THE OF CHEMISTRY Vol. 255, No. 19, hue of October 10, pp ... › content › 255 › 19 › 9465.full.pdf · THE JOURNAL OF BIOLOGICAL CHEMISTRY Printed in U.S.A. Vol. 255, No. 19,

THE JOURNAL OF BIOLOGICAL CHEMISTRY

Printed in U.S.A. Vol. 255, No. 19, h u e of October 10, pp. 9465-9473,1980

Structural and Functional Studies of Hemoglobin Suresnes (Arg 141a2 - His p2) CONSEQUENCES OF DISRUPTING AN OXYGEN-LINKED ANION-BINDING SITE*

(Received for publication, December 18, 1979, and in revised form, April 4, 1980)

Claude PoyartS and Elisabeth Bursaux From the Znstitut National de la Sante et de la Recherche Medicale (Unite 23 , 42, rue Desbassayns de Richemont, 92150 Suresnes, France Arthur Amone From the Department of Biochemistry, University of Iowa School of Medicine, Iowa City, Iowa 52240

Joseph Bonaventurag and Celia Bonaventura From the Marine Biomedical Center, Duke University Marine Laboratory, Beaufort, North Carolina 28516

Hb Suresnes is a human hemoglobin variant in which histidine replaces arginine at the COOH terminus of the a chains. The COOH-terminal arginines of the a chains play a major role in normal human hemoglobin (HbA), both by electrostatic interactions which con- strain deoxygenated hemoglobin in a low affinity quaternary conformation and by involvement in oxygen- linked anion binding known to give rise to a large part of the alkaline Bohr effect. An x-ray crystallographic analysis of deoxyHb Suresnes reveals the loss of the normal intersubunit salt bridge to lysine 127a and a decrease in the occupancy of inorganic anions at the a chain anion-binding site. Relative to normal HbA, Hb Suresnes has a high affinity for oxygen, a reduced cooperativity in oxygen binding, and greatly reduced pH and chloride sensitivity. Similar changes are found in carboxypeptidase B-digested hemoglobin AO where the COOH-terminal arginine is removed by enzymatic digestion. Both equilibrium and kinetic manifestations of the destabilization of the normal low affinity configuration are observed. Inositol hexaphosphate restores cooperativity and pH sensitivity but, even in the presence of this strong allosteric effector, the low and high affinity conformations of Hb Suresnes appear to differ from those of HbA. In both the unliganded and liganded states, appreciable subunit dissociation of Hb Suresnes is apparent, which suggests that the substitution dis- rupts bonds which normally stabilize the tetramer. Structural changes at both the quaternary and tertiary level appear to contribute to alterations in the high and low affinity conformations of this abnormal hemoglobin. Hb Suresnes thus illustrates the consequences of the loss of an important oxygen-linked anion-binding site.

* This work was partly supported by Grant CRL 78-5-0585 from the Institut National de la Sante et de la Recherche Medicale, Grants HL 15.460 and ESO 1908 from the National Institutes of Health, and Grant PCM 77-08453 from the National Science Foundation. The crystallographic research reported here was supported by Grant AM- 17563 from the National Institutes of Health and by a National Institutes of Health Research Career Development Award (to A. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. * To whom correspondence should be addressed.

0 Established Investigator of the American Health Association.

The COOH-terminal arginine of the a chain of normal human hemoglobin forms salt bridges that are critical to the maintenance of normal functional properties (1-4). Addition- ally, a number of experiments suggest an involvement of the NH2- and COOH-terminal residues of the a chain in oxygen- linked anion binding (5-7). It was, consequently, of interest when a new human hemoglobin variant was discovered with a substitution involving the COOH-terminal arginine of the (Y chain. This mutant, Hb Suresnes, was determined to have histidine in place of arginine at position 141 of the a chain (8). The hemoglobin was shown in preliminary studies to have a high oxygen affinity, a reduced Bohr effect, and reduced homotropic interactions in oxygen binding (8). The heterozygous subjects, whose erythrocytes contained approximately 42% of the abnormal component, presented a slight erythro- cytosis but were otherwise free of clinical symptoms. Hb Suresnes has also been reported to occur in another family (9). The preliminary studies of Poyart et al. (8) and Gravely et al. (9) suggested that the functional properties of Hb Suresnes were similar in many respects to those of CPB-HbA’ (1-4) in which the COOH-terminal arginine of the a chain has been removed enzymatically. The following concerns a detailed study of the functional properties of Hb Suresnes and a comparison of its properties with those of carboxypeptidase B-digested hemoglobin. An x-ray crystallographic analysis of deoxyHb Suresnes is also presented which illustrates the degree of structural perturbation which is brought about by this substitution.

MATERIALS AND METHODS

Structural Studies-Crystals of deoxyhemoglobin Suresnes and deoxyhemoglobin A were grown from phosphate-buffered (pH 6.5) solutions of ammonium sulfate according to the procedure described by Perutz (10). Single crystals of native or abnormal deoxyhemoglobin were mounted in quartz capillary tubes in an atmosphere of high purity nitrogen. The diffraction data were then collected to a resolu-

The abbreviations used are: CPB-HbA, carboxypeptidase B-digested hemoglobin Ao; HbA, human hemoglobin Ao; MetHb, ferri- hemoglobin; p50 and P C O ~ , the partial preasure of oxygen or carbon monoxide necessary to half-saturate hemoglobin; y. the fractional saturation of hemoglobin with ligand; h, the Hill coefficient for heme- heme interaction; K1 and K,, the apparent equilibrium constant of oxygen binding at very low and very high levels of saturation; L4. the apparent equilibrium constant of carbon monoxide binding at very high saturation levels; DPG, 2,3-bisphosphoglycerate; IHP, inositol hexaphosphate (Na salt); bis/Tris, [bis(2-hydroxyethyl)amino] tris(hydroxymethy1)methane.

9465

9466 Structure and Function of Hemoglobin Suresnes

tion of 3.5 A (7000 independent reflections) on an Enraf-Nonius CAD4 diffractometer using the w scan mode. Two crystals were used for the native deoxyhemoglobin data set, but only one crystal was required for the deoxyhemoglobin Suresnes data set. Degradation due to radiation damage never exceeded 8% as determined by repeated measurements of four standard reflections. An empirical correction for radiation damage was incorporated into the crystal scaling calculations as described previously (11). The absorption correction method of North et al. (12) was applied. A difference electron density map was calculated using the known phases of deoxyhemoglobin A (11) and the difference amplitudes ( [ F ] Hb Suresnes - [1;1 HbA) and then symmetry averaged about the molecular 2-fold rotation axis that relates the two equivalent a/3 dimers (13). In order to produce a clearer image of the imidazole side chain of histidine 141q we also calculated a 3.5 A difference map of deoxyhemoglobin Suresnes- deoxyCPB-HbA using the deoxyhemoglobin A phases.

Equilibrium Studies-Erythrocytes from heterozygous donors were washed and lysed and the hemolysate was extensively dialyzed prior to ion exchange chromatography. Separation of HbA and Hb Suresnes was accomplished using columns of DEAE-Sephadex and a pH gradient from 8.0 to 7.2 in Tris-HCI buffer (14). Purity of the isolated hemoglobins was verified by isoelectric focusing. The MetHb content of the final stock solutions was less than 2% according to usual spectrophotometric criteria. CPB-HbA was prepared by a digestion of oxyhemoglobin A with carboxypeptidase B (Worthington, diisopropyl fluorophosphate-treated) as described in Ref. 2. The extent of the digestion was checked by amino acid analysis, which showed 0.9 to 1.2 arginine residues released per ab dimers in three different preparations. After the digestion, carboxypeptidase B was removed from the CPB-Hb solution by chromatography (DE52). The modified Hb was further analyzed by thin layer electrofocusing (LKB Multiphor, Ampholines pH range 6 to 8, LKB, Bromma, Sweden), which displayed one single band migrating 0.03 pH unit more anodic than the unmodified HbA.

Oxygen-binding measurements were performed according to a dis- continuous equilibrium technique as described in detail elsewhere (14). Unless otherwise specified, pTKr measurements were carried out a t 25°C with hemoglobin solutions 0.50 to 0.60 mM in heme in 50 m bis/Tris or Tris buffer solutions containing 0.1 M Na chloride. The spectrophotometer was either a Pye Unicam Philips SP 1800 or a Cary model 219.

Carbon monoxide-binding experiments were performed in the same way in hemoglobin solutions 0.8 to 0.9 mM in heme. Absorbances were read a t 554.5 and 538.3 nm and also a t 546.5 nm, an isosbestic point between the deoxy- and fully carboxyHb spectra. Due to the low pro necessary to saturate Hb, each point of the CO-binding curve was equilibrated for at least 45 min in a rotating thermostated water bath a t 25°C. Deoxygenated sodium dithionite solution was added initially to the deoxygenated Hb solution to a final concentration of 1.0 mM. Additions of the gas were performed from either a gas mixture containing 1% CO in argon or from pure CO. The MetHb formation during the 5-h experiment was insignificant as estimated from the ratio A , i/Ai:lx I. For the calculations of the ~ C O , the amount of CO bound to hemoglobin was substracted from the amount of CO added. Lq was estimated by extrapolation to the log P axis of the line with a slope of unity and corresponding to the experimental points.

The compositions of the different media used in these studies are given in the legends to the figures and tables. All reagents were of analytical grade. IHP was used as its sodium salt and diluted in the same buffer and at the same pH as that of the hemoglobin solutions. The gases used in these studies were of the purest grade available commercially.

P,, and the Hill coefficient were estimated from the Hill equation. The free energy of ligand binding to Hb was calculated from

AG = RT ln[O~].,~~ (1)

where AG is the Gibbs free energy of interaction in kcal mol". The difference between the AG in the presence of an effector to the free energy in its absence (AGx = ") gave the AAG which represents an estimate of the variation in the free energy of oxygen binding due to the addition of the effector.

In Equation 1, [02], is the molar concentration of oxygen necessary to half-saturate Hb. The free energy of heme-heme interaction (AF) was calculated from the values of the fmt and fourth association constants as

A F = 4 R T l n Kq kcal mol" (K) (2)

The curves relating logpso to pH were submitted to an iterative curve fitting procedure using a 4' polynomial regression and a standard algorithm for the minimization of the residuals squared (15). The interpolated values were obtained from these curves. The first derivative of this function represents the Bohr effect a t any pH and salt concentration

where pH and (A) indicate constant pH and free anion concentration in the solutions (16). These calculations were performed with an IBM 370-138 computer.

Kinetic Studies-Rapid mixing experiments were performed with a Gibson-Dunum stopped flow apparatus, equipped with a pneumatic drive. The wavelength calibration was checked using the 486.13 and 656.28 nm lines from a deuterium lamp. The apparatus has a dead time of about 2.3 ms. Flash photolysis experiments were performed using dual fast extinguishing (-30 ps) flash tubes and a Xenon Corp. model B Micropulser. The discharge energy was 400 J and the time required for total recovery of the EM1 photomultiplier was less than 1 ms. Sodium dithionite (Merck) was used a t a concentration of about 0.5% to deoxygenate hemoglobin in kinetic experiments. Data collec- tion and analysis were facilitated by use of a data acquisition and storage device (DASAR, American Instrument Co.) and a PDP 11/34 minicomputer (Digital Equipment Corp.)

RESULTS

X-ray Analyses-A sketch of the environment of arginine 141a in HbA is illustrated in Fig. 1 and the native electron density map is shown in Fig. 2A. A portion of the 3.5 A difference map of deoxyhemoglobin Suresnes-deoxyhemoglobin A is shown in Fig. 2, B and C. The strongest feature of the map is an extended negative peak which superimposes over the length of the native electron density for a chain COOH- terminal dipeptide. A region of strong positive difference electron density borders the inner surface of the negative peak (ie. in the direction of the central cavity), while a region of weaker positive density runs along part of its outer surface. Taken together, these features show that (relative to deoxyhemoglobin A) the COOH-terminal group and the penulti- mate residue, tyrosine 140a, are shifted toward the central cavity by as much as 3 8, in deoxyhemoglobin Suresnes, while the imidazole group of histidine 141a overlaps the position of

o c Q N 00

o.-NH,

L I

FIG. 1. Sketch showing the environment of arginine 141~x1 in human deoxyhemoglobin. - - -, salt bridges (1, 24). The COOH terminus of the al chain is labeled al-COO-, and the NH? terminus of the a2 chain is labeled a.l-NH:i'. XC and X:I- represent inorganic anion binding sites. For clarity, only selected side chains are shown, and the peptide backbone between the first and fifth residues of the mp chain has been deleted. In deoxyhemoglobin Suresnes, histidine 141al does not form salt bridges with aspartate 1 2 6 ~ and lysine 1 2 7 ~ and the X.1- anion is lost.

Structure and Function of Hemoglobin Suresnes 9467

FIG. 2. Fourier maps showing contiguous electron density sections which include residues arginine 141a. lysine 127a. tyrosine 140a, and leucine 136a, the H helices, F helices, and heme groups of the a chains, as well as a small part of the al/ll

and a2Pz interfaces. The feature labeled S I corresponds to an inorganic anion shown in Fig. I . All features occur in pairs related by the molecular 2-fold axis of symmetry (only one member of each has been labeled). and the maps have been averaged about this axis. The symmetrv axis is perpendicular t o the plane of the paper and passes through the center of each figure. A, composite of six two-dimensional sections (spaced 1 A apart) of the native human deoxyhemoglobin electron density map at 2.5 A resolution (sections +15 and + I 0 A) . The contours are drawn at intervals of 0.15 eA above the zero level with the upper contours obscuring the lower ones. B, difference electron density at 3.5 A resolution (section +I4 to +I2 A) of deoxyhemoglobin Suresnes (white contours) superimposed on the native electron density (hlnch contours). Positive difference electron density is marked by solid uhite contours and upper case letters; negative

its guanidinium counterpart in deoxyhemoglobin A. The large movement of the COOH-terminal group results in the loss of i t s normal intersubunit salt bridge with lysine 127q but places it closer to the a-amino group of valine la on the adjacent a chain so that an interaction between these two groups may occur in Hb Suresnes.

A number of weaker features (many of which are not shown in Fig. 2 ) can also be seen in the deoxyhemoglobin Suresnes difference map. A series of alternating negative and positive peaks extend from the COOH-terminal dipeptide along the entire length of the H helix, showing that this helix shifts in the general direction of its NH2-terminal end and toward the central cavity. From a similar series of positive and negative

difference electron density is marked hy broken tthite contours and lower case letters. The difference electron density map is contoured at intervals of kO.10 eA I. Negative difference Ped: a superimposes on the native electron density for the n chain COOH-terminal groups and adjacent backbone atoms. Positive difference electron density, labeled as Penks A, straddles Peaks n and shows the COOH terminus of deoxyhemoglobin Suresnes to be disordered to a small degree with its equilibrium position shifted by 2 to 3 A toward the central cavity and away from lysine 127n. The pair of Penlrs B and b show a shift of leucine 1360 away from the n-heme group. Negative Peak c shows the displacement of inorganic anion X., . C. section +12 to 10 A of the deoxyhemoglobin Suresnes difference map. Negative difference Penh n superimposes over the length of the native electron density for the side chain of arginine 141n and is bordered on hoth sides by positive difference electron density. labeled as Penks A . This indicates that the imidazole group of histidine 141n overlaps the position of i t s guanidinum counterpart in deoxyhemoglobin A. Peaks B and b shows a large shift in the position of tyrosine 140n toward the central cavity.

peaks, it is also clear that the F helix moves in a direction perpendicular to i t s major axis and toward the H helix. This in turn may result in a small movement of the a-heme group away from the E helix. The evidence for the latter movement (not shown in Fig. 2) includes an obvious shift of the heme- linked histidine (histidine 87a) , low level negative difference density between the E helix and the a-heme, and positive difference density between the a-heme and the H helix.

A weak negative difference peak is also observed at the position normally occupied by inorganic anion X:)- (Figs. 1 and 2A) . This implies a decrease in the occupancy of sulfate (or phosphate) anions a t this site in the high salt (2.3 M sulfate and 0.3 M phosphate) crystals of deoxyhemoglobin Suresnes.


TABLE I Effects of anionic cofactors on oxygen binding by Hb Suresnes,

CPB-HbA and HbA The hemoglobins were in 0.05 M bis/Tris at a protein concentration

of 0.5 mM in heme, except CPB-HbA, which was 0.25 mM in heme (pH 7.15 and 25°C). The p s ~ value of stripped Hb was measured in 0.02 M bis/Tris and 5 m~ chloride. AG is the Gibbs free energy of oxygen binding to the protein. M G is the difference in AG with and without the effector. hm is the index of heme-heme interaction a tpa . Hbx relates to Hb Suresnes or CPB-HbA.

Protein and cofactor P n mm p n

Hbx Hg h:n AG AAG HbA/p*

kcal, mol-

Hb Suresnes Stripped 0.40 1.42 -33.1 4.8 + 0.1 M NaCl 0.70 1.52 -32.1 1.0 8.2 + 1.5 m~ IHP 9.70 1.76 -25.8 7.3 7.0 + 0.1 M NaCl + 1.5 mM 7.20 1.90 -26.6 6.5 9.4

IHP

CPB-HbA Stripped 0.47 1.18 -32.0 4.0 + 0.1 M NaCl 0.66 1.75 -32.2 0.8 8.7 + 1.5 m~ IHP 6.30 1.75 -26.9 6.15 10.7 + 0.1 M NaCl + 1.5 mM 3.20 2.0 -28.5 4.5 21.0

IHP

HbA Stripped 1.90 2.8 -29.7 + 0.1 M NaCl 5.75 2.9 -27.1 2.6 + 1.5 m~ IHP 67.60 2.0 -21.3 8.5 + 0.1 M NaCl + 1.5 mM 67.60 1.8 -21.3 8.5

IHP

The position of histidine 141a is partially obscured in the deoxyhemoglobin Suresnes-deoxyhemoglobin A difference map by the negative peak which results from the loss of the arginine side chain. However, by calculating a deoxyhemoglobin Suresnes-des(arginine 141a)-deoxyhemoglobin difference map (not shown) we obtained a much clearer image of histidine 141a. The dominant feature of this map, an intense positive difference peak, shows the imidazole ring to be ori- ented parallel to the axis of the H helix and about 1 A further away from the carboxyl group of aspartate 126a than the guanidinium side chain is in deoxyhemoglobin A.

Functional Studies a t Equilibrium-Table I summarizes the effects of anionic effectors on oxygen binding by Hb Suresnes, CPB-HbA, and HbA. In the absence of any effector, both Hb Suresnes and CPB-HbA evidence a 4- to 5-fold larger affinity for oxygen than HbA. One may note also the large difference observed in the effects of anions on the two abnormal hemoglobins compared to HbA. The effect of chloride is about halved as shown by the change in the fiee energy of oxygen binding with and without 0.1 M chloride. The reduced effect of chloride on oxygen binding by Hb Suresnes and CPB- HbA relative to HbA is shown in Fig. 3, which illustrates chloride titrations at 25°C for the three proteins. Comparison of the slopes of the tangents to these curves at 0.1 M chloride gives an estimate of the amount of oxylabile chloride bound to these different hemoglobins. We find that 6 log p d 6 log (Cl-) was 0.5, 0.25, and 0.14 mol/mol of heme oxygenated for HbA, Hb Suresnes, and CPB-HbA, respectively.

The abnormal pH dependence of oxygen binding by Hb Suresnes is illustrated in Fig. 4. At low concentration of chloride (~0 .01 M), the alkaline Bohr effect in Hb Suresnes is completely abolished. It is incompletely restored upon addition of 0.1 M chloride. In the presence of IHP, the pH dependence of oxygen binding in Hb Suresnes appears very similar to that observed in HbA. The Bohr curves corresponding to

the variations of log pso with pH in Hb Suresnes and HbA are shown in Fig. 5A. These demonstrate that in the presence of IHP the alkaline Bohr effect in Hb Suresnes is very close to that observed in HbA. The peak of the two Bohr curves is displaced by the effector to the right with a maximum at pH 8. Fig. 5B illustrates the additional Bohr effect due to IHP binding. The two curves are superimposed except in the alkaline pH range. In this region, it appears that the amount of protons taken up by deoxyHb Suresnes upon binding IHP is larger than in HbA.

In order to further characterize the abnormal function of Hb Suresnes, it was of interest to evaluate the anion effects on ligand binding at very low and very high saturation, corresponding to the “T” and “R” quaternary configuration. K I ,

1 .o

0.5

0

- 0.I

Hb Surrrnes

0

FIG. 3. Titration curves with chloride for oxygen binding to Hb Suresnes, HbA, and CPB-HbA at pH 7.0, 25°C. Heme concentration was 0.5 mM for HbA and Hb Suresnes and 0.25 mM for CPB-HbA. Proteins were diluted in 0.05 bis/Tris buffer except for the fmt point of the curve where 0.02 M bis/Tris was used plus 5 m~ total chloride.pso values expressed in mm Hg and chloride concentrations are molar.

FIG. 4. The pH dependence of the oxygen affinity of Hb Suresnes and HbA at 25°C in 0.02 to 0.05 m bis/Tris or Tris buffer (for pH values greater than 7.6). 0, 5 mM chloride; 0, 100 mM chloride; ., 1.5 mM IHP.


0

1.0,

OI

FIG. 5. The effect of IHP on the alkaline Bohr effect of Hb Suresnes and HbA. A, calculated Bohr curves for Hb Suresnes and HbA, with and without 1.5 nm IHP at 25°C in 0.05 M bis/Tris or Tris buffer, 0.1 M chloride. B, differential proton bin- with and without 1.5 mM IHP as a function of pH. These curves were obtained by difference of the Bohr curves shown in A. -, Hb Suresnes. Calcu- lations are described in the text.

TABLE I1 Effects of anionic cofactors on oxygen binding by Hb Suresnes and

HbA at the lower part of the oxygen binding curue KI is the value of the first equilibrium constant of oxygen binding

to the tetrameric Hb. KI could not be measured in stripped Hb Suresnes as the slope relating log y/1 - y to log PO? repetitively leads to values of 1.2 to 1.3 even for saturation levels lower than 0.8%. Experiments were carried out at pH 7.15 and 25°C.

Protein and cofactor Kt mm Hg" KI Hb Suresnes/ K, HbA

Hb Suresnes Stripped + 0.1 M NaCl 0.90 + 0.4 M NaCl 0.42 + 1.5 mM IHP 0.03

45 28 5

HbA Stripped 0.072 + 0.1 M NaCl 0.020 + 0.4 M NaCl 0.015 + 1.5 mM IHP 0.006

the binding constant at low saturation, and K4, the binding constant at high saturation, were determined spectrophoto- metrically from precise measurements of the small changes in absorbance induced after addition of discrete amounts of oxygen at saturation levels lower than 1.5% ( K I ) or from 98.5 to 99.5% (K4). The details of the method have been described (17). For the K1 determinations, PO2 changes were produced by additions of aliquots of 1 to 5% oxygen diluted in argon (gas mixing pump, Wosthoff M 100 A, Bochum, West Ger- many). When Hb Suresnes was studied, extremely low POz levels (0.01 mm Hg) were sufficient to obtain the 1% saturation level. In these circumstances, the equilibration time for each point was at least 30 to 35 min. For the K4 determinations, POZ changes were produced by equilibrating the hemoglobin solutions with water/vapor/saturated gas mixtures at PO, values ranging from 40 to 200 mm Hg and in the presence of an enzymic reducing system (18). Log Kl and log K4 and the slopes of the two asymptotes were calculated from the linear regression of log y/1- y uersus log PO,. The effects of chloride and IHP on KI and Hb Suresnes are compared to those in HbA shown in Table I1 and Fig. 6. In the presence of 0.1 or

0.4 M chloride, the affinity for oxygen of Hb Suresnes at low saturation is 45 and 28 times larger, respectively, than that of HbA. By comparison, the ratios of thep, for the two proteins are 9 and 12, respectively. The results shown in Table I1 and Fig. 6 indicate that, even in this noncooperative region of ligand binding, chloride and IHP are effective in bringing about changes in oxygen affinity. In the presence of IHP, the oxygen aftinity of Hb Suresnes at low saturation is consider- ably increased but remains 6 times lower than in HbA. One may also note that the absolute changes in K 1 in Hb Suresnes brought about by anion binding are less than in the case of normal human hemoglobin.

The values of K4 for Hb Suresnes and HbA are given in Table 111. In the presence of 0.1 M chloride and at neutral pH, Hb Suresnes in the noncooperative region of ligand binding at high saturation ( R state) has a higher affinity for oxygen than

0

- 0.5

- 1.0

- 1.5

- 2.0

- 2.5

b

FIG. 6. Hill plots at very low saturation for the estimation of the first equilibrium constant of oxygen to Hb Suresnes and HbA. Conditions: 0.6 to 0.8 mM heme, 0.05 M bis/Tris, pH 7.15, at 25°C. Closed symbols, 0.4 M chloride; open symbols, 1.5 m IHP, 10 nm chloride; upper arrows, values of -log KI.

TABLE 111 Oxygen binding parameters measured in Hb Suresnes and in HbA

Conditions: pH 7.0,25"C, 0.1 M chloride, and 0.05 M bis/Tris buffer. K,, and p s ~ are the same as described in the legend to Table 11. K4 is the value of the equilibrium constant of oxygen to the tetrameric hemoglobin at very high saturation. A F is the free energy of heme- heme interaction. Hill coefficients h, the indices of cooperativity, are given in parentheses. The coefficient of variation ({I SD) /m X 100) for the values of the K I and K4 were 10 and 15% respectively (means of four experiments).

K l mm Hg" p,mmHg K4mmHg-' AF

mol- kcall

Hb Suresnes 0.90 0.77 3.5 3.3 (h1 0.98) (hm 1.6) (h, = 0.99)

HbA 0.02 6.8 2.6 11.6 (hl 1.02) (hm 2.95) (h, = 0.99)


TABLE IV Effects of anionic cofactors on carbon monoxide binding to Hb

Suresnes and HbA solutions Conditions: pH 6.7, 25”C, 0.1 M chloride, 0.05 M bis/Tris buffer a t

0.6 to 0.8 m~ heme concentration. L4 is the value of the equilibrium constant of CO to the tetrameric hemoglobins at very high saturation (see text). Values are the means of four different experiments. The coefficient of variation ((SD)/m X 100) was 20% in Hb Suresnes and 30% in HbA with IHP.

Protein and cofactor L, mm Hg” pco,, mm Hg (pco,, Hb Su- (Pco,, HbA)/

resnes)

Hb Suresnes + 0.1 M NaCl 543 0.011 6.5 + 0.1 M NaCl + 1.5 200 0.057 6.75

mM IHP

HbA + 0.1 M NaCl 272 0.072 + 0.1 M NaCl + 1.5 61 0.38

mM IHP

- 2:o - 1:o 0 1 :o

FIG. 7. Hill plots at very high saturation for the estimation of the fourth equilibrium constant of carbon monoxide to Hb Suresnes and HbA. Conditions, 0.8 to 0.9 m~ heme, 0.05 M bis/Tris, pH 6.7,O.l M chloride at 25°C. Circles, Hb Suresnes; squares, HbA; closed symbols, without IHP; open symbols, with 1.5 mM IHP; - - -,

log Pco mmHg

-log K4.

HbA. From the values of the two asymptotes, one may estimate the free energy of heme-heme interaction for the two proteins. The values are given in the last column of Table 111. The AF in Hb Suresnes represents about one third of the free energy of heme-heme interaction calculated in HbA. This is mainly due to the large increase in the KI value of Hb Suresnes. If one assumes a free energy of about 1.5 to 1.7 kcal mol”/salt bridge involved in the stability of the T quaternary structure (19), it appears that only two of the six salt bridges normally present in HbA are still contributing to the stabili- zation of Hb Suresnes in the deoxy configuration.

The effect of IHP on the ligand affinity of Hb Suresnes and HbA in their high affinity state were studied with carbon monoxide as the ligand. This permits full ligation of HbA in the presence of IHP. The IHP effects on the R state of both proteins are given in Table IV and Fig. 7 and are compared with the IHP effects on the values ofpco,. At physiological ionic strength, the affinity for CO of Hb Suresnes in its R state is approximately twice that of HbA. In the presence of IHP, the affinity for CO of the two proteins is much reduced but

that of Hb Suresnes remains higher than that of HbA. It is also worth noting from the changes in thepco,, given in Table IV that for both Hb Suresnes and HbA the effect of IHP on CO binding appears reduced compared to its effect on oxygen binding (Table I).

Another aspect of the abnormality of Hb Suresnes was observed with UV difference spectroscopy. The difference spectrum of deoxy- uersus oxyHbA that we obtained was identical with that described by Perutz et al. (20). Fig. 8 shows the deoxy uersus oxy UV spectrum for Hb Suresnes (-). Addition of IHP into the deoxy cuvette results in spectral perturbations associated with aromatic amino acids as shown by the dashed line. This indicates clearly that in the Hb Suresnes difference spectrum the peaks at 280 and 290 nm were reduced or absent compared to HbA. These were restored by IHP.

Oxygen Kinetics-The process of oxygen dissociation from Hb Suresnes was studied by stopped flow spectrophotometry, monitoring the spectral change which occurs when an air- equilibrated hemoglobin solution is rapidly mixed with an equal volume of buffer containing sodium dithionite. Fig. 9 illustrates time courses of oxygen dissociation from HbA and Hb Suresnes at pH 7.0 and at pH 8.5. As is evident from Fig. 9, the overall rate of the oxygen dissociation reaction is slower and shows less pH sensitivity for Hb Suresnes than for HbA. The time courses shown for Hb Suresnes can be described in

I 260 280 3da 320 g0 360 hnm

FIG. 8. W difference spectra of deoxy- versus oxyHb Su- resnes at 25”C, pH 7.0, in 0.05 M bis/Tris buffer, 0.1 M chloride.. Na dithionite was added to the deoxy samples to insure complete deoxygenation and to record the base line. The reference cuvette was then blown and equilibrated with pure oxygen. -, deoxy versus oxy difference spectrum; - - -, deoxy versus oxy difference spectrum after the addition of deoxygenated IHP (1.5 mM final) to the deoxy cuvette. Circles, significant differences induced upon addition of IHP to deoxyHb Suresnes. Inset, comparison of the deoxy uersus oxy spectra of HbA and Hb Suresnes in the absence of IHP. Proteins were at a concentration of 60 PM heme. Extinction coefficients (e) are in moles of heme.

FIG. 9. Normalized time courses of oxygen dissociation from Hb Suresnes and HbA in the absence of allosteric effectors at pH 7.0 (A) and at pH 8.5 (B). The proteins were in 0.05 M bis/Tris at pH 7.0 and in 0.05 M Tris a t pH 8.5. Experiments were carried out at 20°C with heme concentrations of 40 PM heme (before mixing), with an observation wavelength of 437.5 nm.

Structure and Function of Hemoglobin Suresnes 947 1

TABLE V Kinetic parameters for oxygen dissociation and CO combination

with Hb Suresnes in the presence and absence of IHP Experiments were carried out at 20°C by the rapid mixing tech-

nique with a monitoring wavelength of 437.5 nm. In the absence of IHP, the processes of 02 dissociation and CO combination are biphasic, and the velocity constants given pertain to the two exponential phases of the reaction. In the presence of IHP, there are interactions between the a and B chains which result in an autocatalytic time course of CO combination, and the apparent second order velocity constant for the final phase of the reaction is given.

On dimciation co 'Ornbina- tion Buffer and pH

K K L' L (fast) (slow) (fast) (slow)

S- I 10' M - I S"

0.05 M bis/Tris, pH 7.0 32 10 2.0 0.50 + l m M I H P 75 25 0.35 0.05 M Tris, pH 8.5 25 7 2.0" 0.50

" An additional fast phase with a rate constant of 5 X lo6 M" s-' was observed in dilute solutions (see text).

terms of two exponential processes of equal amplitude, both of which show a significant degree of pH sensitivity. The rates and proportions of the two phases in the oxygen dissociation kinetics of Hb Suresnes are essentially unchanged when the heme concentration is varied from 2 to 72 IJ.M (before mixing). The amplitudes and rates were calculated with an iterative curve fitting program written by Dr. Michael Johnson, Duke University Marine Laboratory, Beaufort, N. C. As shown in Table V, the rates of both phases are increased by IHP. The slower phase, however, is less affected by addition of organic phosphate cofactors. The increased rates observed in the presence of IHP are consistent with the decreased oxygen affinity observed in comparable equilibrium experiments.

The initial phase of the oxygen dissociation kinetics for stripped Hb Suresnes was decreased in rate to about 15 s" when 200 IJ.M CO was added to the buffer containing dithionite. In this case, the reaction observed is a replacement of oxygen by CO, at a rate determined by the rate of oxygen dissociation. The hemoglobin molecule under these conditions stays in its fully liganded configuration (21). The low cooperativity of stripped Hb Suresnes is substantiated by the near equivalence of the time courses observed when CO is present or absent in the dithionite-containing buffer.

Carbon Monoxide Kinetics-The high rate constant for oxygen binding to human hemoglobin and the low quantum yield for photodissociation of the oxyhemoglobin complex precluded a detailed study of the oxygen binding reaction. Consequently, the process of carbon monoxide binding was studied, with the awareness that the ligands may not be completely analogous. TWO techniques were used to examine the process of carbon monoxide combination with Hb Su- resnes. The first method, that of rapidly mixing deoxyHb Suresnes with a buffer containing carbon monoxide, examines the properties of the deoxygenated molecule. When stripped deoxyHb Suresnes is rapidly mixed with carbon monoxide at pH 7.0, the time course of carbon monoxide binding is much faster than that of normal human hemoglobin. As is illustrated in Fig. 10, the time course is distinctly biphasic. This differs markedly from the accelerating rate of CO binding which characterizes CO binding to the interactive binding sites of HbA. Biphasic CO binding to CPB-HbA has also been reported, with "on" constants of 6.6 X LO6 and 1.6 x lo6 M" s" (22). The two phases of the carbon monoxide binding process observed with Hb Suresnes are equal in amplitude at most observation wavelengths and are interpreted as being due to differential CO reactivities of the a and p chains of Hb

Suresnes. As determined by the CO concentration dependence of these rates, the faster phase has a second order rate constant of 2 x lo6 M" s-l, while the slower phase has a second order rate constant of 0.5 X lo6 M" s". These constants are significantly lower than for CPB-HbA. The rates and proportions of these two phases are unaffected by altering the protein concentration of the deoxyhemoglobin over a concentration range from 1 to 40 IJ.M in heme (before mixing). The time course of CO binding at pH 8.5 is essentially identical with that at pH 7.0 except that a still faster component appears whose magnitude is dependent on protein concentration. Thus, when a 20-IJ.M solution of Hb Suresnes is mixed with CO at pH 8.5, the time course reflects the presence of a and p chains within the Hb tetramer with on constants of 2 and 0.5 X lo6 M" s-', and approximately 50% of the observed reaction appears to be due to the presence of high reactivity dimers characterized by on constants of approximately 5 X lo6 M" s-'. Dissociation at high pH was also found for deoxyCPB-HbA (22).

The other technique used to examine the carbon monoxide combination process involves complete flash photolysis of fully liganded carbonmonoxy hemoglobin. In this technique, the dimer-tetramer equilibrium of the hemoglobin sample is characteristic of the liganded form of the molecule. As is shown in Fig. 10, the time course of carbon monoxide combination with Hb Suresnes at pH 7.0, as monitored by the flash photolysis technique, shows a distinctly different time course from that observed by the rapid mixing method. When a 5- PM (heme) solution of Hb Suresnes is subjected to flash photolysis, the time course is dominated by a fast phase which makes up 84% of the reaction. This phase has a second order velocity constant of 5 X lo6 M" s-'. The remainder of the reaction (16%) can be described by the same two phases as in the rapid mixing experiment shown in Fig. 10. The quickly reacting phase observed in the flash photolysis experiments but not in the stopped flow experiments at pH 7.0 may be ascribed to partial dissociation of liganded Hb Suresnes into high reactivity dimers. The proportion of this fast phase is dependent on protein concentration, as expected from simple mass law considerations. A comparison of the protein concentration dependence of the quickly reacting form of Hb Su- resnes with that for HbA indicated that at pH 7.0 the carbon monoxide derivative of stripped Hb Suresnes has a dimer- tetramer dissociation constant roughly 20-fold higher than that measured for HbA under the same conditions.

There is a large effect of IHP on the kinetics of carbon

.051 , ; , , , , , ,

"

. . . -.

.02 20 40 msec

FIG. 10. Time courses of CO combination and Hb Suresnes at pH 7.0 as measured by the rapid mixing technique @ow) or by the flash photolysis technique (Flash). Experiments were carried out with a final heme concentration of 5 p ~ . Experiments were performed at 20°C, with an observation wavelength of 437.5 nm.


monoxide combination with Hb Suresnes. Rapid mixing experiments carried out in the presence of IHP show a autocatalytic time course of carbon monoxide combination that is independent of protein concentration over the range from 20 to 1.2 ph4 heme (before mixing). The dominant rate of carbon monoxide combination (calculated from 75% to 25% of the reaction) has a second order rate constant of approximately 3.5 X IO5 M” s-’. Functional differences between the chains, which were conspicuous in the kinetic behavior of stripped Hb Suresnes, are no longer apparent. An even more dramatic effect of IHP is evident in the carbon monoxide combination behavior following flash photolysis. At 5 p~ (heme), very little quickly reacting material is observed. This contrasts with 84% seen in 5-pM solutions in the absence of inositol hexaphosphate. The addition of inositol hexaphosphate stabilizes the low affilnity state of Hb Suresnes and, in addition, strongly stabilizes the tetrameric form of liganded Hb Suresnes. The decreased second order velocity constant, determined from the carbon monoxide concentration dependence of the reaction as measured by both rapid mixing and flash photolysis techniques, is consistent with the dramatically lowered affinity of Hb Suresnes for ligands that is brought about by organic phosphate binding.

DISCUSSION

The x-ray studies of Perutz showed that the COOH-terminal residue of each a chain, arginine 141, serves as a major cross-link between equivalent al/31 and a2P2 dimers in the native deoxyhemoglobin tetramer (1). This cross-link includes two strong ionic bonds one between a guanidino NH, group of arginine 141al and a carboxyl oxygen of aspartate 126~~2, and the other between a carboxyl oxygen of the al chain COOH terminus and the a-amino group of lysine 127a2 (see Fig. 1). High resolution x-ray analyses (1) showed that the same guanidino NH2 group which interacts with aspartate 126a2 also forms a hydrogen bond across the aI/3z interface with the carbonyl oxygen of valine 34/32. At 3.5 A resolution, the electron density map of human deoxyhemoglobin also indicated the possibility of an additional salt bridge between the COOH-terminal group of arginine 141011 and the a-amino group of valine la2 (23). However, when the resolution of this map was extended to 2.5 A (ll), refinement of the atomic model by Fermi (24) showed these groups to be too far apart (5.3 A) to interact strongly. Instead, it has recently been shown that the second guanidino NH2 group of arginine 141a1 can be indirectly liked to the a-amino group of valine la2

through an inorganic anion bridge (6,25). From our x-ray study of deoxyhemoglobin Suresnes, we

find that a histidine residue at position 141a is not as effective as arginine in acting as a link between a/3 dimers. The deoxyhemoglobin Suresnes-deoxyhemoglobin A difference map (Fig. 2) shows that although the imidazole side chain of histidine 141a is located in approximately the same position as the normal guanidinum side chain, it cannot form as many strong intersubunit interactions. In particular, the interaction with aspartate 126m is lost (or at least weaker) in deoxyhemoglobin Suresnes, as is the inorganic anion bridge with valine lan. Also, the carbonyl oxygen of valine 34/32 does not appear to form a hydrogen bond with the imidazole of histidine 141al. Since the a-carboxyl group of histidine 141a1 is shifted (relative to the COOH terminus of deoxyhemoglobin A) toward the a-amino group of valine lan and away from the €-amino group of lysine 127a2, it may interact with the former residue but not with the latter.

In response to these major changes in a chain structure a number of secondary changes are also detected as weaker features in the deoxyhemoglobin Suresnes difference map.

These include a shift of the COOH-terminal portion of the H helix toward the central cavity and a movement of the F helix toward the H helix. Associated with the first structural change is a large movement of tyrosine 140a toward the central cavity. The phenolic side chain of this residue serves as an important part of the interface between helices F and H (1). Its movement (along with other elements of the H helix) causes the corresponding shift of the F helix so that the heli-helix interface is maintained. This, in turn, results in the coordi- nated movement of F helix residues which are in contact with the a-heme group (i.e. leucine 83a, leucine 86a, histidine 87a, and leucine 91a). Although the difference electron density around the a-heme group is weak, it appears that the heme group may follow helices F and H and move away from the E helix.

The functional studies performed with Hb Suresnes reveal the abnormal characteristics of this protein. Several observations made in this study through direct or indirect measurements point to the large alteration of the T quaternary structure in Hb Suresnes. This is demonstrated by the relatively high affiiity for oxygen of Hb Suresnes at very low levels of oxygen saturation. The mechanism for this alteration may be assigned to the disruption of the cluster of salt bridges dis- cussed above. The recently recognized interaction between the guanidinum group of arginine 141al and the protonated amino group of the NHz-terminal valine az has been related to the presence of a small inorganic anion like chloride (6,25). Therefore, Hb Suresnes should evidence a lower interaction with chloride anion. This is effectively so as it was found that the amount of oxylabile chloride is 1 mol/tetramer in Hb Suresnes compared to 2 mol/tetramer in HbA. This is further reduced in CPB-HbA. This should correspond to the loss of one oxylabile chloride binding site in Hb Suresnes, presum- ably that located in the a chains as described by O’Donnell et al. (25) and Chiancone et al. (5). Our x-ray studies show that the imidazole side chain of histidine 141al cannot form the normal inorganic anion bridge with valine la2. The abnormal interaction between the COOH and NH2 termini of the a chains in Hb Suresnes appears to preclude the expected pK change of the amino group of the NHn-terminal valine which is normally responsible for about 25% of the alkaline Bohr effect (26). The reduction in the pH dependence of the oxygen affinity of Hb Suresnes at physiological ionic strength was observed (Fig. 4).

It was shown previously that DPG increased the p m and the heme-heme interaction of Hb Suresnes, which led to the conclusion that the affiiity constant for DPG was similar in Hb Suresnes and HbA (8). Results presented in Table I confirm the same conclusions for IHP binding. However, it may be noticed that the effect of IHP on Hb Suresnes was slightly but significantly lower than in HbA. This phenomenon was still more pronounced in CPB-HbA. This indicates that the anion-binding sites of the /3 chains (27) are functional but “feel” the substitution at the a chain-anion-binding site. One may recall that the two classes of anionic sites are at opposite ends of the tetramer or approximately 55 A apart.

The right panel in Fig. 4 indicates that, compared to HbA, additional Bohr protons are associated with deoxyHb Su- resnes in the presence of IHP. In this regard, the effect of IHP on deoxyHb Suresnes differs from that observed in CPB-HbA where fewer (28) or equal (3) IHP-linked Bohr protons have been observed relative to HbA. The appearance of extra Bohr protons in Hb Suresnes upon addition of IHP may be interpreted in two ways. Either IHP binding to the /3 chains of Hb Suresnes indirectly triggers conformational changes that result in “extra” Bohr protons, or IHP alternatively may have a direct effect brought about by oxygen-linked binding at the


modifkd NH2-COOH-tenninal region of the Hb Suresnes a chains. The latter possibility cannot be eliminated by the data presently available. A final remark along these lines concerns the observation that Hb Suresnes in the presence of IHP has higher affinity for ligands than HbA, despite the fact that it has similar cooperativity and a similar alkaline Bohr effect under these conditions. This higher affinity is present at the midportion as well as at the lower and upper parts of the ligand binding curve. This indicates clearly that, even in the presence of IHP, the high oxygen affinity of Hb Suresnes cannot be explained solely by changes in the allosteric equilibrium between conformational states but must also involve changes in R and T structures. The R and T terminology retains usefulness for discussion purposes in that hemoglobin variants like Hb Suresnes undergo oxygen-linked changes in conformation. The two-state model must be used cautiously, however, in that the two end points are often found to be susceptible to modulation by pH and buffer conditions. This is clearly the case in Hb Suresnes where an infinite number of R and T states can be generated by varying the concentration of anionic effectors.

The kinetic measurements reported here reveal some inter- esting aspects of the functionaal differences brought about by the arginine + histidine substitution in Hb Suresnes. In both oxygen dissociation and CO combination, we find that stripped Hb Suresnes shows heterogeneous kinetics that are attribut- able, at least in part, to functional differences between a and /3 chains. It has been reported that CPB-HbA also shows biphasic CO binding curves under conditions of low ionic strength (22). In the presence of effectors like IHP, which stabilize the T structure of Hb Suresnes, it appears that the conformational transition occurs at higher levels of saturation. The h = 1 portion of the equilibrium binding curves is then readily measurable (Table 11) and the CO binding kinetics are much slower and show autocatalytic time courses.

When carbonmonoxy Hb Suresnes is subjected to complete flash photolysis at neutral pH, we observe a concentration- dependent fast phase in CO binding that appears to be due to quickly reacting dimers. When unliganded Hb Suresnes at pH 8.5 is mixed with CO, we find a fast phase with the same apparent second order rate constant (5 X lo6 M-’ s-’) that also appears to be due to dimers. These measurements strongly suggest that the substitution in Hb Suresnes involves a destabilization of the tetrameric forms of both oxy and deoxy forms. Previous reports show a similarly increased tendency of deoxyCPB-HbA to dissociate at alkaline pH (22). The structural studies reported here make it possible to interpret the increased dissociation of deoxyHb Suresnes as a direct con- sequence of missing salt bridges (see “Results”). The appar- ently 20-fold increased dissociation of carbonmonoxy Hb Su- resnes relative to HbA cannot be given an unequivocal structural explanation since crystals of liganded Hb Suresnes have not yet been examined. It is of interest, however, that liganded CPB-HbA was found to be significantly less dissociated than HbA (29). The available x-ray crystallographic data indicate that the a141 arginines act as spacers between the two a chains and that the a chains draw closer together in their absence (30). Enhanced dissociation instead of association in liganded Hb Suresnes suggests that the a141 histidines do not allow the a chains to come together but are instead disruptive of some interactions that normally stabilize the tetrameric form. The disruption of these as yet unidentified bonds may contribute to the altered R state observed for Hb Suresnes. To what extent the R state differences are tertiary or quaternary is also open to question. Preliminary experiments with

isolated a chains of Hb Suresnes revealed a somewhat increased affinity for oxygen compared to a chains of HbA. The experimental observations may be related to details of the x- ray structure of the a chains of Hb Suresnes. There is a weak electron density around the a-hemes and an abnormal movement of the F helix residues. These residues are thought to influence the tension exerted upon the heme. Alterations in the tertiary structure of the a chains might alter both the deoxy and oxy configurations of Hb Suresnes and may provide an important clue to its abnormal functional properties.

Acknowledgments-We gratefully acknowledge the skillful tech- nical assistance of Brigitte Bohn, Patrick Briley, and Paul Rogers. We thank also C. Gautheron and J. Grellier for their invaluable assistance in the preparation of this article.

1. 2.

3.

4.

5.

6.

7.

8.

9.

10. 11. 12.

13.

14.

15.

16. 17.

18.

19. 20.

REFERENCES

Perutz, M. F. (1970) Nature (Lond.) 228, 726-739 Bonaventura, J., Bonaventura, C., Brunori, M., Giardina, B.,

Antonini, E., Bossa, F., and Wyman, J. (1974) J. Mol. Biol. 82,

Kilmartin, J. V., Hewitt, J. A., and Wootton, J. F. (1975) J. Mol. Biol. 93, 203-218

Kilmartin, J. V., and Hewitt, J. A. (1971) Cold Spring Harbor Symp. Quant. Biol. 36, 311-314

Chiancone, E., Norne, J. E., F o d n , S., Bonaventura, J., Brunori, M., Antonini, E., and Wyman, J. (1975) Eur. J. Biochem. 55,

Arnone, A., and Williams, D. J. (1977) in Molecular Interactions of Hemoglobin (Labie, D., Poyart, C., and Rosa J., eds) pp. 15- 22, Institut National de la Sante et de la Recherche Medicale, Paris

Rollema, H. S., de Bruin, S. H., Janasen, L. H. M., and van Os, G. A. J. (1975) J. Bwl. Chem. 250, 1333-1339

Poyart, C., Krishnamoorthy, R., Bursaux, E., Gacon, G., and Labie, D. (1976) FEBS Lett. 69, 103-107

Gravely, M. E., Harris, H. F., Stallings, M., Lam, H., Wilson, J. B., and Huisman, T. H. J. (1978) Hemoglobin 2, 187-189

Perutz, M. F. (1968) J. Crystallogr. 2, 54-56 Ten Eyck, L. F., and Arnone, A. (1976) J. Mol. Biol. 100.3-11 North, A. C. T., Phillips, D. C., and Mathews, F. S. (1968) Acta

Muirhead, H., Cox, J. M., Mazzarella, L., and Perutz, M. F. (1967)

Poyart, C. F., Bursaux, E., and Bohn, B. (1978) Eur. J. Biochem.

Pomendale, T. (1968) CERN Computer 6OOO series, Program

Wyman, J. (1964) Adu. Protein Chem. 19,223-286 Poyart, C., Bursaux, E., Guesnon, P., and Bohn, B. (1979) Eur. J.

Hayashi, A., Suzuki, T., and Shin, M. (1973) Biochim. Biophys.

Perutz, M. F. (1976) Br. Med. Bull. 32, 195-208 Perutz, M. F., Ladner, J. E., Simon, S. R., and Ho, C. (1974)

499-511

385-390

Crystallogr. 24,351-359

J. Mol. Biol. 28, 117-156

87, 75-83

Library, Geneva, Switzerland

Biochem. 99,85-88

Acta 310.309-316

Biochemistry 13.2163-2173 21. Gibson, Q. H.,-and-Roughton, F. J. W. (1959) J. Physiol. (Lond.)

22. Hewitt, J. A,, and Gibson, Q. H. (1973) J. Mol. Biol. 74,489498 23. Muirhead, H., and Greer, J. (1970) Nature (Lond.) 228,516-519 24. Fermi, G. (1975) J. Mol. Bwl. 97,237-256 25. O’Donnell, S., Mandaro, R., Schuster, T. M., and Amone, A.

(1979) J. Biol. Chem. 254, 12204-12208 26. Kilmartin, J. V., and Rossi-Bernardi, L. (1969) Nature (Lond.)

27. Amone, A. (1977) Nature 237, 146-148 28. Van Beeck, G. G. M., Zuiderweg, E. R. P., and de Bruin, S. H.

(1978) Eur J. Biochem. 92,309-316 29. Chiancone, E., Anderson, N. M., Antonini, E., Bonaventura, J.,

Bonaventura, C., Brunori, M., and Spagnuolo, C. (1974) J. Biol. Chem. 249,5689-5694

30. Perutz, M. F., and Ten Eyck, L. F. (1971) Cold Spring Harbor Symp. Quant. Biol. 36, 295-310

145,32p

222, 1243-1246

THE OF CHEMISTRY Vol. 255, No. 19, hue of October 10, pp ... › content › 255 › 19 › 9465.full.pdf · THE JOURNAL OF BIOLOGICAL CHEMISTRY Printed in U.S.A. Vol. 255, No. 19,

Documents