4229
Elucidation of the Catalytic Mechanism of Carbonic Anhydrase’
Through the same infrared “window” in the spectrum of water-protein mixtures, we also observed the following absorption peaks due to the asymmetric stretching Sir: of azide ions in aqueous environment: free N3- in Since its proposal by Quastel,2 the “strained molewater, 2049 cm-l; N3- in inert protein solution, 2046 cule” hypothesis has been assumed by many investigacm-1; N3- bound to carbonic anhydrase, 2094 cm-l; tors to account for the relatively low activation energy N3- bound to the Co(I1)-enzyme, 2082 cm-’; N3of enzymatic reactions. By using accurate differential bound to the diethylenetriamine-zinc(I1) complex, infrared spectrometry, we have detected the absorption DETA-Zn(II), 2085 cm-l; N3- bound to DETAat 2341 cm-l (precision A0.5 cm-l, accuracy * 1 Co(II), 2067 cm-’; HN3 in water, 2147 cm-l. These cm-’) due to the asymmetric stretching of the COZ frequencies show that the N3- bound to carbonic anmolecule bound in a hydrophobic cavity at the active hydrase is coordinated to the Zn(I1) of the enzyme. The site of bovine carbonic anhydrase. Because this value concentration of the azide bound to carbonic anhydrase is very close to the corresponding observed frequency in an equilibrium mixture can be calculated directly of 2343.5 cm-’ for COz dissolved in water, we conclude from the intensity of the 2095-cm-l peak, which exists that the C 0 2 bound at the active site of carbonic anhyside by side with the 2046-cm-’ peak due to the free drase is not under appreciable strain. azide in the mixture. In a typical experiment, the infrared sample cell with For a given enzyme solution at constant temperature, 0.075-mm path length and CaFz windows was filled COz pressure, and pH, the difference peak at 2341 cm-l with 33 % by weight aqueous bovine carbonic anhydrase due to bound COz was observed to decrease as the 2095(Worthington) solution at pH 5.5 under equilibrium cm-‘ peak due to bound N3- increases in such a way COZ pressure. The reference cell, which was adjusted that there is a 1 : 1 correspondence between the per cent by interference method to the same path length (within of Zn(I1) coordinated to N3- and the per cent of COz 0.1 p ) , was similarly filled with the same enzyme soludisplaced from the hydrophobic cavity. Knowing that tion containing either the stoichiometric amount of the N3- is coordinated to the Zn(II), we deduce that the Ethoxyzolamide (Upjohn, 6-ethoxybenzothiazo-2-sul- hydrophobic cavity must be right next to the Zn(I1) fonamide) or an excess of sodium azide. During each so that the ligand N3- can protrude into the cavity and measurement, the Perkin-Elmer 125 spectrometer was sterically displace the COZ. continually flushed with NS and the cells were proNitrate and bicarbonate were found in these infrared tected by germanium cut-off filters and water cooling studies to displace both the N3- from the Zn(I1) and jackets. After spectral measurements, the enzyme the COz from the hydrophobic cavity. Consequently solutions were found to retain the initial concentration we conclude that the bicarbonate ion must be coordiof soluble protein (within =k2% OD at 280 mp) and, nated to the Zn(I1) through its negatively charged oxygen for the uninhibited solution, the original enzyme activatom such that its relatively neutral oxygen atom and ity (within *4% by esterase assay3). The intensity of hydroxyl group are placed in the hydrophobic cavity, as this difference infrared absorption peak gives a direct illustrated by I1 in Figure 1. measure of the concentration of carbonic anhydraseCOZcomplex (E-COZ). The enzyme concentration was determined by Ethoxyzolamide titration of the diluted ample.^ A linear plot of l/[E-CO2] us. l/pco9 shows that each enzyme molecule binds only one COz at its active site with a dissociation constant of 3 f 0.5 atm in 33% solution at pH 5 . 5 and 25”, which is eight times the literature value of K, in very dilute solutions.5 Similar experiments with COS-equilibrated aqueous solutions of a-chymotrypsin and ovalbumin, respectively, did not detect any differential infrared spectrum. Ir I The infrared spectra of carbonic anhydrase solutions Figure 1. The catalytic mechanism of carbonic anhydrase. equilibrated with COS NzO mixtures show that these two gases compete with similar affinity for the same binding site. Consequently, in view of the similar size, Therefore in the dehydration reaction proton transfer shape, small dipole moment, and weak ligand field of must accompany the breaking of the C-0 bond to these two molecules, we infer that the C 0 2 (or NzO) is leave an OH- coordinated to the Zn(II), because we bound in a hydrophobic cavity of the protein, not coalready know from the above results that only COSis to ordinated to the Zn(I1) of the enzyme. This conclusion be left in the hydrophobic cavity as a result of the reacis also consistent with the observation that the binding tion. Conversely, because of the principle of detailed of COS to the Co(I1)-substituted carbonic anhydrase at balancing, it must be the OH- on the Zn(I1) which atlow pH does not affect the visible spectrum of the latter. tacks the bound CO2 and converts the latter to HCOain the reverse hydration reaction. (1) This work was supported by research grants from the U. S. Public
*
+
Health Service (GM-04483) and the National Science Foundation (GB3631). (2) J. H. Quastel, Biochem. J., 20, 166 (1926). (3) B. G. Miilmstrom, P. 0. Nyman, B. Strandberg, and B. Tilander in “Structure and Activity of Enzymes,” T. W. Goodwin, et al., Ed., Academic Press, Inc., New York, N. Y., 1964, p 121. (4) S . Linskgg, J. B i d . Chern., 238, 945 (1963). (5) J. C. Kernohan, Biochem. Biophys. Acta, 118, 405 (1966).
(6) National Science Foundation Predoctoral Fellow, 1966-1967.
Mary Ellen Riepe,6 Jui H. Wang Kline Chemistry Laboratory, Yale Unicersity New Haven, Connecticut 06520 Receiaed May 31, 1967
Communications to the Editor
