X-ray and Functional Studies of Hemoglobins Nancy and Cochin-Port-Royal*

The mutations in hemoglobin Nancy Tyr-Asp and His-Arg involve residues which are thought to be essential for the full expression of allosteric action in hemoglobin. Relative to the structure of deoxyhemoglobin A, our x-ray study of deoxyhemoglobin Nancy shows severe disordering of the p chain COOH-terminal tetrapeptide and a possible movement of the fi heme iron atom toward the plane of the porphyrin ring. These structural perturbations result in a high oxygen affinity, reduced Bohr effect, and lack of cooperativity in hemoglobin Nancy. In the presence of inositol hexaphosphate (IHP), the Hill constant for hemoglobin Nancy increases from 1.1 to 2.0. But relative to its action on hemoglobin A, IHP is much less effective in reducing the oxygen affinity and in increasing the Bohr effect of hemoglobin Nancy. This indicates that IHP does not influence the R 2 T equilibrium as much in hemoglobin Nancy as in hemoglobin A, and this probably is

the R 2 T equilibrium as much in hemoglobin Nancy as in hemoglobin A, and this probably is due to the disordering of His 143p which is known to be part of the IHP binding site. IHP is also known to produce large changes in the absorption spectrum of methemoglobin A, but we find that it has no effect on the spectrum of methemoglobin Nancy. In contrast to the large structural changes in deoxyhemoglobin Nancy, the structure of deoxyhemoglobin Cochin-Port-Royal differs from deoxyhemoglobin A only in the position of the side chain of residue This degradation factor never exceeded 8% of the initial intensities.
Two crystals were used for a complete set of data for each mutant hemoglobin.
Difference electron density maps were calculated using the phases from the 2.5 8, native structure factors computed by isomorphous replacement methods3 and the structure amplitudes (/ F,,t,,, 1 -I Fm,,ve 1 ). The mean isomorphous difference, expressed as a percentage of the mean 1 F,,,,,,/ , was 3.8% for hemoglobin Cochin-Port-Royal and 9.5% for hemoglobin Nancy.
The larger value for hemoglobin Nancy reflects larger changes in structure.

RESULTS
Functional Properties-The functional properties of hemoglobin Cochin-Port-Royal have been reported previously (6) and are summarized under "Discussion." The oxygen equilibrium curve of purified hemoglobin Nancy ( Fig. 3 of Ref. 7) shows that it is a non-cooperative hemoglobin (n = 1.1) with a very high oxygen affinity. Table I gives the pH dependence of log pSo and n (measured at half-saturation) for hemoglobin Nancy in the presence and absence of a lo-fold excess of IHP. These data show that IHP partially restores cooperativity by uniformly increasing n from 1.1 to 2.0 over the pH range 5.9 to 7.2. Although the pso is increased markedly by IHF!, the Alog psa is only 60% of that observed with hemoglobin A (For hemoglobin A Alog p5,, is 0.75 at pH 7.2, and 1.0 at pH 6.6). Under the same conditions, a 70-fold excess of the less potent allosteric effector DPG does not increase the value of n and changes log psO by only 0.14 at pH 7.2.
The Bohr effect of hemoglobin Nancy, as measured by direct proton titration (5,12), was recorded in the presence and absence of IHP. These data ( Fig. 1) show that in the absence of IHP the maximum alkaline Bohr effect is reduced from a normal value of 0.50 proton/heme in hemoglobin A to 0.25 proton/heme in hemoglobin Nancy. Addition of a lo-fold excess of IHP increases the maximum value of the Bohr effect of hemoglobin Nancy by 0.13 proton/heme which is much less than the increase of 0.45 proton/heme observed for hemoglobin A under the same conditions. Perutz et al. (15,16,32) have shown that IHP produces large changes in the absorption spectrum of high spin methemoglobin derivatives and have presented evidence that these changes reflect a shift in the structure of methemoglobin from predominantly R state in the absence of IHP to the T state when it forms a complex with IHP. We have conducted the same experiment with aquomethemoglobin Nancy (Fig. 2) and find that IHP produced no change in its visible absorption spectrum, implying that the R + T transition cannot be induced by IHP in this case.
X-ray studies-The main feature of the deoxyhemoglobin Nancy difference map (Figs. 3 and 4 A and B) is a very large and intense region of negative density which starts at the COOH-terminal carboxyl group of the /3 chain and continues without interruption along the polypeptide backbone until it ' L. T. TenEyck and A. Arnone (1976) J. Mol. Biol. 100, 3-11. stops abruptly at His 143p. This negative density extends out from the backbone to include the side chains of the COOH-terminal 4 residues and is particularly intense at the mutation site, the phenolic group of Tyr 145P. Bordering the negative density are much weaker regions of positive density. Together these features show that the COOH-terminal tetrapeptide of the fl chain is severely delocalized as a result of the substitution of an aspartate for Tyr 1450. This disordering results in the rupture of the a,& intersubunit salt bridge between the carboxyl group of His 146@ and the c-amino group of Lys 40a, as well as the intrasubunit salt bridge between the imidazole group of His 146p and the carboxyl group of Asp 94p. It is interesting to note, however, that the side chains of residues  Lys 40~ and Asp 94p are not disordered in deoxyhemoglobin Nancy, but remain fixed in the positions found in deoxyhemoglobin A.
In addition to these strong features, the deoxyhemoglobin Nancy difference map contains two weaker ones which may be very significant. A pair of positive and negative peaks straddle the native density of Cys 930 (the positive peak may be diminished and distorted by the massive negative peak adjacent to it at the position of Tyr 145/3) indicating a movement of Cys 93 toward the space normally occupied by Tyr 145P. Another weak pair of peaks flank the iron atom of the /3 heme group indicating a small movement of the iron toward the heme plane, or possibly a small shift of the entire heme group with the movements of the light atoms unobserved at the resolution of this map. Anderson (17) has observed a similar movement upon oxidation of deoxyhemoglobin A, but in that case the presence of the water ligand produced a massive positive peak (relative to the weak negative peak on the opposite side of the iron atom) not seen in our difference map.
In contrast to the large structural changes caused by the mutation in deoxyhemoglobin Nancy, the difference electron density map of deoxyhemoglobin Cochin-Port-Royal (Fig. 4C) shows only one major feature, a very intense negative peak which is at the position of the imidazole group of His 146p in deoxyhemoglobin A. A few very weak positive contours (not shown) indicate that the guanidinium group of Arg 146p assumes a number of random conformations on the surface of the P subunit.

Hemoglobin Cochin-Port-Royal
Bohr Effect (18) found that the pK of the imidazole group is increased to 8.1 in deoxyhemoglobin from the normal value of 7.1 found in carbonmonoxyhemoglobin.
They estimate from this change in pK that the His 146-Asp 94 salt bridge is responsible for about 40% of the alkaline Bohr Effect.
Our x-ray studies of deoxyhemoglobin Cochin-Port-Royal show that the guanidinium ion of Arg 146p floats freely in solution and does not interact with Asp 940 to reform the intrasubunit salt bridge. Remarkably, the changes in electron density are confined entirely to the side chain of residue 1466. No other structural perturbations are observed. In particular, the intersubunit salt bridge is completely intact in deoxyhemoglobin Cochin-Port-Royal.
The functional studies of Wajcman et al. (6) showed that hemoglobin Cochin-Port-Royal has a normal Hill coefficient, a normal response to DPG, and (at physiological pH) a normal oxygen affinity. The only functional difference is a reduced Bohr effect, 0.35 proton/heme compared with 0.48 proton/ heme for hemoglobin A: a decrease of 27%. It is not clear why the full 40% decrease is not realized in this case, but it may be that the guanidinium ion of 146/3 decreases the pK of some basic group (e.g. His 2& His 1436, or the a-amino group of Val I@) in oxyhemoglobin Cochin-Port-Royal and thereby generates some additional Bohr protons to partially compensate for the loss of the salt bridge in the deoxy structure. mide group attached to Cys 93/3 prevents the salt bridges from forming), the Hill coefficient is reduced to about 2.5 (see Table  II). Our x-ray and functional studies of hemoglobin Cochin-Port-Royal show that the intrasubunit salt bridge is not required to full cooperativity (n = 3.0). However, it is not true that the presence of the intersubunit salt bridge implies full cooperativity. In the case of hemoglobin Hiroshima, replacement of His 1460 by aspartate causes a decrease in n to 2.4 (21), but x-ray studies have shown that the intersubunit salt bridge is not broken in the deoxy state (22).

Hemoglobin Nancy
Stability of T State-In addition to hemoglobin Nancy, an enzymatically modified hemoglobin and two mutant hemoglobins involving Tyr 145fi have been studied in some detail; des-(His 146/3, Tyr 145p) hemoglobin (4,23), hemoglobin Bethesda 145p Tyr-His (24, 25), and hemoglobin Rainier 145p Tyr-Cys (24, 26). In every case, the oxygen affinity is close to the average value for free (Y and P chains, cooperativity is virtually eliminated with n = 1.0 to 1.2, and the Bohr effect is reduced by about 40% (except for des-(His 146/3, Tyr 145p) hemoglobin where it is only one-third the normal value). Our x-ray study of deoxyhemoglobin Nancy shows that the substitution of aspartate for Tyr 145p causes severe disordering of the /3 chain COOH-terminal tetrapeptide which results in the rupture of both the intra-and intersubunit salt bridges associated with His 146p. X-ray studies of hemoglobin Rainier (26) and des-(His 146& Tyr 1450) hemoglobin (23) show the salt bridges to be missing in these hemoglobins as well.
Removing the salt bridges, however, does not totally explain the very high oxygen affinity or the corresponding lack of cooperativity in these four hemoglobins because other hemoglobins, namely des-(His 146p) hemoglobin and N-ethylsuccinimide (NES) hemoglobin (27), which lack these salt bridges but retain Tyr 145/3 have moderately low oxygen affinities and n values of 2.5. Tyr 145p stabilizes the T structure of deoxyhemoglobin by forming van der Waals contacts with portions of helices F and H and a hydrogen bond between its phenolic hydroxyl group and the carbonyl group of Val 98p (3). Using the allosteric model of Monod et al. (28), Baldwin (29) has shown that the parameter L (the ratio of '1' to R structure in deoxyhemoglobin) must be reduced from a normal value of lo5 to about 300 to account for the higher oxygen affinity and lower n value of des-(his 1460) hemoglobin and N-ethylsuccinimide hemoglobin. In order to reduce the value of n to 1.15, L must have a value near unity (29). Translating these equilibrium constants to free energies indicates that in the p chain the salt bridges and Tyr 145 contribute about the same amount of free energy to the stability of the T structure in deoxyhemoglobin.
Bohr Effect-The reduction of the alkaline Bohr effect by  (5) have shown that only 10% of the Bohr effect is related to changes in the R structure upon ligation, the remainder is generated by the R + T transition. Thus, retention of one-half of the Bohr effect is evidence for the R + T transition in solution and this is consistent with the crystallization of deoxyhemoglobin Nancy in the T state. Effect of IHP-Subsequent to the discovery of Benesch et al. (30) of IHP as an extremely potent effector of hemoglobin oxygen affinity, the dimished or absent allosteric properties of several abnormal hemoglobins have been partially (or even fully) restored to normal levels by the addition of IHP (4,5). We find that a lo-fold excess of IHP increases n from 1.1 to 2.0 for hemoglobin Nancy, but that it is only about 60% as effective in raising the log pS,, of hemoglobin Nancy relative to its action on hemoglobin A. It can be shown (29) to a good approximation that the value of Alog psO due to the binding of any heterotopic ligand is related only to the association constants of that ligand (in this case IHP) with deoxy and fully oxygenated hemoglobin. Therefore, relative to hemoglobin A the affinity of IHP for deoxyhemoglobin Nancy must have decreased and/or it must have increased for oxyhemoglobin Nancy. The observed disordering of His 143/3 in deoxyhemoglobin Nancy implies that IHP would probably have a decreased affinity for the deoxy structure since His 1430 is part of the IHP binding site (31). This is consistent with the very small i.ncrease of only 0.13 proton/heme in the Bohr effect of hemoglobin Nancy (compared with 0.45 proton/heme for hemoglobin A) after the addition of a lo-fold excess of IHP. The small increase in the Bohr effect also indicates that the His 146&Asp 94p salt bridge is not reformed in the IHP 'deoxyhemoglobin Nancy complex.
In the presence of IHP the value of L for deoxyhemoglobin A increases from lo5 to 10' (29). The ratio of T to R structure in aquomethemoglobin is given by L* = Lc' (where c = K,/K, is the ratio of equilibrium constants for the reduction reaction in the R and T states). Perutz et al. (32) have estimated that for aquomethemoglobin A the value of L* changes from < 1 to > 1 with the addition of IHP. This inversion of population in favor of the T structure is thought to result in an increase in tension on the heme iron which in turn leads to changes in the adsorption spectrum of aquomethemoglobin (32). If, however, the value of L is much lower than 105, as it is in deoxyhemoglo-bin Nancy, then the value of L* may be so low that addition of excess IHP will not produce a significant concentration of T structure. Moreover, as discussed above, IHP does not shift the R 4 T equilibrium as much in hemoglobin Nancy as in hemoglobin A. These two factors together with a reduction of tension in the p hemes of the hemoglobin Nancy T structure (see below) would explain the inability of IHP to change the visible absorption spectrum of aquomethemoglobin Nancy (Fig. 2).
Release of Heme Tension-High spin heme iron in the ferrous state, and to a smaller degree in the ferric state, is out of the plane of the porphyrin ring because of its large radius (3,33). Perutz (34,35) and Perutz et al. (15,16,321 have presented evidence to support the idea that the low oxygen affinity of deoxyhemoglobin A is due in part to constraints on the globin which in turn give rise to a tension on the heme that pulls the iron atom even further out of the plane of the porphyrin ring. If these particular constraints are relaxed or eliminated in the T structure of some high affinity hemoglobins like hemoglobin Nancy, then a movement of the iron atom toward the porphyrin plane would be expected. Our difference map of deoxyhemoglobin Nancy indicates such a movement and is therefore consistent with Perutz's hypothesis. However, at the present resolution of our map we cannot rule out the possibility that this movement reflects a very small shift of the entire heme group and adjacent atoms which would produce differences in electron density too weak to be observed for the lighter atoms.