Hemoglobin Disorders. Molecular Methods and Protocols

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 <7.0, there is an increased possibility of the formation of methemoglobin (MetHb). Therefore, 1 mM EDTA is included in all the buffers to minimize this formation. 7. For analytical SEC runs, 800 µg of Hb (25 µL of a 0.5 mM solution) is used. At lower concentrations of Hb, attention should be paid to the tetramer-dimer equilibrium. In fact, Manning et al. (13) have described the utility of high-resolution SEC for the determination of dimer-tetramer equilibrium.

Acknowledgments This work was supported by National Institutes of Health Grants HL-38665, HL-55435, HL-58512, and HL-58247 and a grant-in-aid from the American Heart Association Heritage Affiliate.

Purification and Molecular Analysis of Hb by HPLC

47

References 1. Bunn, F. H. and Forget, B. G. (1986) Hemoglobin: Molecular, Genetic and Clinical Aspects, W. B. Saunders, Philadelphia, PA. 2. Adachi, K., Konitzer, P., Lai, C. H., Kim, J., and Surrey, S. (1992) Oxygen binding and other physical properties of human hemoglobin made in yeast. Prot. Eng. 5, 807–810. 3. Shen, T.-J., Ho, N. T., Simplaceanu, V., Zou, M., Green, B. M., Tam, B. F., and Ho, C. (1993) Production of unmodified human adult hemoglobin in Escherchia coli. Proc. Natl. Acad. Sci. USA 90, 8108–8112. 4. Shen T-J., Ho, N. T., Zou M, Sun D. P., Cottam, P. F., Simplaceanu V., Tam, M. F., Bell, J. D. A., and Ho, C. (1997) Production of human normal adult and fetal hemoglobins in Escherichia coli. Prot. Eng. 10, 1085–1097. 5. Manjula B. N., Kumar RA, Sun D.P, Ho, N. T., Ho, C., Rao, M. J., Malavalli, A., and Acharya, A. S. (1997) Correct assembly of human normal adult hemoglobin when expressed in trangenic swine: chemical, conformational and functional equivalence with the human-derived protein. Prot. Eng. 11, 583–588. 6. Rao, M. J. and Acharya, A. S. (1994) Amidation of basic carboxyl groups of hemoglobin. Methods Enzymol. 231, 246–267. 7. Perumalsamy, K., Manjula, B. N., Bookchin, R. M., and Acharya, A. S. (1998) Chemistry of the microenvironment of Glu-43(β) of deoxyhemoglobin S probed by amidation. Blood 92, 11a. 8. Macleod, R. M. and Hill, R. J. Demonstration of the hybrid hemoglobin Z A A S. (1970) J. Biol. Chem. 245, 4875–4879. 9. Manjula, B. N., Malavalli, A., Smith P. K., Chan, N.-L., Arnone, A., Friedman, J. M., and Acharya, A. S. (2000) Cys-93-ββ-succinimidophenyl polyethylene glycol 2000 hemoglobin A. J. Biol. Chem. 275, 5527–5534. 10. Schroeder, W. A., Shelton, J. B., Shelton, J. R., Huynh, V., and Teplov, D. B. (1985) Hemoglobin 9, 461–482. 11. 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. 12. Rao, M. J., Schneider, K., Chait, B. C., Chao, T. L., Keller, H. L., Anderson, S. M., Manjula, B. N., Kumar, R. A., and Acharya, A. S. (1994) Recombinant hemoglobin A produced in transgenic swine: structure equivalence with human hemoglobin A. Artif. Cells, Blood Substitutes Immobil. Biotechnol. 22, 695–700. 13. Manning, L. R., Jenkins, W. T., Hess, J. R., Vandegriff, K. D., Winslow, R. M., and Manning J. M. (1996) Subunit dissociations in natural and recombinant hemoglobins. Protein Sci. 5, 775–781.

48

Manjula and Acharya

O2 Equilibrium Measurements of RBCs

49

4 Oxygen Equilibrium Measurements of Human Red Blood Cells Jean Kister and Henri Wajcman 1. Introduction 1.1. When and Why Blood Oxygen Affinity Should be Measured From a medical viewpoint, the aim of blood oxygen affinity measurement is to determine whether the oxygen binding properties of red blood cells (RBCs) are normal or not. If abnormal, there may be several reasons. One is an abnormality of the hemoglobin (Hb) itself. A second, more frequent reason, is a metabolic defect (or status) leading to change in the intraerythrocytic concentration of 2,3-diphosphoglycerate (DPG). Some toxic mechanisms may also lead to compounds with modified oxygen-binding properties such as methemoglobin (MetHb) or carbonmonoxyhemoglobin. Interpretation of the results therefore requires one to measure the DPG content of the RBCs at the time of the test, as well as the MetHb percentage in a fresh lysate. It is also important to have information concerning the smoking habits of the patient. Only 10–15% of the 800 Hb variants that have been identified to date (http:// globin.cse.psu.edu/cgi-bin/hbvar/counter) display altered oxygen-binding properties. Variants with a clear increase in oxygen affinity lead to polycythemia, while those with a clear decrease to cyanosis and anemia. When, in addition, they are unstable, they lead to chronic hemolytic anemia. Several methods have been proposed to measure the oxygen affinity of RBCs. They are all based on the determination of the fraction of Hb saturated in oxygen, which is measured spectrophotometrically, at various oxygen partial pressures (PO2). In tonometric methods, the RBCs are first deoxygenated (usually by vacuum) and then equilibrated under known PO2; this procedure leads to point-by-point measurements, going from deoxy to oxy state. In the From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

49

50

Kister and Wajcman

continuous methods, usually the RBC suspension (or blood) is fully oxygenated first and then slowly deoxygenated. Continuous methods are of obvious interest for obtaining precise oxygen equilibrium curves (OECs) covering the full PO2 range. From a basic scientific viewpoint, the measurement of oxygen binding in RBC suspensions is the easiest way to evaluate the functional properties of Hb in a physiological range of concentration. Under these conditions, Hb occurs almost only as tetramers, the proportion of the various hybrid molecules present in the cells is not modified, and for some fragile variants the reducing enzymatic systems prevent MetHb formation.

1.2. Discontinuous vs Continuous Methods Measuring the oxygen-binding properties of RBCs requires one to obtain simultaneously an accurate measurement of the PO2 to which the cells are exposed and of their content in oxyhemoglobin (HbO2). When the variation in oxygen pressure is discontinuous, as achieved in a tonometer equipped with an optical cell, the PO2 of the gas mixture in which the RBC suspension is equilibrated is perfectly known, being determined by the proportion of the different gases used. The percentage of HbO2 is measured by comparing, at each point of PO2, the optical spectrum of the suspension with that of the same sample, fully deoxygenated and fully oxygenated. Isobestic points are checked to control the stability of the sample during the assay. The RBC solution needs to be diluted enough to be kept within the range of sensitivity of the spectrophotometer. In addition, the spectrophotometer should be modified to solve the problem of turbidity and light scattering. Studies done by such a point-by-point procedure are time-consuming and may cause alteration of the sample during the assay. Historically, methods of this type were used first. The ancient gasometric method using the Van Slyke and Neill (1) apparatus was a classic way to determine OECs. However, this procedure was so laborious and time-consuming that only a few points could be obtained in one day. Static methods using tonometry and spectrophotometry (2) were the most widely used in many laboratories before the use of oxygen electrodes. These procedures were simple and only a small amount of sample was needed. A variety of other techniques have been used for determination of the OEC of red cell suspensions, following many methods: gasometric, spectrophotometric, polarographic and spectrophotometric-polarographic combinations. Today, automatic methods are available in which the variation in PO2 and HbO2 are simultaneously and continuously recorded. Two different strategies are proposed. In the first one, which is used for the HEM-O-SCAN (Aminco) apparatus, measurements are done on thin layers of blood with experimental parameters designed to be close to “physiological” conditions. In the second

O2 Equilibrium Measurements of RBCs

51

strategy, which has been chosen for the HEMOX ANALYSER (TCS, Southhampton, PA) (3), blood, or washed RBCs, is suspended in a buffer, contained in the optical cell, which is bubbled by the gas mixture to change progressively the PO2. This second technique is more suitable for biological studies, because environmental conditions (e.g., pH, ionic strength, drugs) can be easily modified.

1.3. Oxygen-Binding Parameters That Can Be Determined in RBC Oxygen affinity is characterized by the P50. This parameter corresponds to the PO2 at which the Hb is half saturated. It depends on several factors, including type of Hb, temperature, pH (Bohr effect), and intracellular DPG concentration. From the experimental plot of the HbO2 percentage vs PO 2, the interaction between heme groups, known as the Hill coefficient, or n50, may be obtained. The Bohr effect can be calculated by comparing the log P50 values of OECs determined at various pH values. A rough estimation of the DPG effect can be obtained by comparing the P50 of fully DPG-depleted RBCs with that of the same sample at a known DPG content. OECs obtained from RBCs may be fitted to allosteric parameters, but in the highly complex environment of Hb in the RBCs their meaning is doubtful. 2. Continuous Methods 2.1. Thin-Layer Method: HEM-O-SCAN

2.1.1. Principle In the thin-layer method, a few microliters of blood is applied on a circular microscope slide and covered with a gas-permeable membrane. This sample is introduced into a chamber where, at 37°C, it is submitted to a gas stream, which modifies the PO2, and thus the level of HbO2. A Clark electrode monitors the PO2 in this chamber, while a photomultiplicator determines the percentage of HbO2 by measuring the variations in absorption of a monochromatic light. Curves showing the percentage of HbO2 vs PO2 are drawn on an X-Y recorder. Both axes need to be calibrated before the run: PO2 by air (or gas mixtures) and nitrogen, and HbO2 by the sample fully deoxygenated and oxygenated. Gas mixtures containing CO2 are used to buffer the sample, thus mimicking physiological conditions. This method gives satisfactory results for routine measurements of P50.

2.1.2. Preparation of Sample Blood samples are collected under EDTA-glucose medium and kept at 4°C until assayed, ideally within less than 24 h after venipuncture. If necessary, the blood sample can be shipped by express mail in wet ice, but it should never be frozen. For these samples, it is compulsory to have a normal control sent with them.

52

Kister and Wajcman

2.1.3. Experimental Procedure The sample is transferred to a cover glass and a gas-permeable membrane is placed over the sample. The resulting thin film minimizes the time required for oxygen to be diffused effectively throughout the sample. The sample is placed on a holder at the sample compartment. The holder allows the sample to be loaded, tested, and removed without opening the sample compartment. Initial deoxygenation is accomplished by purging the sample compartment with nitrogen (N2). The oxygen equilibrium curve of Hb is then plotted by exposing the sample to varying PO2s and measuring the response of the sample. The resulting graph of fraction HbO2 as a function of PO2 is a fundamental description of the oxygen transporting capacity of Hb. The N2 purge and PO2 are controlled by a gas delivery system that uses two compressed gas cylinders with regulators, solenoid valves, and calibrated flow restrictors. Both gas cylinders contain 5.6% CO2 in order to maintain the carbon dioxide partial pressure (PCO2) at 40 mmHg. Uniform PO2 is maintained in the sample compartment by a stirring device. The sample compartment contains a water jacket that must be connected to an external constant-temperature bath to obtain the required temperatures. Humidity control is obtained with two gas-conditioning wells and a wick reservoir built into the sample compartment block. The wells contain foam disks saturated with distilled water that condition the incoming N2 and O2 gases to the proper humidity levels. The wick reservoir provides distilled water to a wick located near the sample holder, which ensures that the sample is maintained near 100% humidity during a test run. The PO2 of the sample compartment is monitored by a Clark O2 electrode, and the fraction HbO2 of the sample is monitored by a dual-wavelength spectrophotometric system. The O2 electrode system consists of an O2 electrode and an amplifier that is connected to the X-axis of the recorder. The gain and offset of the amplifier are controlled by the oxygen electrode controls on the top panel. The oxygen electrode calibrate control is adjusted while the sample compartment is completely filled with humidified, thermostated gas from the oxygen cylinder, which has a PO2 of 178.2 mmHg at 760 mmHg atmospheric pressure (37°C, water saturated). The oxygen electrode zero control is adjusted while the sample compartment is completely filled with N2. The use of a true N2 zero instead of an electrical zero cancels out nonspecific electrode current, which would otherwise limit the accuracy of measurements at low PO2. The spectrophotometric system monitors the sample with a light beam from a tungsten-halogen lamp. This light beam passes through the sample to a beam splitter, where it is directed to two separate photodetectors. Each photodetector

O2 Equilibrium Measurements of RBCs

53

has a different interference filter in front of it. One filter transmits a wavelength of 560 nm; the other, 576 nm. The difference in absorbance between the two wavelengths changes substantially when Hb undergoes an oxy-deoxy transition. The two wavelengths are chosen close together to minimize error owing to light-scattering changes. The photodetector outputs are processed by a log ratio amplifier that generates a signal proportional to the difference in sample absorbance at the two wavelengths. This signal is applied to an amplifier whose gain and offset are controlled by the HbO2 controls on the top panel. The output of this amplifier is applied to the Y-axis input of the X-Y recorder.

2.1.4. Limits of the Method The main difficulties encountered with the HEM-O-SCAN are some inaccuracy in the assessment of the percentage of HbO2 for the fully saturated sample and a slight variation in the intraerythrocytic pH during the oxygenation procedure. It is commonly assumed that Hb is completely oxygen saturated by air under normal atmospheric pressure (PO2 = 160 mmHg) at 37°C, but because of the oxygen-binding reaction, Hb within the RBCs is not completely saturated. This means that a slight but systematic error is made assuming that the 100% oxygen saturation level is reached under air pressure. This error resulting from incomplete saturation can be quite large in the case of lowoxygen-affinity samples (4). Another problem is the pH change accompanying oxygenation when CO2-containing gases are used, as for experimental conditions of human blood “standards” (such as PCO2 of 40 mmHg, pH 7.4, 37°C).

2.2. Cell Suspension Method: HEMOX ANALYSER 2.2.1. Principles The cell suspension method is suitable for biological and clinical purposes (3). It consists of the simultaneous recording of PO2 changes during slow deoxygenation using a Clark-type oxygen electrode (polarographic method) and the changes in absorbance of the Hb solution with a double-wavelength sprectrophotometer (sprectrophotometric method) (Fig. 1). In principle, this technique is similar to that described by Imai et al. (5) for the study of Hb solution. The advantage of the HEMOX ANALYSER is the simultaneous measurement of the optical densities at two different wavelengths, which eliminates the light-scattering problem. The OEC is measured in a buffered solution, which maintains the pH constant. Under such experimental conditions, the OEC represents the true oxygen-binding graph. Data obtained by this procedure may be treated in the framework of the general binding polynomial model described by Wyman (6). Therefore, this technique is more appropriate for precise biochemical studies.

54

Kister and Wajcman

Fig. 1. Schematic representation of HEMOX ANALYSER.

The HEMOX ANALYSER may be interfaced to a microcomputer to save each run by sampling 300–500 data points of HbO2 and PO2 values. These data may be further used for automatic calculation of the other parameters for oxygen binding. Since the measurement of OEC by the HEMOX ANALYSER is now the most widely used technique, we present here this method in more detail.

2.2.2. Preparation of Sample Blood samples should be collected and shipped as described in Subheading 2.1.2. After collection, the blood is centrifuged for 5 min at 350g at 4°C to remove the plasma and buffy coat. The packed erythrocytes are washed three times in cold 50 mM Bis-Tris isotonic buffer, pH 7.4. An aliquot of the packed cells (30–50 µL) is suspended in 4 mL of the working buffer solution in the HEMOX cuvet. Theoretically, measurements can be directly done on blood, but in this case, errors may result from hemolytic samples or from samples in which aggregated RBCs cause irregular turbidity.

2.2.3. Experimental Procedure for Determination of OEC Standard oxygen affinity measurement of an RBC suspension is performed in 50 mM Bis-Tris, 140 mM NaCl, pH 7.4, buffer contained in an optical cell. The cell also contains the Clark electrode and a magnetic bar for rapid and even stirring of the sample suspension during measurement. Because of this rapid stirring, the HbO2 content (optical signal) and the PO2 are measured simultaneously at the same location. A special cuvet stopper contains the gas exchange tubing system through which the gas exchange takes place. Temperature is maintained at 37 ± 0.01°C using a water jacket connected to a large temperature-controlled water bath.

O2 Equilibrium Measurements of RBCs

55

Usually, the analysis consists only of performing a deoxygenation curve. Nevertheless, when required, it is possible to do a reoxygenation curve on the same sample; in a normal sample these two curves should be superimposed. Before recording a deoxygenation curve, the red cell suspension is equilibrated for 10–20 min under oxygen (about 600–700 mmHg of PO2) in the optical cuvet of the HEMOX ANALYSER. This procedure ensures that full oxygen saturation of the sample is achieved, and allows the corresponding absorbance value to be memorized by the machine. This value is used by the software to determine the HbO2 scale. The deoxygenation process is initiated with a gas mixture made of 20% O2 in N2, which is similar to the air pressure conditions. When the PO2 value decreases to about 300 mmHg, the deoxygenation is continued with pure nitrogen to obtain a PO2 value near zero. Each deoxygenation recording usually takes about 40–45 min (7). To avoid sample concentration via dehydration, the gas mixtures are water saturated by bubbling through a small bottle of distillated water. Using the reverse procedure, a reoxygenation curve can be recorded following the deoxygenation curve. After 10–20 min of equilibration under pure N2, the 20% oxygen gas mixture is very slowly introduced to reach a P O2 value of about 70 mmHg. Pure oxygen is then used to complete oxygenation to the maximum PO 2 value desired (monitored by computer software). The reoxygenation procedure is usually faster than deoxygenation, requiring 25–30 min. The HEMOX ANALYSER uses the principle of differential absorbance measurement using a dual-wavelength spectrophotometer that requires only one cell. Through this cuvet, the light of two different wavelengths is passed: a “measuring” wavelength of 560 nm for the maximum absorbance of the deoxyhemoglobin; and a “reference” wavelength of 569 nm for the isobestic point between oxy and deoxyhemoglobin spectra, which remains practically unchanged during the deoxygenation process. Thus, the change in the optical property of Hb during deoxygenation is detected by the electronic circuitry as the differential optical change between these two wavelengths. A special, balanced amplifier circuit provides a very high signal-to-noise ratio for recording extremely small changes. Since the differential extinction coefficient of oxygenated and deoxygenated blood is known, it is also possible to determine the Hb concentration of the sample.

2.2.4. Representations and Interpretation of Data Oxygen-binding curves may be represented in different ways. The same data are displayed according to the linear representation in Fig. 2, as the Hill plot in Fig. 3, and as the cooperativity curve in Fig. 4.

56

Kister and Wajcman

Fig. 2. Linear representation of normal fresh red cell suspension OEC with P50 value of 26 mmHg and arteriovenous difference in oxygen saturation (∆Y) of 20%. Other conditions: pH 7.4, 0.05 M bis-Tris, 0.14 M NaCl, 37°C.

Fig. 3. Hill plot representation of deoxygenation and reoxygenation OEC from normal RBC suspension. The two curves are superimposable. Other conditions: pH 7.4, bis-Tris 0.05 M, NaCl 0.14 M, 37°C.

O2 Equilibrium Measurements of RBCs

57

Fig. 4. Oxygen cooperativity representation for normal fresh red cell suspension. Other conditions: pH 7.4, 0.05 M bis-Tris, 0.14 M NaCl, 37°C.

2.2.4.1. LINEAR REPRESENTATION

From the PO2 and absorbance values stored in the computer, the fractional oxygen saturation (Y) vs PO2 curve (linear representation) is directly drawn (Fig. 2). The P50 and n50 values are calculated by linear regression analysis from the experimental points obtained between 40 and 60% oxygen saturation (7). The P50 value represents the oxygen affinity of the RBC suspension in the experimental conditions used. In “standard” conditions (i.e., pH 7.4, 50 mM Bis-Tris, 140 mM NaCl, 37°C), the P50 value from a normal control is 27 ± 0.1 mmHg and the n50 value is 2.6 ± 0.1. A lower P50 means that the oxygen affinity is increased while a higher P50 corresponds to a decreased oxygen affinity. This plot also allows determination of the oxygen-carrying capacity of the blood. This is done by measuring under the standard conditions the difference between the arterial (SAO2) and the mixed venous (SVO2) oxygen saturation, which, on the OEC, corresponds to the saturation level at 90 and 40 mmHg of PO2, respectively. In a normal RBC suspension, the arteriovenous difference in oxygen saturation is about 20–25%. From this value, it is possible to obtain the cardiac output parameter (Q) using the Fick equation: VO2 = 0.136 × Q × Hb × (SAO2 – SVO2)

in which VO 2 is the amount of oxygen released per minute (L/min), Q is the blood flow (L/min), Hb is the patient’s hemoglobin concentration, and (S AO 2 – S VO 2) is the arteriovenous difference of oxygen saturation as already indicated.

58

Kister and Wajcman

2.2.4.2. HILL PLOT REPRESENTATION AND INTERPRETATION

As an alternative, from the data stored in the computer, the OEC may be represented according to the empirical Hill equation as log[Y/(1 – Y)] vs log(PO2) (8). According to this representation, known as the Hill plot, the OEC is a sigmoid. This curve has two asymptotes with slopes equal to unity: the lower one corresponds to the deoxygenated structure (T state in the MonodWyman-Changeux model) (9), and the upper one to the fully oxygenated structure (R state) (Fig. 3). This representation may be clearer than the decimal one to visualize the presence of an abnormal Hb component in the RBCs, as shown in a few typical examples (Fig. 6). 2.2.4.3. COOPERATIVITY REPRESENTATION AND INTERPRETATION

A very powerful method to analyze the cooperative properties of Hb is to represent the variations in the Hill coefficient, nH, vs oxygen saturation, expressed as log[Y/(1 – Y)] (7). This is done by calculating the first derivative of the Hill plot as d log[Y/(1 – Y)]/d log (PO2) by linear regression analysis of the i + 3/i – 3 values for each point of the OEC curve (Fig. 4). The resulting plot, called the cooperativity curve, is classically bell shaped. It starts from an nH of 1, at very low PO2, reaches a maximum nH (nmax) around the half-saturation level (nmax similar to the n50 for a “symmetrical” OEC), and ends with an nH near 1 at high PO2. This representation may also be quite useful to visualize the presence of an abnormal Hb component with altered cooperativity. A normal fresh RBC suspension “cooperativity” curve is shown in Fig. 4. This curve exhibits a large asymmetrical aspect with the nmax (2.8) obtained at a high oxygen level (about 90%) and with a lower n50 value near P50 (2.5). This behavior may be explained by variations in the activity of free DPG during the oxygenation process that are owing to the equimolar concentration of DPG and tetrameric Hb within the normal fresh red cells (7). The fact that the apparent heme-heme interaction (Hill coefficient) is increasing in the physiologically important portion of the red cell OEC (between 90 and 95% oxygen saturation level) may improve oxygen delivery to the tissues. This cooperativity curve could present dramatic perturbations when the DPG concentration is below the normal value; under such nonsaturating concentrations of DPG in red cells, the OEC exhibits a clear biphasic shape with a low value of n50 (7). This is observed during blood storage with aging of erythrocytes and acidosis or in vivo in some pathological conditions.

2.3. Bohr Effect in Case of Red Cell Suspensions The Bohr effect is an important functional property of human Hb. Its physiological role can be described either as the pH dependence of the oxygen affinity or as the pH of the solvent on oxygenation of deoxyhemoglobin.

O2 Equilibrium Measurements of RBCs

59

Fig. 5. Variation in log P50 vs pH: The alkaline Bohr coefficient (–∆H+max) is the maximal slope of this curve.

The Bohr effect parameters are calculated from the variation in the log P50, or log Pm, vs pH (“Bohr curve,” Fig. 5) using the thermodynamic framework of the linked-function theory of Wyman (6): This theory states that the binding of oxygen to tetrameric Hb and the binding of protons are reciprocal so that ∆log P50/∆pH = – ∆H+ / mole heme

in which P50 is the PO2 at half saturation, similar to the theoretical parameter of Pm, median PO2, if the OEC is not too far from the conditions of symmetry: ∆H+ difference in the number of bound protons released per heme. Thus, the Bohr effect can be studied by the measurements of OEC at different pH values between 5.5 and 8.5. Measurements are done in 140 mM NaCl, 50 mM bis-Tris buffer for pH values <7.5 and in Tris buffer for values >7.5. For normal fresh red cell suspensions at 37°C, the alkaline Bohr effect coefficient (–∆H+max) is about 0.66 per heme (10).

2.4. Variation in Erythrocytic DPG Content A red cell OEC has no meaning without the knowledge of the intraerythrocytic DPG content. It can be measured using a commercial kit from Boehringer Mannheim. It is possible to study the effect of varying concentrations of DPG on the OEC. To obtain a complete DPG depletion, washed red cells are incubated at 37°C in an isotonic buffer in the absence of glucose for about 20–24 h. The total absence of DPG results in a twofold increased oxygen affinity and a high value of the Hill coefficient (n50 = 2.8). This procedure may be used to verify whether an increased oxygen affinity observed in the case of a sample with unknown DPG content is owing to an abnormal Hb.

60

Kister and Wajcman

Table 1 Oxygen-Binding Properties of DPG-Depleted and Normal Fresh RBCsa P50 normal RBCs

Experimental condition

P50 (torr)

n50

P50 DPG-depleted RBCs

Fresh red cells ([DPG/Hb4] ~ 1) DPG-depleted red cells ([DPG/Hb4] ~ 0) Increased DPG red cells ([DPG/Hb4] ~ 2)b

26.6

2.5

—

13.5

2.8

2.0

39.0

2.8

0.7

a Other

conditions: pH 7.4, 0.05 M bis-Tris, 0.14 M NaCl, 37°C. in PIGPA buffer (2 h at 37°C, pH 7.8).

b Incubated

Conversely, a two- to four-fold enrichment of the intracellular DPG content can be obtained by a 2- to 6-h incubation of the RBCs in a Krebs-Ringer (pH 7.8) buffer containing 5 mM phosphate (dihydrogeno-Na), 5 mM inosine, 10 mM glucose, 5 mM pyruvate, and 0.5 mM adenine (PIGPA buffer) (11). In RBCs containing a [DPG/Hb4] ratio of about 2, the oxygen affinity is halved and the Hill coefficient is high (n50 = 2.8). In both of these conditions, the cooperativity curve is symmetrical with the maximum Hill coefficient close to the n50 value (Table 1). This results from the fact that there is only one functional Hb species (Hb alone for DPG-depleted RBCs or DPG-bound Hb for RBCs with a large DPG excess relative to Hb) in opposition to fresh red cells where the two functional Hb populations are present (7).

2.5. Allosteric Modifiers of RBC Oxygen Affinity Regulating the allosteric equilibrium of Hb has been of interest in medicine. Pharmaceutical agents that produce a high-affinity Hb have been clinically evaluated as antisickling agents, while those that produce a low-affinity Hb may be useful for the treatment of thalassemia and ischemic problems arising from stroke and cardiovascular diseases (12). It has been demonstrated that bezafibrate, an antilipidemic drug, lowers the oxygen affinity of red cells and Hb solutions (13,14). Later, it was reported that a bezafibrate derivative, RSR-4 or [2-[4-[[(3,5-dimethylalanilino) carbonyl]methyl]-phenoxy]-2-methylpropionic acid], was much more effective in lowering the oxygen affinity of suspensions of fresh intact cells (15). This compound is the most potent allosteric modifier discovered to date that shifts the oxygen equilibrium curve to the right in whole blood and in vivo. These compounds cross the red cell membrane and bind mostly the α-chains of

O2 Equilibrium Measurements of RBCs

61

Fig. 6. Hill plot representation of OECs of RBC suspensions from a normal patient (curve 1), a heterozygous patient with a high-oxygen-affinity Hb variant (curve 3), and a heterozygous patient with a low-oxygen-affinity Hb variant (curve 2). Other conditions: pH 7.4, 0.05 M bis-Tris, 0.14 M NaCl, 37°C.

deoxy-Hb, thus stabilizing the T deoxy state. Since DPG binds to a different site, it acts in a synergetic way when combined with these allosteric effectors. 3. A Few Examples The following sections discuss some typical OEC found in various clinical situations.

3.1. Patient with High-Oxygen-Affinity Hb (Polycythemia) For patients heterozygous for a high-oxygen-affinity Hb variant, the red cell suspension OEC (curve 3) is displaced toward the left in comparison with the normal OEC (curve 1), as illustrated in Fig. 6. In some cases, the Hill plot of this OEC can be biphasic. The lower portion with a low n value (low or noncooperative Hb) corresponds to the high-oxygen-affinity Hb while the upper portion that approaches the normal curve reflects the oxygenation of the normal HbA that coexists in the RBCs. It results in a marked defect in oxygen extraction with erythrocytosis (16). The biphasic aspect of the Hill plot can be more or less difficult to recognize when the abnormal Hb displays only a moderate increase in oxygen affinity and remains cooperative.

62

Kister and Wajcman

3.2. Patient with Low-Oxygen-Affinity Hb (Cyanosis) For patients heterozygous for a low-oxygen-affinity Hb variant, the red cell suspension OEC (curve 2) is displaced toward the right in comparison with the normal OEC (curve 1), as illustrated in Fig. 6. It results in a marked defect in oxygen saturation with cyanosis (17).

3.3. Case of Patient with Sickle Cell Hb (Sickle Cell Disease) In homozygous sickle cells red cell suspensions, evidence for HbS polymerization can be provided by the hysteresis of the OEC between the deoxygenation and reoxygenation curves (Fig. 7A). In the reoxygenation experiment, the amount of polymerized Hb at the beginning of the recording is larger than in the deoxygenation curve, which results in a further shift to the right of the curve (18,19). Additional evidence for the presence of these polymers can be observed in the cooperativity profiles (Fig. 7B). For homozygous SS red cells, an increased nmax is observed, which can be higher than the theoretical limit for normal Hb cooperative behavior (nmax = 3), indicating that an aggregation process is present during the oxygenation. (18,19). References 1. Van Slyke, D. D. and Neill, J. M. (1924) The determination of gases in blood and other solutions by vacuum extraction and manometric measurement. J. Biol. Chem. 61, 523–573. 2. Benesch, R., Macduff, G., and Benesch, R. E. (1965) Determination of oxygen equilibria with a versatile new tonometer. Anal. Biochem. 11, 81–87. 3. Asakura, T. (1979) Automated method for determination of oxygen equilibrium curves of red cell suspensions under controlled buffer conditions and its clinical applications. Crit. Care Med. 7, 391–395. 4. Marden, M. C., Kister, J., Poyart, C., and Edelstein, S. J. (1989) Analysis of hemoglobin oxygen equilibrium curves: are unique solutions possible? J. Mol. Biol. 208, 341–345. 5. Imai, K., Morimoto, H., Kotani, M., Watari, H., Hirata, W., and Kuroda, M. (1970) Studies on the function of adnormal hemoglobins. I. An improved method for automatic measurement of the oxygen equilibrium curves of hemoglobin. Biochim. Biophys. Acta 200, 189–196. 6. Wyman, J. (1964) Linked functions and reciprocal effects in hemoglobin: a second look. Adv. Prot. Chem. 19, 223–286. 7. Kister, J., Poyart, C., and Edelstein, S. J. (1987) An expanded two-state allosteric model for interaction of human hemoglobin A with non saturating concentrations of 2,3 diphosphoglycerate. J. Biol. Chem. 262, 12,085–12,091. 8. Hill, A. V. (1910) The possible effects of the aggregation of the molecules of hemoglobin on the dissociation curves. J. Physiol. (Lond.) 40, iv–vii.

O2 Equilibrium Measurements of RBCs

63

Fig. 7. (A) Linear representation of fresh homozygous sickle red cell suspension OECs: deoxygenation (curve 1) and reoxygenation (curve 2) OECs. Other conditions: pH 7.4, 0.05 M bis-Tris, 0.14 M NaCl, 37°C. (B) Oxygen cooperativity representation of fresh homozygous sickle red cell suspension OECs: deoxygenation (curve 1) and reoxygenation (curve 2) OECs. Other conditions: pH 7.4, 0.05 M bis-Tris, 0.14 M NaCl, 37°C. 9. Monod, J., Wyman, J., and Changeux, J. P. (1965) On the nature of the allosteric transitions: a plausible model. J. Mol. Biol. 12, 88–118. 10. Kister, J., Marden, M. C., Bohn, B., and Poyart, C. (1988) Functional properties of hemoglobin in human red cells: II. Determination of the Bohr effect. Respir. Physiol. 73, 363–378. 11. Lian, C. Y., Roth, S., and Harkness, D. R. (1971) The effect of alteration of intracellular 2,3-DPG concentration upon oxygen binding of intact erythrocytes

64

12.

13. 14.

15.

16. 17.

18.

19.

Kister and Wajcman containing normal or mutant hemoglobins. Biochem. Biophys. Res. Commun. 45, 151–158. Poyart, C., Marden, M. C., and Kister, J. (1994) Bezafibrate derivatives as potent effectors of hemoglobin, in Methods in Enzymology, vol. 232, Hemoglobins. Part C: Biophysical Methods (Everse, J., Vandegriff, K. D., and Winslow, R. M., eds.), Academic, San Diego, pp. 496–513. Perutz, M. F. and Poyart, C. (1983) Bezafibrate lowers oxygen affinity of haemoglobin. Lancet 2, 881, 882. Perutz, M. F., Fermi, G., Abraham, D. J., Poyart, C., and Bursaux, E. (1986) Hemoglobin as a receptor of drugs and peptides: X-ray studies of the stereochemistry of binding. J. Am. Chem. Soc. 108, 1064–1078. Abraham, D. J., Wireko, F. C., Randad, R. S., Poyart, C., Kister, J., Bohn, B., Liard, J. F., and Kunert, M. P. (1992) Allosteric modifiers of hemoglobin: 2-[4[[(3,5-disubstitutedanilino)carbonyl]methyl]phenoxyl]-2-methylpropionic acid derivatives that lower the oxygen affinity of hemoglobin in red cell suspensions, in whole blood, and in vivo in rats. Biochemistry 31, 9141–9149. Wajcman, H. and Galacteros, F. (1996) Abnormal hemoglobins with high oxygen affinity and erythrocytosis. Hematol. Cell Ther. 38, 305–312. Griffon, N., Badens, C., Lena-Russo, D., Kister, J., Bardakdjian, J., Wajcman, H., Marden, M. C., and Poyart, C. (1996) Hb Bruxelles, deletion of Pheβ42, shows a low oxygen affinity and low cooperativity of ligand binding. J. Biol. Chem. 271, 25,916–25,920. Poyart, C., Edelstein, S., Kister, J. and Bohn, B. (1986) Oxygen binding by sickle red cells, in Approaches to the Therapy of Sickle Cell Anaemia, Colloque INSERM, vol. 141 (Beuzard, Y., Charache, S., and Galacteros, F., eds.), Les Editions INSERM, Paris, pp. 67–87. Cohen-Solal, M., Préhu, C., Wajcman, H., Poyart, C., Bardakdjian-Michau, J., Kister, J., Promé, D., Valentin, C., Bachir, D., and Galacteros, F. (1998) A new sickle cell disease phenotype associating Hb S trait, severe pyruvate kinase deficiency (PK Conakry), and an α2 globin gene variant (Hb Conakry). Br. J. Haematol. 103, 950–956.

Rate Constant Measurement for Reactions with Hb

65

5 Measurement of Rate Constants for Reactions of O2, CO, and NO with Hemoglobin John S. Olson, Erin W. Foley, David H. Maillett, and Eden V. Paster 1. Introduction The kinetics of O2, CO, and NO binding to mammalian hemoglobins (Hbs) have been studied for 75 yr, starting with the original rapid mixing experiments of Hartridge and Roughton (1). Over the last 20 yr, these measurements have been extended to time scales ranging from hours to picoseconds. Numerous articles have been written about rapid mixing and photolysis instruments, methods for defining specific association and dissociation rate constants, and algorithms for analyzing the results in terms of specific models for cooperative ligand binding (2–10). A comprehensive review of these techniques and methods, however, is beyond the scope of this book. Instead, a practical guide to determining rate constants for O2, CO, and NO binding to native and recombinant Hbs is presented, with a special emphasis on tetrameric adult human Hb (HbA). First, the basic kinetic expressions for reversible ligand binding to hemecontaining subunits are presented. Second, the differences in reactivity of O2, CO, and NO with heme iron are discussed in terms of the time resolution required for direct measurements of association and dissociation. Third, typical rapid mixing and flash photolysis experiments for O2, CO, and NO binding to human Hb are described. The complications caused by cooperative ligand binding are discussed for each ligand, and approaches to assigning rate parameters for the R and T quaternary states are summarized. Fourth, techniques for measuring the oxidative reaction of NO with oxyhemoglobin (HbO2) are described in view of the importance of this reaction for in vivo NO detoxification and vasoregulation. From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

65

66

Olson et al.

2. Methods 2.1. Simple Kinetic Analysis For simple monomeric proteins, the basic ligand binding mechanism and rate equations are as follows: Hb + X

k'X kX

HbX;

d[Hb] = –k'X[X][Hb] + kX[HbX] dt

(1)

in which Hb represents a heme-containing monomer; k'X and kX are bimolecular association and unimolecular dissociation rate constants; and [X] and [HbX] are the concentrations of free and bound ligand, respectively. To simplify data analysis, the ratio of the initial concentrations of ligand, Xtotal, and heme, Hbtotal, is usually manipulated to maintain pseudo first-order conditions in one of the reactants. Normally, Xtotal is kept ≥10 Hbtotal so that simple exponential time courses are obtained: that is, [Hb]t = [Hb]total exp(–kobst). Under these conditions, the pseudo first-order rate constant, kobs, is given by the sum of the forward and backward rates: kobs = k'XX0 + kX. The values of the association (k'X) and dissociation (kX) rate constants are determined as the slope and y-axis intercept, respectively, from a plot of kobs vs ligand concentration (see Fig. 1). Under some circumstances, it is necessary to keep the protein in excess. If Hbtotal > 10Xtotal, then the observed pseudo first-order rate is given by kobs = k'X[Hb]total + kX. If the concentration of heme groups approaches that of the ligand, then nonexponential, second-order time courses will be observed and more complex analyses are required (see refs. 5 and 6, and Fig. 6). The ligand dissociation rate constant (kx) is usually determined more directly in ligand replacement or consumption reactions. The general schemes for these reactions are as follows: HbX

kX X + Hb + Y k'X

k'Y kY

HbY or HbX

k'X k'X

Hb + X

Consume X with dithionite or other reagents

(2)

In the replacement reactions, the original liganded complex, HbX, is mixed with excess displacing ligand, Y. The displacing ligand is chosen to have much higher affinity for Hb than the bound ligand, X, and to have a much smaller dissociation rate constant (i.e., kY < kX). Under these conditions, the observed replacement rate constant, robs, is given by robs =

kXk'Y[Y]

(3)

(k'Y[Y] + k'X[X])

If the rate of association of Y is much greater than that for the rebinding of X (i.e., k'Y[Y] >> 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 5–10 500–1000 ~0.01 O2 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 <1 pH unit, thus leading to a much higher resolution (8,9). To increase the separation of some Hb fractions, such as HbF or glycated Hbs, some investigators have proposed that the pH gradient in the region where these Hbs migrate be flattened (10). This is achieved by adding, when preparing the gel, separators such as ε-amino caproic acid, or β-alanine.

2.1.1. IEF on Agarose Gel IEF on agarose gels is presently the screening method of choice for both Hbs in cord blood and adult blood samples. Together with cation-exchange high-performance liquid chromatography, it should be among the first tests to be used when an Hb disorder is suspected. IEF is usually performed using a commercially available agarose plate containing a mixture of ampholytes in the pH 6.0–8.0 range. IEF is run on a Multiphor II electrophoresis unit (Amersham Pharmacia Biotech), or on any

Electrophoretic Methods for Study of Hbs

95

similar equipment. The gel is carefully placed on the cooling plate refrigerated at 10°C. On each lane, 3–5 µL of hemolysate containing 10–15% Hb is applied, using a sample application template. This Hb sample may be either eluted from a blood spot collected on dry paper or obtained by hemolysis of whole blood or washed erythrocytes. When the Hb concentration is higher than 15%, the volume used needs to be reduced in proportion. A wick properly wetted with an acid solution is positioned at the anode and another one with a basic solution at the cathode. Couples of solutions generally used are 1 M H3PO4/1 M NaOH or 0.5 M acetic acid/0.5 M ethanolamine. IEF is run with a voltage limited at 1600 V. To avoid formation of spurious bands resulting from the oxidation of oxyHb into methemoglobin (MetHb), the sample may be prepared in a medium containing 0.1% KCN, which will convert the MetHb into cyanMetHb, which displays the same charge as oxyHb. Another possibility is to have the KCN included at a concentration of 0.1% in the catholyte, which would avoid the formation of MetHb during the run. Hb bands are visualized by using a staining system containing o-dianisidine, which is readily oxidized by heme in the presence of hydrogen peroxide. The reaction also forms an insoluble precipitate that intensifies each band and avoids diffusion. Manuals provided by the manufacturer of IEF gels usually detail all the experimental procedures to follow.

2.1.2. IEF on Polyacrylamide Gels A higher resolution of Hb fractions is obtained by using polyacrylamide gels instead of agarose gels, but this is only done as a further step in the investigation of an Hb variant. This IEF procedure allows one to distinguish variants having pIs differing by 0.02 pH unit (11). With the advent of immobilized pH gradients, it became possible to separate Hb species differing by 0.001 pH unit. Variants with such small differences in their pIs usually result from structural changes consecutive to an exchange between amino acids identically charged but differing in their size. In some cases, this could lead to slight modifications in the exposure of charged groups, which are totally or partially buried in the native oxy structure. In our laboratory, we prefer to use homemade polyacrylamide, which permits us to work with a specially designed pH gradient. The actual procedure is as follows: 1. For a 260 × 125 × 0.5 mm plate, in a 50-mL Erlenmeyer flask, add 4.5 mL of 2% acrylamide (19.4 g/dL of acrylamide, 0.6 g/dL of bis-acrylamide, prepared no more than 2 wk earlier and stored in a dark bottle); 0.9 mL of ampholine, pH 6.0– 8.0; and 1.05 mL of ampholine, pH 7.0–9.0, to 11.7 mL of distilled water.

96

Wajcman

2. To prevent formation of MetHb, add 12 mg of KCN to the gel. 3. Degas the solution for 8 min under vacuum and add 5 to 6 µL of TEMED (Sigma Aldrich, Saint Quentin Fallavier, France). 4. Add 250 µL of a freshly prepared 3% ammonium persulfate (Merck, Darmstadt, Germany) solution. 5. Promptly pour this mixture with a 10-mL pipet into a cassette made from a 3-mm glass plate on which is placed a 1-mm glass plate, a 0.5-mm-thick gasket, and a Plexiglas plate, all maintained by 12 LKB clamps. Polymerization takes 2 h at room temperature (see Note 1). Conditions of migration in these gels are almost identical to the previously described (1600 V, 2 h).

2.2. Citrate Agar Electrophoresis In a second step of investigation, all the mutants found through IEF are subjected to citrate agar electrophoresis. This method was introduced for Hb study some 40 yr ago and represents a combination of electrophoresis and chromatography. It is not primarily sensitive to the charge of the mutated residue but to structural modifications of positively charged regions of the Hb molecule interacting with the agaropectin matrix contained in the gel (12). These regions of the molecule are usually those involved in Hb/ion interaction. Position β6, where the structural abnormality of HbS is located, is close to one of these regions and therefore leads to highly specific profiles when modified. The buffer used for this type of electrophoresis is made from 50 mM sodium citrate adjusted to pH 5.8 by a few drops of a 30% citric acid solution. Under a current of 60 mA and a voltage of about 60–100 V, electrophoresis is done at 4°C for 45–60 min. Agar plates are available from several manufacturers (Helena Laboratories, Beaumont, TX) or may be homemade. In the Paragon kit (Beckman), maleate buffer is used instead of citrate buffer. Migration is measured according to a normalized scale in which HbA = 0, HbF = –4.4, and HbC = 10 (see Note 2).

2.3. Cellulose Acetate Electrophoresis at Alkaline pH Cellulose acetate of electrophoresis has a resolution clearly inferior to IEF, but because of its simplicity it remains among the more popular techniques for Hb screening. Several kits are commercially available (among others, Helena Laboratories). In this technique, Hbs are separated according to their charge at pH 8.5 and, to a lesser extent, according to the position of the modified residue in the molecule or to its molecular environment. At this pH, the Hb molecule carries an overall negative charge. Hence, the Hb molecule migrates toward the cathode. Nevertheless, two variants carrying the same amino acid exchange may display large differences in their mobility according to the position of the modified residue toward the exterior of the molecule. Residues, involved in

Electrophoretic Methods for Study of Hbs

97

internal contacts and buried inside the molecule, will affect to a lesser degree the mobility of the protein than residues exposed toward the surrounding water. Electrophoresis is run in Tris (85 mM)-EDTA (2 mM)-borate (50 mM), pH 8.5, buffer. Using a cellulose acetate plate (Titan II-H; Helena), the electrophoresis is done in 20 min at room temperature under 420 V. Proteins are stained with amidoblack, and after being washed with a 5% acetic acid solution, the plates may be dried in methanol or made transparent (see Note 3).

2.4. Electrophoresis of Globin Chains in 6 M Urea at pH 6.0 and 9.0 Treatment of the Hb molecule by high concentrations of urea and β-mercaptoethanol solubilizes the heme group and dissociates the globin chains. Analysis of the globin chains by electrophoresis on a cellulose acetate strip in 6 M urea in Tris-EDTA buffer at pH 9.0 and 6.0 confirms an exchange between residues with different charges. Some patterns of modifications in the electrophoretic behavior of the globin between these two pHs suggest particular types of substitutions such as those involving a histidine residue. This electrophoretic separation may be achieved on cellulose acetate plates. Revelation is done by amidoblack staining (see Note 4). This solution is used for destaining. Migration of the globin chains in urea is estimated from a scale in which α-chain is +10 and β-chain +20 (5). Known variants with mobility similar to that of the unknown sample should be selected as controls for accurate measurements (see Note 5).

2.5. Electrophoresis of Globin in Presence of Triton X-100 This method reveals amino acid exchanges involving alkaline residues and, more interestingly, modifying the hydrophobicity of the polypeptide chain. Triton, a nonionic detergent, masks some charged residues. This is done by a vertical polyacrylamide gel electrophoresis (PAGE) in the presence of 8 M urea with migration in a 5% acetic acid solution according to the technique of Alter et al. (13). A 160 × 210 × 1.5 mm plate of polyacrylamide gel is used. 1. Make the gel as follows: a. Prepare 10 mL of a solution of 60% acrylamide and 0.4% bis-acrylamide. b. Add 37.5 mL of an 8 M urea solution (24.0 g/31.5 mL of water), 2.8 mL of acetic acid, and 1 mL of Triton X-100. c. Degas the mixture for 10 min under vacuum. d. Add 250 µL of TEMED and 300 µL of ammonium persulfate. e. Pour the mixture between two glass plates and place a comb. Cover the top of this with aluminum foil. f. Polymerize for 1 h at room temperature. When polymerized, remove the comb. 2. Fill both compartments of the electrophoresis tank are filled with 5% acetic acid, and conduct a 1-h prerun at 200 V.

98

Wajcman

3. Empty the sampling wells and refill with a 1 M cysteamine solution (227 mg of cysteamine/2 mL of water). Conduct a second 1-h electrophoretic prerun in 5% acetic acid at 150 V. 4. Prepare the sample just before applying on the plate. It is made by adding 5 µL of hemolysate (0.2–0.3 g% Hb) to 10 µL of a dissociating solution (5 mL of 8 M urea, 0.5 mL of β-mercaptoethanol, and 0.5 mL of concentrated acetic acid). 5. Carefully wash the sampling wells with 5% acetic acid and apply the sample (15 µL). Perform migration at room temperature for 18–20 h. 6. At the end of the migration, carefully remove the gel and stain with a solution of Coomassie R-250 blue for 1 h (see Note 6). Then immerse the gel at room temperature in a destaining solution for several days (see Note 7). 7. Calculate chain mobility as proposed by Barwick and Schneider (5) assuming a value of 10 for a normal α-chain and 20 for a normal β-chain.

2.6. Scaled-Up Electrophoretic Methods for Hb Separations For structural or functional studies, milligram amounts of Hb need to be separated. This can be conveniently done by flat-bed preparative IEF as described by Radola (14). We describe here a modified miniaturized technique. 1. Make the gel by a suspension of 0.9 g Ultrodex (Pharmacia) in 24 mL of a 1.7% solution of ampholine, pH 6.0–8.0, (Pharmacia) in water. 2. Pout this mixture into a 26.0 × 2.5 × 0.4 cm homemade Plexiglas cassette and submit to an air stream to reduce the gel weight by 30%. 3. Apply the hemolysate (500 µL of a 4–6% Hb solution) on the gel at 8 cm from the cathode within a 1 × 1 cm metallic mold and carefully remove the mold when the sample has penetrated the gel. 4. Perform IF overnight at 4°C at 1300 V and 2 W. 5. At the end of the focusing, collect the various Hb components with a spatula and elute from the gel with water in small columns.

3. Notes 1. It is difficult to determine with precision by this technique the pI of an Hb variant. It can’t, however, be done, with reasonable approximation, by considering as reference values the isoelectric point that have been experimentally measured for a few Hb markers such as HbA, HbF, MetHb, HbS (β6 [A3] Glu → Val), and HbC (β6[A3] Glu → Lys), which are 6.95, 7.15, 7.20, 7.25, and 7.40, respectively (11). Because the pH gradient is linear, it is possible to express the mobility of any variant by its distance in millimeters from HbA according to some normalized scale. Such a scale has been obtained in our laboratory by using a few selected Hbs. The precision of the measurement is increased when the unknown sample migrates, in the same lane, together with a control sample having a mobility close to it. As an example, under these experimental conditions, values for HbA, HbD Punjab (β121 [GH4] Glu → Gln), HbS, HbE (β6 [B8] Glu → Lys), and

Electrophoretic Methods for Study of Hbs

2.

3.

4.

5.

6.

7.

99

HbC are 0, –7.6, –8.5, –15.1, and –16 (with a confidence limit of ± 0.2), respectively. Further practical data concerning IEF and PAGE are detailed in ref. 15. In this type of electrophoresis, good reproducibility may be difficult to obtain from one run to run. Several factors, such as slight differences in the ionic strength or pH of the electrophoretic buffer, may be responsible. Lack of reproducibility may also result from storage conditions of the agar plate, or from variation in the volume of the sample applied and its concentration in Hb. The amount of agaropectin present in the gel may also vary from one batch to another; the highest amount is found in crude agar, whereas all the agaropectin is removed from the agarose gels used for IEF. In our laboratory, all mutants are analyzed by cellulose acetate electrophoresis at alkaline pH in addition to the other methods. Migration in cellulose acetate electrophoresis is measured according to a comparative scale using as markers HbC = –10, HbA = 0, and HbI Texas (a16 [A14] Lys → Glu) = +10 (5). The main limitation of this technique is its low resolving power: it is impossible to discriminate variants that have almost the same mobility as, e.g., HbS, HbD Punjab, or HbG Ferrara [b57 [E1] Asn → Lys), which could be distinguished by IEF. Staining is obtained by soaking the band in a 0.5g/100 mL solution of amidoblack made up of 1000 mL of methanol, 250 mL of acetic acid, and 1250 mL of distilled water. This solution is also used for destaining. A scaled-up technique of chain separation on cellulose acetate electrophoresis may be used for biosynthetic studies after incubating the reticulocytes with 3H-leucine (16). Staining solution is obtained by dissolving 0.5 g of Coomassie Blue R-250 in 30 mL of methanol under magnetic stirring. Then 7 mL of acetic acid is added to this solution, and its final volume is completed to 100 mL by adding distilled water. Destaining solution is made from methanol/acetic acid/water in the proportion 30/7/63 (v/v/v).

References 1. Pauling, L., Itano, H. A., Singer, S. J., and Wells, I. C. (1949) Sickle cell anemia, a molecular disease. Science 110, 543. 2. Huisman, T. H. J. and Jonxis, J. H. P. (1977) The Hemoglobinopathies Techniques of Identification, Clinical and Biochemical Analysis, vol. 6, Marcel Dekker, New York. 3. Huisman, T. H. J. (1986) Introduction and review of standard methodology for the detection of hemoglobin abnormalities, in The Hemoglobinopathies, Methods in Hematology, vol. 15 (Huisman, T. H. J., ed.), Churchill Livingstone, Edinburgh. 4. Righetti, P. G., Gianaza, E., Bianchi-Bosisio, A., and Cossu, G. (1986) Conventional isoelectric focusing and immobilized pH gradients for hemoglobin separation and identification, in The Hemoglobinopathies, Methods in Hematology, vol. 15 (Huisman, T. H. J., ed.), Churchill Livingstone, Edinburgh, p. 47. 5. Barwick, R. C. and Schneider, R. G. (1980–1981) The computer-assisted differentiation of hemoglobin variants, in Human Hemoglobins and Hemoglobinopathies:

100

6.

7. 8.

9.

10.

11.

12. 13.

14. 15. 16.

Wajcman

A Review to 1981. Texas Reports on Biology and Medicine, vol 40. (Schneider, R. G., Charache, S., and Schroeder, W. A., eds.), University of Texas Medical Branch, Galveston, TX, pp. 143–156. Basset, P., Beuzard, Y., Garrel, M. C., and Rosa, J. (1978) Isoelectric focusing of human hemoglobin: its application to screening, to the characterization of 70 variants, and to the study of modified fractions of normal hemoglobins. Blood 51, 971–982. Righetti, P. G. (1983) Isoelectric Focusing: Theory, Methodology and Applications, Elsevier, Amsterdam. Bjellqvist, B., Ek, K., Righetti, P. G., Gianazza, E., Gorg, A., Westermeier, R., and Postel, W. (1982) Isoelectric focusing in immobilized pH gradients: principle, methodology and some applications. J. Biochem. Biophys. Methods 6, 317–339. Righetti, P. G., Gianazza, E., Bianchi-Bosisio, A., Wajcman, H., and Cossu, G. (1989) Electrophoretically silent hemoglobin mutants as revealed by isoelectric focusing in immobilized pH gradients. Electrophoresis 10, 595–599 Cossu, G., Manca, M., Pirastru, M. G., Bullitta, R., Bianchi-Bosisio, A., Gianazza, E., and Righetti, P. G. (1982) Neonatal screening of β-thalassemias by thin layer isoelectric focusing. Am. J. Hematol. 13, 149–157. Drysdale, J. W., Righetti, P., and Bunn, H. F. (1971) The separation of human and animal hemoglobins by isoelectric focusing in polyacrylamide gel. Biochim. Biophys. Acta. 229, 42–50. Schneider, R. G. and Barwick, R. C. (1982) Hemoglobin mobility in citrate agar electrophoresis: its relationship to anion binding. Hemoglobin 6, 199–208. Alter, B. P., Goff, S. C., Efremov, G. D., Gravely, M. E., and Huisman, T. H. J. (1980) Globin chain electrophoresing: a new approach to the determination of the G/A ratio in fetal haemoglobin and to the studies of globin synthesis Br. J. Haematol. 44, 527–534. Radola, B. J. (1973) Analytical and preparative isoelectric focusing in gel-stabilized layers. N Y Acad. Sci. 209, 127–143. Westermeier, R. (1997) Electrophoresis in Practice, 2nd ed., VCH Verlagsgesellschaft, Weinheim, Germany. Harano, T., Ueda, S., Harano, K., and Shibata, S. (1980) Improved method for quantitation of biosynthesized human globin chains in reticulocytes by use of urea cellulose acetate membrane electrophoresis. Proc. Jap. Acad. 56(B), 230–234.

DNA Diagnosis of Hb Mutations

101

7 DNA Diagnosis of Hemoglobin Mutations John M. Old 1. Introduction The hemoglobinopathies are a diverse group of inherited recessive disorders that include the thalassemias and sickle cell disease. They were the first genetic diseases to be characterized at the molecular level and, consequently, have been used as a prototype for the development of new techniques of mutation detection. There are now many different polymerase chain reaction (PCR)– based techniques that can be used to diagnose the globin gene mutations, including dot-blot analysis, reverse dot-blot analysis, the amplification refractory mutation system (ARMS), denaturing gradient gel electrophoresis (DGGE), mutagenically separated PCR, gap-PCR, and restriction endonuclease analysis (1,2). Each method has its advantages and disadvantages, and the particular one chosen by a laboratory to diagnose point mutations depends not only on the technical expertise available in the diagnostic laboratory but also on the type and variety of the mutations likely to be encountered in the individuals being screened.

1.1. Diagnostic Approaches The main diagnostic approaches for the PCR diagnosis of the hemoglobinopathies are provided in Table 1. The ones well used in my laboratory are gapPCR, ARMS-PCR, and restriction endonuclease analysis of amplified product. Detailed protocols for each of these techniques are presented in this chapter. The alternative well-used method for the diagnosis of hemoglobin (Hb) mutations, that of allele-specific oligonucleotide (ASO) hybridization by dot blotting or reverse dot blotting, is not covered in this chapter, but detailed protocols may be found elsewhere (3). From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

101

102

Old Table 1 DNA Diagnosis of Hemoglobinopathies Disorder and mutation type

Diagnostic methoda

αo-Thalassemia α+-Thalassemia Deletion Nondeletion β-Thalassemia Deletion Nondeletion δβ-Thalassemia HPFH Deletion Nondeletion Hb Lepore HbS HbC HbE HbD Punjab HbO Arab Hb variants

Gap-PCR, Southern blotting Gap-PCR, Southern blotting ASO, RE, DGGE Gap-PCR ASO, ARMS, DGGE Gap-PCR Gap-PCR ASO, ARMS, RE, DGGE Gap-PCR ASO, ARMS, RE ASO, ARMS ASO, ARMS, RE ASO, ARMS, RE ASO, ARMS, RE RT-PCR and DNA sequencing

a Gap-PCR, gap polymerase chain reaction; ASO, allele-specific oligonucleotide; RE, restriction endonuclease; DGGE, denaturing gradient gel electrophoresis; RT-PCR, reverse transcriptase polymerase chain reaction.

1.2. α-Thalassemia Gap-PCR provides a quick diagnostic test for α+-thalassemia and αo-thalassemia deletion mutations but requires careful application for prenatal diagnosis. Most of the common α-thalassemia alleles that result from gene deletions can be diagnosed by gap-PCR. Primer sequences have now been published for the diagnosis five αo-thalassemia deletions and two α+-thalassemia deletions (4–7), as given in Table 2. The αo-thalassemia deletions diagnosable by PCR are the – –SEA allele, found in Southeast Asian individuals; the – –MED and –(α)20.5 alleles, found in Mediterranean individuals; the – –FIL allele, found in Fillipino individuals; and, finally, the – –THAI allele, found in Thai individuals. The two α+-thalassemia deletion mutations are the 3.7- and the 4.2-kb single α-gene deletion mutations, designated -α3.7 and -α4.2, respectively. Amplification of sequences in the α-globin gene cluster is technically more difficult than that of the β-globin gene cluster, requiring more stringent conditions for success owing to the higher GC content of the break-point sequences

DNA Diagnosis of Hb Mutations

103

Table 2 Thalassemia Deletion Mutations Diagnosed by Gap-PCR Disorder αo-Thalassemia

α+-Thalassemia βo-Thalassemia

(δβ)o-Thalassemia

(Aγδβ)o-Thalassemia HPFH

Deletion mutation –SEA

– – –MED -(α)20.5 – –FIL – –THAI -α3.7 -α4.2 290-bp deletion 532-bp deletion 619-bp deletion 1393-bp deletion 1605-bp deletion 3.5-kb deletion 10.3-kb deletion 45-kb deletion Hb Lepore Spanish Sicilian Vietnamese Macedonian/Turkish Indian Chinese HPFH1 (African) HPFH2 (Ghanaian) HPFH3 (Indian)

Reference 4 4 4 6,7 6,7 5 5 22 23 24 25 26 27 28 29 18 18 18 18 18 18 18 18 18 18

and the considerable sequence homology within the α-globin gene cluster. Experience in many laboratories has shown some primer pairs to be unreliable, resulting occasionally in unpredictable reaction failure and the problem of allele dropout. However, the more recently published primer sequences (6,7) seem to be much more robust than the earlier ones, possibly owing to the addition of betaine to the reaction mixture. They are also designed for a multiplex screening test, although in my laboratory they are still used in pairs to test for individual mutations. Figures 1 and 2 show, respectively, the results of screening for the five common αo and the two common α+-thalassemia mutations with these more robust primers. The other αo and α+-thalassemia mutations cannot be diagnosed by PCR because their break-point sequences have not been determined. These deletion

104

Old

Fig. 1. Ethidium bromide (EtBr)–stained gel showing screening of DNA samples for the five common αo-thalassemia mutations by gap-PCR. All the primers used are from ref. 6. The fragment sizes are as follows: normal allele (αα), 1010 bp; – –SEA, 660 bp; – –THAI, 411 bp; – –FIL, 550 bp; – –MED; 875 bp; -(α)20.5, 1180 bp.

mutations are diagnosed by the Southern blotting technique using ζ-gene and α-gene probes. This approach is still very useful because it permits the diagnosis of α-thalassemia deletions and α-gene rearrangements (the triple and quadruple α-gene alleles) in a single test (8). The characteristic abnormal fragments used in my laboratory for the diagnosis of the more common α-thalassemia deletions are given in Table 3. α+-Thalassemia is also caused by point mutations in one of the two α-globin genes. These nondeletion alleles can be detected by PCR using a technique of selective amplification of each α-globin gene followed by a general method of mutation analysis such as DGGE (9) or DNA sequence analysis (10). Several of the nondeletion mutations alter a restriction enzyme site and may be diagnosed by selective amplification and restriction endonuclease analysis in a manner similar to that reported for the mutation that gives rise to the unstable α-globin chain variant Hb Constant Spring (11).

1.3. β-Thalassemia The β-thalassemia disorders are a very heterogeneous group of defects with more than 170 different mutations characterized to date (12). The majority of the defects are single-nucleotide substitutions, insertions, or deletions. Only 13 large gene deletions have been identified, and eight of these can be diagnosed

DNA Diagnosis of Hb Mutations

105

Fig. 2. EtBr-stained gel showing the screening of DNA samples for the two common α+-thalassemia mutations by gap-PCR. The primers used to detect the 3.7-kb deletion are from ref. 1, and the primers used for the 4.2-kb deletion are from ref. 6. The fragment sizes are as follows: for the -α3.7 allele:normal allele (αα), 1800 bp and -α3.7, 2020 bp; for the -α4.2 allele:normal allele (αα), 1510 bp and -α4.2, 1725 bp.

by gap-PCR, as listed in Table 2. The other types of mutations can be diagnosed by a variety of methodologies, but the strategy for identifying β-thalassemia mutations remains the same. The mutations are regionally specific, and each at-risk population has a few common mutations together with a larger variable number of rarer ones. The strategy depends on knowing the spectrum of common and rare mutations in the ethnic group of the individual being screened. The common ones are analyzed first using a PCR method designed to detect specific mutations simultaneously. This approach will identify the mutation in more than 80% of cases for most ethnic groups. Further screening of the known rare mutations will identify the defect in another 10–15% of cases, if necessary. Mutations remaining unidentified at this stage are characterized by DNA sequencing. The first PCR diagnostic method to be developed and gain widespread use was the hybridization of ASO probes to amplified DNA bound to nylon membrane

106

Old

Table 3 The Diagnosis of α-Thalassemia Alleles by Southern Blotting Restriction enzyme/gene probea Allele

BamHI/α

αα

14

ααα

18

αααα

22

-α3.7

10.3

-α4.2

9.8

BglII/α

BamHI/ζ

BglII/ζ

SstI/LO

12.6 7.4 12.6 7.4 12.6 7.4 16

10–11.3 5.9 10–11.3 5.9 10–11.3 5.9 10–11.3 5.9 10–11.3 5.9 5.9 5.9 20 5.9 5.9 5.9 None None

12.6 or 5.2 10–11.3 12.6 or 16 10–11.3 12.6 or 20 10–11.3 16 10–11.3 10–11.3 8.0 10.8 13.9 10.5

5.0

5.0 5.0 5.0

7.0 7.5 None None

5.0 5.0 8.0 7.4

-(α)20.5 – –MED – –SEA

4.0 None None

8.0 7.4 10.8 None None

– –SA – –BRIT – –THAI – –FIL

None None None None

None None None None

a Fragment

5.0 5.0 5.0 5.0

sizes are given in kilobase pairs. Characteristic abnormal fragments are underlined.

by dot blotting (13). Although still in use, the method is limited by the need for separate hybridization steps to test for multiple mutations. This problem was been overcome by the development of the reverse dot-blotting technique, in which amplified DNA is hybridized to a panel of mutation-specific probes fixed to a nylon strip. This technique is compatible with the optimum strategy for screening β-thalassemia mutations, using a panel of the commonly found mutations for the first screening and a panel of rare ones for the second screening (14). The technique used by my laboratory and described here is the ARMS. This technique fits the main requirements of a PCR technology—i.e., speed, cost, convenience, and the ability to test for multiple mutations simultaneously while providing a screening method without any form of labeling of primers or amplified DNA. The simplest way is to screen for mutations with simultaneous PCR assays although the multiplexing of ARMS primers in a single PCR assay is possible (15). Figure 3 shows the results of screening a patient’s DNA sample for the seven common Mediterranean β-thalassemia mutations. The most widely used indirect method to characterize β-thalassemia mutations is DGGE (16). The technique detects at least 90% of β-thalassemia muta-

DNA Diagnosis of Hb Mutations

107

Fig. 3. EtBr-stained gel showing screening of a DNA sample for seven common Mediterranean β-thalassemia mutations by ARMS-PCR. The seven mutations are as follows: lanes 1 and 2, IVSI-110(G → A); lanes 3 and 4, IVSI-1(G → A); lanes 5 and 6, IVSI-6(T → C); lanes 7 and 8, codon 39(C → T); lanes 9 and 10, codon 6 (-A); lanes 11 and 12, IVSII-1 (G → A); lanes 13 and 14, IVSII-745(C → G). The gel shows the amplification products from the DNA of a β-thalassaemia heterozygote (odd-numbered lanes) and products generated by control DNAs (even-numbered lanes). The results show that the patient carries the mutation IVSI-110(G → A). For mutations 1–6, the control primers D and E produced an 861-bp fragment. However, for mutation 7, IVSII-1(G → A), the HindIII/Gγ-gene restriction fragment length polymorphism (RFLP) primers were used, giving a control band of 326 bp. The primers used are given in Table 3.

tions by a shifted band pattern to normal and provides an alternative approach to ASO probes or ARMS in countries where a very large spectrum of β-thalassemia mutations occur (17).

1.4. δβ-Thalassemia and Hereditary Persistence of Fetal Hb δβ-Thalassemia and the hereditary persistence of fetal Hb (HPFH) disorders result from large gene deletions affecting both the β- and δ-globin genes. Restriction enzyme mapping has enabled the characterization of more than 50 different deletions starting at different points between the Gγ gene and the δ gene and extending up to 100 kb downstream of the β-globin gene. In two cases, the Macedonian/Turkish (δβ)o-thalassemia gene and the Indian (Aγδβ)o-thalassemia gene, the mutation is a complex rearrangement consisting of an inverted DNA sequence flanked by two deletions. A small number of these deletions have had their break-point sequences characterized, and these can be diagnosed by gap-PCR (18). Gap-PCR can also be used for the diagnosis of Hb Lepore, created by a deletion of the DNA sequence between the δ- and β-globin genes. Hb

108

Old

Lepore is the product of the δβ fusion gene, and is associated with a severe β-thalassemia phenotype. All the deletion mutations currently diagnosable by gap-PCR are provided in Table 2; the others can only be diagnosed by the identification of characteristic break-point fragments with Southern blot analysis.

1.5. Hb Variants More than 700 Hb variants have been described to date, most of which were identified by protein analysis and have never been characterized at the DNA level. Positive identification at the DNA level is achieved by selective globin gene amplification and DNA sequence analysis. However, the clinically important variants, HbS, HbC, HbE, HbD Punjab, and HbO Arab, can be diagnosed by simpler DNA analysis techniques. Sickle cell disease is caused by homozygosity for HbS and also, in varying degrees of severity, from interaction of HbS with HbC, HbD Punjab, HbO Arab, and β-thalassemia trait. All these variants can be diagnosed by ASO hybridization; the ARMS technique; or, for all except HbC, restriction endonuclease digestion of amplified product. The sickle cell gene mutation abolishes a DdeI recognition site at codon 6, and diagnosis by DdeI digestion of amplified product remains the simplest method of DNA analysis for sickle cell disease. Similarly, the mutations giving rise to HbD Punjab and HbO Arab abolish an EcoRI site at codon 121. However, the HbC mutation at codon 6 does not abolish the DdeI site and is diagnosed by other methods. HbE interacts with β-thalassemia trait to produce a clinical disorder of varying severity ranging from thalassemia intermedia to transfusiondependent thalassemia major. The HbE mutation can be diagnosed by ASO hybridization, ARMS, or restriction endonuclease analysis since the mutation abolishes an MnlI site in the β-globin gene sequence. 2. Materials 1. For ARMS PCR the buffer used is the standard Cetus buffer: 50 mM KCl, 10 mM Tris-HCl (pH 8.3 at room temperature), 1.5 mM MgCl2, 100 µg/mL of gelatin. A 10X stock buffer can be prepared by adding a mixture of 0.5 mL of 1 M Tris-HCl (pH 8.3 at room temperature), 1.25 mL of 2 M KCl, 75 µL of 1 M MgCl2, 5 mg of gelatin, and 3.275 mL of distilled water. The stock buffer is heated at 37°C until the gelatin dissolves and then is frozen in aliquots. 2. For gap-PCR the reaction buffer (10X) recommended for the particular pair of primers being used is required. The buffer for α-thalassemia primers also includes 0.5 M betaine. 3. Stock deoxynucleotide mixture containing each dNTP at 1.25 mM: Mix together 50 µL of a 100 mM solution of each dNTP (as purchased) and 3.8 mL of distilled water. The 1.25 mM dNTP stock solution should be stored in frozen aliquots. 4. PCR reaction stock solution (4 mL): This is made of 0.5 mL of 10X buffer, 0.8 mL of 1.25 mM dNTP stock solution, and 2.7 mL of distilled water.

DNA Diagnosis of Hb Mutations

109

5. Dilute aliquots of primer stock solutions to make a working solution of 1 OD U/mL and store frozen. 6. Taq polymerase: AmpliTaq Gold (PE Biosystems) is the best.

3. Methods (see Notes 1–6) 3.1. Gap-PCR (see Note 1) Gene deletion mutations in the β-globin gene cluster may be detected by PCR using two primers complementary to the sense and antisense strand in the DNA regions that flank the deletion. For small deletions of <1 kb in size, the primer pair will generate two products, the smaller fragment arising from the deletion allele. For large deletions, the distance between the two flanking primers is too great to amplify the normal allele, and product is only obtained from the deletion allele. In these cases, the normal allele is detected by amplifying across one of the break points, using a primer complementary to the deleted sequence and one complementary to the flanking DNA. 1. Set up the reaction mixture to a final volume of 22 µL in a 0.5-mL tube with the following components as required: 1 µL of genomic DNA (100 ng/µL), 1 µL of forward primer—flanking sequence (10 pmol/µL), 1 µL of reverse primer—flanking sequence (10 pmol/µL), 1 µL of primer—deleted sequence (10 pmol/µL), 1 µL of primer—inverted sequence (10 pmol/µL), 2.5 µL of 1.25 mM (dNTP mixture), 2.3 µL of the reaction buffer (10X) as recommended for the primers, and sterile dH2O to a final volume of 22 µL. For the α-thalassemia primers that also contain 0.5 M betaine, add 2.5 µL of 5 M betaine (Sigma, St. louis, MO). 2. Overlay with 25 µL of mineral oil. 3. Prepare the enzyme mixture: 0.2 µL of reaction buffer (10X), 0.1 µL of Ampli Taq (5U/µL) (PE Biostems) for the β-gene primers, 0.1 µL of Platinum Taq (5U/µL) (Life Technologies) for the α-gene primers, and 2.7 µL of sterile dH2O to a final volume of 3 µL. 4. Mix the enzyme mixture and hold on ice. 5. Place the reaction mixture in a thermal cycler and perform one cycle as follows, adding 3 µL of the enzyme mix after 2 min of the 94°C denaturation step: 4 min at 94°C, 1 min at 55°–65°C (as recommended), 1.5 min at 72°C. 6. Continue for 33 cycles with the following steps per cycle: 1 min at 94°C, 1 min at 55°–65°C (as recommended), 1.5 min at 72°C. 7. Finish with one cycle as follow: 1 min at 94°C, 1 min at 55°–65°C (as recommended), 10 min at 72°C. 8. Hold at 15°C until gel electrophoresis. 9. Remove the tubes from the thermal cycler and add 5 µL of blue dye (15% Ficoll/ 0.05% bromophenol blue). Mix and centrifuge. 10. Load a 20-µL aliquot onto a 1–3% agarose gel (depending on expected fragment sizes) and run at 100 V for 45 min in Tris-borate-EDTA (TBE) buffer (89 mM Tris, 89 mM boric acid, 10 mM EDTA, pH 8.0).

110

Old

11. Stain gel in EtBr solution (0.5 µg/mL) for 15–30 min, visualize bands on an ultraviolet light box (312 nm), and photograph with an electronic camera system or a Polaroid CU-5 camera fitted with an orange filter (e.g., Wratten 22A).

3.2. ARMS PCR (see Note 4) Newton et al. (19) first described the ARMS technique for detecting known point mutations. It has been developed for the diagnosis of all the common β-thalassemia mutations and many of the rare ones (8,20). The technique is based on the principle of allele-specific priming of the PCR process; that is, a specific primer will permit amplification to take place only when its 3' terminal nucleotide matches with its target sequence. Thus, to detect the β-thalassemia mutation IVSI-5(G → C), the 3' nucleotide of the ARMS primer is G in order to base pair with the substituted C in the mutant DNA. The primer forms a G-G mismatch with normal DNA, but this is a weak mismatch and will not prohibit extension of the primer by itself. Only strong mismatches (C-C, G-A, and A-A) reduce priming efficiency to zero or below 5% (21). To prevent amplification, a further mismatch with the target sequence is introduced at the second, third, or fourth nucleotide from the 3' end of the primer. As a general rule, if the 3' terminal mismatch is a weak one, a strong secondary mismatch is engineered; if it is a strong one, a weak secondary mismatch is introduced. The mutation-specific ARMS primers used in my laboratory to diagnose the 25 most common β-thalassemia mutations, plus the Hb variants HbS, HbC and HbE, are given in Table 4. All are 30 bases long so that they can all be used at a single high annealing temperature (65°C). Primers for the diagnosis of the normal alleles for many of these mutations are provided in Chapter 8. These are required when both partners of a couple requesting prenatal diagnosis of β-thalassemia carry the same mutation. A typical ARMS test for a single mutation consists of two amplifications in the same reaction mixture using the same genomic DNA as substrate. One amplification product results from the specific ARMS primer and its primer pair (when the mutation is present in the genomic DNA), and the other amplification results from two primers that generate a control fragment in all cases. The generation of control product indicates that the reaction mixture and thermal cycler are working optimally. 1. Prepare a reaction mixture (4 mL) comprising: 0.5 mL of 10X PCR buffer (50 mM KCl; 10 mM Tris-HCl, pH 8.3, 1.5 mM MgCl2, 100 µg/mL gelatin), 1.25 mL of a 1.35 mM dNTP mixture, and 2.65 mL of sterile dH2O. When more than one test is being performed, a primer and the enzyme can be mixed together in a separate tube before addition to the reaction mix. This decreases pipeting errors as larger quantities are used. 2. Pipet 20 µL of the PCR reaction mixture into a 0.5-µL tube.

DNA Diagnosis of Hb Mutations

111

Add 1 µL of each primer (1 OD unit/mL). Add 0.05 µL of Taq DNA polymerase (5 U/µL). Add 1 µL of genomic DNA (100 ng/µL). Overlay with 25 µL of mineral oil. Mix, centrifuge, and place in the thermal cycler. Amplify for 25 cycles as follows: 1 min at 94°C, 1 min at 65°C, 1.5 min at 72°C with a final extension period of 3 min at 72°C following the twenty-fifth cycle. 9. Remove the tubes from the thermal cycler and add 5 µL of blue dye (15% Ficoll/ 0.05% bromophenol blue). Mix and centrifuge. 10. Load a 20-µL aliquot onto a 3% agarose gel and run at 100 V for approx 45 min in TBE (see Subheading 3.1., step 10). 11. Stain with EtBr and visualize the bands as described in Subheading 3.1., step 11. 3. 4. 5. 6. 7. 8.

3.3. Restriction Enzyme Digestion A small number of the β-thalassemia mutations create or abolish a restriction endonuclease recognition site in the globin gene sequence. Provided that the enzyme is commercially available (not always the case) and that there is not another site too close to the mutation, the loss or creation of a site can be used to diagnose the presence or absence of the mutation. This is useful for the diagnosis of a few of the common β-thalassemia mutations, as listed in Table 4, but the main use of this PCR technique is for the diagnosis of the clinically important Hb variants HbS (Fig. 3), HbD Punjab, and HbO Arab. The primer sequences used in my laboratory for diagnosing these Hb variants are given in Table 5. When possible, the amplified product should include a second site for the appropriate restriction enzyme. This site will act as a control for the digestion reaction since it should be fully cleaved in product from both the normal and mutant DNA alleles. This is possible for the HbS and HbE mutations but not for HbO Arab and HbD Punjab, for which the flanking EcoRI sites are too far away from the one in codon 121. 1. To one 0.5-mL tube, add the following: 20 µL of PCR reaction mixture (as detailed in Subheading 3.2.), 1 µL of each primer, 1 µL of genomic DNA (100 ng/µL), 2 µL of sterile dH2O, and 0.05 µL of AmpliTaq DNA polymerase (5 U/µL). 2. Overlay with 25 µL of mineral oil. 3. Place in the thermal cycler and perform 30 cycles of: 1 min at 94°C, 1 min at 65°C, and 1.5 min at 72°C with a final period at 72°C for 3 min after the last cycle. 4. Remove the tubes and add 5–10 U of the appropriate restriction enzyme, plus 2 µL of the corresponding 10X buffer. 5. Incubate at 37°C for a minimum of 1 h. 6. Add blue dye, mix, and spin as in Subheading 3.1., steps 8 and 9. 7. Load a 20-µL aliquot onto a 3% agarose gel consisting of 50% Nusieve GTC agarose and 50% ordinary agarose. 8. Electrophorese, stain, and photograph as in Subheading 3.1., step 11.

112

Oligonucleotide sequence

-88(C → T) -87(C → G) -30(T → A) -29(A → G) -28(A → G) CAP+1(A → G) CD5(-CT) CD6(-A) CD8(-AA) CD8/9(+G) CD15(G → A) CD16(-C) CD17(A → T) CD24(T → A) CD39(C → T) CD41/42(-TCTT) CD71/72(+A) IVSI-1(G → A) IVSI-1(G → T) IVSI-5(G → C) IVSI-6(T → C) IVSI-110(G → A) IVSII-1(G → A)

TCACTTAGACCTCACCCTGTGGAGCCTCAT CACTTAGACCTCACCCTGTGGAGCCACCCG GCAGGGAGGGCAGGAGCCAGGGCTGGGGAA CAGGGAGGGCAGGAGCCAGGGCTGGGTATG AGGGAGGGCAGGAGCCAGGGCTGGGCTTAG ATAAGTCAGGGCAGAGCCATCTATTGGTTC TCAAACAGACACCATGGTGCACCTGAGTCG CCCACAGGGCAGTAACGGCAGACTTCTGCC ACACCATGGTGCACCTGACTCCTGAGCAGG CCTTGCCCCACAGGGCAGTAACGGCACACC TGAGGAGAAGTCTGCCGTTACTGCCCAGTA TCACCACCAACTTCATCCACGTTCACGTTC CTCACCACCAACTTCATCCACGTTCAGCTA CTTGATACCAACCTGCCCAGGGCCTCTCCT CAGATCCCCAAAGGACTCAAAGAACCTGTA GAGTGGACAGATCCCCAAAGGACTCAACCT CATGGCAAGAAAGTGCTCGGTGCCTTTAAG TTAAACCTGTCTTGTAACCTTGATACCGAT TTAAACCTGTCTTGTAACCTTGATACGAAA CTCCTTAAACCTGTCTTGTAACCTTGTTAG TCTCCTTAAACCTGTCTTGTAACCTTCATG ACCAGCAGCCTAAGGGTGGGAAAATAGAGT AAGAAAACATCAAGGGTCCCATAGACTGAT

Product size (bp)

A A A A A A A B A B A B B B B B C B B B B B B

684 683 626 625 624 597 528 207 520 225 500 238 239 262 436 439 241 281 281 285 286 419 634

Altered restriction site +FokI -AvrII +NlaIII

-DdeI -DdeI

+MaeI +MaeI

-BspMI -BspMI +SfaNI

Old

Mutation

Second primer

112

Table 4 ARMS-PCR Primer Sequences Used for the Detection of Common β-Thalassemia Mutationsa

-HphI

GAATAACAGTGATAATTTCTGGGTTAACGT* TCATATTGCTAATAGCAGCTACAATCGAGG* CCCACAGGGCAGTAACGGCAGACTTCTGCA CCACAGGGCAGTAACGGCAGACTTCTCGTT TAACCTTGATACCAACCTGCCCAGGGCGTT

D D B B B

DNA Diagnosis of Hb Mutations

IVSII-654(C → T) IVSII-745(C → G) bsCD6(A → T) bCCD6(G → A) bECD26(G → A)

829 738 207 206 236

+RsaI -DdeI -MnlI

a The above primers are coupled as indicated with either primer A: CCCCTTCCTATGACATGAACTTAA; B: ACCTCACCCTGTGGAGCCAC; C: TTCGTCTGTTTCCCATTCTAAACT; or D: GAGTCAAGGCTGAGAGATGCAGGA. The control primers used were primers D plus E: CAATGTATCATGCCTCTTTGCACC (which yield an 861-bp product as shown in Fig. 1) for all the mutation-specific ARMS primers except the two marked with an asterisk. Control primers used with these two are the Gγ-HindIII RFLP primers (8).

113

Table 5 Oligonucleotide Primers for Detection of βS, βE, βD Punjab, and βo Arab Mutations as RFLPs Mutation and affected restriction enzyme site

Primer sequences (forward and reverse)

Annealing Product Temperature size (°C) (bp)

(A → T) (Loses DdeI site)

ACCTCACCCTGTGGAGCCAC GAGTGGACAGATCCCCAAAGGACTCAAGGA

65 65

443

βECD26 (G → A) (Loses MnlI site) βDPunjab CD121 (G → C) (Loses EcoRI site) βo Arab CD121 (G → A) (Loses EcoRI site)

ACCTCACCCTGTGGAGCCAC GAGTGGACAGATCCCCAAAGGACTCAAGGA CAATGTATCATGCCTCTTTGCACC GAGTCAAGGCTGAGAGATGCAGGA CAATGTATCATGCCTCTTTGCACC GAGTCAAGGCTGAGAGATGCAGGA

65

443 89/56/35/33 861

65 65 65 65

861

376 67

201 175 67 171

231 89/60/35/33 861 252 309 861 552 309

113

βSCD6

Absence Presence of site of site (bp) (bp)

114

Old

4. Notes 1. Gap-PCR seems to work reasonably well for amplifying deletion mutations in the β-globin gene cluster. However, amplification of deletions in the α-gene cluster is technically more difficult, possibly owing to the high GC content of the α-globin gene cluster sequence. Experience in my laboratory has shown that some of the first primer pairs to be published are unreliable, resulting occasionally in unpredictable reaction failures owing to allele dropout, especially when the normal and mutant specific primers are multiplexed together. On the other hand, the recently published multiplex primers seem to give more robust and reproducible results. The addition of betaine to the reaction mixture and the use of Platinum Taq DNA polymerase (which has been developed for automatic hot-start amplification of problematic or GC-rich templates) are key features of their success. The technique is useful for screening for a particular αo-thalassemia deletion mutation, but because of the problem of allele dropout, for a prenatal diagnosis the result is always confirmed by Southern blot analysis. 2. Both positive and negative DNA controls should always be tested alongside any sample. For prenatal diagnosis, this usually means a normal, heterozygous, and homozygous DNA sample for the mutation under study. 3. The relationship between fragment intensities after staining should be constant for all DNA samples. Any deviation to the expected pattern of band intensities in a particular sample should be treated as suspect and the sample retested (e.g., as a result of poor amplification of one of the two alleles, or a partial digestion of the amplified product by the restriction enzyme). 4. Sometimes an ARMS primer may produce a faint positive response with a negative DNA control. This is usually less intense than that of the product observed in the positive DNA control and may be a false positive result. This occurs if there has been a subtle change in the reaction conditions or if the ARMS primer has started to lose its specificity through degradation of the oligonucleotide. Always use small aliquots of primers as a working solution and store the stock solution at –20° or –70°C if available. 5. Failure of the PCR to produce any product may be the result of the genomic DNA sample being too dilute (reprecipitate in a smaller volume) or, more often, the DNA being too concentrated (try a 1:10 dilution). Do not increase the number of cycles to obtain a result because nonspecific amplification products are likely to appear and there could be a problem of coamplifying any contaminating maternal DNA in a fetal sample. Amplification failure may also be owing to an error in the primer sequence. Check the published sequence against the GenBank sequence for typing errors. 6. The unexpected failure of the PCR tests set up in one run usually results from a problem with the dNTP mixture. The first troubleshooting step should be to try new dNTP solutions.

References 1. Embury, S. H. (1995) Advances in the prenatal and molecular diagnosis of the haemoglobinopathies and thalassaemias. Hemoglobin 19, 237–261. 2. Old, J. (1996) Haemoglobinopathies. Prenat. Diagn. 16, 1181–1186.

DNA Diagnosis of Hb Mutations

115

3. Thein, S. L., Eshari, A., and Wallace, R. B. (1993) The use of synthetic oligonucleotides as specific hybridisation probes in the diagnosis of genetic disorders, in Human Genetic Disease Analysis: A Practical Approach (Davies, K. E., ed.), IRL, Oxford, pp. 22–33. 4. Bowden, D. K., Vickers, M. A., and Higgs, D. R. (1992) A PCR-based strategy to detect the common severe determinants of α-thalassaemia. Br. J. Haematol. 81, 104–108. 5. Baysal, E. and Huisman, T. H. J. (1994) Detection of common deletional α-thalassaemia-2 determinants by PCR. Am. J. Hematol. 46, 208–213. 6. Liu, Y. T., Old, J. M., Fisher, C. A., Weatherall, D. J., and Clegg, J. B. (1999) Rapid detection of α-thalassaemia deletions and α-globin gene triplication by multiplex polymerase chain reactions. Brit. J. Haematol. 108, 295–299. 7. Chong, S. S., Boehm, C. D., Higgs, D. R., and Cutting, G. R. (2000) Single-tube multiplex-PCR screen for common deletional determinants of α-thalassemia. Blood 95, 360–362. 8. Old, J. M. (1996) Haemoglobinopathies: community clues to mutation detection, in Methods in Molecular Medicine, Molecular Diagnosis of Genetic Diseases (Elles, R., ed.), Humana, Totowa, NJ, pp. 169–183. 9. Hartveld, K. L., Heister, A. J., G. A. M., Giordano, P. C., Losekoot, M., and Bernini, L. F. (1996) Rapid detection of point mutations and polymorphisms of the a-globin genes by DGGE and SSCA. Hum. Mutat. 7, 114–122. 10. Molchanova, T. P., Pobedimskaya, D. D., and Postnikov, Y. V. (1994) A simplified procedure for sequencing amplified DNA containing the α-2 or α-1 globin gene. Hemoglobin 18, 251–255. 11. Ko, T. M., Tseng, L. H., Hsieh, F. J., and Lee, T. Y. (1993) Prenatal diagnosis of HbH disease due to compound heterozygosity for south-east Asian deletion and Hb Constant Spring by polymerase chain reaction. Prenat. Diagn. 13, 143–146. 12. Baysal, E. (1995) The β- and δ-thalassemia repository. Hemoglobin 19, 213–236. 13. Ristaldi, M. S., Pirastu, M., Rosatelli, C., and Cao, A. (1989) Prenatal diagnosis of β-thalassaemia in Mediterranean populations by dot blot analysis with DNA amplification and allele specific oligonucleotide probes. Prenat. Diagn. 9, 629–638. 14. Sutcharitchan, P., Saiki, R., Fucharoen, S., Winichagoon, P., Erlich, H., and Embury, S. H. (1995) Reverse dot-blot detection of Thai β-thalassaemia mutations. Br. J. Haematol. 90, 809–816. 15. Tan, J. A. M. A., Tay, J. S. H., Lin, L. I., Kham, S. K. Y., Chia, J. N., Chin, T. M., Norkamov, B. T., Aziz, A. O. B., and Wong, H. B. (1994) The amplification refractory mutation system (ARMS): a rapid and direct prenatal diagnostic techniques for β-thalassaemia in Singapore. Prenat. Diagn. 14, 1077–1082. 16. Cai, S. P. and Kan, Y. W. (1990) Identification of the multiple β-thalassaemia mutations by denaturing gradient gel electrophoresis. J. Clin. Invest. 85, 550–553. 17. Losekoot, M. Fodde, R., Harteveld, C. L., Van Heeren, H., Giordano, P. C., and Bernini, L. F. (1991) Denaturing gradient gel electrophoresis and direct sequencing of PCR amplified genomic DNA: a rapid and reliable diagnostic approach to beta thalassaemia. Br. J. Haematol. 76, 269–274.

116

Old

18. Craig, J. E., Barnetson, R. A., Prior, J., Raven, J. L., and Thein, S. L. (1994) Rapid detection of deletions causing δβ thalassemia and hereditary persistence of fetal hemoglobin by enzymatic amplification. Blood 83, 1673–1682. 19. Newton, C. R., Graham, A., and Heptinstall, L. E. (1989) Analysis of any point mutation in DNA. The amplification refractory mutation system (ARMS). Nucl. Acids Res. 17, 2503–2516. 20. Quaife, R., Al-Gazali, L., Abbes, S., Fitzgerald, P., Fitches, A., Valler, D., and Old, J. M. (1994) The spectrum of β-thalassaemia mutations in the U.A.E. national population. J. Med. Genet. 31, 59–61. 21. Kwok, S., Kellogg, D. E., McKinney, N., Spasic, D., Goda, L., Levenson, C., and Sninsky, J. J. (1990) Effects of primer-template mismatches on the polymerase chain reaction: human immunodeficiency virus type I model studies. Nucl. Acids Res. 18, 999–1005. 22. Faa, V., Rosatelli, M. C., Sardu, R., Meloni, A., Toffoli, C., and Cao, A. (1992) A simple electrophoretic procedure for fetal diagnosis of β-thalassaemia due to short deletions. Prenat. Diagn. 12, 903–908. 23. Waye, J. S., Cai, S.-P., Eng, B., Clark, C., Adams, J. G. III, Chui, D. H. K., and Steinberg, M. H. (1991) High haemoglobin A2 βo thalassaemia due to a 532 bp deletion of the 5' β-globin gene region. Blood 77, 1100–1103. 24. Old, J. M., Varawalla, N. Y., and Weatherall, D. J. (1990) The rapid detection and prenatal diagnosis of β thalassaemia in the Asian Indian and Cypriot populations in the UK. Lancet 336, 834–837. 25. Thein, S. L., Hesketh, C., Brown, K. M., Anstey, A. V., and Weatherall, D. J. (1989) Molecular characterisation of a high A2 β thalassaemia by direct sequencing of single strand enriched amplified genomic DNA. Blood 73, 924–930. 26. Dimovski, A. J., Efremove, D. G., Jankovic, L., Plaseska, D., Juricic, D., and Efremov, G. D. (1993) A β o thalassaemia due to a 1605 bp deletion of the 5' β-globin gene region. Br. J. Haematol. 85, 143–147. 27. Lynch, J. R., Brown, J. M., Best, S., Jennings, M. W., and Weatherall, D. J. (1991) Characterisation of the breakpoint of a 3.5 kb deletion of the β-globin gene. Genomics 10, 509–511. 28. Craig, J. E., Kelly, S. J., Barnetson, R., and Thein, S. L. (1992) Molecular characterisation of a novel 10.3 kb deletion causing β-thalassaemia with unusually high Hb A2. Br. J. Haematol. 82, 735–744. 29. Waye, J. S., Eng, B., and Hunt, J. A., et al. (1994) Filipino β-thalassaemia due to a large deletion: identification of the deletion endpoints and polymerase chain reaction (PCR)-based diagnosis. Hum. Genet. 94, 530–532.

Analysis of Prenatal Diagnosis

117

8 Methods for Analysis of Prenatal Diagnosis John M. Old 1. Introduction Prenatal diagnosis of β-thalassemia was first accomplished in 1974, and since then, many countries have developed an extremely successful program for controlling the disorder based on population screening and fetal diagnosis. Initially, this was performed by the measurement of globin chain synthesis in fetal blood, obtained by fetal blood sampling at 18–20 wk of gestation. However, DNA analysis techniques soon began to replace the globin chain synthesis approach, first by the indirect technique of restriction fragment length polymorphism (RFLP) analysis, followed by direct detection of mutations by restriction enzyme digestion and later by oligonucleotide hybridization to DNA fragments on a Southern blot. All of these DNA analysis methods by the Southern blot technique were complex and expensive, and prenatal diagnosis remained inaccessible for developing countries until the discovery of polymerase chain reaction (PCR), which led to the development of simpler, quicker, and less expensive nonradioactive methods of mutation detection (1). Fetal DNA was obtained from cultured amniotic fluid cells until 1982, when chorionic villous sampling (CVS) in the first trimester of pregnancy was developed (2). Currently, prenatal diagnosis by CVS DNA analysis is the method of choice because it is carried out at wk 10–12 of gestation, the risk of fetal mortality associated with the method is acceptably low at 1%, and sufficient DNA is obtained for analysis without culturing (3). Couples at risk of severe hemoglobin (Hb) disorders are first identified by hematological screening tests as directed by published guidelines and flow charts (4). The basic tests are the measurement of the mean corpuscular volume (MCV) and mean corpuscular Hb (MCH) values, the levels of HbA2 and From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

117

118

Old

HbF, and the detection of abnormal Hbs by electrophoresis methods or highperformance liquid chromatography (HPLC). An individual with a reduced MCV and MCH with a normal HbA2 level has α-thalassemia, with a raised HbA2 level has β-thalassemia, and with a raised HbF level of 5–15% has δβ-thalassemia. An individual with normal red cell indices and an HbF level of 15–30% has hereditary persistence of fetal Hb (HPFH). However, such use of these tests has many pitfalls that may lead to the wrong carrier identification. These include the presence of iron deficiency, which also reduces the MCV and MCH; mild β-thalassemia mutations, which are associated with borderline raised HbA2 levels; and the coinheritance of a δ-thalassemia mutation, which reduces the HbA2 level in an individual with β-thalassemia trait to a normal value (5).

1.1. Diagnostic Approaches The main diagnostic approaches for the prenatal diagnosis of the hemoglobinopathies are given in Table 1 in Chapter 7. The PCR-based ones used in my laboratory are gap-PCR, amplification refractory mutation system (ARMS)PCR, and restriction endonuclease analysis of amplified product. Detailed protocols for each of these techniques are presented in Chapter 8. A brief description of the main globin gene disorders for which prenatal diagnosis should be offered is given next.

1.2. α-Thalassemia The most severe form of α-thalassemia results from the homozygous state for αo-thalassemia, known as Hb Bart’s hydrops fetalis syndrome (6). This condition results from a deletion of all four α-globin genes and an affected fetus cannot synthesize any α-globin to make HbF or HbA. Examination of fetal blood by HPLC or isoelectric focusing reveals only the abnormal Hb Bart’s (γ4) and a small amount of Hb Portland (ζ2γ2). The resulting severe fetal anemia leads to asphyxia, hydrops fetalis, and stillbirth or neonatal death. Prenatal diagnosis is always indicated in order to avoid the severe toxemic complications that occur frequently in pregnancy with hydropic fetuses. HbH disease results from the compound heterozygous state of αo- and + α -thalassemia (––/–α) or, more rarely, from the homozygous state of nondeletion α+-thalassemia mutations affecting the more dominant α2 gene (αTα/αTα) (7). Individuals with HbH disease have a moderately severe hypochromic microcytic anemia and produce large amounts of HbH (β4) as a result of the excess β-chains in the reticulocyte. Patients may suffer from fatigue, general discomfort, and splenomegaly, but they rarely require hospitalization and lead a relatively normal life. Therefore, prenatal diagnosis is not normally performed for HbH disease. However, there is also a more severe form of HbH

Analysis of Prenatal Diagnosis

119

disease arising from the compound heterozygous state of αo-thalassemia and nondeletion α+-thalassemia (– –/αTα). Such patients seem to exhibit more severe symptoms with a possible requirement of recurrent blood transfusions and splenectomy. In some situations, couples at risk of this more severe form of HbH disease have opted for prenatal diagnosis and termination of an affected fetus (8).

1.3. β-Thalassemia The β-thalassemias are a heterogeneous group of disorders characterized by either an absence of β-globin chain synthesis (βo type) or a much reduced rate of synthesis (β+ type) (9). The majority of the βo and β+ type of mutations are called severe mutations because in either the homozygous or compound heterozygous state they give rise to the phenotype of β-thalassemia major, a transfusiondependent anemia from early in life. Some β-thalassemia mutations (the mild β+ type, sometimes designated β++ type) in the homozygous state are associated with a milder clinical condition called thalassemia intermedia. Thalassemia intermedia results from a wide variety of disorders including β-thalassemia, δβ-thalassemia, and Hb Lepore. The coinheritance of α-thalassemia or one of the many determinants resulting in a raised HbF level in adult life is also a factor that may cause thalassemia intermedia. Patients with thalassemia intermedia present later in life relative to those with thalassemia major and are capable of maintaining an Hb level above 6 g without transfusion. By contrast, the phenotype of compound heterozygotes when one of these mild mutations is inherited with a severe mutation is less clear and less predictable. Some of these individuals have a mild phenotype, especially if it involves a very mild mutation such as one of the “silent β-thalassemia” mutations, whereas others are more severe and are often transfusion dependent. The unpredictability of the phenotype in compound heterozygotes remains a diagnostic and counseling problem.

1.4. Interaction of Thalassemia with Hb Variants The β-thalassemia mutations and various Hb variant mutations can interact to produce a number of thalassemia and sickle cell disorders for which genetic counseling and prenatal diagnosis should be offered. These interactions are presented in Fig. 1. The interactions are divided into combinations that have a risk of resulting in a serious disorder, those that have a risk of a less serious disorder, and those that pose no risk of a serious disorder. Also included are the carrier combinations that can give rise to a hidden risk (i.e., a risk not easily discernible by simple hematological analysis) of having a fetus affected with homozygous αo-thalassemia. This can occur because the carrier state for various β-thalassemia disorders can mask coexisting αo-thalassemia trait in regions where both disorders are found (e.g., in Southeast Asia). Thus, for couples in whom one partner is diagnosed by hemato-

120

Old

Fig. 1. Diagram showing interactions of various thalassemia disorders and abnormal Hb variants S, C, E, D Punjab, O Arab, and Lepore.

logical screening as a carrier of β-thalassemia and the other a possible carrier of αo-thalassemia, the individual with β-thalassemia should also be screened for αo-thalassemia mutations by DNA analysis. 2. Materials (Sources of Fetal DNA) 2.1. Blood DNA is normally prepared from 5–10 mL of peripheral blood that is anticoagulated with heparin or, preferably, EDTA. The DNA can be isolated by the standard method of phenol-chloroform extraction and ethanol precipitation, or by using one of several available kits on the market based on salt extraction, protein precipitation, and so on. Sufficient DNA is obtained for molecular analysis and subsequent storage in a DNA bank at –20°C. If this is not required, a much smaller quantity of blood may be used for PCR diagnosis of the globin gene disorders. Mutation analysis may be carried out by simply adding 1 µL of boiled whole blood to the PCR reaction mixture (10).

2.2. Amniotic Fluid DNA can be prepared from amniotic fluid cells directly or after culturing. It is prudent to prepare DNA directly from half the sample and to set up the other

Analysis of Prenatal Diagnosis

121

half of the sample for cell culturing as a backup source of DNA. It takes 2 to 3 wk to grow amniocytes to confluence in a 25-mL flask, but culturing has the advantage that a large amount of DNA is obtained (in our experience, the yield from such a flask has varied from 15 to 45 µg, enough DNA for all types of analyses). A diagnosis can be made using DNA from noncultivated cells in most cases. Approximately 5 µg of DNA is obtained from 15 mL of amniotic fluid, which is sufficient for any PCR-based method of analysis. However, for genotype analysis by Southern blotting, it is only enough for one attempt and thus a backup culture is essential in case of failure. The method of DNA preparation for both cultured and noncultivated cells is essentially the same as that for chorionic villi (11).

2.3. Chorionic Villi The two main approaches to CVS—ultrasound-guided transcervical aspiration and ultrasound-guided transabdominal sampling—both provide goodquality samples of chorionic villi for fetal DNA diagnosis. Sufficient DNA is normally obtained for both PCR and Southern blot analysis of the globin genes. For my laboratory’s first 200 CVS DNA diagnoses, the average yield of DNA was 46 µg and only in one instance was <5 µg obtained (11). The main technical problem with this source of fetal DNA is the risk of contamination with maternal DNA, which arises from the maternal decidua that is sometimes obtained along with the chorionic villi. However, by careful dissection and removal of the maternal decidua with the aid of a phase-contrast microscope, pure fetal DNA samples can be obtained, as demonstrated by Rosatelli et al. (12), who reported no misdiagnoses in a total of 457 first-trimester diagnoses for β-thalassemia in an Italian population. Maternal contamination can be ruled out in most cases by the presence of one maternal and one paternal allele following the amplification of highly polymorphic repeat markers (13). The risk of misdiagnosis through maternal DNA contamination can be further reduced by the preparation of DNA from a single villous frond.

2.4. Fetal Cells in Maternal Blood Fetal cells have long been known to be present in the maternal circulation, and they provide an attractive noninvasive approach to prenatal diagnosis. Nevertheless, attempts to isolate the fetal cells using immunological methods and cell sorters have been only slightly successful in providing a population of cells pure enough for fetal DNA analysis. Until recently, analysis of fetal cells in maternal blood could only be applied for the prenatal diagnosis of β-thalassemia in women whose partners carried a different mutation, as reported for the diagnosis of Hb Lepore (14). However, the development of the technique of isolating single nucleated fetal erythrocytes by micromanipulation under microscopic obser-

122

Old

vation (15) has permitted the analysis of both fetal genes in single cells from maternal blood. This approach was shown to be possible for prenatal diagnosis by the report of two successful cases of sickle cell anemia and β-thalassemia (16), but it remains technically very difficult and the method has not been widely adopted.

2.5. Samples for Preimplantation Genetic Diagnosis Preimplantation genetic diagnosis of for the globin gene disorders is now possible either by DNA analysis of single cells biopsied from cleaving embryos or by the analysis of polar body DNA obtained from the two polar bodies extruded during the maturation of the oocyte. Both approaches use a nested PCR technique and appear to be subject to the problem of allele dropout. This problem, like that of maternal contamination, is overcome by the simultaneous analysis of other maternal and paternal markers (17). Only the eggs without the defect are fertilized and implanted in the mother. This approach is appealing to couples whose religious beliefs will not permit the termination of a pregnancy and for those who have already had several therapeutic abortions. However, the preimplantation genetics diagnosis approach is limited in its applicability by the degree of technical difficulty of the amplification procedure and the high costs of the obstetric procedure. 3. Methods (see Notes 1–5) The methods of DNA analysis used for prenatal diagnosis are the same as those used for mutation screening in Chapter 7. However, in addition to the mutation-specific primers described previously, normal sequence-specific primers are required. These are necessary for the diagnosis of normal DNA sequence in cases in which both partners carry the same mutation. The normal ARMS primer sequences for many of the common β-thalassemia mutations are detailed in Table 1. Examples of the prenatal diagnosis of β-thalassemia and sickle cell disease are shown in Figs. 2–4. Further tests are also necessary. A second method of diagnosis is applied whenever possible to confirm the result, and analysis of the inheritance of fetal DNA polymorphisms is necessary for checking maternal DNA contamination. Analysis of the linkage of the β-globin gene haplotype polymorphisms is a useful second approach for confirmation of results in some cases in which a family study is possible. However, it is not so useful for checking the presence of maternal DNA contamination and the occasional case of false paternity; a more general approach is obtained by the use of highly polymorphic markers. These can be short tandem repeat polymorphisms or variable number tandem repeat (VNTR) polymorphisms, the latter approach favored by my laboratory because the method is very simple, using the same gel electrophoresis protocol as the ARMS procedure.

123

Mutation

Oligonucleotide sequence

–87 (CÆG) CD5 (–CT) CD8 (–AA) CD8/9 (+G) CD15 (G → A) CD39 (C → T) CD41/42 (–TCTT) IVSI-1 (G → A) IVSI-1 (G → T) IVSI-5 (G → C) IVSI-6 (T → C) IVSI-110 (G → A) IVSII-1 (G → A) IVSII-654 (C → T) IVSII-745 (C → G) βS CD6 (A → T) βE CD26 (G → A)

CACTTAGACCTCACCCTGTGGAGCCACCCC CAAACAGACACCATGGTGCACCTGACTCCT ACACCATGGTGCACCTGACTCCTGAGCAGA CCTTGCCCCACAGGGCAGTAACGGCACACT TGAGGAGAAGTCTGCCGTTACTGCCCAGTA TTAGGCTGCTGGTGGTCTACCCTTGGTCCC GAGTGGACAGATCCCCAAAGGACTCAAAGA TTAAACCTGTCTTGTAACCTTGATACCCAC GATGAAGTTGGTGGTGAGGCCCTGGGTAGG CTCCTTAAACCTGTCTTGTAACCTTGTTAC AGTTGGTGGTGAGGCCCTGGGCAGGTTGGT ACCAGCAGCCTAAGGGTGGGAAAATACACC AAGAAAACATCAAGGGTCCCATAGACTGAC GAATAACAGTGATAATTTCTGGGTTAACGC TCATATTGCTAATAGCAGCTACAATCGAGC AACAGACACCATGGTGCACCTGACTCGTGA TAACCTTGATACCAACCTGCCCAGGGCGTC

a See

Analysis of Prenatal Diagnosis

Table 1 Primer Sequences Used for the Detection of Normal DNA Sequence by Allele-Specific Priming Techniquea Second Primer

Product size (kb)

A A A B A A B B A B A B B D D A B

683 528 520 225 500 299 439 281 455 285 449 419 634 829 738 527 236

Table 4 legend of Chapter 7 for details of primers A–D and control primers.

123

124

Old

Fig. 2. Prenatal diagnosis for β-thalassemia using ARMS primers to detect mutations IVS1–110 (G → A) and codon 39 (C → T). Lanes 1 and 6, fetal DNA; lanes 2 and 4, maternal DNA; lanes 3 and 5, paternal DNA. Lanes 1–3 show the results with an ARMS primer for codon 39, and lanes 4–6 show the results with an ARMS primer for IVS1–110. The upper band is the 861-bp control product. The results show that the fetus was heterozygous for codon 39.

3.1. Haplotype Analysis Linkage analysis of RFLPs within the β-globin gene cluster can often be used for prenatal diagnosis of β-thalassemia in rare cases in which one or both of the mutations remain unidentified after screening using a direct detection method such as ARMS. The technique can also enable the prenatal diagnosis of uncharacterized δβ-thalassemia deletion mutations through the apparent non-Mendelian inheritance of RFLPs (owing to the hemizygosity created by the inheritance of deleted sequences on one chromosome). Finally, haplotype analysis may provide an alternative approach for the confirmation of a prenatal diagnosis result obtained by a direct detection method such as ARMS and, in very rare instances, has helped to reveal a possible diagnostic error (18). At least 18 RFLPs have been characterized within the β-globin gene cluster (19). However, most of these RFLP sites are nonrandomly associated with each

Analysis of Prenatal Diagnosis

125

Fig. 3. Prenatal diagnosis for β-thalassemia using ARMS primers to detect mutation IVS1–5 (G → C). The diagram of the β-globin gene shows the location of the mutation and its specific ARMS primer (M) plus the positions of the common primer (B) and two control primers (E and D) that generate an 861-bp product. The ethidium bromide–stained gel shows the results of screening of parental DNA samples with mutant primer in lanes 1 and 2, a homozygous IVS1–5 (G → C) control DNA with normal primer in lane 3, and the CVS DNA with mutant primer and normal primer in lanes 4 and 5, respectively. Generation of the specific 285-bp product with normal primer and the absence of product with mutant primer shows that the fetus is normal for the IVS1–5 (G → C) mutation.

other and thus combine to produce just a handful of haplotypes (20). In particular, they form a 5' cluster that is 5' to the δ gene and a 3' cluster that extends downstream from the β-globin gene. The DNA in between the two clusters contains a relative hotspot for meiotic recombination with a rate of approx 1 in 350 meioses (21). The β-globin gene cluster haplotype normally consists of 5 RFLPs located in the 5' cluster (HindII/ε gene, HindIII/Gγ gene, HindIII/Aγ gene, HindII/3'ψβ-gene, and HindII/5'ψβ-gene) and two RFLPs in the 3' cluster (AvaII/β gene, BamHI/β gene) (22). All of the seven RFLPs except BamHI can be analyzed by PCR very simply and quickly using the procedure described in Chapter 7 for PCR and restriction enzyme digestion. The primer sequences and sizes of the fragments generated are provided in Table 2. The BamHI RFLP is located within an L1 repetitive

126

Old

Fig. 4. Diagnosis of sickle cell genotypes by DdeI digestion. The diagram shows a map of the DdeI sites at the 5' end of the β-globin gene together with the results of analysis of heterozygous parental DNA samples (AS) in lanes 1 and 2, the CVS DNA in lane 3, a normal DNA (AA) control in lane 4, and a homozygous (SS) DNA control in lane 5.

element creating amplification problems, and an Hinf I RFLP located just 3' to the β-globin gene is used instead, because these two RFLPs have been found to exist in linkage disequilibrium (23). Three other RFLPs are included in Table 2. An AvaII RFLP in the ψβ gene is extremely useful in haplotype analysis of Mediterranean β-thalassemia heterozygotes. The (–) allele for this RFLP is frequently found on chromosomes carrying the IVSI-110 mutation, whereas it is very rare on normal β-globin chromosomes (24) and thus is a very useful informative marker for individuals heterozygous for this mutation. The RsaI RFLP located just 5' to the β-globin gene is useful for linkage analysis because it appears to be unlinked to either the 5' cluster or the 3' cluster RFLPs and thus may be informative when the 5' haplotype and the 3' haplotype are not. Finally the Gγ-XmnI RFLP, created by the nondeletion HPFH C → T mutation at position –158, is included because of its use in the analysis of sickle cell gene haplotypes and in individuals with thalassemia intermedia.

RFLP site

Primer sequences: 5'–3'

Hind II ε-gene

127

5 TCTCTGTTTGATGACAAATTC 5 AGTCATTGGTCAAGGCTGACC Xmn I Gγ-gene 5 AACTGTTGCTTTATAGGATTTT 5 AGGAGCTTATTGATAACCTCAGAC HindIII Gγ-gene 5 AGTGCTGCAAGAAGAACAACTACC 5 CTCTGCATCATGGGCAGTGAGCTC HindIII Aγ-gene 5 ATGCTGCTAATGCTTCATTAC 5 TCATGTGTGATCTCTCAGCAG HindII 5' ψβ-gene 5 TCCTATCCATTACTGTTCCTTGAA 5 ATTGTCTTATTCTAGAGACGATTT HindII 3' ψβ-gene 5 GTACTCATACTTTAAGTCCTAACT 5 TAAGCAAGATTATTTCTGGTCTCT AvaII ψβ-gene Sequence as for Hind 5'ψβ RFLP

Product Coordinates size on GenBank (bp) sequence U01317 760 657 326 635 795 913 795

RsaI β-gene

5 AGACATAATTTATTAGCATGCATG 5 CCCCTTCCTATGACATGAACTTAA

AvaII β-gene

5 GTGGTCTACCCTTGGACCCAGAGG 328 5 TTCGTCTGTTTCCCATTCTAAACT 5 GGAGGTTAAAGTTTTGCTATGCTGTAT 474 5 GGGCCTATGATAGGGTAAT

HinfI β-gene

1200

18652–18672 19391–19411 33862–33883 34495–34518 35677–35700 35981–36004 40357–40377 40971–40991 46686–46709 47457–47480 49559–49582 50448–50471 46686–46709 47457–47480 61504–61527 62680–62703

62416–62439 62720–62743 63974–64001 64429–64447

Absence of site (bp) 760

Analysis of Prenatal Diagnosis

Table 2 Primers Used for Analysis of β-Globin Gene Cluster RFLPs

Presence Annealing of site temperature (bp) (°C) 55 55 65 65 55 55 55 55

65 55

127

315 445 657 455 202 326 235 91 635 327 308 795 691 104 913 479 434 795 440 355 411 330 Plus constant 81 fragments of Plus 694 and 95 694 and 95 328 228 100 320 213 Plus constant 107 fragment of and 154 154

128

Old

3.2. VNTR Analysis A check for maternal contamination in the fetal DNA sample by polymorphism analysis should always be set up at the same time as the globin gene mutation assays. In my laboratory, we routinely analyze the fetal DNA and parental DNA samples for two VNTR (13) polymorphisms, the Apo B and IgJh VNTRs. In the very rare cases in which neither of the polymorphic markers provides informative results that exclude the possibility of maternal contamination, other VNTR polymorphisms are tried such as the Col2A1gene VNTR locus (25) and the D4S95 marker from the Huntington disease region of chromosome 4 (26). The primer sequences and size range of the PCR products for these four VNTR markers are given in Table 3. All except the IgJh primers use the standard ARMS PCR buffer. The IgJh primers require an (NH4)2SO4 buffer: 75 mM Tris-HCl (pH 9.0), 20 mM (NH4)2SO4, 2.0 mM MgCl2, 0.01% Tween, 10% dimethylsulfoxide, 10 mM β-mercaptoethanol (all final concentrations). 4. Notes 1. The main problem of prenatal diagnosis is diagnostic error leading to misdiagnosis. An audit of the accuracy of prenatal diagnosis for the Hb disorders in the United Kingdom from 1974 to 1999 revealed a diagnostic error rate of 0.41%. Diagnostic errors were recorded to have occurred from the very high sensitivity of PCR to maternal DNA contamination, the failure to amplify the target sequence, false paternity, sample exchange, and various nonlaboratory errors such as incorrect referral or diagnosis by hematological screening of parental phenotypes (18). 2. Nonlaboratory errors are minimized by insisting that fresh blood samples be received for confirmation of the parental phenotypes in every case. In cases when a fresh blood sample from the father is simply not available (in couples at risk of a sickle cell disorder), extra tests for other possible globin gene mutations are carried out. In particular, when a fetal genotype of AS is diagnosed, the fetal DNA is always analyzed for the βC mutation and the common β-thalassemia mutations observed in the ethnic group of the father. 3. Laboratory errors are minimized by performing duplicate tests, and preparing DNA from both a single frond and the bulk CVS material whenever possible. Technical errors such as partial digestion or allele dropout are minimized by using two independent diagnostic methods on each sample whenever possible. 4. Polymorphism analysis is used routinely to exclude error owing to maternal DNA contamination or nonpaternity. Maternal DNA contamination must be excluded in all cases in which the fetal diagnosis is the same genotype as the mother. The risk of maternal DNA contamination is much lower in cases in which the fetus is normal, is homozygous, or has inherited a different mutation from that carried by the mother. 5. The precautions in Notes 1–4 form a best code of practice for minimizing errors in prenatal genetic testing for any genetic disorder. The guidelines for best practice are as follows:

129

VNTR

Primer pair

Apo B

5'-GAAACGGAGAAATTATGGAGGG-3' 5'-TCCTGAGATCAATAACCTCG-3' 5'-GGGCCCTGTCTCAGCTGGGGA-3' 5'-TGGCCTGGCTGCCCTGAGCAG-3' 5'-CCAGGTTAAGGTTGACAGCT-3' 5'-GTCATGAACTAGCTCTGGTG-3' 5'-GCATAAAATGGGGATAACAGTAC-3' 5'-GACATTGCTTTATAGCTGTGCCTCAGTTT-3'

IgJh Col2A1 D4S95

Analysis of Prenatal Diagnosis

Table 3 Primers Used for the Check for Maternal DNA Contamination by VNTR Analysis Annealing temperature (°C)

Repeat length (bp)

Size range of products (bp)

55

30

541–871

68

50

520–1720

55

34 and 31

584–779

60

39

900–1600

129

130

Old a. Ensure that fresh parental blood samples are obtained with the fetal sample in order to check the parental phenotypes and to provide fresh control DNA samples. b. Ensure that the chorionic villus sample has undergone careful microscopic dissection to remove any contaminating maternal decidua. c. Always analyze parental and the appropriate control DNAs with the fetal DNA and always repeat the fetal DNA analysis to confirm the result. d. Whenever possible use an alternative diagnostic method to confirm the diagnosis. e. Use a limited number of amplification cycles to minimize any coamplification of maternal DNA sequences. f. Check for maternal DNA contamination in every case. g. Be sure that the fetal DNA diagnosis report details the types of DNA analysis used and clearly states the risk of misdiagnosis owing to technical errors based on current data.

References 1. Old, J. (1996) Haemoglobinopathies. Prenat. Diagn. 16, 1181–1186. 2. Old, J. M., Ward, R. H. T., Petrou, M., Karagozlu, F., Modell, B., and Weatherall, D. J. (1982) First-trimester fetal diagnosis for haemoglobinopathies: three cases. Lancet 2, 1413–1416. 3. Old, J. M. (1999) Haemoglobinopathies, in Fetal Medicine: Basic Science and Clinical Practice (Rodeck, C. H. and Whittle, M. J., eds.), Churchill Livingstone, London, pp. 483–498. 4. The Thalassemia Working Party of the BCSH General Haematology Task Force (1994) Guidelines for the fetal diagnosis of globin gene disorders. J. Clin. Pathol. 47, 199–204. 5. Cao, A., Rosatelli, M. C., and Eckman, J. R. (2001) Prenatal diagnosis and screening for thalassemia and sickle cell disease, in Disorders of Hemoglobin: Genetics, Pathophysiology, and Clinical Management (Steinberg, M. H., Forget, B. G., Higgs, D. R., and Nagel, R. L., eds.), Cambridge University Press, Cambridge, MA, pp. 958–978. 6. Weatherall, D. J.and Clegg, J. B., ed. (1981) The Thalassemia Syndromes, Blackwell Scientific, Oxford. 7. Higgs, D. R., Vickers, M. A., Wilkie, A. O. M., et al. (1989) A review of the molecular genetics of the human a-globin gene cluster. Blood 73, 1081–1104. 8. Ko, T. M., Tseng, L. H., Hsieh, F. J., and Lee, T. Y. (1993) Prenatal diagnosis of HbH disease due to compound heterozygosity for south-east Asian deletion and Hb constant spring by polymerase chain reaction. Prenat. Diag. 13, 143–146. 9. Olivieri, N. F.and Weatherall, D. J. (2001) Clinical aspect of b-thalassemia, in Disorders of Hemoglobin: Genetics, Pathophysiology, and Clinical Management (Steinberg, M. H., Forget, B. G., Higgs, D. R., and Nagel, R. L., eds.), Cambridge University Press, Cambridge, MA, pp. 277–341 10. Liu, Y. T., Old, J. M., Fisher, C. A., Weatherall, D. J., and Clegg, J. B. (1999) Rapid detection of a-thalassemia deletions and a-globin gene triplication by multiplex polymerase chain reactions. Br. J. Haematol. 108, 295–299. 11. Old, J. M. (1986) Fetal DNA analysis, in Genetic Analysis of the Human Disease: A Practical Approach (Davies, K. E., ed.), IRL, Oxford, England, pp. 1–16.

Analysis of Prenatal Diagnosis

131

12. Rosatelli, M. C., Sardu, R., Taveri, T., Scalas, M. T., Di-Tucci, A., De-Murtas, M., Loudianos, G., Monni, G., and Cao, A. (1990) Reliability of prenatal diagnosis of genetic diseases by analysis of amplified trophoblast DNA. J. Med. Genet. 27, 249–251. 13. Decorte, R., Cuppens, H., Marynen, P., and Cassiman, J.-J. (1990) Rapid detection of hypervariable regions by the polymerase chain reaction technique. DNA Cell. Biol. 9, 461–469. 14. Camaschella, C., Alfarano, A., Gottardi, E., Travi, M., Primignani, P., Cappio, F. C., and Saglio, G. (1990) Prenatal diagnosis of fetal hemoglobin Lepore-Boston disease on maternal peripheral blood. Blood 75, 2102–2106. 15. Sekizawa, A., Watanabe, A., Kimwa, T., et al (1996) Prenatal diagnosis of the fetal RhD blood type using a single fetal nucleated erythrocyte from maternal blood. Obstet. Gynaecol. 87, 501–505. 16. Cheung, M.-C., Goldberg, J. D., and Kan, Y. W. (1996) Prenatal diagnosis of sickle cell anemia and thalassemia by analysis of fetal cells in maternal blood. Nat. Genet. 14, 264–268. 17. Kuliev, A., Rechitsky, S., Verlinsky, O., Ivakhnenko, V., Cieslak, J., Evsikov, S., Wolf, G., Angastiniotis, M., Kalakoutis, G., Strom, C., and Verlinsky, Y. (1999) Birth of healthy children after preimplantation diagnosis of thalassemias. J. Assist. Reprod. Genet. 16, 207–211. 18. Old, J., Petrou, M., Varnavides, L., Layton, M., and Modell, B. (2000) Accuracy of prenatal diagnosis of hemoglobin disorders in the United Kingdom: twentyfive years experience. Prenat. Diagn. 20, 986–991. 19. Kazazian, H. H. Jr. and Boehm, C. D. (1988) Molecular basis and prenatal diagnosis of b-thalassaemia. Blood 72, 1107–1116. 20. Antonarakis, S. E., Boehm, C. D., Diardina, P. J. V., and Kazazian, H. H. J. (1982) Non-random association of polymorphic restriction sites in the b-globin gene cluster. Proc. Natl. Acad. Sci. USA 79, 137–141. 21. Chakravarti, A., Buetow, K. H., Antonarakis, S. E., Waber, P. G., Boehm, C. D., and Kazazian, H. H. (1984) Non-uniform recombination within the human b-globin gene cluster. Am. J. Hum. Genet 36, 1239–1258. 22. Old, J. M., Petrou, M., Modell, B., and Weatherall, D. J. (1984) Feasibility of antenatal diagnosis of b-thalassemia by DNA polymorphisms in Asian Indians and Cypriot populations. Br. J. Haematol. 57, 255–263. 23. Semenza, G. L., Dowling, C. E., and Kazazian, H. H. Jr. (1989) Hinf I polymorphisms 3' to the human b globin gene detected by the polymerase chain reaction (PCR). Nucl. Acids Res. 17, 2376. 24. Wainscoat, J. S., Old, J. M., Thein, S. L., and Weatherall, D. J. (1984) A new DNA polymorphism for prenatal diagnosis of b-thalassemia in Mediterranean populations. Lancet 2, 1299–1301. 25. Berg, E. S. and Olaisen, B. (1993) Characterization of the COL2A1 VNTR polymorphism. Genomics 16, 350–354. 26. Allitto, B. A., Horn, G. T., Altherr, M. R., Richards, B., McClatchey, A. I., Wasmuth, J. J., and Gusella, J. F. (1991) Detection by PCR of the VNTR polymorphism at D4S95. Nucl. Acids Res. 19, 4015.

132

Old

Hemoglobin Fluorescence

133

9 Hemoglobin Fluorescence Rhoda Elison Hirsch 1. Introduction Protein structural analysis took a big leap forward with the discovery of aromatic amino acid and protein fluorescence (1–4). The intrinsic fluorescence of proteins is a highly sensitive reporter of conformational change at or near the fluorescent tryptophans (Trp) and tyrosines (Tyr). Phenylalanine also exhibits ultraviolet (UV) fluorescence excitation and emission, with a low quantum yield that becomes insignificant in proteins containing Tyr and Trp. The binding of specific extrinsic fluorescent probes allows the site-specific probing of other microdomains or nonfluorescent side chains. Fluorescence resonance energy transfer measurements serve as a “spectroscopic ruler” to measure intramolecular and intermolecular distances and may also be used to ascertain the magnitude of conformational change on ligand binding, protein folding, and protein-protein interactions ([5]; for basic principles of fluorescence, see ref. 6). For more than two decades, intact heme-proteins (i.e., the protein with its heme moiety [or moieties] and subunits required for functionality), including natural hemoglobin (Hb) variants, had been excluded from this highly sensitive and informative direct spectroscopic structural probing. This exclusion was based on the general assumption that the fluorescence emission from the Tyr and Trp residues was effectively quenched by the heme moieties (7). Despite the presumed quenching effects, clever utilization of the phenomenon of fluorescence quenching by the hemes provided informative ligand-binding studies of Hbs (e.g., see ref. 8–11). The choice of optics in fluorescence detection contributed significantly to the dogma that heme proteins do not exhibit fluorescence emission. Standard fluorescence instruments typically employ right-angle optics wherein a highly absorptive sample, such as Hb, introduces significant inner-filter effects that From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

133

134

Hirsch

mask the emission. It has been estimated that the hemes give rise to ~99% nonradiative quenching of the aromatic intrinsic fluorophores (7), but this may need to be reevaluated when factoring in corrections for inner-filter effect. Front-face fluorometry and/or alternative excitation and emission detection sources enabled the direct detection of fluorescence emission from Hb. The application of a synchrotron light excitation source yielded the first direct fluorescence lifetime decay measurements in the nanosecond range of Hb and its subunits compared with that of the apoprotein (albeit, as noted by the authors, with poor precision [±0.25 ns]) (12). In 1980, the simultaneous discovery of significant steady-state Hb intrinsic fluorescence emission, by independent laboratories, using more sensitive detectors (13) or front-face optics (14), facilitated the application of fluorescence principles and methodology to provide a powerful tool to probe Hb structure. The use of Trp Hb mutants (with substitutions that increased or decreased the number of tryptophans) demonstrated the significance of the fluorescence emission (Fig. 1). In contrast to theoretical calculations by others (15), the experimental data by different laboratories indicated that β37 Trp at the α1β2 interface is the primary contributor to the fluorescence emission and report on alterations in the R → T transition (for a review, see refs. 16 and 17). With these observations in mind, heme-protein fluorescence was revisited by Alpert, Jameson, Weber, and colleagues (13,18), who concluded that consideration of motions of groups involved in energy transfer mechanisms may dramatically reduce the transfer energy resulting in the observed unquenched steady-state emission. This chapter focuses on details of the methodology regarding front-face fluorometry of Hbs: (1) intrinsic Hb fluorescence and (2) the employment of extrinsic fluorescence probes to explore nonaromatic site-specific microdomains of the Hb tetramer. For in-depth explanations of general fluorescence principles and heme-protein fluorescence (see refs. 6, 16, and 17, respectively).

1.1. Front-Face Fluorometry Front-face fluorometry provides multiple advantages in measuring the fluorescence of any protein solution with a high extinction coefficient of absorption such as Hb. With standard right-angle optics, emission is detected at right angles from the exciting beam through optics focused to the center of the cuvet. Therefore, in a strongly absorbing solution, all absorption takes place near the front surface, with little excitation occurring in the center of the cuvet, and the detector receives little or no light. In essence, using right-angle geometry, the highly absorbent solution itself acts as an inner filter. Inner-filter effects are essentially eliminated by front-face fluorescence measurements. Optimal front-face measurements are made when the incident light

Hemoglobin Fluorescence

135

Fig. 1. Front-face steady-state intrinsic fluorescence emission (uncorrected) of oxyHb tryptophan variants (from Hirsh et al. [14]). H*, HbH β4 [2 β15Trp and 2 β37 Trp], where the sensitivity of the recorder is one-third less than that recorded for the other Hbs (i.e., the actual relative intensity is three times that shown); F, HbF (α2γ2); A, HbA, α2β2 [α14 Trp, β15Trp, β37Trp]; RC, Hb Rothschild (α14Trp, β37 Trp → Arg). More recently, under different conditions and preparations, a low-intensity defined emission maximum near ~330 nm (excitation: 296 nm) has been observed for Hb Rothschild, while the emission spectrum with 280 nm excitation appears the same (A). (B) The intrinsic fluorescence emission spectra of variant hemoglobins obtained with 296 nm excitation that specifically excites tryptophan.

makes an angle of 34° with the normal to the cell face, or 56°, depending on the orientation of the front-face cell adapter (Fig. 2, [19]). This permits the detection of fluorescence emission from optically dense concentrated solutions of Hb. This feature is desirable because the dimer-tetramer dissociation equilibrium (discussed later) is shifted toward the native tetramer at higher (i.e., millimolar or submillimolar) concentrations, in contrast to the micromolar requirements of right-angle optics. Unlike right-angle optics, with front-face optics, there is a certain concentration of fluorophore wherein the fluorescence intensity is no longer dependent on concentration (19). This is advantageous because it eliminates artifacts such as those introduced by small pipeting errors. For HbA, this concentrationindependent plateau is reached at concentrations >0.3 g% (~0.19 mM heme or ~0.05 mM tetramer) (Fig. 3, [14]).

136

Hirsch

Fig. 2. Comparison of fluorescence optics: (A) front-face; (B) right-angle optics. (From ref. 17.)

Front-face measurements may be simply, but suboptimally, achieved in a right-angle configuration with the use of small (millimolar) rounded cuvets or triangular cuvets (14,16–17,20), with the latter providing more sensitive detection. A front-face cell is designed for easy insertion into a standard cuvet holder for 1 × 1 cm cells and orients the sample for the optimal angle requirement (Fig. 4A). The small volume required for this cell (100–200 µL) becomes advantageous when studying heme-proteins with limited availability (e.g., scarce mutants or recombinant mutants). However, an instrument designed with a horizontal orientation of the light source slit may preclude use of this cell. Most companies now offer the option of temperature controlled front-face adapters designed specifically for the fluorometer.

Hemoglobin Fluorescence

137

Fig. 3. Concentration dependence of Hb fluorescence emission intensity plateaus when using front-face optics. Excitation wavelength: 280 nm; oxyHbA: 0.07 mM tetramer, 0.05 M potassium phosphate buffer, 25°C. (From ref. 14.)

Novel variations in front-face optical designs provided further advantage in the study of heme-protein fluorescence. The rhombiform optical cell (Fig. 4B), designed by Horiuchi and Asai (21), simultaneously measures absorption and fluorescence. This allows direct and continuous measurements of the binding of a fluorescent allosteric effector to Hb (at a limited range of concentration), while assessing variations in the partial pressure of oxygen during deoxygenation. The solution is gently stirred for gas exchange. Caution must be exercised when stirring any protein solution, and especially Hbs, which may be subject to mechanical instability (e.g., HbS [22]). With the purpose of eliminating reflections and stray emissions (which may become significant for the relatively low fluorescence emission of heme-proteins), Bucci and colleagues (23,24) developed an optimized shielded cuvet as well as designed an optical cell with a front-face configuration that operates on a free liquid surface (Fig. 4C). This also avoids any possible protein conformational changes induced by a protein-solid interface. However, air-water interfaces do have the potential to induce protein unfolding for some proteins and Hb mutants (25–26). Absolute fluorescence and the determination of quantum yields are not possible with heme-proteins until one can design a true “blank”: the identical globin fold without the heme. Finding a true blank remains a challenge since the respective apoglobins (i.e., globin without the heme) are structurally dis-

138

Hirsch

Fig. 4. Novel fluorescence optical designs for detection of heme-protein fluorescence. (A) Front-face optics is achieved by an insert placed in a standard right-angle cuvet holder. The base plate shown on the left is removable. The key feature is that the exciting light makes an angle of 34° with the normal to the cell face, or, by inverting, one may make the angle of incidence 56°. The central rays of the excitation and emission beams intersect normally at the center of the cuvet holder for either configuration.

Hemoglobin Fluorescence

139

tinct from intact Hb or intact myoglobin and, therefore, unsuitable for this purpose. Apohemoglobin and apomyoglobin exhibit shifted fluorescence emission maximas consistent with findings that apohemoglobin is a dimer (27,28) and apomyoglobin is unfolded (29). Apohemoglobin derived from HbA exhibits a fluorescence emission maximum shifted ~14 nm to longer wavelengths. Obviously, in the quest for a blank, the buffer does not represent a true baseline for a heme-protein. Therefore, until a true blank is designed, absolute fluorescence quantum yields are unattainable. Yet, this limitation has not impeded the utility of heme-protein fluorescence analysis and interpretation. Since fluorescence polarization calculations are based on right-angle optics, only relative polarization and anisotropic measurements are meaningful with front-face fluorescence using adapters with quartz cuvets. Alternative frontface cell designs (Fig. 4B,C) may provide a means to reduce the distortions in anisotropic measurements that are introduced by front-face fluorometry (24). Front-face optics also have been used in time-resolved fluorescence measurements of hemoglobins (30). Generally, with state-of-the-art time-resolved fluorescence measurements, normal Hb exhibits a multiexponential decay that fits to three lifetimes. Since Trp itself exhibits a multiexponential decay, interpretation of the intrinsic fluorescence lifetimes (picosecond, subnanosecond, and nanosecond) remains controversial and is discussed at length in ref. 17. Nevertheless, the intrinsic and extrinsic fluorescence of Hb is an established property that is useful in probing structural perturbations in Hb. Noteworthy is that front-face fluorometry is clinically important for the detection of fluorescent components circulating in blood. The hematofluorometer was designed to measure blood levels of zinc protoporphyrin, which rises in lead poisoning and iron deficiency anemia, and blood levels of protoporphyrin IX, which may rise in porphyria diseases (e.g., erythropoietic protoporphyria) (31–33). Circulating bilirubin may also be detected by this method (34). Furthermore, front-face fluorometry has been useful in detecting the binding of porphyrins to Hb (35); correlating high zinc protoporphyrin levels in sickle Fig. 1. (continued) The front window of the cell is 0.5 mm thick and the sample thickness is 1 mm. This cuvet is advantageous for rare samples, requiring ~100–200 µL. (From ref. 19.) (B) Shown is the schematic of the rhombiform optical cell compartment. Components (a) and (b) are made of quartz, and (c) is the Hb sample. The solid line depicts the incident excitation light beam; the broken and dotted lines show the transmitted and the emitted light, respectively. θ is 52.4°, avoiding light direct excitation beam reflectance. (From ref. 21.) (C) Side view of free-surface cuvet. 1–3, Fixed quartz windows; 4, sliding quartz window; 5, metallic mirror; 6, body of the cover; 7, body of the cuvet; 8, supporting stem; 9, liquid sample; 10, O-rings. (From refs. 23 and 24.) (Composite figure from ref. 17.)

140

Hirsch

cell patients with low HbF levels (36); and detecting circulating fluorescent drugs, such as the antibiotic tetracycline (37), and drugs used in inflammatory bowel disease, such as the aminosalicylic acid derivatives that circulate in blood (38–39).

1.2. Intrinsic Hb Fluorescence Emission Is Sensitive to Tertiary and Quaternary Structural Alterations While Tyr fluorescence may be distinguished in a heme-protein containing both Tyr and Trp (14,29), Trp fluorescence emission generally predominates as a result of a greater quantum yield and fluorescence resonance energy transfer from Tyr to Trp. In general, Tyr and Trp are excited with 280 nm excitation light, while 296 nm selectively excites Trp (40). The contribution by Tyr may be dissected out by the difference spectrum (296 nm excitation emission spectrum –280 nm excitation emission spectrum) (40). The emission maximum of Tyr is ~305 nm, which is blue shifted from the emission maximum of Trp. It is well established that the Trp emission maximum is a function of the microenvironment: Trp in a hydrophilic environment or exposed environment exhibits an emission maximum at 350–353 nm, whereas the maximum for a hydrophobic or buried Trp is at 330–332 nm; Trp in limited contact with water exhibits an emission maximum at 340–342 nm (41). This environmental sensitivity of Trp arises from a large dipole change on excitation (42). Experimentally, the exact emission maximum wavelength may vary (up to ~5 nm) with the specific instrument employed. Evidence that the fluorescence signal emanates from intact Hb is supported in part by the ~330-nm emission wavelength, which is that of a buried Trp, consistent with the location of β37 Trp. By contrast, if α14 Trp and β15 Trp, which lie close to the surface, were the primary emitter, a longer wavelength emission maximum (~345–355 nm) corresponding to partially or fully exposed Trp would be expected. Likewise, if the Hb fluorescence emission originated from an exposed Trp, it should be quenched by KI. This is not the case (14). Similarly, if the emission arose from an apoglobin, the emission maximum would be shifted to longer wavelengths than that observed. An example of a shifted emission maximum when Trp is in an aqueous microenvironment is seen in the recombinant Hb α96 Val → Trp designed by Ho and collaborators (43,44). This Hb exhibits low oxygen affinity, high cooperativity, and no unusual subunit dissociation (43). The steady-state fluorescence emission maximum of Hb α96 Trp is of higher intensity and ~5 nm shifted to longer wavelengths in comparison with HbA (Fig. 5), suggesting that the additional α96 Trp is exposed to an aqueous environment. This prediction (as opposed to the one formulated by molecular dynamics simulation [43]) was confirmed by the high-resolution X-ray crystal structure showing the addi-

Hemoglobin Fluorescence

141

Fig. 5. Steady-state front-face fluorescence emission spectra of a Trp-plus Hb recombinant mutant compared with HbA: COHb α96 Val → Trp and COHbA. The solution was 0.05 M HEPES, pH 6.5, 25°C. Note the ~5-nm shift to longer wavelengths of the emission maximum (relative to COHbA in the same conditions). This wavelength shift indicates that the additional Trp is in a more exposed or hydrophilic environment, as confirmed by the high-resolution X-ray crystal structure (44). The emission maximum of HbA falls at ~320 nm, indicative that the primary emitter is that of a buried or hydrophobic Trp (e.g., β37 Trp). Both Hbs contain α14 Trp, β15 Trp, and β37 Trp, with the recombinant Hb containing the additional α96 Trp.

tional α96 Trp indole side chain directed away from the α1β2 interface and directed toward the water-filled central cavity (44).

1.2.1. Ligand Binding and Quaternary Structure Changes in Hb It is well established, by several different laboratories, that the steady-state fluorescence emission intensity is dependent on the R (oxy) → T (deoxy) transition (for a review, see refs. 16 and 17): an 18–25% increase in the fluorescence intensity is observed on deoxygenation (Fig. 6). This R → T transitional change in fluorescence intensity is sensitive to pH as modulated by inositol hexaphosphate (IHP) for carp Hb (45) and HbA (Fig. 7). These pH effects are observed in the range where Trp and Tyr fluorescence emission is pH insensitive (i.e., pH 3.0–11.0) (46). Therefore, the data presented in Fig. 7 are modulated by the R → T transition, which, to a degree, may be correlated with the Bohr effect.

142

Hirsch

Fig. 6. Hb fluorescence is a sensitive reporter of the R → T transition. The frontface intrinsic fluorescence emission of HbA varies as a function of ligand binding. All solutions are 0.155 mM Hb tetramer (pH 7.35), 0.05 M phosphate, 25°C. The lowest curve is the buffer solution. (From ref. 57.)

Relative fluorescence comparisons of variant hemoglobins to HbA are instructive. Conformational changes in non-Trp-substituted Hb mutants are reflected by emission differences compared to HbA. For example, fluorescence emission intensity differences are seen among R-state HbC (β6 Glu → Lys), HbS (β6 Glu → Val), and HbA. Coupled with functional studies, circular dichroism, differential fluorescence perturbations by allosteric effectors, and UV resonance Raman studies, significant differences in the A-helix and the central cavity of the β6 mutants were revealed (47–50). Spectroscopic solution studies may reveal fluctuations that are not seen in the crystal structure because of lattice constraints. This serves as an example of the importance of spectroscopy to reveal fluctuations that occur in solution but that may be constrained in the crystal structure and, thus, not observed with ~2Å resolution. Fluorescence spectroscopy combined with resonance Raman spectroscopy and functional studies revealed site-specific tertiary and quaternary differences in the Tyr-substituted Hb, Hb Montefiore (α126 Asp → Tyr), which exhibits high oxygen affinity and low cooperativity (51). OxyHb Montefiore exhibits a ~40% increase in fluorescence compared with oxyHbA. This difference is more than would be expected with the additional Tyr. The difference spectrum using

Hemoglobin Fluorescence

143

Fig. 7. The intrinsic fluorescence of HbA is sensitive to pH and is further modulated by IHP. (a) Plotted is the change in fluorescence intensity of deoxyHbA minus oxyHbA in the absence of IHP and presence of IHP (excitation: 280 nm; emission maximum: 330 nm). (b) Plotted is the difference between the change in the fluorescence intensity in the absence of IHP [deoxy – oxy] minus the fluorescence intensity change in the presence of IHP [deoxy – oxy]. Note that Trp or indole fluorescence intensity is not sensitive to pH within this range (46,81) (from Lin and Hirsch, unpublished data; ref. 82).

280-nm excitation exhibits a shoulder at ~308 nm, providing a spectral marker for α126 Tyr. Confirmation of this marker is shown by the difference emission spectrum which arises with 296-nm excitation selective for tryptophans, that does not show the ~308-nm shoulder. The large difference in the fluorescence compared to HbA indicates a conformational change or a change in the fluctuation properties of this Hb. The molecular alteration appears to be more significant in the T state than in the R state since the difference spectrum when subtracting deoxyHb Montefiore from deoxyHbA is significantly greater than

144

Hirsch

Fig. 8. oxyHb intrinsic fluorescence spectra as a function of synthetic allosteric effector DCBAA titration. A titration of oxyHbA with DCBAA, up to a molar ratio of 10:1 (DCBAA:Hb tetramer ratio), is monitored by decreases in the intrinsic fluorescence emission of Hb (pH 7.35, 0.1 M HEPES). Excitation wavelength: 280 nm (unpublished data; see ref. 53).

the difference spectrum of the oxy forms. The T-state fluorescence difference may result from the alteration in the important contact between α126 Asp and the α1β2 interface residues found in normal Hb, specifically β35 Tyr and β34 Val and the C-terminal 141 Arg. The probable perturbation of this contact by α126 Tyr and conformational alteration is likely to extend to the nearby β37 Trp that serves as the reporter of this altered T state. This destabilized T state may bind oxygen with higher affinity than deoxyHbA (51). Binding of allosteric effectors and their perturbation of Hb conformation may be monitored by front-face fluorometry. For example, the synthetic allosteric effector 3,4-dichloro-benzyloxy acetic acid (DCBAA) decreases the oxygen affinity of Hbs by binding specifically at the deoxyHb surface/crevice residue α14 Trp (52). The crystal structure of DCBAA bound to deoxyHbA demonstrates site-specific binding to the A-helix of deoxyHb, specifically α14 Trp (52). Thus, binding of DCBAA to Hb as monitored by Hb intrinsic fluorescence changes serves as a specific reporter of the A helix at α14 Trp. This is seen by probing distally to the β6 site of mutation in R states of HbC and HbS. The differences in oxyHb intrinsic fluorescence in response to DCBAA are shown in Fig. 8; oxyHbC and oxyHbS show a minimal decrease in the intensity of the intrinsic fluorescence emission maximum on titration with DCBAA. These titration results show that the fluorescence intensity changes of oxyHbC and oxyHbS, with increasing concentration of DCBAA, are just at the level of resolution of the instrument. By contrast, oxyHbA exhibits a larger decrease in fluorescence intensity (~6%) with clear resolution of the fluorescence intensity changes as a function of DCBAA titration. Consistent with the findings by

Hemoglobin Fluorescence

145

Mehanna and Abraham (52), the oxyHbA titration data indicate that DCBAA binds to R-state HbA in a 6:1 mole ratio. The titration curve (Fig. 8) exhibits a change of slope at a ratio of 2:1, which may imply two sites with different affinity and capable of allosteric change. The ability of DCBAA to serve as a fluorescence quencher is shown by the significant decrease in fluorescence of liver alcohol dehydrogenase on addition of DCBAA (53). The above data further demonstrate differences in the A-helix of HbC and HbS. A further conclusion that may be drawn from these studies is that the relatively small intrinsic fluorescence quenching of Hb on binding of DCBAA implies that α14 Trp is a small contributor to the overall intrinsic fluorescence signal emanating from HbA. Changes in the relative fluorescence intensity in the presence of allosteric effectors may also be explained as a function of the R → T transition. The 2,3diphosphoglycerate (DPG) analogs—IHP (e.g., Fig. 7) and the fluorescent 1,3hydroxypyrene-trisulfonate (HPT) (see Subheading 1.3.)—both induce changes in fluorescence intensity and have been useful in interpreting Hb fluorescence, comparing structural perturbations in Hb variants, as well as obtaining binding constants (54). Since the natural allosteric effectors, such as DPG and chloride, alter the conformation of Hb that is revealed as a perturbation in the fluorescence emission (49,55), it is important to strip Hb of these factors before embarking on such titration studies. The method to strip Hb is described in Subheading 3. Hb-reducing agents such as sodium dithionite should not be used for deoxygenation, because dithionite exhibits significant UV absorption and fluorescence that overlaps with Hb excitation and intrinsic emission. However, dithionite interference is minimal in the visible light range and may be used to deoxygenate Hb bound with an extrinsic fluorophore that emits visible light (56). Deoxygenation without dithionite is carried out by gently blowing nitrogen or helium over an Hb solution in a closed system for up to 90 minutes if needed, without met (Fe+3) Hb formation.

1.2.2. Hb Oxidation Intrinsic fluorescence emission intensity increases significantly (greater than twofold) on oxidation to methemoglobin compared with oxyHb and deoxyHb (54,57). In addition, the chemical oxidation of Hb may be monitored as a function of the generation of a fluorescent heme degradation end product using H2O2 as the oxidizing agent. The fluorescent end-product exhibits an emission maximum at 465 nm with 321-nm excitation (58–60). The coupling of frontface fluorometry with this method permits comparison of the oxidation propensity of various Hb mutants at concentrations at which the Hb remains a tetramer (55).

146

Hirsch

1.2.3. Hb Dissociation Hb fluorescence studies using right-angle optics require low Hb concentrations. Generally, when right-angle optics are employed, low protein concentrations (on the order of micromolar) are required. However, in the case of intact HbA (a tetramer composed of two identical α- and two identical β-chains) significant dissociation to dimers occurs at low concentration in the oxy or R-state forms: under conditions of moderate ionic strength (~0.1 M NaCl), KD = 1.0 × 10–6 M. In the case of deoxyHb, dissociation is significantly less (KD = 2.0 × 10–11 M) (61–62). Note that a high salt concentration will shift the dissociation equilibrium to dimers (27,63–64). The percentage of dimer dissociation (α) is calculated by (64): α = {KDM4[1 + (16c/KDM4)] – KDM4}/8c

in which KD is the dissociation constant in molarity, M4 is the molecular weight of the tetramer, and c is the concentration in grams/liter. Dimerization results in emission maximum shifts to longer wavelengths as a result of Trp exposure to a more hydrophilic environment (65–66). This concentration-dependent dissociation (1) complicates the interpretation and comparison of Hb fluorescence studies performed under different solution conditions, and (2) highlights the advantage of using front-face optics which reduces inner-filter effects that arise in a strongly absorbing solution. Reversible protein dissociation and unfolding may be investigated using high hydrostatic pressure (up to ~2 bBar) coupled to fluorescence (67). A number of laboratories have utilized steady-state and anisotropic fluorescence measurements in conjunction with the application of high pressure to study the dissociation properties of variant hemoglobins (68–72). In summary, Hb in its native state is a tetramer, and unwarranted solution conditions (e.g., dilute solutions, high salt concentration) can shift the equilibrium to the preponderance of dimers that differ in structure and function; some Hb variants are more prone to dissociation (73); and the apoprotein (i.e., globin without hemes) results in a distinct structure without resemblance to the native tetramer. Dissociation and apoglobin formation may be revealed by front-face fluorometry. The fluorescent moiety 1-anilinonaphthalene-8-sulfonic acid (ANS), known to bind to the empty heme pocket, was first used to demonstrate that the fluorescence arises from intact Hb (13).

1.3. Extrinsic Fluorescence The site-specific labeling of proteins with extrinsic fluorescent probes allows spectroscopic probing of side chains of amino acid residues that are nonfluorescent and may serve as a ruler to measure intra- and intermolecular distances. Front-

Hemoglobin Fluorescence

147

face fluorometry permits direct monitoring of the probe bound to a hemeprotein. For example, iodoacetamidofluorescein covalently modifies β93 Cys (the nearest neighbor to the proximal His of the heme) and is responsive to the R → T transition (74). Fluorescence quenching studies of HPT when bound to Hb demonstrated that HPT serves as a DPG analog, binding noncovalently (one per Hb tetramer) to the central cavity (11,75–77). DPG is the natural allosteric effector of the red blood cell that binds to Hb in a 1:1 ratio, effectively lowering the oxygen affinity so that sufficient oxygen is released to tissues with low oxygen saturation. Direct monitoring of HPT fluorescence is a means to explore the Hb central cavity and has been useful in demonstrating (1) structural alterations distal from the site of mutation, such as HbS (β6 Glu → Val) and HbC (β6 Glu → Lys) (48); and (2) central cavity differences anticipated in crosslinked Hbs designed as potential therapeutic oxygen carriers (78). Another fluorescent DPG analog, 1,3,6,8-pyrenetrisulfonate, may be advantageous for time-resolved fluorescence studies probing the DPG pocket; this application revealed a long-range communication from the positively charged substitution in the middle of the central cavity of Hb Presbyterian (β108 Asn → Lys) to the DPG-binding pocket that lies at the entrance to the ββ cleft (79). In summary, intrinsic and extrinsic fluorescence properties of Hb are valuable reporters of site-specific conformational changes. 2. Materials 1. 2. 3. 4. 5. 6. 7. 8.

Purified hemoglobins. Fluorescence spectrophotometer with UV excitation and emission monochrometers. Front-face adapter or specially designed cuvets for front-face optics. Cuvet washer. 100% Ethyl alcohol contained in glass. Nitric acid. Anaerobic glove box or glove bags. Fluorescence probes (e.g., 5-iodoacetamidofluorescein, cat. no. I-3; Molecular Probes, Eugene OR; or 1,3-hydroxypyrene trisulfonate, H-1529; Sigma, St. Louis, MO).

3. Methods (see Notes 1–9) 3.1. Purification of Hemoglobins For reasons already noted, it is critical to this technique that the Hb be homogeneously purified, and that a concentration be selected wherein it remains intact in its tetrameric native state containing all four hemes. Depending on the Hb mutant, Hb may be separated and purified by liquid column chromatography using anion- and cation-exchange resins. The red cell

148

Hirsch

lysate (hemolysate) contains 95% Hb. It also contains a small percentage of other molecules, cofactors that interact with Hb, and a variety of enzymes and proteins including minor Hbs (e.g., HbF, HbA2, HbA1a1, HbA1a2, HbA1b, HbA1c). Thus, it is imperative that these other proteins and compounds be removed. This is specifically addressed by Pin et al. (70). Therefore, for meaningful comparisons, the proteins should be purified and handled according to one method of choice (see ref. 80). (Hb purification is further developed in Chapters 3, 6, 10, and 14.) Purified Hbs are stripped (i.e., DPG and various ions are removed) by Sephadex G-25 gel filtration chromatography. Stripping twice ensures the preparation of fully stripped Hbs, which may be assessed by oxygen equilibrium methods: it is known that stripped Hb has a higher oxygen affinity compared with nonstripped Hb.

3.2. Covalent Modification with Fluorescent Probes As an example, iodoacetamidofluorescein (5-IAF) (cat. no. I-3, Molecular Probes) covalently binds to sulfhydryl groups. β93 labeling of Hb may be effected by a slight modification of a procedure described earlier (74). A purified Hb solution (~3–5 g%, 5 mL) in the presence of an ~5-IAF:1 heme in the desired buffer is incubated for 3 hours at 4°C with gentle hand rotation every 20–30 min. After 3 h, the solution is added to a Sephadex G-25 column (50-mL volume) equilibrated in the desired buffer. The solution is collected, repurified, and concentrated in Centricon 10 (YM 10, Amicon) three times or until no fluorescence is found in the dialysate as detected by a fluorometer. Isoelectric focusing and mass spectrometry are used to assess complete modification and to ensure purity. An inability to titrate with paramecuricbenzoic acid is another method to verify that the reactive —SH groups are all bound. However, if available, mass spectrometry is the method of choice for determination of site-specific and complete modification. Functional alterations that may arise as a result of site-specific labeling are examined by oxygen-equilibrium methods. 4. Notes 1. Quartz cuvets are required for UV excitation light and emission of intrinsic fluorescence. 2. At millimolar Hb concentrations, the Raman band should not interfere with the emission spectrum as observed with dilute aqueous solutions. 3. Do not store liquid buffers or Hb solutions in plastic tubes for any length of time (hours); plastic derivatives that scatter and fluoresce may be introduced into the solutions. 4. The choice of buffer is important, with 0.05 M HEPES is optimal. Phosphates may bind in the DPG pocket of the central cavity. Tris buffers should be avoided

Hemoglobin Fluorescence

5.

6.

7.

8.

9.

149

because Tris exhibits fluorescence (although this may not be the case with ultrapurified Tris). Therefore, the buffer baseline should always be recorded to ensure no background fluorescence from the buffer or from any residual material in the cuvet. Cuvets may be soaked with 50% nitric acid for several minutes up to a few hours. If soaked longer or if higher concentrations of nitric acid are used longer than a few minutes, etching of the cuvet is possible. The use of detergents is not recommended because they may contribute to the fluorescence background. The cuvet should be rinsed thoroughly with distilled water, followed by a final rinse with 100% ethyl alcohol, and dried on a vacuum-cuvet washer. Pipeting samples into the cuvet should be carefully done without scratching. If necessary, soft tubing may be added to the end of a needle or hard pipet tip. The cuvets should be held at the edges to avoid fingerprints. As for any fluorescence study, the same cuvet should be used for relative comparisons. Moreover, the cuvet must be placed in the identical position during measurements. Excitation sources such as intense lasers or light arising from a synchrotron may cause heating, which can interfere with the stability of the protein. This may be avoided by rotating the cuvet, replacing the sample after every third scan, or recording spectra in a closed flowing cell. Fluorescence intensity is temperature dependent (i.e., the higher the temperature, the lower the intensity), and, therefore, temperature must be kept constant during the course of the experiment and for relative comparisons. For studies of temperature-sensitive mutants, the fluorescence difference as a function of temperature between the wild-type protein and the temperature-sensitive variant must be factored in.

Acknowledgments This work was supported in part by the American Heart Association, Heritage Affiliate Grant-in-Aid No. 0256390T; and the National Institutes of Health R01 HL58038 and RO1 HL58247. References 1. Weber, G. (1953) Rotational brownian motion and polarization of the fluorescence of solutions. Adv. Protein. Chem. 8, 415–459. 2. Teale, F. W. J. and Weber, G. (1957) Ultraviolet fluorescence of the aromatic amino acids. Biochem. J. 65, 476–482. 3. Teale, F. W. J. and Weber, G. (1959) Ultraviolet fluorescence of proteins. Biochem. J. 72, 156. 4. Teale, F. W. J. (1960) The ultraviolet fluorescence of proteins in neutral solution. Biochem. J. 76, 381–388. 5. Stryer L. (1978) Fluorescence energy transfer as a spectroscopic ruler. Annu. Rev. Biochem. 47, 819–846.

150

Hirsch

6. Lakowicz, J. R. (1999) Principles of Fluorescence Spectroscopy, 2nd ed., Kluwer Academic/Plenum, New York. 7. Weber, G. and Teale, F. W. J. (1959) Electronic energy transfer in haem proteins. Disc. Faraday Soc. 28, 134–141. 8. Nagel, R. L. and Gibson, Q. H. (1967) Kinetics and mechanism of complex formation between hemoglobin and haptoglobin. J. Biol. Chem. 242, 3428–3434. 9. Nagel, R. L. and Gibson, Q. H. (1971) The binding of hemoglobin to haptoglobin and its relation to subunit dissociation of hemoglobin. J. Biol. Chem. 246, 69–73. 10. Benesch, R. E., Ikeda, S., and Benesch, R. (1976) Reaction of haptoglobin with hemoglobin covalently cross-linked between the alpha beta dimers. J. Biol. Chem. 251, 465–470. 11. Marden, M. C., Hazard, E. S., and Gibson, Q. H. (1986) Testing the two-state model: anomalous effector binding to human hemoglobin. Biochemistry 25, 7591–7596. 12. Alpert, B. and Lopez-Delgado, R. (1976) Fluorescence lifetimes of haem proteins excited into the tryptophan absorption band with synchrotron radiation. Nature 263, 445, 446. 13. Alpert, B., Jameson, D., and Weber, G. (1980) Tryptophan emission from human hemoglobin and its isolated subunits. Photochem. Photobiol. 31, 1–4. 14. Hirsch, R. E., Zukin, R. S., and Nagel, R. L. (1980) Intrinsic fluorescence emission of intact oxy hemoglobins. Biochem. Biophys. Res. Commun. 93, 432–439. 15. Gryczynski, Z., Tenenholz, T., and Bucci, E. (1992) Rates of energy transfer between tryptophans and hemes in hemoglobin, assuming that the heme is a planar oscillator. Biophys. J. 63, 648–653. 16. Hirsch, R. E. (1994) Front-face fluorescence spectroscopy of hemoglobins. Methods Enzymol. 232, 231–246. 17. Hirsch, R. E. (2000) Heme protein fluorescence, in Topics in Fluorescence Spectroscopy, vol. 6, Protein Fluorescence (Lakowicz, J. R., ed.), Kluwer Academic/ Plenum, New York, pp. 221–255. 18. Fontaine, M. P., Jameson, D. M., and Alpert, B. (1980) Tryptophan-heme energy transfer in human hemoglobin: dependence upon the state of the iron. FEBS Lett. 116, 310–314. 19. Eisinger, J. and Flores, J. (1979) Front-face fluorometry of liquid samples. Analytical Biochem. 94, 15–21. 20. Bucci, E., Malak, H., Fronticelli, C., Gryczynski, I., Laczko, G., and Lakowicz, J. R. (1988) Time-resolved emission spectra of hemoglobin on the picosecond timescale. Biol. Chem. 32, 187–198. 21. Horiuchi, K. and Asai, H. (1980) Binding of β-naphthyl triphosphate to human adult hemoglobin accompanying deoxygenation. Investigated by simultaneous measurements of fluorescence, absorbance and partial pressure of oxygen. Eur. J. Biochem. 131, 613–618. 22. Asakura, T., Agarwal, P. I., Relman, D. A., McCray, J. A., Chance, B., Schwartz, E., Friedman, S., and Lubin, B. (1973) Mechanical instability of the oxy-form of sickle haemoglobin. Nature 244, 437–438.

Hemoglobin Fluorescence

151

23. Bucci, E., Gryczynski, Z., Fronticelli, C., Gryczynski, I., and Lakowicz, J. R. (1992) Fluorescence intensity and anisotropy decays of the intrinsic tryptophan emission of hemoglobin measured with a 10-Ghz fluorometer using front-face geometry on a free liquid surface. J. Fluorescence 2, 29–36. 24. Gryczynski, Z. and Bucci, E. (1993) A new front-face optical cell for measuring weak fluorescent emissions with time resolution in the picosecond time scale. Biophys. Chem. 48, 31–38. 25. Elbaum, D., Harrington, J., Roth, E. F. Jr., and Nagel, R. L. (1976) Surface activity of hemoglobin S and other human hemoglobin variants. Biochim. Biophys. Acta 427, 57–69. 26. Hirsch, R. E., Elbaum, D., Brody, S. S., and Nagel, R. L. (1980) Hemoglobin-A and Hemoglobin-S films at an air-water interface: absorption spectra studies. J. Colloid. Interface Sci. 78, 212–216. 27. Antonini, E. and Brunori, M. (1971) Hemoglobin and Myoglobin in Their Reactions with Ligands. North Holland, Amsterdam. 28. Sassaroli, M., Bucci, E., Leisegang, J., Fronticelli, C., and Steiner, R. F. (1984) Specialized functional domains in hemoglobin: dimensions in solution of the apohemoglobin dimer labeled with fluorescein iodoacetamide. Biochemistry 23, 2487–2491. 29. Hirsch, R. E. and Peisach, J. (1986) A comparison of the intrinsic fluorescence of red kangaroo, horse and sperm whale met-myoglobins. Biochim. Biophys. Acta 872, 147–153. 30. Gryczynski, Z., Lubkowski, J., and Bucci, E. (1997) Intrinsic fluorescence of hemoglobins and myoglobins. Methods Enzymol. 278, 38–69. 31. Blumberg, W. E., Eisinger, J., Lamola, A. A., and Zuckerman, D. M. (1977) The hematofluorometer. Clin. Chem. 23(2 Pt. 1), 270–274. 32. Blumberg. W. E., Doleiden, F. H., and Lamola, A. A. (1980) Hemoglobin determined in 15 microL of whole blood by “front-face” fluorometry. Clin. Chem. 26, 409–413. 33. Lamola, A. A. (1981) Fluorescence methods in the diagnosis and management of diseases of tetrapyrrole metabolism. J. Invest. Dermatol. 77, 114–121 34. Cashore, W. J., Oh, W., Blumberg, W. E., Eisinger, J., and Lamola, A. A. (1980) Rapid fluorometric assay of bilirubin and bilirubin binding capacity in blood of jaundiced neonates: comparisons with other methods. Pediatrics 66, 411–416. 35. Hirsch, R. E., Lin, M. J., and Park, C. M. (1989) The interaction of zinc protoporphyrin with intact oxy hemoglobin. Biochemistry 28, 1851–1855. 36. Hirsch, R. E., Pulakhandam, U. R., Billett, H. H., and Nagel, R. L. (1991) Blood zinc protoporphyrin is elevated only in sickle cell patients with low fetal hemoglobin. Am. J. Hematol. 36, 147–149. 37. Park, C. M, Pulakhandan, U. R., and Hirsch, R. E. (l986) The interference of fluorescent drugs with the determination of zinc protoporphyrin levels in humans: The case of tetracycline. Clin. Res. 34(2), 466A. 38. Hirsch, R. E., Lin, M. J., and Das, K. M. (1990) The estimation of 5-aminosalicylic acid and its metabolite in human serum by front-face fluorometry: a simple and sensitive method. J. Lab. Clin. Med. 116, 45–50.

152

Hirsch

39. Ritland, S. R., Leighton, J. A., Hirsch, R. E., Morrow, J. D., and Gendler, S. J. (1999) Evaluation of 5-aminosalicylic acid (5-ASA) for cancer chemoprevention: Absence of efficacy against nascent adenomatous polyps in the ApcMin mouse. Clin. Cancer Res. 5, 855–863. 40. Eisinger, J. (1969) Intramolecular energy transfer in adrenocorticotropin. Biochemistry 8, 3902–3908. 41. Burstein, E.A, Vedenkina, N. S., and Ivkova, M. N. (1973) Fluorescence and the location of tryptophan residues in protein molecules. Photochem. Photobiol. 18, 263–279. 42. Vivien, J. T. and Callis, P. R. (2001) Tryptophan fluorescence shift mechanisms in proteins: simulation study of Trp rotational conformers. Biophys. J. 80(Pt. 1), 362a. 43. Kim, H. W., Shen, T. J., Ho, N. T., Zou, M., Tam, M. F., and Ho, C. (1995) A novel low oxygen affinity recombinant hemoglobin (α96 Val → Trp): switching quaternary structure without changing the ligation state. J. Mol. Biol. 248, 867–882. 44. Puius, Y. A., Zou, M., Ho, N. T., Ho, C., and Almo, S. C. (1998) Novel water mediated hydrogen bonds as the structural basis for the low oxygen affinity of the blood substitute candidate rHb (α96 Val → Trp). Biochem. USA 37, 9258–9265. 45. Hirsch, R. E. and Noble, R. W. (l987) Intrinsic fluorescence of carp hemoglobin: A study of the R → T transition. Biochim. Biophys. Acta 914, 213–219. 46. Williams, R. T. and Bridges J. W. (1964) Fluorescence of solutions: A review. J. Clin. Pathol. 17, 371–394. 47. Hirsch, R. E., Lin, M. J., Vidugiris, G. J., Huang, S., Friedman, J. M., and Nagel, R. L. (1996) Conformational changes in oxyhemoglobin C (β6 Glu → Lys) detected by spectroscopic probing. J. Biol. Chem. 271, 372–375. 48. Hirsch, R. E., Juszczak, L. J., Fataliev, N. A., Friedman, J. M., and Nagel, R. L. (1999) Solution-active structural alterations in liganded hemoglobins C (β6 Glu → Lys) and S (β6 Glu → Val). J. Biol. Chem. 274, 13,777–13,782. 49. Sokolov, L. and Mukerji, I. (1998) Conformational changes in FmetHbS probes with UV resonance Raman and fluorescence spectroscopic methods. J. Phys. Chem. B. 102, 8314–8319. 50. Juszczak, L.J, Hirsch, R. E., Nagel, R. L., and Friedman, J. M. (1998) Conformational differences in CO derivatives of HbA, HbC (E 6K) and HbS (E 6V) in the presence and absence of inositol hexaphosphate (IHP) detected using ultraviolet resonance Raman spectroscopy. J Raman Spectrosc. 29, 963–968. 51. Wajcman, H., Kister, J., Galacteros, F., Spielvogel, A., Lin, M. J., Vidugiris, G. J. A., Hirsch, R. E., Friedman, J. M., and Nagel, R. L. (1996) Hb Montefiore [α126(H9)Asp → Tyr]: high oxygen affinity and loss of cooperativity secondary to C-terminal disruption. J. Biol. Chem. 271, 22,990–22,998. 52. Mehanna, A. S. and Abraham, D. J. (1990) Comparison of crystal and solution hemoglobin binding of selected antigelling agents and allosteric modifiers. Biochemistry 29, 3944–3952. 53. Hirsch, R. E., Juszcak, L. J., Abraham, D. J., Friedman, J. M., and Nagel, R. L. (1997) Further evidence for solution-active structural differences in the β6 mutants HbC and HbS. Blood 90(10, Suppl. 1, Pt. 1), 126a.

Hemoglobin Fluorescence

153

54. Mizukoshi, H., Itoh, M., Matsukawa, S., Mawatari, K., and Yoneyama, Y. (1982) Tryptophan fluorescence of human hemoglobin. II. Effect of inositol hexaphosphate on the T-R transition. Biochim. Biophys. Acta 700, 143–147. 55. Chen, Q. Y., Bonaventura, C., Nagel, R. L., and Hirsch, R. E. (2002) Distinct domain responses of R-state human hemoglobins A, C, and S to anions. Blood Cells Mol Dis. 29, 119–132. 56. Hirsch, R. E., and Nagel, R. L. (1989) Stopped-flow front-face fluorometer: a prototype design to measure hemoglobin R → T transition kinetics. Anal. Biochem. 176, 19–21. 57. Hirsch, R. E. and Nagel, R. L. (1981) Conformational studies of hemoglobins using intrinsic fluorescence measurements. J. Biol. Chem. 256, 1080–1083. 58. Nagababu, E. and Rifkind, J. M. (1998) Formation of fluorescent heme degradation products during the oxidation of hemoglobin by hydrogen peroxide. Biochem. Biophys. Res. Commun. 247, 592–596. 59. Nagababu, E. and Rifkind, J. M. (2000) Heme degradation during autooxidation of oxyhemoglobin. Biochem. Biophys. Res. Commun. 273, 839–845. 60. Nagababu, E., Chrest, F. J., and Rifkind, J. M. (2000) The origin of red cell fluorescence caused by hydrogen peroxide treatment. Free Radic. Biol. Med. 29, 659–663. 61. Ackers, G. K., Johnson, M. L., Mills, F. C., and Ip, S. H. (1976) Energetics of oxygenation-linked subunit interactions in human hemoglobin. Biochem. Biophys. Res. Commun. 69, 135–142. 62. Imai, K. (1982) Allosteric Effects in Hemoglobin, Cambridge University Press, New York. 63. Bunn, H. F. and Forget, B. G. (1986) Hemoglobin: Molecular, Genetic and Clinical Aspects, W. B. Saunders, Philadelphia. 64. Herskovits, T. T., Cavanagh, S. M., and San George, R. C. (1977) Light-scattering investigations of the subunit dissociation of human hemoglobin A: effects of various neutral salts. Biochemistry 16, 5795–5801. 65. Chothia, C., Wodak, S., and Janin, J. (1976) Role of subunit interfaces in the allosteric mechanism of hemoglobin. Proc. Natl. Acad. Sci. USA 73, 3793–3797. 66. Hirsch, R. E., Squires, N. A., Discepola, C., and Nagel, R. L. (1983) The detection of hemoglobin dimers by fluorescence. Biochem. Biophys. Res. Commun. 116, 712–718. 67. Pin, S. and Royer, C. A. (1994) High-pressure fluorescence methods for observing subunit dissociation in hemoglobin. Methods Enzymol. 232, 42–55. 68. Marden, M. C., Hoa, G. H. B., and Stetzkowski-Marden, F. (1986) Heme protein fluorescence versus pressure. Biophys. J. 49, 619–627. 69. Silva, J. L., Villas-Boas, M., Bonafe, C. F. S., and Meirelles, N. C. (1989) Anomalous pressure dissociation of large protein aggregates. J. Biol. Chem. 264, 15,863–15,868. 70. Pin, S., Royer, C. A., Gratton, E., Alpert, B., and Weber, G. (1990) Subunit interactions in hemoglobin probed by fluorescence and high-pressure techniques. Biochemistry 29, 9194–9202. 71. Hirsch, R. E., Harrington, J. P., and Scarlata, S. F. (1993) The differential effects of carbon dioxide and oxygen on the pressure dissociation of Lumbricus terrestris hemoglobin. Biochim. Biophys. Acta 1161, 285–290.

154

Hirsch

72. Hirsch, R. E., Friedman, J. M., Harrington, J. R., and Scarlata, S. F. (1994) Stability of a potential blood substitute, HbXL99α under high pressure. Biochem. Biophys. Res. Commun. 200, 1635–1640. 73. Sharma, V. S., Newton, G. L., Ranney, H. M., Ahmed, F., Harris, J. W., and Danish, E. H. (1980) Hemoglobin Rothschild (β 37(C3)Trp replaced by Arg): A high/low affinity hemoglobin mutant. J. Mol. Biol. 144, 267–280. 74. Hirsch, R. E., Zukin, R. S., and Nagel, R. L. (1986) Steady-state fluorescence emission from the fluorescent probe, 5-iodoacetamidofluorescein, bound to hemoglobin. Biochem. Biophys. Res. Commun. 138, 489–495. 75. MacQuarrie, R. and Gibson, Q. H. (1971) Use of a fluorescent analogue of 2,3diphosphoglycerate as a probe of human hemoglobin conformation during carbon monoxide binding. J. Biol. Chem. 246, 5832–5835. 76. MacQuarrie, R. and Gibson, Q. H. (1972) Ligand binding and release of an analogue of 2,3-diphosphoglycerate from human hemoglobin. J. Biol. Chem. 247, 5686–5694. 77. Serbanescu, R., Kiger, L., Poyart, C., and Marden, M. C. (1998) Fluorescent effector as a probe of the allosteric equilibrium in methemoglobin. Biochim. Biophys. Acta 1363, 79–84. 78. Gottfried, D. S., Juszczak, L. J., Fataliev, N. A., Acharya, A. S., Hirsch, R. E., and Friedman, J. M. (1997) Probing the hemoglobin central cavity by direct quantification of effector binding using fluorescence lifetime methods. J. Biol. Chem. 272, 1571–1578. 79. Gottfried, D. S., Manjula, B. N., Malavalli, A, Acharya, A. S., and Friedman, J. M. (1999) Probing the diphosphoglycerate binding pocket of HbA and HbPresbyterian (β 108 Asn → Lys). Biochemistry 38, 11,307–11,315. 80. Schroeder, W. A. and Huisman, T. H. J. (1980) The Chromatography of Hemoglobin, Marcel, Dekker, New York. 81. White, A. (1959) Effect of pH on fluorescence of tyrosine, tryptophan, and related compounds. Biochem. J. 71, 217–220. 82. Lin, M. J., Rao, M. J., Friedman, J. M., Acharya, A. S., and Hirsch, R. E. (1991) Inositol hexaphosphate induced pH sensitive conformational changes of the α1β2 interface of hemoglobin. Biophys. J. 59(2), 290a.

Nucleation and Crystal Growth of HbC

155

10 Nucleation and Crystal Growth of Hemoglobins The Case of HbC Peter G. Vekilov, Angela Feeling-Taylor, and Rhoda Elison Hirsch 1. Introduction Hemoglobin (Hb) crystallization is of significance both in vivo and in vitro. Hb crystals form in red blood cells (RBCs), as occurs in the case of patients expressing βC-globin (β6 Glu → Lys). In vitro, high-resolution structural determination by crystallographic methods requires the growth of Hb crystals to approx ~1 mm in diameter, which may be induced by a variety of precipitants. The first indication that oxyHb and deoxyHb exhibit two distinct conformational states and different crystal habits was first observed in 1938 by Haurowitz (1), who noted cracking of the horse deoxyHb crystal on oxygenation. Years later, Perutz and colleagues (2–4) obtained the high-resolution structure of deoxy (T-state) and oxy (R-state) human HbA. Since then, crystal diffraction methodology has continued to provide high-resolution details of structural alterations that occur on point mutations or peptide changes in natural Hb mutants or engineered recombinant variants (see Chapter 1).

1.1. HbC Forms Crystals in RBCs HbC is the second most commonly encountered abnormal Hb in the United States and, next to HbS and HbE, the third most prevalent hemoglobin structural variant worldwide (5,6). Approximately 3 of 100 African Americans carry the HbC gene. Individuals homozygous for HbC exhibit a mild hemolytic anemia, not considered a life-threatening disease. However, double heterozygotes for both HbS and HbC have sickle cell (SC) disease, which results in a reduced life expectancy and significant morbidity. It is life-threatening after the age of From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

155

156

Vekilov et al.

20, and some patients have severe retinal, osteonecrotic, and pulmonary complications (for a review, see ref. 5). It has long been known that HbC (β6 Glu → Lys) forms tetragonal crystals in red cells of CC patients (individuals homozygous for the expression of βC-globin) without resulting in morbid pathophysiology. In the venous blood of splenectomized patients, ~3% of RBCs contain crystals with all of the Hb recruited into a single crystal (7–9). Unknown was the Hb state in which the intraerythrocytic crystal formed. In 1985, it was first demonstrated by video-enhanced microscopy, (1) that these intraerythrocytic HbC crystals formed in the oxygenated (R) liganded state in cells that contained no detectable fetal hemoglobin (HbF), and (2) that deoxygenation (or switching to the T-state) resulted in dissolution of the intraerythrocytic crystal (10). The lack of vasoocclusion in CC patients could be explained as follows: Deoxygenation in the microcirculation (secondary to oxygen delivery to tissues) will result in the dissolution of the oxyHbC crystals, avoiding vasoocclusion by allowing the CC red cell to regain its pliability, necessary to navigate through narrow capillaries. Individuals coexpressing HbS and HbC exhibit a moderately severe disease, SC disease, arising from the polymerization of deoxyHbS induced by the increased intracellular hemoglobin concentration (6,11). Yet, along with RBCs containing HbS polymers, intraerythrocytic crystals are also detected in SC patients (12). To date, little is known about the mechanism of oxy (R-state) HbC crystallization nor why the oxy form of HbC crystallizes in the red cell whereas the deoxy form of HbS polymerizes in the red cell. Our laboratories have taken several different approaches to elucidate the mechanisms of liganded HbC crystallization.

1.2. In Vitro Batch Nucleation Studies Since CC erythrocytes containing HbF did not contain crystals (10), in vitro batch nucleation were undertaken to determine the effects of co-habiting hemoglobins in the RBC on HbC crystallization (13). In vitro, purified HbC forms tetragonal crystals within 15–30 min in concentrated phosphate buffer (1.8 M) at 30°C (Fig. 1). The size of the crystal is dependent on nucleation kinetics (i.e., the faster the nucleation rate, the smaller the crystal). Similar solution conditions were first employed by Adachi and colleagues (14,15) to study deoxyHbS nucleation and polymerization. A lag phase (on the order of 15–30 minutes) is always seen to precede any observation of crystal formation from a purified HbC solution using the methods outlined below (Fig. 2A,B). In agreement with the general expectations about any nucleation process (16–18), the length of the lag phase for nucleation and crystallization is dependent on the supersaturation of the solution (19,20).

Nucleation and Crystal Growth of HbC

157

Fig. 1. Tetragonal oxy or CO (R-state) HbC crystals form in concentrated potassium phosphate buffer (1.8 M, pH 7.35). The Hb concentration is 2 g% (magnification: ×1000).

These batch crystallization methods demonstrated in vitro that HbF inhibits nucleation (Fig. 2) (13), while HbS (sickle cell Hb, β6 Glu → Val) accelerated nucleation and is incorporated into the crystal (21). HbA simply serves as a diluent in these nucleation studies (13,21). The value of using in vitro batch crystallization to identify contact sites (Table 1) and the effects on crystallization of binary mixtures of Hb and cell components soon became apparent (13,21–26). Interestingly, intraerythrocytic crystal morphology is altered in heterozygous individuals expressing HbC and other point mutated globins: those expressing HbC and Hb Korle-bu (β73 Asp → Asn) form cubic crystals (22), while those expressing HbC and Hb αG-Philadelphia (β68 Asn → Lys) form unusually long, narrow crystals (24). Different intertetrameric contacts are implied and may give rise to the altered morphology in a manner analogous to that proposed by Gallagher et al. (27).

158

Vekilov et al.

Fig. 2. (a) Effects of HbA and HbF on kinetics of nucleation. (A) Dependence of the total number of crystals nucleated in fixed solution volume on concentrations of HbA (upper curve) and HbF (lower curve). (B) Time dependence of number of crystals nucleated in a crystallization cell in presence of concentrations of HbA and HbF indicated on plots. Note that HbA and HbF increase the time lag and decrease the number of nucleated crystals; that is, both proteins decrease the rate of nucleation, with HbF having a markedly stronger action. (From ref. 13).

The question of why R-state HbC gives rise to crystals, in contrast to polymer formation when HbS is deoxygenated (T state), is addressed by a variety of approaches. Front-face fluorescence and ultraviolet resonance Raman spectroscopic studies, employed to probe intratetrameric R-state differences in HbS and HbC compared with HbA, at nonaggregating concentrations, suggest intramolecular alterations in the A-helix position and in the central cavity. (For methodological details of front-face fluorescence, see Chapter 9.) Video-enhanced differential interference contrast (DIC) microscopy compared aggregation/crystallization pathways of oxyHbC and deoxyHbC in concentrated salt conditions. It was demonstrated that R-state HbC exhibits a large propensity to crystallize in a tetragonal habit, whereas under similar conditions, deoxyHbC is driven to form a variety of morphological aggregates (e.g., radial arrays and macroribbons) while hexagonal crystal formation is a rare and least favored pathway (28).

1.3. In Vitro Solubility Studies Further quantification of R-state HbC crystal growth parameters, such as protein solubility and its dependence on temperature, was assessed by a novel scintillation method (29). An exploded view of the new scintillation arrangement is shown in Fig. 3A. The solution is contained in a silica microcell. This

Nucleation and Crystal Growth of HbC

159

Table 1 Nucleation and Crystallization Effects by Binary Mixtures of HbC with Hb Variants HbC compound Heterozygote

Site-specific substitution

Effect on oxyHbC nucleation and crystallization

HbS Hb Korle-bu HbF Hb N-Baltimore Hb Riyadh HbαGPhiladelphia

β6 Glu→Val β73 Asp→Asn β87 Thr→Gln β95 Lys→Glu β120 Lys→Asn α68 Asn→Lys

Accelerates Accelerates Inhibits Accelerates Inhibits Accelerates

Hb J-Baltimore

β16 Gly→Asp

Accelerates

Crystal morphology Tetragonal Cubic-like — Tetragonal — Tetragonal and numerous 3X elongated Tetragonal (in vitro)

Reference 13 21 22 23 23 24

25

cell is surrounded by a machined brass jacket that sits on a thermoelectric heat pump (Peltier cooler) connected to a programmable controller. The use of Peltier elements to maintain temperature increases the temperature’s stability to about ± 0.02°C and allows temperature ramps at rates of up to 5°C/min. The temperature of the brass block is monitored with a type-T thermocouple. The attached controller facilitates programmed temperature changes. A laser beam from a self-contained laser diode assembly is directed through the solution. Light scattered normally to the incident beam is detected by an integrated detector/amplifier photodiode through a polished rod capping the microcell. A small segment of latex rubber tubing envelops the microcell and the cap and minimizes evaporation of the solvent. A beam splitter between the laser and cell diverts some of the laser’s output to a second integrated photodiode, whose signal is used to correct for intensity fluctuations of the laser. Backscatter of light, which passes through the cell, is minimized by painting the inside of the brass block black and by a light trap in the cavity formed by a rotating horseshoe magnet. This magnet drives a small nickel wire (0.6 mm in diameter, 5–7 mm long) used as a stirring bar inside the lower part of the solution, not illuminated by the laser. The output from the two integrated photodiodes is amplified and filtered with a four-pole Butterworth-style circuit located on a circuit board mounted to the nylon block, which minimizes the signal path length. Low-noise, low-drift/offset (0.6 mV/°C, 25 mV) precision op-amps, along with low-temperature coefficient (<100, 50 ppm/°C) trimming resistors and surface-mount resistors and capacitors are used.

160

Vekilov et al.

Fig. 3. Determination of Hb solubility using the miniaturized scintillation technique. (a) Schematic of scintillation cell arrangement. (b) Variations in temperature and corresponding scintillation signal changes in a solubility determination run with a solution containing 40 mg/mL of HbC and 1.9 M K2HPO4/K2HPO4 buffer at pH 7.37. Arrow on right ordinate, the temperature axis, indicates the temperature of equilibrium Teq between this solution and the crystals. (Inset) The expanded part of the signal trace illustrates the requirement for steady scintillation signal (i.e., crystals—solution equilibrium), before temperature is changed.

The whole assembly, together with a stirring motor, is housed in a light metal enclosure with a water-cooling loop to extract the heat from the Peltier device. The amplified/filtered output of the two photodiodes and the thermocouple are interfaced to a Macintosh computer via a MacADIOS II, which is

Nucleation and Crystal Growth of HbC

161

also connected to the thermoelectric temperature controller. A custom-made module employing the LabView programming environment changes the cell/ solution temperature according to the changes in the scintillation signal. In this method, the correspondence between temperature and equilibrium concentration is established by searching for the temperature at which a solution with a given protein concentration is in equilibrium with the crystalline material of interest. For this, a number of small crystallites are nucleated at a temperature that ensures sufficiently high supersaturation. The formation of these crystals decreases the solution concentration to roughly the equilibrium value at this temperature. Then the temperature is changed in steps in the direction of undersaturation, until all crystals have dissolved. In this way, the equilibrium is approached from the side of dissolution. This approach provides significant advantages over methods in which equilibrium is approached from the growth side (i.e., batch techniques) The advantages are twofold: (1) during dissolution, layers start retracting from the crystals’ edges, and thus no “dissolution layer source” is needed; (2) impurity pinning of steps (30,31) is believed to be less common during layer retraction. Kinetic hindrances, associated with growth layer generation and with impurity effects at low supersaturations, can lead to cessation of growth in supersaturated solutions, and thus bias equilibrium point determinations from the supersaturated side. An interesting finding, made on separate and numerous occasions using light and DIC video-enhanced microscopy and the scintillation technique, is that nucleation and crystallization do not occur in the presence of direct light. We had hypothesized that the CO was photolyzed, resulting in deoxyHb, which has a significantly longer nucleation lag time and rarely results in crystals (38). However, oxyHb, in which the ligand does not dissociate as a result of photolysis, exhibited the same phenomenon. This phenomenon of light inhibition of nucleation and crystal growth remains to be explained, and does not appear to be a simple artifact. After the solution is introduced in the light-scattering cell (see Fig. 3), the temperature is raised to a value in the range of 30–35°C. Tetragonal crystals, 10–50 µm on an edge, with the typical morphology illustrated by Fig. 1, were observed under the microscope with ×250 magnification within 1 to 2 h. The cell is then inserted in the scintillation apparatus described previously. These small crystals scatter light that is detected by the photodetector. Then the control program lowers the temperature to an operator-defined value, whereon the crystals begin to dissolve and the light scattered from them decreases. Representative temperature changes and the response of the photodiode signal are displayed in Fig. 3B. After 0.5–1 h, the signal reaches a steady value (see Fig. 3, inset), indicating that the remaining crystals are at equilibrium with the solution at this temperature. The temperature is further lowered in steps.

162

Vekilov et al.

Because of dissolution of all crystallites, the steady value of the signal returns to baseline. This indicates that all crystallites formed at the initially raised temperature have dissolved, and the solution concentration is identical to the initial one (verified by spectrophotometric determinations). The equilibrium temperature for the given protein concentration falls between the last two temperature steps. We approximate it by the lower temperature value (see arrow in Fig. 3B). To minimize the error introduced through this assumption, we optimized the magnitude of the final temperature step. These data were used to design a novel strategy for growing HbC crystals for X-ray structure studies by the temperature gradient technique (33). This method provides the following advantages for X-ray ready-crystals: First, crystallization occurs in the X-ray capillaries and crystal handling between growth and diffraction data collection is avoided. Second, only one or very few wellseparated crystals form, which eliminates interference of diffraction patterns of multiple intergrown crystals. Third, growth occurs in a controlled, steadytemperature environment, conducive to higher crystal perfection (33). In this method, a linear temperature gradient is set across an X-ray capillary. The high T value is chosen such that it provides for supersaturation at which crystal nucleation occurs, on average, over 2 to 4 d. For a solution containing 20 mg/mL of HbC, this temperature was found to be 24–25°C. The first HbC crystals grown by this method diffracted to 1.8 Å resolution. This is an improvement over previous studies, in which batch-grown HbC crystals diffracted to 2.1–2.0 Å. Efforts to further increase the diffraction resolution by optimizing the temperature conditions in the capillary are currently under way.

1.4. Atomic Force Microscopy Besides the clinical and biochemical significance of the in vitro studies of HbC crystallization, atomic force microscopy (AFM) is a suitable model for crystal growth of a multisubunit, allosteric protein. AFM of tetragonal HbC crystals, generated in 1.8 M phosphate, is employed to determine crystal growth mechanisms at the molecular level. To prepare samples for imaging, we placed droplets of a crystallizing solution of ~50 µL on 12-mm glass cover slips mounted on iron disks. To avoid evaporation, the droplets were covered with glass covers, hermetically sealed, and kept for a few hours in a controlledtemperature chamber at ~22°C. Typically, this leads to the formation of 3–20 crystals of sizes ranging from 20 to 200 µm firmly attached to the glass bottom. Droplets with three to five crystals were selected and magnetically mounted on the AFM scanner. The fluid AFM cell was filled with the crystallizing solution and imaging commenced. Temperature in the laboratory was stabilized to ~22°C. No additional temperature control of the solution in the AFM fluid cell was employed. Insertion

Nucleation and Crystal Growth of HbC

163

of a thermocouple in the crystallizing solution revealed that its temperature was higher than the room temperature by ~0.5–1.0°C. All images were collected in situ during growth of the crystals using the less intrusive tapping imaging mode (34–37). This allows visualization of adsorbed protein and impurity species; tip impact in the contact imaging mode often prevents such imaging. We used the standard SiN tips, and tapping drive frequency was adjusted in the range of 25–31 kHz to the resonance value for the used tip. Other scanning parameters were adjusted such that continuous imaging affected neither the surface structure nor the process dynamics. For verification, we varied the scan sizes and the time elapsed between image collections and saw that neither the spatial nor the temporal characteristics of the processes changed. For details and tests about the determination of the maximum resolution of 16 Å, and the calibration of the AFM imaging technique with other studied proteins, see ref. 38. The experiments with COHbC crystals in supersaturated solutions revealed that the thickness of a layer on a (101) face is 55 Å, and the periodicity along the c-axis is 195 Å, in agreement with the X-ray structure (32,39) Figure 4 provides examples of data attainable by an AFM investigation. If the width of the scanning area is 10 µm or wider (using the J scanner, it can be as wide as 140 µm), as in Fig. 4A, we can see that the crystals grow by a twostep mechanism: (1) a new layer is generated by a surface nucleation process; (2) these layers incorporate building blocks from the solution and spread to cover the whole facet. Note that the generation of a subsequent layer occurs while the underlying layer is still growing. This leads to many layers spreading and chasing one another on the crystal surface. Volmer (40) published this crystallization mechanism in the 1930’s, and it has been observed for numerous small-molecule, protein and virus crystals (38,41–44). Zooming in on the edge of the growing layer (Fig. 4B) we find that layer thickness equals ~55 Å, and this is the molecular dimension in the a (or b) crystallographic direction of the unit cells. Figure 4C–F reveals that as the molecules attach to the edge of the unfinished top crystal layer, this layer advances and the crystal grows. Furthermore, the edge of the unfinished layer in Fig. 4B is rough, and the characteristic length scale of the roughness equals one molecular dimension. This is only possible if molecules join the crystal one by one. We conclude that the building blocks of CO HbC crystals are a single protein molecule. Accordingly, in supersaturated solutions, the surface lattice parameter is ~200 Å and, within the resolution of our technique, equals the one expected from crystal structure.

1.5. Impurities and Crystal Growth The earlier notion that crystallization is a method for protein purification now requires revision as shown by crystal growth studies wherein crystal

164

Vekilov et al.

Fig. 4. AFM characterization of the mechanisms of incorporation of molecules into CO HbC crystals during growth. (a) Multiple layers spread along the surface, arrow points at a newly nucleated layer. (b) Molecular resolution imaging of the growth interface. The crystallographic unit cell is highlighted in a white box, and the roughness of the edge of the spreading layer is highlighted in white revealing the length scale of a single molecule.

imperfections arise from the incorporation of impurities. It was found that impurities affect growth and are often abundantly incorporated into the crystals. Species that have been considered impurities include microheterogeneous molecules of the crystallizing protein, aggregates and clusters of this protein, other protein molecules remaining in the solution after isolation and purification, ligands or cofactors that naturally bind to the protein, as well as small

Nucleation and Crystal Growth of HbC

165

molecules such as salts and buffers. For instance, the presence of other proteins and a covalently bound molecular dimer were shown to severely affect the shape of the interface, reduce the growth rate by factors of up to five, and completely inhibit growth at low supersaturations (45–49). More important, it was found that these same impurities are preferentially incorporated into the crystals and exhibit severely nonuniform distribution with a region of very high impurity incorporation in the central regions of a crystal (50,51). Therefore, crystallization should not be considered a method of protein purification without further verification. Our studies of the crystallization of Hb revealed that all of these mechanisms apply. The effects of HbS, HbA, and HbF, microheterogeneous molecules affecting the nucleation and crystal growth of HbC, are discussed under Subheading 1.1. and in Table 1. As far as small molecules are concerned, it was found that HbC nucleation is accelerated by intracellular components and analogs such as 2,3-diphosphoglycerate (DPG), inositol hexaphosphate, and Band 3 (23,52). This raises the question, Are erythrocyte components involved in the nucleation and crystallization of HbC in vivo? This in turn introduces the possibility of in vivo correlates to the present concepts of crystal growth and impurities. Video-enhanced DIC microscopy was used to observe the fine structure of the growing crystals in the solution. A detailed account of the findings of this investigation is provided in ref. 28. Within the context of a discussion of the effects that impurities may have on the growth and quality of the Hb crystals, it is important to note that these studies revealed well-pronounced striations parallel to the crystal faces (Fig. 5). Previous studies with other proteins have shown that these striations are owing to temperature variations during growth and are significantly fainter when the impurity content is lower (46,49,53). In analogy to these previous results, we suspect that the striation in also reflects enhanced impurity incorporation of the variation in growth rate caused by temperature instability. 2. Materials 2.1. Purification of Hemoglobins 1. Heparanized vacuum tubes for blood collection (Vacutainer tubes; BectonDickinson, Franklin Parks, NY). 2. Columns for liquid chromatography (Kontes, Vineland, NJ). 3. Preswollen and microgranular anion (DE-52, cat. no. 4057-200) and cation (CM-52, cat. no. 4037-200) exchangers (Whatman, Newton, MA). 4. Sephadex G-25 Fine gel filtration (cat. no. G25-80; Sigma, St. Louis, MO). 5. Protein concentrators (e.g., Centriprep 10, cat. no. 4304; Amicon, Beverly, MA]. 6. Isotonic saline (0.9%) prepared by dissolving 9 g of NaCl in distilled water and brought to 1 L.

166

Vekilov et al.

Fig. 5. Video-enhanced DIC imaging of formation of tetragonal crystals of oxyHbC. Magnification: ×4000; bar = 8.7 µm. (From ref. 28).

2.2. Batch Nucleation Studies 1. 2. 3. 3. 4. 5. 6.

2.1 M potassium phosphate buffer, pH 7.35. Glass test tubes (5 mL, 75 × 12 mm diameter) (cat. no. 55.476; Sarstadt). Pipettors with delivery capabilities of 1–10 µL and several hundred microliters. Incubator (30°C). Hematocytometer or other fixed grid for counting crystals in a reproducible fashion. Timer (minutes). Light microscope.

2.3. Video-Enhanced DIC Microscopy 1. Zeiss Axiovert 35 microscope: Similar to the video-enhanced DIC microscope described by Samuel et al. (54), it is equipped with a polarizer and analyzer, Wollaston prism, 546 interference filter, heat cut and reflecting filters, and a 100-W Hg light source.

Nucleation and Crystal Growth of HbC

167

2. Hamamatsu video camera, Hamamatsu C2400 analog enhancer, Hamamatsu Argus-10 Digital Image Processor, Mitsubishi Diamond Pro video recorder, and Sony monitor: The specimen image is projected onto the video camera and analog enhanced, then passed through the digital image processor and video recorded and displayed on the monitor. 3. Mitsubishi Video Copy Processor and VCR cassettes: A record is obtained with the processor and/or recorded over time onto the cassettes. 4. 35-mm Camera attachment, to photograph unenhanced video images. 5. Standard microscope slides and cover slips. 6. Nitrogen or helium for deoxygenation of Hbs. 7. Sodium dithionite (product 13551, cat. no. 7775-14-6; Sigma, St. Louis, MO) to reduce and deoxygenate the Hbs. 8. Glove box or glove bags, for anaerobic preparation of slides. 9. Cover glass sealant to maintain anaerobic conditions.

2.4. Determination of Solubility 1. Silica microcell (model 37G; Wilmad Glass). 2. Thermoelectric heat pump (Peltier cooler) MI1062T-03AC (12 Wmax heat load; Marlow, Dallas, TX). 3. Programmable controller SE 5010 (Marlow). 4. Type-T thermocouple (Omega). 5. Laser diode assembly VLM 25L, 5 mW, 670 nm (Applied Laser Systems). 6. Detector/amplifier photodiode UDT 455 (United Detector Technology). 7. Beam splitter (stock number A 32,600; Edmund Scientific). 8. Horseshoe magnet AM-300-RH (Active Magnetics). 9. Nickel wire 0.6 mm in diameter, 5–7 mm long. 10. Low-noise, low-drift/offset (0.6 mV/°C, 25 mV) precision op-amps OP27 (Precision Monolithics). 11. Stirring motor, 720 rpm BA P/N 4201-001 (Hurst). 12. Macintosh computer with a MacADIOS II (GW Instruments) interface board. 13. LabView programming environment (National Instruments).

2.5. Atomic Force Microscopy 1. Multimode Atomic Force Microscope Nanoscope IIIa (Veeco, Santa Barbara, CA). 2. J and E scanners (Veeco). 3. Cantilever holder for imaging in fluids (Veeco). 4. Disks (12 mm) cut of glass cover slips. 5. Iron disks (12 mm) (Veeco). 6. Epoxy resin. 7. Glass cup (5 mL). 8. Parafilm. 9. Syringe (1 mL), for solution loading into AFM fluid cell.

168

Vekilov et al.

2.6. Adhering Erythrocytes to AFM Disks 1. 2. 3. 4.

Trimethylchlorosiline (Sigma). Ruby Red Mica Sheets (Electron Microscopy Sciences, Washington, PA). Cover glass (Corning) or Microscope Cover Glass (Fisherbrand). Citrate-phosphate-dextrose (CPD): 87.7 mM citric acid, 15.2 mM sodium citrate (Na3C 6H5O 72H2O), 15.8 mM NaH2PO 4, 139 mM dextrose (anhydrous dextrose), pH ~5.63 (1.45 mL of CPD/10-mL whole-blood collection, which is usually equivalent to one vacutainer).

3. Methods 3.1. Purification of Hbs HbA, normal adult Hb, is obtained from volunteers. Naturally occurring Hb mutants are obtained from patients exhibiting specific hemoglobinopathies. Some mutations are spuriously detected in individuals who exhibit no pathophysiology for the amino acid substitution but may be under treatment for some other condition. HbC is obtained from CC or CA or SC individuals. Heterozygous RBCs ensure an excellent control since it allows for the handling and simultaneous separation and purification of these Hbs. In all these cases, written consent is required according to guidelines of the National Institutes of Health under the specific guidelines of the institution’s committee on clinical investigation. Whole blood must be assumed infectious and handled with proper precautions, such as wearing gloves, protective garments over clothing, and eye shields. Venous blood should be drawn into a vacuum tube containing an anticoagulant. For our studies, heparinized vacuum tubes are employed. On opening, hold the tube away from the body over a sink or basin. Open slowly and deliberately to avoid spurting of blood as air rushes in. Any spillage must be wiped off with a disinfectant such as Chlorox. The blood may be refrigerated overnight. A hemolysate should be prepared within 24 h from the time drawn. Preparation of a hemolysate is as follows: 1. Separate the red cells from the plasma by standing or spinning the whole blood at ~3000–5000 rpm in low-speed tabletop centrifuges. 2. Draw off and discard the plasma and buffy coat (a grayish thin layer containing leukocytes [<1% total blood volume] atop the RBCs). 3. Resuspend RBCs in isotonic saline (0.9%). 4. Wash the cells three times in this manner to ensure removal of the plasma. Recall that the plasma contains proteins and other compounds that bind specifically or nonspecifically to Hb with significant affinity—hence, the need for purified Hbs. 5. After washing the RBCs, lyse the cells by the addition of distilled water (1:1), shaken to break the membranes, then frozen in liquid nitrogen. Repeat the freezethaw step at least three times, each time transferring the supernatant to a clean

Nucleation and Crystal Growth of HbC

169

tube. Finally, spin the material at least three times at high speed (e.g., 10,000 rpm for at least 20 min) or until the membranous material is no longer visualized. This lysate is termed a hemolysate. Alternatively, membranes may be removed with the addition of distilled water:toluene (1:1 to the ratio of RBCs) to ensure lysis of the erythrocyte membrane. However, it is not advisable to use toluene for crystallization studies: it is known that toluene binds to Hb at α14 Trp, and toluene accelerates nucleation and crystallization. In fact, toluene is often added in very small quantities to the precipitants to induce nucleation and crystallization of hemoglobins or Hbs that do not readily crystallize. 6. Depending on the specific Hb mutation and subsequent charge modification, purify the Hb by liquid column chromatography using specific anion- and cationexchange resins, depending on its charge (see ref. 55). 7. Strip the purified hemoglobins of DPG and various ions by Sephadex G-25 gel filtration chromatography. We find that repeating this twice ensures the preparation of stripped Hbs, which may be assessed by oxygen equilibrium methods. Stripped Hb has a higher oxygen affinity. 8. Dialyze the Hb to the desired buffer and concentrate. Freeze drop-size pellets in liquid nitrogen and store under liquid nitrogen or temperatures not higher than ~–136°C.

3.2. Batch Nucleation and Crystallization Studies of HbC and Mixtures of HbC This method is a modification of a procedure introduced by Adachi and Asakura (56–58) to study polymerization and crystallization of HbS and HbC. Note that they employed Rayleigh light scatter to detect Hb aggregates, in contrast to our microscope and scintillation techniques. We clearly observe a lag phase for HbC crystallization in contrast to their reports of no observed lag phase for HbC crystallization. The difference in these reports may arise from their technique used to detect aggregation. It is important to use volumes that ensure a final experimental buffer concentration of at least 1.83 M potassium phosphate in order to observe crystals of HbC within 15–30 min. Therefore, the concentration of Hb must be high (~14 g%) in order that the volumes used to obtain a final concentration of 2 g% will be small. If the Hb availability is limited, then one can work with a total volume of 0.5 mL and up to 2 mL for batch crystallization in the indicated tubes. Surface area may be a likely contributing factor to the kinetics. CO hemoglobins (preferable in order to minimize oxidation) or oxyhemoglobins are mixed, and the tube is held between two fingers, gently tapped about 12 times, and placed in an incubator at 30°C. This is considered time zero. Note that there is no difference in nucleation and crystallization kinetics in the CO or oxy ligand states. At time zero and at 15-min intervals, the tube is removed, tapped as just described, and gently inverted to ensure mixing. The crystals are less dense than the solution. A 10-µL aliquot is removed and placed

170

Vekilov et al.

into a hematocytometer. The four grids of the outer corners are observed and crystals are counted. If a hematocytometer is not available, a slide with a fixed grid for counting may also serve the purpose. Since these are relative comparisons, it is important to use the same method of counting. Aliquots are usually observed for about 2 h. For solutions containing 100% HbC, red tetragonal crystals generally appear within 15–30 min under these conditions (Fig. 1). Mixtures of hemoglobins or HbC mixed with various factors may result in crystallization faster (within the dead time) or slower (hours or days). The log number of crystals is plotted against time (Fig. 2). Should it be desirable to collect the crystals, the batch solution is placed in a syringe with a Millipore filter adapter (micron size) to separate the crystals from the mother liquor. The mixture of crystals and mother liquor is pushed through the syringe. The filter containing the crystals is removed and carefully washed with the concentrated phosphate buffer to remove excess mother liquor. The crystals may then be dissolved in a few microliters of low concentration buffer (e.g., 0.05 M). The Hb solution is harvested and analyzed as necessary (e.g., electrophoresis, isoelectric focusing).

3.3. Growth of Single Crystals >2 mm via Batch Methods 1. Prepare 14 mg/mL of CO HbC in 1.8 M KH2PO4/K2HPO4, pH 7.35. A concentrated Hb solution is used at a volume which brings the 2.1 M potassium phosphate buffer concentration to 1.80–1.82 M final concentration. To measure the pH of a high concentration buffer (i.e., 1.8–2 M), it is recommended that one add 0.1 mL of the buffer to 2 mL distilled water. 2. Add 500-µL aliquots of the aforementioned mixture to a 7-mL test tubes. 3. Seal the test tubes tightly with parafilm. 4. Under a fume exhaust hood, gently add CO gas (~5 min) toward the surface of the aforementioned solutions without bubbling by placing a small hole in the parafilm with a pipet tip. Gentle rotation is allowed. A COHb solution appears cherry red, in contrast to the red oxyHb and purple deoxyHb solutions. 5. Cover the opening with parafilm and the test tubes with aluminum foil. Place the test tubes in a dark drawer at room temperature. 6. Add CO gas every 3 d by the method described in step 4 and re-cover with parafilm. Large crystals (~1 mm and greater) usually form within 10 d.

3.4. Growth of One to Three CO HbC Crystals Inside a Single X-Ray Diffraction Capillary (see Note 1) 1. Prepare 20 mg/mL of CO Hb C in 1.9 M KH2PO4/K2HPO4 in a 7-mL test tube. Excessive agitation may initiate nucleation of numerous small tetragonal crystals. If this occurs, placing the test tube mixture on ice or in a refrigerator for 10 min up to 4 h can dissolve the crystals. 2. Fill the X-ray diffraction capillary to its capacity. Handle these fragile capillaries with a gentle but firm touch.

Nucleation and Crystal Growth of HbC

171

3. Cap the ends of the capillary with melted wax. To achieve this, melt 2 to 3 g of wax in a beaker on a hot plate. Allow the wax by diffusion to enter the capillary, ~2–3 mm. As the wax hardens, add additional wax until a small bulblike structure appears to secure the ends. 4. Place the sealed capillary in the refrigerator for 1 h to dissolve small nuclei that may have formed during handling. 5. Place the capillary in a thermostated solution growth cell (33). The capillary is positioned such that one-quarter to one-third of the capillary sits over the growth sting or hot spot. 6. Program the localized heated growth sting at 24 to 25°C to provide a minimum supersaturated environment; meanwhile, set the thermostated jacket that surrounds the remaining capillary at temperatures <12°C, temperatures that will prevent nucleation. 7. Cover the entire apparatus with aluminum foil to prevent exposure to light. 8. One to three crystals should grow at the growth sting within 2 to 3 d. The crystallization solution will become clear after 1.5–3 wk, at which time the crystal (or crystals) will have reached maximum size.

3.5. Adhering CO HbC Crystals to AFM Scanning Disk 1. Glue an untreated cover glass to the steel AFM scanning disks. 2. Place the disks inside 24-well crystallizing plates. 3. Dispense 20-µL aliquots of 14 and 16 mg/mL of CO HbC in 1.6 M KH2PO4/ K2HPO4 on each disk. 4. Cover and seal the crystallizing plates with parafilm. Numerous small tetragonal crystals will grow overnight (e.g., 14 mg/mL produces few but larger crystals; 16 mg/mL produces smaller but more numerous crystals), some of which will be attached to the glass substrate whereas others will be floating in the crystallizing solution. 5. Remove floating crystals by ciphering off the crystallizing solution with a 22-gage 1-1/2 needle. Discard the crystallizing solution and the floating crystals. Gently add and remove 10 µL of 1.6 M KH2PO4/K2HPO4 repeatedly until the remaining floating crystals are removed. Crystals attached to the glass substrate are used for AFM studies. 6. To test that the CO HbC crystals are firmly attached, place 10 µL of 1.6 M KH2PO4/K2HPO4 over the crystals (as though a bubble of solution is formed over the crystal). Place this disk under a microscope for viewing. Using a 22-gage 1-1/2 needle, add 2 µL of 1.6 M KH2PO4/K2HPO4 while viewing the disk under the microscope. The examiner must determine whether the crystals actually move or remain attached. The disk will appear as though a wave of solution has run across the crystals. If the crystals do not move, they are assumed to be firmly attached to the glass substrate. 7. Place the disk on the scanner piezoelectric tube and xyz translator. For crystal growth studies, add 0.2 mg/mL of CO Hb in 1.6 M KH2PO4/K2HPO4 to the precipitant. This solution must be filtered because clusters of the Hb molecules

172

Vekilov et al. sometimes arise and impair the viewing field. (The clusters interfere with the laser light that diffracts off the “scanning” cantilever tip.) (see Note 2).

3.6. Adhering CC Erythrocytes to AFM Scanning Disk (see Note 3) A cover glass treated full strength with trimethylchlorosiline or mica glued to cover slips can be used. (Use tape to uncover a fresh layer of mica for each new experiment.) A droplet of RBCs treated with CPD is placed on the cover glass and incubated overnight. Depending on the requirements of the experiment, incubation can be done between 4 and 37°C. However, incubations >24 h at 37°C often leave hollow membrane seals and cremated cells. Incubations >24 h should be done at <30°C. The red cell droplet is rinsed with 0.9% saline or phosphate-buffered saline (PBS) and used “as is” for scanning (see Note 3). Intraerythrocytic crystals can be induced by incubating CC red cells with 3% saline (8,12). Higher yields of intraerythrocytic crystals are observed when the whole blood is not treated with CPD, but, rather, is incubated at 37°C for 4 h in 2% NaCl with 80 mM KCl, 20 mM HEPES, 1 mM MgCl2, and 60 mM NaCl. Using the methods mentioned above, these HbC crystal-containing cells can be attached to treated cover slips. These cells can then be viewed by an optical microscope, or by AFM.

3.7. Video-Enhanced DIC Microscopy to Observe In Vitro Crystal Growth 4 µL oxy, CO, or deoxy Hb (2 g% Hb, 1.8 M potassium phosphate buffer [e.g., 28]) are placed on a slide and observed over time (see Subheading 2.3. and refs. 28 and 54 for a description of the microscope configuration). The COHb does not oxidize readily to form met-Hb and, therefore, is advantageous in investigating R-state forms of Hb. For anerobic conditions, it is importatnt to deoxygenate the hemoglobin, check for 100% deoxygenation by absorption spectroscopy, and seal the slide well to prevent air from entering (this is discussed earlier). 4. Notes 1. CO HbC is prepared by gently adding CO gas to >140 mg/mL of concentrated HbC. Be sure to recalculate HbC concentration values; CO gas tends to further concentrate the mixture. For this reason, buffers are made by first bubbling CO gas into H2O and then by adding calculated portions of buffer/precipitants. Adding CO gas directly into the HbC solution results in incorrect concentration values for the solution. 2. The concentration of the solution must remain low to prevent nucleation so that the crystals do not interfere with the laser light.

Nucleation and Crystal Growth of HbC

173

3. Erythrocytes can be fixed by treating cover slips with full-strength poly-L-lysine. Droplets of RBCs are added and incubated. The RBCs are rinsed two to three times with PBS and fixed with 1% glutaraldehyde for imaging (2 to 3% glutaraldehyde has also been used).

Acknowledgments This work was supported in part by the American Heart Association, Heritage Affiliate Grant-in-Aid No. 0256390T; the National Institutes of Health R01 HL58038 and NHLBI 1F31 HL09564; and the Universities Space Research Association Research Contract 03537.000.013. We would like to acknowledge Dr. S.-T. Yau for his efforts with the atomic force microscope imaging. References 1. Haurowitz, F. (1938) Das Gleichgewicht zwischen Hamoglobin und Sauerstoff. Z. Physiol. Chem. 254, 266–274. 2. Perutz, M. F. and Mathews, F. S. (1966) An x-ray study of azide methaemoglobin. J. Mol. Biol. 21(1), 199–202. 3. Perutz, M. F. and Lehmann, H. (1968) Molecular pathology of human haemoglobin. Nature 219(157), 902–909. 4. Perutz, M. F. (1969) The Croonian Lecture, 1968. The haemoglobin molecule. Proc. R. Soc. Lond. B. Biol. Sci. 173(31), 113–140. 5. Nagel, R. L. (1991) The distinct pathobiology of SC disease: therapeutic tmplications, in Hematology/Oncology Clinics of North America, vol. (Nagel, R. L., ed.), W. B. Saunders, Philadelphia, pp. 433–451. 6. Nagel, R. L. and Steinberg, M. H. (2000) Hb SC and Hb C disease, in Disorders of Hemoglobin: Genetics, Pathophysiology, Clinical Management, (Steinberg, M. H., Forget, B. G., Higgs, D. R., Nagel, R. L., eds.), Cambridge University Press, MA. 7. Diggs, L. W. and Kraus, A. P. (1954) Intraerythrocytic crystals in a white patient with hemoglobin C in the absence of other types of hemoglobin. Blood 9, 1172–1184. 8. Kraus, A. P. and Diggs, L. W. (1956) In vitro crystallization of hemoglobin occurring in citrated blood from patients with hemoglobin. C. J. Lab. Clin. Med. 47, 700–705. 9. Charache, S., Conley, C. L., Waugh, D. F., Ugoretz, R. J., and Spurrell, J. R. (1967) Pathogenesis of hemolytic anemia in homozygous hemoglobin C disease. J. Clin. Invest. 46(11), 1795–811. 10. Hirsch, R. E., Raventos-Suarez, C., Olson, J. A., and Nagel, R. L. (1985) Ligand state of intraerythrocytic circulating HbC crystals in homozygote CC patients. Blood 66(4), 775–777. 11. Bunn, H. F., Noguchi, C. T., Hofrichter, J., Schechter, G. P., Schechter, A. N., and Eaton, W. A. (1982) Molecular and cellular pathogenesis of hemoglobin SC disease. Proc. Natl. Acad. Sci. USA 79(23), 7527–7531. 12. Lawrence, C., Fabry, M. E., and Nagel, R. L. (1991) The unique red cell heterogeneity of SC disease: crystal formation, dense reticulocytes, and unusual morphology. Blood 78(8), 2104–2112.

174

Vekilov et al.

13. Hirsch, R. E., Lin, M. J., and Nagel, R. L. (1988) The inhibition of hemoglobin C crystallization by hemoglobin F. J. Biol. Chem. 263(12), 5936–5939. 14. Adachi, K. and Asakura, T. (1981) Aggregation and crystallization of hemoglobins, A., S, and, C., Probable formation of different nuclei for gelation and crystallization. J. Biol. Chem. 256(4), 1824–1830. 15. Adachi, K., Segal, R., and Asakura, T. (1980) Nucleation controlled agregation of deoxyhemoglobin. S. J. Biol. Chem. 255, 7595–7603. 16. Kashchiev, D. (1969) Solution of the non-steady state problem in nucleation kinetics. Surf. Sci. 14, 209–220. 17. Kashchiev, D., (1995) Nucleation, in Science and Technology of Crystal Growth (van Eerden J. P. and Bruinsma, O. S. L., eds.), Kluwer Academic, Norwell, MA, pp. 53–56. 18. Mutaftschiev, B. (1993) Nucleation theory, in Handbook of Crystal Growth, vol. 1 (Hurle, D. T. J., ed.), Elsevier, Amsterdam, pp. 189–247. 19. Weber, P. (1991) Physical principles of protein crystallization, in Advances in Protein Chemistry, vol. 41 (Afinsen, C. B., Richards, F. M., Edsal, J. T., and Eisenberg, D. S., eds.), Academic, New York. 20. Weber, P. C. (1997) Overview of protein crystallization methods, in Methods in Enzymology, vol. 276. (Carter, C. W. Jr. and Sweet, R. M., eds.), Academic, New York, pp. 13–22. 21. Lin, M. J., Nagel, R. L., and Hirsch, R. E. (1989) The acceleration of hemoglobin C crystallization by hemoglobin S. Blood 74, 1823–1825. 22. Nagel, R. L., Lin, M. J., Witkowska, H. E., Fabry, M. E., Bestak, M., and Hirsch, R. E. (1993) Compound heterozygosity for hemoglobin C and Korle-Bu: moderate microcytic hemolytic anemia and acceleration of crystal formation [corrected] [published erratum appears in Blood 1994; 83(10):3105]. Blood 82(6), 1907–1912. 23. Hirsch, R. E., Witkowska, H. E., Shafer, F., Lin, M. J., Balazs, T. C., Bookchin, R. M., and Nagel, R. L. (1997) HbC compound heterozygotes [HbC/Hb Riyadh and HbC/Hb N-Baltimore] with opposing effects upon HbC crystallization. Br. J. Haematol. 97(2), 259–265. 24. Lawrence, C., Hirsch, R. E., Fataliev, N. A., Patel, S., Fabry, M. E., and Nagel, R. L. (1997) Molecular interactions between Hb alpha-G Philadelphia, HbC, and HbS: phenotypic implications for SC alpha-G Philadelphia disease. Blood 90(7), 2819–2825. 25. Hirsch, R. E., Shafer, F., Wajcman, H., Fataliev, N., and Nagel, R. L. (1999) HbJBaltimore Favors Crystallization of OxyHbC, but chain assembly abnormalities blunt the phenotype of the compound heterozygote. Blood 92(10)(Suppl. 1, Pt. 1), 11a. 26. Adachi, K., Lai, C. H., Konitizer, P., Donahee, M., and Campbell, A., S. S. (1994) Crystallization of recombinant hemoglobins with basic amino acid substitutions (Lys and Arg) at the 6 position. Blood 84, 1309–1313. 27. Gallagher, D. T., Eisenstein, E., Fisher, K. E., et al. (1998) Polymorphous crystallization and diffraction of threonine deaminase from Escherichia coli. Acta. Cryst. D. 54, 467–469. 28. Hirsch, R. E., Samuel, R. E., Fataliev, N. A., Pollack, M. J., Galkin, O., Vekilov, P. G., and Nagel, R. L. (2001) Differential pathways in oxy and deoxy HbC aggregation/crystallization [in process citation]. Proteins 42(1), 99–107.

Nucleation and Crystal Growth of HbC

175

29. Feeling-Taylor, A. R., Banish, R. M., Hirsch, R. E., and Vekilov, P. G. (1999) Miniaturized scinitillation technique for protein solubility determinations. Rev. Sci. Instr. 70(6), 2845–2849. 30. Cabrera, N. and Vermileya, D. A. (1958) The growth of crystals form solution, in Growth and Perfection of Crystals (Doremus, R. H., Roberts, B. W., and Turnbul, D., eds.), Wiley, New York. 31. Voronkov, V. V. and Rashkovich, L. N. (1994) Step kinetics in the presence of mobile adsorbed impurity. J. Crystal Growth 144, 107–115. 32. Dewan, J. C., Feeling-Taylor, A., Puius, Y. A., et al. (2002) Structure of mutant human carbonmonoxyhemoglobin C (βE6K) at 2.0 Å resolution. Acta. Crystallogr. D58, 2038–2042. 33. Rosenberger, F., Howard, S. B., Sowers, J. W., and Nyce, T. A. (1993) Temperature dependence of protein solubility—determination and application to crystallization in X-ray capillaries. J. Crystal Growth 129, 1–12. 34. Hansma, H. G. and Hoh, J. H. (1994) Biomolecular imaging with the atomic force microscope. Annu. Rev. Biophys. Biomol. Struct. 23, 115–139. 35. Hansma, P. K., Cleveland, J. P., Radmacher, M., et al. (1994) Tapping mode atomic force microscopy in liquids. Appl. Phys. Lett. 64(13), 1738–1740. 36. Möller, C., Allen, M., Elings, V., Engel, A., and Müller, D. J. (1999) Tappingmode atomic microscopy produces faithful high-resolution images of protein surfaces. Biophys. J. 77, 1150–1158. 37. Noy, A., Sanders, C. H., Vezenov, D. V., Wong, S. S., and Lieber, C. M. (1998) Chemically sensitive imaging in tapping mode by chemical force microscopy: relationship between phase lag and adhesion. Langmuir 14, 1508–1511. 38. Yau, S.-T., Petsev, D. N., Thomas, B. R., and Vekilov, P. G. (2000) Molecularlevel thermodynamic and kinetic parameters for the self-assembly of apoferritin molecules into crystals. J. Mol. Biol. 303(5), 667–678. 39. Fitzgerald, P. M. and Love, W. E. (1979) Structure of deoxy hemoglobin C (beta six Glu replaced by Lys) in two crystal forms. J. Mol. Biol. 132(4), 603–619. 40. Volmer, M. (1939) Kinetik der Phasenbildung. Steinkopff, Dresden. 41. Giesen, M., Schulze Icking-Konert, G., Stapel, D., and Ibach, H. (1996) Step fluctuations on Pt(111) surfaces. Surface Sci. 366, 229–238. 42. Malkin, A. J., Kuznetsov, Y. G., Land, T. A., DeYoreo, J. J., and McPherson, A. (1996) Mechanisms of growth of protein and virus crystals. Nat. Struct. Biol. 2, 956–959. 43. McPherson, A., Malkin, A. J., and Kuznetsov, Y. G. (2000) Atomic force microscopy in the study of macroimoleular crystal growth. Annu. Rev. Biomol. Struct. 20, 361–410. 44. Yip, C. M. and Ward, M. D. (1996) Atomic force microscopy of insulin single crystals: direct visulization of molecules and crystal growth. Biophysical J. 71, 1071–1078. 45. Thomas, B. R., Vekilov, P. G., and Rosenberger, F. (1998) Effects of microheterogeneity on hen egg white lysozyme crystallization. Acta Crystallogr. Sect. D 54, 226–236.

176

Vekilov et al.

46. Thomas, B. R., Vekilov, P. G., and Rosenberger, F. (1996) Heterogeneity determination and purification of commercial hen egg white lysozyme. Acta Crystallogr. Sect. D 52, 776–784. 47. Vekilov, P. G. (1993) Elementary processes of protein crystal growth, in Studies and Concepts in Crystal Growth (Komatsu, H., ed.), Pergamon, Oxford: pp. 25–49. 48. Vekilov, P. G., Monaco, L. A., and Rosenberger, F. (1995) Facet morphology response to non-uniformities in nutrient and impurity supply. I. Experiments and interpretation. J. Crystal Growth 156, 267–278. 49. Vekilov, P. G. and Rosenberger, F. (1996) Dependence of lysozyme growth kinetics on step sources and impurities. J. Crystal Growth 158, 540–551. 50. Stojanoff, V., Siddons, D. P., Monaco, L. A., Vekilov, P. G., and Rosenberger, F. (1997) X-ray topography of tetragonal lysozyme grown by the temperature controlled technique. Acta Crystallogr. Sect. D 53, 588–595. 51. Vekilov, P. G., Monaco, L. A., Thomas, B. R., Stojanoff, V., and Rosenberger, F. (1996) Repartitioning of NaCl and protein impurities in lysozyme crystallization. Acta Crystallogr. Sect. D 52, 785–798. 52. Hirsch, R. E., Rybicki, A. C., Fataliev, N. A., Lin, M. J., Friedman, J. M., and Nagel, R. L. (1997) A potential determinant of enhanced crystallization of Hbc: spectroscopic and functional evidence of an alteration in the central cavity of oxyHbC. Br. J. Haematol. 98(3), 583–588. 53. Monaco, L. A. and Rosenberger, F. (1993) Growth and etching kinetics of tetragonal lysozyme. J. Crystal Growth 129, 465–484. 54. Samuel, R. E., Salmon, E. D., and Briehl, R. W. (1990) Nucleation and growth of fibres and gel formation in sickle cell haemoglobin. Nature 345, 833–835. 55. Schroeder, W. A. and Huisman, T. H. J. (1980) The Chromatography of Hemoglobin, Marcel Dekker, Nw York. 56. Adachi, K. and Asakura, T. (1979) Gelation of deoxyhemoglobin A in concentrated phosphate buffer. Exhibition of delay time prior to aggregation and crystallization of deoxyhemoglobin, A., J. Biol. Chem. 254(24), 12,273–12,276. 57. Adachi, K. and Asakura, T. (1979) Nucleation-controlled aggregation of deoxyhemoglobin S: possible difference in the size of nuclei in different phosphate concentrations. J. Biol. Chem. 254(16), 7765–7771. 58. Adachi, K. and Asakura, T. (1979) The solubility of sickle and non-sickle hemoglobins in concentrated phosphate buffer. J. Biol. Chem. 254(10), 4079–4084.

Semisynthesis of Hb

177

11 Semisynthesis of Hemoglobin Seetharama A. Acharya and Sonati Srinivasulu 1. Introduction 1.1. Recombinant DNA Technology and Total Chemical Synthesis for Generation of Mutant Forms of Proteins Protein engineering, generation of mutant or modified forms of protein, has become the first step for studying the correlation of structure and function of proteins. Design and generation of novel protein molecules with tailor-made properties is the long-range goal of such studies. Such novel protein molecules are now designed and generated by recombinant DNA technology, as long as these protein molecules contain only the naturally occurring 20 amino acids residues (1,2). However, if unnatural amino acids needs to be introduced, cell-free protein expression system and special manipulation of the tRNA is necessary (3,4). Incorporation of the unnatural amino acid residues into a protein in a site-specific fashion could also be achieved through total chemical synthesis. Gutte and Merrifield (5) achieved the total chemical synthesis of RNase-A using solid-phase synthesis, starting from the carboxyl end of the molecule and building, one residue at a time, to its amino terminus. Hoffman et al. (6) introduced an alternate approach that involved the synthesis of a limited number of medium-size, protected segments of the molecule and condensing them to generate the full-length protein. This approach, generally referred to as a segment condensation approach (6), has gained significant attention in recent years. The molecular size of the proteins that biochemists desire to design and assemble are larger compared to RNase A. Accordingly, interest in developing newer and simpler approaches of segment condensation has increased.

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

177

178

Acharya and Srinivasulu

1.2. Semisynthesis of Proteins Semisynthesis of a protein can be considered as a specialized case of total chemical synthesis of protein by the segment condensation approach, where a significantly large portion of the protein is derived from the wild-type protein, and only a small segment of the protein is chemically synthesized with the desired changes in the covalent structure. This segment is then assembled with the complementary segment from the wild-type protein to generate a mutant form of the protein. The assembly may involve splicing of the two segments to establish chain contiguity. These semisynthetic reactions are referred to as covalent semisynthesis. However, establishing chain contiguity may not be necessary if the complementary segments exhibit strong noncovalent interaction among them that facilitates the assembly of the native folding of the protein (fragment-complementing systems). The latter represents the case of noncovalent semisynthesis.

1.2.1. Noncovalent Semisynthesis So far, in choosing the appropriate segment of the parent protein suited for chemical synthesis, information on the permissible discontinuity region of the protein has provided the road map (7). Many proteins have been converted into functionally active fragment-complementing systems either by limited proteolysis or site-specific chemical cleavage of the protein (8–10). In the modified protein, generally referred to as the fragment-complementing system, two (or more) segments of the protein are held together by strong noncovalent interactions, and accordingly, the modified protein conserves most of the conformational aspects of the parent protein. However, the interacting segments of the fragment-complementing system could be separated under denaturing conditions. In addition, a functional unit having the conformational and functional properties similar to that of the starting material could be generated by reassembling the complementary segments of the protein under the physiological conditions. In such a system, the amino acid sequence of segments of the protein is readily accessible for chemical manipulation through peptide synthesis. The modified peptide segment could be assembled with other segments of the system. The assembled product carrying a chemically synthesized segment is a semisynthetic protein and is analogous to the fragment-complementing system of the parent protein (11).

1.2.2. Covalent Semisynthesis Over the years, it has been established that it is now possible to induce the protease that generated the fragment-complementing system of a protein to religate the discontinuity in the polypeptide chain of the new semisynthetic

Semisynthesis of Hb

179

fragment-complementing system (12). The approach of splicing the discontinuity has been referred to as covalent semisynthesis. There are many examples of such covalent semisynthetic reactions in the literature to date (13–15). A recent modification of this approach introduced by Proudfoot et al. (16) is to have a very reactive group on the α-carboxyl group of the discontinuity site. In such a fragmentcomplementing system, the nucleophilic attack of the activated carboxyl group by the α-amino group of the discontinuity site established the contiguity to the polypeptide chain. Since the noncovalent interaction of the complementary segments facilitating the generation of native-like conformation in the fragmentcomplementing system is pivotal to the splicing reaction, this reaction is referred to as a conformationally assisted protein ligation reaction (17).

1.3. α-Globin Semisynthetic Reaction The V8 protease catalyzes the splicing of the complementary segments of α-globin (α1–30 and α31–141) (18), and has been used by our laboratory group for the preparation of many chimeric α-globins. A schematic representation of the generation of chimeric (semisynthetic) α-globin by exchanging one of the complementary segments of human α-globin with that of the animal α-globin is shown in Fig 1. Similarly, an exchange between the two animal α-globin chains could also be carried out to generate animal-animal chimeric α-globin chain. The V8 protease–catalyzed splicing reaction (semisynthetic reaction) is novel and very distinct from the previously described covalent protein semisynthetic reactions. This ligation reaction is facilitated by the conformation aspects of the product, rather than that of the “native-like” conformational aspects of reactants (i.e., the fragment complementing system). The continuity of the polypeptide chains established in the mixture of the complementary segments facilitates the induction of α-helical conformation into the contiguous chain in the presence of the organic cosolvent. This secondary structure of the ligated segment protects the Glu30 Arg31 peptide bond from proteolysis (19,20). There are several unique features of this α-globin semi synthetic reaction. First, there is the absence of noncovalent interactions between the two reacting peptides that establish a native-like structure in the fragment-complementing system. Second, the splicing reaction requires the presence of 30% propanol (or other α-helix-inducing organic solvents). Third, an extensive excision of the amino-terminal and the carboxyl-terminal region of α1–30 and α31–141 can also be made without influencing the equilibrium yields of the splicing reaction. Therefore, reaction has been exploited for the generation of semisynthetic hemoglobins (Hbs) (21–24).

1.4. Semisynthesis of Hbs Hb tetramer contains four polypeptide chains, two copies each of two subunits, the α- and β-chains. Each of these chains is composed of the protein part,

180

Acharya and Srinivasulu

Fig. 1. Schematic illustration of V8 protease–catalyzed slicing and splicing of α-globin to generate chimeric (semisynthetic) α-chains. The complementary segments (α1–30 and α31–141). of α-chains with single or multiple sequence differences in either the α1–30 or α31–141 segments or in both segments are spliced together to reduce or increase the number of sequence differences in the α-chain as compared with that of human α-chain.

globin, and the non-protein part, heme. Thus, the generation of tetrameric Hb from semisynthetic α-globin involves the assembly of the heme with semisynthetic α-globin, and the hybridization of the semisynthetic α-chain (hemebound polypeptide chain) with the appropriate β-chain to generate the α2β2 structure. Accordingly, the protocol for semisynthesis of Hb involves two major sections: (1) covalent semisythesis of desired variant or chimera of α-globin, and (2) assembly of semisynthetic α-globin with heme and β-chain to generate the tetramer through the alloplex intermediate pathway (25) (see Note 1). 2. Materials 2.1. Preparation of α and β-Globin Chains of HbA and/or Animal Hb 1. 2. 3. 4. 5.

Chromatographically purified HbA or HbS (refer to Chapter 3 for procedures). p-Hydroxymercuribenzoate (HMB) (see Note 2). 0.1 N NaOH (Fischer). 0.1 N Acetic acid (Fischer Scientific). Chromatographic columns (Rainin or Pharmacia), 1.5 × 30 cm for loads of 50–200 mg of protein, and 2.5 × 50 cm for loads of 1–2.5 g and higher. 6. Ion-exchange resin, CM-52 cellulose (Whatman). 7. Phosphate buffers: 10 mM phosphate buffer, pH 6.5, and 15 mM phosphate buffer, pH 8.3.

Semisynthesis of Hb

181

8. 0.5% Acid acetone (prepared by mixing 5 mL of concentrated HCl with 995 mL of acetone). 9. Corex centrifugation flasks (200–250 mL). 10. Sorval centrifuge RC-5C, LKB fraction collector, and a manual linear gradient marker when using a single-pump system or gravity for chromatography. Alternatively, a fast protein liquid chromatogrphy (FPLC) or an Acta system from Pharmacia can be used.

2.2. Preparation of α1–30 and α31–141 from Human and/or Animal α-Globins 1. 10 mM Ammonium acetate buffer, pH 4.0. 2. Lyophilized α-globin (about 100 mg). 3. Staphylococcus aureus V8 protease (Pierce, Rockford, IL). The V8 protease from Pierce comes as a lyophilized sample. Generally, a stock solution of the enzyme is made in water and stored at –80°C. One milligram of the enzyme is dissolved in water; the exact concentration of the enzyme is determined spectrophotometrically. The absorbance of a 1 mg/mL solution at 280 nm is 0.67. 4. Reverse-phase C4 column, analytical reverse-phase high-performance liquid chromatography (RP-HPLC) column (Vydac) for analytical runs, and semipreparative or preparative columns, for isolation of the complementary segments of α-globin using RP-HPLC. 5. Sephadex G-50, for purification of the complementary fragments of α-globin under denaturing conditions (0.1% trifluoroacetic acid [TFA] in water). 6. Urea (98+% purity) (Sigma, St. Louis, MO). 7. Buffer A: 5 mM phosphate buffer, pH 7.0, containing 8 M urea and 50 mM β-mercaptoethanol. 8. Buffer B: 25 mM phosphate buffer, pH 7.0, containing 8 M urea and 50 mM β-mercaptoethanol. 9. Buffer C: 50 mM phosphate buffer, pH 7.0, containing 8.0 M urea and 50 mM β-mercaptoethanol. 10. HPCL setup (we use a Shimadzu system), for the RP separation of the complementary fragments 11. LKB fraction collector, FPLC or Acta (both from Pharmacia).

2.3. V8 Protease–Catalyzed Ligation of α1–30 with α31–141 and Purification of Semisynthetic α-Globin 1. An α1–30 segment that is to be spliced, generated either by chemical synthesis incorporating the desired sequence differences (mutation) or by the V8 protease digestion of an animal α-globin carrying a number of sequence differences. 2. Desired α31–141 segment from the α-globin of either human or mammal. 3. 50 mM Ammonium acetate buffer, pH 6.0 containing 30% n-propanol. 4. V8 protease: a stock solution prepared as discussed in Subheading 2.2. is used. 5. C-4 RP-HPLC column (Vydac). 6. CM-52 cellulose (Whatman).

182

Acharya and Srinivasulu

7. Sephadex G-50. 8. Urea.

2.4. Assembly of Semisynthetic Hb and Its Purification 1. Mutant semisynthetic α-globin, or internally or externally deleted semisynthetic α-globin, or semisynthetic chimeric α-globin. 2. Chromatographically purified HMB βA- or βS-chains. 3. Catalase. 4. 100 mM EDTA in water (37.2 mg/mL of water). 5. 100 mM Dithiothreitol (DTT) (15.5 mg/mL) (Sigma). 6. Hemin (Sigma) solution: Dissolve 5 mg of hemin in 200 µL of 0.1 N NaOH. 7. Sodium cyanide: 1 mg in 100 µL of water. 8. Hemin dicyanide solution: 200 µL of hemin solution mixed with 60 µL of sodium cyanide solution and made up to 5 mL with double-distilled water (see Note 3). 9. Chromatographic columns: 1.5 × 50 cm and 0.9 × 30 cm (Rainin). 10. Sephadex G-25. 11. CM-52 cellulose (Whatman). 12. 100 mM Tris-HCl buffer, pH 7.4. 13. Globin-dissolving buffer: 50 mM Tris-HCl buffer, pH 7.4, containing 8 M urea, 1 mM EDTA, and 2 mM DTT.

2.5. Chemical and Functional Characterization of the Semisynthetic Hb 1. C4 column (Vydac), for RP-HPLC analysis of semisynthetic Hb. 2. Mass spectral analysis of semisynthetic α-globin of semisynthetic Hb isolated by RP-HPLC (API-III Triple-Quadrupole Mass Spectometer, Perkin-Elmer Sciex®). 3. Isoelectric focusing of semisynthetic Hb (using the instrument from Isolab). 4. Hem-O-Scan or Hem-Ox-Analyzer, for analysis of O2 affinity of semisynthetic Hb.

3. Methods 3.1. Preparation of α- and β-globin Chains from HbA and/or Animal Hb

3.1.1. α- and β-HMB Chains of HbA or HbS Reaction of HbA or HbS with HMB around pH 6.0 in the presence of 200 mM NaCl generates HMB α- and β-chains (26), which can be separated by CM-cellulose chromatography. Reaction of HbA or its mutant or chemically modified forms with HMB is carried out at a final concentration of 0.1 mM (6.5 mg/mL) in 20 mM phosphate buffer, pH 6.0, and containing 200 mM NaCl. A stock solution of Hb, generally at a concentration of 2.5 mM, is diluted with distilled water to a concentration of 0.2 mM (approx 12–15 mg/mL). When all the reagents are added,

Semisynthesis of Hb

183

and the pH is readjusted to 6.0, the volume of the reaction mixture is nearly doubled and protein concentration will be approx 6.5 mg/mL. A volume of 1.0 M phosphate buffer, pH 6.0, equivalent to one-fiftieth of the volume of the 0.2 mM Hb solution is added to this solution. Similarly, an amount of 1.0 M NaCl in 20 mM phosphate buffer, equivalent to one-fifth of the final volume (1:2.5 with respect to the 0.2 mM Hb solution) of the reaction mixture is added to this solution. Enough HMB, to make the final concentration of the reagent in the final reaction mixture 1.0 M, is dissolved in a minimum amount of 0.1 N NaOH. Once the reagent is completely dissolved, giving a clear solution, it is neutralized with 0.1 N acetic acid until a light whitish turbidity appears in the solution; it should not precipitate. This freshly prepared HMB solution is added completely to the Hb solution and mixed gently. The final volume of the reaction mixture is made up by adding water (the final concentration of Hb is 0.1 mM in tetramer). The pH of the reaction mixture is adjusted to 6.0 with dilute acetic acid. The reaction mixture is kept on an ice bath in a cold room overnight (generally 16–20 h). After the overnight HMB reaction with Hb, generally some precipitation is seen in the reaction mixture. Such a precipitate is removed by centrifuging of the reaction mixture (7000 rpm for 20 min in a Sorval centrifuge at 4°C), and the supernatant, which contains the HMB α-and β-chains, is dialyzed against three changes of 10 mM phosphate buffer, pH 6.0. The sample is then concentrated to approx 50 mg/mL. The concentrated sample is chromatographed on a CM-cellulose column (0.9 × 30 cm), equilibrated to pH 6.0. The chromatogram is developed with a linear gradient generated from an equal volume of 10 mM phosphate buffer, pH 6.0 (starting buffer), and 15 mM phosphate buffer, pH 8.5 (final buffer). A 0.9 × 30 cm column can easily purify the chains from a load of 200 mg of HMB-reacted Hb employing a gradient generated by 250 mL each of starting (10 mM phosphate buffer, pH 6.0) and final buffers (15 mM phosphate buffer, pH 8.5). Elution of the protein is monitored by following the absorption of the effluent at 540 nm. The HMB β-chain elutes first from the column, and the HMB α-chain elutes at the end of the gradient. The identity as well as the purity of the HMB chains as they elute from the CM-cellulose column are assessed by RP-HPLC of the samples (the HMB globin chains as well as the HMB-free globin chains elute at the same position in RP-HPLC). The chromatographically purified HMB α-and β-chains are pooled and concentrated and stored at –80°C until needed.

3.1.2. α-and β-Globin Chains of Hb The heme-free α-and β-globin chains can be prepared from the respective HMB α-and β-chains by acid acetone preparation (27). Alternately, the α- and β-globin chains can be prepared by CM-52 cellulose–urea column chromatography

184

Acharya and Srinivasulu

(28) of the total globin prepared by the acid acetone precipitation of the Hb chosen for the study (see Note 4). In our experience, both methods yield globin chains that work well in the semisynthetic reaction. Since the HMB reaction of animal Hbs has not been successful in generating the α-and β-chains, we generally follow CM-52 cellulose–urea column chromatography for preparation of the globin chains of animal Hbs. 3.1.2.1. ACID ACETONE PRECIPITATION OF TOTAL HB

Under strong acidic conditions, heme dissociates from the globin chain, and acetone precipitates the globin chains (27). Thus, the heme remains in solution, and globin chains precipitate. The protocol for the precipitation involves diluting of the Hb sample first with water to a concentration of about 10 mg/mL and keep it cold in an ice bath. The Hb solution is transferred to Corex centrifuging tubes or flasks. To this Hb solution 10 vol of acid acetone is slowly added. Until the addition of nearly 7 to 8 vol, the solution remains clear; afterward the addition of acid acetone results in precipitation of the protein as fluffy white material. The remaining volume of the acid acetone is added to the solution while the solution is mixed by using a glass rod. After the mixing of the Hb solution and acid acetone are completely mixed, the tube (or flask) is left in the ice bath for 30 min. All the protein precipitates as fluffy material. The sample is centrifuged to pellet down the precipitated globin. The supernatant, which is colored because it contains the heme, is decanted and discarded. The globin pellet is redissolved in 0.1 M acetic acid (in a volume equivalent to that of the original Hb solution, 10 mg/mL), and the acid acetone protocol is repeated two or three times until no more heme is extracted into the acid acetone phase. The final pellet of globin is dissolved in 0.1 M acetic acid (about 1 mg/mL) and lyophilized. Globin lyophilizes as white fluffy material. 3.1.2.2. CM-52-CELLULOSE-UREA CHROMATOGRAPHY OF ACID ACETONE PRECIPITATED GLOBIN

Thirty grams of CM-52 cellulose is first equilibrated with buffer A (50 mM phosphate buffer, pH 7.0, containing 8.0 M urea and 50 mM β-mercaptoethanol). The equilibrated resin is packed into a 2.5 × 15 cm column and further equilibrated by washing the column with 5 bed vol of buffer A. This column can resolve 150–200 mg of acid acetone–precipitated globin. The lyophilized sample of acid acetone–precipitated globin is dissolved in buffer A (10–15 mg/mL) and dialyzed against the same buffer (20 times over the volume of the protein solution) for 3 to 4 h. This dialyzed sample is loaded onto the CM-52 column prepared fresh, and the protein sample is eluted with a linear gradient of buffers B and C (250 mL each). The protein elution is moni-

Semisynthesis of Hb

185

tored by measuring the optical density of the fractions at 280 nm. The proteincontaining fractions are pooled and their identity is established by RP-HPLC. The pooled fractions containing α- and β-globins are dialyzed extensively to get rid of both the urea and the salts and then lyophilized. The lyophilized material is stored at –20°C until needed.

3.2. Preparation of α1–30 and α31–141 from Human and/or Animal α-Globin Chains The human α-globin is readily and quantitatively digested by V8 protease at pH 4.0 and 37°C to generate α1–30 and α31–141 (29). The animal α-globin is also digested in the same fashion (30,31).

3.2.1. V8 protease Digestion of Human or Animal α-Globin The α-globin sample is taken in 10 mM acetate buffer, pH 4.0, at a concentration of 0.5 mg/mL and placed in a 37°C water bath. When the α-globin solution is equilibrated to 37°C , protease digestion is initiated by adding the required amount of V8 protease. A substrate-to-enzyme ratio of 200:1 (w/w) is used for the digestion (2.5 µL of V8 protease solution [1 mg/mL]/1 mL of α-globin solution). Aliquots of the reaction mixture are analyzed by RP-HPLC to determine completion of the digestion (Fig. 2, inset A). When the digestion is complete, the ratio of the integrated area of α1–30 to α31–141 is nearly 1:3. Complete digestion of α-globin takes nearly 1–3 h. Once digestion is complete, the digest is lyophilized.

3.2.2. Purification of Complementary Fragments of α-Globin V8 protease digestion of 5–10 mg of α-globin has been routinely purified by our laboratory using RP-HPLC. A semipreparative Vydac C-4 column is used for such a purification. For a large-scale purification of the complementary segments of α-globin, size-exclusion chromatography (SEC) on Sephadex G-50 under denaturing conditions has been found very convenient. The lyophilized digest of α-globin is dissolved in 0.1% TFA in water (20–30 mg/mL). This is loaded onto a Sephadex G-50 column (1.5 × 70 cm) equilibrated with 0.1% TFA, and size-exclusion chromatographic separation of the fragments is achieved by eluting the column with 0.1% TFA at a flow rate of 0.5 mL/min. A load of 100 mg of V8 protease digest of α-globin is well resolved into complementary fragments in about 6 h. A typical purification profile of one such preparation is shown in Fig. 2. The identity and purity of the resolved complementary segments of α-globin can be established by RP-HPLC (Fig. 2, insets B and C). The fractions representing α1–30 and α31–141 are pooled and lyophilized. The purity of α31–141 is further established by tryptic digestion and RP-HPLC of the tryptic peptides. The absence of the tryptic peptides αT1, αT2, and αT3 in the tryptic digest establishes the purity of this fragment.

186

Acharya and Srinivasulu

Fig. 2. Large-scale purification of α1–30 and α31–141 from a V8 protease digest of human α-globin on a sephadex G-50 column (1.5 × 100 cm). The size-exclusion chromatography was developed using 0.1% TFA. Elution of the complementary segments from the column was followed by monitoring the absorption of the fractions at 280 nm. The high molecular weight fraction (fraction a) has been identified as α31–14 (inset B) and the lower molecular weight fraction as α1–30 (inset C). (Inset A) The RP-HPLC pattern of the total V8 protease digest is shown. Insets B and C are the RP-HPLC patterns of pooled fractions of a and b isolated from the Sephadex G-50 column. RP-HPLC analysis was carried out using a Vydac C4 column employing a linear gradient of 5–70% acetonitrile in 0.1% TFA in 130 min. The flow rate was 1 mL/min, and elution of the fragments was monitored at 210 nm.

3.2.3. Chemical Synthesis of Mutant α1–30 Segment The desired α1–30 segments with deletion/replacement/addition have also been generated by chemical synthesis. The fragments that are chemically synthesized include α1–30 (H2OQ), α1–30 (K16E), and α1–30des23–26. Prior to using these synthetic peptides, they were subjected to purification by RP-HPLC using a semipreparative column.

3.3. V8 Protease–Mediated Splicing of α1–30 Carrying Desired Structural Modification with α31–141 Splicing of the desired complementary segments is carried out in 50 mM ammonium acetate buffer (pH 6.0) containing 30% n-propanol. n-propanol

Semisynthesis of Hb

187

induces α-helical conformation into the contiguous segments generated through protease-mediated splicing (18). Other organic cosolvents, such as propanediol and butanediol, are also suitable for the splicing reaction.

3.3.1. Splicing of α1–30 with α31–141 Nearly a 1.2-fold molar excess of α1–30 or its mutant form over the desired α31–141 is mixed in 0.1% TFA and lyophilized to obtain a fluffy material. The lyophilized material is taken up in 50 mM ammonium acetate buffer (pH 6.0) containing 30% n-propanol, and a clear solution should be obtained. If a clear solution is not obtained with this level of organic solvent, the concentration of n-propanol can be increased to obtain a clear solution. If this also does not facilitate solubility of the mixture of the complementary segments, other organic solvents such as propanediol or butanediol can be tried. The final concentration of the complementary fragments is kept at about 10 mg/mL. The solution is kept at 4°C, and the semisynthetic reaction is initiated by adding the required amount of a stock solution of the V8 protease. An enzyme-to-substrate ratio of 1:200 is maintained in the reaction mixture. Progress of the ligation of the complementary segments is monitored by RP-HPLC analysis of the sample at different time intervals. The decrease in the amount of α1–30 in the reaction mixture as a function of time reflects the progress of the splicing reaction. Nearly 50% of the α1–30 is spliced with α31–141 within the initial 24 h of the reaction. Nonetheless, the incubation is generally continued up to 48 h. Once the reaction attains equilibrium, the reaction mixture is lyophilized and stored at –20°C until further processing.

3.3.2. Purification of Unreacted α1–30 from Ligated and Unligated α31–141 The unreacted α1–30 in the reaction is generally recovered in order to subject it to a second cycle of the splicing reaction. Recovery of the unreacted α1–30 is achieved by subjecting the lyophilized sample to an SEC on a Sephadex G-50 column as described in Subheading 3.2.2. The unreacted α1–30 and the mixture of the semisynthetic chimeric α-globin and α31–141 are pooled separately and isolated by lyophilization. The α1–30 is subjected to a second round of the splicing reaction using a new sample of α31–141 and processed similarly.

3.3.3. Purification of Semisynthetic α-Globin The high molecular weight fraction from the SEC of the spliced sample on Sephadex G-50 column is a mixture of α31–141 and the semisynthetic mutant or chimeric α1–141. The resolving power of size-exclusion chromatographic column is not sufficient to resolve the two, and accordingly, both elute together from the column. However, semisynthetic α-globin can be purified by CM-52 cellulose–urea column chromatography (28), as explained in Subheading

188

Acharya and Srinivasulu

3.1.2.2. The fractions from the chromatography containing the purified chimeric α-globin and α31–141 or mutant α1–141 are extensively dialyzed against 0.1% acetic acid and lyophilized. The lyophilized material is stored at –20C°.

3.4. Assembly of Semisynthetic Hb and its Purification To generate a functional tetrameric molecule, the semisynthetic (chimeric) α-globin has to be assembled with the complementary β-chain in the presence of heme (used as hemin dicyanide) (see Note 3). Depending on the objectives of the investigation, we have assembled the semisynthetic α-globin with either βA- or βS-chains. For these assembly reactions, the alloplex intermediate pathway, originally developed by Yip et al. (25) is used. This protocol takes advantage of the propensity of the β-chains of Hb (with free thiols) to interact noncovalently with heme-free α-globin to generate a tetramer, referred to as alloplex intermediate (only the β-chains of the tetramer containing heme). When hemin dicyanide is added to this, it occupies the heme-binding pocket of the molecule. The dithionite reduction of such a complex generates a functional tetramer.

3.4.1. Regeneration of the Sulfhydryl Groups of HMB β-Chain The HMB βA- or βS-chain, as established by the experimental protocol, is diluted into 50 mM Tris-HCl buffer, pH 7.4, containing 1 µg/mL of catalase, 2 mM DTT, and 1 mM EDTA to get a solution of 5 mg/mL. This solution kept at 4°C for about 45 min. The DTT removes the HMB groups from the thiol group of Cys-residue, regenerating the free sulfhydryl group.

3.4.2. Regeneration of Sulfhydryl Groups of the Semisynthetic α-Globin The semisynthetic α-globin is dissolved in 50 mM Tris-HCl buffer (pH 7.4) containing 8.0 M urea, 2 mM DTT, and 1 mM EDTA to get a solution of 5 mg/mL. This solution is incubated for 30–45 min at room temperature.

3.4.3. Preparation of Half-Filled Molecules (Alloplex Intermediate) Half-filled molecules are prepared by slowly adding the α-globin and β-chain solutions to the dilution buffer (50 mM Tris-HCl buffer, pH 7.4, containing 1 mM DTT, 1 mM EDTA, and 1 µg/mL of catalase). The two solutions are added to the dilution buffer (nearly 10 times the volume of the solution of semisynthetic α-globin) simultaneously with constant stirring at 4°C so that the urea concentration is lowered to 0.8 M. After the dilution, the final concentration of the protein is about 1 mg/mL (0.5 mg/mL of each chain). This mixture is incubated at 4°C for 30 min to facilitate the formation of half-filled molecules.

Semisynthesis of Hb

189

3.4.4. Generation of Cyanomet Semisynthetic Hb To the solution of half-filled molecules generated in Subheading 3.4.3., an equivalent amount of freshly prepared hemin dicyanide is added dropwise, to generate a 1.1-fold molar excess of hemin dicyanide in solution over that of globin. This addition is carried out with mild shaking of the sample. The reaction between hemin dicyanide and half-filled molecules should then be allowed to proceed for about 60 min at 4°C. This will generate the semisynthetic molecule in the cyanomet form. This solution containing the cyanomet form of semisynthetic Hb is subjected to an extensive dialysis against 50 mM Tris-Hcl buffer, pH 7.4. During dialysis, some amount of precipitation occurs. This is removed by centrifuging the sample at 4°C for 20 min at 7000 rpm. The clarified dialyzed sample is concentrated by ultrafiltration to a concentration of 20 mg/mL.

3.4.5. Dithionite Reduction of Cyanomet Form of Semisynthetic Hb The semisynthetic Hb generated in Subheading 3.4.4. is in the ferric state. To convert it to the ferrous state, the sample has to be reduced by either an enzymatic procedure (32) or a chemical method (33). In our laboratory, we have routinely used the sodium dithionite to reduce the methemoglobin sample (MetHb). This is done using a Sephadex G-25 column under anaerobic conditions. Preswollen Sephadex G-25 column equilibrated with 10 mM phosphate buffer, pH 7.0 is packed into a column (1.5 × 50 cm) and the column is equilibrated with 10 mM phosphate buffer (pH 7.0) that is constantly purged with N2 gas. Once the column is washed with 2 to 3 bed vol of the buffer, the dithionite solution is freshly prepared. Sodium dithionite, approx 1.1 equivalents over the semisynthetic α-globin used in the reconstitution, is dissolved in 10 mM phosphate buffer, pH 7.0, degassed with N2 gas. The volume used to dissolve the sodium dithionite is equal to that of the semisynthetic cyanoMetHb solution that needed to be reduced. This solution is placed on the top of the Sephadex G-25 column very gently without disturbing the gel and allowed to enter the bed completely. Once the dithionite solution enters the column, the column is washed with a volume of buffer (degassed with nitrogen) equivalent to that of the dithionite solution. The column is then loaded with semisynthetic cyanoMetHb. About 2 to 3 mL of 10 mM phosphate buffer, pH 7.0, previously degassed with nitrogen gas is used to rinse the top of the column, and the chromatogram is developed with 10 mM phosphate buffer. During the elution, the buffer tank is continuously purged with N2 gas. The effluent is collected as a 1-mL fraction, and elution of Hb is monitored by measuring the absorption at 540 nm. The fractions containing Hb are analyzed for the concentration of MetHb. Samples of semisynthetic Hb containing <2 to 3% of the MetHb are pooled and concentrated by an ultrafiltration unit.

190

Acharya and Srinivasulu

3.4.6. Chromatographic Purification of Semisynthetic Hb The semisynthetic Hb is purified by CM-52 cellulose column chromatography. The chromatographic conditions used for the separation of HMB α- and β-chains are also appropriate for purification of semisynthetic Hb. Chromatography is carried out on a column (0.9 cm × 30 cm) generally equilibrated with 10 mM phosphate buffer, pH 6.5. The dithionide-reduced semisynthetic Hb is dialyzed against 10 mM phosphate buffer, pH 6.5, for just by couple of hours in the cold, before loading onto the column. The column is developed using a linear gradient generated from 250 mL each of 10 mM phosphate buffer, pH 6.5, and 15 mM phosphate buffer, pH 8.5. Approximately 3.4-mL fractions are collected, and elution of the protein is monitored by measuring the absorption at 540 nm. The fractions containing protein are analyzed by RP-HPLC to establishing the identity of assembled semisynthetic Hb. The fractions containing semisynthetic Hb are pooled and concentrated by ultrafiltration.

3.5. Chemical and Functional Characterization of Semisynthetic Hb The semisynthetic Hb thus generated has to be subjected to (1) RP-HPLC analysis of the globin chains, (2) mass spectral and/or tryptic peptide mapping analysis of the semisynthetic α-globin chain isolated from the RP-HPLC runs of the semisynthetic Hb, (3) isoelectric focusing analysis of the tetramer, and (4) analysis of the functional properties of the semisynthetic Hb. The methods for each of these are not discussed here. The reader should refer to Chapters 2, 3, and 6 for further details. 4. Notes 1. Site-directed mutagenesis is a powerful tool for protein design and engineering. However, the protein chemists’ desire is to expand the range of the modification of the covalent structure of proteins by incorporating unnatural amino acids into the protein or by introducing unique rigid elements of three-dimensional structure into the protein. Since only limited variations in the amino acid side chains are possible through genetic approaches, the search for innovative approaches for construction of proteins has continued. Protease-mediated semisynthesis of proteins is one such approach. Such approaches, however, are generally very specific to a given protein system. The α-globin semisynthetic approach discussed here has been very useful system in the study of structure-function relationships in Hb, particularly in exposing the synergy of the intermolecular contacts of deoxyHb polymer (24). 2. HMB is a light sensitive material. Accordingly, while preparing the HMB solution, it is necessary to avoid exposing the reagent to direct light. After adding the HMB solution to the Hb solution and adjusting of the pH to 6.0 (see Subheading 3.1.1.), the flask containing the reaction mixture is wrapped with aluminum foil, to avoid any further decomposition of the reagent during the overnight reaction.

Semisynthesis of Hb

191

3. Extreme care should be exercised during preparation of hemin dicyanide (see Subheading 2.4.) and handling of the hemin dicyanide solution (see Subheading 3.4.). Hand gloves should be used during these procedures, to avoid any contact of cyanide with the body. All the solutions are prepared in a fume hood. 4. Urea solutions for the CM-52 urea chromatography must be prepared freshly with good-quality urea. The ultrapure molecular biology–grade material from Sigma is used in all our studies. The pH of all the urea-containing solutions must be adjusted with dilute phosphoric acid after complete dissolution of the urea and before adding β-mercaptoetanol to the solution. Use fume hoods when working with β-mercaptoethanol. The addition of mercaptoethanol will not affect the pH of the buffer. The slurry of the CM-52 cellulose must be prepared with ureacontaining buffer a couple of hours before use, and this preparation cannot be stored for further use. We generally run the CM-52 urea chromatography inside a fume hood.

References 1. Mathews, B. W. (1993) Structural and genetic analysis of protein stability. Ann. Rev. Biochem. 62, 139–161. 2. Smith, M. (1994) Synthetic DNA and biology. Angew. Chem. Int. Ed. Eng. 33, 1214–1221. 3. Hecht, S. M. (1992) Probing the synthetic capabilities of a center of biochemical catalysis. Acc. Chem. Res. 25, 545–555. 4. Mendel, D., Cornigh, V. W., and Schultz, P. G. (1995) Site directed mutagenesis with an expanded genetic code. Annu. Rev. Biophys. Biomol. Struct. 24, 435–462. 5. Gutte, B. and Merrifield, R. B. (1969). The total synthesis of an enzyme with ribonuclease A activity. J. Am. Chem. Soc. 91, 501, 502. 6. Hoffman, K., Kisser, J. P., and Finn, F. M. (1969) Studies of Polypeptides: XLII Synthesis of S-peptide1–20 by two routes. J. Am. Chem. Soc. 91, 4883–4887. 7. Anfinsen, C. B. and Scheraga, H. A. (1975) Experimental and theoretical aspects of protein folding. Adv. Protein Chem. 29, 205–294. 8. Richards, F. M. and Vithayathil, P. J. (1959) The preparation of subtilisin modified ribonuclease A and the separation of peptide and protein component. J. Biol. Chem. 234, 1459–1465. 9. Tanuichi, H., Parr, G. R., and Juillerat, M. A. (1986) Complementation in folding and fragment exchange. Methods Enzymol. 131, 185–217. 10. Vita, C., Dalzoppa, D., and Fontana, A. (1985) Limited proteolysis of thermolysine by subtilisin: Isolation and characterization of partially active enzyme derivative. Biochemistry 24, 1798–1806. 11. Wallace, C. J. (1995) Pepide ligation and semisynthesis. Curr. Opin. Biotechnol. 6, 403–410. 12. Homandberg, G. A. and Laskowski, M. (1979) Enzymatic synthesis of the hydrolyzed peptide bond(s) of ribonuclease S. Biochemistry 18, 586–592. 13. Wallace, C. J. (1993) Understanding cytochrome C function: engineering protein structure by semisynthesis. FASEB J. 7, 505–515.

192

Acharya and Srinivasulu

14. Wallace, C. J. (1993) The curious case of protein splicing: mechanistic insights suggested by protein semisynthesis. Protein Sci. 2, 697–705. 15. Ni, X. and Schachman, H. K. (2001) In vivo assembly of aspartate transcarbamylase from fragmented and circularly permutated catalytic polypeptide chains. Protein. Sci. 10, 519–527. 16. Proudfoot, A. E., Rose, K., and Wallace, C. J. (1989). Conformation directed recombination of enzyme activated peptide fragments: a simple and efficient means to protein engineering: its use in the creation of cytochrome C analogs for structure function studies. J. Biol. Chem. 264, 8764–8770. 17. Kullman, W. (1987) Enzymatic Peptide Synthesis, CRC, Boca Raton, FL. 18. Sahni, G., Cho, Y., Iyer, K. S., Khan, S., Seetharam, R., and Acharya, A. S. (1989) Synthetic hemoglobin A: reconstitution of functional tetramers from semisynthetic α-globin. Biochemistry 28, 5456–5461. 19. Roy, R. P., Khandke, K. M., Manjula, B. N., and Acharya, A. S. (1992) Helix formation in the enzymatically ligated peptides as a driving force for the synthetic reaction. Biochemistry 31, 7249–7255. 20. Sahni, G., Khan, S., and Acharya, A. S. (1998). Chemistry of the ‘Molecular Trap’ of protease catalyzed splicing reaction of the complementary segments of α-subunit of hemoglobin A. J. Protein Chem. 17, 669–678. 21. Roy, R. P., Nagel, R. L., and Acharya, A. S. (1993). Molecular basis of the inhibition of βS chain polymerization reaction by mouse α-chain: semisynthesis of chimeras of human and mouse α-chains. J. Biol. Chem. 268, 16,406–16,412. 22. Nacharaju, P., Roy, R. P., White, S. P., and Acharya, A. S. (1997). Inhibition of sickle β-chain (βS) chain dependent polymerization by non-human α-chain: a super inhibitory mouse-horse chimeric α-chain. J. Biol. Chem. 272, 27,869–27,876. 23. Rao, M. J., Malavalli, A., Manjula, B. N., Kumar, R. Prabhakaran, R., Sun, D. P., Ho, N. T., Ho, C., Nagel, R. L., and Acharya, A. S. (2000) Interspecies hybrid HbS: complete neutralization of Val-6(β) dependent polymerization of human β-chain by pig α-chain. J. Mol. Biol. 300, 1389–1406. 24. Srinivasuslu, S., Malavalli, A., Prabhakaran, M., Nagel, R. L., and Acharya, A. S. (1999) Inhibition of βS chain dependent polymerization by synergistic complementation of contact site perturbations of α-chain: application of chimeric α-chain. Protein Eng. 12, 1105–1111. 25. Yip, K. Waks, M., and Beychock, S. (1977). Reconstitution of native human hemoglobin from separated globin chains and alloplex intermediates. Proc. Natl. Acad. Sci. USA 74, 64–69. 26. Bucci, E. (1981) Preparation of isolated chains of human hemoglobin. Methods Enzymol. 76, 97–106. 27. Rose-Fanelli, A., Antonini, E., and Caputo, A. (1958) Studies on the structure of hemoglobin: 1. Physicochemical properties of hemoglobin. Biochem. Biophys. Acta. 30, 608–615. 28. Clegg, J. B., Naughton, M. A., and Weatherall, D. J. (1966). Abnormal hemoglobins: separation and characterization of the α and β chains by chromatography and determination of two variants, Hb-Chesapeak and HbJ (Bangkok). J. Mol. Biol. 19, 91–108.

Semisynthesis of Hb

193

29. Iyer, K. S. and Acharya, A. S. (1987) Conformational studies of α-globin in 1-propanol: propensity of the alochol to limit the sites of proteolytic cleavage. Proc. Natl. Acad. Sci. USA 84, 7014–7018. 30. Roy, R. P., and Acharya, A. S. (1994). Semisynthesis of hemoglobin. Methods Enzymol. 231, 194–215. 31. Nacharaju, P. and Acharya, A. S. (000) Hemoglobin semisynthesis, in Semisynthesis of Proteins, (Wallace, C. J., ed.), CRC, Boca Raton, FL, pp. 151–198. 32. Hayashi, A., Suzuki, T., and Shin, M. (1973). An enzymatic reduction system for met myoglobin and met hemoglobin and its application to functional studies of oxygen carriers. Biochim. Biophys. Acta. 310, 309–316. 33. Dixon, H. B. F. and McIntosh, R. (1967) Reduction of met hemoglobin hemoglobin samples using gel filtration for continuous removal of reaction products. Nature 213, 399,400.

194

Acharya and Srinivasulu

β-Globin-like Gene Cluster Haplotypes

195

12 β-Globin-like Gene Cluster Haplotypes in Hemoglobinopathies Shanmugakonar Muralitharan, Rajagopal Krishnamoorthy, and Ronald L. Nagel 1. Introduction The pioneering work of Kan and Dozy (1) revealed by restriction endonuclease mapping a genetic variation in an HpaI recognition site about 5000 nucleotides from the 3' end of the β-globin gene. Instead of a normal 7.6-kb fragment containing the β-globin gene, 7.0- and 13.0-kb variants were detected and were found in African Americans, Asians, and Caucasians. The 13.0-kb variant (HpaI+) was frequently associated with the sickle hemoglobin (Hb) mutation. Kan and Dozy (1) predicted that polymorphisms in a restriction enzyme site could be considered a new class of genetic marker and may offer a new approach to linkage analysis and anthropological studies. Based on limited data (2), they reported that linkage to the wild-type HpaI-positive site was characteristic of West Africans while an HpaI-negative site typified East Africans. This finding did not show definitively that the mutation had occurred in two different chromosomal backgrounds, because a secondary mutation at the HpaI site could have postdated the sickle mutation. Orkin (3) demonstrated in samples from three regions of Africa that the distribution of the HpaI linkage disequilibrium was more complex. An HpaI polymorphism was territorially segregated in Africa in three geographical locations (4). Atlantic West Africa and Bantu-speaking Central Africa had HpaInegative -linked βS genes, while Central West Africa had HpaI-positive-linked βS genes. This observation increased the possibility that the βS mutation could have been multicentric in origin. Nevertheless, it would still be possible, albeit with diminishing probability, that these mutations occurred after the sickle From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

195

196

Muralitharan et al.

mutation. The next stage of this development came when Mears et al. (5) developed the concept of globin haplotype based on a series of restriction endonuclease–defined polymorphisms in and around the β-globin-like gene cluster. In these studies, DNA sequence differences in the β-globin gene cluster among different individuals were detectable by restriction enzymes that fail to cut DNA if the enzyme target sequence is not exactly correct. Pagnier et al. (6) used haplotypes to link different thalassemia mutations to particular haplotypes with the hope that haplotypes could serve as diagnostic tools. This approach was not totally successful since it rapidly became evident that more than one thalassemic mutation was linked to the same haplotype and that more than one mutation was linked to the same mutation. Nevertheless, the effort to understand the relationship between thalassemia and β-globin-like gene cluster haplotypes stimulated others to apply this technique to detect the origins of the sickle mutation. A breakthrough in this quest came when 11 restriction sites were examined (7) to define haplotypes in sickle cell (SS) anemia patients from three regions in Africa previously studied by Mears et al. (5,6). This work established that the βS-globin gene was present on three distinctly different chromosomes. Each was identifiable by its specific array of DNA polymorphism haplotypes, and each was localized exclusively to one of the three separate geographical areas: Atlantic West Africa, Central West Africa, and Bantu-speaking Central and Southern Africa. Hence, the βS gene in Africa is principally distributed around these three main geographical locations, each exhibiting a center of very high frequency surrounded by regions of declining frequencies. The conclusion from these findings was that the βS gene originated independently at least three times with subsequent expansion of the frequency of the abnormal gene in each of these geographical areas (8). The strongest argument for this interpretation is the geographical segregation of these distinct haplotypes with only haplotype associated with the βS gene in each of these locales. Typical haplotypes, such as low-frequency haplotypes different from the major haplotypes associated with the βS gene, are all explainable by crossing-over events around a “hot spot” of recombination 5' to the β-globin gene (9). Differences between atypical and typical haplotypes of each geographical area are generally found in the region 5' to this “hot spot” (10). This picture has gotten more complex with more recent findings, as discussed later. In addition to the three major haplotypes linked to the βS gene, a fourth minor African haplotype, the Cameroon (described later) has been found to be a “private” haplotype restricted to the Eton ethnic group, indicating a fourth independent origin of the HbS gene in Africa (11). In this group, the Cameroon haplotype reached polymorphic frequencies, but expansion beyond

β-Globin-like Gene Cluster Haplotypes

197

the original ethnic group did not occur. Further evidence regarding the separate origin of the three major haplotypes linked to the βS-globin gene came from the studies of Chebloune et al. (12) on variable repeats of the ATTTT motifs found about 1.5 Kb 5' to the β globin gene and the AT repeats (followed by T runs of different size) found about 0.5 kb 5' of the β-globin gene. Early DNA sequencing in this region suggested that these repeats were polymorphic. Chebloune et al. (12) found that ATTTT repeated either four or five times with (AT)X TY probes having X = 7 or 11 and Y = 7 or 3. Their results were complemented by sequence data from –1080 bp 5' to the cap site of the β-globin gene. It is apparent that the combination of these polymorphic areas is unique for each haplotype, supporting the independent origin of the Arab-Indian haplotype found today among the tribals of India and among Arabs living in the eastern oasis of Saudi Arabia and in Oman.

1.1. β-Globin-like Gene Cluster Haplotypes in Biology, Medicine, and Anthropology Haplotypes of the β-globin-like clusters have been used for the following purposes: 1. To provide anthropological correlations: They have been used to give evidence of and/or define the common origin and the likelihood of an ancestral home for the tribals of India (13) and their potential origin in the Harappa culture, in the margins of the Indus River; to give a biological basis to the linguistical basis of the Bantu expansion hypothesis in Africa (8); and determine the Indian tribal origin and east African origin of the sickle gene, respectively, in Indian and African inhabitants of Mauritius Island (14). 2. To provide a source of clinical diversity among SC patients: Evidence exists that the linkage of the βS gene to the Senegal and Arab Indian India haplotypes is associated with higher expression of HbS in SS and more benign hematological profile (15–17). Conversely, the Bantu haplotype has the most severe course (18).

1.2. Study of Gene Flow 1.2.1. Slave Trade–Based Gene Flow of Sickle Gene to America Strikingly, of more than 20 haplotypes associated with the βS-globin gene in Jamaica whose population was generated by forced African migration from all over Africa to the Americas, the three haplotypes described in Africa (7) comprised more than 95% of the cases (19,20). This demonstrated that the three geographically segregated haplotypes were the major βS-linked haplotypes in Africa and suggested that the rest might represent “private” haplotype linkages represented by fresh mutations, gene conversion, or more classic crossing-over events around the putative hot spot 5' to the β-globin gene. Nevertheless, since

198

Muralitharan et al.

Jamaica was populated from Africa, these three major haplotypes and other atypical haplotypes have coexisted only since the eighteenth century limiting the potential for the suggested crossing-over events.

1.2.2. Determination of the African Origin of Sickle Mutation in Sicily (21) The question had been raised whether the sickle mutation was autochthonous to Sicily or was imported by gene flow from Africa. The haplotype analysis demonstrated that the sickle mutation was linked to the Benin haplotype which is characteristic of the African population of central west Africa. Nevertheless, using an “African marker” located between the two gamma genes, it was possible to ascertain that the incidence of this polymorphism was less than 1% of the Sicilian population, hence the sickle gene was imported in small numbers to Sicily from Africa and latter expanded by the selective pressure of p. falciparum malaria.

1.2.3. Expansion of the Sickle Gene in Middle East During the Sassanian Empire (22) The presence of the Arab-India haplotype in the populations of the Arabian peninsula reinforces the hypothesis that this particular mutation originated in the Harappa culture or in a nearby population and, in addition, reveals that the Sassanian Empire might have been the vehicle by which this Indo-European sickle mutation migrated (gene flow) to the present-day Arabian peninsula (22). More recently, it has been established that the sickle gene is linked to the ArabIndia haplotype in Central Iran, giving further support to this hypothesis (23). Recent sickle gene flow studies have included Venezuela, Southern Tunisia, the Medenka population in Senegal, Colombia, United Arab Emirates, Guadaloupe (FWI), the Afro-Brazilian population from the Amazon region, Madagascar, Cuba and Guinea (Conackry) (23,24).

1.3. Biochemical Inquiries of Mechanisms Involved in Diversity of Haplotype/Phenotype Associations Gilman and Huisman (35) proposed that the -158 site 5' to the G γ gene (detected by XmnI + site) determines the G γ expression after the first 4 mo of life. DNA from SS patients from Africa and β-thalassemic homozygotes from Algeria demonstrated that the XmnI site is strongly linked to the Senegal haplotype among SS patients, to haplotype IX (most probably identical to the Sengal haplotype), and to haplotype III among the Algerian thalassemics (36,37). It was concluded that, although highly correlated, the –158 C→T substitution does not perfectly predict the presence of high G γ expression. Further study of the relationship between these markers and Fetal hemoglobin (HbF) expression judged by F-cell levels in unrelated nonanemic AS het-

β-Globin-like Gene Cluster Haplotypes

199

erozygotes from Sicily, all linked to the Benin haplotype and differing only by their βA chromosomes, were informative (38). F-cell expression was more strongly associated with LCR-HS2 polymorphism than with XmnI polymorphism. The observed association between XmnI polymorphism and HbF expression is very likely owing to linkage disequilibrium with LCR-HS2 sequences. It is clear that the expression of HbF is under polygenic control involving determinants both linked and unlinked to the β-globin gene cluster on chromosome 11.

1.4. Haplotype Studies of Other Hemoglobinopathies Haplotypes have allowed determination of the unicentric origin of HbC in Africa, with its epicenter in Burkina-Faso (39), although isolated appearances of this mutation elsewhere have been linked to different haplotypes. In addition, correlations between the phenotype and the haplotype have been reported in SC patients (double heterozygotes for HbC and HbS) (40). Extensive studies of haplotype definition in different thalassemic mutations have been studied in different ethnic groups (41–51).

1.5. Miscellaneous Haplotype Research Applications From the panoply of applications of haplotype research there are some recent interesting findings that merit mention: HbG-Coushatta [β-22 (B4) Glu→Ala] is found in the Silk Road region of China but is also present in the North American Coushatta Indians, and a commonality of origin was suspected. Nevertheless, haplotype studies revealed that the Chinese and Louisiana Coushatta had different haplotypes associated with the identical HbG mutation (52). On the other hand, the strong previous evidence of the connection between Amerindian populations and Asian populations has been further strengthened by β-globin gene cluster haplotype analysis of the Huichol Indians of Mexico (53). Haplotype analysis with grouped worldwide populations showed Native Americans as the population closest to the Huichols, followed by Pacific Islanders and Asians. Present observations are consistent with an important Asian contribution to the Huichol genome in this chromosomal region. The relationship between growth in children with SS and the different β-globin haplotypes, as well as components of the insulin-like growth factor (IGF)/IGF-binding protein (IGFBP) axis, has been studied (54). Patients with the Bantu/Bantu haplotype had significantly lower mean growth velocity compared with those with Ben/Ben, and total IGF levels in Bantu/Bantu patients were also lower compared with the Ben/Ben genotype. A positive correlation was found between hematocrit and total IGF-1 and between HbF percentages and the Z-scores for total IGF-1 and IGFBP-3. These interactions suggest that the delayed growth of these patients may be linked to the low circulating concentrations of the various elements of the GH/IGF-laxis, and that the

200

Muralitharan et al.

decrease in total IGF-1 concentrations in patients with Bantu/Bantu haplotype is secondary to the severity of the disease associated with this haplotype. HbF modulates the phenotypic diversity of SS, and the prevalence of many disease complications are related to the level of HbF. Hydroxyurea increases HbF levels in many patients. While the beneficial effects of hydroxyurea are probably the product of many factors, modulation of γ-globin gene expression is only one of them. In a multicenter trial of hydroxyurea (55), F-cells increased in HU-treated patients compared with control subjects, and the increases in the HbF level at 2 yr were greatest in patients who had the highest baseline counts of reticulocytes and neutrophils, two or more episodes of study-defined myelotoxicity, and absence of a Bantu haplotype. The Senegal haplotype is associated with higher expression of HbF in children with sickle cell anemia, compared to the Benin and Bantu haplotypes. Finally, the almost 500 publications that include β-gene-like cluster haplotype research, since its first description, attest to the variety of applications of this technology. 2. Materials 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

Whole blood collected in an anticoagulant (EDTA)-coated tube. NH4Cl. NH4HCO3 (Sigma, St. Louis, MO). NaCl. EDTA. Triisopropylnaphthalene sulfonic acid (TPNS). Butanol. Sodium dodecyl sulfate (SDS). Tris-HCl. Phenol. Chloroform. Octanol. Ethanol. RNase A. Spectrophotometer (Pharmacia). Oligonucleotides. Taq DNA polymerase. dNTP mix (dA, dG, dC, dT). Thermal cycler (Perkin Elmer). Agarose. TAE Buffer (Tris-acetic acid-EDTA). Restriction enzymes. Water bath. Refrigerated centrifuge.

β-Globin-like Gene Cluster Haplotypes

201

3. Methods The methods detailed here pertain to the analysis of genetic variation by restriction endonuclease polymorphism in the β-globin gene that defines a particular haplotype. The methods consist of genomic DNA extraction, polymerase chain reaction (PCR) amplification of β-globin gene, and restriction enzyme digest analysis of the β-globin gene.

3.1. Genomic DNA Extraction The first step in haplotype analysis is the preparation of high molecular weight chromosomal DNA from the nucleated white blood cells (WBCs) from the patient’s whole blood; it is described in Subheadings 3.1.1–3.1.9. The procedure includes preferential lysis of red blood cells (RBCs), lysis of WBCs, phenol chloroform extraction, and ethanol precipitation (see Note 1).

3.1.1. Lysis of RBCs 1. Centrifuge whole blood collected in an anticoagulant (EDTA)-coated tube at low speed (2000g) for 15 min in a refrigerated centrifuge (4°C). 2. Carefully decant the supernatant to avoid the loss of WBCs. After approximate estimation of the volume of the RBC pellet, add five times of RBC lysis solution (1.5 mM NH4Cl and 0.5 mM NH4HCO3). 3. Gently invert the tubes several times and incubate on ice for 10 min. 4. Centrifuge the mix at 2000g for 15 min and decant the lysed RBCs. 5. Repeat the lysis procedure until the pellet is white. The WBC pellet is ready for immediate DNA extraction or can be stored for several months at –70°C.

3.2.1. Lysis of WBCs 1. Resuspend the WBC pellet red in WBC lysis solution (0.15 M NaCl and 0.1 M EDTA, pH 10.5). 2. Approximately 0.8 mL for a pellet from 10–15 mL of whole blood. 3. Add 1 mL of WBC lysis solution (6% TPNS, 8% 2-butanol, and 3% SDS). Complete lysis of WBC is judged by the progressive increase in viscosity.

3.2.2. Phenol-Chloroform Extraction 1. Add an equal volume of neutralized phenol to the WBC lysate. 2. Mix very well and centrifuge at 2000g for 10 min. 3. Recover the aqueous phase.

3.1.4. Octanol Extraction of Chloroform 1. Add an equal volume of a mixture of chloroform:octanol (24:1), mix very well, and centrifuge at 2000g for 10 min. 2. Recover the aqueous phase.

202

Muralitharan et al.

3.1.5. Precipitation of DNA 1. To the aqueous phase, add approx 5 vol of cold absolute ethanol. As the ethanol is being added, the DNA fibers solve out of the solution and are visible. 2. Spool out the filamental DNA using a glass Pasteur pipet. 3. Air-dry the retrieved DNA and suspend in 1 mL of TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.4). 4. Allow the high molecular weight DNA to dissolve completely in a rotating wheel for 6–8 h.

3.1.6. Removal of RNA and Protein Contaminants 1. Mix the dissolved DNA with an appropriate volume of 20X saline sodium citrate (SSC) to reach a final concentration of 2X SSC. 2. Add RNase A to a final concentration of 50 µg/mL and incubate at 37°C for 1 h. 3. Add 20% SDS to a final concentration of 0.1%, and then add proteinase K to a final concentration of 120 µg/mL and incubate at 37°C for 3 h.

3.1.7. Cleaning of DNA 1. After proteinase K digestion, ethanol precipitate the DNA again as in Subheading 3.1.5. Resuspend the DNA in TE buffer as in Subheading 3.1.5., step 4. 2. Extract the dissolved DNA with phenol and chloroform/octanol as in Subheading 3.1.4. 3. After ethanol precipitation and drying the DNA pellet, dissolve the pellet in TE buffer.

3.1.8. Quantitation of DNA 1. After the DNA is completely dissolved in TE buffer, make up a dilution (e.g., 2 µL in 300 µL). 2. Read the optical density (OD) in an ultraviolet (UV) spectrophotometer at 260 and 280 nm. One OD of DNA at 260 nm corresponds to a concentration of 50 µg/mL. An OD260/OD280 ratio in the range of 1.6–2.0 represents a clean DNA with less contaminants.

3.1.9. Agarose Gel Electrophoresis 1. 2. 3. 4. 5. 6. 7. 8. 9.

Analyze the quality of the DNA in a 0.8% agarose gel with TAE buffer (Fig. 1). Dissolve 0.8 g of agarose in 100 mL of 1X TAE buffer. Boil this mixture in a microwave oven. After cooling the melted agarose, add 0.5 mg/mL of ethidium bromide and mix well. Pour the gel in a gel-casting tray with a comb. After the gel is set (15 min), immerse the gel in 1X TAE buffer in the gel tank. Mix 5 µL of extracted DNA with the loading dye and apply on the gel. Carry out electrophoresis at 100 V using a power pack. After electrophoresis, visualize the gel in a UV transilluminator.

β-Globin-like Gene Cluster Haplotypes

203

Fig. 1. Analysis of extracted genomic DNA in 0.8% agrose gel with 1X TAE buffer. Lane 1, λ HindIII marker.

3.2. β-Cluster Haplotype Determination by PCR PCR is a powerful technique in which the target DNA is amplified in vitro exponentially through the enzymatic action of DNA polymerase for specific amplification of a DNA region. The chemical reaction is based on the annealing and extension of two oligonucleotide primers that flank the target region in duplex DNA. The method employs repeated cycles of denaturation (95°C), annealing at 50–64°C of oligonucleotide primers to the target DNA, and enzymatic primer extension (72°C) to amplify the DNA flanked by the primers. In each subsequent cycle, the amplification product serves as a template, and hence the process is exponential. The primary step in haplotype analysis is the PCR amplification of the β-globin gene by using specific oligonucleotides (see Note 2). This includes PCR reaction, analysis of the PCR product in an agarose gel with 1X TAE buffer, restriction enzyme digestion of the PCR product, and analysis of the restriction digest in agarose gel.

204

Muralitharan et al.

Fig. 2 Organization of β-globin gene cluster.

3.2.1 PCR Amplification of β-Globin Gene The β-globin gene cluster consists of a DNA region of 60 kb encompasing six different β-globin-like genes (Fig. 2): (starting from the 5' end of the cluster) the embryonic ε-globin gene, two fetal Gγ- and Aγ-globin genes, the pseudogene (Ψβ), and the adult δ- and β-globin genes. As in the case of the α-globin genes, the β-globin genes share the same configuration with 3 exons (peptideencoding regions) and 2 introns (intervening sequences). Both Gγ and Aγ polypeptides are identical with the exception of a single amino acid at position 136, which is alanine (Aγ) in Aγ, while it is glycine (Gγ) in Gγ. For haplotype analysis, a series of polymorphic sites in the β-globin gene is amplified, the success of the PCR reaction is confirmed by agarose gel electrophoresis analysis of the PCR product, and the polymorphisms are identified by restriction endonuclease enzyme cleavage (Figs. 3 and 4) (Table 1). A standard PCR reaction consists of the following components to a final volume of 50 µL (see Note 3): 5 µL of 10X PCR buffer, 1 µL (10 pm/µL) of forward primer, 1 µL (10 pm/µL) of reverse primer, 0.25 µL (25 mM) of dNTP mix, 1.5 mM (25 mM) MgCl2, 1 µL (100 ηg) of DNA, 0.25 µL (5 U/µL) of Taq polymerase, and 40 µL of H2O. A standard PCR cycling condition consists of the following steps: 94°C for 5 min; 94°C for 30 s; 55°C for 1 min; 72°C for 45 s; 72°C for 7 min.

3.2.2. Restriction Enzyme Digest Analysis For restriction enzyme analysis, 10 µL of the PCR product is used. A standard restriction enzyme analysis consists of the following components to a final volume of 15 µL (see Note 4): 10 µL of PCR product, 1 µL (10 U) of restriction enzyme, 1.5 µL of 10X buffer, and 2.5 µL of H2O. This mixture is incubated at 37°C for a minimum period of 3 h. After the incubation is complete, the restriction analysis is carried out in an agarose gel electrophoresis with 1X TAE buffer. 4. Notes 1. During ethanol precipitation of the DNA, in most of the samples you can very easily observe that the nucleic acid is precipitated and floats like a fiber. In some samples, the DNA is not visible because of the poor yield owing to partial lysis of the leukocytes. In these samples, it is absolutely necessary that the samples be

β-Globin-like Gene Cluster Haplotypes

205

Fig. 3. Analysis of restriction enzyme digests (HincII and HindIII) in agarose gel with 1X TAE buffer. +/+, presence of the restriction site in both chromosomes (homozygous). -/-, absence of the restriction site.

Fig. 4. Diagrammatic representation of different β-globin-like gene cluster haplotypes.

206

Table 1 Primer Sequences for Different Polymorphic Sites Restriction site HincII ε XmnI Gγ HindIII Gγ

206

TaqI HindIII Aγ HincII 5'ψβ HincII 3'ψβ

β-HinfI

TCT CTG TTT GAT GAC AAA TTC AGT CAT TGG TCA AGG CTG ACC AAC TGT TGC TTT ATA GGA TTT T AGG AGC TTA TTG ATA ACT CAG AC AGT GCT GCA AGA AGA ACA ACT ACC CTC TGC ATC ATG GGC AGT GAG CTC CCT GAC CAG GAA CCA GCA GA CTT ATC GGA GGC AAG CTG TAT CT TGC TGC TAA TGC TTC ATT ACA A TAA ATG AGG AGC ATG CAC ACA C TCC TAT CCA TTA CTG TTC CTT GAA ATT GTC TTA TTC TAG AGA CGA TTT TCT GCA TTT GAC TCT GTT AGC GGA CCC TAA CTG ATA TAA CTA ACT CCC AGG AGC AGG GAG GGC AGG TTC GTC TGT TTC CCA TTC TAA ACT AGT AGA GGC TTG ATT TGG AGG GTT AAG GTG GTT GAT GGT AAC

Absence of site (bp)

Presence of site (bp)

Annealing temperature (°C)

760

760

446 + 314

58

655

655

450 + 205

54

328

328

91 + 237

65

960

960

753 + 207

64

761

761

435 + 325

65

794

794

104 + 690

64

620

620

540 + 80

54

1152

413 + 644 + 95

331 + 82 + 644 + 95

65

638

336 + 302

123 + 213 + 302

58

Muralitharan et al.

RsaI 5' β

Primer sequence (5'–3')

Product size (bp)

β-Globin-like Gene Cluster Haplotypes

207

kept in –20°C for 1 h and then centrifuged at 2000g to pellet the DNA. After centrifugation, the aqueous phase is removed and the DNA pellet is dried. 2. Different polymorphic sites in the β-globin gene cluster are amplified by using appropriate pair of primers. The annealing temperature of the particular pair of primers depends on the sequence composition of the primers. Care should be taken to set up the PCR cycling condition with an appropriate annealing temperature. 3. Each PCR reaction is standardized by varying the concentration of MgCl2, dNTP and Taq polymerase. Based on the standardization, an optimum concentration of the different components of the PCR is decided for the final reaction. The final volume of the PCR reaction can be changed into different volumes such as 25 and 10 µL and the different components of the PCR reaction are used in an appropriate concentration ratio to the final volume. 4. Usually, restriction enzyme analysis is carried out according to the conditions provided by the enzyme’s manufacturer. For example, some enzymes need bovine serum albumin in the reaction (0.1%). The optimum incubation temperature for most of the restriction enzymes is 37°C, but some enzymes need to be incubated at a higher temperature. The appropriate reaction conditions and the optimum temperature are usually given in the manufacturer’s instructions.

References 1. Kan, Y. W. and Dozy, A. M. (1978) Polymorphism of DNA sequence adjacent to human β-globin structural gene : relationship to sickle mutation. Proc. Natl. Acad. Sci. USA 75, 5631–5635. 2. Kan, Y. W. and Dozy, A. M. (1980) Evolution of the hemoglobin S and C genes in world populations. Science 209, 388–391. 3. Orkin, S. H. (1978) Selective restriction endonuclease cleavage of human globin genes. J. Biol. Chem. 253, 12–15. 4. Orkin, S. H., Kazazian, H. H., Antonarakis, S. E., Goff, S. C., Boehm, C. D., Sexton, J. P., Waber, P. G., and Giardina, P. J. (1982) Linkage of β-thalassaemia mutations and beta-globin gene polymorphisms with DNA polymorphisms in human β-globin gene cluster. Nature 296, 627–631. 5. Mears, J. G., Beldjord, C., Benabadji, M., Belghiti, Y., Baddou, M. A., Labie, D., and Nagel, R. L. (1981) The sickle gene: polymorphisms in North Africa. Blood 58, 599–601. 6. Mears, J. G., Lachman, H. M., Cabannes, R., Amegnizin, K. P., Labie, D., Nagel, R. L. (1981). Sickle gene: its origin and diffusion from West Africa. J. Clin Invest. 68, 606–610. 7. Pagnier, J., Mears, J. G., Dunda-Belkhodja, O., Schaefer-rego, K. E., Beldjord, C., Nagel, R. L., and Labie, D. (1984) Evidence for the multicentic origin of the sickle cell hemoglobin gene in Africa. Proc. Natl. Acad. Sci. USA 81, 1771–1773. 8. Nagel, R. L. and Steinberg, M. H., (2000) Genetic variability in sickle cell anemia, in Disorders of Hemoglobin: Genetics, Pathophysiology, Clinical Management (Steinberg, M. H., Forget, B. G., Higgs, D. R., and Nagel, R. L., eds.), Cambridge University Press, Cambridge, MA.

208

Muralitharan et al.

9. Chakravarti, A., Buteow, K. H., Antonarkis, S. E., Aber, P. G., Boehm, C. D., and Kazazian H. H. (1984) Non uniform recombination within the human β globin gene cluster. Am. J. Hum. Genet. 36, 1239–1235. 10. Srinivas, R., Dunda, O., Krishnamoorthy.R., Fabry, M. E., Georges, A., Labie, D., and Nagel, R. L. (1988) Atypical haplotypes linked to the βS gene in Africa are likely to be the product of recombination. Am. J. Hematol. 29, 60–62. 11. Lapoumerouli, C., Dunda, O., Ducrocq, R., Trabuchet, G., Mony-Lobe, M., Bodo, J. M., Carnevale, P., Labie, D., Elion, J., and Krishnamoorthy, R. (1992) A novel sickle cell mutation of yet another origin in Africa. Hum. Genet. 89, 333–337. 12. Chebloune, Y., Pagnier, J., Trabuchet, G., Faure, C., Verdier, G., Labie, D., and Nigon, V., (1988) Structural analysis of the 5' flanking region of the β-globin gene in African sickle cell anemia patients: further evidence for three origins of the sickle cell mutation in Africa. Proc. Natl. Acad. Sci. USA 85, 4431–4435. 13. Labie, D., Srinivas, R., Dunda, O., Dode, C., Lapomeroulie, C., Devei, V., Devi, S., Ramasami, K., Elion, J., Ducroq, R., Krishnamoorthy, R., and Nagel, R. L. (1989) Haplotypes in tribal Indians bearing the sickle gene: evidence for the unicentric origin of the β S mutation and the unicentric origin of the tribal populations of India. Hum. Biol. 61, 479–491. 14. Kotea, N., Baligadoo, S., Surran, S., Ramasawmy, R., Lu, C. Y., Ducrocq, R., Labie, D., Krishnamoorthy R., and Nagel, R. L. (1995) Bicentric origin of sickle hemoglobin among the inhabitants of Mauritius Island. Blood 86, 407–408. 15. Nagel, R. L., Fabry, M. E., Pagnier, J., Zohoum, I., Wajcman, H., Baudin, V., and Labie, D (1985) Hematologically and genetically distinct forms of sickle cell anemia in Africa: the Senegal type and the Benin type. N. Engl. J. Med. 312, 880–884. 16. Nagel, R. L., Rao, S. K., Dunda-Belkhodja, O., Connolly, M.M, Fabry, M. E., Georges, A., Krishnamoorthy, R., and Labie, D. (1987) The Hematologic characteristics of sickle cell Anemia bearing the Bantu haplotype: the relatioship between G γ and HbF level. Blood 69, 1026–1030. 17. Steinberg, M. H., Lu, Z. H., Nagel, R. L., Venkataramani, S., Milner, P. F., Huey, L., Safaya, S., and Reider, R. F. (1998) Hematological effects of atypical and Cameroon β-globin gene haplotypes in adult sickle cell anemia. Am. J. Hematol. 59, 121–126. 18. Powars, D. R. (1991) βS-gene cluster haplotypes in sickle cell anemia. Clinical and hematological features. Hematol. Oncol. Clin. North Am. [review] 5(3), 475–493. 19. Wainscoat, J. S., Bell J. I., Thein, S. L., Higgs, D. R., Sarjeant, G. R., Peto, T. E., and Weatheral, D. J. (1983) Multiple origins of the sickle mutation: evidence from βS-globin gene cluster polymorphisms. Mol. Biol. Med. 1, 191–197. 20. Antonarakis, S. E., Boehm, C. D., Serjeant, G. R., Theisen, C. E., Dover, G. J., and Kazazian H. H. Jr. (1984) Origin of the βS-globin gene in blacks: the contribution of recurrent mutation or gene conversion or both. Proc. Natl. Acad. Sci. USA 81, 853–856. 21. Ragusa, A., Lombardo, M., Sortino, G., Lombardo, T., Nagel, R. L., and Labie, D.(1988) βS gene in Sicily is in linkage disequilibrium with the Benin haplotype: Implications for gene flow. Am. J. Hematol. 27, 139–141.

β-Globin-like Gene Cluster Haplotypes

209

22. Dar, S., Hussain, H. M., Gravel, D., Nagel, R. L., and Krishnamoorthy, R.(2000) Genetic epidemiology of HbS in Oman: multicentric origin of the βS gene. Am. J. Hematol. 64, 39–46. 23. Rahgozar, S., Poorfathollah, A. A., Moafi, A. R., and Old, J. M. (2000) βS gene in central Iran is in linkage disequilibrium with the Indian-Arab haplotype. Am. J. Hematol. 65, 192–195. 24. Currat, M., Trabuchet, G., Rees, D., Perrin, P., Harding, R. M., Clegg, J. B., Langane, A., and Excoffier, L. (2002) Molecular analysis of the β-globin gene cluster in the Niokholo Mandenka population reveals a recent origin of the βS Senegal mutation. Am. J. Hum. Genet. 70, 207–223. 25. Cuellar-Ambrosi, F., Mondragon, M. C., Figueroa, M., Prehu, C., Galceteros, F, and Ruiz-Linares, A. (2000) Sickle cell anemia and β-globin gene cluster haplotypes in Colombia. Hemoglobin 24, 221–225. 26. Romana, M., Keclard, L., Froger, A., Lavocat, E., Saint-Martin, C., Berchel, C., and Merault, G. (2000) Diverse genetic mechanisms operate to generate atypical βS-haplotypes in the population of Guadeloupe. Hemoglobin 24, 77–87. 27. Arends, A., Alvarez, M., Velazquez, D., Bravo, M., Salazar, R., Guevara, J. M., and Castillo, O.(2000) Determination of β globin gene cluster haplotypes and prevalence of α-thalassemia in sickle cell anemia patients in Venezuela. Am. J. Hematol. 64, 87–90. 28. Pante-DeSousa, G., Mousinho-Ribeiro, R. C., Dos Santos, E. J., Guerreiro, J. F. (1996) β-globin haplotypes analysis in Afro-Brazilians from the Amazon region: evidence for a significant gene flow from Atlantic West Africa. Ann. Hum. Biol. 26, 365–373. 29. Frikha, M., Fakhfakh, F., Mseddi, S., Gargouri, J., Ghali, L., Labiadh, Z., Harrabi, M., Soussi, T., and Ayadi, H. (1998) Hemoglobin βS haplotype in the Kebilli region (Southern Tunisia). Transfus. Clin. Biol. 5, 166–172. 30. El-Kalla, S. and Baysal, E. (1998) Genotype-phenotype correlation of sickle cell disease in the United Arab Emirates. Pediatr. Hematol. Oncol. 15, 237–242. 31. Hewitt, R., Krause, A., Goldman, A., Cambell, G., and Jenkins, T. (1996) Beta globin haplotype analysis suggests that a major source of Malagasy ancestry is derived from Bantu-speaking Negroids. Am. J. Hum. Genet. 58, 1303–1308. 32. Keclard, L., Ollendorf, V., Berchel, C., Loret, H., and Merault, G. (1996) Haplotypes, α-globin gene status and hematological data of sickle cell disease patients in Guadeloupe (F. W. I.). Hemoglobin 20, 63–74. 33. Muniz, A., Corral, L., Alaez, C., Svarch, E., Espinosa, E., Carbonell, N., Di Leo, R., Fellicetti, L., Nagel, R. L., and Martinez, G. (1995) Sickle cell anemia and β-globin gene cluster haplotypes in Cuba. Am. J. Hematol. 49, 163–164. 34. Sow, A., Peterson, E., Josifovska, O., Fabry, M. E., Krishnamoorthy, R., and Nagel, R. L. (1995) Linkage-disequilibrium of the Senegal haplotype with the βS-gene in the Republic of Guinea. Am. J. Hematol. 50, 301–303. 35. Gilman, J. G. and Huisman, T. H. (1984) Two independent genetic factors in the β-globin gene cluster are associated with high Gγ levels in the HbF of SS patients. Blood 64, 452–457. 36. Labie, D., Dunda-Belkhodja, O., Rouabhi, F., Pagnier, J., Ragusa, A., and Nagel, R.L (1985) The -158 site 5' to the Gγ gene and G gamma expression. Blood 66, 1463–1465.

210

Muralitharan et al.

37. Perichon, B., Ragusa, A., Lapomeroulie, C., Romand, A., Moi, P., Ituka, T., Labie, D., Elion, J., and Krishnamoorthy, R. (1993) Inter ethnic polymorphism of the β-globin gene locus control region (LCR) in sickle cell anemia patients. Hum. Genet. 91, 464–468. 38. Merghoub, T., Perichon, B., Maier-Redelsperger, M., et al. (1997) Dissection of the association status of two polymorphisms in the β-globin gene cluster with variations in F-cell number in non-anemic individuals. Am. J. Hematol. 56, 239–243. 39. Boehm, C. D., Dowling, C. E., Antonarakis, S. E., Honig, G. R., and Kazazian, H. H. Jr. (1985) Evidence supporting a single origin of the βC-globin gene in blacks. Am. J. Hum. Genet. 37, 771–777. 40. Steinberg, M. H., Nagel, R. L., Lawrence, C., Swaminathan, V., Lu, Z. H., Plonczynski, M., and Harrel, A. (1996) Beta-globin gene haplotype in Hb SC disease Am. J. Hematol. 52, 189–191. 41. Perrin, P., Bouhassa, R., Mselli, L., Garguier, N., Nigon, M., Bennani, C., Labie, D. and Trabuchet, G. (1998) Diversity of sequence haplotypes associated with β-thalassemia mutations in Algeria: implications for their origin. Gene 213, 169–177. 42. Tadmouri, G.O, . Garguier, N., Demont, J., Perrin, P., and Basak, A. N. (2001) History and origin of β-thalassemia in Turkey: sequence haplotype diversity of β-globin genes. Hum. Biol. 73, 661–674. 43. Zahed, L., Qatanani, M., Nabulsi, M., Taher, A. (2000) β-thalassemia mutations and haplotype analysis in Lebanon. Hemoglobin 24, 269–276. 44. Kotea, N., Ramasawmy, R., Lu, C. Y., Fa, N. S., Gerard, N., Beesoon, S., Ducrocq, R., Surrun, S. K., Nagel, R. L., and Krishnamoorthy, R. (2000) Spectrum of β-thalassemia mutations and their linkage to β-globin gene haplotypes in the IndoMauritians. Am. J. Hematol. 63, 11–15. 45. Bandyopadhyay, A., Bandyopadhyay, S., Chowdhury, M. D., and Dasgupta, U. B. (1999) Major β-globin gene mutations in eastern India and their associated haplotypes. Hum. Hered. 49, 232–235. 46. Samakoglu, S., Philipsen, S., Grosveld, F., Luleci, G., and Bagci, H.(1999) Nucleotide changes in the γ-globin promoter and the (AT)X NY(AT)Z polymorphic sequence of beta LCRHS-2 region associated with altered levels fo HbF. Eur. J. Hum. Genet. 7, 345–356. 47. Garner, C., Mitchell, J., Hatzis, T., Reittie, J., Farrall, M., and Thein, S. L. (1998) Haplotype mapping of a major quantitative-trait locus for fetal hemoglobin production on chromosome 6q23. Am. J. Hum. Genet. 62, 1468–1474. 48. Perrin, P., Bouhassa, R., Mselli, L., Garguier, N., Nigon, V. M., Benani, C., Labie, D., and Trabuchet, G.(1998) Diversity of sequence haplotypes associated with β-thalassemia mutations in Algeria: Implications for their origin. Gene 213, 169–177. 49. Furumi, H., Firdous, N., Inoue, T., Ohta, H., Winichagoon, P., Fuchaaroen, S., and Fukumaki, Y. (1998) Molecular basis of β-thalassemia in the Maldives. Hemoglobin 22, 141–151.

β-Globin-like Gene Cluster Haplotypes

211

50. Rund, D., Oron-Karni, V., Filon, D., Goldfarb, A., Rachmilewitz, E., and Oppenheim, A. (1997) Genetic analysis of β-thalassemia intermedia in Israel: diversity of mechanisms and unpredictability of phenotype. Am. J. Hematol. 54, 16–22. 51. Pacheco, P., Loureiro, P., Faustino, P., and Lavinha, J.(1996) Haplotypic heterogenity of β thalassemia IVS1(G A) mutation in Southern Portugal. Br. J. Hematol. 94, 767. 52. Li, J., Wilson, D., Plonczynski, M., Harrel, A., Cook, C. B., Scheer, W.D, Zeng, Y.T Coleman, M. B., and Steinberg, M. H. (1999) Genetic studies suggests a multicentric origin for Hb G-Coushatta [β22(B4)Glu Ala]. Hemoglobin 23, 57–67. 53. Villalobos-arambula, A. R., Rivas, F., Sandoval, L., Perea, F. J., Casas-Castaneda, M., Cantu, J. M., and Ibara, B. (2000) β(A) globin gene haplotypes in Mexican Huichols: Genetic relatedness to other populations. Am. J. Human Biol. 12, 201–206. 54. Luporini, S. M., Bendit, I, Manhani, R., Bracco, O. L., Manzella, L., and GianellaNeto, D. (2001) growth Harmone and insulin like growth factor I axis and growth of children with different sickle cell anemia haplotypes. Pediatr. Hematol. Oncol. 23, 357–363. 55. Steinberg, M. H. Determinants of fetal hemoglobin response to hydroxyurea. Semin. Hematol. 199: 34, 8–14. 56. Green, N. S., Fabry, M. E., Kaptue-Noche, L., and Nagel, R. L. (1993) Senegal haplotype is associated with higher HbF than Benin and Cameroon haplotypes in African children with sickle cell anemia. Am. J. Hematol. 44, 145–146.

212

Muralitharan et al.

Transgenic Mice and Hemoglobinopathies

213

13 Transgenic Mice and Hemoglobinopathies Mary E. Fabry, Eric E. Bouhassira, Sandra M. Suzuka, and Ronald L. Nagel 1. Introduction A wide variety of transgenic mouse models expressing human globin genes have been generated in the last decade and a half. The majority of these models have been sickle transgenic models because they can advance our understanding of the pathophysiology of sickle cell disease, aid in development of therapeutic approaches, aid in development of preclinical strategies for antisickling gene therapy, and may be used for the identification and study of pleiotropic and epistatic genes. One of the most interesting features of the current sickle cell models is the extent to which they reproduce the pleiotropic aspects of sickle cell disease. That is, simply introducing a single amino acid mutation (β6 glu→val, which results in an abnormal β globin, βS) results in a polymerizable hemoglobin (Hb) and generates a wide range of pathology in mice that mimics that found in human sickle cell disease. Another important aspect of pathology in mice as well as in humans are epistatic effects (see Note 1). That is, the phenotype of a given genotype is influenced by the interaction of many genes. Furthermore, these genes may be polymorphic. In some inbred strains of mice, polymorphisms (see Note 2) may be fixed, and the combinations of inbred strains that are used to generate transgenic mice may result in an array of epistatic effects that would not be present if the mice were bred onto a genetically defined background. Epistatic effects in mouse models may differ from those in humans due to different polymorphisms and/or different isozymes or tissue expression of genes (see Note 3). From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

213

214

Fabry et al.

1.1. Early Transgenic Models for Sickle Cell Disease The earliest attempts at producing a murine model for sickle cell disease introduced only the mutated human β globin (βS) (1) and produced a mouse without pathology. It was subsequently shown that mouse α globin inhibits polymer formation as effectively as human γ globin (2). All successful transgenic models of sickle cell disease have approximately equal expression of human α globin and βS.

1.1.1. Early Models The first models expressing moderately high levels of both human α- and βS-globin chains were reported by Greaves et al. (3) and Ryan et al. (4). The mice produced by Greaves et al. (3) failed to propagate but those produced by Ryan et al. (4) were bred with thal mice that have a deletion of mouse βmajor; this increases expression of βS by reducing competition from the mouse β globins. Ryan et al. (4) used two constructs: a locus control region (LCR) with the human α1 gene and an LCR with the βS gene. These were then coinjected and the founder mice were bred with thal mice to produce mice that were hemizygous for the thal deletion. The HbS/thal exhibited intracellular polymer formation and sickling but had normal Hb and reticulocyte counts.

1.1.2. SAD Mouse The SAD mouse (5) expresses SAD-Hb at a low level (about 55% human α and 19% HbSAD), but this Hb has super-HbS properties. In HbSAD the sickle mutation is accompanied by two other mutations that enhance polymer formation (5). Sickling, polymer formation, and irreversibly sickled cells (ISCs) have been observed. The SAD mouse has normal hematocrit and reticulocyte count. The red cells have increased density, and this property has been used to test clotrimazol, a Gardos channel blocker that inhibits sickle dense cell formation (6) (dense cells are known to participate in vasoocclusion) and to test Mg++ supplementation, aimed at inhibiting red cell K-Cl cotransport (7), also involved in dense sickle cell formation. A form of clotrimazol, modified to eliminate the anti-P450 activity, is now in clinical trials. A more severe form of the SAD mouse was generated by breeding to hemizygosity with the βmajor deletion to produce a SAD-thal mouse (5). The SAD-thal mouse has 26% SAD as opposed to 19% in the SAD mouse and an increase in reticulocytes from 2.6 to 6.2; however, it is more difficult to breed and has not been used in the majority of work reported. The effect of adding γ to the SAD mouse has been studied by breeding in lines expressing two different levels of human γ (8). Since the percentage of reticulocytes and the level of Hb are normal, no changes in these features were found; however, the life-span

Transgenic Mice and Hemoglobinopathies

215

of SAD mice is shortened from the 700 d found for C57BL to 474 d, and in the SAD mouse with 18.8% γ the life-span is increased to 552 d, the percentage of very dense cells was decreased, and the glomerular profile area was decreased. At high levels of γ, the SAD transgene became homozygous, suggesting that competition with γ had reduced the percentage of βSAD below the level required to produce pathology.

1.1.3. NY1DD Model The NY1DD model (9,10) expresses about 75% βS and 55% human α-chains (the balance are murine globins) and presents a moderate phenotype, owing to partial inhibition of polymerization by murine α-chains. Intracellular polymers of HbS are generated on deoxygenation and the red cells sickle. A deoxy potassium efflux can also be detected and linked to increased red cell density, a property that is unique to sickle cell disease and that increases pathology by increasing the intracellular Hb concentration, thereby favoring polymer formation and sickling. The model exhibits a hypoxia-inducible urine concentration defect and constitutively increased glomerular filtration rate, features of sickle cell anemia. Peripheral retinopathy and choroid infarcts were also detected (11), the latter not previously recognized as a complication of this disease, but later confirmed in human autopsies. Interestingly, this model also provided the first in vivo evidence of malaria protection afforded by HbS (12), as well as in vivo evidence of oxidative stress in hypoxia-exposed sickle mice (13).

1.1.4. The S+S-Antilles Model The S+S-Antilles model (14) is similar to the first three models, in that it is not anemic but, in contrast to the first three models, it does have a mild reticulocytosis and a constitutive urine concentration defect. Introduction of 11% human γ (Aγ+Gγ/Aγ+Gγ+βS+βS-Antilles+βmin) into this mouse line reduces reticulocytosis, normalizes red cell density, and corrects the urine-concentrating defect. In vivo adhesion of sickle cells to the endothelium was demonstrated for the first time in this model (15). Kaul et al. (16) showed an increased endothelial nitric oxide synthase (eNOS) expression, low mean arterial blood pressure, and constitutive microcirculatory vasodilation by video microscopy in an in vivo cremaster preparation of the S+S-Antilles mouse. A nonselective inhibitor of NOS (L-NAME) caused a significant increase in mean arterial blood pressure and decrease in the diameters of cremaster muscle arterioles that was reversible after administration OF L-arginine in both control and transgenic mice, confirming NOS activity. Based on data on nitric oxide (NO)-mediated vasodilators and forskolin (a cyclic adenosine monophosphate–activating agent), Kaul et al. (16) concluded that increased eNOS/NO activity results in lower blood pressure and diminished arteriolar responses to NO-mediated

216

Fabry et al.

vasodilators. Although the increased NOS/NO activity may compensate for flow abnormalities, it may also cause pathophysiological alterations in vascular tone. S+S-Antilles mice have a reduced plasma arginine level (17) similar to that reported for human sickle cell patients (18). Arginine is required for the production of NO, a vasodilator that plays a crucial role in maintaining vascular tone. In arginine-supplemented S+S-Antilles mice red cell density, the percentage of high-density red cells and the Vmax of the Ca++-activated K+-channel (K[Ca] or Gardos channel) are all reduced (17). Specifically, the clotrimazolesensitive, deoxy-stimulated K+ efflux was reduced in red cells from supplemented vs nonsupplemented mice, suggesting the involvement of the Gardos channel. This was confirmed by direct measurements of the significant reduction in the channel Vmax. Animals taken off the diet for 2 mo and retested returned to baseline. Hence, 5% arginine supplementation in sickle transgenic mice significantly reduces deoxy K+ efflux and normalizes red cell density by inhibiting K(Ca) channel activity. Poorly oxygenated tissues are at increased risk of vasoocclusion in the presence of polymerizable Hbs. Fabry et al. (19) have detected deoxyHb in kidney and liver of the sickle cell disease mouse model S+S-Antilles by using blood oxygen level–dependent magnetic resonance imaging (BOLD-MRI), in which image intensity is compared in tissues while the animals first breathe room air and then 100% O2. If elevated deoxyHb is the result of to a reduction in flow or partial obstruction by sickled cells, then infusion of oxygen-carrying material with small particle size (<0.4 µm) such as a perfluorocarbon emulsion (PFCE) could improve oxygenation and flow. A PFCE comprised primarily of Perflubron® was administered to S+S-Antilles mice at volumes equivalent to 5, 10, and 20% of blood volume (BV) by tail vein and BOLD-MRI images and T2 maps were obtained. The intensity of change was compared with that seen in C57 mice and C57 mice injected with 10% BV of PFCE. PFCE at 10% of BV results in a larger reduction in the change of image intensity in S+S-Antilles mice when compared to that observed in C57BL control mice. These observations demonstrate that infusion of PFCE results in reduction in deoxyHb in S+SAntilles mouse kidney and liver because the excess percentage change in signal intensity is proportional to the deoxyHb present and PFCE reduced that percentage in S+S-Antilles mice to that seen in control mice. Infusion of PFCE may reduce risk of extending and/or alleviate sickle cell vasoocclusion during sickle cell painful crisis or sickle liver crisis.

1.2. Sickle Knockout Mice that Express Exclusively Human Hbs Four groups have reported sickle transgenic mouse models in which all of the murine globin genes are “knocked out” and express only human globins

Transgenic Mice and Hemoglobinopathies

217

(20–23). Although the earlier models successfully reproduced much of the pathology of sickle cell disease, they did not exhibit the anemia that gives the disease its name. All of the knockout (KO) mice have significant anemia, reticulocytosis, and abundant irreversibly sickled cells. The first two sickle knockout models were described by Ryan et al. (20) and Paszty et al. (21). Ryan et al. (20) coinjected two constructs, LCRα1 and LCRAγβS, and reported six independent lines of mice. The resulting mice had adult HbF between 3.2 and 7.7%. There was some amelioration of reticulocytosis and anemia in mice with higher levels of γ; however, since these are independent lines with different integration sites and potential differences in globin chain expression and balance, it is difficult to make a direct comparison. The mice described by Paszty et al. (21) (referred to hereafter as BERK mice) were generated by coinjecting three constructs: the LCR, α1, and GγAγδβS. This resulted in one successful line with minimal expression of HbF in mature mice. Chang et al. (22) generated a transgenic line of sickle mice using two copies of the human α2 gene linked to an LCR and a βS yeast artificial chromosome (YAC). This line of mice is severely anemic (Hct of 22 vs 47, sickle vs control, respectively), has a low mean corpuscular Hb (MCH) (10.4 vs 16.1, sickle vs control, which suggests the possibility of β-thalassemia-like characteristics), and has a high reticulocyte count (20 vs 1.1, sickle vs control). This is the first model to incorporate the majority of the β-globin locus and is particularly suitable for studying regulation of the γ genes. Finally, Fabry et al. (23) used the previously described sickle transgenic line (NY1) generated by Costantini and introduced murine α- and βglobin KOs and three different levels of human HbF. These mice express exclusively human Hb and we refer to them as NY1KO mice.

1.2.1. Birmingham Mouse and Berkeley Mouse The Birmingham mouse described by Ryan et al. (20) is very anemic with a decrease in Hb that is almost double that seen in sickle cell anemia. When averaging the Hbs of Birmingham mice in which more than one animal was studied, the value is 4.9 g/dL, a decrease of 10.1 g/dL from the control C57BL. The average Hb for sickle cell anemia patients is 8.4 g/dL, a decrease of 5.6 g/dL. Although the original publication describing the BERK mice did not report a Hb level (24), we find a value of 6.1 g/dL. The NY1KO mice described in Subheading 1.2.2. have higher values for Hb, 9.8 g/dL and 11.9 g/dL for the NY1KO-γM and NY1KO-γH mice, respectively. As described in a previous publication (23), the BERK and the NY1KO-γM mice have very similar hematocrits and reticulocyte counts, so the difference in Hb is the result of to the large difference in mean corpuscular hemoglobin concentration (MCHC) between the two types of mice. The imbalance in chain synthesis and very low MCHC of the BERK mice suggests that they have a thalassemic component.

218

Fabry et al.

The organ damage reported in BERK mice is more severe than that reported in earlier models, with grossly enlarged spleens that are characteristic of all sickle KO mice. The bad news is the very low MCHC (18.5 vs 33.8 g/dL [control]) observed in the BERK line that provides a form of protection against sickling not found in human disease. This feature, plus chain imbalance and the decrease in mean corpuscular hemoglobin (MCH), strongly suggests the presence of a coexisting thalassemia-like defect (23). Hence, how much of the anemia is thalassemia-like and how much is sickle is not clear. The presence of thalassemia in this model suggests that the amelioration of the phenotype by thalassemia is indispensable for the survival of an otherwise too-severe model. This interpretation is further supported by the usually high number of matings necessary to obtain homozygotes; by the presence in the mouse of hyperosmolar plasma (330 vs 290, mouse vs human), and the high intraerythrocytic 2,3-diphosphoglycerate (DPG) (about double that found in humans) that facilitates polymer formation (25). These last three features are characteristics of all sickle KO transgenic mice studied to date. In contrast, it has been relatively easy to generate mice that express exclusively human nonsickling Hbs (such as HbA or HbC (26)). In effect, in the absence of ameliorating features such as thalassemia (low MCHC) or added HbF, sickle cell anemia is lethal in the KO mice that express exclusively human Hbs studied up to now.

1.2.2. KO Mice Expressing Various Levels of Human γ Globin (NY1KO) Fabry et al. (23) have recently reported on mice expressing exclusively human sickle Hb with three levels of HbF. Mice with the least adult HbF expression were the most severe. A progressive increase in HbF from <3 to 20 to 40% (which are called NY1KO-γL, -γM, and -γH, respectively; see Table 1) correlated with a progressive increase in hematocrit (22 to 34 to 40%) and a progressive decrease in reticulocyte count (60 to 30 to 13%). High HbF normalized urine-concentrating ability, and tissue damage detected by histopathology and organ weight were ameliorated by increased HbF. The introduction of γ into the BERK model also progressively corrected its “thalassemic” aspects. Mean survival time reflects the cumulative effects of chronic organ damage and acute ischemic events. Kaplan-Meier mean survival times for NY1KO mice were calculated, and those with the least adult HbF had the shortest mean survival. Higher levels of HbF resulted in progressively longer mean survival. In each class, an approximately equal number of mice were withdrawn from the calculation owing to sacrifice for experimentation. In the NY1KO-γM group only, males had a shorter survival. This excludes the effect of HbF on survival because all γM mice express the same transgenes and have the same level of HbF expression. Hence, this mouse model clearly points to gender effects other than HbF.

Transgenic Mice and Hemoglobinopathies

219

1.3. What Is the Perfect Transgenic Model for Sickle Cell Anemia? Is there a perfect model? By definition, it is not possible. HbS is a mutation of a single gene, while sickle cell anemia is the product of many genes (multigenic) because the phenotype is affected by epistatic (modifier) genes in addition to the indispensable primary mutation (sickle gene). These epistatic genes vary among individuals, explaining the interindividual variation in severity in sickle cell anemia; can ameliorate or aggravate different components of the phenotype; and can interact with each other, in an additive, a neutral, or a canceling fashion. In transgenic mice, potential epistatic genes may differ from those in humans: for example, the red cells of the KO mouse contain only human Hbs, but they are literally half the size of human red cells, and their membrane and cytosolic components are murine.

1.4. Transgenic Mice Are Useful for Identification and Study of Pleiotropic Effects and Epistatic Genes The KO mice have now completed the panoply of phenotypic intensities. For testing therapeutics and inducible pathology, the NY1DD model is useful because it is easy to breed and has hypoxia-inducible pathology. The S+SAntilles and the SAD-mouse models are useful in pathophysiological studies and are relatively easy to breed. The KO mice are excellent models for longterm pathology studies and, particularly if the thalassemic features are corrected, for gene therapy dry runs.

1.5. Transgenic Mice Expressing HbC The homozygous form of HbC (CC disease) increases red cell density, a feature that is the major factor underlying the pathology in patients with sickle cell (SC) disease (27). The basis for the increased red cell density has not yet been fully defined. An HbC founder mouse expressing 56% human α and 34% human βC was bred to KOs of mouse α and β globins in various combinations including full KO of all mouse globins (26). All partial KOs have normal MCH. Full KOs, which express exclusively HbC, have minimally reduced MCH and a ratio of β- to α-globin chains of 0.88 determined by chain synthesis; hence, these mice are not thalassemic. Mice with βC >30% have increased MCHC, dense reticulocytes, and increased K:Cl cotransport. Red cell morphology studied by scanning electron microscopy is strikingly similar to that of human CC cells with bizarre folded cells. Red cells of these mice have many properties that closely parallel the pathology of human disease in which HbC is the major determinant of pathogenesis. These studies also establish the existence of interactions with other gene products that are necessary for pleiotropic effects (red cell dehydration, elevated K:Cl cotransport, morphological changes) that

220

Table 1 Mouse Nomenclature

Nickname THAL NY1DD SAD SAD-thal S-AntillesDD S+S-Antilles

220

NY1KO γL NY1KO γM NY1KO γH BERKd BERK γMd

αβ-Globin transgene name

Description of transgene

— NY1 SAD SAD S-Antilles NY1 and S-Antilles NY1 NY1 NY1 BERK BERK

— miniLCRα2, miniLCRβS a miniLCRα2, miniLCRβSAD b miniLCRα2, miniLCRβSAD HS2α2, HS2βS-Antilles c miniLCRα2, miniLCRβS + HS2α2, HS2βS-Antilles miniLCRα2, miniLCRβS miniLCRα2, miniLCRβS miniLCRα2, miniLCRβS miniLCR, α1, GγAγδβSd miniLCR, α1, GγAγδβSd

γ Construct

Mouse α-knockout

Mouse β-knockout or deletion

— — — — — —

+ // + + // + + // + + // + + // + + // +

Hbbth-1//Hbbth-1 h Hbbth-1//Hbbth-1 h + // + +//Hbbth-1 h Hbbth-1//Hbbth-1 h Hbbth-1//Hbbth-1 h

γLe γMf γHf — γMf

Hba0//Hba0 g Hba0//Hba0 g Hba0//Hba0 g Hba0//Hba0 g Hba0//Hba0 g

Hbb0//Hbb0 i Hbb0//Hbb0 i Hbb0//Hbb0 i Hbb0//Hbb0 j Hbb0//Hbb0 j

a See

ref. 10. ref. 5. c See ref. 66. d See ref. 21. e See ref. 67. f See ref. 23. g See ref. 69. h See ref. 68. i See ref. 70. jSee ref. 55. b See

Fabry et al.

Transgenic Mice and Hemoglobinopathies

221

are also present in these transgenic mice, validating their usefulness in the analysis of pathophysiological events induced by HbC in red cells. K:Cl cotransport (KCl) is elevated in human SS and CC disease. In the case of SS, at least part of this elevation is correlated with increased reticulocyte count since density fractions rich in reticulocytes have higher KCl activity. The case for HbC disease is less clear, since KCl in CC red cells is nearly as high as SS red cells, but retic counts are lower. KCl activity was studied in the mice described above. HbAKO mice (mice expressing exclusively HbA) have a small volume-stimulated KCl, which is stimulated by NO3– as reported in previous studies of early transgenic lines such as NYIDD, transgenic mice expressing 56% human α, 75% βS, and residual mouse globins that are insensitive to DIOA and stimulated by NO3–. In KO mice expressing exclusively human HbC or HbS+γ, the results are very different: KCl had a strong volume-stimulated component for C and S+γ mouse red cells vs sulfamate that was partially inhibited by NO3– (48 vs 37% for C and S+γ mouse red cells, respectively) and also by DIOA. A similar and even larger effect was observed for KCl activity measured at pH 7.0, which was more pronounced in the HbC mice. To eliminate the contribution of elevated reticulocyte, K:Cl cotransport in founder mice expressing 56% human α, 33% βC, and residual mouse globins that have low (3–5%) reticulocyte counts was studied. A strong volume dependence and sensitivity to NO3– (34%) and DIOA (30%) in these mice was found. Therefore βC interacts differently and more strongly with the transporter and/or its regulators than does βS.

1.6. Use of Transgenic Mice in Study of Innate Resistance to Malaria Shear (28) has reviewed the use of transgenic mice in the study of innate resistance to malaria. More recently, γ-expressing transgenic mice have been used to test the hypothesis based on in vitro studies (29) that the growth of Plasmodium falciparum in cells containing fetal hemoglobin (HbF = α2γ2) is retarded. Transgenic (γ) mice expressing human Aγ and Gγ chains resulting in 40–60% αM2γ2 Hb were infected with rodent malaria. Three species of rodent malaria were studied: P. chabaudi adami, which causes a nonlethal infection; P. yoelii 17XNL, which causes a nonlethal infection; and P. yoelii 17XL, a lethal variant of P. yoelii 17XNL that causes death in mice. Data indicate that this strain may cause a syndrome resembling cerebral malaria caused by P. falciparum (30). Results suggest that HbF does indeed have a protective effect in vivo, which is not mediated by the spleen (31). In terms of mechanisms, light microscopy showed that intraerythrocytic parasites develop slowly in HbF erythrocytes, and electron microscopy showed that hemozoin formation was defective in these transgenic mice. Finally, digestion studies of HbF by

222

Fabry et al.

recombinant plasmepsin II demonstrated that HbF is digested only half as well as HbA. Hence, HbF provides protection from P. falciparum malaria by the retardation of parasitic growth. The mechanism involves resistance to digestion by malarial hemoglobinases based on the data presented and with the wellknown properties of HbF as a superstable Hb. These studies also established HbF as the mechanism for the strong innate resistance to malaria by neonates. 2. Generation of Transgenic Mice Expressing Human Globins The human β-like globin genes have strict developmental and tissue-specific regulation. The tissue specificity is conferred by regulatory elements proximal to the genes, and the developmental regulation is conferred in part by proximal cis-acting elements (32) but depends, in addition, on the combination of genes present in the construct and on the distance of the genes from the LCR. For instance, the β-globin gene is more correctly developmentally regulated in the presence of a γ-globin gene in the construct than in its absence (33,34). The LCR, a group of hypersensitive sites (HS) located 50–15 kb upstream of the genes, regulates the level of expression of all the globin genes but has a minimal impact on developmental and tissue specificity. α-Globin genes are also developmentally regulated.

2.1. Overall Construct Design The design of constructs for the creation of transgenic mice expressing human globin depends greatly on the purpose of the experiments. When the experimental goal is to produce mutant globin proteins, simple constructs containing a segment of the LCR attached to a short genomic fragment containing the globin genes and their proximal regulatory sequences are sufficient. When the experimental goal is to study the cis-acting regulatory elements of the locus, or to study the pharmacological induction of the γ-globin genes, larger constructs containing the intact locus are preferable because the relative arrangement of the different genes and their regulatory elements (including the LCR) have been shown to be critical.

2.1.1. Short Constructs Production of transgenic mice by injection of DNA constructs into the male pronucleus of fertilized mouse ovocytes leads to the integration of multiple copies of the transgenes at random integration sites in the mouse genome (usually one site per founder). Multiple copies of short genomic fragments containing the globin genes and their proximal regulatory sequences integrated at random sites in the mouse genome are expressed at variable but usually very low levels because of position effects caused by the influence of the site of integration and the presence of multiple copies. Typical genomic fragments

Transgenic Mice and Hemoglobinopathies

223

used for these experiments include the gene itself plus 1–3 kb of sequence upstream and downstream. cDNA fragments are very poorly expressed. Most investigators therefore have almost always used entire intact genomic genes as transgenes. Inclusion of the LCR dramatically improves the level of expression (35). In the presence of LCR fragments, the level of expression per integrated copy, measured at the RNA level, can be as high as that of the endogenous globin genes.

2.1.2. LCR Derivatives Numerous LCR derivatives that can confer high-level expression to globin transgene have been produced (36–40). These LCR derivatives are assembled from genomic fragments containing one or more of the hyper-sensitives (HS) sites making up the LCR plus a variable amount of sequence flanking the HS sites. Usually, the larger the amount of flanking sequence the stronger the LCR. Spacing of the various HS sites within the constructs has also been reported to be important for expression. Because a weak LCR derivatives can be compensated for by a higher number of integrated copies, most LCR derivatives can provide adequate enhancing activities to achieve high-level expression in some founders. However, if the size of the construct is not a concern, a larger LCR derivative is preferable (i.e., miniLAR). The human LCR is usually used for creation of transgenic mice expressing abnormal Hbs. Other LCRs (i.e., mouse or rabbit) would probably be adequate too.

2.2. Controlling Expression Even in the presence of an LCR, the level of expression of the globin transgenes cannot be predicted because the number of integrated copies cannot be controlled. If a specific level of expression is sought, several founders can be screened until one with the requisite level of expression is identified. An alternate strategy to vary the level of expression of a globin transgene is to include one Lox site in the construct. Breeding mice expressing a Lox site with mice expressing the cre recombinase in the germ line will lead to reduction (complete or partial) of the arrays of integrated copies and therefore vary the level of expression (41) (see Notes 4 and 5).

2.3. Position Effects In the presence of LCR fragments, position effects are attenuated but are not eliminated (42–44). Position effects can be stable (every red cell expresses the transgene but at a lower level than the endogenous genes) or variegating (the transgene is silenced in a fraction of the red cells). If expression of the transgene in all red cells is important for the experiment, expression of the transgene

224

Fabry et al.

should be tested by a method that allows detection of the transgenic globin in single cells (FACS, in situ hybridization, reverse transcriptase polymerase chain reaction) (see Note 6).

2.4. Balanced α- and β-Globin Expression Unbalanced expression of β-globin chains is toxic because of precipitation of the globin chains and release of heme and iron leading to oxidative damage. To obtain adult mice with a high level of circulating transgenic globin, it is critical to express α- and β-like globin genes at approximately equal levels and at the same time. Constructs for the α-like globin genes can be based on the same model as those for the β-like genes since LCR derivatives activate the α-globin gene as effectively as the β-globin gene (45). For the purpose of balancing β-globin expression, a construct containing an α-globin genomic fragment plus a couple of kilobases of flanking sequences on either side and driven by an LCR derivative is perfectly adequate (see Note 7).

2.5. Use of Embryonic Stem Cells for Modification of Endogenous Locus An alternative method of producing mutant globin chains in mice is to modify the endogenous globin genes via homologous recombination in embryonic stem cells. The Ley laboratory (44) has recently reported the creation of a mouse that expresses a mutant globin (β6I) in which the sixth codon of the mouse β-major chain has been mutated in an attempt to mimic the human βS-chain that is responsible for sickle disease. Although technically more delicate than microinjection in the pronucleus, this approach could be generalized. In theory, the coding sequences of one or more endogenous globin genes could be substituted by the coding sequence of any other globin. Such an approach would probably facilitate the precise control of the level of transgene expression (see Note 8).

2.6. Lentiviral Vectors Expressing Globin Genes Moloney-based retroviruses containing globin transgenes were produced more than ten years ago, but could not be used for the production of transgenic mice because of poor expression owing to silencing, either early in development if the virus was used to infect blastocysts (46) or during differentiation if bone marrow stem cell infection followed by transplantation in irradiated recipient was the chosen gene transfer methodology (47). Difficulties in infecting quiescent multipotential stem cells with retroviral vectors compounded the problems associated with the latter approach. However, recent advances in viral gene transfer could revolutionize these approaches. Lois et al. (48) have recently reported that self-inactivating, VSV-Gpseudotyped lentiviral vectors appear to escape developmental silencing. It should therefore be possible to produce transgenic mice with mutant globins

Transgenic Mice and Hemoglobinopathies

225

by infecting the blastocyst with virus containing a globin gene and an LCR derivative. Blastocyst or ovocyte infection requires less equipment and is easier to perform than microinjection. In addition, controlling the number of copies of the integrated transgenes should be easier than with microinjection, since lentiviral vectors always integrate as single copies. Production of transgenic mice by infection of hematopoietic stem cells followed by bone marrow transplantation has also been revolutionized by the use of VSV-G-pseudotyped lentiviral vectors since these vectors can infect quiescent murine hematopoietic stem cells at high efficiency. Using such an approach, two teams have cured mice of β-thalassemia and sickle cell disease (49,50). Although these vectors were produced for gene therapy purposes, they could be used to generate mice producing any globin transgene. While technically quite demanding, this approach could allow laboratories with the requisite expertise in bone marrow transplantation and lentiviral vector production to produce several vectors to test various combinations of α- and β-globin antisickling transgenes or to test globin chains that might be too toxic to allow survival during the fetal or perinatal periods. Of course, transgenic mice produced by bone marrow transplantation will not transmit their transgenes through their germ lines.

2.7. Large Constructs The main advantage of large constructs (yeast artificial chromosomes [YACs] or bacterial artificial chromosomes [BACs]) is that distance between the genes themselves, and between the genes and the regulatory elements, is normal. When large constructs are used, expression at the mRNA level of the human globin genes is often equivalent to that of their mouse counterparts (51–53) although position effects are clearly not eliminated (see Subheading 2.3.). However, at the protein level, on a per integrated-copy basis, the level of expression is much lower than that of the endogenous genes, probably because of poor translation (54). When human YACs are introduced into the mouse genome, the developmental regulation of the human adult globin genes is recapitulated. Silencing of the γ-globin gene in adult life is also recapitulated. However, expression of the human globin genes in mouse early embryos and fetal liver differs from expression of these genes in human cells since the γ-globin genes are expressed very early during development and the ε gene is expressed after the γ genes (or at the same time-depending on reports [44,51–53] rather than prior to the γ genes.

2.8. KO Mice Mice generated by homologous recombination with deletion of their adult α- or β-globin genes are available (20,54–57). Mice with deletion of Ey or bh1

226

Fabry et al.

are also available (Steve Fiering, personal communication). Mice with deletions of various size in the globin clusters that were induced by irradiation are also available from Jackson Laboratory. The most commonly available of these is the βmajor-deletion described by Skow et al. (68). 3. Breeding Mice with Severe Sickle Phenotypes In general, mice with severe phenotypes cannot be bred by mating mice homozygous for the transgene and also homozygous for the KOs or deletions necessary for enhancing Hb expression to each other (see Note 9). Since the breeders are not homozygous, the pups will have a variety of genotypes, and testing them to determine Hb expression and genotype is a necessary part of any breeding program. The first step in successful breeding is accurately characterizing the mice that you have, and the second step is characterizing the offspring. In this section, we discuss determining Hb composition and genotype and breeding strategies for mice with severe phenotypes.

3.1. Determining Hemoglobin Composition and Genotype 3.1.1. Validating the Mutation In many cases, mice with mutant Hbs will be obtained from commercial laboratories or originating laboratories that have already established the identity of the Hb. However, in newly created lines of transgenic mice, the identity of the globin chains should be confirmed. In the case of known common human Hbs and mutants, authentic samples may be available for testing by addition of a known sample to the test sample followed by isoelectric focusing (IEF) or high-performance liquid chromatography (HPLC). A more general technique is mass spectroscopy (MS). MS can determine the molecular weights of individual globin chains to within ±one atomic mass unit. This level of accuracy allows detection of posttranslational modification and verification of mutant and recombinant Hbs, and, in contrast to electrophoretic techniques, MS can separate similarly charged samples. Individual bands or peaks from IEF or HPLC can also serve as the starting sample. Sample requirements are very small—a few picomoles of protein. MS has also been used for complete sequencing of proteins based on identification of fragments. This approach is particularly useful for mutant Hbs since sequence homology allows good working approximations to be formulated. Fragments can be produced classically by proteolytic digestion and then exposed to matrix-assisted laser desorption ionization MS. The abnormal fragments can be identified and, in many cases, the mutation deduced. With a few exceptions, whole blood hemolysates can be subjected to digestion without further separation. Separation is necessary when the variant is present at a low

Transgenic Mice and Hemoglobinopathies

227

level, a mass difference of 1 Dalton is suspected, or high levels of HbF are present. In these cases, the sample can be purified by IEF or HPLC prior to digestion (58). Alternatively, the protein can be fragmented in the spectrometer itself without resorting to wet chemistry, and the fragments produced in the spectrometer can be used to identify the portion bearing the mutant amino acid (59–61).

3.1.2. Initial Screening 3.1.2.1. STORING AND SHIPPING RED CELLS AND HB

Murine red cells are more fragile than human red cells, and therefore greater care needs to be exercised in both storage and shipping. The best way to store or ship cells is in their own plasma at 4°C or on wet ice with the cells separated from direct contact with the ice. Murine Hbs are also less stable than human Hbs. The best way to store murine Hb for 1 or 2 d is in red cells in autologous plasma; the cells can then be washed and hemolyzed as needed. If the cells have already been hemolyzed, the hemolysate should be stored in a low-temperature freezer, –135°C, or liquid nitrogen. If neither is available, a –80°C freezer is the next best choice. Do not leave hemolysate or red cells at room temperature. Do not freeze hemolysates in a refrigerator freezer. A timedependent degradation occurs under poor storage conditions which may generate shifted or even additional bands or peaks. 3.1.2.2. POLYMERASE CHAIN REACTION

PCR is useful for identifying the first-founder mice; however, since the work on transgenic mice expressing human/mutant Hbs is focused on the effects of the Hb itself, direct determination of mutant Hb levels is a necessary part of identifying the mice. In addition, the correlation between protein expression and either the presence of the gene may depend on multiple factors (Notes 4, 6, 16, and 17). 3.1.2.3. ELECTROPHORESIS

At pH 8.6, all human Hbs have a net negative charge and migrate in an electric field toward the positive pole or anode. Separation of Hbs by electrophoresis is based on the relative charge of the αβ dimer; therefore, mutations that do not alter the charge may be “silent” and not detectable by electrophoresis. Most electrophoretic methods separate Hb tetramers, but only tetramers composed of two identical αβ dimers, or homotetramers, are seen at the end of the separation process (see Notes 10 and 11). Cellulose acetate electrophoresis at pH 8.4 is a standard method and kits are inexpensive and easy to use. Charts accompanying these kits can be used to

228

Table 2 Characteristics of Selected Sickle Transgenic Micea βS (%)

228

γ (%) (>9 wk)

MCH MCHC (pg/cell) (g/dL)c

Reticsulocytes (% Sysmex)

48

3147a 1688b

32.3

2820a

3.6

44

3147a

b

3.3

c

—

14.1

35.7

3.2

47.0

—

14.3

36.2

11.1

44.5

2821a 1302a 1350b

59

10

14.5

35

7.7

45.8

2810a

100

<3

14.2

20.8

63.2

22.4

—

—

—

—

14.5

33

—

—

—

12.7

24.9

24.6

βSAD 21

52

—

15.1

37.3

βSAD NA

NA

5.2

NA

75 βS 42 βS-Ant 38 βS 36 βS-Ant 34

56

—

58

97

Urine Hemaconc. tocrit (mOsm)

2.2

Fabry et al.

C57BL CONTROL C57BL Hbbth-1//Hbbth-1 THAL α2 βSAD SAD α2 βSAD SAD γ α2βS Hbbth-1//Hbbth-1 NY1DD α2βS α2βS-Antilles Hbbth-1//Hbbth-1 S+S-Antilles α2βS α2βS-Antilles γ(G100)Hbbth-1//Hbbth-1 S+S-Antilles γH α2βSγ(F1352) Hba0//Hba0 Hbb0//Hbb0 NY1KO γ L α2βSγ(G203) Hba0//Hba0 Hbb0//Hbb0

αH (%)

a Twenty-four-hour b Eight-hour

water deprivation water deprivation.

Transgenic Mice and Hemoglobinopathies

229

NY1KO γM 80 α2βSγ(G100) Hba0//Hba0 Hbb0//Hbb0 NY1KO γH 60 α1GγAγδβS Hba0//Hba0 Hbb0//Hbb0 BERK >99 α1GγAγδβSγ(G203) Hba0//Hba0 Hbb0//Hbb0 BERK γM 79

100

20

13.7

24

30.1

34.0

1176b

100

40

14.4

31

12.9

41.1

3285a

100

<1

9.3

18.5

36.5

28.7

898b

100

21

10.8

22.9

37.2

41.6

1710a

229

230

Fabry et al.

identify common Hbs. A set of reference Hbs, which could be HbA, HbS, and HbC and a relevant mouse Hb, should be included with each run. It is useful to mix known standards with the test sample when the presence of mutants with electrophoretic properties similar to normal and common variants is suspected. Electrophoresis can also be carried out on separated globin chains if denaturing conditions are used (62,63). In these cases, urea is added to the electrophoresis buffer and the diluent in which the Hb is dissolved. Another approach that sometimes allows separation of overlapping Hbs, one of which contains cysteine, is treatment of the sample with cystamine (64). This simplifies analysis for samples with multiple α- and β-globin chains; however, at this point it is reasonable to turn to a higher-resolution technique such as IEF, HPLC, or MS. 3.1.2.4. ISOELECTRIC FOCUSING

IEF is capable of much higher resolution than cellulose acetate electrophoresis. Proteins and amino acids all have a pH at which the net charge is zero that is called the isoelectric point, or pI. At this pH there is no net movement in the presence of an externally applied electric field. To separate proteins based on their isoelectric points, a stable pH gradient needs to be created. This is achieved by applying a set of ampholites with pIs that cover the range of pIs of the proteins that are to be separated on a support matrix. During the initial period after the current is applied, both the ampholites and the proteins to be separated move as the pH gradient is formed. If a protein molecule finds itself on the acidic side of its pI it will migrate to the cathode, and if it finds itself on the basic side of its pI it will migrate toward the anode (hence the term isoelectric focusing). Sharp bands of individual proteins are thus formed. If focusing is continued, the pH gradient is eventually degraded and the protein bands begin to spread. A major factor that degrades the pH gradient is the effect of heat; therefore, efficient cooling is a crucial aspect of all IEF systems (see Notes 12 and 13).

3.1.3. More precise Determination HPLC HPLC, and in particular denaturing HPLC, gives both quantitative and qualitative information of mouse globins. The combination of screening with IEF and further characterizing the sample with HPLC can eliminate most ambiguities in both overlapping bands and identification of partial KOs and deletions. The equipment used for HPLC is much more expensive, sophisticated, and difficult to maintain than that used for electrophoresis or IEF, but it is still well within the reach of individual research laboratories. In HPLC, ionic and hydrophobic interactions of the sample with the supporting matrix are the basis of separation. The sample is applied as a thin layer to the top of the column under conditions wherein it interacts strongly with the matrix. The proteins are then

Transgenic Mice and Hemoglobinopathies

231

eluted with a developing solution (buffer) of gradually increasing strength, until all of the proteins are eluted. In cation- and anion-exchange chromatography, the properties of the developing solution that are varied are pH and ionic strength (salt concentration) and in reverse-phase chromatography, the hydrophobicity (organic solvent) content is also varied. Hb may be separated as the intact tetramer or, under denaturing conditions, the individual globin chains can be separated (see Notes 14–16). Human Hbs usually have a relatively small number of possible homotetramers, and analysis of the intact tetramer generally yields readily interpretable results. However, Hb from transgenic mice, because of the possible formation of human-mouse chimeric αβ-dimers and hence a wide variety of tetrameric forms, yields a more readily interpretable chromatogram when denaturing conditions are used and the isolated α- and β-chains are detected. Very small internal diameters increase column resolution and decrease the amount of sample required at the cost of increased time per sample (65) (see Note 17).

3.1.4. Refining the Screen Complete blood counts are sometimes useful for detection of deletions or KOs. The red cells may have lower MCH or MCV. To make use of this approach, one needs to establish reference values for mice of known genotype (see Note 18).

3.2. Breeding Strategies for Mice with Severe Phenotypes 3.2.1. Choice of Breeders As stated earlier, mice with severe phenotypes usually cannot be bred by mating mice homozygous for the transgene and also homozygous for the KOs or deletions necessary for enhancing Hb expression to each other. In general, the best strategy is a male of the desired genotype and a healthy female of a related genotype. Use of Punnett squares for hybrid and dihybrid crosses to calculate expected ratios of offspring is desirable because some crosses that look possible may have very small predicted yields. Punnett squares can be used to choose the pairs yielding the highest predicted percentage of desired mice and, when possible, regenerate desirable breeders (see Notes 19–22). In the breeding schemes described, the best possible outcome is 50% of the pups born with the desired genotype, and a more common predicted outcome is that between 12.5 and 25% of the pups born will have the desired genotype. Even in models with moderate severity, such as the S+S-Antilles mouse, many fewer pups are born and survive to 10 d than would be predicted. From these observations, it is clear that anticipated production will be low at best and that misidentification of breeding mice can cut productivity even further. The need for efficient screening and rigorous identification of all breeders should be clear.

232

Fabry et al.

3.2.2. Epistatic Efects, Polymorphisms, Founder Effects and the Need to Backcross Founder effects (loss of genetic variation when a new colony is formed by a very small number of individuals from a larger population) are present in all transgenic lines because generation of transgenic mice usually occurs on mixed backgrounds of several inbred lines, but they are frequently diluted in the initial phase of expanding a newly created transgenic line and in breeding in the KOs or deletions needed to enhance expression of the new transgenes. However, founder effects may reappear (with a different set of polymorphisms) when one successful male fathers most of the mice of a given genotype in a colony. This scenario is probable when severe mice are bred. Very often the first few (10–20) mice with a given severe genotype are more severe than their third- and fourth-generation descendants. This is an expected outcome since founders are frequently the combination of two or more inbred lines and the usual breeding strategy is to choose the most successful males and females for further breeding. This strategy, while necessary for efficient propagation, can rapidly screen out deleterious genes that impact the pathology of the selected Hb and enhance the frequency of ameliorating genes. Frequently a single male may father the majority of pups in a particular line of mice. These mice will generally have a less severe phenotype than their predecessors that are identical in the selected genes but may differ substantially in genes contributing to epistatic effects. One way to exert partial control over these effects is to backcross onto an inbred background.

3.3. Record Keeping Computerized record keeping is the only way to maintain accurate records once a mouse colony begins to grow. The best choice of databases are relational databases, which allow a great deal of flexibility in setting up and subsequently modifying tables and the ability to query one or more tables in a very sophisticated way. Several popular relational databases are available. Data can also be transferred to a personal digital assistant for portable access to detailed information about individual mice while in the animal colony. 4. Notes 1. Epistasis is the interaction between genes at two or more loci, such that the phenotype differs from that that would be expected if the loci were expressed independently. 2. Polymorphism is a condition in which a population possesses more than one allele at a locus. 3. Examples of epistatic effects in human sickle cell disease are the effect of α-thalassemia and genetically determined levels of HbF expression.

Transgenic Mice and Hemoglobinopathies

233

4. The level of globin chains in mouse red cells is affected by differential assembly of the globin tetramers and by the fact that human globin mRNA is poorly translated when compared with the endogenous mouse globin (71). Ratios of β-globin mRNA to protein in transgenic mice ranging from 5 to 10 have been reported. As a consequence, mRNA measurements are poor predictors of the mutant globin in peripheral blood. 5. Some transgenes can be bred to homozygosity, thereby doubling the level of expression. However, many transgenes cannot be bred to homozygosity either because their insertion in the genome led to a recessive lethal mutation or because highlevel expression of the protein is not compatible with either fetal or adult life. 6. Both stable and variegating position effects can be age dependent, with expression of the transgene decreasing as the animal ages. It is therefore important to measure expression levels at the age at which the experiment will be performed 7. There are two ways to obtain expression of two genes at the same time. The two genes can be placed in the same construct and either share the same LCR derivative or be controlled by two different LCRs. Alternatively, two independent constructs each with its own LCR can be coinjected into the fertilized eggs. Such coinjection generally results in cointegration of the two transgenes at the same site of integration, although the proportion of each globin chain produced will vary in each founder. 8. Recently, methods to perform site-specific integration such as recombinasemediated cassette exchange (RMCE) (72,73) have been developed. RMCE may be combined with homologous recombination and could provide a powerful way to produce a mouse model expressing reproducible and predictable levels of transgenic globins. 9. When an Hb is introduced into a transgenic mouse, it must compete with the murine Hb. Higher levels of expression can be attained by introducing deletions, such as the β major deletion (68), or KOs of the murine α- and β-globins (22,55,69,70). In some cases, mice with exclusively human Hb can be readily produced (mice with HbA or HbC); however, in other cases (mice with HbS), the phenotype is severe and production of pups with the desired genotype occurs at a much lower frequency than would be predicted. 10. Hb is a tetramer composed of two αβ dimers. The αβ dimers are formed at the time of Hb synthesis and are stable under all physiological conditions; denaturing or near-denaturing conditions, such as low pH, are required to break up αβ dimers. The dimers exchange readily under solution conditions (in the presence of oxygen), but only homodimers are seen at the end of most techniques used to separate Hbs. 11. In mice expressing both murine and human Hbs, in addition to αβ dimers that are composed of αmouseβmouse or αhumanβhuman, there are αhumanβmouse and αmouseβhuman dimers. Some inbred lines have multiple α or β globins, further confusing the picture, especially when some of the bands overlap. The best way to deal with multiple and overlapping bands is to isolate the bands (most easily done from an IEF plate) and run them on a denaturing HPLC to identify the globins present.

234

Fabry et al.

12. Human globins usually focus much faster than mouse globins. Mouse globins may require run times nearly twice as long as human globins. Heating and distortion owing to contamination of the sample or the gel with the anodic or cathodic buffer solutions are common sources of error. In long runs, it may be necessary to stop the run and remove accumulated moisture from the top of the apparatus. 13. When evaluating an unfamiliar Hb that may contain a mixture of both mouse and human hemoglobins, it is useful to isolate the bands on the IEF (usually by applying the same sample to several slots in the IEF gel, combining the bands at the same level, and extracting them with buffer) and then characterize these bands on the HPLC. 14. Resolution of commercially supplied columns may depend on the characteristics of the individual column and may require adjustment of the developing buffer to cope with variation. Variations among columns from the same manufacturer with the same specification may be large. For example, the composition of the starting buffer may need to be varied by 10% or more to achieve the same elution profile. In addition, peaks that are partially separated on one column may overlap completely on another. Manufacturers frequently allow testing and return of individual columns if they fail to meet the user’s requirements. 15. Protect columns and extend column life by centrifuging and filtering solutions prior to application and by using a guard column. 16. There are many factors that may affect expression of Hb transgenes in mice. Those occurring at or prior to translation into protein such as, insertion point, copy number, silencing, translation, and transcription were described in Subheading 2. Those occurring posttranslation such as instability, competition for insufficient α- or β-chains, and ineffective assembly are also factors that can affect protein expression and interfere with producing healthy mice. Unanticipated Hb instability can occur when mutations that do not occur in nature are introduced. Most of the known mutations of human Hb and their effect on stability, p50, and many other factors are listed in the Globin Gene Server (http://globin.cse.psu.edu). Imbalance between α- and β-chains results in thalassemia, and balanced production of chains is largely a matter of luck using the most common current methods for producing transgenic mice (see Subheading 2.) for alternatives. When there are an insufficient number of either α- or β-chains and two or more types of the complementary chains present, some chains may be more successful in the competition, leading to a smaller percentage of the poor competitor than would be predicted from chain synthesis or other types of measurement. 17. The percentage of individual globins in peripheral blood may vary with age and either decrease or increase. Silencing, as described in Subheading 2., usually leads to a decrease in expression. In sickle transgenic mice that express HbF in F-cells (i.e., some cells have HbF and some do not), the F-cells may be enriched. That is, cells with HbF may survive longer in the circulation and the percentage of HbF may increase with time (23). As a consequence Hb expression should always be measured at the time of the experiment. 18. Commercial instruments for determination of CBCs usually have a means of bypassing automatic dilution (which requires a large sample) and accept prediluted samples

Transgenic Mice and Hemoglobinopathies

235

with hematocrits between 8 and 12. A good minimum volume is 300 µL of hematocrit 10, which will allow repeat measurements, although smaller samples can be used. Prediluting the sample allows smaller, mouse-sized samples to be used. It is important to remember that the cells will ultimately be suspended in a solution with an osmolarity suitable for human red cells (usually 290–300 mOsm) which is lower than that of mouse plasma (about 330 mOsm). This will result in a systematically higher MCV and hematocrit and lower MCHC than would be measured by hand. 19. A simple Punnett square for breeding two mice hemizygous for the NY1 transgene will be represented as an “S” and the absence of the transgene will be represented by a “–”: –

S

–

––

–S

S

–S

SS lethal

This square predicts that of the mice conceived, 25% will have no transgene, 50% will be hemizygous for the transgene, and 25% will be homozygous for the transgene that is lethal. A Punnett square for breeding a male S+S-Antilles mouse that is hemizygous for the NY1 transgene or “S,” hemizygous for the S-Antilles transgene or “A,” and the absence of a transgene will again be represented by “–” with a female mouse that is hemizygous for the NY1 transgene: ––

––

–S

–S

––

––––

––––

–S– –

–S– –

A–

A– – –

A– – –

AS– –

AS– –

–S

–S

–S– –

–S–S lethal

–S–S lethal

AS

AS– –

AS– –

AS–S lethal

AS–S lethal

This square predicts that of the mice conceived, 12.5% will have no transgene, 12.5% will be hemizygous for the S-Antilles transgene, 25% will be hemizygous for the NY1 transgene “S,” 25% will be hemizygous for S+S-Antilles, 12.5% will be homozygous for NY1 that is lethal, and 12.5% will be homozygous for NY1 and hemizygous for S+Antilles that is lethal. 20. Example 1: To breed S+S-Antilles mice several strategies could be used (all of these strategies call for using mice homozygous for the βmajor deletion and all of the offspring will also be homozygous for the βmajor deletion): a. They could be bred directly to each other. This is undesirable for two reasons: S+S-Antilles mice are a moderately severe phenotype and the mothers some-

236

Fabry et al.

times die, have small litters, or fail to nurture their pups; and although the S-Antilles line can be homozygous for the transgene, the NY1 line cannot be homozygous for the transgene under any condition, and the combination of NY1 and homozygosity for S-Antilles is also lethal in the presence of homozygosity for the βmajor deletion. This breeding scheme would result in 25% of the desired offspring and 25% lethals of various forms. b. A homozygous S-Antilles female (also homozygous for the βmajor deletion) can be bred to an NY1DD male. This is the most efficient mating and results in 50% S+S-Antilles pups and 50% S-Antilles pups. c. Most other crosses, S+S-Antilles x NY1DD or homozygous S-Antilles x S+SAntilles, yield 25% S+S-Antilles, but the last cross also generates 25% homozygous S-Antilles, which regenerates the breeding stock. 21. Example 2: NY1KO-γL mice have a very short survival (about 40 d) and the females either fail to survive, produce small litters (one pup), or fail to nurture the pups born. The solution here is the same as that given for breeding S+S-Antilles mice. Use a male of the desired phenotype and a healthy female. A productive cross uses an NY1KO-γL male and an HbAKO-γLγL female. This cross yields 25% of the desired phenotype (NY1KO-γL) divided between mice that are hemiand homozygous for the γL transgene. The remaining mice are 25% lethal (owing to the absence of an α transgene), 25% HbAKO mice, and 25% HbAHbSKO mice, which can also be used as breeders. If only one of the mice is hemizygous for the γL transgene, then the yield of NY1KO-γL mice drops to 12.5%. 22. Example 3: BERK mice (21) are bred using a similar approach. Once again, female BERK mice are unsuccessful breeders. A male BERK full KO mouse (which expresses only human Hb) is mated to a female BERK mouse that is hemizygous for the βmajorβminor KO and may be either hemi- or homozygous for the BERK transgene. The BERK transgene is weakly expressed in mice that are hemizygous for the βmajorβminor KO and the percentage of βS of all β-chains reaches only 15 or 30% for the hemi- or homozygote for the BERK transgene, respectively. The low level of βS expression in mice hemizygous for the βmajorβminor KO is consistent with the observed thalassemia of these mice and the lack of severity in mice hemizygous for the βmajorβminor KO. Both types of female yield 25% of the desired BERK full KO and a mix of mice hemizygous for the βmajorβminor KO.

References 1. Rubin, E. M., Lu, R. H., Cooper, S., Mohandas, N., and Kan, Y. W. (1988) Introduction and expression of the human Bs-globin gene in transgenic mice. Am. J. Hum. Genet. 42, 585–591. 2. Rhoda, M. D., Domenget, C., Vidaud, M., Bardakdjian Michau, J., Rouyer Fessard, P., Rosa, J., and Beuzard, Y. (1988) Mouse α-chains inhibit polymerization of hemoglobin induced by human βS or βS-Antilles chains. Biochim. Biophys. Acta. 952, 208–212. 3. Greaves, D. R., Fraser, P., Vidal, M. A., Hedges, M. J., Ropers, D., Luzzatto, L., and Grosveld, F. (1990) A transgenic mouse model of sickle cell disorder [see comments]. Nature 343, 183–185.

Transgenic Mice and Hemoglobinopathies

237

4. Ryan, T. M., Townes, T. M., Reilly, M. P., Asakura, T., Palmiter, R. P., and Behringer, R. R. (1990) Human sickle hemoglobin in transgenic mice. Science 247, 566–568. 5. Trudel, M., Saadane, N., Garel M-C, Bardakdjuan-Michau, J., Blouquit, Y., Guerquin-Kern J-L, Rouyer-Fessard, P., Vidaud, D., Pachniss, A., Romeo P-H, Beuzard, Y., and Costantini, F. M. (1991) Towards a transgenic mouse model of sickle cell disease: hemoglobin SAD. EMBO J. 10, 3157–3168. 6. de Franceschi, L., Saadane, N., Trudel, M., Alper, S. L., Brugnara, C., and Beuzard, Y. (1994) Treatment with oral clotrimazole blocks Ca(2+)-activated K+ transport and reverses erythrocyte dehydration in transgenic SAD mice. A model for therapy of sickle cell disease. J. Clin. Invest. 93, 1670–1676. 7. de Franceschi, L., Villa-Moruzzi, E., Fumagalli, L., Brugnara, C., Turrini, F., Motta, R., Veghini, E., Corato, C., Alper, S. L., and Berton, G. (2001) K-Cl cotransport modulation by intracellular Mg in erythrocytes from mice bred for low and high Mg levels. Am. J. Physiol. Cell Physiol. 281, C1385–C1395. 8. Blouin, M. J., Beauchemin, H., Wright, A., De Paepe, M., Sorette, M., Bleau, A. M., Nakamoto, B., Ou, C. N., Stamatoyannopoulos, G., and Trudel, M. (2000) Genetic correction of sickle cell disease: insights using transgenic mouse models [in process citation]. Nat. Med. 6, 177–182. 9. Fabry, M. E., Costantini, F., Pachnis, A., Suzuka, S. M., Bank, N., Aynedjian, H. S., Factor, S. M., and Nagel, R. L. (1992) High expression of human βS- and α-globins in transgenic mice: erythrocyte abnormalities, organ damage, and the effect of hypoxia. Proc. Natl. Acad. Sci. USA 89, 12,155–12,159. 10. Fabry, M. E., Nagel, R. L., Pachnis, A., Suzuka, S. M., and Costantini, F. (1992) High expression of human βS- and α-globins in transgenic mice: hemoglobin composition and hematological consequences. Proc. Natl. Acad. Sci. USA 89, 12,150–12,154. 11. Lutty, G. A., McLeod, D. S., Pachnis, A., Costantini, F., Fabry, M. E., and Nagel, R. L. (1994) Retinal and choroidal neovascularization in a transgenic mouse model of sickle cell disease. Am. J. Path. 145, 490–497. 12. Hood, A. T., Fabry, M. E., Costantini, F., Nagel, R. L., and Shear, H. L. (1996) Protection from lethal malaria in transgenic mice expressing sickle hemoglobin. Blood 87, 1600–1603. 13. Kaul, D. K. and Hebbel, R. P. (2000) Hypoxia/reoxygenation causes inflammatory response in transgenic sickle mice but not in normal mice [see comments]. J. Clin. Invest. 106, 411–420. 14. Fabry, M. E., Sengupta, A., Suzuka, S. M., Costantini, F., Rubin, E. M., Hofrichter, J., Christoph, G. W., Manci, E. A., Culberson, D., Factor, S. M., and Nagel, R. L. (1995) A second generation transgenic mouse model expressing both hemoglobin S (HbS) and HbS-Antilles results in increased phenotypic severity. Blood 86, 2419–2428. 15. Kaul, D. K., Fabry, M. E., Costantini, F., Rubin, E. M., and Nagel, R. L. (1995) In vivo demonstration of red cell-endothelial interaction, sickling and altered microvascular response to oxygen in the sickle transgenic mouse. J. Clin. Invest. 96, 2845–2853.

238

Fabry et al.

16. Kaul, D. K., Liu, X. D., Fabry, M. E., and Nagel, R. L. (1998) Increased nitric oxide synthase expression in transgenic sickle mice is associated with relative hypotension and altered microvascular response. Blood 92, 330a. 17. Romero, J. R., Suzuka, S. M., Nagel, R. L., and Fabry, M. E. (2002) Arginine supplementation of sickle transgenic mice reduces red cell density and Gardos channel activity. Blood 99, 1103–1108. 18. Enwonwu, C. O. (1989) Increased metabolic demand for arginine in sickle cell anemia. Med. Sci. Res. 17, 997–998. 19. Fabry, M. E., Kennan, R. P., Paszty, C., Costantini, F., Rubin, E. M., Gore, J. C., and Nagel, R. L. (1996) Magnetic resonance evidence of hypoxia in a homozygous α-knockout of a transgenic mouse model for sickle cell disease. J. Clin. Invest. 98, 2450–2455. 20. Ryan, T. M., Ciavatta, D. J., and Townes, T. M. (1997) Knockout-transgenic mouse model of sickle cell disease [see comments]. Science 278, 873–876. 21. Paszty, C., Brion, C. M., Manci, E., Witkowska, H. E., Stevens, M. E., Mohandas, N., and Rubin, E. M. (1997) Transgenic knockout mice with exclusively human sickle hemoglobin and sickle cell disease [see comments]. Science 278, 876–878. 22. Chang, J. C., Lu, R., Lin, C., Xu, S. M., Kan, Y. W., Porcu, S., Carlson, E., Kitamura, M., Yang, S., Flebbe-Rehwaldt, L., and Gaensler, K. M. (1998) Transgenic knockout mice exclusively expressing human hemoglobin S after transfer of a 240-kb betas-globin yeast artificial chromosome: a mouse model of sickle cell anemia. Proc. Natl. Acad. Sci. USA 95, 14,886–14,890. 23. Fabry, M. E., Suzuka, S. M., Weinberg, R. S., Lawrence, C., Factor, S. M., Gilman, J. G., Costantini, F., and Nagel, R. L. (2001) Second generation knockout sickle mice: the effect of HbF. Blood 97, 410–418. 24. Paszty, C. (1997) Transgenic and gene knock-out mouse models of sickle cell anemia and the thalassemias [review]. Curr. Opin. Hematol. 4, 88–93. 25. Poillon, W. N., Kim, B. C., and Castro, O. (1998) Intracellular hemoglobin S polymerization and the clinical severity of sickle cell anemia. Blood 91, 1777–1783. 26. Fabry, M. E., Romero, J. R., Suzuka, S. M., Gilman, J. G., Feeling-Taylor, A., Odunusi, E., Factor, S. M., Bouhassira, E. E., Lawrence, C., and Nagel, R. L. (2000) Hemoglobin C in transgenic mice: effect of HbC expression from founders to full mouse globin knockouts [In Process Citation]. Blood Cells Mol. Dis. 26, 331–347. 27. Fabry, M. E., Kaul, D. K., Raventos Suarez, C., Chang, H., and Nagel, R. L. (1982) SC erythrocytes have an abnormally high intracellular hemoglobin concentration: pathophysiological consequences. J. Clin. Invest. 70, 1315–1319. 28. Shear, H. L. (1993) Transgenic and mutant animal models to study mechanisms of protection of red cell genetic defects against malaria. Experientia 49, 37–42. 29. Pasvol, G., Weatherall, D. J., and Wilson, R. J. (1977) Effects of foetal haemoglobin on susceptibility of red cells to Plasmodium falciparum. Nature 270, 171–173. 30. Kaul, D. K., Nagel, R. L., Llena, J. F., and Shear, H. L. (1994) Cerebral malaria in mice: demonstration of cytoadherence of infected red blood cells and microrheologic correlates. Am. J. Trop. Med. Hyg. 50, 512–521.

Transgenic Mice and Hemoglobinopathies

239

31. Shear, H. L., Grinberg, L., Gilman, J., Fabry, M. E., Stamatoyannopoulos, G., Goldberg, D. E., and Nagel, R. L. (1998) Transgenic mice expressing human fetal globin are protected from malaria by a novel mechanism. Blood 92, 2520–2526. 32. Starck, J., Sarkar, R., Romana, M., Bhargava, A., Scarpa, A. L., Tanaka, M., Chamberlain, J. W., Weissman, S. M., and Forget, B. G. (1994) Developmental regulation of human g. Blood 84, 1656–1665. 33. Peterson, K. R. and Stamatoyannopoulos, G. (1993) Role of gene order in developmental control of human g. Mol. Cell Biol. 13, 4836–4843. 34. Hanscombe, O., Whyatt, D., Fraser, P., Yannoutsos, N., Greaves, D., Dillon, N., and Grosveld, F. (1991) Importance of globin gene order for correct developmental expression. Genes Dev. 5, 1387–1394. 35. Grosveld, F., van Assendelft, G. B., Greaves, D. R., and Kollias, G. (1987) Position-independent, high-level expression of the human beta-globin gene in transgenic mice. Cell 51, 975–985. 36. Forrester, W. C., Novak, U., Gelinas, R., and Groudine, M. (1989) Molecular analysis of the human beta-globin locus activation region. Proc. Natl. Acad. Sci. USA 86, 5439–5443. 37. Sadelain, M., Wang, C. H., Antoniou, M., Grosveld, F., and Mulligan, R. C. (1995) Generation of a high-titer retroviral vector capable of expressing high levels of the human beta-globin gene. Proc. Natl. Acad. Sci. USA 92, 6728–6732. 38. Emery, D. W., Chen, H., L. i., Q., and Stamatoyannopoulos, G. (1998) Development of a condensed locus control region cassette and testing in retrovirus vectors for A gamma-globin. Blood Cells Mol. Dis. 24, 322–339. 39. Einerhand, M. P., Antoniou, M., Zolotukhin, S., Muzyczka, N., Berns, K. I., Grosveld, F., and Valerio, D. (1995) Regulated high-level human beta-globin gene expression in erythroid cells following recombinant adeno-associated virusmediated gene transfer. Gene Ther. 2, 336–343. 40. Hargrove, P. W., Vanin, E. F., Kurtzman, G. J., and Nienhuis, A. W. (1997) Highlevel globin gene expression mediated by a recombinant adeno-associated virus genome that contains the 3' gamma globin gene regulatory element and integrates as tandem copies in erythroid cells. Blood 89, 2167–2175. 41. Garrick, D., Fiering, S., Martin, D. I., and Whitelaw, E. (1998) Repeat-induced gene silencing in mammals. Nat. Genet. 18, 56–59. 42. Porcu, S., Kitamura, M., Witkowska, E., Zhang, Z., Mutero, A., Lin, C., Chang, J., and Gaensler, K. M. (1997) The human beta globin locus introduced by YAC transfer exhibits a specific and reproducible pattern of developmental regulation in transgenic mice. Blood 90, 4602–4609. 43. Alami, R., Greally, J. M., Tanimoto, K., Hwang, S., Feng, Y. Q., Engel, J. D., Fiering, S., and Bouhassira, E. E. (2000) Beta-globin YAC transgenes exhibit uniform expression levels but position effect variagation in mice. Hum. Mol. Genet. 9(4), 631–636. 44. Kaufman, R. M., Pham, C. T., and Ley, T. J. (1999) Transgenic analysis of a 100-kb human beta-globin cluster-containing DNA fragment propagated as a bacterial artificial chromosome. Blood 94, 3178–3184.

240

Fabry et al.

45. Ryan, T. M., Behringer, R. R., Townes, T. M., and Palmiter, R. D., Brinster, R. L. (1989) High-level erythroid expression of human alpha-globin genes in transgenic mice. Proc. Natl. Acad. Sci. USA 86, 37–41. 46. Jahner, D., Stuhlmann, H., Stewart, C. L., Harbers, K., Lohler, J., Simon, I., and Jaenisch, R. (1982) De novo methylation and expression of retroviral genomes during mouse embryogenesis. Nature 298, 623–628. 47. Rivella, S. and Sadelain, M. (1998) Genetic treatment of severe hemoglobinopathies: the combat against transgene variegation and transgene silencing. Semin. Hematol. 35, 112–125. 48. Lois, C., Hong, E. J., Pease, S., Brown, E. J., and Baltimore, D. (2002) Germline transmission and tissue-specific expression of transgenes delivered by lentiviral vectors. Science 295, 868–872. 49. May, C. and Sadelain, M. (2001) A promising genetic approach to the treatment of beta-thalassemia. Trends Cardiovasc. Med. 11, 276–280. 50. Pawliuk, R., Westerman, K. A., Fabry, M. E., Payen, E., Tighe, R., Bouhassira, E. E., Acharya, S. A., Ellis, J., London, I. M., Eaves, C. J., Humphries, R. K., Beuzard, Y., Nagel, R. L., and Leboulch, P. (2001) Correction of sickle cell disease in transgenic mouse models by gene therapy. Science 294, 2368–2371. 51. Peterson, K. R., Zitnik, G., Huxley, C., Lowrey, C. H., Gnirke, A., Leppig, K. A., Papayannopoulou, T., and Stamatoyannopoulos, G. (1993) Use of yeast artificial chromosomes (YACs) for studying control of gene expression: correct regulation of the genes of a human beta-globin locus YAC following transfer to mouse erythroleukemia cell lines. Proc. Natl. Acad. Sci. USA 90, 11,207–11,211. 52. Gaensler, K. M., Kitamura, M., and Kan, Y. W. (1993) Germ-line transmission and developmental regulation of a 150-kb yeast artificial chromosome containing the human beta-globin locus in transgenic mice. Proc. Natl. Acad. Sci. USA 90, 11,381–11,385. 53. Strouboulis, J., Dillon, N., and Grosveld, F. (1992) Developmental regulation of a complete 70-kb human beta-globin locus in transgenic mice. Genes Dev. 6, 1857–1864. 54. Chang, J., Lu, R. H., Xu, S. M., Meneses, J., Chan, K., Pedersen, R., and Kan, Y. W. (1996) Inactivation of mouse alpha-globin gene by homologous recombination: mouse model of hemoglobin H disease. Blood 88, 1846–1851. 55. Ciavatta, D. J., Ryan, T. M., Farmer, S. C., and Townes, T. M. (1995) Mouse model of human beta zero thalassemia: targeted deletion of the mouse beta majand beta min-globin genes in embryonic stem cells. Proc. Natl. Acad. Sci. USA 92, 9259–9263. 56. Yang, B., Kirby, S., Lewis, J., Detloff, P. J., Maeda, N., and Smithies, O. (1995) A mouse model for beta 0-thalassemia. Proc. Natl. Acad. Sci. USA 92, 11,608–11,612. 57. Shehee, W. R., Oliver, P., and Smithies, O. (1993) Lethal thalassemia after insertional disruption of the mouse major adult beta-globin gene. Proc. Natl. Acad. Sci. USA 90, 3177–3181. 58. Witkowska, H. E., Lubin, B. H., Beuzard, Y., et al. (1991) Sickle cell disease in a patient with sickle cell trait and compound heterozygosity for hemoglobin S and hemoglobin Quebec-Chori. N. Engl. J. Med. 325, 1150–1154. 59. Light-Wahl, K. J., Loo, J. A., Edmonds, C. G., Smith, R. D., Witkowska, H. E., Shackleton, C. H., and Wu, C. S. (1993) Collisionally activated dissociation and

Transgenic Mice and Hemoglobinopathies

60.

61.

62. 63.

64. 65. 66.

67.

68.

69.

70. 71.

72.

73.

241

tandem mass spectrometry of intact hemoglobin beta-chain variant proteins with electrospray ionization. Biol. Mass Spectrom. 22, 112–120. Bakhtiar, R., W. u., Q., Hofstadler, S. A., and Smith, R. D. (1994) Charge state specific facile gas-phase cleavage of Asp 75-Met 76 peptide bond in the alphachain of human apohemoglobin probed by electrospray ionization mass spectrometry. Biol. Mass Spectrom. 23, 707–710. Ferranti, P., Carbone, V., Sannolo, N., Fiume, I., and Malorni, A. (1993) Mass spectrometric analysis of rat hemoglobin by FAB-overlapping: primary structure of the alpha-major and of four beta constitutive chains. Intl. J. Biochem. 25, 1943–1950. Ueda, S. and Schneider, R. G. (1969) Rapid differentiation of polypeptide chains of hemoglobin by cellulose acetate electrophoresis of hemolylsates. Blood 34, 230–235. Schneider, R. G., Hosty, T. S., Tomlin, G., and Atkins, R. (1974) Identification of hemoglobins and hemoglobinopathies by electrophoresis on cellulose acetate plates impregnated with citrate agar. Clin. Chem. 20, 74–77. Whitney, J. B. (1978) Simplified typing of mouse hemoglobin (Hbb) phenotypes using cystamine. Biochem. Genet. 16, 667–672. Huisman, T. H. (1987) Separation of hemoglobins and hemoglobin chains by highperformance liquid chromatography [review]. J. Chromatog. 418, 277–304. Rubin, E. M., Witkowska, H. E., Spangler, E., Curtin, P., Lubin, B. H., Mohandas, N., and Clift, S. M. (1991) Hypoxia-induced in vivo sickling of transgenic mouse red cells. J. Clin. Invest. 87, 639–647. Arcasoy, M. O., Romana, M., Fabry, M. E., Skarpidi, E., Nagel, R. L., and Forget, B. G. (1997) High levels of human gamma-globin gene expression in adult mice carrying a transgene of deletion-type hereditary persistence of fetal hemoglobin. Mol. Cell. Biol. 17, 2076–2089. Skow, L. C., Burkhart, B. A., Johnson, F. M., Popp, R. A., Popp, D. M., Goldberg, S. Z., Anderson, W. F., Barnett, L. B., and Lewis, S. E. (1983) A mouse model for beta-thalassemia. Cell 34, 1043–1052. Paszty, C., Mohandas, N., Stevens, M. E., Loring, J. F., Liebhaber, S. A., Brion, C. M., and Rubin, E. M. (1995) Lethal alpha-thalassaemia created by gene targeting in mice and its genetic rescue. Nat. Genet. 11, 33–39. Yang, B., Kirby, S., Lewis, J., Detloff, P. J., Maeda, N., and Smithies, O. (1995) A mouse model for β°-thalassemia. Proc. Natl. Acad. Sci. USA 92, 11,608–11,612. Alami, R., Gilman, J. G., Feng, Y. Q., Marmorato, A., Rochlin, I., Suzuka, S. M., Fabry, M. E., Nagel, R. L., and Bouhassira, E. E. (1999) Anti-beta s-ribozyme reduces beta s mRNA levels in transgenic mice: potential application to the gene therapy of sickle cell anemia. Blood Cells Mol. Dis. 25, 110–119. Bouhassira, E. E., Westerman, K., and Leboulch, P. (1997) Transcriptional behavior of LCR enhancer elements integrated at the same chromosomal locus by recombinase-mediated cassette exchange. Blood 90, 3332–3344. Feng, Y. Q., Seibler, J., Alami, R., Eisen, A., Westerman, K. A., Leboulch, P., Fiering, S., and Bouhassira, E. E. (1999) Site-specific chromosomal integration in mammalian cells: highly efficient CRE recombinase-mediated cassette exchange. J. Mol. Biol. 292, 779–785.

242

Fabry et al.

Single Globin-Chain Expression and Purification

243

14 Recombinant Single Globin-Chain Expression and Purification Kazuhiko Adachi 1. Introduction The development of molecular biological techniques to selectively replace individual amino acids has furthered our understanding of the relationship between the structure and function of hemoglobin (Hb). Initial reports described production of normal and modified human globin chains in bacteria employing a fusion-protein expression vector (1). An expression system was later developed in which α- and β-globin chains were coexpressed, resulting in the formation of soluble tetrameric Hb in yeast (2,3). Coexpression of human α and β globin in Escherichia coli that resulted in the formation of soluble tetrameric Hbs was described (4), as well as a system for high expression of insoluble β-globin chains in E. coli. (5). These last two systems, however, result in globin chains containing an N-terminal methionine that may affect the functional properties of Hb. In 1993, Shen et al. (6) expressed soluble Hb tetramers lacking the N-terminal methionine in bacteria by coexpression of α- and β-globin cDNAs with methionine aminopeptidase (MAP) cDNA. To further understand Hb assembly and folding, production of soluble single-chain Hb variants is critical; however, expression of recombinant, soluble individual globin chains has not been realized to date (4,5,7,8). Globins expressed in cells with and without additional hemin often form insoluble inclusion bodies that require harsh denaturing conditions for solubilization. After solubilization, chains must then be renatured in vitro and form correctly folded native globin chains, which then must properly assemble to form authentic Hb tetramers (1,7–9). This process is labor-intensive, and efficiency of tetramer reconstitution from denatured globin chains is very low. In addition, insoluble apo-α-globin chains expressed in E. coli were refolded with heme after solubilization in 0.1 M From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

243

244

Adachi

NaOH in the absence of native β-chains (10). However, the functional properties and heme environment of these resolubilized recombinant α-chains were not exactly the same as those of normal human α-chains (10). Furthermore, differences between Thr at γ112 at the α1γ interaction site instead of Cys at β112 make it difficult to isolate individual native, heme-intact γ-chains from native α2γ2 tetramers using p-mercuribenzoate (11). It is therefore critical to express soluble globin chains in vivo for evaluation of folding and assembly of α-globin chains with non-α-globin chains. We can now express soluble authentic β-, α-, and γ-globin chains employing an E. coli expression system that contains cDNAs for human globin chains and MAP (12–14). 2. Materials 2.1. Generation of Plasmids for Soluble Recombinant Human Globin Chains in E. Coli The original pHE2 plasmid (kindly provided by Drs. C. Ho and T.-J. Shen, Carnegie Mellon University, Pittsburgh, PA) was constructed to coexpress α- and β-globin chains with MAP (10) under transcriptional control of a ptac promoter in order to obtain soluble authentic human HbA without N-terminal methionine (6). 1. Expression vector pHE 2β: To obtain authentic β-globin chain expression vector pHE 2β, the α-globin cDNA is removed from pHE2 (10) by digestion with XbaI (Gibco-BRL, Gaithersburg, MD), and the 6.3-kb fragment that contains β globin and MAP cDNAs is purified using a Gene Clean II kit (BIO 101, Vista, CA) and ligated by incubation with T4 DNA ligase (Gibco-BRL). A diagram of the β globin expressing plasmid pHE2β is shown in Fig. 1. 2. Expression vector pHE 2α: To obtain expression α-globin chain expression vector pHE 2α, β-globin cDNA is removed from pHE2 plasmid (10) by digestion with PstI and NheI, and the NheI end is filled by the Klenow fragment. The PstI end is cleaved by T4 DNA polymerase and the 6.3-kb fragment that contains α globin and MAP cDNAs is then purified using a Gene Clean II kit and ligated by incubation with T4 DNA ligase (Fig. 1). 3. Expression vector pHE 2γ: To obtain expression γ-globin chain expression vector pHE 2γ, human Gγ-globin cDNA (441 bp) can be generated by polymerase chain reaction (PCR) from pGS 389γ, which we previously used as a yeast expression vector and which contains α- and Gγ-globin cDNAs (15). Two primers, no. 1 (5'-TACCGTTCTGACTTCGAAATA-3' linked to 5'-sequence that contains a PstI site and complementarity to a ribosome binding site [RBS] present in pHE2) and no. 2 (5'-TGTGAAATGACCCATATGTTATTCCTCCT-3', which contains a 3'-end sequence corresponding to the RBS in pHE2 followed by a 24-bp complementary overlap to primer no. 3), can be used to generate PCR fragment 1. Fragment 1 therefore contains a PstI site linked to an RBS. Two other primers,

Single Globin-Chain Expression and Purification

245

Fig. 1. Expression vector pHE2β containing human β-globin and mAP cDNAs. Ptac, Ori, 5S, and T1T2 refer to tac promoter, origin of DNA replication, 5S rRNA gene, and transcriptional terminators, respectively.

no. 3 (5'-GAATAACATATGGGTCATTTCACAGAGGA-3', which is complementary to the RBS and 5'-untranslated region of γ-globin cDNA in pGS 389γ) and no. 4 (5'-GACCGCTTCTGCGTTCGTA-3', which contains an NheI site followed by the 3'-end γ-cDNA sequence), can be used to generate PCR fragment 2. We have created a third PCR product that contains PCR fragments 1 and 2 by overlap PCR using primers no. 1 and no. 4. Annealing of PCR fragments 1 and 2 can be facilitated by sequence complementarity of primers no. 2 and no. 3 originally used to generate PCR fragments 1 and 2, respectively. The resultant PCR fragment 3 contains a 5' PstI site linked to an RBS followed by full-length human γ-globin cDNA linked to a 3'NheI site. PCR fragment 3 is isolated, digested with PstI and NheI, and then exchanged for β-globin cDNA in the vector pHE 2 (10) after PstI and NheI digestion to remove the β-globin cDNA insert. The resulting vector pHE 2αγ contains both α- and γ-globin cDNAs for expressing HbF in bacteria. The α-globin cDNA insert is removed following digestion with XbaI, and the remaining fragment is isolated and religated to generate pHE2γ, which can be used for expression of individual γ-chains in bacteria. A diagram of plasmid pHE2γ which expresses γ-globin chain would be the same as that shown in Fig. 1.

246

Adachi

3. Methods (see Notes 1–3) 3.1. Expression of β-Globin Chains To achieve expression of authentic β-globin chain alone, the plasmid is transfected into E. coli (JM 109) (Promega, Madison, WI), and the bacteria are grown at 30˚C with shaking at 225 rpm in 1 L terrific broth (TB) containing 10 µM ampicillin to a density of about 3 × 1010 bacteria/mL. β-Globin chain expression is induced for 2 h at 30˚C by the addition of 0.2 mM isopropyl-β-D-thiogalactoside (IPTG) (Fisher, Fair Lawn, NJ), and cultures should then be supplemented with 10 µM hemin (Aldrich, Inc., Milwaukee, WI) and 0.1% (w/v) glucose.

3.2. Isolation and Purification of Soluble β-Globin Chains After a 2-h incubation, bacterial cultures should be saturated with CO gas to convert expressed β globin to the CO form. Bacteria are pelleted by centrifuging at 2300g for 10 min; resuspended in 10 mM phosphate buffer, pH 8.6; lysed by sonication at 4˚C; and then centrifuged at 4°C for 45 min at 27,000g. Soluble β-globin chains are purified as described as follows at 4°C, and CO gas should be introduced at each purification step to maintain β globin in the CO form. The supernatant is then applied to a DEAE-cellulose (Sigma, St. Louis, MO) column equilibrated with 10 mM phosphate buffer, pH 8.6. This column is washed with 5 column vol of buffer and thereafter fractions are eluted with 50 mM phosphate buffer, pH 6.3. The partially purified β-globin fraction is rechromatographed on a Mono Q column equilibrated with 10 mM phosphate buffer, pH 8.6, and further purified with a linear gradient from 10 mM phosphate buffer, pH 8.6, to 50 mM phosphate buffer, pH 6.3. The purified soluble β globin in the CO form is then concentrated by Centriprep 10 (Amicon, Beverly, MA) and stored at –70˚C before use. Separation of polymeric and monomeric forms of β globin can be achieved by gel filtration on a Superose 12 column in 100 mM potassium phosphate buffer, pH 7.0.

3.3. Expression of α-Globin Chains To achieve expression of individual α-globin chains, the plasmid (pHE 2α) is transfected into E. coli (JM 109), and the bacteria are grown at 30˚C with shaking at 225 rpm in 1 L of TB containing 10 µM ampicillin to a density of about 3 × 1010 bacteria/mL. Expression of α globin is then induced for 2 h at 30˚C by the addition of 0.2 mM IPTG, and cultures are supplemented with 30 µM hemin and 0.1% (w/v) glucose.

3.4. Isolation and Purification of Soluble α-Globin Chains After a 2-h incubation and before cell lysis, bacterial cultures are saturated with CO gas to convert expressed α globin to the CO form. Bacteria is then

Single Globin-Chain Expression and Purification

247

pelleted by centrifuging at 2300g for 20 min; resuspended in 10 mM phosphate buffer, pH 6.0; lysed by sonication at 4˚C, and centrifuged at 4°C for 45 min at 27,000g. Soluble α-globin chains can be purified as described below at 4˚C, and CO gas should be introduced at each purification step to maintain α globin in the CO form. The supernatant is then applied to a CM-52 column equilibrated with 10 mM phosphate buffer, pH 6.0. The column is washed with 5 column volumes of buffer, and thereafter fractions are eluted with 50 mM phosphate buffer, pH 8.3. The partially purified α-globin fraction is rechromatographed on a Source 15S column (Pharmacia Biotech, Piscataway, NJ) equilibrated with 40 mM Bis-Tris buffer, pH 5.8. α Globin is eluted following a linear gradient from 40 mM Bis-Tris buffer pH 5.8, to 40 mM Bis-Tris buffer, containing 0.2 M NaCl, pH 6.3. The purified soluble α globin in the CO form can be concentrated by Centriprep 10 and be stored at –70°C before use.

3.5. Expression of γ-Globin Chains For expression of individual γ-globin chains, human Gγ-globin cDNA (441 bp) can be generated by the same conditions are used for expression of γ-chain variants as for expression of β- and α-chains.

3.6. Isolation and Purification of Soluble γ-Globin Chains After a 2-h induction and before cell lysis, bacterial cultures are saturated with CO gas to convert expressed γ-globin chains to the CO form. Bacteria are then pelleted by centrifuging at 2300g for 20 min; resuspended in 10 mM phosphate buffer, pH 6.0; lysed by sonication at 4°C; and centrifuged at 4°C for 45 min at 27,000g. Soluble γ-globin chains can be purified as described as follows at 4°C. CO gas should be introduced at each purification step to maintain γ globin in the CO form. The supernatant containing γ-globin chains is applied to a Q-Sepharose column equilibrated with 10 mM phosphate buffer, pH 8.0. The column is washed with 5 column vol of buffer, and the γ-globin fraction is eluted with 50 mM phosphate buffer, pH 6.3. The partially purified γ-globin fraction is rechromatographed on a Source 15S column equilibrated with 40 mM Bis-Tris/HCl buffer, pH 5.8, and γ globin are eluted following a linear gradient to 40 mM Bis-Tris/HCl buffer containing 0.2 M NaCl, pH 6.3. The purified soluble γ globin is then concentrated by Centriprep 10 (Amicon) and stored at –70°C before use.

3.7. Biochemical Characterization of Purified Globin Chains Molecular mass and sample purity of expressed globin chains can be assessed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDSPAGE) as previously described (16). In addition, mass determination should be done to confirm independently results from SDS-PAGE and to evaluate

248

Adachi

N-terminal methionine cleavage from globin chains. Electrospray ionization mass spectrometry can be performed on a VG BioQ triple quadrupole mass spectrometer (Micromass, Altrincham, Cheshire, UK) (17). Purified globin chains should also be analyzed by cellulose acetate electrophoresis and mobilities compared with those of authentic human globin chains on Titan III membranes at pH 8.6 with Super-Heme buffer (Helena Laboratories, Beaumont, TX). Absorption spectra of the purified globin chains in the CO form should be recorded using a spectrophotometer. CO of Hbs can be removed by first blowing oxygen across the surface of the Hb solution on ice in a rotary evaporator under a 150-W floodlight bulb. Assembled tetramers using recombinant α and non-α single chains can be characterized after separation from excess free globin chains by fast protein liquid chromatography using a Source 15S column. Formation of α2β2 tetramers can also be monitored by high-performance liquid chromatography using a POROS HQ (10 × 0.46 cm) column (PerSeptive, Framingham, MA) and by cellulose acetate electrophoresis on Titan III membranes at pH 8.6 with Super-Heme buffer. Oxygen-dissociation curves of assembled and human native tetramers can be determined in 50 mM Bis-Tris buffer containing 0.1 M NaCl and 5 mM EDTA, pH 7.2, at 20˚C using a Hemox Analyzer (TCS, Huntingdon Valley, PA) (18). 4. Notes 1. Major critical factors in the production of single globin chains with high yields are (1) the use of fresh replated bacterial cells with the plasmids using fresh media rather than frozen bacterial cells, and (2) the prevention oxidation of Hb by bubbling with CO gas after expression and prior to each purification step to maintain globin chains in the CO form. 2. Purified β-chains migrate predominantly as 32-kDa β-chain dimers caused mainly by oxidization of β93 Cys, which results in formation of disulfide-linked dimers (12). Addition of 20 mM dithiothreitol to β-chains in solution converts the dimers to monomers. 3. Globin-chain variants can be constructed and expressed using the single-chain pHE 2β, 2α, or 2γ plasmid vectors, which contain cDNAs coding for each globin chain and MAP. The basic strategy for generation of amino acid variants by sitespecific mutagenesis of normal single chains involves recombination/PCR as described previously (18). Clones should be subjected to DNA sequence analysis of the entire globin cDNA region using site-specific primers and fluorescently tagged terminators in a cycle-sequencing reaction in which extension products are analyzed on an automated DNA sequencer.

References 1. Nagai, K., and Thogerson, H. C. (1984) Generation of β-globin by sequence-specific proteolysis of a hybrid protein produced in Escherichia coli. Nature 309, 810–812.

Single Globin-Chain Expression and Purification

249

2. Wagenbach, M., O’Rourke, K., Vitez, L., Wieczorek, A., Hoffman, S., Durfee, S., Tedesco, J., and Stetler, G. (1991) Synthesis of wild type and mutant human hemoglobins in Saccharomyces clevisiae. Bio/Technology 9, 57–61. 3. Adachi, K., Konitzer, P., Lai, C. H., Kim, J., and Surrey, S. (1992) Oxygen binding and other physical properties of human hemoglobin made in yeast. Protein Engin. 5, 807–810. 4. Hoffman, S. J., Looker, D. L., Roehrich, J. M., Cozart, P. E., Durfee, S. L., Tedesco, J. L., and Stetler, G. L. (1990) Expression of fully functional tetrameric human hemoglobin in Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 8521–8525. 5. Herman, R. A., Hui, H. L., Andracki, M. E., Nobel, R. W., Sligar, S. G., Walder, J. A., and Walder, R. Y. (1992) Human hemoglobin expression in Escherichia coli: importance of optimal codon usage. Biochemistry 31, 8619–8628. 6. Shen, T-J., Ho, N. T., Simplaceanu, V., Zou, M., Green, B. N., Tam, M. F., and Ho, C. (1993) Production of unmodified human adult hemoglobin in Escherichia coli. Proc. Natl. Acad. Sci. USA 90, 8108–8112. 7. Groebe, D. R., Busch, M. R., Tsao, T. Y. M., Luh, F. Y., Tam, M. F., Chung, A. E., Gaskell, M., Liebhaber, S. A., and Ho, C. (1992) High-level production of human α-and β-globins in insect cells. Protein Expression Purif. 3, 134–141. 8. Fronticelli, C., O’Donnell, J. K., and Brinigar, W. S. (1991) Recombinant human hemoglobin: expression and refolding of β-globin from Escherichia coli. J. Protein Chem. 10, 495–501. 9. Nagai, K., and Thogerson, H. S. (1987) Synthesis and sequence-specific proteolysis of hybrid proteins produced in Escherichia coli. Methods Enzymol. 153(Pt. D), 461–481. 10. Sanna , M. T., Razynska, A., Karavitis, M., Koley, A. P., Friedman, F. K., Russu, I. M., Brinigar, W. S., and Fronticelli, C. (1997) Assembly of human hemoglobin: studies of Escherichia coli-expressed α-globin. J. Biol. Chem. 272, 3478–3486. 11. Huehns, E. R., Dance, N., Jacobs, M., Beaven, G. H., and Shooter, E. M., (1965) Isolation and properties of the βA- and γF-chain subunits from normal and foetal haemoglobins. J. Mol. Biol. 12, 215–224. 12. Yamaguchi, T., Pang, J., Reddy, K. S., Witkowska, H. E., Surrey S., and Adachi, K. (1996) Expression of soluble human β-globin chains in bacteria and assembly in vitro with α-globin chains. J. Biol. Chem. 271, 26,677–26,683. 13. Adachi, K., Zhao, Y., Yamaguchi, T., and Surrey, S. (2000) Assembly of γ- with α-globin chains to form human fetal hemoglobin ion vitro and in vivo. J. Biol. Chem. 275, 12,424–12,429. 14. Adachi, K., Yamaguchi, T., Yang, Y., Konitzer, P. T., Pang, J., Reddy, K. S., Ivanova, M., Ferrone, F., and Surrey, S. (2000) Expression of functional soluble human alphaglobin chains of hemoglobin in bacteria. Protein Expr. Purif. 20, 37–44. 15. Adachi, K., Konitzer, P., Kim, J., Welch N., and Surrey, S. (1993) Effects of β6 aromatic amino acids on polymerization and solubility of recombinant hemoglobins made in yeast. J. Biol. Chem. 268, 21,650–21,656. 16. Laemmli, U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685.

250

Adachi

17. Shackleton, C. H. and Witkowska, H. E. (1994) Mass spectometry in the characterization of variant hemoglobins, in Mass Spectrometry: Clinical and Biomedical Applications, vol. 2 (Desiderio, D. M., ed.), Plenum, New York, pp. 135–199. 18. Festa, R. S. and Asakura, T. (1979) The use of an oxygen dissociation curve analyzer in transfusion therapy. Transfusion 19, 107–113.

Nuclear Magnetic Resonance of Hb

251

15 Nuclear Magnetic Resonance of Hemoglobins Jonathan A. Lukin and Chien Ho 1. Introduction 1.1. Theory of Nuclear Magnetic Resonance Nuclear magnetic resonance (NMR) spectroscopy detects the interaction of radiofrequency (rf) radiation with the nuclear spins of molecules placed in an applied magnetic field. Because the spins are sensitive to their environment, and may be coupled to one another both through chemical bonds and through space, NMR can provide a wealth of information on the structure and dynamics of macromolecules. In particular, NMR has proven to be a powerful technique for investigating the structure-function relationship of hemoglobin (Hb). In this chapter, we focus on the procedures involved in applying one-dimensional and two-dimensional NMR spectroscopy to Hb and give examples of the information that may be obtained from this method. We begin with a brief outline of theory; a more complete treatment can be found in several excellent books (1–8). A typical NMR sample may contain approx 3 × 10–7 mol of Hb, which include ≈1021 hydrogen atoms, each of which has a nucleus (i.e., a proton) with a nuclear spin 1/2. In the presence of an applied magnetic field B0, each nucleus will occupy one of two possible energy levels, corresponding to the z-component of the spin being either parallel or antiparallel to B0, which conventionally points along the z-axis. At thermal equilibrium, a slight excess (a few parts in 105) of spins will occupy the lower energy level and be parallel to B0. This small population difference is crucial to the NMR signal resulting from the absorption of rf energy, which excites spins from the lower to the upper energy state. The equilibrium population distribution is restored by spin-lattice relaxation, which takes place with a characteristic relaxation time T1. From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

251

252

Lukin and Ho

Fig. 1. Illustration of sample magnetization M, applied static magnetic field B0, and rf field B1. In the absence of B1, the magnetization will precess about B0 at the Larmor frequency ω0 = γB0.

The individual spins add up to give the bulk magnetization M of the sample, which at equilibrium is parallel to the applied field B0. However, if M is tilted at an angle θ to B0, the magnetization will experience a torque perpendicular to both M and B0 (see Fig. 1). This torque causes M to precess about B0, at a frequency (in radians/seconds) ω0 = γB0, in which γ is the gyromagnetic ratio characteristic of the type of nucleus, and ω0 is called the Larmor frequency. We now consider a rotating reference frame (x', y', z'), in which z' points along B0, while the x'- and y'-axes rotate so as to keep pace with the precessing magnetization. Viewed in this frame, M appears motionless, so the effective magnetic field vanishes in the rotating frame. Now introduce a rotating magnetic field B1, along the x' axis. As seen in the rotating frame, B1 will appear to be a static field, so M will precess about the x'-axis at an angular frequency ω1 = γB1. The motion of M with respect to the laboratory frame will be a combination of its precession about B0 and B1. In a modern NMR spectrometer, the static field B0 is provided by a superconducting solenoid, and the oscillating rf field B1 is generated by a small coil in the probe. The strength of the magnet is often expressed in terms of the Larmor frequency of protons converted to Hertz; i.e., ν0 = ω0/2π. For example, a 600-MHz spectrometer has a 14.0-T magnet. In the simplest one-pulse NMR experiment, the sample begins at equilibrium, with its magnetization along z. The rf field B1 is applied for a time τ such that γB1τ = π/2 and is then switched off. This 90° pulse rotates M into the x-y plane. The subsequent precession of M at the Larmor frequency causes a changing magnetic flux through the same coil used to generate B1. The oscillating voltage induced in the coil is amplified, mixed with the rf reference frequency

Nuclear Magnetic Resonance of Hb

253

in a phase-sensitive detector, digitized by an analog-to-digital converter (ADC), and stored in a computer’s hard drive. The time-dependent NMR signal is called the free induction decay (FID). Its amplitude decays as a result of relaxation processes, which are classified into two types. Spin-lattice, or longitudinal relaxation (mentioned earlier), involves the exchange of energy between spins and the molecule’s motional degrees of freedom. This restores the Boltzmann population distribution among spins, so that the net magnetization returns exponentially, with a characteristic time T1, to the +z axis. The second type of relaxation is spin-spin, or transverse relaxation, which causes the spins to “fan out” in the x-y plane following a 90° pulse. Thus, the net transverse magnetization (the vector sum) decays with a characteristic time T2. In a typical 1D-NMR experiment, the FID is acquired for ≈1 s. Then, the pulse sequence is repeated, and the second FID is added to the first in computer memory. This process is repeated until an adequate signal-to-noise ratio is achieved. The time-dependent NMR signal is converted to the frequency-domain spectrum by Fourier transformation. According to the theory we have already described, this would yield a single resonance at the Larmor frequency, with a Lorentzian line shape of width ∆ν = 1/(πT2). Fortunately, a real NMR spectrum is more informative than this. Within each molecule, the static magnetic field B0 induces circulation of the electrons, so as to produce an opposing magnetic field. Thus, the surrounding electrons will partially shield, or screen, a nucleus from the influence of the field B0. Each nucleus experiences an effective field B = B0(1 – σ), in which the shielding constant σ depends on the environment of the proton in the molecule. This effect, called the chemical shift, is expressed as the fractional difference between a given resonance frequency and that of a reference NMR peak, such as the methyl 1H resonance of the sodium salt of 2,2-dimethyl-2-silapentane-5-sulfonate (DSS). Thus, for a resonance frequency ν, the corresponding chemical shift δ in parts per million is δ = 106 × (ν – νDSS)/vDSS. In practice, DSS is not added to NMR samples of Hb. Instead, we calibrate the 1H chemical shift scale using the easily measured NMR peak of water, which resonates at 4.76 ppm from DSS at 29°C, with a temperature coefficient of –0.01 ppm/°C. The existence of a range of chemical shifts among the spins in a sample implies that a single rf pulse cannot be exactly on-resonance for all of the spins. There is a reciprocal relationship between the duration of a pulse and the bandwidth of frequencies that it excites (its excitation profile). A high-power, “hard” 90° pulse lasting 10 µs gives an effective range of frequencies of ≈105 Hz (200 ppm at 500 MHz), enough to uniformly excite all the protons in the sample. Sometimes, it is desirable to use a long, low-power “soft” pulse in order to irradiate a small part of the spectrum.

254

Lukin and Ho

Fig. 2. The 300-MHz proton NMR spectrum of HbCO A in 0.1 M phosphate at pH 7.1 in H2O at 29°C is shown. Water suppression was achieved by the jump-and-return pulse sequence, which yields a spectrum where signals upfield and downfield of the water resonance appear with opposite sign. For clarity, the negative signals have been inverted.

1.2. NMR Spectrum of Hb The 1H NMR spectrum of normal human adult hemoglobin (HbA) can be divided into spectral regions that have been used to monitor structural changes associated with the ligation of HbA, changes in pH (the Bohr effect), and the addition of allosteric effectors such as 2,3-bisphosphoglycerate or inositol hexaphosphate (9). Figure 2 shows the 300-MHz 1H spectrum of carbonmonoxyhemoglobin A (HbCO A) in 0.1 M phosphate at pH 7.1 in H2O. The peaks at 12.9 and 12.1 ppm from DSS have been assigned to the Hε2 protons of α122His and α103His, respectively, which participate in H-bonds across the α1β1-subunit interface (10). Resonances at 10.7, 10.4, and 10.1 ppm have been assigned to the Hε2 protons of β37Trp, α14Trp, and β15Trp, respectively (10). β37Trp is of particular interest because it lies within the α1β2 interface, a region essential to the cooperative oxygenation of HbA (11). The large number of

Nuclear Magnetic Resonance of Hb

255

overlapping peaks between +6 and +10 ppm originates from aromatic and amide protons. The intensity of the latter can be greatly reduced by recording the spectrum of a sample in D2O. Resonances in the most crowded spectral region (from 0 to +4 ppm) arise from aliphatic protons. The most upfieldshifted resonances in Fig. 2 (i.e., those with chemical shifts <0 ppm) originate from ring current–shifted protons located above or below either the heme porphyrins or aromatic amino acid residues. The two barely resolved peaks at –1.8 ppm are assigned to the side-chain γCH3 protons of E11Val (α62Val and β67Val), which are sensitive markers for the tertiary structure of the heme pocket (9). In both oxyhemoglobin (HbO2 A) and HbCO A, the heme iron atoms are in the low-spin, diamagnetic ferrous state. However, in other ligated states, as well as deoxyHb, the iron atoms adopt nonzero values of S, corresponding to various degrees of paramagnetism (9,12). Methemoglobin (MetHb) is in the high-spin ferric state, with S = 5/2. The low spin (S = 1/2) ferric state is occupied by cyanomet-Hb and azidomet-Hb. Unliganded (deoxy) Hb adopts the high-spin ferrous state with S = 2. The presence of one or more paramagnetic ions in a protein affects both the chemical shifts and relaxation properties of nearby nuclear spins. Unpaired electron spins on the iron induce paramagnetic hyperfine shifts by means of two contributions (9,12). The scalar or Fermi contact contribution δcon arises from unpaired spin delocalization onto resonating protons through chemical bonds or hyperconjugation. The pseudo-contact or dipolar shift δdip operates through space with a 1/r3 distance dependence as well as a complicated dependence on the polar coordinates of the nucleus in an iron-centered coordinate system. Both contributions to the hyperfine shift are proportional to S(S + 1)/T, in which T is the absolute temperature. The paramagnetic contribution to the longitudinal relaxation rate R1 ≡ 1/T1 varies as 1/r6, so that a non-selective T1 measurement can yield the ratio of distances of two nuclei to the iron. A similar distance dependence is seen in the paramagnetic contribution to the spin-spin (transverse) relaxation rate. This effect, which contributes to the line width, also contains a term proportional to B02. Because of the strong field dependence of the line width of hyperfine-shifted resonances, these NMR lines are best observed on a 300- or 400-MHz spectrometer, rather than a higher-field instrument (9). The large spread of chemical shifts induced by unpaired electrons can be seen in Fig. 3, which shows the 300-MHz 1H-NMR spectrum of deoxyHbA. The range of chemical shifts (≈100 ppm) provides the selectivity and resolution necessary to investigate specific regions of the Hb molecule (9). The broad peaks at 75 and 63 ppm from DSS arise from the hyperfine-shifted NδH protons of the proximal histidyl residues. They can be used to monitor the binding of O2 to the α- and β-hemes of Hb. Resonances in the region +10 to +23 ppm

256

Lukin and Ho

Fig. 3. The 300-MHz proton NMR spectrum of deoxyHbA under the same conditions as in Fig. 2 is shown. Spectral regions in dashed boxes are shown expanded in the insets.

from DSS are owing to two sources: hyperfine-shifted protons on the heme groups and nearby amino acid residues, and exchangeable protons owing to intra- and interresidue hydrogen bonds. As in the spectrum of HbCO A (Fig. 2), peaks from +6 to +10 ppm from DSS originate from aromatic and amide protons, while those from 0 to +4 ppm arise from aliphatic protons. Resonances from 0 down to -20 ppm arise from ring-current- and hyperfine-shifted protons on or near the hemes.

1.3. Spin-Spin Coupling and 2D NMR Two nuclear spins may be coupled both through chemical bonds and through space. Consider two spin-1/2 nuclei, A and B, in atoms that are covalently bonded. The nuclear magnetic moment of spin A produces a local field that perturbs the surrounding electrons. Since the electron orbitals extend to the other atom participating in the bond, spin B experiences a slightly different

Nuclear Magnetic Resonance of Hb

257

effective field depending on whether spin A points “up” or “down.” This electron-mediated scalar coupling is expressed as J IA·IB, in which J (expressed in Hertz) is independent of the applied field B0. Its effect is to split the resonance lines at νA and νB by J. In heteronuclear NMR experiments (discussed in Subheading 1.4.), it is usually desirable to remove the effects of coupling between each proton and its bonded 13C or 15N nucleus during acquisition, so that the resonance lines appear as singlets. Such broadband decoupling can be achieved by specifically designed, composite pulse sequences (3,6). Generally, NMR can detect scalar couplings between nuclear spins separated by up to three bonds. Nuclear spins are also coupled through space by means of the dipoledipole interaction. Consider a spin IA subjected to a continuous selective pulse. The population of the two energy levels will be equalized, so the intensity of the corresponding resonance line will vanish (saturation). A nearby spin IB (within ≈5 Å of IA) will cross-relax with IA through the mechanism of mutual spin-flips. This will perturb the population of the energy levels of IB, resulting in a fractional change in the intensity of the corresponding resonance, called the nuclear Overhauser effect (NOE) (13). The initial buildup of the NOE is proportional to 1/r6, in which r is the distance between spins. Depending on the arrangement of nuclei, a two-step pathway for cross-relaxation IA → IC → IB may be more efficient than the direct pathway IA → IB. In this case, the NOE at IB is no longer directly related to the distance rAB. This effect, termed spin diffusion, may be reduced by using short NOE mixing times, and by replacing most of the protons in the protein with deuterons (2H). The NOE experiment can determine the distance between a given nucleus and other nearby nuclei in a molecule. The convenience of measuring all pairwise distances in a single experiment is provided by 2D NOE spectroscopy (2D-NOESY). The 2D-NOESY experiment, illustrated schematically in Fig. 4, is worth describing in some detail, because it provides a good example of a general 2D-NMR experiment. The 2D-NOESY experiment uses only nonselective 90° rf pulses. The pulse sequence begins with a preparation period consisting of an initial delay time during which the spins relax to their equilibrium state, followed by a 90° rf pulse which rotates the magnetization into the x'y' plane. During the evolution period t1, each spin becomes “frequency labeled” as it precesses by an angle that depends on its chemical shift. Then, the second 90° pulse rotates the spins from the x'y' plane to the x'z' plane. During the mixing period τ, the spins cross-relax as already described, so that their level populations (and corresponding z-components of magnetization Mz) change. The third 90° pulse converts Mz to detectable transverse (x'y'-plane) magnetization. The FID is digitized and recorded as a function of the detection time t2, as each spin precesses according to its chemical shift. However, the amplitude of the FID also depends on the chemical shifts of NOE-coupled

258

Lukin and Ho

Fig. 4. Schematic illustration of 2D-NOESY pulse sequence. This is an example of a general 2D experiment, consisting of preparation, evolution, and mixing periods followed by detection (acquisition). See the text for a detailed discussion of this experiment.

spins, which have evolved during t1. The pulse sequence is repeated several times, keeping τ fixed but incrementing t1 systematically, generating a signal that is a function of both t 1 and t2. These data are Fourier transformed with respect to both time variables to yield a 2D spectrum, which correlates the precession frequencies during the evolution time with those during the detection time. In a 2D-NOESY spectrum, usually displayed as a contour plot, a cross-peak will appear at coordinates (δA, δB) if two protons with those chemical shifts are within ≈5 Å of each other. While NOESY detects couplings through space, 2D correlation spectroscopy (COSY) reveals couplings among protons linked to one another by one, two, or three chemical bonds. The COSY pulse sequence differs from that of the NOESY experiment, but may also be divided into preparation, evolution, mixing, and detection periods.

1.4. Heteronuclear NMR 2D 1H-NMR spectra can provide useful information on pairwise interactions between protons, provided that the corresponding resonances are resolved in the 1D 1H spectrum. However, most of the resonances in the spectrum of Hb are unresolved. Large proteins, such as Hb, undergo relatively slow tumbling motion in solution, leading to short transverse relaxation times, causing broadening of 1H line widths. This effect, together with the large number of hydrogen atoms in Hb, contributes to the extreme overlap of resonances seen in Figs. 2 and 3. In addition, rapid transverse relaxation causes spin coherence to decay during the time required for the transfer of magnetization through bonds. If the 1H line width exceeds the typical three-bond coupling 3J HH ≈ 3–5 Hz, then COSY-type experiments become very inefficient. In this situation, a more successful technique uses the large scalar coupling between a proton and its

Nuclear Magnetic Resonance of Hb

259

directly bonded heteronucleus (carbon or nitrogen). The spin-1/2 isotopes desirable for NMR, namely 13C and 15N, occur at low natural abundance. However, these isotopes (as well as 2H) can be incorporated at high levels in Hb through the use of a bacterial expression system for the protein (14,15). Isotopically labeled Hb samples allow the acquisition of heteronuclearedited spectra, which have two main advantages over homonuclear (1H) experiments (3). First, the chemical shift dispersion of 13C or 15N resonances is inherently better than that of 1H resonances. Second, magnetization transfer between a proton and a heteronucleus is much more efficient than between two protons, since the former involves a much larger through-bond coupling J. Because the delay required for magnetization transfer is proportional to 1/J, a one-bond heteronuclear transfer can be achieved in a shorter period, during which less transverse relaxation occurs, relative to the case of two- or three-bond 1H-1H transfer. Thus, heteronuclear NMR remains sensitive for relatively large proteins. Examples of such NMR experiments are 2D heteronuclear single- and multiple-quantum coherence (HSQC and HMQC, respectively), which correlate the chemical shifts of protons and their directly bonded nitrogens. Applications of these experiments to Hb are discussed in Subheading 3. In general, a 3D-NMR pulse sequence can be performed by combining two 2D sequences, leaving out the detection period of the first experiment and the preparation period of the second. A useful 3D-NMR experiment is 15N-edited NOESY-HSQC, in which the 2D-NOESY spectrum is spread into a third dimension corresponding to the chemical shift of the 15N nucleus bonded to one of the protons. This 15N editing helps resolve the NOESY spectrum, which in 2D would be impossibly crowded. To interpret the NOESY cross-peaks in terms of distances between protons, the 1H and 15N chemical shifts must be assigned to specific nuclei in the protein. The modern strategy for resonance assignment (16) uses 3D triple-resonance experiments to sequentially transfer magnetization along specific pathways in a (15N,13C)-labeled protein. These experiments provide overlapping sets of chemical shift correlations, which collectively cover the polypeptide backbone. However, even heteronuclear experiments lose sensitivity in proteins as large as Hb, because of rapid transverse relaxation. This problem can be somewhat alleviated by 2H labeling, which lengthens 13C relaxation times by reducing the 13C-1H dipolar interactions. The efficiency of heteronuclear NMR in high-field spectrometers has been significantly improved by the recent development of transverse relaxation-optimized spectroscopy (TROSY) (17). TROSY-based experiments are now being applied to Hb in order to assign the backbone resonances, and to determine the structure and dynamic behavior of the protein in solution.

260

Lukin and Ho

2. Materials The materials used in bacterial growth and purification of Hb are given in the original references (14,15). Here, we list the buffers and gases commonly used for NMR samples of Hb. Buffers, obtained from Sigma-Aldrich, include sodium phosphate, HEPES and Bis-Tris. The preparation of chloride-free HEPES at different pH values is described in Busch et al. (18). D2O (99.9% in deuterium content) is purchased from Cambridge Isotope. Once opened, a bottle of D2O should be reclosed as soon as possible and stored in a desiccator. Hb in deoxy, oxy, and carbonmonoxy forms is prepared using N2, O2, and CO gas, respectively. The application of CO to Hb samples should be carried out in a fume hood, with an external CO detector installed. 3. Methods 3.1. Sample Preparation The development of a bacterial expression system for the growth and purification of unmodified HbA has been described previously (14,15). This system has enabled the expression of recombinant HbA (rHbA) as well as mutant rHbs in quantities sufficient for study by a number of biophysical methods, including NMR. While 1H NMR spectroscopy is a powerful technique for investigating both normal and mutant Hbs, the amount of information available from NMR is greatly increased when samples can be isotopically labeled with 15N, 2H and/or 13C. However, a great deal of resonance overlap remains in the 2D (15N,1H) correlation spectrum of 15N-labeled HbA, because of the similarity of the α- and β- chains. This degeneracy can be reduced through the use of chainselectively labeled samples of Hb. Detailed protocols for such labeling have recently been described (19); the basic procedure is to express and purify uniformly 15N-, 2H-, and/or 13C-labeled rHbA; separate the α- and β-subunits; and recombine them with the complementary subunits (β and α, respectively) of unlabeled HbA. Samples of rHbs from our bacterial expression system are subjected to an oxidation-reduction procedure to ensure that the hemes are inserted correctly into their native conformation (14,15,20). The resulting samples are subjected to heteronuclear-edited NMR experiments as described in Subheading 3.3. Following the oxidation-reduction procedure, the rHb samples are further purified through a fast protein liquid chromatography Mono-S column (14); they are subsequently handled in the same way as samples obtained from blood donors. HbCO samples are exchanged into the desired buffer using a Centricon centrifugal concentrator with a membrane cutoff of 30 kDa. An HbCO sample can be converted to HbO2 by passing a stream of oxygen over the sample in a rotary flask under bright light in an ice-water bath for 45 min. If a deoxyHb

Nuclear Magnetic Resonance of Hb

261

sample is desired, the oxygen ligand is removed by filling the rotary flask with pure N2 gas for another 45 min. DeoxyHb samples are transferred from the rotary flask to an NMR tube previously flushed with N2, under pressure of the same gas. HbO2 and HbCO samples may be pipeted directly into the NMR tube. HbCO samples are prevented from becoming O2 liganded by blowing CO gas into the tube, over the surface of the sample, for a few minutes. NMR samples of proteins are commonly kept sterile with ≈0.2 mM sodium azide; however, the use of azide is not recommended with Hb, because its presence in solution will convert MetHb to the azidomet form. Instead, Hb samples are sterilized by filtering through a 0.22-µm membrane. NMR tubes and caps can be sterilized by overnight exposure to ultraviolet light, while other accessories used in sample handling can be autoclaved. NMR samples of Hb generally consist of 0.4–0.6 mL of 3–7% Hb (about 0.5–1.0 mM) that have been transferred to a clean, dry, scratch-free NMR tube of 5 mm outer diameter. Smaller quantities of Hb (≈0.3 mL) are adequate, if a Shigemi tube is used. As soon as the NMR sample is prepared, a 1D 1H-NMR spectrum with a spectral width of about 200 ppm should be obtained, to check the purity of the desired ligation state. The presence of met-Hb is indicated by the appearance of several broad peaks between 30 and 70 ppm from DSS. The oxidation of HbO2 to MetHb can be inhibited by the use of a basic buffer (pH ≥ 8.0) and by running the NMR experiments at or below room temperature. A barely resolved doublet of peaks at –1.8 ppm originates from the E11Val methyl protons of HbCO A; the corresponding signals of HbO2 A appear at –2.4 ppm. Therefore, a weak resonance at –2.4 ppm in the spectrum of a (nominal) sample of HbCO A indicates the presence of O2, which can be eliminated by blowing CO gas over the sample.

3.2. Spectrometer Setup The first step in setting up the spectrometer (after loading the sample) involves tuning the probe. Each rf channel of the probe may be thought of as a resonant circuit, which must be tuned to the desired observed frequency and matched to the network impedance. Recently manufactured spectrometers include a “wobble generator,” which sweeps the transmitter frequency back and forth while displaying the reflected power vs frequency on the workstation monitor. The user must adjust the “matching” and “tuning” capacitors so as to minimize the reflected power as the probe’s resonant frequency matches the observe frequency. This condition is indicated by the appearance of a sharp V-shaped dip centered horizontally on the display. As noted in Subheading 1. high-resolution NMR requires a strong applied magnetic field B0, provided by a superconducting solenoid. With the exception of hyperfine-shifted resonances in paramagnetic proteins, the spectral resolu-

262

Lukin and Ho

tion increases linearly with B0, while the sensitivity is proportional (3) to B03/2. Thus, a high-field spectrometer operating at 500 MHz or higher is desirable for most NMR studies of Hb. To detect narrow resonance lines and carry out multidimensional NMR experiments over several days, the static field must be extremely homogeneous and stable. The stability of the applied field is maintained by the deuterium field-frequency lock. A separate, dedicated channel of the spectrometer continuously monitors the resonance of 2H in the sample and forms part of a feedback loop. Any drift of B0 will cause a change in the 2H resonance frequency, which activates a current in an auxiliary electromagnet (coaxial with the superconducting solenoid) so as to compensate for the drift. Meanwhile, the amplitude of the 2H absorption signal is displayed on the workstation monitor. The integrated intensity of the lock signal will be constant, so if the 2H line can be made narrower by increasing the homogeneity of the field, the signal amplitude will increase. Thus, the level of the 2H lock is usually monitored while the homogeneity of the field is optimized, in a process called shimming. This involves adjusting the current in a set of room temperature coils. These shim coils create magnetic field gradients proportional to z, z2, z3, x, xz, xy, and so on that can compensate for the residual inhomogeneity of the superconducting magnet. State-of-the-art NMR spectrometers are equipped with pulsed-field gradients, which can be used to map the static field and adjust the shims automatically. Less advanced instruments must be shimmed manually, by adjusting the shim currents in an iterative fashion (3,5). In the absence of 2H in the sample, the FID is monitored, and the shim currents as adjusted to produce a smooth exponential decay in the envelope of the FID. After the probe has been tuned and the magnet shimmed, the next step is to calibrate the 90° pulse width for protons. At a given power level, the 90° pulse width will be the duration of the rf pulse that results in an FID of maximum amplitude. Rather than repeat the one-pulse experiment while varying the pulse width to search for the maximum signal, it is more convenient to search for the null signal resulting from a 360° pulse, and to divide the pulse width by four. The experiment can be repeated several times in quick succession without waiting for longitudinal relaxation to occur, since a 360° pulse returns the magnetization to equilibrium. The NMR signal of a sample of Hb in water will be dominated by the protons in the solvent, which are present at a concentration of 110 M, while the concentration of the protein is only ≈1 mM. This represents a problem for the dynamic range of the spectrometer. If the receiver gain is set too high, then the intense signal at the beginning of the FID will overload the digitizer. The resulting “clipping” of the FID causes severe distortion of the Fourier-transformed NMR spectrum. On the other hand, if the gain is low enough that the full FID is within the range of the digitizer, then the small

Nuclear Magnetic Resonance of Hb

263

fraction (≈10–5) of the signal that originates from the protein will be coarsely digitized by a few least-significant bits. Additionally, many protein resonances will be obscured by the enormous water peak. To alleviate these problems, several methods of water suppression have been developed. In general, the rf carrier frequency is set at the water resonance. The simplest solvent suppression technique is presaturation of the water resonance by a weak rf field applied during the recycle delay. This has the disadvantage of reducing the intensities of exchangeable proton resonances (such as amide protons) through saturation transfer with water. A useful alternative is the jump-and-return pulse sequence (21), which consists of a 90° pulse followed by a short precession period τ and then another 90° pulse of opposite phase. Since the water line is on resonance, it undergoes no precession (in the rotating frame) during τ, so that the second pulse returns it to the z-axis and it does not contribute to the FID. However, all off-resonance spins retain a component in the x-y plane. The signal amplitudes change sign depending on whether they are upfield or downfield of the water resonance. This method of water suppression was used to obtain the spectra shown in Figs. 2 and 3. To accurately represent the NMR spectrum over a given range of frequencies, it is necessary to set the sampling rate at which the digitizer (ADC) records discrete points of the FID. Note that the signal reaching the receiver from the probe consists of a superposition of harmonic components with frequencies νi = ν0(1 + 10–6δi), in which ν0 is the Larmor frequency of the reference compound (e.g., DSS) and δi is the chemical shift of spin i. This signal is split into two separate channels. The high-frequency (megahertz) signals in the two channels are modulated by the carrier (reference) frequency νref to generate audio (kilohertz) signals with frequencies νi – νref. Thus, we detect the signal relative to the rotating frame of reference. The reference signals in the two channels are 90° out of phase; therefore, the output consists of cosine- and sine-modulated FIDs, referred to as the real and imaginary parts of the spectrum, respectively. This quadrature detection scheme allows us to determine the sign of the frequency offset with respect to νref. In the following discussion, we refer to the frequency offset νi – νref simply as the frequency. The range of frequencies that we can detect is determined by the time interval between sampling points, called the dwell time. To uniquely represent a sinusoidal signal, it must be sampled at least twice per cycle. Therefore, a dwell time, DW, will yield an accurate representation of frequencies within a spectral width SW = 1/(2DW). A frequency ν0 outside this range will appear at an aliased frequency νa = mSW + ν0, in which m is an integer such that νa appears within the spectral width. In modern NMR spectrometers, the user specifies the spectral width, SW, and the computer calculates DW. For

264

Lukin and Ho

example, consider a 300-MHz NMR spectrum of HbCO A taken with the carrier frequency set at the water resonance (4.76 ppm) and a SW = 4680 Hz, or 15.6 ppm. This will provide an accurate picture of peaks that resonate in the range 4.76 ± (15.6/2 ppm), i.e., from –3.04 to +12.56 ppm. The peak at 12.9 ppm will be aliased to 12.9 – 15.6 = –2.7 ppm. The digital resolution of the Fourier-transformed spectrum depends on the amount of computer memory used to record the FID. If the memory size is SI, then SI/2 real and SI/2 imaginary points are obtained in the spectrum. The frequency separation between these data points is given by the digital resolution DR = 2SW/SI, in which SW (as earlier) is the spectral width in hertz. The total acquisition time, AQ, for a single FID is the product of the DW and the number of points collected: AQ = DW × SI = 1/(2SW) × (2SW/DR) = 1/DR. Thus, the digital resolution is simply the reciprocal of the acquisition time. This implies that if all of the resonances are broader than 2 Hz, it is useless to acquire the FID for much longer than (0.5 s—the signal will have decayed by this time, and continued acquisition will only add noise to the spectrum. On the other hand, if the acquisition time is too short, a meaningful portion of the FID will be truncated, leading to artifacts in the Fourier-transformed spectrum. As noted in Subheading 1., the NMR pulse sequence is repeated several times, and the FIDs are added cumulatively in computer memory. Thus, the last two parameters to be set are the recycle delay between transients and the number of scans, Ns. The signal-to-noise ratio of the summed spectrum is proportional to √ Ns . As a compromise between waiting long enough to achieve complete recovery of the longitudinal magnetization and acquiring more transients in a given time, the recycle delay is usually set at 1 to 2 s.

3.3. 2D Heteronuclear NMR: HSQC and HMQC As already noted, resonances which overlap in the 1D proton spectrum of Hb can be resolved through the application of 2D heteronuclear NMR to isotopically labeled samples. A full description of the theory and experimental aspects of 2D NMR is beyond the scope of this chapter. However, several of the spectral parameters for the 15N dimension of a (1H,15N) correlation experiment are analogous to those of a 1D 1H spectrum. The duration of a 90° pulse for the 15N channel can be calibrated by maximizing the signal of a 1D version of the 15N-edited HSQC spectrum. The carrier frequency, spectral width, and number of time domain points acquired are set independently for the 15N and 1H dimensions. The 2D (1H,15N) HSQC spectra of chain-selectively (2H,13C,15N) edited samples of HbCO A (19) are shown in Fig. 5. It is clear that a significant degeneracy of chemical shifts would occur in the 2D spectrum of fully labeled Hb. However, by labeling the α- and β-chains in different samples, most of the resonances can be resolved. Note that only 15N labeling is required for the

Nuclear Magnetic Resonance of Hb

265

Fig. 5. The 600-MHz HSQC spectra of chain-selectively labeled HbCO A in 95% H2O/5% D2O with 0.1 M phosphate at pH 7.0 and 29°C is shown. (A) Spectrum of a sample of tetrameric HbA with (2H,15N,13C)-labeled α-chains and natural-abundance β-chains; (B) spectrum of a sample with natural abundance α-chains and (2H,15N,13C)labeled β-chains. Both spectra were acquired with a proton acquisition size of 2048 points and spectral width of 20 ppm. Two hundred fifty-six complex points were acquired in the indirect 15N dimension, with the 15N carrier frequency and spectral width set at 118 and 50 ppm, respectively. Signal averaging was carried out over 16 scans. Light gray contours indicate cross-peaks that are folded in the 15N dimension.

266

Lukin and Ho

HSQC experiment, although 2H labeling leads to narrower line widths. The spectra of Fig. 5 were obtained after the samples had been left in 95% H2O/5% D2O solution for several months. By this time, all but a few deeply buried backbone amides are expected to have exchanged their original 2H atoms with solvent protons, which would yield cross-peaks in the spectrum. On the other hand, NH groups that undergo fast exchange on the millisecond time scale do not exhibit cross-peaks, since their resonances are broadened beyond detection or eliminated by water suppression. Some side-chain NHs are protected by hydrogen bonding from rapid exchange and thus exhibit cross-peaks in Figs. 5A,B. These include a few Asn, Gln, and Arg side chains as well as the three nonequivalent Trp residues in HbA, the proximal histidines of the α- and β-chains, and two histidines (α103His and α122His) that participate in H-bonds in the α1β1 interface. Additional His and Trp side-chain resonances are revealed by a 2D HMQC experiment performed without 15N decoupling during acquisition. The echo antiecho HMQC pulse sequence uses pulsed-field gradients for coherence selection, suppressing the solvent signal while preserving nearby protein resonances (10,22). In the spectrum, broad doublet cross-peaks appear for directly bonded (H,N) groups, while sharp cross-peaks originate from protons coupled to 15N through two or three bonds. Thus, for Trp residues, crosspeaks appear correlating Nε 1 with both Hε 1 and the carbon-bound Hδ1. Histidine imidazole side-chain resonances appear in a characteristic region, with 15N chemical shifts >150 ppm. In these side chains, the coupling of both carbon-bound protons to both nitrogens creates a rectangular pattern of four cross-peaks, as shown in Fig. 6. A weak or missing cross-peak is diagnostic of the weak three-bond (Hδ2, Nδ1) coupling. The C2 proton (Hε1) resonates at a higher chemical shift than the C4 proton (Hδ2), whereas the protonated 15N of the imidazole ring resonates at a lower chemical shift than the bare nitrogen (23). Thus, the appearance of the pattern of cross-peaks allows the identification of both protons and nitrogens, as well as the tautomeric state of the histidine. The HMQC spectra of chain-selectively labeled samples have been used to confirm and extend the assignments of all 38 histidines in HbCO A. These residues are of particular interest because of their importance to the Bohr effect in Hb. Each histidine’s contribution to the Bohr effect can be calculated from the difference in its pK between deoxy and oxy (or carbonmonoxy) Hb. Accurate pK values are determined from a nonlinear leastsquares fit of the chemical shift of each C2 proton as a function of pH. NMR measurements have established that a global network of electrostatic interactions plays a dominant role in the Bohr effect, with some histidyl residues opposing the net Bohr effect (18,20,24,25). These results illustrate the use-

Nuclear Magnetic Resonance of Hb

267

Fig. 6. Surface histidine region of the 600-MHz echo antiecho HMQC spectrum of HbCO A in water with 0.1 M phosphate at pH 6.86 and 29°C. Lines connect crosspeaks originating from the same residue; solid and dashed lines are used for α- and β-chain residues, respectively. For this experiment, the proton time-domain acquisition size was 8192 points, and the 1H spectral width was 22 ppm. Eighty scans were averaged for each of 256 complex points in the 15N dimension. To cover both the backbone amides and histidine side-chain resonances without folding, the 15N spectral width was set at 240 ppm and the carrier at 160 ppm. (From Fig. 4 of ref. 19.)

fulness of NMR techniques in elucidating the relationship between structure and function in Hb. Acknowledgments We thank Nancy T. Ho and Virgil Simplaceanu for helpful discussions. Our Hb research is supported by a grant from the National Institutes of Health (R01HL-245215).

268

Lukin and Ho

References 1. Atta-ur-Rahman and Choudhary, M. I. (1996) Solving Problems with NMR Spectroscopy, Academic, San Diego. 2. Becker, E. D. (2000) High Resolution NMR: Theory and Chemical Applications, Academic, San Diego. 3. Cavanagh, J., Fairbrother, W. J., Palmer, A. G., III, and Skelton, N. J., (1996) Protein NMR Spectroscopy Principles and Practice, Academic, San Diego. 4. Croasmun, W. R. and Carlson, R. M. K. (eds.) (1994) Two-Dimensional NMR Spectroscopy: Applications for Chemists and Biochemists, VCH, New York. 5. Derome, A. E. (1987)Modern NMR Techniques for Chemistry Research, Pergamon, Oxford, UK. 6. Freeman, R. (1997) Spin Choreography: Basic Steps in High Resolution NMR, Spektrum, Oxford, UK. 7. Sanders, J. K. M. and Hunter, B. K. (1993) Modern NMR Spectroscopy, A Guide for Chemists, Oxford University Press, Oxford, UK. 8. van de Ven, F. J. M. (1995) Multidimensional NMR in Liquids, VCH, New York. 9. Ho, C. (1992) Proton nuclear magnetic resonance studies on hemoglobin: cooperative interactions and partially ligated intermediates. Adv. Protein Chem. 43, 153–312. 10. Simplaceanu, V., et al. (2000) Chain-selective isotopic labeling for NMR studies of large multimeric proteins: application to hemoglobin. Biophys. J. 79, 1146–1154. 11. Dickerson, R. E. and Geiss, I. (1983) Hemoglobin: Structure, Function, Evolution, and Pathology., Benjamin/Cummings, Menlo Park, CA. 12. La Mar, G. N., Satterlee, J. D., and De Ropp, J. S. (2000) Nuclear magnetic resonance of hemoproteins, in The Porphyrin Handbook, vol. 5 (Kadish, K. M., Smith, K. M., and Guilard, R., eds.), Academic, New York, pp. 185–298. 13. Neuhaus, D. and Williamson, M. P. (1989) The Nuclear Overhauser Effect in Structural and Conformational Analysis, VCH, New York. 14. Shen, T.-J., Ho, N. T., Simplaceanu, V., et al. (1993) Production of unmodified human adult hemoglobin in Escherichia coli. Proc. Natl. Acad. Sci. USA 90, 8108–8112. 15. Shen, T.-J., Ho, N. T., Zo, M., et al. (1997). Production of human normal adult and fetal hemoglobins in Escherichia coli. Protein Eng. 10, 1085–1097. 16. Ikura, M., Kay, L. E., and Bax, A. (1990) A novel approach for sequential assignment of 1H, 13C, and 15N spectra of proteins: heteronuclear triple-resonance three-dimensional NMR spectroscopy. application to calmodulin. Biochemistry 29, 4659–4667. 17. Salzmann, M., Pervushin, K., Wider, G., Senn, H., and Wuthrich, K. (1998) TROSY in triple-resonance experiments: new perspectives for sequential NMR assignment of large proteins. Proc. Natl. Acad. Sci. USA 95, 13,585–13,590. 18. Busch, M. R., Mace, J. E., Ho, N. T., and Ho, C. (1991) Roles of β146 histidyl residue in the molecular basis of the bohr effect of hemoglobin: a proton nuclear magnetic resonance study. Biochemistry 30, 1865–1877.

Nuclear Magnetic Resonance of Hb

269

19. Simplaceanu, V., Lukin, J. A., Fang, T. Y., Ho, N. T., and Ho, C. (2000) ChainSelective Isotopic Labeling for NMR Studies of Large Multimeric Proteins: Application to Hemoglobin. Biophys. J. 79, 1146–1154. 20. Sun, D. P., Zou, M., Ho, N. T., and Ho, C. (1997) The contribution of surface histidyl residues in the α-chain to the bohr effect of human adult normal hemoglobin: roles of global electrostatic effects. Biochemistry 36, 6663–6673. 21. Plateau, P. and Gueron, M. (1982) Exchangeable proton NMR without base-line distortion, using new strong-pulse sequences. J. Am. Chem. Soc. 104, 7310–7311. 22. Lukin, J. A., Simplaceanu, V., Zou, M., Ho, N. T., and Ho, C. (2000) NMR reveals hydrogen bonds between oxygen and distal histidines in oxyhemoglobin. Proc. Natl. Acad. Sci. USA 97, 10354–10358. 23. Pelton, J. G., Torchia, D. A., Meadow, N. D., and Roseman, S. (1993) Tautomeric states of the active-site histidines of phosphorylated and unphosphorylated IIIGlc, a signal-transducing protein from Escherichia coli, using two-dimensional NMR techniques. Protein Sci. 2, 543–558. 24. Ho, C., and Russu, I. M. (1987) How much do we know about the Bohr effect of hemoglobin? Biochemistry 26, 6299–6305. 25. Fang, T. Y. et al. (1999) Assessment of roles of surface histidyl residues in the molecular basis of the Bohr effect and of β143 histidine in the binding of 2,3bisphosphoglycerate in human normal adult hemoglobin. Biochemistry 38, 13,423–13,432.

270

Lukin and Ho

Solubility of Sickle Polymer

271

16 Solubility Measurement of the Sickle Polymer Mary E. Fabry, Seetharama A. Acharya, Sandra M. Suzuka, and Ronald L. Nagel 1. Introduction In sickle hemoglobin (HbS), a valine is substituted for glutamic acid in the sixth codon of the globin chain. This change endows deoxyHbS, but not oxyHbS, with a new property: the capacity to polymerize. This new property reduces the pliability of the red cell, an indispensable property to navigate the microcirculation. The polymerization of deoxyHbS is the primary and indispensable (but not sufficient) event in the molecular pathogenesis of sickle cell disease. Following deoxygenation, HbS-containing cells assume a variety of distorted shapes readily appreciated by light microscopy and even more clearly by scanning electron microscopy (1). Studies at higher resolution using transmission electron microscopy have provided information concerning the structure and packing of the sickle fiber (2).

1.1. Polymer Structure The structure of Hbs has been ascertained from X-ray diffraction studies that subsequently refined the structure to a resolution of 0.2 nm (2.0 Å) and provided information on molecular orientation and contacts sites (3–5). Studies of the solubility of mixtures of HbS with Hbs A, A2, F, and α- and β-globin mutants (6,7), and more recently, using genetically engineered site-directed globin mutants (8–11) have provided further evidence on the contact sites critical for the formation of polymer. The basic polymer structure is a double-stranded filament. Each strand is a string of deoxyHbS beads aligned in a head-to-tail (or axial) array. Seven filaments combine into a fiber organized such that adjacent double strands have antiparallel orientation. The twisted structure is composed of an inner core of From: Methods in Molecular Medicine, vol. 82: Hemoglobin Disorders: Molecular Methods and Protocols Edited by: Ronald L. Nagel © Humana Press Inc., Totowa, NJ

271

272

Fabry et al.

4 strands surrounded by a sheath of 10 strands (12). Individual fibers have an elliptical cross-section of about 23 by 18 nm (11). The helix has a high pitch with a periodicity of about 300 nm. The double strands are slightly twisted in the fiber. Stretching of the outer strands in the fiber limits its size to seven pairs. Identification of the double strands (of the HbS crystal) as assembly units of the sickle fiber (6,7) was defined by the use of naturally occurring mutants in hybridized copolymerization experiments (binary mixtures), first involving β mutants (13) and later extended to α mutants (14,15).

1.2. Intermolecular Bonding More important, the 6 Val of the β1-subunit does not participate in intermolecular bonding, a feature predicted by binary mixtures involving the β6 and β2 sites (Hbs Leiden, Makassar and Deer Lodge) determining that only one β-chain in the HbS tetramer was involved in a contact area of the fiber, since their mutations did not affect polymerization (16,17). Furthermore, the contact between the β2 β6 Val and an acceptor site on the partner strand is only possible when HbS is in the T or deoxy conformation. R or oxygenated molecules cannot fit into the polymeric structure.

1.3. Other Amino Acid Residues Also Affect Bonding and Polymerization First, the β-globin mutants that alter gel formation tend to be on the β1 or “trans” subunit and involve either lateral contacts between partners of the double strand or axial contacts between members of a single strand. The α-chain contacts, which are fewer, are predominantly axial but sometimes lateral. Second, the amino acid residue β6 Val is not indispensable for polymerization. Replacement of β6 Glu (in HbA) by another hydrophobic residue, isoleucine, results in an abnormal Hb that polymerizes even more readily than HbS (18). Third, the double mutant β6 Val, β121 Gln polymerizes much more readily than does HbS (17). This finding fully supports the earlier observation of enhanced polymerization in a mixture of HbS and Hb O-Arab (β121 Gln) confirming this as a site of contact in the fiber. Finally, interactions involving the amino acid residues of proline at α114 and threonine at β87 also appear to be important for polymerization (9). As with other contacts that stabilize the polymer, amino acid replacement results in a marked decrease in polymerization, which has important implications for gene therapy. Although β87 threonine does not directly interact with the β6 Val in deoxyHbS polymers (10), it does, however, play a critical role in formation of the hydrophobic acceptor pocket that then promotes protein-protein interactions facilitating the formation of stable nuclei and polymers of deoxyHbS.

Solubility of Sickle Polymer

273

1.4. Sickle Hb Polymerization The polymerization of sickle Hb involves the self-association of identical molecules; no accessory molecules are involved. As a result, the assembly process obeys classic chemical rules of kinetics and thermodynamics (19–21).

1.5. Solubility of HbS (CSAT) The CSAT, or the Hb concentration at saturation, is the concentration of deoxyHb in equilibrium with the polymer phase. The polymerization of HbS is a simple phase change from solution to gel. When a gelled solution of deoxyHbS is examined, large polymers (fibers) and free tetramers can be readily demonstrated, but species of intermediate size cannot be detected. Thus, the equilibrium between solution and gel can be studied by measuring the concentration of free Hb in solution after separating the phases by centrifugation (19–21). For pure deoxyHbS at pH 7.0 and 20°C, the solubility is 20 g/dL, significantly less than the concentration of Hb inside the red cell. The rate of polymerization of deoxyHbS is dependent on the HbS concentration. Polymerization-induced activation of volume-regulating transport systems (22,23) leading to enhanced dehydration results in the presence of a substantial population of very dense cells that have a very high intracellular Hb concentration or mean corpuscular Hb concentration (MCHC). This dehydration of sickle erythrocytes involves increased activity of the membrane K-Cl cotransporter and the calciumdependent (Gardos) potassium channel.

1.6. Solubility of HbS Is Also Affected by Several Parameters Other Than Concentration 1.6.1. Other Hbs Of particular and practical importance are the effects of coexistent Hbs F, A, and C. Information on the copolymerization of HbS with these Hbs has provided important insights into the pathogenesis and clinical severity of the various sickle syndromes (SS) including SS with increased levels of HbF, S-βo thalassemia, S-β+ thalassemia, sickle cell (SC) disease, and AS (sickle trait). The solubility of a mixture of equal amounts of HbS and HbA (and that of HbS and HbC) is only about 40% higher than that of HbS alone. In this mixture, half of the Hb comprises asymmetric hybrid tetramers (α2βSβA). Since only one of the two β6 valines is engaged in an intermolecular contact, there is a 50% chance that the hybrid tetramer will enter the polymer in such a way that all the proper contacts are made. Incorporation of HbA into the sickle polymer has been experimentally documented (24). By contrast, the HbA tetramers (α2β2), which comprise 25% of the mixture, fail to be incorporated into the

274

Fabry et al.

polymer. Nevertheless, by virtue of their excluded volume, the solubility is further lowered (25). In contrast to HbA and HbC, HbF and HbA2 inhibit polymerization (7). Thus, the hybrid tetramers α2βSγ and α2βδ fail to be incorporated into the sickle polymer. Since HbF (α2γ2) affects polymerization by means of the asymmetrical hybrid α2βSγ (16), the inhibition is trans to the β6 Val contact (24,26). γ87 is one of the important inhibitory sites for both HbF and HbA2 (7). This residue constitutes one of the lateral contacts in the double strand of the sickle fiber.

1.6.2. pH The solubility of deoxyHbS is lowest between pH 6.0 and 7.2 and rises sharply at higher and lower pH values (27). Thus, in the extraerythrocytic pH range of 7.5–7.2 (corresponding to intracellular pHs of ~pH 7.2 and 7.0), the alkaline Bohr effect is enhanced in concentrated solutions of HbS and in SS red cells (27). Alkalosis, by shifting the equilibrium toward the oxy conformation, tends to retard sickling but also impairs oxygen release. The significant difference in the slope of the Bohr effect in sickle cells means that a drop in SS blood pH below 7.4 in tissue capillaries yields twice the normal decrease in oxygen affinity and a large release of oxygen from red cells, increasing significantly their risk of sickling. Clinically, even mild transient acidosis (with the corresponding drop of intracellular pH) would be hazardous for patients with sickling disorders. The latter was actually experimentally and unethically demonstrated when patients with sickle cell anemia were infused with low pH isotonic solutions, and painful crises ensued.

1.6.3. Temperature The polymerization of HbS is an endothermic process consistent with the importance of hydrophobic interactions (28–30). Polymer formation is therefore entropically driven, resulting from the release of ordered water molecules from the surface of free Hb. Sickle polymers are melted by cooling. Thus, a temperature jump is a simple and effective way of initiating polymerization, thereby enabling kinetic measurements. Although low temperatures do inhibit sickle polymerization, the use of cooling therapeutically is not advisable since vasoconstriction and other effects that are provasoocclusion might predominate.

1.6.4. Ionic Strength The solubility of deoxyHbS is altered by salt and buffer conditions. At salt concentrations spanning the physiological range, solubility increases with ionic strength but decreases markedly at high ionic strength (31,32).

Solubility of Sickle Polymer

275

1.6.5. Organic Phosphates The primary modulator of O2 affinity in the red cell is 2,3-diphosphoglycerate (DPG). An increase in red cell DPG favors HbS polymerization in three ways: lowered O2 affinity, a reduction in intracellular pH (both of which increase deoxyHbS), and a direct effect on the conformation of deoxyHbS (33,34).

1.6.6. Kinetics The polymerization of sickle Hb is a remarkably dynamic event. Measurements of the kinetics of polymer formation, both in pure HbS solutions and in sickle erythrocytes, have contributed to the understanding of the pathogenesis of vasoocclusive crises (35,36). These studies led directly to information on the nucleation mechanism responsible for fiber formation, and studies on intact red cells have provided an explanation at the molecular level of the morphological changes that are observed following the deoxygenation of cells, both in vitro and in vivo. Understanding the kinetics of polymerization has provided a novel and workable approach to the assessment of new antisickling therapies.

1.6.7. HbS Solution Studies The time course for polymerization of a concentrated solution of HbS can be monitored after either rapid removal of ligand (e.g., by photolysis) or by rapidly increasing the temperature of deoxyHbS, taking advantage of the markedly endothermic nature of the process. The formation of sickle fibers can be documented by a variety of physicochemical techniques including turbidity, light scattering, calorimetry, and nuclear magnetic resonance spectroscopy. The subsequent alignment of fibers is best monitored by measurement of birefringence. Following ligand removal or a temperature jump, there is a clearly measurable lag or delay time before a signal reflecting the presence of detectable polymer is detected. After the delay time, the progress of polymer formation is exponential. During the delay time nucleation occurs. Nucleation is a necessary first step in polymer formation. The number of molecules in the nucleus appears to be proportional to the slope of the concentration dependence of the delay time (19). When molecular crowding (nonideality) is taken into account, the slope of the concentration dependence of the delay time predicts a nucleus of about 15 molecules (19). Aggregates smaller than the critical nucleus are thermodynamically not favored. By contrast, once the nucleus is formed, subsequent addition of molecules is highly favored and fiber growth becomes very rapid (approx 250 Hb tetramers) (37). Studies using statistical thermodynamic modeling have suggested that there are two pathways for the nucleation of sickle Hb fibers (36,38): (1) Homogeneous

276

Fabry et al.

nucleation refers to nucleation of individual fibers occurring in the bulk solution phase; (2) heterogeneous nucleation refers to nucleation occurring on the surface of existing polymers, which leads to the autocatalytic formation of fibers, and therefore the delay period. In highly concentrated solutions of deoxyHbS, homogeneous nucleation is favored. Polarizing microscopy reveals multiple domains of polymers giving rise to birefringent spherulites. These are probably the tactoids first observed in solutions of deoxyHbS (39). In less concentrated solutions of deoxyHbS, heterogeneous nucleation predominates, leading to fewer domains of aligned sickle fibers. Studies with video-enhanced differential interference contrast microscopy show that fibers originate both from centers that produce many radially distributed fibers and on the surface of preexisting fibers (37).

1.6.8. Intracellular Studies Extension of the equilibrium and kinetic studies of polymerization to erythrocytes is greatly complicated by marked heterogeneity of SS cells, owing to a wide range of oxygen affinity, an even wider distribution of intracellular Hb concentration (20–50 g/dL), and the heterogeneous distribution of DPG and HbF. However, when these variables are taken into account, the delay times of SS red cells and the amount of polymer per cell at equilibrium are remarkably close to what is encountered in Hb solutions, indicating that the cytosolic surface of the red cell membrane has no significant effect on the delay time (40,41). The kinetics of polymerization plays a critical role in the rheology and morphology of circulating red cells (42). Because the range of transit times in the microcirculation is short relative to the range of delay times of SS red cells, the great majority (perhaps 95%) of cells fail to form polymers during their flow through arterioles and capillaries (42). By contrast, if these cells were equilibrated at the oxygen tensions in the microcirculation, virtually all of them would contain polymer and, as a result, would have markedly decreased deformability. Thus, kinetics is the critical determinant of cell shape and morphology (19). The mechanisms by which red cells containing sickle polymer can induce pathology are complex (interacting with other pleiotropic effects of sickle cells, such as endothelial adhesion of young sickle cells, regulation of rheological changes, and regulation of shunting), and organ-by-organ variations in vascular activity, particularly when lung, retina, and brain circulation are compared with other microcirculatory beds, play an important role in determining pathology. First, when SS red cells are deoxygenated slowly, they form classic elongated (sickle) shapes. This is the result of homogeneous nucleation in which one domain propagates by fiber growth and aligns to distort the cell into the

Solubility of Sickle Polymer

277

classic sickle shape. With somewhat more rapid deoxygenation, several independent domains form and will induce a more irregular shape (43). Second, when deoxygenation is rapid, multiple spherulitic domains result in a granular or cobblestone texture (sack-of-potatoes aspect) with no gross distortion of cell shape. Because the shape of the sickled cell is so dependent on the number of independent polymer domains, it is possible to convert a holly leaf cell into an elongated sickle shape by partial reoxygenation (44). The distortion of cell shape by projections of aligned HbS fibers plays a critical role in the pathogenesis of the membrane lesion. 2. Materials and Methods

2.1. Classic CSAT: The Reference Method The CSAT or solubility is the concentration of deoxy HbS in equilibrium with the polymerase phase. The reference method for determining CSAT requires several hundred microliters of concentrated Hb solution that is completely deoxygenated and then centrifuged to separate the solution and polymer phases. The concentration of Hb in the solution phase and, in some cases, the Hb composition of the supernatant are then determined. Since complete deoxygenation using nitrogen is difficult in concentrated, viscous Hb solutions, diothionite is often used to ensure that the sample is fully deoxygenated. This frequently results in a low-pH solution (6.8 or lower), and, as a consequence, the CSAT is decreased; that is, the Hb concentration in the solution phase is reduced. A common type of experiment mixes HbS with other Hbs to determine whether these have an inhibitory effect on polymer formation that is evidenced by an increase in the Hb in the solution phase, i.e., an increase in the CSAT. CSAT values are determined by preparing Hb samples in a 0.1 M potassium phosphate buffer, pH 7.35 at 25°C (see Note 1). Deoxygenation is accomplished in a mixing tube by the addition of enough Na dithionite solution to equal three times the final concentration of Hb (see Notes 2 and 3). Samples are transferred to the tube in which they will be centrifuged and allowed to gel overnight at 25°C (see Notes 4 and 5). They are centrifuged the next day at 25°C 2 h at 140,000g. Purified HbS in the same buffer is always run as a control. The supernatants are removed anaerobically, and concentrations and deoxy pHs are determined (see Notes 6 and 7). Gradients in concentration or pH above the polymer phase are a potential source of error. The solution phase should be completely removed and mixed prior to measurement of concentration or pH. To eliminate the problems inherent with dithionite, some CSATs are run without it. The Hb samples are prepared as before in 0.1 M potassium phosphate, pH 7.35 at 25°C. The samples are then placed in stoppered vials and

278

Fabry et al.

Fig. 1. Two-bottle system for preparation of dithionite.

deoxygenated on ice by N2 flow overnight and are then transferred anaerobically to CSAT tubes and allowed to gel overnight at 25°C. The Hb samples are then centrifuged for 2 h at 140,000g at 25°C. Purified HbS is used as a control. The supernatants are removed anaerobically, and concentrations and deoxy pHs are determined.

2.1.1. Preparation of Dithionite Sodium dithionite (sodium hydrosulfite, Na2S2O4) solutions deteriorate rapidly, owing to interaction with oxygen, and should be freshly prepared daily. Failure to do so is a major cause of irreproducible results. Material from commercial sources can vary considerably in quality because of deterioration from moist air during manufacture, packaging, and storage. As soon as it is received, the fine granular powder without lumps should be divided into 5- to 10-g portions under a nitrogen atmosphere. Small airtight bottles are useful for this purpose and may be kept in a desiccator also flushed with nitrogen. Without specialized glassware, we have found that it is convenient to use two small bottles connected in series (Fig. 1). The first bottle contains dithionite, and its stopper is pierced by a needle to allow the passage of nitrogen. Its outlet is fed into 5–20 mL of solvent in a second bottle, and the gas is released through another needle. Bubbling is allowed to take place for 30–60 min, after which the direction of gas flow is reversed: nitrogen is introduced through the second bottle to drive the liquid back into the first (with dithionite). To take aliquots of this solution for use, an airtight, glass syringe is first flushed with nitrogen gas and then a sample is withdrawn, the first portion of which may be discarded.

2.2. CSAT by P50 for Solutions The CSAT method described in Subheading 2.1. is regarded as the reference method (29,45). Relatively large amounts of Hb are required, the final pH is

Solubility of Sickle Polymer

279

low (about 6.8, compared with the intracellular pH of sickle cells of about 7.2 in the venous circulation), and the measurement is made under fully deoxygenated conditions in the presence of dithionite. Other methods requiring smaller amounts of Hb are also in use, such as the high phosphate method developed by Adachi and Asakura (46,47) and the dextran method developed by Bookchin et al. (48). Although these methods have the advantage of consuming smaller amounts of Hb, they deviate even further from physiological conditions and, particularly for the high phosphate method, may systematically deviate from values obtained with the CSAT method. In solution studies of HbS, Benesch et al. (49) demonstrated that p50, the point at which Hb is half saturated with oxygen, is proportional to the amount of polymer formed and could be used to measure CSAT. The method requires very small amounts of concentrated Hb (about 2 µL for each point) and can also be applied to intact red cells by varying extracellular osmolarity, which, in turn, varies intracellular Hb concentration (MCHC). Benesch et al. (49) demonstrated that in HbS solutions, the p50 is proportional to the amount of polymer formed because polymer has a much lower oxygen affinity than HbS tetramer (49). In the Benesch method, p50 is measured for a series of Hb concentrations with values below and above the expected CSAT, the Hb concentration at which polymer formation begins can be detected by the onset of rapidly, linearly increasing p50s in a plot of p50 vs Hb concentration. The concentration at which the break in the plot occurs is the CSAT. It is the CSAT because it is the maximum concentration of Hb that can exist without polymer formation. The increase in p50 is linear because, under conditions in which polymer is formed, the concentration of Hb in both the solution and polymer phases is constant and the variable is the proportion of Hb in these phases. Correlation of the CSAT determined by p50 with that determined by the centrifugation method has been demonstrated to be good for a wide range of samples. Note that using the p50 method for determining CSAT and extending it to modified forms of HbS involves the assumption that polymers of modified HbS have exactly the same structure as those of unmodified HbS, and that the influence of the polymer’s structure on the oxygen affinity for all forms of HbS is nearly the same. Under these ideal conditions, the slope of the second phase of the oxygen affinity vs protein concentration curves for modified forms of HbS would be the same as that of HbS. However, this has not been the case for all of the various mutant and/or modified forms of HbS that have been examined. The slopes of the oxygen affinity vs concentration for some of the modified forms of HbS may exhibit significant differences in slope when compared with that of HbS, although many modified forms exhibit similar slopes.

280

Fabry et al.

This method is particularly useful for measuring CSAT of Hbs that have a very high solubility and hence require a very high Hb concentration to detect polymerization. Under these circumstances, the classic CSAT becomes difficult to execute owing to high solution viscosity. The method is also useful when only small amounts of the Hb can be obtained. In a classic CSAT, about 200 mg of protein is needed for each measurement. If a Hemoscan (Aminoco, Silver Spring, MD) is used to determine p50, one needs only 2 µL for each determination, and 40–50 µL of a 40 g/dL solution of the sample (about 10–20 mg) is sufficient for the entire analysis. Hb samples are prepared in 100 mM phosphate buffer at pH 6.8. Since the solubility of a given sample will be unknown, samples are generally concentrated to about 40 g/dL. Centricon microconcentrators from Millipore are useful for this purpose. Oxygen affinity of samples is tested in the range of 6–40 g/dL. Generally, in the range of 6–16 g/dL, two or three points are chosen that will give the slope of the curve in the nonpolymerizing phase of the sample. In the range of 30–40 g/dL, another three points are selected to give the slope of the line in the polymerizing phase (see Note 8). For each concentration, about 10 µL of the diluted stock solution is prepared by mixing with buffer; two microliters are used for the Hemoscan measurement, 2 µL are used to determine the concentration of the solution by diluting 1/100 with the buffer and obtaining a spectrum from which the concentration is determined by the optical density (OD) at 540 nM, and the percentage of methemoglobin (MetHb) is determined by the OD at 630 nM. If the percentage of MetHb exceeds 4 to 5%, the entire sample is reduced with dithionite, the dithionite is removed by column chromatography, and the sample is reconcentrated. Hemoscan measurements are made by deoxygenating the sample, waiting 7 min, and then slowly reoxygenating the sample; this protocol minimizes the kinetic effects associated with polymerization. The approximate point at which p50 begins to increase rapidly is estimated visually. Data below this point are fitted by linear least squares regression, and data above this point are similarly fitted using a linear regression program. The intersection of the two lines is the concentration at which onset of polymer formation occurred and is hence the CSAT.

2.3. CSAT by P50 for Intact Red Cells Because CSAT is a measure of the onset of polymer formation, it may be correlated with the probability of vasoocclusion in humans and transgenic mice. However, not only is CSAT usually measured under nonphysiological conditions, but these measurements are also not sensitive to other important in vivo factors such as red cell MCHC and DPG content. Intracellular components that either do or may affect polymerization are generally not present because they

Solubility of Sickle Polymer

281

have been removed by dialysis. They can, of course, be added back, but accurate replication of intracellular conditions is difficult. To use the p50 method to measure CSAT in red cells, intracellular Hb concentration (MCHC) is varied by changing extracellular osmolarity (50). The MCHC at which onset of polymer formation occurs (the intracellular CSAT) is the point at which p50 begins to rapidly increase in a plot of p50 vs osmolarity. This is an estimate of the polymer-forming tendency of the Hbs in the red cell with all red cell components (such as DPG, antisickling Hbs, or added antisickling agents) present. Make high- and low-osmolarity buffers containing 10 mM HEPES, 5 mM KCl, and 5 mM glucose, and add NaCl to adjust the osmolarity to 160 or 450 mOsm. Titrate the buffers to pH 7.4 at 37°C and prepare intermediate osmolarities by mixing the two buffers and measuring osmolarity with a MicroOsmette (Precision Systems, Natick, MA) (see Note 9). Begin with 200–300 µL of cells suspended in isotonic buffer at hematocrit 50. This is the cell stock solution. Add 30 µL of the cell stock to 300 µL of buffer at the desired osmolarity. Centrifuge gently and suspend thoroughly. After the third wash remove 230 µL and set aside for the osmolarity measurement (see Notes 10 and 11). Remove aliquots for measurement of p50 using a Hemoscan (Aminoco) (see Note 12) on the slowest change in percent oxygen program to allow equilibration of polymer and solution phases. Hemoscan measurements are made by deoxygenating the sample, waiting 7 min, and then slowly reoxygenating the sample; this protocol minimizes the kinetic effects associated with polymerization. Three aliquots are also removed for determination of MCHC by measurement of hematocrit using a microhematocrit centrifuge (MicroHematocrit, Damon/IEF Division, Needham Heights, MA) and Hb concentration with Drabkin’s reagent (Sigma, St. Louis, MO). The approximate point at which p50 begins to increase rapidly is estimated visually. Data below this point are fitted by linear least squares regression, and data above this point are similarly fitted using the program Statgraphics Plus (Manugistics, Rockville, MD). The intersection of the two lines is the osmolarity at which onset of polymer formation occurred. This osmolarity is correlated with MCHC measured on cells equilibrated in the same buffer.

2.3.1. Interpretation of Measurements These measurements yield two values: the osmolarity and MCHC at which polymer formation begins. The latter is a measure of the effect of cell contents on polymer formation while the former relates both whole-animal physiology (plasma and renal osmolarity) and red cell physiology (factors affecting red cell density/ MCHC) to polymer formation. The onset of polymer formation in AS cells (Fig. 2) occurs at an osmolarity (330 mOsm) that is higher than physiological

282

Fabry et al.

Fig. 2. MCHC and p50 for representative AA, AS, and SS patients. (A) Measurement of intracellular CSAT by plotting p50 versus extracellular osmolarity for AA, SS, and AS red cells. (B) Determination of MCHC at the onset of polymer formation for SS red cells by plotting MCHC vs extracellular osmolarity. As expected, patient-topatient variation is observed (data not shown). The CSAT for SS red cells can also be estimated from the intersection of the SS p50 line with the horizontal line since the p50s of HbA and HbS are the same in the absence of polymer formation. The intracellular CSAT for this patient who had 9.7% HbF is 18.2 g/dL. The AS patient did not have α-thalassemia trait and had an estimated CSAT of 32.2 g/dL. Under fully deoxygenated conditions, the traditional method of measuring CSAT yields values of 15.8 g/dL for purified SS (5).

(280–300 mOsm), but within the range found in kidney. This is consistent with the relative clinical severity of AS (AS has only a urine-concentrating defect). Red cell heterogeneity is a potential complication for these measurements; however, because each cell acts like an independent container of well-defined solution, the effective 50 in the presence of red cell heterogeneity (MCHC or HbF) is a linear combination of all cell types present and therefore yields an average value for all cells present. This technique may be particularly suitable for measuring relative CSATs for evaluation of antisickling agents with mul-

Solubility of Sickle Polymer

283

tiple effects and comparing different strains of transgenic mice that may express complex mixtures of Hbs. 3. Notes 1. Use a final concentration of HbS that will be just sufficient to gel (e.g., 22 g/dL for the HbS standard). If the concentration is too high, there will not be enough supernatant to measure the pH and concentration. If the concentration is too low, it will not gel. 2. Since deoxyHbS at high concentrations will gel under the right conditions, you must keep everything ice cold at all times. Low temperature melts the polymer. The syringes used to measure the volumes of Hb to be added to the mixing tubes must be wrapped in small bags filled with ice. The tubes also must be on ice (cutoff fingers from disposable gloves work well for this purpose). Do not touch the tubes after chilling; the warmth may be sufficient to initiate polymer formation. If localized polymer formation occurs, it will not be possible to properly mix in all of the components. In general, it is not be possible to salvage samples that have gelled prematurely. 3. After filling the mixing tubes with Hb, buffer, or another component other than dithionite, they must be capped with a tight-fitting rubber stopper (cut off the excess sleeve part of the stopper to make it easier to work with). The space above the solution must be displaced with nitrogen or argon gas to get rid of any oxygen before the addition of dithionite. The gas should flow for 2 min through a 20-gage needle and flow out through a 23-gage needle. Always pull out the outlet needle before pulling out the gas needle. Immediately add the dithionite solution and mix the tube—which is kept in its ice “finger”—vigorously. Keep on ice until you can transfer the contents to the CSAT tubes. 4. The CSAT tubes should be filled with paraffin oil and the deoxy Hb added under the oil to prevent oxygen from getting in. The tubes should be placed in a container, capped with a rubber stopper and flushed with nitrogen, then stored overnight at the temperature you need until you can centrifuge them. 5. Quartz Electron Paramagnetic Resonance (EPR) tubes cut short are suitable for CSAT determinations. 6. To facilitate removal of the solution phase for concentration determinations, you must distinguish it clearly from the gel phase. To visualize the solution phase vs the gel phase, cut a slit about 1 mm wide in a white index card. Hold it against the CSAT tube with a high-intensity lamp behind it. This will make it easier to distinguish between the two phases. 7. Use positive displacement pipets for determining concentration. Use airtight gas syringes for measuring volumes of deoxyHb, buffer, and so on. Since Hb in solutions containing dithionite is converted to the met form on exposure to oxygen, exposure to oxygen after the sample has been centrifuged is avoided unless aliquots are taken for determination by the cyanmet method (Drabkin’s reagent). 8. Measurements made near the CSAT are frequently unreliable, owing to the long delay times at this concentration. Once the probable location of the CSAT has

284

Fabry et al.

been found, it is best to cluster the points at concentrations above the CSAT. 9. By measuring hematocrit, Hb, pH, and osmolarity while the next sample is equilibrating and scanning, it is possible to measure five to six osmolarities in a day. Intermediate points can be measured on the same blood sample that has been preserved in plasma on the next day; however, reliable measurements usually cannot be made on the third day. This is particularly true of red cells with pathology of any kind. 10. A stock solution of whole blood in isotonic buffer is maintained on ice and the cells are washed with the desired osmolarity just before use. Keeping cells for extended and variable periods of time under nonisotonic conditions may lead to other changes in the red cell. The cells in buffer can only be used for 1 d. The very small (2 µL) sample of red cells quickly reaches thermal equilibrium in the Hemoscan. 11. This value should be used rather than the nominal value of the buffer used to wash the cells, since it represents the final osmolarity in equilibrium with the cells. 12. Although a Hemoscan, an instrument no longer in commercial production, was used for these measurements and is very convenient because of the small sample size, any instrument capable of measuring p50, in a user-specified buffer, would be suitable; however, sample sizes would need to be scaled accordingly.

References 1. White, J. W. (1974) Ultrastructural features of erythrocyte and hemoglobin sickling. Arch. Intern. Med. 133, 545–562. 2. White, J. G. (1968) The fine structure of sickled hemoglobin in situ. Blood 31, 561–579. 3. Wishner, B. C., Ward, K. B., Lattman, E. E., and Love, W. E. (1975) Crystal structure of sickle-cell deoxyhemoglobin at 5 Å resolution. J. Mol. Biol. 98, 179–194. 4. Padlan, E. A. and Love, W. E. (1985) Refined crystal structure of deoxyhemoglobin. S. I. Restrained least-squares refinement at 3.0-A resolution. J. Biol. Chem. 260, 8272–8279. 5. Harrington, D. J., Adachi, K., and Royer, W. E. Jr. (1997) The high resolution crystal structure of deoxyhemoglobin S. J. Mol. Biol. 272, 398–407. 6. Nagel, R. L., Johnson, J., Bookchin, R. M., Garel, M. C., Rosa, J., Schiliro, G., Wajcman, H., Labie, D., Moo-Penn, W., and Castro, O. (1980) β chain contact sites in the hemoglobin S polymer. Nature 283, 832–834. 7. Nagel, R. L., Bookchin, R. M., Johnson, J., Labie, D., Wajcman, H., Isaac-Sodeye, A. W., Honing, G. R., Schiliro, G., Crookston, J. H., and Matsutomo, K. (1979) The structural bases of the inhibitory effects of Hb F and Hb A2 on the polymerization of Hb S. Proc. Natl. Acad. Sci. USA 76, 670–672. 8. Adachi, K., Kim, J., Ballas, S., Surrey, S., and Asakura, T. (1988) Facilitation of Hb S polymerization by the substitution of Glu for Gln at beta 121. J. Biol. Chem. 263, 5607–5610. 9. Ho, C., Willis, B. F., Shen, T. J., et al. (1996) Roles of alpha114 and beta 87 amino acid residues in the polymerization of hemoglobin S: implications for gene therapy. J. Mol. Biol. 263, 475–485.

Solubility of Sickle Polymer

285

10. Reddy, L. R., Reddy, K. S., Surrey, S., and Adachi, K. (1997) Role of beta87 Thr in the beta6 Val acceptor site during deoxy Hb S polymerization. Biochemistry 36, 15,992–15,998. 11. Li, X., Mirza, U. A., Chait, B. T., and Manning, J. M. (1997) Systematic enhancement of polymerization of recombinant sickle hemoglobin mutants: Implications for transgenic mouse model for sickle cell anemia. Blood 90, 4620–4627. 12. Edelstein, S. J., Telford, J. N., and Crepeau, R. H. (1973) Structure of fibers of sickle cell hemoglobin. Proc. Natl. Acad. Sci. USA 70, 1104–1107. 13. Nagel, R. L. and Bookchin, R. M. (1978) Areas of interaction in the HbS polymer, in Biochemical and Clinical Aspects of Hemoglobin Abnormalities (Caughey, W. S., ed.), Academic, New York, pp. 195–203. 14. Benesch, R. E., Kwong, S., and Benesch, R. (1982) The effects of alpha chain mutations cis and trans to the beta6 mutation on the polymerization of sickle cell haemoglobin. Nature 299, 231–234. 15. Crepeau, R. H., Edelstein, S. J., Szalay, M., Benesch, R. E., Benesch, R., Kwong, S., and Edalji, R. (1981) Sickle cell hemoglobin fiber structure altered by alphachain mutation. Proc. Natl. Acad. Sci. USA 78, 1406–1410. 16. Bookchin, R. M. and Nagel, R. L. (1973) Molecular interactions of sickling hemoglobins, in Symposium on Sickle Cell Disease (Abramson, H., Bertles, J., and Wethers, D., eds.), CV Mosby, St. Louis. 17. Nagel, R. L. and Bookchin, R. M. (1975) Mechanism of hemoglobin S gelation: structural restrictions to supramolecular models of the polymerization of hemoglobin S., in Sickle Cell Anemia and Other Hemoglobinopathies (Levere, R. D., ed.), Academic, New York, pp. 51–66. 18. Baudin-Chich, V., Pagnier, J., Marden, M., Bohn, B., Lacaze, N., Kister, J., Schaad, O., Edelstein, S. J., and Poyart, C. (1990) Enhanced polymerization of recombinant human deoxyhemoglobin beta 6 Glu→Ile. Proc. Natl. Acad. Sci. USA 87, 1845–1849. 19. Eaton, W. A. and Hofrichter, J. (1990) Sickle cell hemoglobin polymerization. Adv. Protein Chem. 40, 63–279. 20. Eaton, W. A. and Hofrichter, J. (1987) Hemoglobin S gelation and sickle cell disease. Blood 70, 1245–1266. 21. Briehl, RW. (1991) Sickle-cell hemoglobin. Encyclopedia Hum. Biol. 7, 1. 22. Canessa, M., Romero, J. R., Lawrence, C., Nagel, R. L., and Fabry, M. F. (1994) Rate of activation and deactivation of K:Cl cotransport by changes in Cell volume in hemoglobin SS, CC and AA red cells. J. Memb. Biol. 142, 349–362. 23. Brugnara, C. and Tosteson, D. C. (1987) Cell volume, K transport, and cell density in human erythrocytes. Am. J. Physiol. 252, C269–C276. 24. Goldberg, M. A., Husson, M. A., and Bunn, H. F. (1977) Participation of hemoglobins A and F in polymerization of sickle hemoglobin. J. Biol. Chem. 252, 3414–3421. 25. Minton, A. P. (1977) Non-ideality and the thermodynamics of sickle-cell hemoglobin gelation. J. Mol. Biol. 110, 89–103. 26. Bookchin, R. M., Nagel, R. L., and Balazs, T. (1975) Role of hybrid tetramer formation in gelation of haemoglobin S. Nature 256, 667–668.

286

Fabry et al.

27. Ross, P. D., Hofrichter, J., and Eaton, W. A. (1977) Thermodynamics of gelation of sickle cell deoxyhemoglobin. J. Mol. Biol. 115, 111–134. 28. Ross, P. D., Hofrichter, J., and Eaton, W. A. (1975) Calorimetric and optical characterization of sickle cell hemoglobin gelation. J. Mol. Biol. 96, 239–253. 29. Magdoff-Fairchild, B., Poillon, W. N., Li, T., and Bertles, J. F. (1976) Thermodynamic studies of polymerization of deoxygenated sickle cell hemoglobin. Proc. Natl. Acad. Sci. USA 73, 990–944. 30. Adachi, K., Ozguc, M., and Asakura, T. (1980) Nucleation-controlled aggregation of deoxyhemoglobin S: participation of hemoglobin A in the aggregation of deoxyhemoglobin S in concentrated phosphate buffer. J. Biol. Chem. 255, 3092–3099. 31. Poillon, W. N. and Bertles, J. F. (1979) Deoxygenated sickle hemoglobin: Effects of lyotropic salts on its solubility. J. Biol. Chem. 254, 3462–3467. 32. Ueda, Y., Nagel, R. L., and Bookchin, R. M. (1979) An increased Bohr effect in sickle cell anemia. Blood 53, 472–480. 33. Swerdlow, P. H., Bryan, R. A., Bertles, J. F., et al. (1977) Effect of 2, 3-diphosphogylcerate on the solubility of deoxy sickle hemoglobin. Hemoglobin 1, 527–537. 34. Poillon, W. N., Kim, B. C. (1990) 2,3-Diphosphoglycerate and intracellular pH as interdependent determinants of the physiologic solubility of deoxyhemoglobin S. Blood 76, 1028–1036. 35. Hofrichter, J., Ross, P. D., and Eaton, W. A. (1974) Kinetics and mechanism of deoxyhemoglobin S gelation: a new approach to understanding sickle cell disease. Proc. Natl. Acad. Sci. USA 71, 4864–4868. 36. Hofrichter, J. (1986) Kinetics of sickle hemoglobin polymerization. III. Nucleation rates determined from stochastic fluctuations in polymerization progress curves. J. Mol. Biol. 189, 553–571. 37. Samuel, R. E., Salmon, E. D., and Briehl, R. W. (1990) Nucleation and growth of fibres and gel formation in sickle cell haemoglobin. Nature 345, 833–835. 38. Ferrone, F. A., Hofrichter, J., Sunshine, H. R., and Eaton, W. A. (1980) Kinetic studies on photolysis- induced gelation of sickle cell hemoglobin suggest a new mechanism. Biophys. J. 32, 361–380. 39. Harris, J. W. (1950) Studies on the destruction of red blood cells. VII. Molecular orientation in sickle cell hemoglobin solutions. Proc. Soc. Exp. Biol. Med. 75, 197–201. 40. Coletta, M., Hofrichter, J., Ferrone, F. A., and Eaton, W. A. (1982) Kinetics of sickle haemoglobin polymerization in single red cells. Nature 300, 194–197. 41. Goldberg, M. A., Lalos, A. T., and Bunn, H. F. (1981) The effect of erythrocyte membrane preparations on the polymerization of sickle hemoglobin. J. Biol. Chem. 256, 193–197. 42. Mozzarelli, A., Hofrichter, J., and Eaton, W. A. (1987) Delay time of hemoglobin S polymerization prevents most cells from sickling in vivo. Science 237, 500–506. 43. Horiuchi, K., Ballas, S. K., and Asakura, T. (1988) The effect of deoxygenation rate on the formation of irreversibly sickled cells. Blood 71, 46–51.

Solubility of Sickle Polymer

287

44. Horiuchi, K. and Asakura, T. (1989) Oxygen promotes sickling of SS cells. Ann. NY Acad. Sci. 565, 395–397. 45. Hofrichter, J., Ross, P. D., and Eaton, W. A. (1976) Supersaturation in sickle cell hemoglobin solutions. Proc. Natl. Acad. Sci. USA 73, 3035–3039. 46. Adachi, K. and Asakura, T. (1979) The solubility of sickle and non-sickle hemoglobins in concentrated phosphate buffer. J. Biol. Chem. 254, 4079–4084. 47. Adachi, K. and Asakura, T. (1979) Nucleation-controlled aggregation of deoxyhemoglobin S: Possible difference in the size of nuclei in different phosphate concentrations. J. Biol. Chem. 254, 7765–7771. 48. Bookchin, R. M., Balazs, T., Wang, Z., Josephs, R., and Lew, V. L. (1999) Polymer structure and solubility of deoxyhemoglobin S in the presence of high concentrations of volume-excluding 70-kDa dextran. J. Biol. Chem. 274, 6689–6697. 49. Benesch, R. E., Edalji, R., Kwong, S., and Benesch, R. (1978) Oxygen affinity as an index of hemoglobin S polymerization: a new micromethod. Anal. Biochem. 89, 162–173. 50. Fabry, M. E., Desrosiers, L., and Suzuka, S. M. (2001) Direct intracellular measurement of deoxygenated HbS solubility (CSAT). Blood 98, 883–884.

288

Fabry et al.