protection in place. Blue arrows indicate direction of over perfusion inflow and outflow. (j1) are the magnetic perfusion holders (MPH), and (k1-2) showing the tissue anchor (k1) and slice (k2) assembly. (k3) shows the beveled end of the suction cannula. See also Fig. 1. REF reference electrode

270

Mino D. C. Belle et al.

to assemble the whole unit, except for the aCSF reservoirs water bath, in a Faraday cage to shield from electric fields electromagnetic radiation that can contribute noise to recordings. For each of the pMEA chambers there are two perfusion tems; an “upper” and “lower” perfusion (Figs. 1 and 2). 3.2.1 Upper Perfusion System

and and the sys-

The upper perfusion supplies the brain slices with a continuous “inflow” of fresh oxygenated aCSF and a “suction outflow” to remove excess aCSF waste. To drive the inflow and outflow of each pMEA chamber we use, and recommend using, two separate peristaltic perfusion systems (MCS; PPS2). Each PPS2 contains two independent pumps. Use one for inflow and the other for outflow (Figs. 1 and 2). The aCSF solution is guided into the pMEA chamber by a cannula bearing an inline heating system (PH01). Position the inflow and outflow cannulas as shown in Figs. 1d, d1 and 3j, k to minimize turbulence during solution exchange and maximize drug accessibility to the slices. The components are: Inflow, in sequence; l

Two 2 L bottles filled with aCSF (place these in a water bath. Use larger aCSF reservoirs for recordings over 8 h. Estimate the volume needed by your flow rate; Figs. 1a and 2a).

l

Two ramps of stopcocks Luer-lock (e.g., 4 one-way stopcocks with a 3-way stopcock at one end; Vygon Ltd., REF873.15) to hold and vent the drug reservoirs (can be fitted with plastic barrels from 20 mL syringes: Figs. 1b and 2b, l).

l

Two hypodermic needle shafts or small metal barrels (see Note 2 and Fig. 1a4).

l

Two PPS2 peristaltic pumps.

l

Two PH01 temperature units (connect these to a TC02 temperature controller) with cannulas and magnetic perfusion holders (MPH).

l

Tubing (2 ~1.6 mm internal diameter; MCS code: PPS2-SetF).

l

Connectors for tubing assembling (MCS code: PPS2-Set-F). Outflow, in sequence;

l

Two metal cannulas/suction tubes with beveled end (e.g., Warner Instruments, ST-1 L/R 64-1401: Fig. 3k3).

l

Two magnetic perfusion holders (MPH).

l

Tubing (2 ~1.6 mm internal diameter; MCS code: PPS2-Set-F).

l

Large waste bottle (store on the floor).

Perforated Multi-Electrode Array Recording

271

Fig. 4 Photograph of 6  10 (a) pMEA electrode layout. Notice the perforation lattices of various sizes distributed among the electrodes. (b) Photograph of SCN slice being recorded in situ on a 6  10 pMEA probe. Notice that the pMEA electrodes can be visually identified from underneath the slice. (c) 1 s snapshot of extracellular action potential (AP) discharge at the electrodes from the SCN slice shown in (b). APs are highlighted as they cross the threshold line (here set at 17.5 μV, shown as enlarged inset in (c1). (d) shows 1 s snapshot of the recording during 1 μM TTX application (~5–10 min), indicating the absence of APs. Notice the threshold lines in black underneath the traces. These can be individually set, or the same value can be applied to all channels. (e) shows time series rate histograms (Hz) obtained from measuring AP frequency over 800 s with the acute rise in activity depicting examples of AMPA treatment which transiently excited neurons across the SCN slice. One channel of this located in the ventral SCN (marked with a green box) is enlarged in e1 with top black line indicating the duration of the AMPA application (150 s). (c1 and d1) are enlarged insets of channels taken at regions marked by the blue boxes in c and d, respectively, showing a 1 s recording during control aCSF (c1) and in aCSF containing TTX (d1). (c2) shows an example of SCN multiunit AP activity waveforms. REF: reference electrode; t (in d1): example of threshold line

Other accessories l

A binocular microscope.

l

Halogen cold light source fitted with dual-gooseneck light guards. If you have a dual head-stage MEA system, using two of these is advisable to provide illumination from above and underneath the pMEA chambers. The light sources are important for MEA setup, tissue placement, and in situ photographing of tissue–pMEA chamber assembly (see Figs. 2, 4, and 5).

l

Tissue anchor (e.g., Warner Instruments; code 64-1421: see Fig. 3k1).

272

Mino D. C. Belle et al.

Fig. 5 Photograph of 8  8 (a) pMEA electrode layout. Notice the perforation lattices of various sizes distributed among the electrodes. (b) Photograph of MBH slice being recorded in situ on a 8  8 pMEA probe. Notice that the pMEA electrodes can be visually identified from underneath the slices. (c) 1 s snapshot of extracellular action potential (AP) discharge at the electrodes from the MBH slice shown in (b). APs are highlighted as they cross the threshold line (here set at 15.5 μV, shown as enlarged inset in c1). (d) shows 1 s snapshot of the recordings during 1 μM TTX application (~5–10 min), indicating the absence of APs. c1 and d1) are enlarged insets of channels taken at regions marked by the blue boxes in c and d, respectively, showing a 1 s recording during control aCSF (c1) and in aCSF containing TTX (d1). Notice the threshold lines in black underneath the traces. These can be individually set, or the same value can be applied to all channels. c2 shows an example of MBH multiunit AP activity waveforms. REF: reference electrode; t (in d1): example of threshold line 3.2.2 Lower Perfusion System

The lower perfusion is partly driven by gravity and partly by the CVP pump (Figs. 1 and 2). Its main purpose is to provide continuous gentle suction underneath the pMEA chamber using fresh and oxygenated aCSF. This draws the slices firmly onto the electrode contacts at the bottom of the pMEA chamber, providing several benefits (see introduction and Fig. 1d1). The components are as follows: Inflow, in sequence; l

The two aCSF-filled 2 L bottles mentioned above (if running two pMEA chambers set these so that one bottle supplies fresh aCSF to the upper and the other bottle to the lower perfusion systems, as shown in Figs. 1 and 2).

l

Two 3-way stopcocks Luer-lock taps (e.g., Kendall 170,030; see also Fig. 1 labeled as Tap-A and -B).

l

Two 20 mL syringes (see Tap configuration key in Fig. 1).

l

Two perfusion elements (MEA2100-PE2x60: Figs. 1d1 and 3a).

l

Two rubber O-rings (o-Ring-PGP: Figs. 1d1 and 3b1).

Perforated Multi-Electrode Array Recording

273

l

Tubing (2 ~1.6 mm internal diameter; MCS code: PPS2-Set-F).

l

Connectors for tube assembling (MCS code: PPS2-Set-F). Outflow, in sequence;

l

Tubing (2 ~1.6 mm internal diameter; MCS code: PPS2-Set-F).

l

Two 3-way stopcocks Luer-lock taps (as above, labeled as Tap-C and -D).

l

Two 18G hypodermic needles (file down the sharp beveled end and attach to one end of the 3-way tap (see Fig. 1 and Note 3).

l

A Y-tubing connector (see Fig. 1).

l

A Constant Vacuum Pump (CVP, see Figs. 1 and 2), and appropriate tubing (>2 mm internal diameter) and connectors.

3.3 MEA Recording Setup

To protect the MEA system and to prime the tubing pipework initiate the MEA setup using distilled water. Ensure that the system is running smoothly before switching to aCSF. The overall aim is to run the upper and lower perfusion systems to establish a smooth solution exchange in the pMEA chambers. To achieve near noisefree, high-quality recordings with lasting stability and durability, three crucial conditions must be attained: (1) the pMEA chambers must be tightly sealed onto the blue rubber O-ring and perfusion elements (perfusion elements, MEA2100-PE2x60, are part of the baseplate); (2) there must be NO air bubbles in any part of the tubing and within the water pool that is contained by the rubber O-ring and sandwiched between the underside of the pMEA chambers and the perfusion elements; and (3) the water pool underneath the pMEA probe is well contained by the rubber O-ring and has not overspill during setup (regions outside the rubber O-ring MUST BE completely DRY; see Note 4 and see Figs. 1d1 and 3).

3.3.1 Priming the Perfusion and Vacuum Tubing

Prepare the lower perfusion first. Addition of a binocular microscope (not shown in the diagram, but present in the photograph in Fig. 2) is strongly advised, as this allows the setup process to be supervised to greater precision as well as permitting more accurate brain slice positioning and subsequent image capture. Inflow of the lower perfusion l

Immerse the ends of the four perfusion tubes (two for upper and lower perfusion inflows, respectively) into a 2 L bottle filled with distilled water (use a third and separate bottle designated to circulate water through the system only). Ensure that the tubes are fully immersed with their ends resting at the bottom of the bottle.

l

Open the lid of the MEA (see Note 5) to expose the perfusion elements (MEA2100-PE2x60).

274

Mino D. C. Belle et al. l

Thoroughly dry two blue rubber O-rings and place each one snuggly into the grove/notch of the perfusion elements (see Figs. 1d1 and 3b, c, and Note 6).

l

Fit the Taps (A–D) with 20 mL syringes (see Figs. 1 and 2).

l

Start setting up the perfusion element of head-stage 1, then move to perfusion element of head-stage 2 (Fig. 1), if you plan to set up and run two pMEA probes in parallel.

l

Place Tap-A in Position-A and half-fill the attached syringe with water by a gentle pull on the plunger.

l

Place Tap-A in Position-B and gently push the plunger until a water column starts to appear at the inlet of the perfusion element, forming a concave meniscus (see Fig. 3c).

l

Using Tap-B perform similar preparation to the second perfusion element if you plan to record from two pMEA probes in parallel. Outflow of the lower perfusion

3.3.2 pMEA Placement on MEA Perfusion Element

l

Half-fill the syringe attached to Tap-D with distilled water (notice that Tap-D is on the same perfusion system as Tap-A, but attached to the outlet tubing connected to the CVP).

l

Place Tap-D in Position-A and gently push the plunger until a water column starts to appear at the outlet of the perfusion element, forming a concave meniscus (Fig. 3c).

l

Continue to gently push water through until the two water pools merge, engulfing both the inlet and outlet of the perfusion element. It is IMPORTANT that the surface tension of the water pool is maintained and that the area immediate to the inside region of blue rubber O-ring is kept dry (Fig. 3c).

l

Using Tap-C perform similar preparation to the second perfusion element.

l

During setup simply use your index finger to “flick” at the tubing and connectors to remove trapped air bubbles.

l

After a careful and thorough rinse with distilled water (see Note 7), meticulously dry the electrode contact pads (Fig. 3e1) and underside of the pMEA chamber with soft tissue paper. AVOID any contact with the area of the pMEA that bears the perforations.

l

Follow the guidance of MCS for where the pMEA reference electrode should be pointing. For example, for the 6  10 and 8  8 layouts, the reference electrode should point to the left, aligning with electrodes 14 and 15, respectively (see Figs. 3–5). Once the reference electrode is identified and correctly aligned, carefully lay down the pMEA onto the blue rubber O-ring by

Perforated Multi-Electrode Array Recording

275

doing the following (see Fig. 3d, e; we call this the “coverslipping-like” lay-down as the action to accomplish this resembles microscope slides coverslipping): Place one edge of the pMEA chamber onto the rubber O-ring and keep it in place with two fingers while gently lowering the opposite end of the chamber toward the water pool at the center of the perfusion element (Fig. 3c, d). Once the underside of the pMEA chamber touches the water pool, keep lowering the unit in a “coverslipping-like” fashion until it forms a watertight seal onto the blue rubber O-ring (see Fig. 3e). Do this gently, slowly, and gradually until a perfect seal is achieved. Before completing the seal, sufficient water should be added (use the syringes). The aim is to have the entire volume encased between the underside of the pMEA chamber and the perfusion element completely filled, leaving no gaps or air bubbles once the pMEA is sealed onto the rubber O-ring (compare Fig. 3e, f). This is a critical step and it is tricky to get it right, requiring diligence and dexterity to avoid overspills and/or introducing air bubbles. The appropriate volume of water (Fig. 3c) needed to do this comes with practice. Use the syringes to add extra or remove excess water during pMEA placement, but be careful not to introduce air bubbles or to cause an overspill outside the confinement of the blue rubber O-ring. l

Repeat the above steps to prep the second pMEA chamber.

l

Finally leave Tap-C and -D in Position-B to close the under flow system.

l

Inspect the assembly. IMPORTANT: make sure that all the water is contained within the confinement of the blue rubber O-ring and that there is no overspill. Make sure that the electrical contact pads on the pMEA are dry and debris free.

l

Slowly close and lock the MEA. At this stage two sheets of Parafilm, one for each pMEA probe, can be added. These guard the interior of the MEA and gold connector pins against potential pMEA chamber floods which if happen during the setup and/or recording process can cause acute and long-term issues, such as severe damage to the equipment. This is therefore strongly advised, and if you wish to do this, see Note 8 and Fig. 3g–i.

l

Use the two syringes attached to Tap-C and -D to fill the tube connected to the CVP.

l

Place Tap-C and -D in Position-A to maintain the water column in the CVP tube in place.

l

Open Tap-A and -B by placing them in Position-C, to slowly begin filling the pMEA chambers with distilled water by gravity (see Note 9). Once the base of the pMEA chambers is water-

276

Mino D. C. Belle et al.

coated (supervise this under the binocular microscope), close the inflow by returning these taps to Position-B. l

The inflows for the lower perfusion systems are set to run very slowly (~0.8 mL/min). Inflow and outflow of the upper perfusion

l

Set up the inflow and outflow for the over perfusion system as shown in Figs. 1–3, using the magnetic perfusion holders provided (see also Note 10).

l

Place the inflow cannulas as shown in Figs. 1d, d1 and 3j, k. This minimizes the mechanical strain on the brain slice produced by the incoming inflow solution currents.

l

Place the outflow cannulas as shown in Figs. 1d, d1 and 3j, k. It is best if the tip of the cannulas is beveled. This ensures continuous skimming of the aCSF solution surface without any intermittent make-break contact with the solution, which can introduce electrical noise and/or cause turbulence in the aCSF-filled chamber that supports the tissue.

l

Switch on the PPS2 peristaltic pumps. For optimum stability during slice recording set the inflow to 1.8 mL/min and outflow at 17 mL/min.

l

Turn on the CVP (usually set at 40–43 mbar, depending on slice thickness and brain region, see Note 11).

l

Open the taps (Tap-A to -D in Position-C) and turn on the peristaltic pumps to start running the over and under perfusion systems. Allow water to circulate for at least 10 min.

l

Inspect the system making sure that all is running smoothly, then switch to aCSF.

l

Run with aCSF for at least 20 min to ensure complete aCSF/ water exchange.

l

Turn on the MEA head-stage, heater controller and MCS software (see below). IMPORTANT: Do not turn on the MEA head-stage while running with water. To do so may reduce the life of the pMEA electrodes.

l

Run the MCS acquisition software, making sure that the hardware baseline is stable and within acceptable noise level (~5 μV, except reference electrode number 14 or 15 which should be considerably less; ~1 μV, see also Note 12).

l

Leave the software to run while you place and position the brain sections in the pMEA chambers.

Perforated Multi-Electrode Array Recording 3.3.3 Positioning and Aligning the Brain Slice onto the Recording Electrodes

277

l

Momentarily stop the under perfusion by closing Tap-A to -D (Position-B for Tap-A and -B, and Position-A for Tap-C and -D).

l

Cut off the tip of a 3 mL disposable plastic Pasteur pipette (for example Sigma, Z331740-1PAK) to create an opening with a wider aperture (~8 mm in diameter). Carefully draw the required brain section into this modified pipette by using suction, then, transfer the section into one of the chambers. Avoid using a paintbrush to do this as the bristles could damage the slice. The Pasteur pipette transfer method also ensures that the slice is always maintained in a pool of gassed aCSF.

l

Allow the slice to rest flat at the bottom of the pMEA chamber, then with a pair of fine forceps gently place a slice anchor over the tissue (see Fig. 3k3).

l

Using a paintbrush with very fine natural bristles and under the visual guide of the binocular microscope, maneuver the anchored slice until the brain region you want to record from lies directly above and is appropriately aligned with the electrodes.

l

If you are running two pMEA probes, repeat the above for the second chamber.

l

Leave the slices to settle for a few minutes.

l

Slowly open Tap-C (Position-C). This exposes the slice to the gentle negative suction produced by the CVP. While opening the tap, you should therefore observe the slice being drawn down toward and onto the electrode array. Within a few minutes, extracellular AP activity should start appearing on the computer screen (see Figs. 4 and 5). If you do not see any of these, consult Note 13.

l

Repeat this for the second pMEA chamber by opening Tap-D (Position-C).

l

Except for your reference electrode, your noise level should now be at ~8–10 μV.

l

Then, carefully open Tap-A and -B (Position-C) to allow the CVP to draw some of the aCSF directly from the underflow inlet/aCSF reservoir.

l

If all is well, the recording should now be stable and there should be a gradual increase in AP discharge frequency as the slices stabilize and settle on the electrodes.

l

Leave the slices to acclimatize for at least 30 min before starting to record the data. If you choose to start saving the data straight away, do not use the first 30 min of the data in analysis as this typically represents instability in recordings.

l

It is advisable to always capture an image of all recorded slices at an appropriate magnification in situ (e.g., Figs. 4 and 5). For

278

Mino D. C. Belle et al.

this, a high-definition webcam is sufficient. Take a photograph of the electrode array with the same magnification. This can be done prior to tissue placement. You can then align and superimpose the two images to visualize the precise location of each recording electrodes, if they are not already visible through the brain slice (see Figs. 4 and 5). Tips: Use anatomical landmarks in the slice, such as ventricular space, tissue edges, and fiber bundles, to consistently align and position the recording region of interest with a predetermined electrode (see Figs. 4 and 5). This will approximately maintain anatomical consistency across your recordings which may become important during analysis. Further, try to perform your recordings in slices cut at roughly the same level on the rostral-caudal axis.

4

Software, Data Preparation and Analysis The software manuals from Multi Channel Systems (MCS; http:// www.multichannelsystems.com/downloads/software) provide detailed instructions for software installation, setup, and use. Here, we provide start-up instructions for the older MC_Rack software, but there is now the newer and improved MC_Suite program for data acquisition, visualization, and basic analysis. If you choose to use MC_Rack, download MC_Data tool for data management. If you have also purchased a PPS2 and TC01/TC02 systems, installing the appropriate software (PPS2 and TCX-Control, respectively) will allow easy access and control of these devices via USB. What follows below includes an example of a useful way to build a “Rack,” which is a recording configuration provided in the MC_Rack software (see Figs. 4 and 5).

4.1 Data Acquisition: Action Potential/ Spikes Recording

l

Load the MC_Rack software.

l

In general, it is advisable to have the online display showing: (1) long-term and short-term (30 s and 1 s, respectively) raw unfiltered data streams; (2) high-pass filtered data (1 s) with threshold-line indicator to isolate spikes (Figs. 4 and 5); and (3) multiunit activity rate histograms (Fig. 4e). The latter are critical as they provide live second-by-second updates of your recording, alerting you of any recording problems. The online rate histograms also provide the perfect visual guide for appropriate timing of drug treatment, cues for duration of drug actions and recovery to baseline activity (Fig. 4e, e1).

l

To create a Rack containing these displays, do the following: In the Rack, click on the hardware’s name (e.g., MEA2100 (1/2) SN) and add Filter, Display, and Long-term display. Click on the Filter icon and add Spike sorter, Analyzer, and Parameter display

Perforated Multi-Electrode Array Recording

279

(rate histograms). Click on Filter and choose the Butterworth second order high-pass filter with cutoff frequency set at 300 Hz. One of the options in spike sorter allows you to set spike threshold (please refer to the appropriate part of the manual for more details). Once you have built a rack you can save it for future use. If you are running two pMEA chambers you will need to run two separate MC_Racks side by side. Ensure that each one communicates with the appropriate pMEA chamber. l

l

4.2 Offline Data Extraction, Assembly, and Analysis

This system can sustain continuous high-speed USB data transfer with a maximum sampling rate of up to 50 kHz/channel (for most recordings we recommend using 25 kHz). Although we strongly recommend saving the raw data files during acquisition, this may not always be practical, especially for longer-term recordings (>10 h). This is because raw data files acquiring at high sampling rate require large amount of storage space (>100 GB/day). One way to circumvent this is to record spikes and spike-time stamps only (at 50 kHz these require ~5 GB/ day). The down-side of this is you need to establish which online threshold values to use before you start recording. This can have disadvantages if you encounter a prolonged period of sporadic electrical noise during your recording, for example. In these circumstances you will not be able to rethreshold your spike data to exclude such noise after recording or in future analysis. Recording at lower sampling rate will also reduce storage space needed, but if too low the spike waveform will become distorted, thereby rendering spike sorting more difficult to achieve. At the end of your experiment apply tetrodotoxin (TTX, 1 μM; Tocris) for at least 10 min. This silences AP discharge and produces an activity-free segment of the recording trace that can be used to guide analysis (see Figs. 4d, d1 and 5d, d1, and Note 14).

l

During your experiment, MC_Rack saves your data in discrete files. So before extracting spike frequency and time-stamps, you need to concatenate the files with MC_Data tool.

l

Build a Rack as described above. Remember, this time you will be replaying and rerecording the data, extracting the spikes and spike time-stamps in the process (consult the manual) for further analysis. So, in your Rack, there must be an additional “replay” icon.

l

Once extracted, spike frequency and time-stamps within the file can be read by other commercially available software for more sophisticated analysis, such as “NeuroExplorer” for multiunit activity (MUA) frequency analysis and “Offline Sorter” for spike sorting (dissecting single units from MUA; Plexon; http:// www.plexon.com). The data can also be exported to MATLAB

280

Mino D. C. Belle et al.

(https://uk.mathworks.com) and/or Python (https://www. python.org/)/ANACONDA (https://www.continuum.io) for customized scripts /algorithms-based analysis. l

5

MUA frequency analysis from large datasets is labor-intensive and requires long periods of time to complete. Further, some of the commercially available software are expensive and they are not designed for circadian data analysis in mind (such as accurate adjustment for Zeitgeber- or circadian-time, and/or cull-, and recording-time). We are preparing a MATLAB-based circadian tool for fast, user-friendly and systematic analysis of baseline MUA acquired across multiple brain slices over the circadian day. This tool also provides speedy and highly accurate measurements of acute drug effects on MUA across all 60 electrodes. This will be available in due course.

Drug Tests To investigate whether all aspects of the recording setup are operating as should be, we recommend performing a drug application test (e.g., Fig. 4e, e1). This allows the user to approximately time drug action-onset, duration, and washout (this depends on the drug being used). For most brain circuits, exogenous application of fast-acting/washout agonists such as (RS)-α-amino-3hydroxy-5-methyl-4-isoxazolepropionic acid hydrobromide (a water-soluble form of AMPA) or γ-aminobutyric acid (GABA) (Tocris) will do, as an initial one-off guide. We strongly recommend setting up the ramps of the stopcock Luer-lock system to avoid drug cross-contamination. This also provides accurate control when switching from control aCSF to drug-containing aCSF (see Figs. 1b and 2l). For our setup, and as a rule of thumb, most drugs should take 1–2 min to reach the pMEA chambers.

6

Shutdown and Cleaning l

Capture and image your slices in situ (as shown in Figs. 4b and 5b).

l

Stop the acquisition software.

l

With a fine pair of forceps carefully remove the tissue anchor.

l

Close Tap-C and -D (Position-A) to relieve the slices from the CVP suction. The gravity-led inflow should now start to lift the tissue off the pMEA chamber base, away from the electrode contacts.

Perforated Multi-Electrode Array Recording

7

281

l

With a fine paintbrush lift the slices completely off the electrodes, removing them from the pMEA chamber. The slices can be used for further processing, such as immunocytochemistry.

l

Close Tap-A and -B (Position-A or -B) and momentarily stop the PPS2 peristaltic pumps.

l

Wash and switch the inflow tubes (upper and lower) to a 2 L bottle filled with distilled water.

l

Restart the upper perfusion by switching the PPS2 peristaltic pumps back on.

l

Place Tap-A and -B in Position-A, and using the attached syringes withdraw fluid until the distilled water enters the syringes.

l

Place Tap-A to -D in Position-C.

l

Slowly the distilled water will replace the circulating aCSF. Run for at least 1 h. We recommend running 70% ethanol through the tubing system for at least 10 min after each experiment to clear-off stubborn contaminants, especially if drug treatments were used.

l

Withdraw the tubing ends from the distilled water and thoroughly drain the tubing.

l

Open the MEA lid and carefully lift and remove the pMEA chambers.

l

Remove the Parafilm, if this has been applied, and store the pMEA chambers in 1% Terg-a-zyme at 4  C. Store the tissue anchor and blue rubber O-ring in distilled water at room temperature.

l

Dry the perfusion elements, inspect the gold pins for signs of damage/spillage/aCSF salt residue, and close the MEA lid.

l

Carefully and diligently inspect the area around the MEA headstage, pumps, tubing etc., ensuring that they are clean and dry with no aCSF salt residue, and tubing/wires etc. are not out of place.

Notes 1. Figures 1 and 2 show how to set up two sealed reservoir-flow system. Use rubber bungs that can completely seal off the aCSF reservoirs, preventing dissolved gas escaping from the aCSF and leaking from the bottles (Figs. 1a and 2j). Bore three holes in each rubber bung, two for the tubing, and in the third insert a hollow glass tube (Figs. 1a1 and 2j). This glass tube must be long enough to span the distance from the bung to near the base of the bottle. Connect this tube to a leakresistant gas bag to equilibrate negative pressure build-up

282

Mino D. C. Belle et al.

during aCSF flow (Figs. 1a3 and 2i). Fill, but do not pressurize this bag with gas containing 95% O2 and 5% CO2. 2. Cut off the hub of four 18G hypodermic needles and file down the sharp beveled end. Cut an appropriate section of the inflow tubes, ahead of the peristaltic pumps, and under inflows, and insert the needle shafts as shown in Fig. 1 (enlarged in 1a4). Ensure that where you cut the tubes allows the tube–needle– tube assemblies to lie within the Faraday cage. Use this point to earth the system by appropriately attaching crocodile clips and earth wires (see Note 3). 3. These provide earth-points for aCSF leaving the MEA, reducing electrical noise. Attach these to crocodile clips and earth wires (see also Notes 2 and 12). 4. Any liquid overspills beyond the confinement of the blue rubber O-rings will introduce and magnify electrical noise, significantly reducing signal-to-noise ratio. Excessive overspills can damage the electronic components of the MEA head-stage. 5. Using both hands gently press down on the MEA lid and slide off the lid retainer/lock. 6. This is a critical step. Ensure that the blue rubber O-ring sits snuggly in the groove of the perfusion element. Failure to align this properly will cause the MEA to leak rendering it unusable (see Note 4 and Figs. 1d1 and 3b, c). 7. The pMEAs are stored in detergents. So they must be properly washed before use. Place each pMEA in a small beaker filled with distilled water. Flush additional distilled water through the beakers for at least 2–3 min, emptying the beakers twice or three times during this process. Carefully examine the pMEA probes under a binocular microscope, ensuring that there are no residues or debris on the electrode terminals or contact pads. DO NOT at any point allow the pMEA electrodes to get totally dry. 8. The compactness of the MEA2100-HS2x60 system makes it extremely vulnerable to accidental flooding or stray water/ aCSF spray/droplets, which if come into contact with any of the MEA’s electronic components, including the gold pin contacts, will introduce noise. These floods/spray/droplets can cause long-term damage to the system. To circumvent this, we have developed an in-house method for adding an extra layer of protection against accidental spills. Excise a piece of Parafilm (Bemis Company Ltd., USA: PM996) 2 (~10  10 cm ) and stretch it slightly. Center it over the pMEA chamber and fix in place by gripping a rubber O-ring around the outside wall of the chamber, firmly holding the Parafilm in place (see sequence in Fig. 3g–i). For this to properly work, the internal diameter of the O-ring must be slightly

Perforated Multi-Electrode Array Recording

283

less than the external diameter of the pMEA chamber. With a razor blade, cut off the top section of the Parafilm exposing the opening of the pMEA chamber (see Fig. 3g–i, k show the finished outcome). 9. Place the aCSF reservoirs a few centimeters higher than the perfusion elements. 10. Besides using the magnetic perfusion holder, provide additional support/reinforcement to the cannula–PH01 temperature unit assemblies to ensure that these always remain in place. This can simply be done by applying “blu tack” (Bostik Ltd., UK) under the PH01 units. 11. Appropriate setting for this depends entirely on your setup and which pMEA chambers you are using (6  10 or 8  8). The optimum value can only be achieved in situ by trial and error during the initial one-off setup. Start with 42 mbar and increase this pressure if the tissue is not being firmly pulled down onto the pMEA and hold in place. Success in achieving good contact between the slice and the electrode can be identified visually from the level of AP activity emanating from the slices. Thicker sections need higher negative pressure, for longer recordings (more than 10 h) tend to err toward lower values. This is because the negative pressure does exert some mechanical hold on the tissue and lower suction values during long-term recordings ensuring that the sections are not overly stretched or stressed. However, keep in mind that lower negative pressure settings may compromise the signal-to-noise ratio, but this slight offset in recording quality contributes toward slice longevity. 12. These values depend on how well the rig is earthed. How to properly and successfully earth a rig requires systematic checking. Electrical noise can be a real challenge to completely eliminate and requires experience. In Fig. 1 and Note 2 we have indicated strategic locations along the setup that can be used as earth-points. These helps to maintain low electrical noise. Connect these earth-points to the Faraday cage and table top using crocodile clips and electrical wires. The best rule of thumb to avoid electrical noise is to always maintain the rig clean, dry, and salt residue free. 13. If your slice is in good physiological condition (alive and undamaged), but you cannot detect any AP discharge, this means that there is not enough contact between your slice and the pMEA electrodes. Slowly increase the CVP suction. If the issue is not resolved, you may have air bubbles trapped in the underflow tubing which are preventing negative pressure transfer from the CVP to the pMEA chamber and slice. Remove the slices from the pMEA chamber and use the

284

Mino D. C. Belle et al.

syringes to try and rectify this. Please note that the inflow and outflow/negative pressure for the under perfusion system need to be carefully balanced to accommodate for tissue thickness and recording duration (see Note 11). 14. TTX treatment at the end of your experiments confirms AP activity and provides a highly reliable and accurate way to determine signal threshold. Use this TTX segment of your recording to determine and set threshold level for spike extraction (see Figs. 4d, d1 and 5d, d1).

8

Conclusion Here, we provide a step-by-step procedural overview for setting up and recording from brain slices using pMEAs. In particular, we describe fitting methods for the new MEA2100-HS2x60 system, but similar approaches can be used to set up and record with other pMEA devices currently available. We find that this way of the pMEA system assembly, regimented setup, and recording methods provide stable, highly reproducible, and high-quality recordings that can last for several hours. When appropriately fine-tuned (see Note 11), this setup method can be used to interrogate neuronal activity well over the circadian day.

Acknowledgments This work is supported by project grant funding to H.D.P and M. D.C.B. by the BBSRC (BB/L007665/1). B.B.O. was funded by a Postdoctoral Fellowship from the Spanish Fundacio´n Se´neca (19701/PD/14). We thank Dr. J. Turner for his invaluable comments on an earlier draft of the manuscript. References 1. Franke F, Jackel D, Dragas J et al (2012) Highdensity microelectrode array recordings and real-time spike sorting for closed-loop experiments: an emerging technology to study neural plasticity. Front Neural Circuits 6:105 2. Shaban H, O’Connor R, Ovsepian SV et al (2017) Electrophysiological approaches to unravel the neurobiological basis of appetite and satiety: use of the multielectrode array as a screening strategy. Drug Discov Today 22:31–42 3. Hanna L, Walmsley L, Pienaar A et al (2017) Geniculohypothalamic GABAergic projections gate suprachiasmatic nucleus responses to retinal input. J Physiol 595:3621–3649. https:// doi.org/10.1113/JP273850

4. Walmsley L, Hanna L, Mouland J et al (2015) Colour as a signal for entraining the mammalian circadian clock. PLoS Biol 13:e1002127 5. Van Gelder RN, Herzog ED, Schwartz WJ et al (2003) Circadian rhythms: in the loop at last. Science 300:1534–1535 6. Freeman GM Jr, Krock RM, Aton SJ et al (2013) GABA networks destabilize genetic oscillations in the circadian pacemaker. Neuron 78:799–806 7. Honma S, Ono D, Suzuki Y et al (2012) Suprachiasmatic nucleus: cellular clocks and networks. Prog Brain Res 199:129–141 8. Honma S, Nakamura W, Shirakawa T et al (2006) Monitoring the clock neuron’s tick: circadian rhythm analysis using a multi-

Perforated Multi-Electrode Array Recording electrode array dish. In: Taketani M, Baudry M (eds) Advances in network electrophysiology using multi-electrode arrays. Springer, Singapore 9. Ono D, Honma S, Honma K (2013) Cryptochromes are critical for the development of coherent circadian rhythms in the mouse suprachiasmatic nucleus. Nat Commun 4:1666 10. Reinhard K, Tikidji-Hamburyan A, Seitter H et al (2014) Step-by-step instructions for retina recordings with perforated multi electrode arrays. PLoS One 9:e106148

285

11. Egert U, Okujeni S, Nisch W et al (2005) Perforated microelectrode arrays optimize oxygen availability and signal-to-noise ratio in brain slice recordings. In: 5th international meeting on Substrate-integrated Micro Electrode Arrays, Freiburg 12. Moyer JR Jr, Brown TH (1998) Methods for whole-cell recording from visually preselected neurons of perirhinal cortex in brain slices from young and aging rats. J Neurosci Methods 86:35–54

Chapter 21 Collection of Mouse Brain Slices for Bioluminescence Imaging of Circadian Clock Networks Jennifer A. Evans, David K. Welsh, and Alec J. Davidson Abstract Circadian rhythms in cellular function can be monitored in real time with bioluminescence imaging. In this approach, bioluminescence is produced by an enzymatic reaction, which can be used to report dynamic changes in gene or protein expression in living cells. Bioluminescence imaging in circadian experiments typically uses an ex vivo slice preparation, with the most commonly studied structure being the master clock in the suprachiasmatic nucleus (SCN) of the anterior hypothalamus. Here we describe procedures for dissecting and collecting SCN slices for bioluminescence imaging experiments. Key words Bioluminescence imaging, PER2::LUC, Mouse, Suprachiasmatic nucleus, Circadian clock

1

Introduction Insight into the function of the SCN network has benefited from the advent of optical reporters for clock gene and protein expression. Using time-lapsed imaging, this technology allows one to track clock gene/protein expression as it changes in real time, which has many advantages over traditional methods for detecting daily rhythms. This approach uses transgenic rodents in which the production of a fluorescent or bioluminescent protein is under the control of clock gene regulatory elements. Relative to fluorescence imaging, bioluminescence imaging is better suited for the longterm recordings needed in circadian studies due to the absence of photobleaching, phototoxicity, and background. In circadian reporter animals, luciferase is produced whenever and wherever a specific clock gene or protein is expressed (e.g., luciferase driven by promoters for Period1 [1, 2]; Bmal1 [3, 4], or Cry1 [5] genes). In addition, the PERIOD2::LUCIFERASE knockin mice harbor a fusion protein that serves as a translational reporter [6]. Often, bioluminescence imaging is used to investigate SCN network properties in an ex vivo slice preparation. Here we describe procedures for collecting SCN slices for bioluminescence imaging.

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_21, © Springer Science+Business Media, LLC, part of Springer Nature 2021

287

288

2

Jennifer A. Evans et al.

Materials

2.1 Dissecting Medium

Hanks Balanced Salt Solution (HBSS, -CaCl2, -MgCl2, -MgSO4) supplemented with HEPES, NaHCO3, B27, and penicillin/ streptomycin. Recipe for Supplemented HBSS (1 L) 300 mL Milli-Q water

100 mL HBSS 10
stock (Gibco 14065) 5 mL HEPES (Gibco 15630) 10 mL NaHCO3 (2
rinse, Gibco 25080) 10 mL pen/strep (2
rinse, Gibco 15140) Adjust volume to 1 L by adding Milli-Q water. Adjust pH to 7.18  0.02. Filter into 500 mL bottles with Corning filter system. Store at 4  C (should be good for at least 1 month). 2.2 Recording Medium

Dulbecco’s Modified Eagle Medium (DMEM) supplemented with HEPES, NaHCO3, B27, penicillin/streptomycin, and luciferin. Recipe for Supplemented DMEM (1 L) 600 mL Milli-Q water

1 pkg./bottle DMEM (2
rinse, Sigma D2902—no phenol red) Alterative DMEM with higher glucose: Gibco 12100-046 (with phenol red). 20 mL B27 (2 bottles, avoid forming foam, Gibco 17504) 4.7 mL NaHCO3 (2
rinse, Gibco 25080) 10 mL HEPES (Gibco 15630) 2.5 mL pen/strep (Gibco 15140) 3.5 g D-(+)Glucose (2
rinse, Sigma G7021)) Adjust volume to 1 L by adding Milli-Q water. Adjust pH to 7.18  0.02. Filter and aliquot into 50 mL Falcon tubes. Store at 4  C (should be good for at least 6 months). On day of culture, add 50 μL luciferin. Wrap tube in aluminum foil to protect from light. Store DMEM +LUC at 4  C (should be good for at least 2 weeks). You can prewarm DMEM+LUC prior to dissections (always check for growth before using).

Collecting SCN for Bioluminescence Imaging

2.3 Substrate for Bioluminescence Reaction: Luciferin

289

Recipe for Beetle LUCIFERIN with K SALT 50 mg D-luciferin bottle (GoldBio LUCK-1G)

Add 1.67 mL of autoclaved Milli-Q water. Aliquot into 50 μL tubes. Store at 80  C.

3

Methods

3.1 Methods for Different Age Mice

For adult mice, SCN slices are cut using a motorized vibraslicer at 4  C. SCN slices can also be collected from postnatal mice 4–7 days old with a tissue chopper (see Note 1). The procedures below describe collection of SCN slices from adult mice.

3.2 Remove the Brain

Mice can be euthanized with CO2 or cervical dislocation under isoflurane anesthesia. Decapitate the mouse using a large pair of sharp scissors. Cut the skull along the midline and laterally at each of the temporal lobes. Then the skull can be removed easily by slipping the point of medium scissors under the midline incision and applying a counterforce to snap the skull plates away from the underlying brain. When the skull is unroofed, remove the brain ventral side up. Cut the optic nerves anterior to the optic chiasm with small spring scissors, avoiding contact with the hypothalamic region or excessive traction on the optic nerves. Cut the optic nerves as far from the chiasm as possible. If it is difficult to visualize or cut the optic nerves, remove the skull over the olfactory bulb. Cut the skull overlying the olfactory bulb by inserting the point of the medium scissors into the cavity along the lateral edge of each olfactory bulb. It should be easy to cut the optic nerves and remove the brain when both olfactory bulbs are fully unroofed.

3.3

Transfer the brain to a culture dish containing Hanks Balanced Salt Solution (HBSS, see recipe below) and a piece of submerged filter paper. Keep the culture dish with supplemented HBSS on ice until needed so that the solution is cold while blocking the brain. Block the brain on top of the filter paper using halved double-edged razor blades. When purchased, the razors should be soaked in acetone and sterilized in 70% EtOH prior to use. Perform a caudal cut to remove the brainstem and a rostral cut just anterior to where the optic nerves begin (see Fig. 1). Care should be taken to make the caudal cut as straight as possible and oriented orthogonal to the cranial–caudal axis so that the brain can be mounted flat onto the stage of the vibraslicer (see Note 2). Dorsal and lateral cuts can be made as well, depending on the plane of section and which brain region is being collected for imaging (see Note 3).

Block the Brain

290

Jennifer A. Evans et al.

Fig. 1 Location of caudal and rostral cuts recommended for blocking the brain 3.4 Mount the Brain on the Stage for Sectioning

Once the blocked brain is ready, gently scoop it onto a spatula ventral side up. Remove excess liquid from the underside of the brain block by dabbing the brain block onto filter paper using another spatula to lift the tissue. Once dry, place the spatula with the blocked brain off to the side. Apply fast-drying cyanoacrylic adhesive (“Super Glue”) to the stage provided with the vibraslicer and spread evenly into a thin layer. Prior to use, the stage should be cleaned of all glue with a utility razor and 70% EtOH. For coronal sections, place the blocked brain onto the stage caudal side down, ventral side facing forward (see Notes 3 and 4). Apply a small drop of supplemented HBSS to the chiasm area with a clean paintbrush, avoiding contact with the stage.

3.5 Insert the Stage into the Vibraslicer

Allow the glue to dry completely before placing the stage into the bath chamber of the vibraslicer. The bath chamber should be filled with supplemented HBSS cooled to 4  C. Keeping the supplemented HBSS at a low temperature during sectioning is important to maintain the viability of adult SCN slices. This can be accomplished with a circulating chiller, an ice pack surrounding the bath chamber, or a frozen slurry of supplemented HBSS added to the bath chamber.

3.6 Sectioning the Brain Block

Lower the blade to the surface of the brain, then retract the blade, and start sectioning. Discard slices until the contralateral projection of the anterior commissure disappears, at which point it is wise for a novice to start to retain slices. Take 100–300 μm slices, trying to obtain at least one slice before and after the SCN (see Note 5). The optic nerves will join into a small triangular shape that flattens across the ventral surface of the hypothalamus at the level of the

Collecting SCN for Bioluminescence Imaging

291

Fig. 2 Illustration of SCN slices collected during brain sectioning. Top: Midsagittal slice of mouse brain illustrates the anteroposterior position of each coronal slice labeled a–e. Nissl stain imaged from Allen Brain Atlas. Bottom: Coronal sections illustrating morphology of SCN using cresyl violet staining (section thickness: 12 μm). On the right hand side of each coronal image, the SCN and optic nerve are labeled with a dashed line and a solid line, respectively. (a) Section anterior to the SCN, with a thin strip of SCN neurons. (b) Anterior SCN. (c) Middle SCN. (d) Posterior SCN. (e) Section posterior to the SCN

SCN (see Fig. 2). Transfer slices with a broad paintbrush to a 60 mm dish containing supplemented HBSS kept on ice. The slicing procedure takes 15–30 min depending on the slice thickness and the advance speed. It is best to advance the blade slowly and set the lateral vibration frequency to the fastest setting. 3.7 Select and Trim Slices

Examine retained slices under a dissecting microscope. When lit from below, the SCN should appear as a dense nucleus positioned immediately adjacent to the optic chiasm, at the midline, on either side of the third ventricle. Typically, one can obtain three 150 μm slices containing SCN per adult animal (see Fig. 2). Sections not containing the SCN are discarded. Slices containing the SCN are trimmed with a curved scalpel blade to ~2 mm
2 mm. A small notch can be made while trimming to help orient a specific side of the slice toward the camera.

3.8 Transfer SCN to Culture Dish

Each trimmed slice is placed gently on a Millipore membrane insert (PICMORG50) within a 35 mm culture dish containing 1.2 mL of recording medium supplemented with luciferin (see recipes below). Take care to avoid damaging the SCN or optic tissue during this

292

Jennifer A. Evans et al.

process. One effective method for transferring the slice is to use a pipettor that has had the tip cut at an angle with a scalpel. Another method is to use a paintbrush once the membrane is moistened with a small pool of supplemented HBSS. Using either of these two methods should prevent the morphology of the SCN from being distorted, which may limit data analyses. Once transferred, the slice should be positioned flat and near the center of the membrane using a small paintbrush. Remove any excess HBSS with a pipettor and/or filter paper. Finally, the dish is sealed using a thin ring of silicone vacuum grease. 3.9 Acclimate Slices Prior to Imaging

Once sealed, the culture dish containing the SCN slice may be retained in a standard tissue culture incubator (36–37  C, 0% CO2) until ready for imaging. It is best to leave the dish in the incubator for at least 2 h to allow the slice to stabilize before imaging. This will reduce the amount of medium under the slice and prevent the slice from going out of focus or drifting out of frame over the first few hours of imaging (see Note 6).

3.10

For SCN slices collected from wild-type mice, the bioluminescence signal should be evident immediately upon recording. If no signal is detected, check that the recording medium contains luciferin. For circadian studies, slices should be imaged for a minimum of 3 days, during which the bioluminescence rhythm should persist. If the bioluminescence rhythm vanishes abruptly, then it is likely that the dish has become contaminated or the recording medium has evaporated due to a nonoptimal vacuum seal. Because the bioluminescence signal is much lower than that provided by fluorescent probes, there are special considerations for microscope and camera components (see Note 7).

4

Image Slices

Notes 1. For SCN dissections of early postnatal mice, coronal slices can be obtained with a tissue chopper. Once removed from the skull, the brain should be blocked down to the hypothalamus on filter paper moistened with dissection medium and resting on the lid of a 100 mm tissue culture dish. Transfer the brain block on the moist filter paper to the stage of the tissue chopper to collect 200–400 μm slices at room temperature. Lubricate the blade of the tissue chopper for each section. Drop the blade arm from a height of 1 cm to avoid bouncing on the stage. Collect 3–4 slices once the ventral edge appears straight rather than V-shaped. Select and culture SCN slice(s) as described above for adult slices.

Collecting SCN for Bioluminescence Imaging

293

2. When collecting SCN slices 250 μm, the anatomical position of the slices obtained from different mice should be as consistent as possible because the rhythmic properties of SCN neurons depend on their anteroposterior location [7]. The consistency of sectioning can be improved by blocking the brain in an identical manner across mice. Another important consideration is to use mice of the same age and sex in experiments. When using adult mice (i.e., after the onset of puberty), it is best to maintain 4 weeks variation in age within and across experimental groups. 3. If the width of the brain block is less than its height, it may be unstable while cutting. This will cause the brain block to lean during slicing and likely result in damage to the SCN. This can be addressed by placing a small agar block (2% agarose) along the ventral surface of the brain block during mounting to stage, which will support the brain block while slicing. When using this method, place the brain block on the stage first, and then quickly place the agar block before the glue on the stage is dry. Be sure that the underside of the agar block is dry before mounting, as you did for brain block, otherwise it will not adhere to the stage. While mounting the agar block, do not allow it to damage the SCN physically or by pushing excess glue up along the ventral side of the brain. 4. It is helpful to orient the brain block on the stage so that landmarks are visible that permit one to easily monitor progression of sectioning. Also, if the tissue block is not properly centered when the stage is mounted into the vibraslicer, it may be difficult to obtain clean slices with no damage to the optic chiasm and SCN. If using the agar block method, the agar is typically the last thing to be cut. This will allow you to visualize the SCN in its upright orientation to track progress while cutting through the nucleus working rostral-to-caudal in the coronal plane. 5. If the brain block will not remain attached to the stage, slice thickness is variable, or no slices can be obtained, it is likely that the brain block and/or agar block is not securely mounted to the stage. The underside of the tissue block or agar may have been wet when mounted or the glue may have already dried by the time the tissue was placed on the stage. It is often difficult to re-mount without damaging the sample, especially those from adults. 6. If the acute slice preparation needs to be very stable for immediate imaging, there are additional steps that can reduce tissue movement. Strips of filter paper can be used to wick away the medium from the edges of the slice, although care must be

294

Jennifer A. Evans et al.

taken not to touch the slice itself. This should be relatively easy to accomplish with slices 150 μm when done under a dissecting microscope. 7. For bioluminescence imaging, the microscope and camera used should collect and detect as much light as possible. Special considerations for microscope optics and camera specifications for bioluminescence imaging are described in detail in a previous review [8]. References 1. Yamaguchi S, Mitsui S, Miyake S, Yan L, Onishi H, Yagita K, Suzuki M, Shibata S, Kobayashi M, Okamura H (2000) The 50 upstream region of mPer1 gene contains two promoters and is responsible for circadian oscillation. Curr Biol 10(14):873–876 2. Yamazaki S, Numano R, Abe M, Hida A, Takahashi R, Ueda M, Block GD, Sakaki Y, Menaker M, Tei H (2000) Resetting central and peripheral circadian oscillators in transgenic rats. Science 288(5466):682–685 3. Noguchi T, Ikeda M, Ohmiya Y, Nakajima Y (2012) A dual-color luciferase assay system reveals circadian resetting of cultured fibroblasts by co-cultured adrenal glands. PLoS One 7(5): e37093 4. Myung J, Hong S, Hatanaka F, Nakajima Y, De Schutter E, Takumi T (2012) Period coding of Bmal1 oscillators in the suprachiasmatic nucleus. J Neurosci 32(26):8900–8918

5. Maywood ES, Drynan L, Chesham JE, Edwards MD, Dardente H, Fustin JM, Hazlerigg DG, O’Neill JS, Codner GF, Smyllie NJ, Brancaccio M, Hastings MH (2013) Analysis of core circadian feedback loop in suprachiasmatic nucleus of mCry1-luc transgenic reporter mouse. Proc Natl Acad Sci 110(23):9547–9552 6. Yoo SH, Yamazaki S, Lowrey PL, Shimomura K, Ko CH, Buhr ED, Siepka SM, Hong HK, Oh WJ, Yoo OJ, Menaker M, Takahashi JS (2004) PERIOD2::LUCIFERASE real-time reporting of circadian dynamics reveals persistent circadian oscillations in mouse peripheral tissues. Proc Natl Acad Sci 101(15):5339–5346 7. Evans JA, Leise TL, Castanon-Cervantes O, Davidson AJ (2011) Intrinsic regulation of spatiotemporal organization within the suprachiasmatic nucleus. PLoS One 6(1):e15869 8. Welsh DK, Noguchi T (2012) Cellular bioluminescence imaging. Cold Spring Harb Protoc 2012(8):pdb.top070607

Chapter 22 Computational Analysis of PER2::LUC Imaging Data Tanya L. Leise Abstract Advances in imaging technology, combined with the development of bioluminescent reporters for core clock genes, has enabled the observation of spatiotemporal patterns of circadian rhythms in the suprachiasmatic nuclei (SCN). In particular, the PERIOD2::luciferase (PER2::LUC) knockin mouse has led to novel approaches for studying the heterogeneous circadian network in the SCN. This chapter describes how to automate the processing of PER2::LUC imaging data from SCN slices for generating spatiotemporal maps of circadian parameters like phase, period, and amplitude. These maps can be aligned and averaged to produce composite maps displaying common features across multiple slices. In addition, we describe a method for automated detection of cell-like regions of interest, to support the study of the neural network in the SCN. Key words Image processing, SCN, Bioluminescence, Spatiotemporal maps, Circadian oscillations, Automated ROI identification

1

Introduction Bioluminescence imaging allows for a detailed examination of circadian dynamics in the mammalian circadian pacemaker, the suprachiasmatic nuclei (SCN) [1]. In particular, the PERIOD2:: luciferase (PER2::LUC) knockin mouse has spurred a host of novel approaches for studying how the heterogeneous circadian network in the SCN is organized both spatially and temporally. Imaging data collected from SCN explants, typically coronal slices, has required the development of computational analyses for both spatial and time dimensions [2–4]. See the chapter by J. Evans on preparation of PER2::LUC SCN slices for collecting high-quality bioluminescence imaging data. After a sequence of images has been recorded, often stored as a stack of TIFF files, spatiotemporal maps can be generated that display measures like phase, period, or amplitude of the time series associated with each pixel in the sequence of images. Similar methods can be applied to other imaging data, such as intracellular calcium levels [5] and other tissue cultures [6]. A further step can be taken to align and average these maps to create

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_22, © Springer Science+Business Media, LLC, part of Springer Nature 2021

295

296

Tanya L. Leise

composite maps to reveal patterns occurring across multiple slices [2, 7]. Another approach is to use automated detection of cell-like regions of interest (ROIs) in stable image sequences [1, 2, 8, 9], which is particularly useful for neural network analysis of the SCN. For reliable results in assessing period and phase, imaging data should contain at least 3 cycles exhibiting good consistency, not damping too strongly or varying too greatly in cycle length. Images are typically collected at 10–60 min intervals. Longer exposures may produce images with a higher signal-to-noise ratio, while higher sampling rates may provide more precise phase estimates. However, the number of good quality cycles is generally more important than the sampling rate in generating reliable estimates of period and phase from oscillatory time series [10].

2

Computational Software Multiple sources of software are now available for image processing and time series analysis. Many researchers use custom scripts in MATLAB (The MathWorks, Inc) or other commercial software such as Igor (WaveMetrics). An excellent free resource for processing biological images is ImageJ (U.S. National Institutes of Health), which has many available plug-ins for a wide variety of image processing tasks. Custom plug-ins for ImageJ can also be programmed using Java.

3

Computational Methods These computational methods assume that the imaging data takes the form of an array of matrices indicating the brightness at each pixel location at each time point, for example, a recording stored as a stack of TIFF images. Individual images in the image sequence will sometimes be referred to as frames; each image or frame corresponds to a particular time point.

3.1 Generation of Pixel-Based Maps

1. To remove effects of cosmic rays and other extreme outlier noise from the images, set the highest and lowest 0.0001 quantiles of pixel values in the entire image stack to the mean value of the adjacent pixels in the previous, same, and next frames (that is, average over both spatial and temporal neighboring pixels) [3]. 2. Determine a typical cell diameter in terms of pixels and the desired resolution of the output map. Some researchers prefer each pixel in the output map to be roughly the size of a cell, say, 10 μm in diameter, while others prefer a finer resolution with cells covering multiple pixels. Determine an integer value

Analysis of Imaging Data

297

R such that each pixel in the desired output map represents R
R pixels in the original imaging data. 3. Convolve each frame with a kernel, such as a Gaussian function like  
4 ln 2 f ðx, y Þ ¼ exp  2 x 2 þ y 2 , R to reduce noise and adjust the resolution. That is, center the kernel function over a pixel location, multiply the kernel by the values at each of the pixels in a neighborhood of at least radius R, and then sum these values. This process more heavily weights values near the center pixel, while smoothing out noise. Break the resulting image into R
R squares and retain only one pixel from each R
R square, such as the central pixel (see Note 1). 4. Take the time series corresponding to each pixel in the processed image stack (each frame represents a time point) and determine whether it meets a set of criteria for being considered part of the SCN. One possible set of criteria is the following (where the thresholds should be adjusted based on the characteristics of a particular data set and imaging system): (a) Local maximum in autocorrelation (rhythmicity index) occurs at a lag between 18 and 30 h, with value above the 5% significance level (see Note 2). (b) Signal-to-noise ratio is above 1 (see Note 3). (c) The circadian component of the time series (obtained using a wavelet transform, filtering method, or other smoothing method) exhibits at least three peaks with time between successive peaks not differing by more than 2 h and amplitudes not excessively damping (ratios of second to first and of third to second remain above some threshold, say, 0.25). (d) The mean brightness is significantly above background values (compare brightness distributions sampled from regions in each frame that are clearly not SCN to those clearly from SCN to determine a threshold). 5. Calculate the period, phase, and amplitude [11] for the pixels with significantly rhythmic time series as determined by the criteria in the previous step. Maps can now be generated by graphing these values as a heat map. 3.2 Creation of Composite Maps

1. A set of SCN slices with sufficient consistency in location (see Note 4) can be aligned. Select a representative reference slice with a typical shape for that particular position in the SCN. 2. For each slice, sum the frames across the second day of recording (to avoid the spike in brightness that can occur during the

298

Tanya L. Leise

Fig. 1 Spatiotemporal maps. (a) Profile of a summed image, with the SCN outer region having value 1 and the brightest inner region having value 2. (b) Alignment of 10 slice profiles. Full overlap of the inner regions yields value 20 and full overlap of outer regions yields value 10. Partial overlaps have lower values. (c) Map showing peak time on the second cycle at each rhythmic pixel location for a single slice, created as described in Subheading 3.1. (d) Composite map showing circular mean of peak times, averaging across the aligned slices, created as described in Subheading 3.2. Graphs show reanalyzed LD18:6 data from [7]. Original frame size was 512
640 pixels, and processed frames were generated using R ¼ 2. Axes ticks indicate pixel locations in processed frames

first day). Choose two percentile levels of brightness that correspond to the SCN overall and to the brightest inner SCN region in these summed images, such as 84% and 92% (these percentiles depend on how much area the SCN region covers in the image). To create a profile of a slice, assign value 0 to pixels for which the summed value falls below the first percentile level (non-SCN background), assign value 1 if the summed value is between the two percentile levels (SCN outer region), and assign value 2 if the summed value is above the second percentile level (SCN inner region). See Fig. 1a for an example of a profile image. 3. Align the other slices to the reference slice by minimizing the sum of squared differences between the reference profile and the other profiles. Figure 1b shows a set of 10 profiles aligned to the reference profile shown in Fig. 1a. 4. Identify aligned pixel locations for which a large majority of slices exhibit significant rhythmicity, say, at least ¾ of the slices. These locations do not have to lie in the profiles, which are only used here for alignment purposes. Measures of interest can be averaged at each of these pixel locations across the aligned maps

Analysis of Imaging Data

299

(using only values from the slices with significant rhythmicity at each location). Examining the standard deviations is recommended to detect cases where slices are too dissimilar for valid averaging. For phase maps, check that periods are consistent across the slices, and use the circular mean to ensure a valid result when averaging phases. Figure 1c shows the phase map for the reference slice, and Fig. 1d shows the composite phase map, averaged across the aligned set of slices. 3.3 Identification of Cell-Like ROIs

1. Create a summed image highlighting areas that may be cells. Take the processed frames generated in Subheading 3.1 and sum the frames corresponding to the second day of recording. Also find the difference between the 10th and 90th percentiles at each pixel location across the time points of the second and third days. Multiply these two new images together. The resulting highlight image has largest values where the bioluminescence is both bright and varying widely over time, which is likely to correspond to cell locations. See Fig. 2a for an example of a highlight image. 2. Remove local background to isolate cell-like ROIs. A two-dimensional discrete wavelet transform (2DWT) may be applied to accomplish this, for example, using the WaveLab 850 MATLAB Toolbox (http://statweb.stanford.edu/ ~wavelab/). Apply the forward 2DWT using the Coiflet filter with six vanishing moments. Determine the wavelet level with resolution corresponding to roughly cell size. Reduce wavelet coefficients corresponding to other levels to 2% of their value, then apply the inverse 2DWT to this altered result. In the resulting image, set any values below the universal threshold [12] to zero. Figure 2b shows the result of this process applied to the summed image in Fig. 1a (see Note 5). 3. For each iteration of the ROI-picking process, identify the highest remaining value in the highlight image. This location will be classified as a rhythmic cell-like ROI if the following criteria are satisfied: (a) That pixel is located within a cluster of pixels with positive values forming a cell-sized area. (b) The bioluminescence during the second and third days at that location exhibits a significant cross-correlation with a 24 h-period sinusoidal curve. (c) The first 3 days of the circadian component contain three clear peaks. Alternatively, similar criteria to those used in Subheading 3.2 can also be employed. 4. To remove that location from consideration in the next iteration (whether or not it was identified as a valid ROI), subtract a

300

Tanya L. Leise

Fig. 2 Identification of cell-like ROIs. (a) Highlight image, indicating likely cell locations. (b) Cleaned image following wavelet transform procedure, exposing candidate cell locations. (c) Locations of cell-like ROIs generated by iterative process applied to image in B, as described in Subheading 3.3. (d) ROI peak times on second cycle. (e) Time series extracted from processed frames at ROI locations. Graphs show reanalyzed LD18:6 data from [7], as in Fig. 1. Cell size here is assumed to be roughly 5
5 pixels in the processed frames for computational purposes

template centered at that location from the highlight image. The template can be the Gaussian function given in Subheading 3.1 (using a value of R corresponding to cell diameter) multiplied by 1.05 times value of the highlight image at that location. 5. Repeat steps 3 and 4 until no locations remain in the highlight image that satisfy the required conditions for cell-like ROIs. See Fig. 2c for the set of ROI locations determined by this process for the summed image shown in Fig. 2b. 6. Use these ROI locations to create the time series for calculating period, phase, and amplitude of the ROIs. If a typical cell will cover multiple pixels, sum over an equivalent cell-sized area centered at each ROI location to generate the time series. Figure 2d shows the peak times of the second cycle for the ROIs with time series shown in Fig. 2e.

Analysis of Imaging Data

4

301

Notes 1. The convolution only needs to be calculated for the retained pixels, which should be at the same relative location in each R
R square. The decay rate in the Gaussian can be adjusted. For instance, increase the decay rate if a more localized domain is desired for the convolution. 2. A higher threshold can be used, such as rhythmicity index above 0.1, as the 5% significance level is a very low threshold, testing whether the time series is significantly different from white noise, rather than directly testing for rhythmicity. 3. The signal-to-noise ratio is the log of the ratio of the sum of squares of the signal over the sum of squares of the noise. Because the signal of interest here is circadian, one way to calculate the signal-to-noise ratio is to calculate the discrete wavelet transform of the time series [13] and treat the resulting circadian component as the signal and the sum of the remaining components as the noise. 4. See [2] for a metric to verify that slice placement is sufficiently consistent to allow for reliable alignment. Averaging across slices will not be valid unless all of the slices are from very similar positions in the SCN. 5. Other background subtraction methods may also be applied to the highlight image instead of the wavelet transform method described here. ImageJ, for example, has procedures for removing local background noise from images that may be effective for this purpose.

Acknowledgments The author gratefully acknowledges the assistance of Jennifer Evans and other colleagues, who have been very generous with their time and their data as we developed these techniques. References 1. Davidson AJ, Castanon-Cervantes O, Leise TL, Molyneux PC, Harrington ME (2009) Visualizing jet lag in the mouse suprachiasmatic nucleus and peripheral circadian timing system. Eur J Neurosci 29(1):171–180. https://doi.org/10.1111/j.1460-9568.2008. 06534.x 2. Evans JA, Leise TL, Castanon-Cervantes O, Davidson AJ (2011) Intrinsic regulation of spatiotemporal organization within the suprachiasmatic nucleus. PLoS One 6(1):e15869.

https://doi.org/10.1371/journal.pone. 0015869 3. Foley N, Tong T, Foley D, Lesauter J, Welsh D, Silver R (2011) Characterization of orderly spatiotemporal patterns of clock gene activation in mammalian suprachiasmatic nucleus. Eur J Neurosci 10:1851–1865. https://doi. org/10.1111/j.1460-9568.2011.07682.x 4. Myung J, Hong S, Hatanaka F, Nakajima Y, De Schutter E, Takumi T (2012) Period coding of Bmal1 oscillators in the suprachiasmatic

302

Tanya L. Leise

nucleus. J Neurosci 32(26):8900–8918. https://doi.org/10.1523/JNEUROSCI. 5586-11.2012 5. Enoki R, Kuroda S, Ono D, Hasan MT, Ueda T, Honma S, Honma K (2012) Topological specificity and hierarchical network of the circadian calcium rhythm in the suprachiasmatic nucleus. Proc Natl Acad Sci U S A 109 (52):21498–21503. https://doi.org/10. 1073/pnas.1214415110 6. Guenthner CJ, Luitje ME, Pyle LA, Molyneux PC, Yu JK, Li AS, Leise TL, Harrington ME (2014) Circadian rhythms of Per2::Luc in individual primary mouse hepatocytes and cultures. PLoS One 9(2):e87573. https://doi. org/10.1371/journal.pone.0087573 7. Evans JA, Leise TL, Castanon-Cervantes O, Davidson AJ (2013) Dynamic interactions mediated by nonredundant signaling mechanisms couple circadian clock neurons. Neuron 80(4):973–983. https://doi.org/10.1016/j. neuron.2013.08.022 8. Brancaccio M, Maywood ES, Chesham JE, Loudon AS, Hastings MH (2013) A Gq-Ca2 + axis controls circuit-level encoding of circadian time in the suprachiasmatic nucleus.

Neuron 78(4):714–728. https://doi.org/10. 1016/j.neuron.2013.03.011 9. Buijink MR, Almog A, Wit CB, Roethler O, Olde Engberink AH, Meijer JH, Garlaschelli D, Rohling JH, Michel S (2016) Evidence for weakened intercellular coupling in the mammalian circadian clock under long photoperiod. PLoS One 11(12):e0168954. https://doi.org/10.1371/journal.pone. 0168954 10. Cohen AL, Leise TL, Welsh DK (2012) Bayesian statistical analysis of circadian oscillations in fibroblasts. J Theor Biol 314:182–191. https://doi.org/10.1016/j.jtbi.2012.08.038 11. Dowse HB (2009) Analyses for physiological and behavioral rhythmicity. Methods Enzymol 454:141–174. https://doi.org/10.1016/ s0076-6879(08)03806-8 12. Donoho DL, Johnstone JM (1994) Ideal spatial adaptation by wavelet shrinkage. Biometrika 81(3):425–455 13. Leise TL, Harrington ME (2011) Waveletbased time series analysis of circadian rhythms. J Biol Rhythm 26(5):454–463. https://doi. org/10.1177/0748730411416330

Chapter 23 Electrophysiological Approaches to Studying the Suprachiasmatic Nucleus Stephan Michel, Takahiro J. Nakamura, Johanna H. Meijer, and Christopher S. Colwell Abstract In mammals, the part of the nervous system responsible for most circadian behavior can be localized to a bilaterally paired structure in the hypothalamus known as the suprachiasmatic nucleus (SCN). Understanding the mammalian circadian system will require a detailed multilevel analysis of neural SCN circuits ex vivo and in vivo. Many of the techniques and approaches that are used for the analysis of the circuitry driving circadian oscillations in the SCN are similar to those employed in other brain regions. There is, however, one fundamental difference that needs to be taken into consideration, that is, the physiological, cell, and molecular properties of SCN neurons vary with the time of day. In this chapter, we will consider the preparations and electrophysiological techniques that we have used to analyze the SCN circuit focusing on the acute brain slice and intact, freely moving animal. Key words Biological clock, Brain slice, Circadian, In vivo electrophysiology, Neural activity rhythms, Suprachiasmatic nucleus

1

Introduction The ability of the SCN neural population to generate neural activity rhythms in isolation from the rest of the organism has been documented in numerous studies using an acute brain slice preparation [1, 2]. Data obtained from the brain slice provided critical evidence in support of the view that the generation of circadian electrical activity rhythms is not a neural network property in which synaptic feedback information from one neuron to the other is required for the presence of circadian rhythmicity. Instead, these studies are all consistent with the idea that many SCN neurons are stable, selfsustained oscillators that have the intrinsic capacity to generate circadian rhythms. That said, there is evidence that circuit properties contribute to the robustness and precision of SCN circadian oscillations [3, 4]. In vivo recording of SCN neural activity has been critical to determine how these neural rhythms are influenced by

Steven A. Brown (ed.), Circadian Clocks: Methods and Protocols, Methods in Molecular Biology, vol. 2130, https://doi.org/10.1007/978-1-0716-0381-9_23, © Springer Science+Business Media, LLC, part of Springer Nature 2021

303

304

Stephan Michel et al.

the intact input [5–7] and, for example, has established how the behavioral state of the organism can feedback to regulate SCN activity [8]. One of the critical challenges of ongoing work is the development of an understanding of how a molecular feedback loop occurring at the level of transcription and translation interacts with the neuronal plasmalemmal membrane to produce physiological rhythms. It is not known how the molecular feedback loop drives the electrical rhythm in membrane processes, and this is a critical question for the field [2, 9]. In addition, there is a growing body of evidence that disruptions in the SCN neural activity rhythms are common in aging and diseases of the nervous system [10, 11]. Understanding the underlying mechanisms could explain how pathological processes alter the electrical output as well as lead to the development of possible treatment strategies. The techniques and approaches that are used for the analysis of the circuit driving circadian oscillations in the SCN are similar to those used in other brain regions. There is one fundamental difference that needs to be taken into consideration, that is, the physiological, cell, and molecular properties of SCN neurons vary with time of day. These cells continue to oscillate in the dish so the daily cycle needs to be a consideration independent on whether the experiments are conducted in vivo, using ex vivo brain slices or in cultured explants or neurons. For example, circadian oscillations in γ-aminobutyric acid (GABA)-mediated synaptic transmission in both presynaptic release and postsynaptic responses have been described [12, 13]. Moreover, the neurons show spontaneous circadian oscillations in electrical activity driven, at least in part, by daily rhythms in intrinsic membrane properties in these cells [2, 14]. Signaling pathways within these cells oscillate, including dynamic changes of cytosolic Ca2+ and cyclic adenosine monophosphate/protein kinase-A (cAMP/PKA) activity [15–18]. Since circadian oscillations are generated by transcriptional–translational feedback loops, a wide range of genes are rhythmically regulated including those involved in secretion and synaptic transmission critical for circuits. So for the circadian system, time of day is a critical variable that always has to be considered. The extent to which these circadian variations influence other brain circuits is not yet known, although it is well established that molecular oscillations in clock gene expression are widespread within the nervous system [19–22]. Given that the SCN system comprises a circuit of endogenous oscillators that is synchronized to light, the external lighting conditions for all of these experiments need to be carefully considered. When animals are in a light–dark (LD) cycle, they may exhibit day– night variation in a number of parameters. These rhythms are influenced by the internal circadian system and by the direct effects of light and dark. Consequently, a rhythm recorded under these conditions can be considered a diurnal but not necessarily a

Electrophysiology of the SCN

305

circadian rhythm, that is, an internally generated rhythm. An important—though little employed—experiment is to reverse the timing of the LD cycle and show that the resulting rhythm follows the change in the light schedule. This confirms that the diurnal rhythm is at least driven by the environmental conditions but does not necessarily address whether the rhythm has an endogenous origin. In order to isolate the endogenous circadian component driving these differences, it is necessary to hold the organism in constant conditions, that is, constant dark (DD) and constant temperature. In the intact system, the animal’s behavior and locomotor activity can be followed with the convention that the onset of activity for a nocturnal organism is defined as circadian time (CT) 12. While activity onset is a convenient and reliable way to follow the endogenous rhythm, other robust circadian outputs could be used including body temperature or heart rate. For generation of brain slices, the LD cycle in which the animal was held prior to being sacrificed is an excellent predictor of the activity phase of the SCN population. For the rest of this chapter, we will consider the techniques used to analyze the SCN circuit focusing on acute brain slice and in vivo recordings. The materials, methods, and approaches in employing each of these techniques including a short description of the advantages and disadvantages will be described. Certainly, other techniques can be usefully applied, but we will focus on the preparations and techniques with which we have working experience.

2

Materials

2.1 Materials for Brain Slice

1. Solutions (concentration in mM): Artificial cerebrospinal fluid (ACSF): NaCl 130, NaHCO3 26, KCl 3, MgCl2 2, NaH2PO4 1.25, CaCl2 2, glucose 10. Modified ACSF (for vibratome slicing): same as ACSF except MgCl2 5, CaCl2 1. The solutions are prepared the day before the experiment, gassed with 95% O2 5% CO2 (pH adjusted to 7.2–7.4, osmolality 290–300 mOsm) and kept at 4
C. For long-term recording of slices (>24 h), gentamycin (5 μg/ml) is added to the ACSF. 2. Anesthetic agents (4% isoflurane) and anesthetic chamber. 3. Sterilized dissection tools, guillotine for decapitation, dissecting scissors, forceps, spatulas, and razor blade for trimming. 4. Slice cutting: vibratome (e.g., Microslicer, DSK Model 1500E, Ted Pella, Redding, CA, USA or Leica VT 1200S, Wetzlar, Germany) or tissue chopper (e.g., McIlwain Tissue Chopper, The Mickle Laboratory Engineering, Guildford, UK).

306

Stephan Michel et al.

5. Electrophysiological recording equipment (patch clamp): vibration isolated table (e.g., TMC, Peabody, MA, USA), Faraday cage, manipulator (e.g., Sutter, Novato, CA, USA or Luig&Neumann, Ratingen, Germany), slice recording chamber (e.g., Warner Series 20, Hamden, CT, USA), gravity fed perfusion control (e.g., ALA perfusion systems, Farmingdale, NY, USA), gas (5 % CO2/95% O2) control and distribution, vacuum control, patch amplifier (e.g., Axon Axopatch 200B or HEKA EPC10, Lambrecht, Germany), control software (e.g., Axon pClamp or HEKA Patchmaster). 6. Electrophysiological recording equipment (long-term multiunit recording): vibration damped table, Faraday cage, temperature-controlled perfusion chamber, water bath, perfusion pump, vacuum control, extracellular amplifier (e.g., 5113 Low Noise Voltage Preamplifier, Signal Recovery, Oak Ridge, TN, USA), digitizer (e.g., CED 1401 with Spike2 software, CED, Cambridge, England). 7. Infrared (IR)-patch: Upright water immersion microscope equipped with IR-DIC (e.g., Zeiss Axioskop FS2plus, Carl Zeiss, Go¨ttingen, Germany) or Olympus BX51W, Olympus, Center Valley, PA, USA), IR-camera (e.g., Hamamatsu C2400 or Watec, 902H, Newburgh NY, USA), video monitor. 8. Calcium (Ca2+) imaging: Light source (e.g., X-Cite, XLED1, Excilitas Technologies, Waltham MA, USA or Axon DG4), cooled CCD camera (e.g., Princeton Instruments, Trenton, NJ, USA, Microview model 1317  1035pixel format or TILL Imago QE cooled interline CCD, 1376  1040 pixel), software (e.g., Metafluor, Molecular Devices). 2.2 Material for In Vivo Recordings

1. Anesthesia for mice C57/bl6: ketamine, 100 mg/kg; xylazine, 10 mg/kg; and atropine, 0.1 mg/kg). 2. Dental drill (Dremel, Racine, WI). 3. Stereotactic setup. 4. NaCl solution for rinsing. 5. Dental cement. 6. Cotton swab. 7. EtOH for disinfection of electrodes, metal screws, and surgical equipment. 8. Electrodes for MUA recordings, 2 twisted insulated stainless steel wires, 125 μm diameter, cut to 8 mm length plus one non-insulated ground electrode 125 μm diameter cut to 4 mm length (Plastics One, M333/3-B). 9. Electrode pedestal forming the connector for the electrodes.

Electrophysiology of the SCN

307

10. Preamplifier, custom-made (INA 101 AM, Burr-Brown; gain, 10 and AC amplifier band-pass, 500 Hz to 5 kHZ; gain, 10,000). Alternatively, preamplification can be used close to the head connector. 11. Window discriminator (custom-made or NeuroLog). 12. Power-1401 data acquisition (CED Instruments) using Sipke2 software. 13. Recording cage with swivel/commutator (Plastics One, SL3 + 3C) and counterbalance.

3

Methods

3.1 Acute SCN Brain Slices 3.1.1 General Features of SCN Slices

3.1.2 Methods for Generating SCN Slices

The acutely isolated brain slice has the advantages of offering accessibility and control while at the same time preserving many of the synaptic connections of the SCN circuitry. Previous workers have reported that the phase of the rhythms expressed in the SCN brain slice is predicted well by the prior LD cycle. For these reasons, it is an excellent ex vivo preparation to characterize synaptic communication in the SCN and to search for diurnal (i.e., day versus night) variations. The first concern in the use of acute SCN brain slices is the duration of time for which neural activity can be recorded. For the majority of brain regions, brain slices seem to remain healthy for 6–12 h. Due to this technical limitation, most experiments look for day–night variation by comparing properties of SCN neurons prepared in the day with those prepared during the night. For these experiments, animals are placed in constant conditions and the behavioral rhythm with wheel running activity. The brain slice can then be prepared at different phases of the daily cycle to evaluate circadian regulation. The same basic procedure for the preparation of brain slices is used for extracellular or intracellular sharp microelectrode or whole-cell patch-clamp recordings and Ca2+ imaging experiments. Animals are kept in LD regimes for a time sufficient to synchronize or entrain their circadian rhythm. Entrainment is typically achieved within 2–3 weeks, but it is advisable to validate the phase of the circadian clock of the animal at the time of preparation by monitoring the animal’s locomotor activity rhythm. This is essential when using manipulations of the LD cycle, such as shifting the phase in a “jet lag” experiment or switching to constant darkness. The animals are killed by decapitation at times determined by the LD cycle (“Zeitgeber” time, ZT) or the behavioral rhythms (subjective or circadian time, CT). The effect of the time of preparation itself on the subsequent phase of the rhythm in electrical activity is small [23, 24], but caution should be taken when slices are prepared at night. Procedures need to be performed either in dim far-red light or under infrared (IR) illumination with the help of an IR viewer.

308

Stephan Michel et al.

The brain slice procedures that we use are similar to those previously described [25–28]. Successful recording requires fast removal of the brain. The longer the time between decapitation and submersion of the brain into the ice-cold slice solution, the poorer the quality of the neurons in terms of resting membrane potential and other properties. The goal is to complete this part of the procedure in 1–2 min. Similarly, the success rate in long-term (>24 h) recordings of acute brain slices is improved by minimizing the total time it takes to transfer the slices to the recordings chamber (4–6 min). It is also critical to cut the optic nerves very carefully in an early stage of dissection to prevent any strain on the optic chiasm during preparation. For students learning this approach, pulling or tearing the optic nerves during the dissection is one of the most reasons for poor outcome. Specifically, brains are dissected and placed for ~1 min in ice-cold carbogenated (95% O2, 5% CO2) modified (ACSF) containing less CaCl2 (1 mM) and more MgCl (5 mM) to inhibit synaptic transmission and reduce excitatory amino acid toxicity. After trimming the brain to a block of tissue containing the hypothalamus, a vibratome (e.g., Microslicer, DSK Model 1500E, Ted Pella, Redding, CA, USA) is used for making coronal slices with a thickness of 250–350 μm for patch-clamp and Ca2+ imaging experiments under visual control. Successful sectioning of SCN slices using the vibratome requires a firm attachment of the brain to the cutting block. This is a common problem for beginners. If the brain moves when the blade hits the tissue, you will face difficulties. Therefore, a thin, even layer of cyanoacrylate is applied to the bottom of the vibratome chamber and the trimmed brain is glued to the cutting block with the optic chiasm facing the blade (see Note 1). The right combination of forward speed of the knife, vibration frequency, and, if applicable, amplitude of oscillation is critical for cutting smoothly through the tissue. Brains from older mice or especially rats require slower forward speed and higher frequency oscillation of the blade. For >24 h MUA recordings of electrical activity from SCN neurons, a tissue chopper (McIlwain Tissue Chopper, The Mickle Laboratory Engineering, Guildford, UK) is preferred over a vibratome to minimize the time of preparation and produce 400–500 μm thick slices. Brain slices containing the SCN are then placed for at least 1 h before recording in our standard ACSF. This superfusate has a pH of 7.2–7.4 due to continuous gassing with carbogen while the osmolarity ranges between 290 and 300 mOsm. Patch-clamp and Ca2+ imaging experiments can be performed at room temperature (22–25
C), but slices need to be warmed up to 36
C for 30 min right after cutting to activate endopeptidases. For MUA recordings, slices are immediately transferred to a recording chamber that is continuously superfused (1.5 ml/min) with warm (35.5
C) oxygenated ACSF with added

Electrophysiology of the SCN

309

antibiotic (gentamycin, 40 mg/ml). SCN slices are mechanically stabilized with either a metal fork or a nylon grid, preventing movement but allowing for access for recording electrodes. While most acute slices will be used for a short period of time, the analysis of circadian rhythms requires longer recording times (at least 32 h). The cultured slice/organotypic explant would be more appropriate for longer recording times, but the acute slice offers many advantages when interested in photic entrainment and resetting mechanisms. To optimize the time that acute slices stay healthy in the recording dish, conditions must be as stable as possible (see Note 2). We have used several techniques to analyze the cellular or circuit properties of the SCN in an acute brain slice preparation. In the next section, we will briefly describe the techniques that we have found most useful for the analysis of the circadian circuit.

IR-Sensitive Camera

CCD Camera Controller

C

A

B D + 5 pA 50 ms

E

Slice in Chamber

0.5 nA

F

Infrared Filter

100 ms

20 mV

Halogen Lamp

200 ms

Fig. 1 Patch-Clamp recordings from SCN brain slices. (a) Upright-microscope equipped with infrareddifferential interference contrast (IR-DIC) and an IR-sensitive camera. Brain slice is placed in a recording chamber on the microscope stage. (b) Computer-controlled image processing showing IR DIC image of the mouse SCN. We have labeled the regions that we take to be dorsal or ventral in our experiments. The third ventricle (III) and optic chiasm (OC) are used as references. Most of our electrophysiological recording in the brain slice are made using this imaging technology to localize the recording site. (c) Recording of action potentials in cell-attached voltage-clamp mode with zero current injected. (d) Postsynaptic membrane currents. (e) Whole-cell voltage-dependent currents recorded as a response to a series of voltage steps. On the left, Na+ inward current (downward reflections) and transient K+ outward currents are visible. At the end of the pulses, the current consists mainly of steady state K+ current. (f) Current-clamp recording of action potentials

310

Stephan Michel et al.

3.2 Optical Approaches 3.2.1 IR-DIC Videomicroscopy

3.2.2 Ca2+ Imaging

CCD camera

Videomicroscopy allows for visualization of live, unstained cells in brain slice preparations [29, 30]. The brain slice rests in a perfusion chamber and is illuminated with IR light. Images are taken with differential interference contrast (DIC) optics and contrastenhanced video microscopy. Figure 1a shows a view of SCN neurons in a brain slice as seen with IR-DIC videomicroscopy. The optic chiasm, third ventricle, and SCN are clearly visible. At higher magnification, it is possible to distinguish the borders and major processes of cells from the surface to about 150 μm deep into a brain slice. This imaging technology is capable of targeting specific cells for electrophysiological analysis. The experimenter can clearly identify cells of the SCN and even distinguish different cell populations within the SCN. In addition, IR-DIC microscopy has improved the frequency of successful recordings and allows for the careful visual positioning of iontophoretic and stimulating electrodes. These are significant technological advantages for this type of study. Ca2+ imaging techniques allow for dynamic measurements of Ca2+ levels inside neurons and have been particularly important in understanding the relationship between membrane events and CCD Camera Controller

GABA

A

B

Monitor

C

fibre optics Slice in Chamber

20 nM

LED Lamp

Fig. 2 Calcium imaging to record cell responses to neurotransmitter agonist. (a) Fluorescent microscope using an LED lamp for excitation light coupled into the microscope light path by fiber optics. Images are recorded by a cooled CCD camera mounted at the microscope. Focal pressure application of neurotransmitter agonists— like GABA—can be performed through a quartz cannula (100 μm) (b) Computer-assisted imaging software controls light exposure, drug administration and image processing. Shown are Fura-2 AM-labeled SCN neurons which were imaged with a 40 water immersion objective. (c) Right panel shows GABA-induced Ca2+ transients, which indicate that in the SCN some cells respond with “excitatory” responses (red lines) and some with “inhibitory” responses (blue lines). The true nature of the responses has to be confirmed with electrophysiological techniques

Electrophysiology of the SCN

311

transcriptional regulation within the SCN. Furthermore, these techniques are useful in recording the neuronal response in a neuronal network to brief applications of neurotransmitters or other agents [31] (Fig. 2). To carry out Ca2+ imaging in the brain slice, we use a cooled charge-coupled device (CCD) camera (Princeton Instruments, Microview model 1317  1035 pixel format, Trenton, NJ, USA) that is added to the Olympus fixed stage microscope. Two different methods are used to load the dye. In the first approach, slices are incubated in 10 μM of the membrane-permeant acetoxymethyl (AM) form of the Ca2+-sensitive fluorescent dye Fura-2 (Fura-2 (AM); Invitrogen, Life Technologies, Carlsbad, CA, USA) at 37
C for 10 min (see Note 3). In a recent modification of this technique [32, 33], the SCN slice is exposed briefly (1 min) to stock solution of fura-2-AM (1 mM) and subsequently loading is continued for 1 h at room temperature using the diluted fura-2-AM in ACSF (10 μM). We used this modified loading protocol successfully in older rats (6 month) and old mice (2 years). The alternative approach uses membrane-impermeable fura2 (1 mM pentapotassium salt) which is loaded into cells via the whole-cell-recording patch pipette [33]. The fluorescence of Fura2 (AM) is excited alternatively at wavelengths of 357 nm and 380 nm by means of a high-speed wavelength switcher (1.2 ms switching capability, Sutter, Lambda DG-4, Novato, CA, USA). With the optics in many microscope objectives, there is a big difference in the transmission of light at 340 and 380 nm. This makes it difficult to bring fluorescence measurements from both 340 nm and 380 nm into the dynamic range of the camera. Accordingly, the dye can be excited with 357 nm light instead of 340 nm in these experiments. Image analysis software (MetaFluor, Universal Imaging, Molecular Devices, Sunnyvale, CA, USA) allows for the selection of several “regions of interest” within the field from which measurements of dynamic changes of Fura-2(AM) fluorescence intensity are taken. To minimize bleaching of the dye, the intensity of excitation light and sampling frequency should be kept as low as possible. Studies with Fura-2 (AM) are technically easy but have potential problems. Firstly, it is not possible to confidently resolve from which cell types (neurons vs. glia) the data are collected. Secondly, in practice, it appears to be more difficult to load cells in slices from older animals and, like with all AM-dyes applied via the bath, only cells located at the surface of the slice take up the dye. Loading dye via the patch pipette can solve these problems. However, this approach is much more laborious and it may be best to use this technique only when necessary to confirm a result obtained with the membrane-permeant form of this dye. Although Fura-2 is the standard dye we use for ratiometric Ca2+ measurements, there are some good arguments for employing other dyes. We have made

312

Stephan Michel et al.

some use of the visible light Ca2+ indicator Oregon Green BAPTA (Invitrogen, Life Technologies, Carlsbad, CA, USA), Fluo-2 (TEFLabs, Austin TX, USA) as well as CAL-520 AM (Abcam, Cambridge, UK or AAT Bioquest, Sunnyvale, CA, USA). We obtain good loading of SCN slices with both dyes and one should see less phototoxicity. The use of visible rather than UV excitation should produce less autofluorescence. Since the cells in the SCN are known to undergo daily rhythms in metabolism, the possibility of daily rhythms in autofluorescence must be considered although we have found that autofluorescence is responsible for less than 5% of the signal. Finally, there are good arguments for the use of 2-photon laser scanning microscopy (2PLSM). 2PLSM offers several important advantages compared to conventional fluorescent microscopy with the CCD camera. All forms of imaging in thick brain tissue are limited by light scatter and longer wavelength light scatters less than shorter wavelength light. For this reason, the ability of the 2PLSM to make use of long wavelength light offers significant advantages for the resolution of structures deep in a brain slice. Furthermore, in 2PLSM the indicator dye is excited in a nonlinear manner. The requirement for two near coincident photons to achieve excitation of the dye means that only focused light reaches the required intensities and that scattered light does not excite the dye. This nonlinear excitation results in a significant reduction in photodamage and bleaching. This reduction in the damage caused by the excitation light is likely to be a critical advantage for experiments that involve taking measurements of Ca2+ for extended periods of time. For the measurements using Fura-2, the free concentration of cytosolic Ca2+ can be calculated from the ratio (R) of fluorescence intensities at 357 and 380 nm, using the following equation: [Ca2 + ] ¼ Kd  Sf  (R  Rmin)/(Rmax  R) [34]. Value for Kd is set at 135 nM, while values for Rmin and Rmax are determined with both in vitro and in vivo calibration methods. Initially, the in vitro method is used to make estimate values. With this method, rectangle glass capillaries are filled with high Ca2+ (Fura-2 + 10 mM Ca2 + ), low Ca2+ (Fura-2 + 10 mM EGTA) and a control saline without Fura-2. The fluorescence (F) at 380 nm excitation of the low Ca2+ solution is imaged and the gain of the camera adjusted to maximize the signal. These camera settings will then remain fixed and background subtracted measurements made with 380 and 357 nm excitation of the three solutions: Rmin ¼ F357 nm in low Ca2+/ F380 in low Ca2+; Rmax ¼ F357 in high Ca2+/F380 in high Ca2+; Sf ¼ F380 in low Ca2+/F380 in high Ca2+. For the more accurate calibration methods [35], cells are loaded via the patch pipette using solutions inside the electrode similar to the normal internal solution but containing either no Ca2+ (20 mM EGTA) or 10 mM Ca2+ for Rmin and Rmax, respectively. To obtain estimates of the

Electrophysiology of the SCN

313

effective Kd, three different EGTA/Ca2+ solutions are used with calculated free Ca2+ of 111 nM (10 mM EGTA/3.5 mM Ca2+), 207 nM (10 mM EGTA/5 mM Ca2+), and 483 nM (10 mM EGTA/7 mM Ca2+). By calibrating in the slice preparation with solutions of ionic strength similar to those used for measurements, some of the uncertainties associated with calibration of Ca2+ indicators can be avoided.

3.3.1 Long-Term Multiunit and Single-Unit Extracellular Recordings

Extracellular electrical activity of SCN neurons in freshly prepared brain slice can be measured by extracellular glass, metal, or suction electrodes. Glass microelectrodes have been used to record the neuronal activity of single SCN neurons for short times (typically 20 MΩ) or tail currents which do not decay rapidly with a single exponential should not be used. It is important to be aware of these problems and carefully monitor the adequacy of space clamp conditions. Another technical concern is that while the whole-cell patch-clamp technique offers improved voltage control, it also dialyzes the inside of the cell. This can result in the loss of membrane currents that are, for example, highly dependent upon phosphorylation. One solution to this problem is, for comparison, to run some experiments with sharp-electrode intracellular recording techniques and without ion channel blockers. An alternative possibility is to use the “perforated patch” technique [26]. 3.3.4 Evoked EPSCs and Stimulation Techniques

Electrical stimulation can be used to induce local excitatory postsynaptic currents (EPSCs) that are mediated by ionotropic glutamate receptors [25, 43]. Local stimulation is applied with bipolar electrodes constructed from twisted Teflon-coated silver wires (0.2 mm diameter exposed at the tip, tip separation 0.2–0.5 mm). The electrode is placed 0.5–1.0 mm from the recording pipette. Constant current square pulses (50–1000 μA intensity, 10–100 μs duration, 1 pulse/4–5 s) are used to induce short latency, graded amplitude EPSCs. In some cases, the stimulation frequency can be increased to 100 Hz. This should induce temporally summated long-duration EPSCs which will likely have a larger NMDA component. The threshold for EPSC onset is determined and a series of intensities from just above threshold for the EPSC to maximum amplitude is used. Changes in EPSC characteristics are determined prior to and at selected intervals after drug administration. Local

318

Stephan Michel et al.

stimulation activates cut input and output fibers as well as local neurons. There is a component of the EPSC mediated by activation of GABAA receptors when the stimulus is applied in the SCN. To reduce the contribution of activation of GABAA receptors, the blocker gabazine can be added to the ACSF. To prevent action potentials, QX-314 (20 mM) can be used in the recording pipette. The bath application of specific glutamate receptor antagonists (e.g., AP5 or CNQX) can be used to produce EPSCs mediated mainly by activation of NMDA or AMPA/KA receptors. The optic nerve is stimulated with the use of a custom-made suction electrode; constant current pulses (biphasic square wave) of 0.1–1.0 mA intensity and 0.5–2.0 ms duration are applied at 30 s intervals. 3.4

In Vivo Methods

3.4.1 Animals

3.4.2 Experimental Setup

As discussed, it is critical to house the animals in a robust LD cycle for at least 2 weeks prior to the experiments. Mice are housed individually, and food and water are available ad libitum. Running wheels were often provided prior to the experiments, and running discs can be used as an alternative during the recordings (see Note 4). In vivo equipment for measuring SCN activity in freely moving mice requires a well-designed, dedicated system. The mouse is recorded within a Faraday cage and is connected to a cable but must be able to move in all directions (Fig. 4). The in vivo measurement system has been constructed in a way that it is causing the least possible tension to the mouse by neutralizing the weight of the cable and headstage. This ensures minimal impact on the behavior of the mouse. SCN neurons produce very small extracellular electrical signals, due to the extremely small size of the cells (~10 μm). The small signals are buffered (impedance correction or preamplified), and subsequently fed into a differential amplifier. The in vivo setup allows for stable recordings from an animal for a long time. This is important for recording long-lasting effects of changes in photoperiod [41] or exposure to constant light [5–7]. If the recordings are to be made on moving animals, the in vivo physiological system will require a swivel and balancing system. The swivel consists of a slip ring (Moog or Moflon) mounted with a set of small ball bearings in a light-weight housing (26 g custommade). The design should ensure very low friction, a low signal noise and a long life. In addition, a swivel balancing system is required. The swivel balancing system compensates the weight of the swivel with cable and follows the vertical movement of the mouse. With a damped long spring, the balancing system ensures a very smooth balance and minimal friction at the onset of the animal’s movement. It is based on a long spring with contra weights (custom-made).

Electrophysiology of the SCN

Output to amplifiers (4) Balancing system

319

A

B

(3) swivel

(2) pre-amplifiers/ cable connector

(1) Head connector

AB

Implanted electrodes

Faraday cage

Fig. 4 Schematic diagram of an in vivo SCN recording setup. The implanted electrodes (1) are attached to the head via a customized connector. The preamplifiers in the head connector (2) are custom-built in our lab and are designed to have extremely low noise (below 9 nV/√Hz at 1 kHz), low input current leakage (