Allosteric transition in triply met-haemoglobin

J. Mol. Biol. (1991) 217, 303-306

Allosteric Transition in Triply Met-Haemoglobin Michael C. Marden, Jean Kister, Brigitte Bohn and Claude Poyart INSERM U299, Hdpital de B&he 94275 Le Kremlin Bic&tre, France (Received 29 June 1990; accepted 17 September 1990) Methaemoglobin undergoes a transition to a T-like form at acid pH in the presence of strong effecters such as inositol hexakisphosphate (IHP), as evidenced by spectroscopic and oxidation potential measurements. Since oxygen and CO do not bind to the ferric haems, it is dithcult to compare the properties of the R-met and T-met forms with those of ferrous haemoglobin. We have therefore prepared 90% oxidized samples, where the dominant signal for ligand (oxygen or CO) binding is due to tetramers with three met haems. Measurements were made of the oxygen equilibrium curves and CO rebinding kinetics after photodissociation. Without effecters, the partially oxidized samples show mainly R-state properties. Addition of IHP at acid pH induces an increase in T-state behaviour, as indicated by a lower oxygen affinity and a higher fraction of the slow bimolecular component for CO rebinding.

ligand affinity and the bimolecular recombination rates, similar to those for ferrous Hb.

1. Introduction Spectroscopic studies (Perutz, 1972; Perutz et al., 1974; Olson, 1976; Noble e2 al., 1989) and oxidation potential measurements (Kilmartin, 1973) have indicated that below pH 7 there is a conformational change in ferric haemoglobin (Hbt) upon addition of strong effecters such as inositol hexakisphosphate (IHP). In analogy with the oxy (R-state) and deoxy (T-state) ferrous forms, met Hb in the presence of effecters was considered to be a T-met conformation. Unfortunately, no comparison can be made to the oxygen or CO binding properties which characterize the R and T ferrous forms, as these ligands do not bind to ferric haems. We therefore prepared partially oxidized Hb to investigate ligand binding to the remaining ferrous chains. Above 90% oxidation, the main tetrameric forms are those with three or four ferric subunits. Symmetrical valency hybrids also provide useful information (Nagai, 1977), but the triply met species is the closest approximation to fully met that allows measurements of the oxygen binding affinity. Under these conditions, oxygen or CO binding to the single ferrous subunit can be observed. The allosteric transition to a T-like state is apparent upon addition of strong effecters; we report here equilibrium and kinetic results for the partially met forms. The allosteric transition in the triply met hybrid shows large changes in the

2. Materials and Methods Hb A was prepared from fresh red blood cell haemolysate by DEAE-Sephadex chromatography and stripped of anions as described (Lee et al., 1988); isoelectric focusing showed a single band. Partially deoxygenated Hb solutions (2 mM in haem) were allowed to autoxidize at 37°C in 50 mM-bis Tris buffer (pH 65), 100 mr+NaCl. Chloramphenicol (20 &ml) was added to prevent bacterial growth: 90% oxidation was achieved in less than 48 h. In order to observe differences in the tl or B-chains, symmetrical valency hybrids, where only one type of chain is in the oxidized form, were prepared (Banerjee & Cassoly, 1969; Petrich et al., 1988) and then allowed to autoxidize to above 90% overall met Hb. From the visible absorption spectrum (Gary 219, Varian, U.S.A.), the percentage met Hb was calculated using the known absorption coefficient at a given pH (Benesch et al., 1973; van Assendelft & Zijlstra, 1975). Oxygen equilibrium curves (OEC) were measured using a Hemox analyser (TCS, Southampton, PA) as described (Kister et al., 1987). Samples were 62 mre (total) haem in 100 mM-NaCl at 25°C in 50 mre-bis Tris buffer (pH 6%). The contribution of dimeric forms should be small above 60 pM total haem (Hensley et aZ., 1975). The percentage met Hb was little changed during the 45 min necessary to record the oxygen equilibrium curve. Recombination kinetics were measured after photodissociation by a 10 ns laser pulse at 532 nm (Quantel, France) as described (Marden et aZ., 1988). Detection was at 436 nm with 92 mm total haem samples in 1 mm cuvettes equilibrated with 10,132 Pa CO. The signal size, relative to that for a ferrous Hb sample, served as a

t Abbreviations used: Hb, haemoglobin; IHP, inositol hexakisphosphate; OEC, oxygen equilibrium curve; L345, 2-[4-(3,4,5-dichlorophenylureido)phenoxy]-2methylpropionic acid. 0022-2836/91/020303-04

$03.00/O

303

0

1991 Academic Press Limited

304

M. C. Marden

-

0.1 I

et al.

I

1

I

I

I

4

6

12

16

20

4

8

Time (ms)

Figure 1. Bimolecular recombination kinetics of CO to 90% met Hb, with 1 mM-IHP, and with both effectors IHP and L345 (02 mnr), pH 6.8, 25°C. Curves are normalized on a log-linear plot. IHP and L345 cause additive effects for the transition to a state with a larger fraction of the slow (T-like) rebinding rate. Samples of 266 PM (total haem) were equilibrated with 10,132 Pa (01 atm) CO.

control of the percentage of the haems in the ferrous form. It has been demonstrated that under CO, met Hb undergoes a slow auto-reduction (Bicker et al., 1984; Young & Caughey, 1987); the measurements with and without effecters were made within minutes of each other, to minimize the difference in percentage met. Effector concentrations were 1 mM for IHP (Sigma) and 6.2 mM for L345 (2-[4-(3,4,5-dichlorophenylureido) phenoxyl-2-methylpropionic acid) a gift from Dr Lalezari (Lalezari et al., 1996).

3. Results (a) Photodissociation Recombination kinetics for 90% met Hb equilibrated with 10,132 Pa CO at 25°C and pH 6.8 are shown in Figure 1. The curves are the normalized change in absorption presented on a log-linear plot; actual signal sizes were typically 605 absorbance unit. Without effector, there is mainly a component with a rapid rate characteristic of R-state Hb. With the effector IHP, there is a large increase in the percentage (80%) with a rate similar to that of ferrous T-state Hb (Gray & Gibson, 1971). IHP and L345 have different binding sites on Hb and show an additive effect (to over 90 o/o slow phase) for the shift in equilibrium towards the T state. The fraction occurring as the slow (T-state) phase, with or without IHP, is greatly diminished if the ferric ligand is low spin such as cyanide or hydroxide (data not shown), as previously observed (Brunori et al., 1974). For a Hb tetramer with four ferrous haems, the slow fraction increases as the level of photodissociation is raised; at nearly complete dissociation producing deoxy Hb, The T-state is heavily favoured, while at low levels producing mainly triply liganded Hb a large fraction of rapid phase (R-state) is observed. The CO rebinding signal for 90% met Hb samples is due mainly to tetramers with a single ferrous haem (about 75% signal from triply ferric tetramers and 25% signal from doubly ferric tetramers, using a binomial distribution of

12 Time (ms)

16

20

Figure 2. Recombination of CO to 92% met Hb, prepared from valency hybrids with one type of chain comuletelv oxidized. Samnles were 206 UM in haem at pH 68, 25’C, equilibratedlwith 10,132 Pa CO.

met haems). Samples between 90 and 97 yo. met showed little change in the fraction of slow CO recombination. Since the dominant signal is for tetramers with a single ferrous haem, little dependence on the photodissociation level is expected and little change was, in fact, observed. In a similar study, a pulse radiolysis method was employed to photo-reduce a small fraction of the haems (Rollema et al., 1976); in the presence of IHP, the fraction of the slow (T-state) phase was also independent of the reduction level. For hybrids where the CO recombines to only one type of chain, similar levels of the slow (T-state) kinetics were observed when the ferrous chain was a or /l (Fig. 2). Without IHP, the rate of the slow phase for CO rebinding to an a-chain was twice that for rebinding to a p-chain. With IHP, the difference in rate decreased. (b) Oxygen equilibrium Data

for ferrous

Hb

curves

and for 90%

met Hb

are

shown in Figure 3. The fraction saturation refers to the ferrous haems: 50% saturation (Y = 95) for a 90% met Hb sample means that 5% of total haems are oxygen bound. With or without the effector IHP (1 mM), the 90 o/o met Hb oxygenation curves show a slope of nearly unity on the Hill plot; that is, they do not show co-operativity of hgand binding. This is expected, since the signal is dominated by tetramers with only a single ferrous haem. Without effector, the oxygen affinity for the triply met species is 1 mmHg (1 mmHg z 133.3 Pa); note that there is a crossover at the upper portion of the curves for 90% met Hb and ferrous Hb (Fig. 3). With IHP, the oxygen affinity for triply met Hb is 50 mmHg. 4. Discussion In order to determine the different conformations of the Hb tetramer, we have compared the ferric and ferrous R and T states. By using triply met

tetramers, we can study the influence of the met haems on the oxygen affinity of the single ferrous haem.

Allosteric Transition

lop (PO,) Figure 3. Oxygen equilibrium without (- - -) and with pH 68, 25°C. Also shown as a ferrous Hb (---) at 50 mlvr-bis-‘l’ris buffer.

curves for 90% met Hb, (-.-.-) 1 mM-IHP at reference is the curve for 100 mM-NaCl, pH 68,

(a) Allosteric equilibrium For ferrous Hb without effecters, simulations using the two-state model (Monod et al., 1965) indicate that the allosteric equilibrium (Li = Lc’, where i is the number of ligands bound and e = KJKT) is shifted toward the R state by a factor of about 100 (l/c) for each oxygen bound, with log(L) = 5 and KR = 0.35 mmHg at pH 6.8. For a single ferrous site per tetramer, the Hill plot is necessarily a straight line; the observed oxygen affinity is intermediate to the R and T-state limits: Kobs = Kdl

+L,)l(l

+L,)

= f&P +LWTIKR)I/(~ +Ld

The observed asymptote, as for the upper asymptote for fully ferrous Hb, may not represent the true R-state affinity, but depends on the R-T equilibrium. For triply met tetramers, the coefficient L, = Lm3 refers to the met (m) haems, and not the ferrous coefficient L, = Lc3. Since the fourth ligand is oxygen, the ratio of the allosteric equilibria coeffcients L,/L, = I/c. Within this two-state framework, we calculate a shift for each aquo-met haem that is three times less than that for oxygen (m/c = 3). This implies a factor of 27 increase in the allosteric equilibrium L, = Lc3 = [T3]/[R3] for triply aquo-met tetramers relative to triply liganded Hb (oxygen or CO), which is consistent with the increase in T-state behaviour in the CO recombination kinetics. This explains the crossover in the OEC (Fig. 3); triply aquo-met Hb shows an oxygen affinity for an R/T mixture that is lower than that for the nearly pure R-state afinity

in met Hb

305

derived from the upper asymptote of the curve for ferrous Hb. The calculated equilibrium for fully aquo-met Hb is L, = Lm4 = @08, which is nearly two orders of magnitude greater than for oxy Hb (Lc4). These values are consistent with the observations of a large allosteric transition upon addition of strong effecters for aquo-met Hb, but not for oxy Hb. For example, an effector inducing a factor of 10 shift towards the T state would lead to nearly 50 y. of the T-met form, but less than 1 y. of liganded T-state. This may provide a new method to study liganded T-state Hb, normally difficult as ligation induces the transition towards the R-state. As mentioned above, the observed affinities for the triply met Hb depend on the R and T affinities, and the equilibrium between these two states. Without knowing the allosteric equilibrium, it is difficult to determine whether the T-met properties are the same as those for the ferrous T-state. The observed oxygen affinity for triply met Hb with IHP is 50 mmHg, as compared to ferrous T-state Hb (KT = 30 mmHg, without IHP) and ferrous Hb in the presence of IHP (KT’ = 160 mmHg; Kister et al., 1987). This suggests that the T-met conformation is similar to the ferrous T structure, rather than the R-state (of affinity 0.35 mmHg). The present results are specifically for pH values below 7, where the ferric haems bind water, a high spin ligand. Preliminary results (not shown) indicate that cyanide and oxygen shift the allosteric equilibrium by a similar factor (m/c = l), as suggested previously (Perutz et al., 1974; Szabo & Karplus, 1975). (b) Chain differences There is evidence that the a-chains autoxidize more rapidly than the B-subunits (Mansouri & Winterhalter, 1973; Tomoda et al., 1981). This would change the distribution of met species and cause the oxygen binding signals to be more heavily weighted by the j haems. Samples with only a or only P-chains being ferrous both showed the R to T transition (increase in the fraction of slow bimolecular kinetics (Fig. 2) upon addition of IHP). These results are also relevant to the question of whether a B-chain may have a T-state affinity, It has been suggested that the p-chains have an extremely low affinity in T-state Hb, with or without IHP (DiCera et al., 1987). The present results with completely met a haems show that the b-chains have normal T-state properties, although the slower rebinding rates for the B-chains may indicate a lower affinity relative to the a-chains. The OEC for triply met Hb samples, where the ferrous haem may be an a or fl haem, show a slope of nearly 1 (a single affinity), suggesting that there is not a large difference in affinity for the two chains.

(c) Conclusions These results confirm the R to T transition in met Hb and provide measurements of the oxygen and

M. C. Marden

306

CO binding properties to triply met species. The results indicate that the T-met conformation is similar to the ferrous T-state, although there may be small differences, as for the R and T states with and without effector. Each aquo-met haem induces a shift in the allosteric equilibrium about three times less than for oxygen. We thank L. Kiger for help with the kinetic measurefor providing samples of the effector L345. This work was supported by funds from the Institut National de la Sante et de la Recherche Medicale (INSERM), the Air Liquide Co, and the Fondation pour la Recherche Medicale.

ments, and Dr Lalezari

References Banerjee, R. & Cassoly, R. (1969). J. Mol. Biol. 42, 351-361. Benesch, R. E., Benesch, R. & Yung, S. (1973). Anal. Biochem. 65, 245-248. Bicker, D., Bonaventura, C. & Bonaventura, J. (1984). J. Biol. Chem. 259, 10777-10783. Brunori, M., Giardina, B. & DiIorio, E. (1974). FE&S’ Letters, 46, 312-316. DiCera, E., Robert, C. H. & Gill, S. J. (1987). Biochemistry, 26, 4003-4008. Gray, R. D. & Gibson, Q. H. (1971). J. Biol. Chem. 246, 7168-7174. Hensley, P., Moffat, K. & Edelstein, S. J. (1975). J. Biol. Chem. 250, 9391-9396. Kilmartin, J. V. (1973). B&hem. J. 133, 727-733. Edited

et al.

Kister, J., Poyart, C. & Edelstein, S. J. (1987). Biophys. J. 52, 527-535. Lalezari, I., Lalezari, P., Poyart, C., Marden, M. C., Kister, J., Bohn, B., Fermi, G. & Perutz, M. F. (1990). Bioc&mistry, 29, 1515-1523. Lee, A. W., Karplus, M., Poyart, C. & Bursaux, E. (1988). Bioc&mistry, 27, 1285-1301. Mansouri, A. & Winterhalter, K. H. (1973). Biochemistry, 12, 4946-4949. Marden, M. C., Kister, J., Bohn, B. & Poyart, C. (1988). Biochemistry, 27, 1659-1664. Monod, J., Wyman, J. & Changeux, J.-P. (1965). J. Mol. Biol. 12, 88-l 18. _ Nagai, K. (1977). J. Mol. BioZ. 111, 41-53. Noble, R. W., DeYoung, A., V&ale, S., Cerdinio, M. & DiIorio, E. E. (1989). Biochemietry, 28, 5288-5292. Olson, J. S. (1976). J. BioZ. Chem. 251, 447458. Perutz, M. F. (1972). Nature (London), 237, 495-499. Perutz, M. F., Fersht, A. R., Simon, S. R. & Roberts, C. K. (1974). Biochemistry, 13, 2174-2186. Petrich, J. W., Poyart, C. & Martin, J.-L. (1988). Biochemistry, 27, 40494060. Rollema, H. S., Scholberg, H. P. F., de Bruin, S. H. & Raap, A. (1976). B&hem. Biophys. Res. Commun. 71, 997-1003. Szabo, A., & Karplus, M. (1975). Biochemistry, 14, 931-940. Tomoda, A., Yoneyama, Y. & Tsuji, A. (1981). B&hem. J. 195, 485-492. van Assendelft, 0. W. $ Zijlstra, W. G. (1975). Anal. B&hem. 69, 4348. Young, L. J. & Caughey, W. S. (1987). J. BioZ. Chem. 262, 1501915025.

by A. Klug