Proton transfer between hemoglobin and the carbonic anhydrase active site.

The exchange of ‘*O between COP and water (type I) and the exchange of ‘*O between Wand W-containing species of CO, (type II) at 25” near neutral pH have been taken as a measure of the hydration activity of bovine red cell carbonic anhydrase. The rate constant for type I exchange increased sharply by 70% upon addition of bovine oxyhemoglobin and reached a plateau at 5 x lo-‘; M hemoglobin. The rate constant for type II exchange decreased sharply and leveled off at 5 x lo-” M oxyhemoglobin. In the same range of concentration, L-histidine had a similar effect on type I and II exchange. Carbonic anhydrase was present at 2 or 3 x 10m9 M. In analogy with the interpretation of similar observations at 5 X lo-” M of buffers such as imidazole and 2,4lutidine, it is proposed that hemoglobin enhances catalytic activity of bovine carbonic anhydrase by donating protons to and accepting protons from the enzyme active site. The observation that hemoglobin causes a maximal increment of hydration activity at concentrations lOOO-fold less than for imidazole and many other buffers of low molecular weight cannot be explained solely by summing the expected proton transfer capabilities of residues on the surface of the protein. Furthermore, this effect need not be related to surface properties or the large size of hemoglobin since rate constants for type I and II exchange also level off at concentrations near 5 X lo-” M of L-histidine. In these equilibrium experiments, it is possible that there is cycling of protons between hemoglobin, or r,-histidine, and the carbonic anhydrase active site during which more than 1 atom of I80 is exchanged from COP to water. Hemoglobin and carbonic anhydrase are the two most abundant proteins in red cells and we anticipate that proton transfer between them, including possibly the Bohr proton, occurs in part as a result of the direct encounter of hemoglobin with carbonic anhydrase in the cell.

The exchange of '*O between COP and water (type I) and the exchange of '*O between W-and W-containing species of CO, (type II) at 25" near neutral pH have been taken as a measure of the hydration activity of bovine red cell carbonic anhydrase. The rate constant for type I exchange increased sharply by 70% upon addition of bovine oxyhemoglobin and reached a plateau at 5 x lo-'; M hemoglobin.
The rate constant for type II exchange decreased sharply and leveled off at 5 x lo-" M oxyhemoglobin.
In the same range of concentration, L-histidine had a similar effect on type I and II exchange. Carbonic anhydrase was present at 2 or 3 x 10m9 M. In analogy with the interpretation of similar observations at 5 X lo-" M of buffers such as imidazole and 2,4lutidine, it is proposed that hemoglobin enhances catalytic activity of bovine carbonic anhydrase by donating protons to and accepting protons from the enzyme active site. The observation that hemoglobin causes a maximal increment of hydration activity at concentrations lOOO-fold less than for imidazole and many other buffers of low molecular weight cannot be explained solely by summing the expected proton transfer capabilities of residues on the surface of the protein. Furthermore, this effect need not be related to surface properties or the large size of hemoglobin since rate constants for type I and II exchange also level off at concentrations near 5 X lo-" M of L-histidine. In these equilibrium experiments, it is possible that there is cycling of protons between hemoglobin, or r,-histidine, and the carbonic anhydrase active site during which more than 1 atom of I80 is exchanged from COP to water. Hemoglobin and carbonic anhydrase are the two most abundant proteins in red cells and we anticipate that proton transfer between them, including possibly the Bohr proton, occurs in part as a result of the direct encounter of hemoglobin with carbonic anhydrase in the cell. Oxygen-18-labeled and carbon-13-labeled potassium bicarbonate were prepared as described previously (12). Initial enrichment of ia0 was as high as 70% so that the isotope exchange processes could be followed with good precision for many half-times.
Carbon-13 enrichment was near 50%. Methods -The activity of bovine carbonic anhydrase was followed as a function of buffer or protein concentration by measuring the rate of exchange of IRO between CO, and water (type I exchange) and between "C-and i3C-containing species of CO, (type II exchange).
Type I exchange occurs as a result of the hydrationdehydration cycle (13): C'"O'"0 + HZ'60 -HC'XO'HO'GO~ + H+ -C'60'H0 + HZ'*0 and is catalyzed by carbonic anhydrase. Type II exchange in the absence of enzyme occurs as a result of the chemical reaction between CO, and C0,2m in which an oxygen is exchanged (14); this exchange is also catalyzed by carbonic anhydrase (6,12,14). Eight milliliters of a solution containing buffer or protein, ixOlabeled bicarbonate, and 'Glabeled bicarbonate, making a total concentration of 0.015 M in all species of CO,, were placed in the inlet vessel at 25". Ionic strength was maintained at 0.2 with Na,SO, and pH was adjusted with 0.1 N NaOH or H,SO,. The inlet vessel had as its bottom a membrane permeable to CO, which allowed the measurement of isotopic content of CO, by a mass spectrometer. This apparatus is described in Ref. 12 under the heading "Low pH Range." Following a period of 2 or 3 min to approach chemical equilibrium between CO, and HCO,-, rates of '*O exchange were measured from the atom fraction of I*0 in i*C-and '"C-containing CO,. These procedures are described in Ref. 12. 0 is the first order rate constant for type I exchange, obtained as the slope of a plot of the logarithm of the atom fraction of '*O in all CO, against time. + is the first order rate constant for the exchange of '"0 between 12C-and i3C-containing species of CO,, type II exchange, obtained as described in Ref. 12. Both rate constants 0 and d, can be separated into catalyzed and uncatalyzed components: 8 = tica1 +-H,,,,,,, 6 = &,r + 4 ""Cat.
Following this measurement, bovine carbonic anhydrase was added in a volume that was less than 0.1 ml to give a total concentration near 3 x 1O-9 M carbonic anhydrase. Values of Beat and 4 cat reported here are obtained by subtraction of 0 and 4 measured before the addition of carbonic anhydrase from the value of 0 and 4 obtained after addition of carbonic anhydrase. For all cases except the hemoglobin experiments, this simply amounts to subtracting 0 ""Cd and hncat from % and $. The same procedure was followed in hemoglobin experiments. For these cases, however, the rate constants 8 and 6 measured before the addition of carbonic anhydrase in the described procedure were slightly greater than the rate constants for the uncatalyzed exchanges because of the small amount of carbonic anhydrase impurity in the hemoglobin sample.  bearing on its surface many amino acid side chains with buffering capabilities.

In
Bovine hemoglobin has a high content of histidyl residues, 38 out of a total of 572 residues (16) Fig. 1 have been interpreted in terms of the catalytic mechanism (5, 6, 12). Most notably, the pattern for bat is indicative of a change in rate-determining step: as the concentration of proton-transfer agent increases the catalysis moves from a region in which buffer-assisted proton transfer is rate-limiting to a region in which another step in the catalysis is rate-limiting.
The pattern for the variation of 4J eat is symmetrical with that for oeat. It can be accounted for by the hypothesis that the residence time for iSO in the active site is roughly the same order of magnitude as the time for one catalytic cycle (6, 12); this results in the catalytic reaction of I80 with WO,. As the concentration of proton transfer agent increases, the residence time for '*O in the active site decreases leading to the decrease in &,t. Treating the data of Fig. 1 in a manner described earlier (51, it is possible to estimate the bimolecular rate constant for the proton exchange between imidazole and the carbonic anhydrase active site: k, 2 1 x lo* M-' s-i. This proton transfer is probably diffusion-controlled or close to it. The data for the changes in f& and &,, with concentration of bovine oxyhemoglobin, human serum albumin, and L-histidine (pK, -6.0) at a constant concentration of bovine carbonic anhydrase are shown in Figs. 2, 3, and 4. These data display the same symmetrical pattern for ecat and &,t and the same evidence for a change in rate-determining step as observed for buffers of low molecular weight. Also, the magnitude of the increase in 8,,, upon addition of hemoglobin or L-histidine is similar to that observed for buffers of low molecular weight. Consequently, it is a reasonable hypothesis that the changes in the I80 exchange rates upon addition of hemoglobin and histidine result from their action as proton transfer agents, I.....

Proton Transfer
between Hemoglobin and Carbonic Anhydrase groups; any of these, if exposed, would act as buffers in the range of pH 6 to 8. Nazaki determined that 28 groups in native bovine hemoglobin are titrated in the neutral pH region.' Human serum albumin, although similar to hemoglobin in molecular weight, contains 16 histidines out of about 610 residues (18) and requires about 12 eq of acid to be titrated from pH 8 to 6 (19). Other amino acid side chains containing carboxyl (pK, -4.5), e-amino (pK, -lo), or phenolic (pK, -10) groups would not be expected to exert a proton transfer effect as great as the imidazole (pK, -6 to 7) side chain of histidine under the conditions of these experiments. Consider a carboxylate anion as a proton acceptor and the enzyme active site as donor (pK, -7). Such a proton transfer is a positive free energy process and would occur with a bimolecular rate constant several orders of magnitude below that for a diffusion-controlled process (20). On the other hand proton transfer from the carboxylic acid moiety to the enzyme active site would be expected to proceed with a rate constant close to diffusion-controlled limits; however, the fraction of all car-boxy1 groups in the protonated form is very small near neutral PH.
Considering these facts and the data of Figs. 1 and 2 it appears that the efficiency of hemoglobin as a proton transfer agent in activating carbonic anhydrase is greater than the sum of the expected effects of the individual buffering groups on the surface of the hemoglobin molecule. Bovine hemoglobin activates carbonic anhydrase at concentrations lOOO-fold less than the concentrations required of imidazole and certain other buffers of low molecular weight (4, 5), although the maximal activities attained are nearly the same with hemoglobin and imidazole. Even if all titratable groups on hemoglobin are considered, the factor of 1000 is not met: it requires about 170 eq of hydrogen ions to titrate 1 mol of bovine hemoglobin from pH 13 to 1.5. Hence, the second order rate constant for the transfer of protons between hemoglobin and the carbonic anhydrase active site, calculated from the data of Fig. 2, is greater by roughly a factor of 10 than can be accounted for on the basis of residues of hemoglobin which are buffers near neutral pH. This order of magnitude is a conservative estimate since these arguments have not yet considered that hemoglobin has a smaller diffusion coefficient (6 x 10m7 cm%) than buffers of small molecular weight (about 5 x 1Om6 cm%).
The low concentration at which hemoglobin acts as a proton transfer agent with the carbonic anhydrase active site in these experiments, then, is due not only to the buffering groups on the surface of hemoglobin but also to additional factors. Whatever the nature of these additional factors, they need not rely on the large size of hemoglobin since L-histidine exhibits rate constants &, and &,t which also level off at a low concentration, near 2 x lo-" M, and cause an enhancement of ticat similar to that observed with hemoglobin, as evident in comparing Figs. 2 and 4. L-Histidine exemplifies the nature of the problem and is more simple to consider than hemoglobin. From Fig. 4, the increase in ecat caused by the addition of histidine is about 3.0 x lo-" s-l, corresponding to an increase in the rate of the catalyzed hydration reaction at equilibrium of 1.4 x lo-' M sL (calculated using Equation 8 of Ref. 5). The rate constant k, for proton transfer between the imidazole group of histidine and the enzyme active site must be large enough to account for this increase in rate. [Bl is the concentration of the buffer histidine. The value of k, obtained in this calculation is larger by a factor of 10' to lo3 than expected for a diffusion-controlled proton transfer between a small molecule and a macromolecule (20). The rate constants for the transfer of protons between the carbonic anhydrase active site and buffers such as 2,2-diethylmalonate, N-methylimidazole, and imidazole are near lo* M-' s-l (4, 5). Assuming that the changes in the rate constants for "0 exchange observed in these experiments are due to proton transfer effects, the efficiency of micromolar concentrations of histidine and hemoglobin in enhancing the activity of carbonic anhydrase can be explained. One possibility is that more than one intermolecular proton transfer step occurs for each encounter of hemoglobin or histidine with carbonic anhydrase. Since these are equilibrium experiments, it is possible that there is cycling of protons between these proton transfer agents and the active site during which more than one catalytic hydration-dehydration cycle occurs and more than 1 atom of '*O is exchanged from CO, to water. This implies that histidine provides more efficient proton cycling than imidazole alone or His 63 in the active site of carbonic anhydrase. This hypothesis requires binding of hemoglobin and histidine to carbonic anhydrase of duration comparable to the time for many catalytic cycles. An association or binding of hemoglobin with carbonic anhydrase has not been demonstrated to this date. If this explanation that hemoglobin and histidine cycle protons during the catalytic hydration-dehydration cycle is correct, we would not expect to observe the 1090-fold greater efficiency of these proton transfer agents compared to imidazole under initial velocity conditions for which there is unidirectional catalysis. Work to investigate this point is in progress.
The plateau regions of Figs. 2 to 4 show that neither bovine hemoglobin nor human serum albumin nor histidine inhibit bovine carbonic anhydrase significantly at the concentrations used. Furthermore, the maximal value of 0,,, is nearly as great in Fig. 2 Fig. 1. This exchange relies on the entry into the active site of '*O-labeled substrate. Hence, it appears that the accessibility of CO, and HCO,-into the carbonic anhydrase active site is not significantly hampered by the presence of hemoglobin. From Fig. 4, a similar result was obtained for histidine. The maximal increment in 0,,, caused by the presence of human serum albumin, Fig. 3, is not as great as that observed with hemoglobin, an observation which we cannot explain at this time.

as in
About half of human oxyhemoglobin is dissociated into dimers at concentrations of lo-" M; a smaller fraction is dissociated at higher concentrations (21). At concentrations as low as those of Fig. 2, dissociation of hemoglobin occurs and will result in exposure of more residues to the solvent. This may cause a small effect on the rate of proton transfer involving hemoglobin. On the other hand, carbamate formation is not expected to influence the rate constants obtained from Fig. 2. First, the ratio of hemoglobin to carbon dioxide concentrations is low (lo-" to lo-'); and second, the mechanism of carbamate formation is such that it is not expected to cause an exchange of oxygen between species of CO2 or between CO, and water (22). We conclude that hemoglobin is an efficient proton transfer agent for the carbonic anhydrase active site; in these equilibrium experiments it is more efficient than can be accounted for by summing the expected effects of the buffering groups on its surface. It follows that, under nonequilibrium conditions, hemoglobin can act as an acceptor of protons for the carbonic anhydrase active site during the catalytic hydration of CO, and as a donor of protons during catalytic dehydration of HCO:,-. Furthermore, these protons can be transferred at diffusion-controlled rates as a result of the encounter of hemoglobin with carbonic anhydrase, although this transfer may be mediated by water bridges or amino acid side chains. Hemoglobin and carbonic anhydrase are the two most abundant proteins in red cells, the former present at about 5 mM and the latter at about 150 PM; hence we anticipate that proton transfer occurs between them in red cells. This may have significance in the Bohr shift, the effect of pH on the affinity of hemoglobin for 0,. Forster and Steen showed that in red cells the Bohr off-shift in response to an increase in external (H+) relies on the production of protons in the cell by the hydration of CO, catalyzed by carbonic anhydrase (23). In the presence of carbonic anhydrase inhibitors, the rate of the Bohr shift is dramatically decreased in the same experiment (23), demonstrating a link in the functions of carbonic anhydrase and hemoglobin. The work reported here points out that the transfer of the Bohr proton at diffusion-controlled rates may occur, at least in part, as a result of the direct encounter of hemoglobin with carbonic anhydrase.