Hemoglobin Disorders. Molecular Methods and Protocols [1st Edition.] 9781617372780, 1617372781

Citation preview

M E T H O D S I N M O L E C U L A R M E D I C I N E TM

Hemoglobin Disorders Molecular Methods and Protocols Edited by

Ronald L. Nagel, MD

Humana Press

X-ray Crystallography of Hemoglobins

1

1 X-ray Crystallography of Hemoglobins Martin K. Safo and Donald J. Abraham 1. Introduction X-ray crystallography has played a key role in understanding the relationship between protein structure and physiological function. In particular, X-ray analysis of hemoglobin (Hb) crystals has been pivotal in the formulation of basic theories concerning the behavior of allosteric proteins. Methemoglobin (MetHb) from horse was the first three-dimensional (3D) structure of liganded Hb to be solved (1–4). It was followed by crystallographic determination of the unliganded (deoxygenated) form nearly a decade later (5). The X-ray analyses provided 3D atomic resolution structures and confirmed that Hb was tetrameric, containing two subunit types (α and β), and one oxygen-binding heme group per subunit. John Kendrew (myoglobin) and Max Perutz (Hb) received the Nobel Prize for their pioneering work, being the first to determine the 3D structures of proteins, using X-ray crystallography. Since the crystallographic determination of these structures, there has been an almost exponential increase in the use of X-ray crystallography to determine the 3D structures of proteins, i.e., as evidenced by the history of structures deposited in the protein data bank. Comparison of the quaternary structures of liganded and deoxygenated horse Hb clearly showed significantly different conformational states. The Hb X-ray structures were the first to confirm the two-state allosteric theory put forward by Monod et al. (6), which is referred to as the MWC model. The liganded Hb conformation conformed to the MWC relaxed (R) state, while unliganded Hb conformation conformed to the MWC tense (T) state. The source of the tension in the T state was attributed to crosslinking salt bridges and hydrogen bonds between the subunits. The relaxed (R) state has only a few intersubunit hydrogen bonds and salt bridges. From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

1

2

Safo and Abraham

Muirhead and Greer (7) published the first structure of human adult deoxygenated hemoglobin (deoxyHbA). Several years later, Baldwin and Chothia (8,9) and Baldwin (9) published the structure of human adult carbonmonoxyhemoglobin (HbCOA), and Shaanan (10) published the structure of human adult oxyhemoglobin (oxyHbA). Interestingly, the structure of oxyHbA was delayed because of complications resulting from heme iron autoxidation. Subsequently, a new quaternary ligand-bound Hb structure known as R2 (11) or Y (12,13) provided another relaxed structure. R2 was proposed to be a low-energy intermediate in the T-to-R allosteric transition. However, further analysis has revealed that R2 is not an intermediate but, rather, another relaxed end-state structure (14). Quite recently, our laboratory discovered two more novel HbCO A relaxed structures (R3 and RR2); RR2 has a structural conformation between that of R and R2 (unpublished results). The quaternary structural difference between T and R3 is as large as that of T and R2. However, R2 and R3 have very different conformations. The quaternary difference is determined by superimposing the α1β1 subunit interfaces and calculating the rotation angle between the nonsuperimposed α2β2 dimers (8,9). The first 3D structures of horse Hb were solved using isomorphous replacement techniques (1–3,5). A number of published Hb structures also crystallize isomorphously, thus making it possible to use phases from the known isomorphous Hb structure for further structural analysis. The development of molecular replacement methods (15,16) for the solution of protein structures enabled routine structure solutions for nonisomorphous Hb crystals. When the structure horse Hb was determined, no computer refinement programs existed. Therefore, the atomic positions were refined visually against the electron density map. With isomorphous mutant crystals (17) or isomorphous crystals with bound allosteric effectors (18), simple electron density difference map calculations have been shown to be powerful tools in analyzing structural differences. Currently, all new protein structures are refined using modern, faster computing methods, such as CNS (19) and REFMAC (20). The crystal structures of more than 250 Hbs have been solved and published, including mutants and Hb cocrystalized with allosteric effector molecules. Selected examples of native and mutant Hbs including quaternary states, crystallization conditions, and unit cell descriptions are given in Tables 1–3. The structures of mutant Hbs provided the first concrete correlation between structural changes and disease states, while Hb cocrystallized with small effector molecules has advanced our understanding of the fundamental atomic-level interactions that regulate allosteric function of an important protein. The general methodologies for isolating, purifying, crystallizing and crystal mounting for data collection follow. The X-ray structure solution of Hb and variants is routine and employs the techniques discussed above: isomorphous

Name

Quater- Chemical nary state form

3

Crystallization condition

Unit cell characteristicsa a = 63.2, b = 83.5, c = 53.8 Å, β = 99.3°, SG = P21, AU = 1 tetramer a = 97.1, b = 99.3, c = 66.1 Å, SG = P21212, AU = 1 tetramer a = 63.2, b = 83.6, c = 53.9 Å, β = 99.2°, SG = P21, AU = 1 tetramer SG = P21, AU = 1 tetramer a = 53.7, b = 53.7, c = 193.0 Å, SG = P41212, AU = 1 dimer a = 53.7, b = 53.7, c = 193.8 Å, SG = P41212, AU = 1 dimer a = 97.5, b = 101.7, c = 61.1 Å, SG = P212121, AU = 1 tetramer a = 62.8, b = 62.8, c = 320.9 Å, SG = P43212, AU = 1 tetramer a = 61.5, b = 61.5, c = 176.3 Å, SG = P4122, AU = 1 dimer a = 65.5, b = 154.6, c = 55.3 Å, SG = P21212, AU = 1 tetramer a = 106.1, b = 86.2, c = 64.3 Å, SG = P212121, AU = 1 tetramer

DeoxyHbA

T

Normal

DeoxyHbA

T

Normal

RSR13-deoxy T HbA complexb T DeoxyHbFc OxyHbA R

Normal

2.2–2.8 M NH4 phosph/sulfate, pH 6.5 10–10.5% PEG 6000, 100 mM KCl, 10 mM K phosph, pH 7.0 2.5–2.9 M NH4 phosph/sulfate, pH 6.5

Fetal Normal

2.2–2.8 M NH4 phosph/sulfate, pH 6.5 2.25–2.75 M Na/K phosph, pH 6.7

HbCO A

R

Normal

2.25–2.75 M Na/K phosph, pH 6.7

HbCO A

R2

Normal

CO Gower II (α2ε2)d HbCO Ae

R2 R3

16% PEG 6000, 100 mM, Na cacodylate, pH 5.8 Embryonic 21% MME PEG 5000, 0.2 M TAPS-KOH, pH 8.5 Normal 2.34–2.66 M Na/K phosph, pH 6.4–6.7

HbCO Ae

RR2

Normal

2.34–2.66 M Na/K phosph, pH 6.4–6.7

CNMetHbAf

Y

Normal

16–17% PEG 8000, 0.1 M Tris, 0.12% BOG

X-ray Crystallography of Hemoglobins

Table 1 Crystallization Conditions and Structural Properties of Selected Human Hbs Resolution (Å)

Reference

1.7

25

2.15

26

1.85

27

2.5 2.1

28 10

2.7

8

1.7

11

2.9

29

2.65

Unpublished data Unpublished data 13

2.18 2.09

a SG

, space group; AU, and asymmetric unit. is an allosteric effector. c The authors of deoxyHbF did not provide the cell constants, however, the crystal is isomorphous to the high-salt deoxyHbA crystal (25). d The quaternary structure of carbonmonoxy embryonic Gower II Hb lies between that of R and R2 states, though closer to the R2 state. e Relaxed end-state structures (see text). f The quaternary structures of Y and R2 state Hbs are similar. b RSR13

3

4

Table 2 Crystallization Conditions and Structural Properties of Selected Natural Mutant Human Hbs Name

Quater- Chemical nary state form

4

Unit cell characteristicsa

a = 52.9, b = 185.7, c = 63.3 Å, β = 92.6°, SG = P21, AU = 2 tetramers a = 63.2, b = 83.6, c = 53.8 Å, β = 99.4°, SG = P21, AU = 1 tetramer a = 97.1, b = 99.3, c = 66.1 Å SG = P21212, AU = 1 tetramer a = 63.2, b = 83.6, c = 53.8 Å, β = 99.4°, SG = P21, AU = 1 tetramer a = 93.1, b = 93.1, c = 144.6 Å SG = P3221, AU = 1 tetramer a = 54.38, b = 54.38, c = 195.53 Å, SG = P41212, AU = 1 dimer SG = P21, AU = 1 tetramer SG = P21, AU = 1 tetramer

DeoxyHbA Sickle cell

T

Glu6βVal

33% PEG 8000, 5.5 mM citrate, pH 4.0–5.0

Catonsville

T

2.2–2.8 M NH4 Phosph/sulfate pH 6.5

Rothschild

T

Pro37α-GluThr38α Trp37βArg

Thionville

T

10–10.5% PEG 6000, 100 mM KCl, 10 mM K phosph, pH 7.0 2.2–2.8 M NH4 phosph/sulfate, pH 6.5

Cowtown

R

His146βLeu

2.25–2.75 M Na/K phosph, pH 6.7

Knossosb GrangeBlancheb Brocktonb Suresnesb Kansas

T T

Ala27βSer Ala27βVal

2.2–2.8 M NH4 phosph/sulfate, pH 6.5 2.2–2.8 M NH4 phosph/sulfate, pH 6.5

T T T

Ala138βPro Arg141αHis Asn102βThr

2.2–2.8 M NH4 phosph/sulfate, pH 6.8 2.2–2.8 M NH4 phosph/sulfate, pH 6.5 2.2–2.8 M NH4 phosph/sulfate, pH 6.5

SG = P21, AU = 1 tetramer SG = P21, AU = 1 tetramer a = 63.4, b = 83.6, c = 53.9 Å, β = 99.3 o, SG = P21, AU = 1 tetramer

2.05

24

1.7

30

2.0

26

1.5

31

3.0

12

2.3

32

2.5 2.5

33 33

3.0 3.5 3.4

34 35 36

Safo and Abraham

COYpsilanti Y

Val1αGlu AcetMet(-1)1α Asp99βTyr 2.25–2.30 M Na/K phosph, pH 6.7

a SG,

Resolution (Å) Reference

Crystallization condition

space group and; AU, asymmetric unit. authors of Hb Knossos, Grange-Blanche, Brockton, and Suresnes did not provide the cell constants, however, the crystals are isomorphous to the high-salt deoxyHbA crystal (25). b The

Quater- Chemical nary state form

Name Yα42H

5

Crystallization condition

Unit cell characteristicsa a = 62.4, b = 81.2, c = 53.3 Å, β = 99.65°, SG = P21, AU = 1 tetramer a = 63.3, b = 83.4, c = 53.8 Å, β = 99.5°, SG = P21, AU = 1 tetramer a = 54.3, b = 54.3, c = 194.1 Å SG = P41212, AU = 1 dimer a = 62.9, b = 81.3, c = 111.4 Å SG = P212121, AU = 1 tetramer a = 62.9, b = 82.0, c = 53.9 Å, β = 99.0°, SG = P21, AU = 1 tetramer a = 102.5, b = 115.2, c = 56.7 Å SG = P212121, AU = 1 tetramer a = 63.5, b = 83.2, c = 54.0 Å, β = 99.15°, SG = P21, AU = 1 tetramer a = 96.7, b = 98.7, c = 66.0 Å, SG = P21212, AU = 1 tetramer a = 63.2, b = 83.4, c = 53.8 Å, β = 99.4°, SG = P21, AU = 1 tetramer a = 63.2, b = 83.7, c = 53.8 Å, β = 99.4°, SG = P21, AU = 1 tetramer

T

Tyr42αHis

2.2–2.8 M NH4 phosph/sulfate, pH 6.5

rHb(α96Val→Trp) T

Val96αTrp

2.2–2.8 M NH4 phosph/sulfate, pH 6.5

rHb(α96Val→Trp) R

Val96αTrp

2.25–2.75 M Na/K phosph, pH 6.7

Deoxy-Hbβ6W

T

Glu6βTrp

Deoxy-rHb1.1

T

CNmet-rHb1.1

B

Deoxy-βV67T

T

Des-Arg141αHbA T

4–7 uL of 33 % PEG 8000, 5 uL of Na citrate, pH 4.8 Asn108βLys 2.2–2.8 M NH4 phosph/sulfate, pH 6.5 α1-Gly-α2 Asn108βLys 13 % PEG 3350, 10 mM KCN, α1-Gly-α2 150 mM NH4 acetate, pH 5.0 Val67βThr 2.2–2.8 M NH4 phosph/sulfate, pH 6.5

Bulltown

T

des-Arg141α 10–10.5 % PEG 6000, 100 mM KCl, 10 mM K phosph, pH 7.0 His146βGln 2.2–2.8 M NH4 phosph/sulfate, pH 6.5

Deoxy-βV1M

T

Val1βMet

a SG,

2.2–2.8 M NH4 phosph/sulfate, pH 6.5

X-ray Crystallography of Hemoglobins

Table 3 Crystallization Conditions and Structural Properties of Selected Artificial Mutant Human Hbs

Resolution (Å) Reference 1.8

38

1.9

39

2.5

39

2.0

40

2.0

41

2.6

41

2.2

42

2.1

43

2.6

44

1.8

45

space group; AU, asymmetric unit.

5

6

Safo and Abraham

replacement, difference electron density calculations, molecular replacement, and structure refinement (for details, see references). 2. Materials 2.1. Purification of Human Hb for Crystallization 1. HbA is purified from outdated human red blood cells (RBCs) unsuitable for transfusion (~500 mL). Sickle cell Hb (HbS) is purified from sickle cell blood, normally obtained from homozygote sickle cell patients who receive blood-exchange transfusions. To avoid clotting, blood samples are normally stored with about 1/10 vol of an anticoagulant agent, such as EDTA, heparin, or potassium citrate. 2. Buffer stock solution (5–10 L) containing 50 mM Tris buffer (pH 8.6) with EDTA: The solution is made by mixing 50 vol of 0.1 M Trizma base, 12.4 vol of 0.1 N Trizma hydrochloride, adjusting the volume to 100 mL with deionized water containing 4 g of EDTA (see Note 1). 3. Stock saline solutions (3 and 1 L) of 0.9% (9 g/L) and 1.0% NaCl (10 g/L), respectively. 4. DEAE sephacel and chromatography column equipment. 5. Cellulose dialysis tubes (Fisher Pittsburgh, PA). 6. Carbon monoxide gas cylinder (Matheson, Joliet, IL) (see Note 2). 7. NaCl, Na dithionite, and K2HPO4. 8. Three Erlenmeyers or side arm flasks (1 L).

2.2. Crystallization of Human Hb 1. Cyrstallization procedures will be described for deoxyHbA, deoxyHbS, and COHb A. These methods are also applicable to other HBs. HbA and HbS isolated and purified as described in Subheading 3.1.2. are used for all crystallization setups.

2.2.1. High-Salt Crystallization of T-State deoxyHbA 1. HbA solution (12 mL) (60 mg/mL or 6g%): Dilute the protein with deionized water if necessary to obtain the above concentration. 2. 3.6 M precipitant solution (50 mL) (pH 6.5): This is made by mixing 8 vol of 4 M (NH4)2SO4, 1.5 vol of 2 M (NH4)2HPO4, and 0.5 vol of 2 M (NH4)H2PO4. 3. Deionized water (100 mL). 4. Ten 8-mL sterile interior vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ). 5. Stoppered glass jar (Aldrich, St. Louis, MO). 6. Parafilm. 7. Pipets and pipet tips (100 and 1000 mL). 8. Three 15- to 25-mL beakers or volumetric flasks. 9. Graduated cylinders (10- and 50-mL). 10. Mixture of FeSO4 (2 g) and Na citrate (1.5 g). 11. A few grains of Na dithionite. 12. Test tube rack.

X-ray Crystallography of Hemoglobins

7

2.2.2. High-Salt Crystallization of R-State HbCO A 1. HbA solution (12 mL) (40 mg/mL or 4g%) in a 50-mL round-bottomed flask equipped with a stir bar and a greased stopcock adapter. 2. 3.4 M precipitant solution (40 mL) (pH 6.4): This is made by mixing 7 vol of 3.4 M NaH2PO4 and 5 vol of 3.4 M K2HPO4 (see Note 3). 3. Deionized water (100 mL). 4. Toluene (50 µL). 5. Ten 8-mL sterile interior vacutainer tubes (Becton Dickinson). 6. Stoppered glass jar (Aldrich). 7. Pipets and pipet tips (100 and 1000 mL). 8. A few grains of Na dithionite. 9. Carbon monoxide gas cylinder (Matheson) (see Note 2) and nitrogen gas cylinder. 10. Test tube rack. 11. Vacuum pump and rubber tubing.

2.2.3. Low-Salt Crystallization of T-State deoxyHbS 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

HbS solution (1.2 mL) (120 mg/mL or 12g%). 50% (w/v) polyethylene glycol (PEG) 6000 (12 mL). 0.2 M citrate buffer (1 mL), pH 4.0–5.0 (Hampton Research, Laguna Hills, CA). Deionized water (10 mL). Ten 3-mL sterile interior vacutainer tubes (Becton Dickinson). Parafilm. Stoppered glass jar (Aldrich). Pipets and pipet tips (100 and 1000 mL). Two 15- to 25-mL beakers. A few grains of Na dithionite.

2.3. Crystal Preparation and Mounting The methods described here are for deoxyHbA and COHb A, and are also applicable to other Hb cystals.

2.3.1. Room Temperature Data Collection 1. Vacutainer tube containing T- or R-state crystals. 2. Capillary sealant, such as epoxy or paraffin wax or any wax with a low melting point. 3. Disposable pipets and pipet rubber bulb. 4. Stainless steel blunt-end needles (Fisher). 5. Disposable syringes (3–5 mL) (Fisher Scientific). 6. Sterilized paper wicks (Hampton Research). 7. Thin-walled quartz or borosilicate capillaries (Charles Supper, Natick, MA), ranging in size from 0.1 to 1.2 mm. 8. Soldering iron. 9. Sharp tweezer.

8

Safo and Abraham

2.3.2. Cryogenic Temperature Data Collection 2.3.2.1. T-STATE DEOXYHBA CRYSTAL 1. 2. 3. 4. 5. 6. 7. 8.

Vacutainer tube containing T-state crystals. Glycerol (100 µL). Small Dewar flask with liquid nitrogen. Thin fiber loop with diameter slightly larger than longest crystal dimension (Hampton Research). Cryovial and cryovial tong (Hampton Research). Disposable pipets and pipet rubber bulb. Glass slides. A few grains of Na Dithionite.

2.3.2.2. R-STATE COHB CRYSTAL 1. Vacutainer tube containing R-state crystals. 2. Cryoprotectant solution made by mixing 60 µL of mother liquor and 5–8 µL of glycerol. 3. Thin fiber loop with diameter slightly larger than longest crystal dimension. 4. Disposable pipets and pipet rubber bulb. 5. Glass slides.

3. Methods 3.1. Purification of Human Hb for Crystallization About 90% of RBC content is made up of Hb, and in healthy human adults, HbA accounts for more than 90% of the human Hb protein, while other minor components, such as fetal HbF (~1%) and hemoglobin HbA2 (2 to 3%), make up the remainder. The method described here for isolating HbA and HbS from blood or RBCs, and further purification by ion-exchange chromatography, is a modified version of Perutz’s (21) protocol. This procedure, using appropriate buffer eluents, has also been used to separate other variant forms of human Hb and Hb from other species.

3.1.1. Purification of HbA 1. Place three Erlenmeyer or side-arm flasks in a walk-in refrigerator and chill to 4°C. 2. Centrifuge the RBCs at 600g for 20 min at 4°C. 3. Gently aspirate the supernatant solution (debris, plasma, and excess serum) from the centrifuge bottles and discard. 4. Wash the RBCs three times with an excess volume of 0.9% NaCl, and then once with 1.0% NaCl, each time centrifuging and discarding the supernatant solution. 5. Pool the RBCs into a chilled flask and lyse the cells by adding 1 to 2 vol of 50 mM Tris buffer, pH 8.6 (containing EDTA) (see Note 4). 6. Allow the mixture to stand on ice for 30 min with occasional gentle stirring. 7. Centrifuge the Hb solution at 10,000g for 2 h at 4°C.

X-ray Crystallography of Hemoglobins

9

8. Pool the supernatant Hb solution, which is free of cell debris, into a chilled flask, and slowly add NaCl (40–60 mg/mL of Hb solution) while stirring the solution. 9. Centrifuge the Hb solution at 10,000g for 1 to 2 h at 4°C to remove any remaining cell stroma. 10. Pool the clear supernatant Hb solution into a chilled flask and discard the “syrupy” pellet. 11. Dialyze the Hb solution against 50 mM Tris buffer, pH 8.6 (containing EDTA), at 4°C to remove NaCl or other low molecular weight impurities (see Note 5). 12. Further purify the dialyzed Hb by ion-exchange chromatography using DEAE sephacel to separate the HbA from other Hb components (see Note 6): a. Equilibrate the resin with 50 mM Tris buffer, pH 8.6. b. Run the Hb solution through the column with 50 mM Tris buffer, pH 8.6 (containing EDTA), to allow the various Hb bands to separate. HbA2 (light band color) elutes first, followed by HbA (dark band color). The HbA fractions can be examined for purity by electrophoresis and only pure fractions (dark band) pooled together. 13. Concentrate the pooled fractions (40–100 mg/mL) with an Amicon stirred cell (Model 402) to a final HbA concentration of about 80–120 mg/mL (see Note 7). 14. Store the concentrated HbA, which is essentially the oxygenated form, at –80°C or freeze in liquid nitrogen. Hb stored at this temperature can remain suitable for crystal growth experiments for several years.

3.1.2. Purification of HbS HbS from homozygous sickle cell blood is isolated and dialyzed as described for HbA in Subheading 3.1.1. (steps 1–11). The HbS solution is further purified on a DEAE sephacel ion-exchange column using a buffer gradient of 50 mM Tris buffer, pH 8.6 (containing EDTA), and 50 mM Tris buffer, pH 8.4 (containing EDTA) (see Note 1). 1. Elute first HbA2 Tris buffer at pH 8.6, then HbS at pH 8.4. 2. Concentrate the pure HbS, identified by electrophoresis and store as indicated for HbA in Subheading 3.1.1. (steps 13 and 14).

3.2. Crystallization of Human Hb DeoxyHbA crystallizes from either high-salt or low-salt precipitants (7,21). The ligand-bound R-state Hbs, such as oxyHbA, HbCO A, and MetHbA; generally crystallize under high-salt conditions (8–10,21), while the ligand-bound R2- or Y-state HbAs also crystallize mainly under low-salt conditions (11,13). The most common approach to crystallizing Hb is the Perutz’s (21) batch method. Alternatively, the vapor diffusion method of hanging or sitting drop (22) is used, especially when only a small amount of protein is available. Here, detailed crystallization is described for both T- and R-state human HbA and

10

Safo and Abraham

Table 4 High-Salt Crystallization of deoxyHbA Tube

3.6 M NH4 phosph/ sulfate (mL)

Deionized H2O (mL)

0.5 M Fe citrate (mL)

1 2 3 4 5 6 7 8 9 10

4.90 4.80 4.70 4.60 4.50 4.40 4.30 4.20 4.10 4.00

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90

0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

6g% deoxy- Final salt HbA (mL) conc. (M) 1 1 1 1 1 1 1 1 1 1

2.94 2.88 2.82 2.76 2.70 2.64 2.58 2.52 2.46 2.40

includes the high-salt crystallization of deoxyHbA and HbCO A and the lowsalt crystallization of deoxyHbS. The crystallization methods described are modified batch methods by Perutz (21) and Wishner et al. (23) and can also be applied to Hb mutants and Hb from other species. See Notes 8 and 9 for important precautions regarding setting up T- and R-state crystals, respectively.

3.2.1. High-Salt Crystallization of T-State deoxyHbA 1. The materials in Subheading 2.2.1., with the exception of the stoppered glass jar and the parafilm, are put in an antechamber of a glove box. The vacutainer tubes should be unstoppered, labeled as shown in Table 4, and arranged on a test tube rack. All containers, including those with solvents, should be left open. 2. Alternately evacuate and fill the antechamber with nitrogen while stirring the HbA solution for 10–20 min to obtain completely deoxyHbA, water, and precipitant solutions (see Note 10). 3. Purge the anaerobic chamber of the glove box with nitrogen to ensure a complete anaerobic condition. 4. Transfer all materials from the antechamber to the anaerobic chamber. 5. Add 25 mL of deionized water to the FeSO4 and Na citrate mixture and shake for about 30 s. 6. Allow the solution to settle and decant. Use the supernatant (Fe citrate) solution for all experiments (see Note 11). 7. Measure the volume of precipitant solution with a graduated cylinder, and add water to restore to the original volume of 50 mL (3.6 M), if necessary. 8. Measure the volume of deoxyHbA solution with a graduated cylinder, and add water to restore to the original volume of 12 mL (60 mg/mL), if necessary. 9. Add a few grains of Na dithionite (or ~2 mM) to the deoxyHbA solution to reduce any ferric heme that may be present.

X-ray Crystallography of Hemoglobins

11

Table 5 High-Salt Crystallization of HbCO Aa Tube

3.4 M Na/K phosph (mL)

4g% HbCO A (mL)

Final salt conc. (M)

1 2 3 4 5 6 7 8 9 10

3.80 3.60 3.40 3.20 3.00 2.80 2.60 2.40 2.20 2.00

1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0

2.69 2.66 2.63 2.59 2.55 2.51 2.46 2.40 2.34 2.27

aA

drop or two of toluene is added to each tube.

10. Measure the precipitant solution and water and add to the vacutainer tubes as indicated in Table 4. 11. Measure 1- and 0.1-mL aliquots of deoxyHbA and Fe citrate, respectively, and add to each vacutainer tube. 12. Stopper each vacutainer tube, and tilt at least twice to mix the solution. 13. Remove all the materials from the glove box and wrap parafilm around the stopper of each vacutainer tube. 14. Store the sealed vacutainer tubes in greased, stoppered glass jars filled with nitrogen. Crystals normally appear within 3–10 d and vary in size from microscopic to as large as 8 mm in any direction. The crystals belong to space group P21 with approximate unit cell constants of a = 63 Å, b = 83 Å, c = 53 Å, and β = 99°.

3.2.2. High-Salt Crystallization of R-State HbCO A 1. Add a few grains of Na dithionite to 12 mL of HbA (40 mg/mL) in a roundbottomed flask (three to five times the size of the volume of the HbA solution) fitted with a stopcock adapter and connected to both a vacuum pump and a nitrogen gas source with rubber tubing. 2. Alternately evacuate and flush with nitrogen for about 10 min. 3. Connect a CO source to a disposable pipet with rubber tubing (see Note 2). 4. Open the flask containing the deoxyHbA solution, and quickly bubble CO through the solution to make the HbCO A derivative. 5. Reconstitute the volume to 12 mL (40 mg/mL) with CO-purged deionized water. 6. Bubble CO through the precipitant solution. 7. Measure the precipitant solution and add to the vacutainer tubes as indicated in Table 5. 8. Measure 1-mL aliquots of HbCO A and add to each vacutainer tube.

12

Safo and Abraham

Table 6 Low-Salt Crystallization of deoxyHbS Tube

50% PEG 6000 (mL)

Deionize water (mL)

0.2 M Citrate (mL)

1 2 3 4 5 6 7 8 9 10

1.5 1.4 1.3 1.2 1.1 1.0 0.9 0.8 0.7 0.6

0.15 0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 1.05

0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05

12g% deoxyHbS (mL) 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

9. Add a drop or two of toluene to each vacutainer tube (see Note 12). 10. Slowly bubble CO through each vacutainer tube, stopper, and tilt at least twice to mix the solution. 11. Seal the vacutainer tubes with rubber stoppers and store in greased, stoppered glass jars filled with nitrogen to minimize formation of MetHbA. Crystals normally appear within 3–10 d. The crystals are octahedral and belong to space group P41212, with approximate unit cell constants of a = 53 Å, b = 53 Å, and c = 193 Å. The method described to crystallize HbCO A is applicable to both oxyHbA and MetHbA (see Note 13).

3.2.3. Low-Salt Crystallization of T-State deoxyHbS 1. Place all materials (except stoppered glass jar and parafilm) in the antechamber of the glove box. The vacutainers should be unstoppered and labeled as shown in Table 6. All containers, including those of solvents, should be left opened (see Note 14). 2. Deoxygenate the HbS and other solutions (5–10 min) in the antechamber of the glove box. 3. Purge the anaerobic chamber and transfer all materials from the antechamber into the anaerobic chamber. 4. Add deionized water to restore the volume of the HbS to 1.2 mL, if necessary. 5. Add a few grains of Na dithionite (or ~2 mM) to the HbS solution. 6. Add deionized water to restore the volume of the precipitant solution to 12 mL, if necessary. 7. Measure the precipitant solution and deionized water and add to the vacutainer tubes as shown in Table 6. 8. Measure 0.1 mL-aliquots of deoxyHbS and add to each vacutainer tube. 9. Stopper each vacutainer tube and tilt at least twice to mix the solution.

X-ray Crystallography of Hemoglobins

13

10. Store the vacutainer tubes and contents as described in Subheading 3.2.1. (steps 13 and 14). Crystals grown by this method are twinned (23,24) and must be separated before X-ray data can be obtained. Final crystals have the symmetry of the monoclinic space group P21, with approximate cell constants of a = 53 Å, b = 184 Å, c = 63 Å, and β = 93° (see Note 15).

3.3. Crystal Preparation and Mounting Hb crystals, like most other protein crystals, are fragile because of their high solvent content and should be handled with care. For room temperature data collection, Hb crystals are mounted and sealed in a thin-walled glass capillary about twice the size of the crystal. For cryogenic data collection, crystals are mounted in a thin fiber loop with a layer of suitable cryoprotectant around the crystal.

3.3.1. Room Temperature Data Collection T-state crystals are prepared and mounted in the glove box, while R-state crystals are mounted outside the glove box. However, to minimize autoxidation, mount R-state oxyHbA crystals as described for T-state crystals. 1. Select at least two 8-cm-long capillaries, and, using a soldering iron, melt a ring of wax close to the middle of the capillary. 2. Use a sharp tweezer to cut the bottom part of the capillary, just below the ring of wax. The top part of the capillary with the wide mouth is retained. Seal the cut bottom (with the ring of wax) with melted wax or epoxy (see Note 16). 3. Using a microscope, select a few good crystals by marking outside the vacutainer tube where those crystals are. 4. For R-state crystals, proceed to step 9. 5. For T-state crystals, place the materials in Subheading 2.3.1., in addition to the prepared capillaries, in the antechamber of the glove box. 6. Alternately evacuate and fill the antechamber with nitrogen for 5–10 min. 7. Transfer all the materials to a nitrogen-purged anaerobic chamber (see Note 17). 8. With a blunt-end needle, introduce a small amount of mother liquor from the vacutainer tube into the upper third of the capillary (all the way to the top). 9. Using a disposable pipet with a rubber bulb, suck a suitable marked crystal up onto the solution in the capillary. Allow the crystal to flow down to the air space. If the crystal is less dense than the mother liquor, invert the capillary to allow the crystal to flow to the air space. 10. Carefully push the crystal with a thin fiber or the blunt end of the needle into the air space. 11. Remove the solution from the capillary with a syringe and needle. 12. Carefully dry excess liquid from the crystal with a filter paper strip, a smaller cut capillary, or even the tip of the blunt-end needle. Leave a thin film of mother liquor between the crystal and the capillary wall (see Note 18).

14

Safo and Abraham

13. Reintroduce a small amount of mother liquor into the capillary, about 5 mm from the crystal (~5-mm-long liquid). Do not fill all the way to the top (see Note 19). 14. Close the capillary with melted wax or epoxy.

3.3.2. Cryogenic Temperature Data Collection T-state crystals are prepared and mounted in the glove box, and R-state crystals are mounted outside the glove box. Slightly different procedures are used, so a protocol for each is given next. 3.3.2.1. T-STATE DEOXYHBA CRYSTAL 1. Place the materials in Subheading 2.3.2.1. in the anaerobic glove box as already described in Subheading 3.3.1. (steps 5–7). 2. Submerge the cryovial in the Dewar liquid nitrogen using the cryovial tong. 3. Prepare cryoprotectant solution by mixing 50 µL of mother liquor, 10–16 µL of glycerol, and a few grains of Na dithionite (see Note 20). 4. Pick up a crystal with the disposable pipet, and place it into 5 µL of cryoprotectant solution on a glass slide for about 30 s. 5. Transfer the crystal to another 5 µL of cryoprotectant solution for another 30 s. 6. Use a fiber loop to scoop the crystal. 7. Plunge the loop containing the crystal and the drop of cryoprotectant directly into the cryovial which is submerged in the liquid nitrogen. 8. Take the closed cryovial out of the glove box and mount the crystal on the goniometer head in the cold nitrogen gas stream (see Note 21).

3.3.2.2. R-STATE COHB A CRYSTAL 1. Pick up a crystal with a disposable pipet, and introduce it into 5 µL of cryoprotectant solution on a glass slide for about 30 s. 2. Transfer the crystal to another 5 µL of cryoprotectant solution for another 30 s. 3. Scoop up the crystal, which has a protective cover of cryoprotectant liquid, with a fiber loop. 4. Place the fiber loop on the goniometer head in the cold nitrogen gas stream.

4. Notes 1. For HbS purification, prepare an additional 3–5 L of buffer stock solution containing 50 mM Tris buffer (pH 8.4) with EDTA. The solution is made by mixing 50 vol of 0.1 M Trizma base and 17.2 vol of 0.1 N Trizma hydrochloride, and adjusting the volume to 100 mL with deionized water containing 4 g of EDTA. 2. CO should be handled with great care; it is extremely toxic. All experiments involving CO should be done in a fume hood in a well-ventilated room. 3. Alternatively, a precipitant solution consisting of equal volumes of 3.4 M NaH2PO4 and 3.4 M K2HPO4 (pH 6.7) may be used and 0.2 mL of distilled water added to each tube. 4. EDTA helps prevent oxidation of ferrous heme to ferric heme by chelating any heavy metals that may act as catalysts for the autoxidation process. The final

X-ray Crystallography of Hemoglobins

5.

6. 7.

8.

9.

10.

15

concentration of the purified HbA will depend on the amount of buffer added to lyse the cell. Strips of standard cellulose dialysis tubing that have been washed three or four times and boiled for 10 min in deionized water are used for the dialysis. This is done to remove traces of impure compounds that may contaminate the HbA. The dialyzing buffer should be 50- to 200-fold of the HbA volume and should be continuously stirred overnight. If possible, the buffer should be changed every 2 to 3 h. Alternatively, HbA is dialyzed with 10 mM phosphate buffer, pH 7.0. The same type of buffer is then used to purify the HbA, as described in the text, using G25 Sephadex (fine) column. Alternatively, HbA is concentrated by ultrafiltration through an Mr 10,000 pellicon cassette. The concentration of HbA can be determined using the Perutz (21) procedure. The concentration is measured by taking 1 mL of HbA solution and diluting it with 19 mL of deionized water and 80 mL of 0.07 M K2HPO4. Na dithionite powder (0.2 g) is then added to the solution to generate the fully reduced deoxyhemoglobin derivative. CO is then bubbled through the solution to produce the COHb A derivative. The extinction coefficient is measured at 540 nm, and the concentration of HbA is calculated by dividing the optical density by 8.03. All crystallization steps for deoxyHbA are performed under rigorous anaerobic conditions in a nitrogen atmosphere glove box. It is critical that all crystallization solvents be purged of oxygen and stored under nitrogen. These precautions are necessary to prevent formation of oxyHbA or MetHbA. Human R-state COHb A , oxyHbA, and MetHbA crystallize isomorphously, and the corresponding structures are very similar. OxyHbA is very susceptible to autoxidation, which leads to formation of MetHbA during crystallization and data collection. To slow autoxidation, EDTA (1 mM) is added to the precipitating agents to chelate traces of heavy metals that catalyze the autoxidation process. Autoxidation of oxyHbA proceeds very rapidly when deoxyHbA is present in the solution; therefore, oxygen should be bubbled through the HbA solution to completely oxygenate all the HbA. In addition, crystallization should be performed at a low temperature, preferably 4°C, to slow down autoxidation. Even though HbCO A is fairly stable for a long period, the presence of oxygen leads to gradual oxidation of the ferrous heme. Therefore, crystallization of HbCO A should be under a CO atmosphere to avoid possible oxidation of the heme. All solutions for HbCO A crystallization should be purged with CO before use. A simple glove bag or Plexiglas box with gloves can be substituted for a more expensive glove box. If a glove bag or Plexiglas box is used, the HbA solution has to be deoxygenated outside the glove box. The HbA solution is put in a roundbottomed flask (three to five times the size of the volume of the HbA solution) and then connected by rubber tubing to both a vacuum pump and a nitrogen gas source with a glass stopcock adapter. The HbA is alternately evacuated and flushed with nitrogen for 30–60 min to obtain a deoxyHbA solution. (For smaller volume, the deoxygenation time is decreased.) A larger flask prevents boiling

16

11.

12.

13.

14. 15.

16. 17.

18.

Safo and Abraham HbA solution from getting into the vacuum line during the evacuation cycle. Additionally, to avoid undue boiling and splashing of the HbA, the flask containing the HbA solution may be cooled briefly in an ice bath before evacuation. Next, all materials are put into the glove bag or Plexiglas box. With the exception of the deoxyHbA solution, all other solution-containing flasks (precipitant, water, and buffer) should be left open. Once all the materials are put in the glove bag or Plexiglas box, it is then purged continuously with nitrogen for at least 40 min before the flask containing the HbA solution is opened. If the chamber is not airtight, it should be purged continuously with nitrogen during the crystallization experiments (Subheading 3.2.1., steps 5–14). Fe Citrate solution is prepared in situ from FeSO4 and Na citrate in the glove box and used fresh because the compound is unstable and easily oxidizes to ferric citrate. Fe citrate is a mild reducing agent and helps prevent oxidation of the iron; it also acts as an antimicrobial agent to prevent growth of bacteria and fungi. Toluene, like similar organic solvents, reduces the effective electrostatic shielding between the macromolecules by decreasing the electrostatic properties of the precipitating solutions. This facilitates increased contact between the macromolecules and serves to induce crystallization. The presence of toluene is also effective in preventing microbial growth. Recently, we have discovered two new crystal forms of HbCO A (R3 and RR2; see Table 1) that grows under the same crystallization conditions. One crystal form is rectangular and needle-like and belongs to the space group P4122. The other crystal form, which is also needle-like, belongs to the space group P212121. Alternatively, the HbA is deoxygenated outside the glove box as indicated above. For a small quantity of solution, the deoxygenation time is reduced accordingly (see Note 10, and continue from Subheading 3.2.3., steps 4–10). Crystals must be transferred to a stabilizing solution made of glutaraldehyde, which strengthens the crystals before cutting. Glutaraldehyde stabilizes the crystals by crosslinking the subunits. Soak the crystals for 1 d in a mixture of 35% (v/v) PEG stock solution, 20% (v/v) 0.2 M citrate buffer (pH 5.6), 45% (v/v) of 2% Drabkin’s buffer, and 10 mM of Na dithionite. The temperature of the solution is subsequently lowered to 3°C, and glutaraldehyde solution (50% [w/v]) is then added. The mixture is allowed to stand overnight at 3°C. Without the wax, the capillary may shatter when cut. If a glove bag or Plexiglas box is used, make sure that all necessary materials are put in the chamber and then purged continuously with nitrogen for at least 40 min before the vacutainer tube containing the crystals is opened. If the chamber is not airtight, it should be purged continuously with nitrogen during the experiments. A large amount of mother liquor around the crystal may decrease the resolution and increase mosaicity and background noise. The crystal can also move freely or slip. While making sure that as much liquid as possible is removed, do not completely dry the crystal. Excess drying will dehydrate the crystal, which may result in cracking, increased mosaicity, poor diffraction, disorder, and a large reduction in cell volume.

X-ray Crystallography of Hemoglobins

17

19. Mother liquor in the capillary ensures that the crystal is kept in the saturated vapor of the mother liquor during room temperature data collection to prevent drying. 20. Paraffin oil (Hampton Research) can also be used as a cryoprotectant. After putting the crystal in the paraffin oil, make sure that all excess mother liquor in the paraffin oil drop is removed by passing the crystal back and forth in the paraffin oil. The drop should form a perfectly clear glass under the cold stream. White patches may lead to reduction in resolution and increase mosaicity. 21. Simple freezing of the crystal will result in the formation of ice in the interior of the crystal and will render it useless. The cryoprotectant forms a noncrystalline glass, which protects the crystal from freeze shock.

References 1. Perutz, M. F., Rossmann, M. G., Cullis, A. F., Muirhead, H., Will, G., and North, A. C. T. (1960) Structure of haemoglobin. A three-dimensional Fourier synthesis at 5.5Å resolution obtained by x-ray analysis. Nature 185, 416–422. 2. Perutz, M. F., Muirhead, H., Cox, J. M., Goaman, L. C., Mathews, F. S., McGandy, E. L., and Webb, L. E. (1968) Three-dimensional Fourier synthesis of horse oxyhaemoglobin at 2.8 Å resolution: (1) x-ray analysis. Nature 219, 29–32. 3. Perutz, M. F., Muirhead, H., Cox, J. M., and Goaman, L. C. (1968) Three-dimensional Fourier synthesis of horse oxyhaemoglobin at 2.8 Å resolution: the atomic model. Nature 219, 131–139. 4. Ladner, R. C., Heidner, E. J., and Perutz, M. F. (1977) The structure of horse methaemoglobin at 2.0 Å resolution. J. Mol. Biol. 114, 385–414. 5. Bolton, W. and Perutz, M. F. (1970) The three dimensional Fourier synthesis of horse deoxyhaemoglobin at 2.8 Å resolution. Nature 228, 551, 552. 6. Monod, J., Wyman J., and Changeux J.-P. (1965) On the nature of allosteric transitions: a plausible model. J. Mol. Biol. 12, 88–118. 7. Muirhead, H. and Greer, J. (1970) Three-dimensional Fourier synthesis of human deoxyhaemoglobin at 3.5 Angstrom units. Nature 228, 516–519. 8. Baldwin, J. and Chothia, C. (1979) Haemoglobin: the structural changes related to ligand binding and its allosteric mechanism. J. Mol. Biol. 129, 175–220. 9. Baldwin, J. (1980) The structure of human carbonmonoxy haemoglobin at 2.7 Å resolution. J. Mol. Biol. 136, 103–128. 10. Shaanan, B. (1993) Structure of oxyhaemoglobin at 2.1 Å resolution. J. Mol. Biol. 171, 31–59. 11. Silva, M. M., Rogers, P. H., and Arnone, A. (1992) A third quaternary structure of human Hb at 1.7 Å resolution. J. Biol. Chem. 267, 17248–17256. 12. Smith, F. R., Lattman, E. E., and Carter, C. W. Jr. (1991) The mutation β99 AspTyr stabilizes a new composite quaternary state of human Hb. Proteins 10, 81–91. 13. Smith, F. R. and Simmons, K. C. (1994) Cyanomet human Hb crystallized under physiological condition exhibits the Y quaternary structure. Proteins 18, 295–300. 14. Janin, J.,and Wodak, S. J. (1993) The quaternary structure of carbonmonoxy Hb Ypsilanti. Proteins 15, 1–4. 15. Rossmann, M. G. and Hodgkin, D. C. (1972) in The Molecular Replacement Method (Rossmann, M. G., ed.), Gordon & Breach, New York, pp. 36–38.

18

Safo and Abraham

16. Navaza J. (1994) AMoRe: an automated package for molecular replacement. Acta Crystallogr. D50, 157–163. 17. Luisi, B. F. and Nagai, K. (1986) Crystallographic analysis of mutant human haemoglobins made in Escherichia coli. Nature 320, 555, 556. 18. Wireko, F. C., Kellogg, G. E., and Abraham, D. J. (1992) Allosteric modifiers of hemoglobin. 2. Crystallographic determined binding sites and hydrophobic binding/interaction analysis of novel hemoglobin oxygen effectors. J. Med. Chem. 34, 758–767. 19. Brunger, A. T., Adams, P. D., Clore, G. M., et al. (1998) Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. D54, 905–921. 20. Murshudov, G., Vagin, A., and Dodson, E. (1997) Application of maximum likelihood methods for macromolecular refinement. Acta Crystallogr. D53, 240–255. 21. Perutz, M. F. (1968) Preparation of haemoglobin crystals. J. Crystal Growth 2, 54–56. 22. McPherson, A. (1982) Preparation and Analysis of Protein Crystals (McPherson, A., ed.), John Wiley & Sons, New York. 23. Wishner, B. C., Ward, K. B., Lattman, E. E., and Love, W. E. (1975) Crystal structure of sickle-cell deoxyHb at 5 Å resolution. J. Mol. Biol. 98, 179–194. 24. Harrington, D. J., Adachi, K., and Royer, W. E. Jr. (1997) The high resolution crystal structure of DeoxyHb S. J. Mol. Biol. 272, 398–407. 25. Fermi, G., Perutz, M. F., Shaanan, B., and Fourme, R. (1984) The crystal structure of human deoxyHb at 1.7 Å resolution. J. Mol. Biol. 175, 159–174. 26. Kavanaugh, J. S., Rogers, P. H., Case, D. A., and Arnone, A. (1992) High-resolution X-ray study of deoxyhemoglobin Rothschild 37β Trp ∏ Arg: a mutation that creates an intersubunit chloride-binding site. Biochemistry 31, 4111–4121. 27. Safo, M. K., Moure, C. M., Burnett, J., Joshi, G. S., and Abraham, D. J. (2001) High resolution crystal structure of deoxy T-state hemoglobin complexed with a potent allosteric effector. Protein Science 10, 951–957. 28. Frier, J. A., and Perutz, M. F. (1977) Structure of human foetal deoxyhaemoglobin. J. Mol. Biol. 112, 97–112. 29. Sutherland-Smith, A. J., Baker, H. M., Hofmann, O. M., Brittain, T., and Baker, E. D. (1998) Crystal structure of a human embryonic haemoglobin: the carbonmonoxy form of Gower II (α2ε2) haemoglobin at 2.9 Å resolution. J. Mol. Biol. 280, 475–484. 30. Kavanaugh, J. S., Moo-Penn, W. F., and Arnone, A. (1993) Accommodation of insertions in helices: the mutation in hemoglobin Catonsville (Pro 37α-Glu-Thr 38α) generates a 3(10) → α bulge. Biochemistry 32, 2509–2513. 31. Vasseur, C., Blouquit, Y., Kister, J., Prome, D., Kavanaugh, J. S., Rogers, P. H., Guillemin, C., Arnone, A., Galacterose, F., Poyart, C., Rosa, J., and Wajcman, H. (1992) Hemoglobin Thionville: An alpha-chain variant with a substitution of a glutamate for valine at NA-1 and having an acetylated methionine NH2 terminus. J. Biol. Chem. 267, 12,682–12,691.

X-ray Crystallography of Hemoglobins

19

32. Derewenda, Z., Dodson, G., Emsley, P., Harris, D., Nagai, K., Perutz, M., and Reynaud, J.-P. (1990) Stereochemistry of carbon monoxide binding to normal and Cowtown haemoglobins. J. Mol. Biol. 211, 515–519. 33. Huang, Y., Pagnier, J., Magne, P., Bakloute, F., Kister, J., Delaunay, J., Poyart, C., Fermi, G., and Perutz, M. F. (1990) Structure and function of hemoglobin variants at an internal hydrophobic site: consequences of mutation at the beta 27 (B9) position. Biochemistry 29, 7020–7023. 34. Moo-Penn, W. F., Jue, D. L., Johnson, M. H., Olsen, K. W., Shih, D., Jones, R. T., Lux, S. E., Rodgers, P., and Arnone, A. (1988) Hemoglobin Brockton [β138(H16) Ala → Pro]: an unstable variant near the C-terminus of the b-subunits with normal oxygen-binding properties. Biochemistry 27, 7614–7619. 35. Poyart, C., Bursaux, E., Arnone, A., Bonaventura, J., and Bonaventura, C. (1980) Structural and functional studies of hemoglobin Suresnes (Arg 141α2 → His α2): consequences of disrupting an oxygen-linked anion-binding site. J. Biol. Chem. 255, 9465–9473. 36. Anderson, N. L. (1975) Structures of deoxy and carbonmonoxy haemoglobin Kansas in the deoxy quaternary conformation. J. Mol. Biol. 94, 33–49. 37. Tame, J. R. H. and Vallone, B. (2000) The structures of deoxy human haemoglobin and the mutant Hb Tyra42His at 120 K. Acta Crystallogr. D56, 805–811. 38. Puius, Y. A., Zou, M., Ho, N. T., Ho, C., and Almo, S. C. (1998) Novel watermediated hydrogen bonds as the structural basis for the low oxygen affinity of the blood substitute candidate rHb(α96Val → Trp). Biochemistry 37, 9258–9265. 39. Harrington, D. J., Adachi, K., and Royer, W. E. Jr. (1997) Crystal structure of deoxy-human hemoglobin Gluβ6 → Trp. Implication for the structure and formation of the sickle cell fiber. J. Biol. Chem. 273, 32,690–32,696. 40. Kroeger, K. S. and Kundrot, C. E. (1997) Structures of Hb-based blood substitute: Insights into the function of allosteric proteins. Structure 5, 227–237. 41. Pechik, I., Ji, C., Fidelis, K., Karavitis, M., Moult, J., Brinigar, W. S., Fronticelli, C., and Gilliland, G. L. (1996) Crystallographic, molecular modeling, and biophysical characterization of the Valineβ67 (E11) → Threonine variant of hemoglobin. Biochemistry 35, 1935–1945. 42. Kavanaugh, J. S., Chafin, D. R., Arnone, A., Mozzarelli, A., Rivetti, C., Rossi, G. L., Kwiatkowski, L. D., and Noble, R. W. (1995) Structure and oxygen affinity of crystalline desArg141 alpha human hemoglobin A in the T state. J. Mol. Biol. 248, 1136–1150. 43. Shih, D. T. B., Luisi, B. F., Miyazaki, G., Perutz, M. F., and Nagai, K. (1993). A mutagenic study of the allosteric linkage of His(HC3)146β in haemoglobin. J. Mol. Biol. 230, 1291–1296. 44. Kavanaugh, J. S., Rogers, P. H., and Arnone, A. (1992) High-resolution X-ray study of deoxy recombinant human hemoglobins synthesized from β-globins having mutated amino termini. Biochemistry 31, 8640–8647.

20

Safo and Abraham

Analysis of Hbs by HPLC

21

2 Analysis of Hemoglobins and Globin Chains by High-Performance Liquid Chromatography Henri Wajcman 1. Introduction In recent years, high-performance liquid chromatography (HPLC) has become a reference method for the study of hemoglobin (Hb) abnormalities. This technique is used in two distinct approaches. The first is quantitative analysis of the various Hb fractions by ion-exchange HPLC, which is now done in routine hospital laboratories mostly by using fully automated systems. The second is reverse-phase (RP)-HPLC, which is of interest for more specialized studies (see Note 1). 2. Materials and Methods 2.1. Ion-Exchange HPLC Separation of Hbs Cation-exchange HPLC is the method of choice to quantify normal and abnormal Hb fractions (1–4). This is the method of reference for measuring glycated Hb for monitoring diabetes mellitus. It is also generally used for measuring of the levels of HbA2, HbF, and several abnormal Hbs. According to some researchers, this method could even replace electrophoretic techniques for primary screening of Hbs of clinical significance (3,5–7) or, at least, should be an additional tool for the identification of Hb variants (8). Automated apparatuses have been developed for large series measurement. I describe the Bio-Rad Variant Hemoglobin Testing System (Bio-Rad, Hercules, CA), using the β Thalassemia Short program as an example of this type of equipment.

From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

21

22

Wajcman

2.1.1. Bio-Rad Variant Hb Testing System The Bio-Rad apparatus is a fully automated HPLC system, using double wavelength detection (415 and 690 nm). The β Thalassemia Short program is the most widely used system for HbA2 and HbF measurements, but other elution methods, including specific columns, buffers, and software, are available from the manufacturer according to the test to perform. This program has been designed to separate and determine, in 5 to 6 min, area percentage for HbA2 and HbF and to provide qualitative determinations of a few abnormal Hbs. Windows of retention time have been established for presumptive identification of the most commonly occurring Hb variants. The β Thalassemia Short program uses a 3.0 × 0.46 cm nonporous cation-exchange column that is eluted at 32 ± 1°C, with a flow rate of 2 mL/min, by a gradient of pH and an ionic strength made of two phosphate buffers provided by the manufacturer. This material and procedure have been used worldwide in many laboratories over the last several years. Since recommendations for experimental procedure are fully detailed by the manufacturer, I describe only a few additional notes of practical import. 1. Blood is collected on adenine citrate dextrose (ACD). 2. Samples for analysis (about 0.2% Hb) are obtained by hemolysis of 20 µL of blood in 1 mL of a buffer containing 5 g/L of potassium hydrogenophthalate, 0.5 g/L of potassium cyanide, 2 mL of a 1% solution of saponine, and distilled water. This procedure for sample preparation, which is currently used for HPLC determination of HbA1c, avoids some of the Hb components present in low amounts (about 1%) eluted together with HbF in the HbF retention time window (8). 3. Twenty microliters of hemolysate is applied onto the column for analysis.

Under these experimental conditions an excellent agreement is found between chromatographic measurement of HbF, down to 0.2%, and resistance to alkali denaturation, up to 15% (9). Presumptive identification of the most commonly occurring variants (Hb S, HbC, HbE, and HbD Punjab) is made using the retention time windows named S-Window, D-Window, A2-Window, and C-Window, which have been specified by the manufacturer. Aged Hb specimens display some degraded products that are eluted in the P2 and P3 windows (e.g., glutathione-Hb) (Table 1). Slight differences in the elution time of the various Hb components are observed from column to column and from one reagent batch to another, which should be taken into account by a program supplied by the manufacturer. The elution time of an Hb component varies also slightly according to its concentration in the sample. For a given column, a more accurate calibration than that proposed by the manufacturer could be obtained using HbA2 as reference. The concentration of this Hb, which varies between narrow limits, prevents significant modification of its elution time.

Analysis of Hbs by HPLC

23

Table 1 Analyte Identification Windowa Analyte name F P2 P3 A0 A2 D-window S-window C-window a Example

Retention time (min)

Band (min)

Window (min)

1.15 1.45 1.75 2.60 3.83 4.05 4.27 5.03

0.15 0.15 0.15 0.40 0.15 0.07 0.15 0.15

1.00 -1.30 1.30-1.60 1.50-1.90 2.20-3.30 3.68-3.98 3.98-4.12 4.12-4.42 4.88-5.18

provided by manufacturer.

Two methods are available for comparing data when the elution time of HbA2 differs between two runs done with a different column or reagent batch. The first consists of slightly modifying the experimental procedure (temperature or pH) to reproduce exactly the elution times of the previous runs. The second method consists of establishing a normalized retention scale taking as references two Hbs eluted within a linear part of the gradient. The elution patterns of more than 100 variants have been published, but, in my opinion, these data should be used as a confirmatory test for characterization of a variant after a careful multiparameter electrophoretic study (8) rather than as a primary identification method.

2.1.2. Alternative Methods When a dedicated machine is not available for Hb analysis, or when the chromatographic separation is done for “preparative” purposes, alternative techniques have to be used. These procedures are suitable for conventional HPLC equipment. Several anion-exchange and cation-exchange HPLC columns may be used for Hb separation; some are silica based and others are synthetic polymers. These methods have been well standardized for several years (10,11). PolyCat A (Poly LC, Columbia, MD) is one of the more popular phases for Hb separations (6). It consists of 5-µm porous (100-nm) spherical particles of silica coated with polyaspartic acid. For analytical purposes, a 5.0 × 0.40 cm column is used; elution is obtained at 25°C with a flow rate of 1 mL/min, by developing in 20 min at pH 6.58 a linear gradient of ionic strength from 0.03 to 0.06 M NaCl in a 50 mM Bis-Tris, 5 mM KCN buffer. The presence of KCN is necessary to convert methemoglobin into cyanmethemoglobin, which displays ion-exchange chromatographic properties similar to those of oxyhemoglobin (see Note 2).

24

Wajcman

2.2. HPLC Analysis of Globin Chains 2.2.1. Analysis of HbF Composition (see Note 3) The solvent system, acetonitrile–trifluororoacetic acid (TFA), which is used for RP-HPLC, dissociates the Hb molecule into its subunits and removes the heme group. This method is therefore used to analyze or separate the globin chains. This kind of study may be useful in the investigation of many human Hb disorders. For instance, the determination of HbF composition (Gγ:Aγ ratio) is of interest in several genetic and acquired disorders. A good separation is obtained between the Gγ and AγI, with most of the RP columns by using a very flat acetonitrile gradient. By contrast, it is often much more difficult to separate Gγ from AγT, a frequent allele of AγI. Among the procedures that have been successfully proposed for this analysis, one of the most popular is the RP-HPLC method described by Shelton et al. (12). They used a Vydac C4 column (The Separation Group, Hesperia, CA) eluted at a flow rate of 1 mL/min by developing in 1 h a linear gradient from 38 to 42% acetonitrile in 0.1% TFA with detection at 214 nm. Under these conditions, the chains were eluted in the following order: β, α, AγT, Gγ, and AγI. In recent years, a modification introduced in the manufacturing process of this type of column (13) made necessary the use the higher acetonitrile concentrations to elute the γ-chains. Unfortunately, it also resulted in the low resolution of AγT. 2.2.1.1. RP PERFUSION CHROMATOGRAPHY

Perfusion chromatography involves a high-velocity flow of the mobile phase through a porous chromatographic particle (14–16). The Poros R1® media (Applied Biosystems, Foster City, CA) used in this technique consists of 10-µm-diameter particles. These particles are made by interadhering under a fractal geometry poly(styrene-divinylbenzene) leading to throughpores of 6000- to 8000-Å-diameter microspheres with short, diffusive 500- to 1000-Ådiameter pores connected to them. As a result, relatively low pressures are obtained under high flow rates. The Poros R1® beads may be considered a fimbriated stationary phase having retention properties somewhat similar to those of a classic C4 support (15). The column (10 × 0.46 cm) is packed on a conventional HPLC machine at a flow rate of 8 mL/min using the Poros selfpack technology® according to the manufacturer’s protocol. More than a thousand runs may be performed without alteration of the resolution. 2.2.1.1.1. Sample Preparation 1. Samples containing about 0.1 mg of Hb/mL are obtained by lysis, in 1 mL water (or 5 mM KCN), of 2–5 µL of washed red blood cells (RBCs). 2. Membranes are removed by centrifuging at 6000g for 10 min.

Analysis of Hbs by HPLC

25

3. According to the HbF level, 20–100 µL of these hemolysates are applied onto the column. To avoid additional chromatographic peaks owing to glutathione adducts, 10 µL of a 50 mM solution of dithiothreitol in water is added per 100 µL of sample. An in-line stainless steel filter (0.5-µm porosity) needs to be used to protect the column.

2.2.1.1.2. Equipment. Any conventional HPLC machine can be used. In the method described here, the analyses were performed on a Shimadzu LC-6 HPLC machine equipped with an SCL-6B system controller, an SIL-6B autoinjector, and a C-R5A integrator (Shimadzu, Kyoto, Japan). A flow rate of 3.0–4.5 mL/min was convenient for synchronization of injection, integration, and column equilibration. 2.2.1.1.3. Experimental Procedure (see Note 4). Using a flow rate of 3 mL/min, the various γ-chains are isolated by developing in 9 min a linear gradient from 37 to 42% acetonitrile in a 0.1% solution of TFA in water. In practice, this is done by using two solvents (A: 35% acetonitrile, 0.1% TFA in water; B: 50% acetonitrile, 0.1% TFA in water) and a linear gradient from 15 to 45% B. Before injection, the column is equilibrated by a 10 column volume wash with the starting solvent, thus allowing completion of a cycle of analysis every 14 min. Elution is followed at 214 nm (wavelength at which double bonds absorb), and the recorder is set to 0.08 AUFS. Higher flow rates may be used, but the slope of the gradient will need to be increased in proportion. Keeping the same initial and final acetonitrile concentrations as above, elution is achieved in 6 min at a flow rate of 4.5 mL/min and in 4 min at a flow rate of 6.0 mL/min.

2.2.2. RP-HPLC Analysis of Globin Chains (see Note 5) Globin chain analysis is also important as an additional test that allows discrimination between Hb variants for the identification of structural abnormalities. Several RP-HPLC procedures have been proposed (10,14,17,18). On a conventional HPLC apparatus, a 20 × 0.46 cm column packed with Lichrospher 100 RP8 (Merck, Darmstadt, Germany) is used. Samples are prepared as described in Subheading 2.2.1.1.1. Elution is obtained at 45°C with a flow rate of 0.7 mL/min using a 90-minute linear gradient of acetonitrile, methanol, and NaCl made by a mixture of two solvents (18). Solvent A contains acetonitrile, methanol, and 0.143 M NaCl, pH 2.7 (adjusted by a few drops of 1 N HCl), in the proportion of 24, 38, and 36 L/L, respectively. Solvent B is made from the same reagents but in the proportion of 55, 6, and 39 L/L, respectively. The gradient starts with 10% B and ends with 70% B. The design of the gradient may be modified according to the machine, the geometry of the column, and the separation to be achieved. Elution can be followed at 214 or 280 nm. Globin chains are eluted in the same order as on the Vydac C4 column.

26

Wajcman

A kit for globin chain analysis with similar performance is also commercially available from Bio-Rad (ref. 270.0301).

2.2.3. Scaled Up Methods for Chain Separation For biosynthetic or structural studies, milligram amounts of globin chains need to be separated. This can be achieved either by scaling up the RP-HPLC procedure using semipreparative size columns or by cation-exchange -HPLC done in the presence of dissociating concentrations of urea. 2.2.3.1. SEMIPREPARATIVE SIZE RP-HPLC COLUMNS

2.2.3.1.1. Samples. Globin solution rather than Hb solution is used. Globin is prepared from a 1% Hb solution obtained by hemolysing washed RBCs in distilled water. Stromas are removed by centrifuging at 6000g for 30 min, and the globin is precipitated by the acid acetone method. Usually, the sample is made from 1 to 2 mg of globin dissolved in 250 µL of 0.1% TFA, which requires the use of a 500-µL injection loop. 2.2.3.1.2. Chromotographic Procedure. A 240 × 10 mm Vydac C4 column (ref. 214TP510) is used. Elution is obtained by a gradient of acetonitrile in 0.1% TFA made by two solvents (solvent A contains 35% acetonitrile and solvent B 45%). A typical elution program, using a flow rate of 1.2 mL/min, consists of a 10-min equilibration at 35% B, 70 min of a linear gradient from 35 to 55% B, 30 min of a linear gradient from 55 to 90% B, and 5 min of an isocratic step at 90% B for cleaning the column. Elution of the column is followed at 280 nm with a full scale of 0.16 absorbance units (AU). 2.2.3.2. CATION-EXCHANGE HPLC IN PRESENCE OF 6 M UREA USING A POLYCAT COLUMN

Procedures that are modified from the classic CM cellulose chromatography described by Clegg et al. (19) may be transposed to the HPLC technology (20). The retention capacity of this type of column is higher than that of RP supports, allowing the handling of larger samples. I describe here a method using a PolyCat 300-Å, 10-µm particle column (150 × 4 mm). 2.2.3.2.1. Reagents and Buffers. Two buffers are used. Buffer A consists of 6 M urea, 0.1 M sodium acetate, and 0.4% β-mercaptoethanol, with the pH adjusted to 5.8 by acetic acid. Buffer B consists of 6 M urea, 0.25 M sodium acetate, and 0.35% β-mercaptoethanol, with the pH adjusted to 5.8 by acetic acid. Both buffers need to be filtered through a membrane with 0.45-µm porosity before being used. In addition, an in-line stainless steel filter (0.5-µm porosity) is needed to protect the column. 2.2.3.2.2. Samples. Up to 5–10 mg of globin, prepared by the acid acetone method, is dissolved in 200–600 mL of buffer A.

Analysis of Hbs by HPLC

27

2.2.3.2.3. Chromatographic Procedure. Elution is obtained by a gradient of ionic strength developed with the two buffers. A typical elution program, using a flow rate of 1.0 mL/min, consists of a 10-min equilibration at 0% B, 5 min of a linear gradient from 0 to 25% B, 50 min of a linear gradient from 25 to 100% B, and 5 min of an isocratic step at 100% B for cleaning the column. Elution of the column is followed at 280 nm with a full scale of 0.32 UA. 3. Notes 1. Why should one method be preferred over another? The choice of a separation method between RP or ion-exchange chromatography depends on the purpose of the separation. Ion-exchange is the only chromatographic method that allows preparation of native Hb fractions. The presence of cyanide ions in the buffers (or during sample preparation) will nevertheless hinder any further oxygenbinding study. If the aim of the separation is to obtain Hbs suitable for functional studies, the technique will have to be modified accordingly by removing cyanide from all the steps. It may be of interest in some cases to work with carbonmonoxyhemoglobin, since Hb is very stable under this form and procedures are available to return to the oxyform. For several applications, salts in excess also need to be removed. RP separation methods always lead to denatured proteins that cannot be used for functional studies. Techniques involving an ionic strength gradient can only be used for analytical purposes. By contrast, using fully volatile buffers, such as the acetonitrile-TFA system, the isolated globin fractions can be vacuum dried and readily used for further structural studies such as mass spectrometry measurements. 2. To isolate amounts of Hb in the milligram range, larger columns (15.0 × 0.46 cm) may be used. According to the separation to be achieved, the dimensions of the column, and the apparatus used, slightly different experimental conditions may have to be designed. Elution is followed at 415 nm for analytical purposes or at 540 nm in preparative runs. This buffer system is not suitable for ultraviolet (UV) detection. The use of an in-line stainless steel filter (0.5-µm porosity) is recommended to increase the column life expectancy. Reproducibility requires careful preparation of the buffers and temperature control. Since in these chromatographic methods the elution is recorded at one of the wavelengths of absorption of the heme, any factor modifying the absorption spectrum of the Hb molecule will hinder accurate quantitative measurement. For instance, unstable Hb variants, which lose their heme groups or lead to hemichrome formation, will be underestimated. HbMs, which are hardly converted into cyanmethemoglobin, display a much higher extinction coefficient than oxyhemoglobin at 415 nm and a lower one at 540 nm. As a consequence, HbMs will be overestimated when measured at the first wavelength, and underestimated at the second one. A modified experimental procedure allowing for a simultaneous measurement of HbF, glycated Hb, and several other Hb adducts has been proposed by using a combination of pH and ionic gradients (11).

28

Wajcman

3. In my laboratory, for routine determination of the γ-chain composition, we replaced this procedure with an RP perfusion chromatography using a Poros R1® column (Applied Biosystems) (14). 4. To obtain good reproducibility, we recommend using the same glassware for preparing the solvents. Solvents may be kept refrigerated at 4°C for a few days. Accurate balance of the TFA between both solvents is important to avoid baseline drift. Acetonitrile must be of HPLC grade with low UV absorbancy in the 210-nm region. With this Poros R1 column, the α-chain is eluted before the β-chain. Resolution may be improved by modifying the geometry of the column or the design of the gradient. A 10 × 0.2 cm column may be used to improve separation between the various γ- or adult chains. In this case, with a flow rate of 1 mL/min, after 5 min of equilibration at 5% B, the column is eluted using a 15-min linear gradient between 5 and 25% B of the described solvents. This is followed by a 2-min isocratic elution at 25% B. 5. Several columns may be used, but I have found that a method adapted from that described in ref. 17 leads to a good resolution. Other columns or techniques may nevertheless be more appropriate for some specific separations. When chromatographic methods are used for globin chain quantification, it is important to consider the absorption coefficient of the various chains at the wavelength of detection. In some cases, it may be identical, such as when comparing the various γ-chains. In other cases, the absorption may differ considerably; for example, at 280 nm, γ-chains, because of their 3 Trp residues, have a higher ε coefficient than β-chains (2Trp) and α-chains (1 Trp). Abnormal Hbs containing a number of aromatic residues different from the normal may also display modified absorption coefficient.

References 1. Wilson, J. B., Headlee, M. E., and Huisman, T. H. J. (1983) A new high-performance liquid chromatographic procedure for the separation and quantitation of various hemoglobin variants in adults and newborn babies. J. Lab. Clin. Med. 102, 174–185. 2. Kutlar, A., Kutlar, F., Wilson, J. B., Headlee, M. E., and Huisman, T. H. J. (1984) Quantitation of hemoglobin components by high-performance cation-exchange liquid chromatography: its use in diagnosis and in the assessment of cellular distribution of hemoglobin variants. Am. J. Hematol. 17, 39–53. 3. Rogers, B. B., Wessels, R. A., Ou, C. N., and Buffone, G. J. (1985) High-performance liquid chromatography in the diagnosis of hemoglobinopathies and thalassemias. Am. J. Clin. Pathol. 84, 671–674. 4. Samperi, P., Mancuso, G. R., Dibenedetto, S. P., Di Cataldo, A., Ragusa, R., and Schiliro, G. (1990) High performance liquid chromatography (HPLC): a simple method to quantify HbC, O-Arab, Agenogi and F. Clin. Lab. Haematol. 13, 169–175. 5. Shapira, E., Miller, V. L., Miller, J. B., and Qu, Y. (1989) Sickle cell screening using a rapid automated HPLC system. Clin. Chim. Acta 182, 301–308.

Analysis of Hbs by HPLC

29

6. Ou, C. N. and Rognerud, C. L. (1993) Rapid analysis of hemoglobin variants by cation-exchange HPLC. Clin. Chem. 39, 820–824. 7. Papadea, C. and Cate, J. C. (1996) Identification and quantification of hemoglobins A, F, S, and C by automated chromatography. Clin. Chem. 42, 57–63. 8. Riou, J., Godart, C., Hurtrel, D., Mathis, M., Bimet, C., Bardakdjian-Michau, J., Préhu, C., Wajcman, H., and Galactéros, F. (1997) Evaluation of cation-exchange high-performance liquid-chromatography for presumptive identification of hemoglobin variants. J. Clin. Chem. 43, 34–39. 9. Préhu,C., Ducrocq, R., Godart, C., Riou, J., and Galactéros, F. (1998) Determination of HbF levels: the routine methods. Hemoglobin 22, 459–467. 10. Huisman, T. H. J. (1998) Separation of hemoglobins and hemoglobin chains by high performance liquid chromatography. J. Chromatogr. 418, 277–304. 11. Bisse, E. and Wieland, H. (1988) High-performance liquid chromatographic separation of human hemoglobins. Simultaneous quantitation of fetal and glycated hemoglobins. J. Chromatogr. 434, 95–110. 12. Shelton, J. B., Shelton, J. R., and Schroeder, W. A. (1984) High-performance liquid-chromatographic separation of globin chains on a large-pore C4 column. J. Liq. Chromatogr. 7, 1969–1977. 13. Vydac. (1994–1995) HPLC columns and separation materials, Technical Bulletin. 14. Wajcman, H., Ducrocq, R., Riou, J., Mathis, M., Godart, C., Préhu, C., and Galacteros, F. (1996) Perfusion chromatography on reversed-phase column allows fast analysis of human globin chains. Anal. Biochem. 237, 80–87. 15. Afeyan, N. B., Gordon, N. F., Mazsaroff, I., Varady, L., Fulton, S. P., Yang, Y. B., and Regnier, F. E. (1990) Flow-through particles for the high-performance liquid chromatographic separation of biomolecules: perfusion chromatography. J. Chromatogr. 519, 1–29. 16. Afeyan, N. B., Fulton, S. P., and Regnier, F. E. (1991) Perfusion chromatography material for proteins and peptides. J. Chromatogr. 544, 267–279. 17. Leone, L., Monteleone, M., Gabutti, V., and Amione, C. (1985) Reversed-phase high performance liquid chromatography of human hemoglobin chains. J. Chromatogr. 321, 407–419. 18. Wajcman, H., Riou, J., and Yapo, A. P. (2002) Globin Chains Analysis by RP-HPLC: recent developments. Hemoglobin 26, 271–284. 19. Clegg, J. B., Naughton, M. A., and Weatherall, D. J. (1966) Abnormal human hemoglobins: separation and characterization of the a and b chains by chromatography, and the detereminatioin of two new variants, Hb Chesapeake and Hb J (Bangkok) J. Mol. Biol. 19, 91–108. 20. Brennan, S. O. (1985) The separation of globin chains by high pressure cation exchange chromatography. Hemoglobin 9, 53–63.

30

Wajcman

Purification and Molecular Analysis of Hb by HPLC

31

3 Purification and Molecular Analysis of Hemoglobin by High-Performance Liquid Chromatography Belur N. Manjula and Seetharama A. Acharya 1. Introduction Hemoglobin (Hb) is a tetrameric protein (mol wt = 64,500) and is the major protein component of red blood cells (RBCs). In normal human erythrocytes, HbA composes about 90% of the total Hb. It is made up of two identical α-chains and two identical β-chains. Besides HbA, human erythrocytes contain small amounts of other forms of Hb as fetal hemoglobin (HbF, α2γ2) and HbA2 (α2δ2), and products of posttranslational modifications as HbA1c. HbS, the sickle cell Hb, is a genetic variant of HbA and is the most widely studied pathological form of Hb (1). Hb is a subject of active research not only for its molecular, genetic, and clinical aspects, but also as a prototype of allosteric proteins. Purification and characterization of Hbs has become easier and faster with the advent of highpressure and high-performance instrumentation, high-sensitivity detectors, and the availability of a wide variety of high-resolution column-packing materials. Methodological development using very small quantities of the protein is feasible, and the analytical methods are readily scalable. Here, we describe three different modes of high-performance liquid chromatography (HPLC) that are used in our laboratory for the purification and characterization of Hb, and modified or mutant Hb. Hb is purified by ion-exchange chromatography (IE-HPLC), its size is analyzed by size-exclusion chromatography (SEC-HPLC) (under native and dissociating conditions), its globin chain separation is accomplished by reverse phase HPLC (RP-HPLC), and tryptic peptide mapping of globin chains is also carried out by RP-HPLC. Preparative runs are generally carried out on an From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

31

32

Manjula and Acharya

AKTA Protein Purification System (Amersham Pharmacia Biotech), which is also used for analytical runs. Other instrumentation used for the analytical runs includes a fast protein liquid chromatography (FPLC) system (Amersham Pharmacia Biotech) for SEC and ion-exchange chromatography, and a Shimadzu Liquid Chromatography System for RP-HPLC. Examples of ionexchange chromatographic purifications are given for analytical-scale runs (100 µg to 1 mg), small-scale preparative runs (up to 50 mg), and large-scale preparative runs (up to 3 g). The analytical-scale runs are useful not only for methodological development, but also for characterization purposes and for monitoring the progress of a chemical modification reaction. The SEC-HPLC runs are illustrated with analytical (1 mg) and semipreparative runs (100 mg). Examples of RP-HPLC are for analytical-scale runs (120 µg). 2. Materials 2.1. Purification of Human Hb by Ion-Exchange Chromatography (see Notes 1 and 2)

2.1.1. Anion-Exchange Chromatography 2.1.1.1. PURIFICATION OF HB ON DEAE-SEPHAROSE FAST FLOW: SMALL-SCALE PURIFICATION (SEE NOTES 1 AND 2) 1. 2. 3. 4. 5.

XK 16/10 chromatographic: column (Amersham Pharmacia Biotech). DEAE-Sepharose Fast Flow anion exchanger: (Amersham Pharmacia Biotech). Buffer A: 50 mM Tris-Ac, pH 8.5. Buffer B: 50 mM Tris-Ac, pH 7.0. Amersham Pharmacia Biotech AKTA Protein Purification System.

2.1.1.2. PREPARATIVE-SCALE PURIFICATION OF HB ON Q-SEPHAROSE HIGH PERFORMANCE CHROMATOGRAPHIC COLUMN 1. 2. 3. 4. 5.

XK26/70 column (Amersham Pharmacia Biotech) (see Notes 1 and 2). Q-Sepharose High Performance column-packing material (Amersham Pharmacia Biotech). Buffer A: 50 mM Tris-Ac, pH 8.5. Buffer B: 50 mM Tris-Ac, pH 7.0. Amersham Pharmacia Biotech AKTA Protein Purification System.

2.1.1.3. CATION-EXCHANGE CHROMATOGRAPHY: RECHROMATOGRAPHY OF Q-SEPHAROSE HIGH PERFORMANCE PURIFIED HBA ON CM-SEPHAROSE FAST FLOW 1. 2. 3. 4. 5.

XK26/70 chromatographic column (Amersham Pharmacia Biotech). CM-Sepharose Fast Flow cation exchanges (Amersham Pharmacia Biotech). Buffer A: 10 mM potassium phosphate, pH 6.35, 1 mM EDTA. Buffer B: 15 mM potassium phosphate, pH 8.5, 1 mM EDTA. Amersham Pharmacia Biotech AKTA Protein Purification System.

Purification and Molecular Analysis of Hb by HPLC

33

2.2. Ion-Exchange Chromatography as an Analytical Tool 2.2.1. Characterization of Recombinant Hb by Cation-Exchange Chromatography on a Mono S Column 1. Mono S HR5/5 column (1 mL) (Amersham Pharmacia Biotech) (see Notes 3 and 4). 2. Buffer A: 10 mM potassium phosphate, pH 6.5. 3. Buffer B: 15 mM potassium phosphate, pH 8.5. 4. Pharmacia FPLC protein purification system. 5. Shimadzu UV-VIS detector at 540 nm. 6. Shimadzu Chromatopac CR7A plus data processor.

2.2.2. Monitoring Progress of a Chemical Modification Reaction Analysis of Amidated HbS by Analytical Anion-Exchange Chromatography on HiTrap Q Column 1. 2. 3. 4.

HiTrap Q, 1 mL column (Amersham Pharmacia Biotech) (see Note 3). Buffer A: 50 mM Tris-Ac, pH 8.5. Buffer B: 50 mM Tris-Ac, pH 7.0. Amersham Pharmacia Biotech AKTA Protein Purification System.

2.3. SEC of Hb 2.3.1. Analytical SEC 1. Pharmacia Superose 12 HR 10/30, two columns in series (column volume [CV], 47 mL) (see Note 4). 2. Buffer: 50 mM Bis-Tris (pH 7.4) or phosphate-buffered saline (PBS), pH 7.4, for analysis of tetrameric and size-enhanced Hbs; 50 mM Bis-Tris and 0.9 M MgCl2 (pH 7.4), for evaluating the stabilization of the tetrameric structure of Hb. 3. Pharmacia FPLC Protein Purification System. 4. Detector: Shimadzu UV-VIS detector at 540 nm. 5. Shimadzu CR7A plus data processor.

2.3.2. Semipreparative SEC 1. 2. 3. 4.

Pharmacia XK26/70 column. Superose 12 prep-grade packing material (Pharmacia). Buffer: PBS, pH 7.4. Pharmacia Biotech AKTA Protein Purification System.

2.4. Globin Chain Analysis of Hb by RP-HPLC Analysis 1. 2. 3. 4.

Column: Vydac Protein C4 column (4.6 × 250 mm). Solvent A: H2O, 0.1% trifluoroacetic acid (TFA). Solvent B: acetonitrile, 0.1% TFA. Shimadzu Liquid Chromatography System consisting of two LC-6A pumps, an SPD-6A UV detector, an SCL-6B System Controller, and Class VP chromatography software.

34

Manjula and Acharya

3. Methods 3.1. Purification of Human HbA by Ion-Exchange Chromatography HbA is purified from erythrocytes obtained from adult human blood. The erythrocytes are gently washed with cold PBS, pH 7.4, and lysed with 4 volumes of water. The lysate containing the Hb is separated from the cell debris by centrifugation. The lysate is dialyzed extensively against PBS, pH 7.4, to strip the protein of 2,3-diphosphoglycerate. Because Hb can exist as an anion or a cation, depending on buffer conditions, it can be purified by either anion- or cation-exchange chromatography, or a combination of the two. Routinely, the erythrocyte lysate is first purified on a Q-Sepharose High Performance column or on a DEAE-Sepharose Fast Flow column followed by a second chromatography on a CM-Sepharose Fast Flow column. All purifications are carried out at 4°C.

3.1.1. Anion-Exchange Chromatography 3.1.1.1. PURIFICATION OF HB ON DEAE-SEPHAROSE FAST FLOW: SMALL-SCALE PURIFICATION

A typical elution profile of a human RBC lysate (25 mg of human RBC lysate injected in 500 µL of buffer A) is shown in Fig. 1. This represents an analytical-scale run of the same sample for which a preparative run is given in Subheading 3.1.1.2. on a Q-Sepharose High Performance column (Fig. 2). Runs like this are useful for the evaluation of the run conditions prior to preparative runs. The total run time is 5 h 6 min, the total volume is ~612 mL, and the gradient time/volume is 2 h/240 mL. 1. Pack the column (1.6 cm × 6 cm, CV = 12 mL) according to the manufacturer’s directions. 2. Wash the column with 1 CV each of water, buffer A, and buffer B. 3. Equilibrate the column with 10–25 CV of buffer A at a flow rate of 2 mL/min. 4. Inject the sample and wash the column with 1 CV of buffer A to elute unbound protein. 5. Elute the bound protein with a linear gradient of 0–100% buffer B in 20 CV. 6. Monitor the column effluent at 540, 600, and 630 nm (see Note 5). 7. Clean the column with 10 CV of buffer B. 8. Reequilibrate with 20 CV of buffer A.

3.1.1.2. PURIFICATION OF HB ON Q-SEPHAROSE HIGH PERFORMANCE: PREPARATIVE RUN

A typical chromatographic profile of a human erythrocyte lysate (load: ~40 mL containing ~3 g of Hb) is shown in Fig. 2. The protein eluting at 1500 mL (~65% buffer B) corresponds to HbA. The fractions corresponding to this peak

Purification and Molecular Analysis of Hb by HPLC

35

Fig. 1. Small-scale anion-exchange chromatography of human red cell lysate on a DEAE-Sepharose Fast Flow column (1.6 × 6 cm) at 4°C. Buffer A: 50 mM Tris-Ac, pH 8.5; buffer B: 50 mM Tris-Ac, pH 7.0. The column was equilibrated with buffer A, and a decreasing pH gradient of 0–100% buffer B over 20 CV was used for elution of the protein. Protein load: 25 mg.

are pooled, concentrated, and subjected to further purification by cationexchange chromatography on a CM-Sepharose Fast Flow column (Subheading 3.1.2.). The following run takes about 25 to 26 h. 1. Pack the Q-Sepharose High Performance ion-exchange column at 4°C in a Pharmacia XK26/70 column according to the manufacturer’s directions. Typically, a 2.6 × 58 cm column (~290-mL column volume) is used for the purification of 2.5–3 g of Hb. 2. Wash the column first with 1 CV of water, followed by 1 to 2 CV each of 20% buffer B and 100% buffer B. 3. Equilibrate the column with at least 10 CV of 20% buffer B, at a flow rate of 1.5 mL/min. 4. Dialyze the red cell lysate extensively against 20% buffer B, and filter through a 0.2-µm filter. 5. Load the lysate onto the column manually using line A18 of Pump A. 6. Elute the protein with a linear gradient of decreasing pH consisting of 20–100% buffer B in 8 column volumes (2320 mL). 7. Monitor the column effluent simultaneously at three wavelengths; 540, 600, and 630 nm (see Note 5).

36

Manjula and Acharya

Fig. 2. Preparative-scale anion-exchange chromatography of human red cell lysate on Q-Sepharose High Performance column (2.6 × 58 cm) at 4°C. Buffer A: 50 mM Tris-Ac, pH 8.5; buffer B: 50 mM Tris-Ac, pH 7.0. The column was equilibrated with 20% buffer B, and a decreasing pH gradient of 20–100% buffer B over 8 CV was used for elution of the protein. Protein load: 3.0 g; fraction size: 20 mL.

3.1.2. Rechromatography of Q-Sepharose High Performance Purified HbA on a CM-Sepharose Fast Flow Column A typical chromatographic profile is shown in Fig. 3. The protein eluting at 1960 mL (~78% buffer B) corresponds to HbA. Pool the HbA-containing fractions, concentrate in an Amicon stirred cell to a concentration of 64–128 mg/mL, dialyze against the buffer of choice, and store either in liquid nitrogen or at –80°C. 1. Dialyze the HbA obtained from the Q-Sepharose High Performance column (~1.3 g in 60 mL) against 10 mM potassium phosphate buffer; 1 mM EDTA, pH 6.35 (see Note 6). 2. Load the dialyzed HbA on the CM-Sepharose Fast Flow column (2.6 cm × 59 cm), preequilibrated with the same buffer. The large sample volume is not a consideration, since the protein binds to the column at the initial conditions. Up to 3 g of Hb can be purified on a 2.6 × 59 cm column. 3. Elute the protein with a linear gradient of increasing pH, consisting of 0–100% buffer B over 8 column volumes (~2500 mL).

Purification and Molecular Analysis of Hb by HPLC

37

Fig. 3. Repurification of HbA purified on Q-Sepharose High Performance column (see Fig. 2) by cation-exchange chromatography on a CM-Sepharose Fast Flow column (2.6 × 59 cm) at 4°C. Buffer A: 10 mM potassium phosphate, 1 mM EDTA, pH 6.35; buffer B: 15 mM potassium phosphate, 1 mM EDTA, pH 8.5. The column was equilibrated with buffer A and an increasing pH gradient of 0–100% buffer B over 8 CV was used for elution of the protein. Protein load: ~1.3 g; fraction size: 12 mL. The effluent was monitored at 540, 600, and 630 nm. Elution profile for 540 nm is shown.

3.2. Ion-exchange Chromatography as an Analytical Tool 3.2.1. Characterization of HbA Expressed in Transgenic Swine by Cation-Exchange Chromatography on a Mono S Column The ion-exchange chromatographic procedures are also valuable as fast techniques for the analysis of Hb variants and recombinant hemoglobins. The elution positions are dependent on the surface topology of the Hb. The Mono S column (Amersham Pharmacia Biotech) distinguishes between the correctly folded and misfolded forms of recombinant HbA (rHbA) (2–4). Adachi et al. (2) have reported that the rHbA obtained from their yeast expression system contains a misfolded form of HbA in addition to the correctly folded form. The misfolded and the correctly folded forms of rHbA exhibit distinct elution positions on a Mono S column. Studies by Shen et al. (3,4) have shown that rHbA containing incorrectly inserted heme can be resolved from the species containing the correctly inserted heme on a Mono S column. In our studies, the chromatographic profile of the transgenic swine HbA on a Mono S column is identical to that of wild-type HbA (Fig. 4), which, in conjunction with NMR

38

Manjula and Acharya

Fig. 4. Comparison of elution profiles of wild-type HbA and HbA expressed in transgenic swine, on a Mono S HR5/5 column. The flow rate was 1 mL/min. Protein load: 1 mg. Buffer A: 10 mM potassium phosphate, pH 6.5; buffer B: 15 mM potassium phosphate, pH 8.5. After injection of the protein, the column was washed with 2 CV of buffer A, and the bound protein was eluted with a linear gradient consisting of 0–100% buffer B over 45 CV. The effluent was monitored at 540 nm.

and functional studies (5), has established the absence of misfolded forms in this preparation. 1. Equilibrate the Mono S column with 10 mM potassium phosphate, pH 6.5 (buffer A), at a flow rate of 1 mL/min. 2. Inject 1 mg of the HbA or TgHbA in 25 µL of buffer A. 3. Wash the column with 2 CV of buffer A. 4. Elute the protein with a linear increasing pH gradient consisting of 0–100% buffer B over 45 CV. 5. Monitor the column effluent at 540 nm. 6. Regenerate the column in situ by washing first with ~5 CV of 100% buffer B followed by reequilibration with 25 CV of buffer A (0% buffer B).

Purification and Molecular Analysis of Hb by HPLC

39

Fig. 5. Chromatography of amidated HbS on a 1-mL HiTrap Q column at room temperature. Buffer A: 50 mM Tris-Ac, pH 8.5; buffer B: 50 mM Tris-Ac, pH 7.0. The column was equilibrated with 10% buffer B. The column was washed with 5 CV of 10% buffer B after injection of the sample, and a decreasing pH gradient of 0–100% buffer B over 20 CV was used for elution of the protein. Protein load: 1 mg. The effluent was monitored at 540 nm.

3.2.2. Monitoring Progress of a Chemical Modificattion Reaction: Analysis of Amidated HbS by Analytical Anion-Exchange Chromatography on HiTrap Q 3.2.2.1. PREPARATION OF AMIDATED HBS

HbS was amidated with ethanolamine, through a carbodimide and sulfo-Nhydroxy-succinimide-mediated reaction, according to the previously described procedures (6,7). 3.2.2.2. CHROMATOGRAPHY OF AMIDATED HBS ON HITRAP Q COLUMN

The chromatographic profile of an HbS preparation amidated with ethanolamine is illustrated in Fig. 5. As can be seen, the amidated HbS can be separated well from the unreacted HbS. Thus, this profile illustrates the feasibility of establishing conditions for the separation of modified and unmodified HbS using small amounts of the protein and within a short period of time. Protein loads as little as 100 µg are sufficient for such runs. Thus, these columns are highly useful for methodological development as well as for monitoring the time course of a protein modification reaction.

40

Manjula and Acharya

Total run time is 60 min, total volume is ~60 mL, and sample run time/ volume is 25 min/25 mL. 1. Equilibrate the column with 10% buffer B, at a flow rate of 1 mL/min. 2. Inject a sample of 1 mg of HbS amidated with ethanolamine, and wash the column with 5 CV of 10% buffer B. 3. Elute the protein with a gradient of 10–100% buffer B in 20 CV. 4. Monitor the column effluent at 540 nm. 5. Regenerate the column in situ by washing first with 10 CV of 100% buffer B, followed by reequilibration with 25 CV of 10% buffer B.

3.3. SEC of Hb Three applications of SEC on Pharmacia Superose 12 are described; two applications are at an analytical level and the third is at a preparative level.

3.3.1. Analytical SEC 3.3.1.1. ESTABLISHING STABILIZATION OF TETRAMERIC STRUCTURE OF HB BY INTERDIMERIC (I.E., INTRATETRAMERIC) CROSSLINKING

The samples are generally injected in a volume of 25 µL. Since the SEC runs are under isocratic conditions, unlike the ion-exchange columns, no column regeneration step is necessary. Once the protein is eluted from the column and the baseline is stable, the column is ready for the analysis of the next sample. The run time for HbA is approx 70 min. Thus, it is possible to analyze several samples during the course of a working day. Under nondenaturing conditions, SEC analysis of Hb serves as a highly useful tool to establish the tetrameric structure and to analyze polymeric forms of Hb. In the presence of 0.9 M MgCl2, Hb dissociates into its constituent dimers (8). However, if the like chains are crosslinked, then the tetrameric structure is stabilized and the protein elutes as a tetramer. Thus, SEC analysis under dissociating conditions is a valuable tool to monitor the stabilization of the tetrameric structure of Hb by interdimeric (i.e., intratetrameric) crosslinking. These two modes of the SEC analyses are illustrated in Fig. 6A and 6B, from an analysis of HbA reacted with the bifunctional maleimide, bis-maleidophenyl PEG2000 (Bis-Mal-PEG2000) (9). SEC analysis of Bis-Mal-PEG2000-reacted HbA in 50 mM Bis-Tris-Ac, pH 7.4, a low-ionic-strength buffer, revealed that >95% of the protein elutes in the same position as the tetrameric HbA, and only a trace amount is present as an octameric species (Fig. 6A). By contrast, SEC analysis of the same BisMal-PEG2000-reacted HbA preparation on the same Superose 12 column but in the presence of 0.9 M MgCl2 revealed that nearly all of the Bis-MalPEG2000-reacted HbA still elutes at the 64,000-Dalton position (Fig. 6B, upper

Purification and Molecular Analysis of Hb by HPLC

41

Fig. 6. SEC of Bis-Mal-PEG2000-reacted HbA on Superose 12 at room temperature. Two Superose 12 HR 10/30 columns connected in series were used. (A) the buffer used was 50 mM Bis-Tris-Ac, pH 7.4; (B) the buffer used was 50 mM Bis-Tris-Ac, 0.9 M MgCl2, pH 7.4. For both (A) and (B) the flow rate was 0.5 mL/min, the protein load was 800 µg and the effluent was monitored at 540 nm (see Note 7).

panel), whereas the control HbA is completely dissociated into its constituent αβ dimers (Fig. 6B, lower panel). Thus, SEC analyses establish the stabilization of the tetrameric structure of HbA by intramolecular crosslinking with Bis-Mal-PEG2000. 3.3.1.1.1. SEC Analysis in 50 mM Bis-Tris-Ac, pH 7.4: Nondenaturing Conditions 1. Equilibrate the column with 50 mM Bis-Tris-Ac, pH 7.4, at a flow rate of 0.5 mL/min. 2. Inject the sample. 3. Elute the column with 50 mM Bis-Tris-Ac, pH 7.4.

42

Manjula and Acharya 3.3.1.1.2. SEC Analysis in 50 mM Bis-Tris-Ac; 0.9 M MgCl2, pH 7.4: Denaturing Conditions

1. Equilibrate the column with 2 CV of 50 mM Bis-Tris-Ac, 0.9 M MgCl2 (pH 7.4) at a flow rate of 0.5 mL/min. 2. Inject the sample. 3. Elute the column with 50 mM Bis-Tris-Ac, 0.9 M MgCl2, pH 7.4.

3.3.1.2. DETERMINING HYDRODYNAMIC VOLUME OF SIZE-ENHANCED HBS

Conjugation of large nonprotein molecules like PEG increases the hydrodynamic volume of the protein. SEC analysis serves as a useful tool to determine the increase in the hydrodynamic volume of Hb as a function of the chain length of the conjugated PEG molecule, and as a function of the number of PEG chains conjugated to the protein. 3.3.1.2.1. Preparation of PEGylated Hb

HbA in PBS, pH 7.4, is reacted with maleidophenyl derivatives of PEG 5000, PEG 10000, and PEG 20000 (unpublished). This results in the surface decoration of HbA at its two β93 cysteines. Homogeneous preparations of Hb carrying two copies of PEG 5K, 10K, and 20K were isolated by ion-exchange chromatography. 3.3.1.2.2. SEC Analysis

SEC analysis of the PEGylated Hbs is carried out on an analytical Superose 12 column, equilibrated and eluted with PBS, pH 7.4, as described in Subheading 3.3.1.1. The results are shown in Fig. 7. The PEGylated HbAs elute earlier than HbA on the Superose 12 column. The actual molecular mass of the three surface-decorated HbAs carrying two PEG chains of 5, 10, and 20 kDa is 74, 84, and 104 kDa, respectively. Thus, no resolution among three surface-decorated HbAs can be expected on the Superose 12 column based on the differences in their actual mass. Nevertheless, the three PEGylated HbAs are well resolved from each other, indicating an apparent increase in their size. The retention time of the PEGylated HbA decreases with the increase in the length of the attached PEG chain, indicating a progressive increase in the apparent size of HbA on surface decoration with PEG molecule of increasing chain length. Intertetrameric crosslinking of HbA with Bis-Mal-PEG600 results in the formation of defined oligomeric forms of HbA with molecular weights that are multiples of 64 kDa (see Subheading 3.3.2. for an example of a preparative run). Comparison of the retention times of surface-decorated HbAs with those of oligomerized HbAs indicated that the size enhancement is a linear function of the mass of the PEG chains attached and is approx 8 to 10 times that anticipated based on the actual molecular size of the attached PEG chain.

Purification and Molecular Analysis of Hb by HPLC

43

Fig. 7. Comparison of hydrodynamic volumes of PEGylated HbAs by SEC on Superose 12 at room temperature. Two Superose 12 HR 10/30 columns connected in series were used. Elution buffer: PBS, pH 7.4; flow rate: 0.5 mL/min; protein load: 800 µg in each case.

3.3.2. Semi-Preparative SEC: Purification of Oligomeric Forms of Hb SEC can also be used in a preparative mode for the purification of Hb based on its size. In the example shown here, HbA, stabilized against dissociation by intratetrameric crosslinking, was oligomerized by intertetrameric crosslinking with Bis-Mal-PEG600, and the products of the reaction were separated by SEC on a semipreparative Superose 12 column. A typical chromatographic profile is shown in Fig. 8. Reanalysis of the fractions from this run on an analytical column confirmed that the fractions eluting at the peak positions of 204, 182, and 169 mL represent tetrameric, octameric, and dodecameric Hb, respectively. Fractions containing the respective oligomeric forms of Hb are pooled, concentrated, and further purified by rechromatography on the same Superose column. 1. Equilibrate a semipreparative column of Superose 12 (2.6 × 65 cm, CV = 325 mL) with 2 to 3 CV of PBS, pH 7.4, at a flow rate of 1 mL/min. 2. Load ~100 mg of protein in 1 mL of PBS, pH 7.4. 3. Elute the column with PBS, pH 7.4.

3.4. Globin Chain Analysis of Hb by RP-HPLC Analysis The globin chains of Hb can be separated by RP-HPLC (10,11). Thus, RP-HPLC analysis is a useful technique for determining the purity of an HbA

44

Manjula and Acharya

Fig. 8. Purification of intertetramerically crosslinked HbA on a semipreparative Superose 12 column (2.6 × 65 cm) at 4°C. Elution buffer: PBS, pH 7.4; flow rate: 1 mL/min; protein load: ~100 mg in 1 mL of PBS, pH 7.4; fraction size: 1.5 mL. The effluent was monitored at 540 nm.

preparation; identifying of the chain modified after a chemical reaction; and isolating of the globin chains for peptide mapping, mass analysis, amino acid analysis, and sequencing A comparison of the RP-HPLC profiles of HbA and Bis-Mal-PEG2000reacted HbA is shown in Fig. 9. This result shows that the β-chains of HbA are completely modified after reaction with Bis-Mal-PEG2000. The modified β-globin eluted as a distinct peak, after the α-chain. Thus, RP-HPLC analysis is a useful tool to identify the globin chain modified after a chemical modification reaction. The globin chains can be isolated for mass analysis and peptide mapping to identify the location of the modification on the polypeptide chain. For such applications, a semipreparative C4 column (10 × 250 cm) is used. One to two milligrams of Hb can be applied on the semipreparative column. A flow rate of 2 mL/min and the same gradient as for the analytical run can be employed. The fractions containing the separated globin chains are collected, and the solvent is removed either in a Speedvac or by lyophilization. Peptide

Purification and Molecular Analysis of Hb by HPLC

45

Fig. 9. RP-HPLC analysis of globin chains of Bis-Mal-PEG2000-reacted HbA on a Vydac C4 column (4.6 × 250 mm, 300 °A). Solvent A: water, 0.1% TFA; solvent B: acetonitrile, 0.1% TFA. The column was equilibrated with 35% solvent B, and a linear gradient of 35–50% solvent B in 100 min was employed for the elution. Flow rate: 1 mL/min. The effluent was monitored at 210 nm.

mapping of the globin chains is carried out by RP-HPLC on a Vydac C18 column (12). 1. Equilibrate a Vydac C4 column with 35% acetonitrile, 0.1% TFA at a flow rate of 1 mL/min. 2. Mix 10–100 µL of HbA (50–150 µg) with 1 mL of 0.3% TFA in a 1.5 mL microfuge tube, vortex, and centrifuge at 12,000 rpm for 4 min to clarify the sample. 3. Inject the supernatant onto the column. 4. Elute the globin chains with a linear gradient of 35–50% acetonitrile, 0.1% TFA in 100 min. 5. Monitor the effluent at 210 nm. 6. Regenerate the column in situ by washing with 100% acetonitrile, 0.1% TFA for 15 min, followed by reequilibration with 35% acetonitrile, 0.1% TFA (about 30 min).

46

Manjula and Acharya

4. Notes 1. The DEAE-Sepharose Fast Flow, the Q-Sepharose High Performance, and the CM-Sepharose Fast Flow columns described here were run using an Amersham Pharmacia Biotech AKTA Protein Purification System. However, these runs can also be carried out on other chromatography systems such as the Pharmacia FPLC system or even by conventional techniques using a peristaltic pump and a twochamber gradient system. 2. The Fast Flow resins have greater mechanical strength than the cellulose-based resins and thus permit higher flow rates. Hence, the run can be completed within a much shorter duration than the Whatman cellulose ion exchangers. In the examples shown, the run time on the Q-Sepharose High Performance and the CM-Sepharose Fast Flow ion exchangers is approximately one-third that on corresponding cellulose-based ion-exchange columns. The High Performance and the Fast Flow ion exchangers can be regenerated and reequilibrated within the column after each run, and the columns can be reused for several runs. 3. Typically, for the analytical- and small-scale ion-exchange columns, the column cleaning and reequilibration steps are programmed as part of the method. Depending on the length of the gradient, the run time for a 1-mL column can vary from 30 to 60 min. Thus, these procedures permit evaluation of several run conditions within a short period of time, utilizing only small quantities of protein (as little as 100 µg of protein per run). 4. Care in the solvent and sample preparation is a crucial step especially for the high-pressure columns such as Mono S and Superose 12 HR 10/30. Routinely, all buffers are freshly prepared and filtered through a 0.2-µm filter. Larger samples are also filtered through a 0.2 µm-filter prior to loading on the column. Smallervolume samples for the Mono S, HiTrap, and analytical Superose 12 are clarified by centrifuging in a microfuge at 12,000 rpm for 4 min. 5. The AKTA Protein Purification System permits simultaneous monitoring of the effluent at three wavelengths. Routinely, all the Hb purifications are monitored at 540, 600, and 630 nm. Monitoring at 600 nm is useful in cases in which the absorbance at 540 nm is too high, and monitoring at 630 nm is useful for the determination of the relative amounts of met-Hb in the column fractions. 6. Since the starting pH for the CM-Sepharose Fast Flow column is > k'X[X]), the observed replacement rate becomes equal to kX. The same principle applies if a reagent is added to consume the dissociated ligand.

Rate Constant Measurement for Reactions with Hb

67

Fig. 1. Time course for CO binding to human deoxyHb in 0.1 M phosphate, pH 7.0, 20°C. In a Gibson-Dionex stopped-flow apparatus equipped with a 2-cm path length, 100 µM CO was mixed 1:1 with 10 µM deoxyHb. The reactant concentrations after mixing were [CO]total = 50 µM and [Hb]total = 5 µM, and the time course was followed at 436 nm. (䊊) Observed data points; solid line —— a fit to single exponential expression with an observed rate constant equal to 7.9 s–1. (Top) Plotted differences between observed data and fitted line (residuals). (Inset) Dependence of fitted pseudo firstorder observed rate constant on [CO]total after mixing.

In ligand consumption experiments, an excess of consuming agent is usually added so that the observed rate equals that for ligand dissociation. The rate of oxygen dissociation, kO2, is normally measured by reacting HbO2 (deoxygen complex with reduced hemoglobin) with either excess CO or high concentrations of dithionite to consume free O2 (2,5,11). The rate of CO dissociation, kCO, is measured by reacting carbon monoxide hemoglobin (HbCO) with excess NO or a protein that scavenges CO (2,5,11,12). The rate of NO dissociation, kNO,

68

Olson et al.

Table 1 Typical Rate Constants for CO, O2, and NO Binding to Human Hb at pH 7.0, 20–25°Ca T-state parameters Protein

k'T (µM –1s–1)

kT (s–1)

R-state parameters KT (µM –1)

Hb (tetramer) CO 0.12 0.16 0.75 O2 5–10 500–1000 ~0.01 NO 25 (~0.001) (~25,000) α-Subunit (tetramer) CO 0.14 0.15 0.9 O2 7.1 2000 0.004 β-Subunit (tetramer) CO 0.10 0.17 0.6 O2 6.4 1500 0.004

k'R kR (µM –1s–1) (s–1)

KR (µM –1)

6.0 66 60

0.008 20 0.00003

750 3.2 2,000,000

5.0 40

0.008 15

620 2.7

7.0 90

0.007 31

1000 2.9

a The T-state parameters were derived from measurements of the first step in ligand binding (i.e., k'1 = 4k'T and k1 = kT). The R-state parameters were derived from measurements of the last step in ligand binding (i.e., k'4 = k'R and k4 = 4kR). The values for O2 and CO in the first two rows were taken from analyses of the data in Figs. 1–5, Sawicki and Gibson (31), and Mathews and Olson (8). The values for NO were taken from Cassoly and Gibson (17), Moore and Gibson (13), and Eich et al. (41). The O2 and CO parameters for the individual α- and β-subunits in T- and Rstate Hb were taken from Unzai et al. (38) and the references therein. In the case of metal hybrid Hbs, the T state was often defined as in the presence of inositol hexaphosphate at pH 6.5, which can give abnormally high dissociation rate constants (see ref. 38).

can be measured by reacting HbNO with both excess CO to displace the bound ligand and excess dithionite to consume the newly released NO (13). The value of kNO can also be determined by mixing HbNO with excess molecular oxygen, which both displaces the NO and, when bound to the heme group, consumes free NO by dioxygenation (see Eq. 11; [14]).

2.2. Time Resolution: Rapid Mixing vs Flash Photolysis In general, the association of carbon monoxide with heme proteins is markedly slower than that of dioxygen or nitric oxide (NO) and is the easiest ligandbinding reaction to measure (Table 1). The rate-limiting step for CO binding is internal bond formation with the heme iron. CO enters and leaves the protein hundreds of times before it finally forms a bond with the iron atom, and, as a result, the overall bimolecular rate constant is normally small. By contrast, NO is so reactive that every ligand molecule that enters the protein combines with the iron atom before it has a chance to escape. Thus, NO binding is limited only by the rate of ligand entry into the protein. Dioxygen shows intermediate behavior,

Rate Constant Measurement for Reactions with Hb

69

being limited in part by the rate of movement into Hb and by the the rate of Fe-O2 bond formation (for a more complete discussion and references see ref. 15). The association rate constant for CO binding to Hb is usually small enough (0.1 to 5 µM–1s–1; Table 1) to allow measurement in simple rapid mixing, stopped-flow instruments with dead times of 2 to 3 ms. For example, at 100 µM free CO, the observed pseudo first rate for CO binding to low concentrations (≤20 µM) of human deoxyhemoglobin A (deoxyHbA) is 100 µM · ~0.1 µM–1s–1 = 10 s–1, which prescribes a half-time of ~70 ms. If the Hb is in the high-affinity, rapidly reacting conformation, the observed rate increases ~50-fold, yielding a half-time of ~1.5 ms. However, lower concentrations of CO can be used to increase the half-time of the reaction to ≥3 ms to allow measurements in rapid mixing devices. Alternatively, these more rapid CO reactions can be measured by the photolysis techniques described in Fig. 2. The association rate constants for NO and O2 binding to Hb are quite large (5–100 µM –1s–1) and normally can only be measured using laser photolysis techniques (7). For example, at 100 µM ligand, the observed rates of O2 and NO binding to high-affinity forms (R state) of human deoxyHbA are 100 µM · ~50 µM–1s–1 = 5000 s–1, yielding half-times of ~0.1 ms. These reactions are too rapid to be detected in rapid mixing experiments but are readily measured using laser photolysis techniques with excitation pulses ≤0.5 µs. In the case of O2 binding to the low-affinity, slowly reacting forms of Hb, the value of k'O2 is much smaller, ~5 µM–1s–1, which would lower the association rate to ~500 s–1 at 100 µM free ligand. Unfortunately, the dissociation rate constant for the low-affinity form of human Hb is on the order of 1000 s–1 (see Table 1). Since the observed pseudo first-order rate constant is the sum of the forward and backward rates (Eq. 1), the value of kobs is ≥1500 s–1, and t1/2 for the reaction is ≤0.5 ms, which precludes rapid mixing experiments. As result, there is no simple way of measuring O2 binding to deoxyHb by rapid mixing, and photolysis techniques with short excitation pulses are required. In the case of NO binding, the dissociation rate constants for either the lowor high-affinity forms of human hemoglobin are very small, 0.001–0.00001 s–1 (Table 1). Thus, NO binding to either the R or T states of deoxyHb can be measured by stopped flow, rapid mixing techniques. However, very low Hb (1–5 µM) and NO (1–10 µM) concentrations must be used to keep the observed rates in the range of 50–500 s–1 (16,17). 3. Results 3.1. CO Binding The association rate constant for CO binding can be measured by mixing a solution of deoxyHb with anaerobic buffer solutions containing various

70

Olson et al.

Fig. 2. Time courses for CO rebinding to human deoxyHb after photolysis with a 1-ms excitation pulse from a photographic flash lamp. The conditions are the same as those in Fig. 1, and the reactions were followed at 436 nm. Photolysis was carried out with two Sunpack 540 Strobes as described in Mathews and Olson (8). (A) Photolysis of 150 µM HbCO at four different excitation intensities. (Inset) A rapid, monophasic time course is observed when the extent of photolysis is ≤10%. (B) Complete photolysis at lower concentrations of HbCO.

amounts of dissolved ligand and measuring the disappearance of the deoxyHb absorption peak at 430 nm or appearance of the HbCO peak at 420 nm. In these experiments, small amounts of sodium dithionite are often added to ensure removal of all oxygen from both reactant solutions. A sample time course for CO binding to human deoxyHbA at pH 7.0 is shown in Fig. 1. The dependence of the apparent pseudo first-order rate constant on ligand concentration is shown in the inset. There is a linear dependence of the overall pseudo firstorder rate constant on [CO] after mixing. The slope of the curve is 0.15 µM–1s– 1 for HbA in 0.1 M phosphate buffer, pH 7.0, 20°C. The intercept of the k obs vs [CO] plot is effectively 0 since the values for CO dissociation are very small (from 0.1 to 0.005 s–1; Table 1). As noted by Gibson and Roughton 45 yr ago (reviewed in ref. 2), the time course for CO binding to deoxyHbA shows complex accelerating behavior. A fit to a single exponential decay expression shows systematic deviations from the observed data (Fig. 1, top). It is clear that the observed rate is increasing as

Rate Constant Measurement for Reactions with Hb

71

the reaction proceeds. A quantitative analysis of CO binding to human deoxyHb requires analysis in terms of the four-step Adair scheme: Hb4 + X

k'1 k1

Hb4X + X

k'2 k2

Hb4X2 + X

k'3 k3

Hb4X3 + X

k'4 k4

Hb4X4

(4)

Under most circumstances, the back reactions for CO binding can be neglected since the absolute value of the dissociation rate constants, k1 to k4, are very small compared to the rates of association (i.e., k'1[CO] >> k1; k'2[CO] >> k2; and so on). In effect, CO binding can be described by four consecutive irreversible reactions, whose analytical solutions under pseudo first-order conditions can be represented by sums of exponentials (see refs. 2 and 6). Assignment of rate constants to all four steps in CO binding is extremely difficult because the observed time course for CO binding is not very different from a simple exponential decay (Fig. 1). An independent experimental determination of the rate constant for the last step, k'4, is required. Gibson was the first to solve this problem by using partial photolysis techniques to measure this rate constant directly (reviewed in refs. 2 and 4). He constructed a Xe flash lamp (pulse length of ~10 µs) that could be used to flash photolyze the Fe-CO bond and drive the ligand out into the solvent. After the flash, CO rebinds to the newly produced Hb4(CO)3, Hb4(CO)2, Hb4(CO), and Hb4 intermediates, depending on the extent of photolysis. At 100% photolysis, fully deoxygenated Hb tetramers are generated; at ≤10% photolysis, the only reactive species is Hb4(CO)3, and the last step in ligand binding can be followed directly. Over the last 20 yr, these photolysis experiments have been extended to nanosecond and picosecond time regimes using ultrafast lasers. On these very short time scales, first-order, internal ligand rebinding is observed. These ultrafast processes are called geminate rebinding because the same iron/ligand pair is involved in bond reformation. Geminate recombination provides detailed information on the factors governing iron-ligand bond formation (15) and allows mapping of ligand movement into and out of the protein (18). In all of the following discussion, however, only long laser (~1-µs) or photographic flash (~1-ms) pulses are considered. In these experiments, only ligand rebinding from the solvent is being measured, all the processes are bimolecular, and the observed rates depend on the first power of the external ligand concentration. Sample time courses for CO rebinding to deoxyHbA after photolysis by a 1-ms pulse from photographic Xe flash lamps are shown in Fig. 2. When the total protein concentration is kept high ([Hb]total ≈ 150 µM; Fig. 2A), the time course for CO rebinding to human HbA after complete photolysis resembles that seen in rapid-mixing experiments (Fig. 1). Acceleration is seen in both cases. Decreasing the excitation light intensity with neutral-density filters leads to biphasic time courses (Fig. 2A, lower curves). The first phase represents rebinding to a rapidly

72

Olson et al.

reacting form of Hb, with an apparent bimolecular rate constant of ~6 µM–1s–1. At 5–10% photolysis, only the rapid phase is observed. This time course is assumed to represent the binding of the last CO molecule to hemoglobin (i.e. Hb4X3 + X → Hb4X4 in Eq. 4), and the observed bimolecular rate constant represents k'4 in the Adair scheme.

3.2. Analysis in Terms of Two States The deoxyheme group in the Hb4X3 intermediate was originally designated as Hb*, owing to its rapid reaction with CO (2). In the early 1970s, the twostate model of Monod, Wyman, and Changeux (19) was adopted as the standard for interpreting cooperative ligand binding. It is still the best first approximation for comparing mutant and naturally occurring Hbs. In this model, deoxyHb starts out in the low-affinity T state where the reactivity toward ligands is low. As ligands successively bind, the tetramer switches to the high-affinity R state so that the reactivities of the remaining deoxyheme group in Hb4X3 toward O2 and CO are ~300 to 1000 times greater, respectively, than any of the groups in Hb4 (Table 1). In the simple two-state model, the equilibrium constant for T-to-R isomerization in the completely unliganded Hb4 species is defined as L = [T]/[R] and is on the order of 1,000,000 for native HbA at pH 7.0 (20,21). The ratio of the ligand association equilibrium constants for binding to the T vs the R states is KT/KR ≈ 0.002–0.005. As the reaction proceeds, the tetramer isomerization constant decreases by L(KT/KR)n in which n is the number of ligands bound. For example, the binding of three ligands causes the isomerization constant to decrease from 106 to 10–2, and Hb4X3 is predominantly in the R state. Thus, under physiological conditions, measurement of the kinetics of the first step in the Adair scheme (Eq. 4) provides an estimate of the rate parameters for the T quaternary state, and time courses for the last step provide estimates of the corresponding R-state rate constants. Using these definitions, the Hb* species can be equated with the rapidly reacting R-state conformation (5,8,11).

3.3. Dimers and Monomers Are Rapidly Reacting In the late 1950s, Gibson (22) reported that the rapidly reacting form of Hb could also be seen after complete photolysis when the protein concentration was low (≤100 µM for human Hb, Fig. 2B). This apparent anomaly was resolved when Antonini and coworkers (11) were able to show that tetrameric, fully liganded Hb dissociates into dimers with an equilibrium dissociation constant K4,2 = 2–10 µM (for a review see ref. 11). In flash photolysis experiments with dilute HbCO (≤10 µM), a large fraction of the heme groups is present as dimers since the equilibrium dissociation constant for Hb4(CO)4 is ~1–5 µM under physiological conditions (11,21,23). By

Rate Constant Measurement for Reactions with Hb

73

contrast, native deoxyHb has an equilibrium dissociation constant ≤10–10 µM at neutral pH and is completely tetrameric at micromoar heme concentrations (21). After photodissociation of Hb2(CO)2, the newly formed deoxy dimers remain in the rapidly reacting or Hb* conformation and have to reaggregate to tetramers before switching back to the slowly reacting, T-state form (23). Dimer aggregation (k'2,4 ≈ 0.1 µM–1s–1) is relatively slow compared to CO rebinding at high ligand concentrations (11,24,25). The mechanism for interpreting time courses at low Hb concentrations and after complete photolysis is as follows:

(5)

The fraction of rapidly reacting Hb after complete photolysis is a measure of the amount of Hb2(CO)2 dimers present in the original solution. The fraction of heme groups that is present as dimers is given by Eq. 6:

(6)

Edelstein et al. (23) have measured the fraction of rapidly reacting species as a function of total heme concentration and shown that the K4,2 value determined from these kinetic analyses is identical to that determined by molecular weight measurements using ultracentrifugation. At roughly the same time, Antonini, Brunori, and coworkers (11) showed that isolated α- and β-chains are also rapidly reacting, and that all three species—triliganded tetramers, dimers, and monomers—show roughly the same R-state ligand-binding parameters (for newer reviews, see refs. 5,8).

3.4. CO Dissociation Rate constants for CO dissociation from Hb are measured by mixing HbCO complexes with a high concentration of NO. In general, the rate constant for NO binding to Hb is ≥10 times that for CO binding. As a result, the observed replacement rate is a direct measure of kCO (see Eq. 3, where k'CO[CO]/ k'NO[NO] is ~0 as long as [CO] ≤ [NO]). For these types of replacement reac-

74

Olson et al.

tions, the intermediates containing unliganded heme groups are Hb4(CO)3, Hb4(CO)2(NO), Hb4(CO)(NO)2, and Hb4(NO)3. All these species contain three bound ligands. Consequently, the ligand replacement reaction measures only the first step in dissociation and, in general, shows simple behavior. Sharma et al. (12,26) have used microperoxidase (a small heme-containing degradation product of cytochrome-c) to scavenge CO from fully liganded Hb and attempted to determine all four rate constants for CO dissociation from Hb4(CO)4. Although it is difficult to analyze these data quantitatively owing to complex side reactions of microperoxidase, Sharma et al.’s (26) work has defined the value of k1 for CO dissociation from Hb4CO to be ~0.1 s–1.

3.5. O2 Binding As described previously, O2 reacts too rapidly with both the high- and lowaffinity states of Hb to allow direct measurement of the association rate constant by rapid-mixing methods. Instead, laser photolysis techniques must be employed. The quantum yield for complete O2 dissociation is ~0.05, compared with ~0.7 for photodissociation of CO (27,28). As result, complete photolysis cannot be achieved by conventional photographic flash lamps. The best choice is a flash lamp–driven dye laser with a pulse length on the order of 500 ns and an energy output of 1–3 J (see ref. 7). Although easier to use, YAG lasers have a pulse length of ≤10 ns, after which substantial internal geminate recombination can occur. In the case of HbO2, the extent of internal recombination is ≥50%. Consequently, complete photolysis of O2 cannot be achieved with a 9-ns pulse regardless of the energy output of the laser. By contrast, a dye laser pulse of ~0.5 µs is long enough to pump all the O2 out of the protein if sufficient energy is available in the excitation pulse. Sample time courses for O2 rebinding to human HbA at low and high free [O2] are shown in Figs. 3A and 3B, respectively. Even at high protein concentration, the observed time courses are biphasic at all levels of photolysis. The fastest process represents O2 association with rapidly reacting Hb* or R-state forms of Hb. The slower processes represent rebinding to low-affinity T-state forms. At low [O2], the rapid phase comprises about 30% of the absorbance change after complete photolysis. This result for O2 rebinding contrasts with that seen for CO rebinding after complete photolysis using a longer pulse. In the latter case, only one slow phase is observed at high protein concentration (100% photolysis; Fig. 3A). The persistence of rapid O2 rebinding occurs because the switch from the rapidly reacting R-state conformation to the slowly reacting T-state tetrameric form is not instantaneous (29–31). The rates for the R-to-T switch are between 1000 and 10,000 s–1 and on the same order as the apparent rates of rebinding at 100 µM O2 (~1000 s–1 to the T state and ~10,000 s–1 to the R state; see Table 1). This interpretation is shown schematically in Eq. 7:

Rate Constant Measurement for Reactions with Hb

75

Fig. 3. Photolysis of human HbO2 at low and high [O2] in 0.1 M phosphate, pH 7.0, 20°C. A Phase-R 2100B dye laser was used to produce a 0.5-µs excitation pulse at 577 nm, and the extent of photolysis was attenuated with neutral-density filters (8). O2 rebinding was monitored at 436 nm. (A) Time courses at 100 µM free O2. Under these conditions, the majority of the rebinding reaction is slow at 100% photolyis. (B) Time courses at 1260 µM O2. Under these conditions, most of the reaction is fast. (Insets) At ≤10% photolysis, the reaction is monophasic and very rapid.

(7)

After the excitation pulse, there is competition between rapid O2 rebinding to the Hb* form of the newly formed deoxy tetramers and the conformational transition to the more slowly reacting T-state tetramer (Eq. 7). If the rate of O2 association is increased by raising its concentration, the percentage of slowly reacting form decreases because more of the O2 molecules rebind to the R-state form before it

76

Olson et al.

can switch to the slowly reacting T-state conformation. Thus, at 1250 µM free O2 (1 atm), the percentage of rapid rebinding increases to ~65% at full photolyis (Fig. 3B). Complete analyses of O2 rebinding time courses are very complex and require, in addition to the normal ligand-binding parameters, the assignment of multiple R-to-T conformational change rate constants for all the various Adair intermediates in Eq. 4 plus consideration of dimerization (see refs. 9 and 10). However, the results in Fig. 3 do allow a qualitative estimation of O2 association rate constants. The time courses can be fitted to two exponential expressions, and the observed fast and slow rate parameters can be assigned to the R and T forms. Plots of kobs (fast and slow) vs [O2] are roughly linear at high ligand concentrations, and the apparent rate constants are k'TO2 ≈ 9 µM–1s–1 and k'RO2 ≈ 60 µM–1s–1. The rate of the slower phase does show a complex dependence on [O2] at low levels of ligand and incomplete saturation, making it difficult to assign exact values for k'1 in the Adair scheme (10,30,31). By contrast, at high [O2] and ≤10% photolysis, the value of k'4 is readily obtained since only the rapid phase of rebinding is observed (Fig. 3B, inset).

3.6. O2 Dissociation Time courses for oxygen dissociation from Hb can be measured in stopped flow, rapid-mixing experiments using the ligand replacement and consumption reactions described in Eq. 2. When HbO2 is reacted with anaerobic buffer containing very high concentrations of sodium dithionite, a single phase is observed with an overall apparent rate constant of ~60–100 s–1 at pH 7.0, 20°C (Fig. 4, lower curve). If the reaction is carried out again with CO in the dithionite solution, the observed rate is two- to threefold smaller, ~30 s–1 (Fig. 4, upper curve). This difference is a reflection of the cooperative nature of O2 release in the absence of replacing ligands. This situation is shown schematically in Eq. 8.

(8)

Rate Constant Measurement for Reactions with Hb

77

Fig. 4. Time courses for O2 dissociation from human Hb in 0.1 M phosphate, pH 7.0, 20°C. HbO2 (10 µM) in air-equilibrated buffer was mixed with anaerobic buffer containing excess sodium dithionite (lower curve) and with buffer equilibrated with 1 atm of CO containing excess dithionite (upper curve). The reaction was monitored by absorbance increases at 424 nm.

When O2 is consumed by dithionite in the absence of other ligands, the rate of dissociation increases as the protein switches from the R to the T conformation. This switch occurs after ~2 ligands are lost. As a result, the first O2 dissociates at the R-state rate, which is ~30 s–1. When the second ligand dissociates, the remaining sites lose O2 immediately because their rates of dissociation are 20- to 50-fold higher. Thus, the second rate is two to three times that of the first rate, and the time course shows acceleration. Consequently, the overall firstorder rate constant is two to three times greater than that for R-state Hb. When CO is present, the protein remains fully liganded, except for the transient formation of triliganded intermediate with one empty site. Under these conditions, only O2 dissociation from fully liganded tetramers is being measured (see Eq. 4). The reaction of dithionite with free O2 leads to the formation of hydrogen peroxide, which can cause secondary oxidative reactions. In addition, dithionite can also reduce methemoglobin (MetHb) impurities and degradation products. Both of these side reactions can cause large spectral changes that can interfere

78

Olson et al.

with measurement of the “true” O2 dissociation reaction. In our view, the best method for measuring the value of k4 for O2 dissociation is to mix HbO2 with buffer equilibrated with 1 atm of CO and then to vary the free [O2] in the Hb sample. The observed replacement rate should be given by the expression in Eq. 3, in which X = O2 and Y = CO. A sample time course for the reaction of human HbO2 with CO is shown in Fig. 5A. The dependence of robs on [CO]/[O2] is shown in the inset. The expression used to fit these data comes from rearranging Eq. 3:

(9)

kRO2 is defined as the intrinsic rate constant for O2 dissociation from a heme group in fully liganded Hb. The value of k4 in the Adair scheme (Eq. 4) would be 4kRO2 since there are four possible sites from which the O2 can be dissociated in Hb4(O2)4 (for more complete discussion of statistical factors see refs. 2, 5, and 6) The fitted value for the limiting rate in the inset to Fig. 5A is 29 s–1 in 0.1 M phosphate, pH 7.0, 20°C. This value of kRO2 is equal to the observed rate obtained when HbO2 was mixed with buffer containing both CO and high concentrations of sodium dithionite (Fig. 4, upper curve). This analysis also allows a quantitative determination of the ratio k'RO2/k'RCO, which can be used to confirm assignments made in partial photolysis measurements with the corresponding Hb4(O2)4 and Hb4(CO)4 complexes. Thus, analysis of the CO replacement reaction can also be used to obtain values for k'RO2 if the rate of CO binding to R-state Hb has already been measured.

3.7. Differences Between α- and β-Subunits Close examination of time courses for both O2 replacement and rebinding after ≤10% photolysis shows systematic deviations from simple exponential behavior (Fig. 5A, B, top). The pattern of residuals indicates two components, one reacting ~2 times faster than the other. This systematic deviation from simple monophasic behavior was first noted for the replacement reaction by Olson and Gibson in 1971 (32) and attributed to differences between the α- and β-subunits within tetrameric Hb. Their interpretation has been confirmed by four different sets of experiments over the past 30 yr. Large ligands such as alkyl isocyanides enhance subunit differences (33–35). Chemical modification of the β Cys93 with either mercurials or alkylating agents selectively increases the rate of O2 dissociation from β-subunits (32). Hybrid recombinant Hbs have

Rate Constant Measurement for Reactions with Hb

79

Fig. 5. Time courses for O2 displacement by CO and O2 rebinding after partial photolysis (≤10%) of HbO2 at pH 7.0, 20°C. (A) HbO2 in air was mixed with buffer equilibrated with 1 atm of CO. The concentrations after mixing were [HbO2] = 5 µM, [O2] = 131 µM, and [CO] = 500 µM. (䊊) Observed data; solid line —— a fit to a single exponential expression with kobs = 9.5 s–1. (Top) Differences between observed data and fitted line (residuals). (Inset) Dependence of apparent pseudo first-order rate constant on [CO]/[O2]. The circles represent the observed rate constants and the line a fit to Eq. 9. (B) Time courses for O2 rebinding after 10% photolysis of 100 µM HbO2 taken from the inset in Fig. 3B. (Inset) The circles represent observed data and the line a fit to a single exponential expression with kobs = 76,000 s–1.

been constructed in which one type of subunit is mutated to be either more or less reactive toward ligands, and the other is kept as the wild-type chain (36,37). These mutant hybrids have been used to define the ligand-binding parameters of the “normal” α- and β-subunits (8). Metal hybrid Hbs have been constructed to examine the individual subunits in both the high- and low-affinity quaternary states. The most definitive hybrids contain Cr- and Ni-porphyrin substitutions in one type of subunit and Fe-porphyrin in the other (see ref. 38 and references therein). The Cr(III)-porphyrin groups are inert to ligands but remain 6-coordinated owing to covalently bound water that promotes the high-affinity R state. Ni(II)-porphyrin is also inert but remains 5- or 4-coordinated, biasing the tetramer toward the T

80

Olson et al.

Table 2 Simple Methods for Assigning Rate Constants to T (Initial Step) and R (Final Step) Forms of Human Hb Ligand

Rate parameter

Technique

CO

k'T (association) k'R (association)

Stopped flow Conventional flash Stopped flow Laser photolysis photolysis Laser photolysis Stopped flow Stopped flow

O2

kR (dissociaton) k'T (association)

NO

k'R (association) kR (dissociation) k'T (association) k'R (association) kR (dissociation)

Laser photolysis Conventional mixing

k'NO,ox

Stopped flow

Reaction Hb4 + CO – simple binding Hb4(CO)4 – partial (≤10%) photolysis Hb4(CO)4 + NO – ligand replacement Hb4(O2)4 – slow phase after 100% Hb4(O2)4 – partial (≤10%) photolysis Hb4(O2)4 + CO – ligand replacement Hb4 + NO – simple binding at very low [Hb],[NO] Hb4(NO)4 – partial (≤10%) photolysis Hb4(NO)4 + CO(excess DT) or Hb4(NO)4 + O2 – ligand replacement and NO scavenging (very slow) Hb4(O2)4 + NO – NO dioxygenation, Hb oxidation at very low [HbO2], [NO]

quaternary structure. Unzai et al. (38) have used (α(Fe)β(Cr))2 and (α(Cr)β(Fe))2 to assign rate constants to the individual subunits in the R conformation and (α(Fe)β(Ni))2 and (α(Ni)β(Fe))2 to assign T-state rate parameters. Quantitative analyses of time courses for ligand binding to Hb become difficult if subunit differences are taken into account. Twenty different intermediates must be considered if the two quaternary states and their rates of interconversion are considered explicitly. This situation is made even more difficult if low concentrations of heme are used and dissociation into dimers occurs. However, this complexity should not inhibit investigators from making the kinetic measurements shown in Figs. 1–5, particularly if the goal is to survey mutant or species differences. Simple experiments that are readily carried out are provided in Table 2. Rate constants for the ligand binding to the R state (defined here as the last step in the Adair scheme in Eq. 4) are readily determined from simple exponential analysis of ligand replacement time courses and rebinding time courses after ≤10% photolysis. The rate constant for CO association to T-state deoxyHb can be estimated by simple mixing experiments. The rate constant for O2 association to the T state can be obtained from analysis of the slow phases observed after total photolysis of HbO2 at high ligand concentrations. The dissociation rate constants for the first step in ligand binding to the T state are much more difficult to measure for native tetrameric Hb.

Rate Constant Measurement for Reactions with Hb

81

In the case of O2, the T-state dissociation rate constant is often estimated by analyzing oxygen equilibrium curves, fixing the R-state parameters and the T-state association rate constants from direct kinetic measurements, and then fitting for the allosteric constant (L = [Hb4(T)]/[Hb4(R)]) and kT. In general, simple analyses that assume subunit equivalence and only two states are inadequate for understanding the detailed structural mechanisms underlying cooperative ligand binding. However, the results in Table 1 show that the differences between the subunit rate constants are less than a factor of 2 for native human HbA. Unzai et al. (38) have argued that even though the structural mechanism for the change in ligand affinity differs significantly between the α- and β-subunits, the two subunits have evolved similar rate and equilibrium constants for O2 binding in order to maximize the amount of observed cooperativity. Thus, the parameters for a simple two-step scheme do provide a useful framework for understanding O2 transport (first three rows in Table 1). The efficiency of O2 delivery depends primarily on the extent and rate of changes in fractional saturation of Hb at different oxygen tensions. As a result, the simple set of parameters, which assumes subunit equivalence, is primarily sufficient to simulate the O2 transport properties of native Hb in capillary experiments (39,40).

3.8. Reversible NO Binding The reactions of NO with Hb are important for detoxifying NO and preventing its transport. The two key reactions are the reversible binding of NO to the ferrous iron atom in deoxyHb and the oxidative reaction of NO with bound O2 to produce nitrate and metHb. In both cases, the rate-limiting step is the capture of NO in the distal pocket of Hb subunits (41). Once inside the protein, NO reacts extremely rapidly, presumably on picosecond time scales, with either the heme iron atom or bound O2 atoms. As a result, both reactions are very rapid: k'NO and k'NO,ox ≈ 60 µM–1s–1 at pH 7.0, 20°C (41). The reaction of NO with deoxyHb is very rapid even when the protein is fully unliganded and in the T quaternary state. In 1975, Cassoly and Gibson (17) reported that the bimolecular rate constant, k'NO, is ~40 µM–1s–1 for both the first and last steps in ligand binding. More recent measurements suggest that k'1 for NO may be twofold less than k'4 (unpublished data), and there may also be small subunit differences (see ref. 41). However, to first approximation, Cassoly and Gibson (17) were correct. The rate of NO binding to deoxyHb is governed exclusively by the rate of ligand entry into the protein and not the R-to-T transition, which affects primarily reactivity at the iron atom. Since the NO dissociation rate constants are very small, 0.001–0.00003 s–1 (13,42), it is possible to measure the rate of the overall binding reaction in rapid-

82

Olson et al.

Fig. 6. Time courses for the reactions of human HbO2 and sperm whale MbO2 with equimolar amount of NO in 0.1 M phosphate, pH 7.0, 20°C. The concentrations of reactants after mixing were 1 µM. If the reaction is second order and irreversible, plots of 1/[HbO2] vs time should be linear, and the slope determines the bimolecular rate of NO dioxygenation, k'NO,ox. (Inset) Plots for all three proteins.

mixing experiments. However, very low protein and ligand concentrations must be used, and only small, very rapid absorbance changes can be observed. The half-time of the reaction at 1 µM deoxyHb and 1 µM NO is 1/(~40 µM–1s–1 · 1 µM) ≈ 0.025 s. If the NO concentration is raised to 10 µM, the rate becomes ~400 s–1, the half time is ~1.6 ms, and >70% of the absorbance change occurs in the dead time of the apparatus (for equivalent time courses, see Figs. 6 and 7). In addition, great care must be taken to maintain anaerobic conditions in the absence of any added dithionite since both O2 and dithionite will consume the small amounts of NO required for the experiments. NO rebinding can be measured at high ligand concentrations using laser photolysis techniques. The problem in this case is that the quantum yield for complete photodissociation of NO into the solvent phase is ~0.001 at room

Rate Constant Measurement for Reactions with Hb

83

Fig. 7. Reactions of HbO2 and MbO2 with excess NO. The conditions are the same as those in Fig. 6. (Top panel) Observed time courses for reaction of 10 µM NO with human HbO2 and sperm whale MbO2. The secondary phases represent binding of NO to newly formed ferric forms of the proteins. (Bottom panel) Dependence of observed pseudo first-order rate constants for the fast phases on [NO] after mixing. The reaction with the V68F MbO2 mutant is included to show that the linearity is observed over a wide range of NO concentrations (see ref. 41).

temperature. The practical consequence is that no more than 20–30% photolysis can be achieved, even with a 0.5-µs excitation pulse of ~2 to 3 J (28). Thus, only partial photolysis time courses can be measured. The observed rates under these R-state conditions are ~60 µM–1s–1 at pH 7.0, 20°C (41), which is about 1.5- to 2-fold greater than the association rate constant measured by mixing fully deoxygenated Hb with NO at very low heme concentrations (17).

84

Olson et al.

NO dissociation from fully liganded Hb can be measured in three ways. First, Moore and Gibson (13) showed that kRNO can be measured directly by mixing HbNO with a concentrated solution of dithionite containing 1000 µM (1 atm) CO. As shown in Eq. 10, the added CO will occupy the deoxy sites formed by NO dissociation long enough for the NO to be consumed by dithionite:

(10)

In the absence of CO, the reaction of released NO with the newly formed deoxyHb site will compete effectively with the reaction with dithionite, and the observed rate will be much smaller than the true value of kNO. Second, Sharma and Ranney (42) used excess deoxyMb to scavenge NO from nitrosylHb; however, this method is complex owing to the need for very high concentrations of deoxyMb, which makes simple absorbance measurements difficult. Third, an equally valid method is to expose HbNO to high concentrations of O2 and measure the rate of autoxidation to MetHb. The mechanism of this reaction is as follows:

(11)

In this case, the newly formed deoxyHb site reacts rapidly with O2 since the ratio k'O2[O2]/k'NO[NO] is always kept ≥100 (note that in Table 1, k'O2 ≈ k'NO for R-state Hb). The newly dissociated NO will react rapidly with bound O2 to form metHb and nitrate. NO can also react with free O2 to produce nitrite. The latter reaction is very slow at low NO concentrations (t1/2 ≈ minutes at 1 µM NO (43)) when compared with the reactions of NO with Hb and HbO2 (t1/2 ≈ 0.1–10 ms at 1–10 µM heme (41)). Foley et al. (unpublished observation) and Eich (14) have shown that there is a 1:1 correlation between the rates of autoxidation of ~30 different mutant nitrosylmyoglobins and the corresponding rate constants for NO dissociation

Rate Constant Measurement for Reactions with Hb

85

measured using the CO/dithionite method of Moore and Gibson (13). The same correlation holds for nitrosylHb. NO binding at low levels of Hb saturation is complicated by the formation of pentacoordinate NO-heme complexes in α-subunits, particularly at low pH and in the presence of organic phosphates (44,45). The formation of the pentacoordinate complex is slow (t1/2 ≈ 1 s) compared to ligand binding (47). Its formation also causes ligand reequilibration from an equal mixture of αNO and βNO to predominately αNO complexes in the first Adair intermediate, Hb4NO (46). These complications are manifested as slow, small absorbance changes when deoxyHb is mixed with subsaturating levels of NO. Fortunately, these secondary changes have little effect on the measured rate constants as long as NO is in excess in the mixing or photolysis experiments.

3.9. NO Dioxygenation by HbO2 It has been known for a long time that the addition of NO to either HbO2 or oxymyoglobin causes a very rapid and stoichiometric oxidation of the heme group and formation of nitrate (see references in refs. 47 and 48). This reaction has been used for more than 20 yr as a simple assay for NO synthase activity (for a review, see ref. 49). However, the physiological importance of this process has become clear only within the last 10 yr. HbO2 in red cells and oxymyoglobin in muscle tissue detoxify NO by converting it to NO3–. This scavenging function prevents NO from being transported into actively respiring tissue, where it would inhibit both aconitase and cytocrome oxidase and shut down oxidative phosphorylation (50–53). Gardner (54–56) has called this activity NO dioxygenation and has shown that it is catalyzed efficiently by flavohemoglobins from various microorganisms, in which the expression of the gene is turned on by the addition of toxic levels of NO. Extracellular Hb scavenges NO much more rapidly than Hb packaged in red cells owing to its closer proximity to the endothelium and extravasation into the vessel walls (57,58). The net results are loss of NO signal molecules, little activation of guanylyl cyclase, sustained smooth muscle contraction, and elevated of blood pressure. Thus, we and others have looked for ways to reduce the reactivity of HbO2 with NO in order to design safer and more effective Hb-based blood substitutes (58–61). Time courses for the reaction of 1 µM NO with 1 µM HbO2, oxymyoglobin, and a slowly reacting mutant of myoglobin (V68F, Val68[E11] to Phe) are shown in Fig. 6. Plots of 1/[HbO2]remaining or 1/[MbO2]remaining vs time are linear, indicating simple, irreversible reactions. The bimolecular rate constants for HbO2, wild-type MbO2, and V68F MbO2 are 65, 35, and 10 µM–1s–1, respectively. Further proof that the reaction is bimolecular and depends directly on the first power of [NO] is shown in Fig. 7B. When the reactions are carried out

86

Olson et al.

under pseudo first-order conditions, [NO] ≥ 5 · [HbO2], there is a linear dependence of kobs on [NO], and the slopes of these plots give rate constants equal to those determined in the equimolar-mixing experiments. The problems associated with the NO dioxygenation reaction are shown in Fig. 7. First, the reactions are very fast and difficult to measure, and it is hard to get more than 2 or 3 points for kobs vs [NO] plots. At 10 µM NO, well over half the reaction with HbO2 is “lost” in the dead time of the apparatus (Fig. 7A). Second, excess free NO can react with newly formed MetHb and MetMb species, causing slow absorbance changes that often occur in the opposite direction of the oxidation reaction (Fig. 7A, right). Third, care must be taken to keep O2 out of the NO solutions and the plastic portions of the mixing device. Otherwise, NO will react slowly with free O2 to produce nitrite in a reaction that consumes 4 NO mol/O2. The resultant nitrite will eventually oxidize HbO2 but at a much slower rate. Fourth, at high pH (≥8.5), a spectral intermediate can be observed and has been assigned to an Fe3+-peroxynitrite complex by Herold and coworkers (63,64) (Fig. 8). Thus, the kinetic scheme for NO dioxygenation reactions needs at least two steps at alkaline pH:

(12)

The peroxynitrite intermediate, Hb(Fe3+OONO–), is formed by a bimolecular process with a rate constant similar to that observed for the overall reaction at pH 7.0, ~50–70 µM–1s–1 (62). This intermediate decays by a first-order process into MetHb and nitrate with no release of peroxynitrite or any ferryl heme formation (63). The rate of decay of the intermediate for both HbO2 and MbO2 increases with decreasing pH. At pH 7.0, it decays too rapidly to be observed. The overall reaction is limited by the bimolecular capture of NO, and simple monophasic kinetic behavior is observed (Fig. 6). In the case of Hb at high pH, the time course of the intermediate decay is biphasic, and Herold (63) has suggested that this is owing to subunit differences. Although difficult to measure, the physiological importance of the NO dioxygenation reaction requires that it be examined routinely in studies of both native and recombinant Hbs. The simplest approach is that shown in Figs. 6 and 7. For example, Eich et al. (41) used these types of experiments to show

Rate Constant Measurement for Reactions with Hb

87

Fig. 8. Reaction of HbO2 with NO in 0.1 M borate, pH 9.0, 20°C. HbO2 (50 µM after mixing) was reacted with an equimolar amount of NO in an OLIS RSM stoppedflow apparatus, and spectra were collected as rapidly as possible (one spectrum every 1 millisecond) in the visible wavelength region (470–670 nm). As shown originally by Herold et al. (64), a spectral species (– – –) resembling the complex of nitrite with MetHb is formed in the dead time of the apparatus. This intermediate decays in a firstorder process to the final hydroxymethemoglobin product (· · ·) with an overall halftime of ~0.020 s.

that Val(E11) to Phe or Trp substitutions in β-subunits and Leu(B10) to Phe or Trp substitutions in α-subunits reduce the rate of NO dioxygenation by HbO2 markedly, up to 10- to 30-fold. Herold et al. (63) used similar methods to characterize the chemistry of the NO dioxygenation reaction with both oxymyoglobin and HbO2. Acknowledgments This research was supported by United States Public Health Service grants GM 35649 and HL 47020 and grant C-612 from the Robert A. Welch Foundation. DHM and EWF were recipients of predoctoral fellowships from NIH Biotechnology Training Grant GM 08362.

88

Olson et al.

References 1. Hartridge, H. and Roughton, F. J. W. (1923) A method of measuring the velocity of very rapid chemical reactions. Proc. Roy. Soc. A 104, 376–394. 2. Gibson, Q. H. (1959) The kinetics of reactions between haemoglobins and gases, in Progress in Biophysical Chemistry, vol. 9 (Butler, J. A. and Katz, B., eds.), Pergamon, New York, pp. 1–53. 3. Gibson, Q. H. and Milnes, L. (1964) Apparatus for rapid and sensitive spectrophotometry. Biochem. J. 91(1), 161–171. 4. Gibson, Q. (1978) Flash photolysis techniques. Methods Enzymol. 54, 93–101. 5. Olson, J. S. (1981) Stopped-flow, rapid mixing measurements of ligand binding to hemoglobin and red cells. Methods Enzymol. 76, 631–651. 6. Olson, J. S. (1981) Numerical analysis of kinetic ligand binding data. Methods Enzymol. 76, 652–667. 7. Sawicki, C. A. and Morris, R. J. (1981) Flash photolysis of hemoglobin. Methods Enzymol. 76, 667–681. 8. Mathews, A. J. and Olson, J. S. (1994) Assignment of rate constants for O2 and CO binding to alpha and beta subunits within R- and T-state human hemoglobin. Methods Enzymol. 232, 363–386. 9. Henry, E. R., Jones, C. M., Hofrichter, J., and Eaton, W. A. (1997) Can a twostate MWC allosteric model explain hemoglobin kinetics? Biochemistry 36(21), 6511–6528. 10. Gibson, Q. H. (1999) Kinetics of oxygen binding to hemoglobin A. Biochemistry 38(16), 5191–5199. 11. Antonini, E., and Brunori, M. (1971) Hemoglobin and myoglobin in their reactions with ligands, in Frontiers in Biology (Neuberger, A. and Tatum, E. L., eds.), North-Holland, Amsterdam. 12. Sharma, V. S., Ranney, H. M., Geibel, J. F., and Traylor, T. G. (1975) A new method for the determination of ligand dissociation rate constant of carboxyhemoglobin. Biochem. Biophys. Res. Commun. 66(4), 1301–1306. 13. Moore, E. and Gibson, Q. (1976) Cooperativity in the dissociation of nitric oxide from hemoglobin. J. Biol. Chem. 251, 2788–2794. 14. Eich, R. F. (1997), Reactions of nitric oxide with myoglobin, PhD thesis, Rice University, Houston, TX. 15. Olson, J. S., and Phillips, G. N. Jr. (1996) Kinetic pathways and barriers for ligand binding to myoglobin. J. Biol. Chem. 271(30), 17,593–17,596. 16. Gibson, Q. H. and Roughton, F. J. (1965) Further studies on the kinetics and equilibria of the reaction of nitric oxide with haemoproteins. Proc. R. Soc. Lond. B. Biol. Sci. 163(991), 197–205. 17. Cassoly, R. and Gibson, Q. (1975) Conformation, co-operativity and ligand binding in human hemoglobin. J. Mol. Biol. 91(3), 301–313. 18. Scott, E. E., Gibson, Q. H., and Olson, J. S. (2001) Mapping the pathways for O2 entry into and exit from myoglobin. J. Biol. Chem. 276(7), 5177–5188. 19. Monod, J., Wyman, J., and Changeux, J.-P. (1965) On the nature of allosteric transitions: a plausible model. J. Mol. Biol. 12, 88–118.

Rate Constant Measurement for Reactions with Hb

89

20. Gibson, Q. H. and Edelstein, S. J. (1987) Oxygen binding and subunit interaction of hemoglobin in relation to the two-state model. J. Biol. Chem. 262(2), 516–519. 21. Ackers, G. K. (1998) Deciphering the molecular code of hemoglobin allostery. Adv. Protein Chem. 51, 185–253. 22. Gibson, Q. H. (1959) The Photochemical formation of a quickly reacting form of hemoglobin. Biochem. J. 71, 293–303. 23. Edelstein, S. J., Rehmar, M. J., Olson, J. S., and Gibson, Q. H. (1970) Functional aspects of the subunit association-dissociation equilibria of hemoglobin. J. Biol. Chem. 245(17), 4372–4381. 24. Andersen, M. E., Moffat, J. K., and Gibson, Q. H. (1971) The kinetics of ligand binding and of the association-dissociation reactions of human hemoglobin: properties of deoxyhemoglobin dimers. J. Biol. Chem. 246(9), 2796–807. 25. McGovern, P., Reisberg, P., and Olson, J. S. (1976) Aggregation of deoxyhemoglobin subunits. J. Biol. Chem. 251(24), 7871–7879. 26. Sharma, V. S., Schmidt, M. R., and Ranney, H. M. (1976) Dissociation of CO from carboxyhemoglobin. J. Biol. Chem. 251(14), 4267–4272. 27. Gibson, Q. H., Olson, J. S., McKinnie, R. E., and Rohlfs, R. J. (1986) A kinetic description of ligand binding to sperm whale myoglobin. J. Biol. Chem. 261(22), 10,228–10,239. 28. Olson, J. S., Rohlfs, R. J., and Gibson, Q. H. (1987) Ligand recombination to the alpha and beta subunits of human hemoglobin. J. Biol. Chem. 262(27), 12,930–12,938. 29. Sawicki, C. A. and Gibson, Q. H. (1976) Quaternary conformational changes in human hemoglobin studied by laser photolysis of carboxyhemoglobin. J. Biol. Chem. 251(6), 1533–1542. 30. Sawicki, C. A. and Gibson, Q. H. (1977) Properties of the T state of human oxyhemoglobin studies by laser photolysis. J. Biol. Chem. 252(21), 7538–7547. 31. Sawicki, C. A. and Gibson, Q. H. (1977) Quaternary conformational changes in human oxyhemoglobin studied by laser photolysis. J. Biol. Chem. 252(16), 5783–5788. 32. Olson, J. S., Andersen, M. E., and Gibson, Q. H. (1971) The dissociation of the first oxygen molecule from some mammalian oxyhemoglobins. J. Biol. Chem. 246(19), 5919–5923. 33. Reisberg, P. I. and Olson, J. S. (1980) Kinetic and cooperative mechanisms of ligand binding to hemoglobin. J. Biol. Chem. 255(9), 4159–4169. 34. Reisberg, P. I. and Olson, J. S. (1980) Rates of isonitrile binding to the isolated alpha and beta subunits of human hemoglobin. J. Biol. Chem. 255(9), 4151–4158. 35. Reisberg, P. I. and Olson, J. S. (1980) Equilibrium binding of alkyl isocyanides to human hemoglobin. J. Biol. Chem. 255(9), 4144–4150. 36. Mathews, A. J., Olson, J. S., Renaud, J. P., Tame, J., and Nagai, K. (1991) The assignment of carbon monoxide association rate constants to the alpha and beta subunits in native and mutant human deoxyhemoglobin tetramers. J. Biol. Chem. 266(32), 21,631–21,639. 37. Mathews, A. J., Rohlfs, R. J., Olson, J. S., Tame, J., Renaud, J. P., and Nagai, K. (1989) The effects of E7 and E11 mutations on the kinetics of ligand binding to R state human hemoglobin. J. Biol. Chem. 264(28), 16,573–16,583.

90

Olson et al.

38. Unzai, S., Eich, R., Shibayama, N., Olson, J. S., and Morimoto, H. (1998) Rate constants for O2 and CO binding to the alpha and beta subunits within the R and T states of human hemoglobin [in process citation]. J. Biol. Chem. 273(36), 23,150–23, 159. 39. Lemon, D. D., Nair, P. K., Boland, E. J., Olson, J. S., and Hellums, J. D. (1987) Physiological factors affecting O2 transport by hemoglobin in an in vitro capillary system. J. Appl. Physiol. 62(2), 798–806. 40. Page, T. C., Light, W. R., and Hellums, J. D. (1998) Prediction of microcirculatory oxygen transport by erythrocyte/hemoglobin solution mixtures. Microvasc. Res. 56(2), 113–126. 41. Eich, R. F., Li, T., Lemon, D. D., Doherty, D. H., Curry, S. R., Aitken, J. F., Mathews, A. J., Johnson, K. A., Smith, R. D., Phillips, G. N. Jr., and Olson, J. S. (1996) Mechanism of NO-induced oxidation of myoglobin and hemoglobin. Biochemistry 35(22), 6976–6983. 42. Sharma, V. S. and Ranney, H. M. (1978) The dissociation of NO from nitrosylhemoglobin. J. Biol. Chem. 253(18), 6467–6472. 43. Beckman, J. S. and Koppenol, W. H. (1996) Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and ugly. Am. J. Physiol. 271(5 Pt. 1), C1424–C1437. 44. Perutz, M. F., Kilmartin, J. V., Nagai, K., Szabo, A., and Simon, S. R. (1976) Influence of globin structures on the state of the heme. Ferrous low spin derivatives. Biochemistry 15(2), 378–387. 45. Hille, R., Olson, J. S., and Palmer, G. (1979) Spectral transitions of nitrosyl hemes during ligand binding to hemoglobin. J. Biol. Chem. 254(23), 12,110–12,120. 46. Hille, R., Palmer, G., and Olson, J. S. (1977) Chain equivalence in reaction of nitric oxide with hemoglobin. J. Biol. Chem. 252(1), 403–405. 47. Doyle, M. P., Pickering, R. A., DeWeert, T. M., Hoekstra, J. W., and Pater, D. (1981) Kinetics and mechanism of the oxidation of human deoxyhemoglobin by nitrites. J. Biol. Chem. 256(23), 12,393–12,398. 48. Wade, R. S. and Castro, C. E. (1996) Reactions of oxymyoglobin with NO, NO2, and NO2- under argon and in air. Chem. Res. Toxicol. 9(8), 1382–1390. 49. Stuehr, D. J. (1997) Structure-function aspects in the nitric oxide synthases. Annu. Rev. Pharmacol. Toxicol. 37, 339–359. 50. Gladwin, M. T., Ognibene, F. P., Pannell, L. K., Nichols, J. S., Pease-Fye, M. E., Shelhamer, J. H., and Schechter, A. N. (2000) Relative role of heme nitrosylation and beta-cysteine 93 nitrosation in the transport and metabolism of nitric oxide by hemoglobin in the human circulation. Proc. Natl. Acad. Sci. USA 97(18), 9943–9948. 51. Brunori, M. (2001) Nitric oxide moves myoglobin centre stage. Trends Biochem. Sci. 26(4), 209–210. 52. Brunori, M. (2001) Nitric oxide, cytochrome-c oxidase and myoglobin. Trends Biochem. Sci. 26(1), 21–23. 53. Thomas, D. D., Liu, X., Kantrow, S. P., and Lancaster, J. R. Jr. (2001) The biological lifetime of nitric oxide: implications for the perivascular dynamics of NO and O2. Proc. Natl. Acad. Sci. USA 98(1), 355–360.

Rate Constant Measurement for Reactions with Hb

91

54. Gardner, P. R., Gardner, A. M., Martin, L. A., Dou, Y., Li, T., Olson, J. S., Zhu, H., and Riggs, A. F. (2000) Nitric-oxide dioxygenase activity and function of flavohemoglobins. sensitivity to nitric oxide and carbon monoxide inhibition. J. Biol. Chem. 275(41), 31,581–31,587. 55. Gardner, P. R., Gardner, A. M., Martin, L. A., and Salzman, A. L. (1998) Nitric oxide dioxygenase: an enzymic function for flavohemoglobin. Proc. Natl. Acad. Sci. USA 95(18), 10,378–10,383. 56. Gardner, A. M., Martin, L. A., Gardner, P. R., Dou, Y., and Olson, J. S. (2000) Steady-state and transient kinetics of Escherichia coli nitric-oxide dioxygenase (flavohemoglobin). The B10 tyrosine hydroxyl is essential for dioxygen binding and catalysis. J. Biol. Chem. 275(17), 12,581–12,589. 57. Liu, X., Miller, M. J., Joshi, M. S., Sadowska-Krowicka, H., Clark, D. A., and Lancaster, J. R. Jr. (1998) Diffusion-limited reaction of free nitric oxide with erythrocytes. J. Biol. Chem. 273(30), 18,709–18,713. 58. Doherty, D. H., Doyle, M. P., Curry, S. R., Vali, R. J., Fattor, T. J., Olson, J. S., and Lemon, D. D. (1998) Rate of reaction with nitric oxide determines the hypertensive effect of cell-free hemoglobin [see comments]. Nat. Biotechnol. 16(7), 672–676. 59. Olson, J. S. (1994) Genetic engineering of myoglobin as a simple prototype for hemoglobin- based blood substitutes. Artif Cells Blood Substit Immobil Biotechnol 22(3), 429–441. 60. Olson, J. S., Eich, R. F., Smith, L. P., Warren, J. J., and Knowles, B. C. (1997) Protein engineering strategies for designing more stable hemoglobin-based blood substitutes. Artif. Cells Blood Substit Immobil. Biotechnol. 25(1–2), 227–241. 61. Tsai, C. H., Fang, T. Y., Ho, N. T., and Ho, C. (2000) Novel recombinant hemoglobin, rHb (beta N108Q), with low oxygen affinity, high cooperativity, and stability against autoxidation. Biochemistry 39(45), 13,719–13,729. 62. Herold, S. (1999) Kinetic and spectroscopic characterization of an intermediate peroxynitrite complex in the nitrogen monoxide induced oxidation of oxyhemoglobin. FEBS Lett. 443(1), 81–84. 63. Herold, S., Exner, M., and Nauser, T. (2001) Kinetic and mechanistic studies of the NO*-mediated oxidation of oxymyoglobin and oxyhemoglobin. Biochemistry 40(11), 3385–3395. 64. Herold, S. (1999) Mechanistic studies of the oxidation of pyridoxalated hemoglobin polyoxyethylene conjugate by nitrogen monoxide. Arch. Biochem. Biophys. 372(2), 393–398.

92

Olson et al.

Electrophoretic Methods for Study of Hbs

93

6 Electrophoretic Methods for Study of Hemoglobins Henri Wajcman 1. Introduction Electrophoresis, a technique consisting of the migration of electrically charged molecules under an applied electric field, occupies one of the most important places in the history of the study of hemoglobin (Hb). HbS, the first abnormal Hb described, was discovered in 1949 by Pauling et al. (1), using moving boundary electrophoresis. Later, Hb variants were detected by zone electrophoresis on paper, starch gel, or cellulose acetate (2,3). With the exception of cellulose acetate electrophoresis, which is still used in some laboratories, these procedures have been replaced by isoelectric focusing (IEF) (4). In IEF, a pH gradient is established by carrier ampholytes subjected to an electric current. The Hb molecule migrates across this gradient until it reaches the position where its net charge is zero (isoelectric point [pI]). It then concentrates into a sharp band. The most traditional and largely used methods of identifying and studying normal and mutant Hbs are the panoply of electrophoretic methods. This chapter describes the strengths and weakness of the most commonly used electrophoretic methods to separate Hbs. This electrophoretic approach needs to be assessed by other criteria, taking into accounts geographic and ethnic distribution as well as hematological and clinical presentation. In some cases additional tests such as biophysical or functional properties or mass spectrometry determinations may be required. For the last 20 yr in the Hemoglobin laboratory of Henri Mondor Hospital (Créteil, France) we have taken a multicriteria approach to the study of Hb, leading to a presumptive diagnosis for most of the known Hb variants. The electrophoretic mobility of more than 400 Hb variants is stored in a data bank under a format convenient for comparison based on an approach close to that From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

93

94

Wajcman

proposed some 20 yr ago by Barwick and Schneider (5). This strategy includes tests done on native Hbs (IEF on polyacrylamide gel, electrophoresis on cellulose acetate at alkaline pH, and citrate agar electrophoresis), and tests on dissociated globins (electrophoreses of globin chains in 6 M urea at pH 6.0 and 9.0 or in the presence of Triton X-100). The electrophoretic mobility of the variants is measured according to a quantified method that is described next. In our laboratory, a wide collection of identified rare variants is stored in liquid nitrogen and may be used, in a last step, as specific controls. 2. Materials and Methods 2.1. Isoelectric Focusing IEF studies of Hb may be done on polyacrylamide or agarose gels containing free ampholytes. These gels can be homemade (6–7), but it is more convenient to use IEF plates polymerized on support films, which are commercially available from several manufacturers (e.g., Perkin-Elmer, Norwalk, CT; Wallac, Akron, OH; Amersham Pharmacia Biotech, Uppsala, Sweden; Serva, Heidelberg, Germany). Polyacrylamide gels were the first ones used (6). Agarose gels suitable for IEF became only available later, when chemical treatments were developed to remove or mask the charged agaropectin residues present in the raw material. Agarose gels exhibit stronger electroendosmosis than polyacrylamide gels. Pores within this gel are also larger, making them more suitable for large proteins. In addition, agarose gels are also selected for routine work because they are not toxic and do not contain catalysts, which could interfere with the Hb molecule and thus lead to separation artifacts. Later, procedures were developed to cast polyacrylamide gels with covalently bound immobilized pH gradients. This technique allows the preparation of gel plates with pH ranges of