Proton magnetic resonance studies of carbonic anhydrase. I. Identification of histidine resonances.

Nuclear magnetic resonance (nmr) spectra of human carbonic anhydrase B recorded in deuterium oxide reveal seven discrete single proton resonances between 7 and 9 ppm downfield from sodium 2,2-dimethyl-i-silapentane-5-sulfonate. Simplification of spectra by use of Fremy's salt, comparison of peak widths at intersections, and evaluation of the results of inhibition and modification experiments permit determination of the pH dependencies of these resonances. Five of these peaks change position with increasing pH; three move upfield by approximately 95 Hz and two move downfield by 10 and 23 Hz. The first three reflect residues with pK values of 7.23, 6.98, and 6 and can be assigned to the C-2 protons of histidines. The two remaining pH dependent resonances reflect groups with pK values of 8.2 and 8.24. Their line widths and T1 values are comparable to those of the first group, and they also appear to reflect C-H protons of histidines. Despite the structural and functional similarities of the B and C isozymes of human carbonic anhydrase, few of the low field resonances appear to be common to both. Six histidine C-2 protons are observed in the C enzyme and reflect groups with pK values of approximately 7.3, 6.5, 5.7, 6.6, 6.6, and 6.4. A seventh peak contains two protons and moves upfield with increasing pH without titrating. A final resonance to low field moves downfield with increasing pH and reflects a group with a pK between 6 and 7. Its behavior resembles that of peak 1 of the human B enzyme, and it also appears to be a histidine C-H proton. This peak may reflect a conserved residue in the two isozymes that plays an important role in enzymatic function, as discussed in the following paper.     

Proton magnetic resonance studies of carbonic anhydrase. III. Binding of sulfonamides.

Resonances of the histidine region of human carbonic anhydrase B have been studied by proton magnetic resonance spectroscopy in the presence of seven sulfonamide inhibitors. Results of difference spectroscopy and observation of the C-2 resonance of an additional titratable histidine in some of these spectra suggest a conformational change in the enzyme, while the large number of unaltered resonances indicates involvement of only a few residues. Inhibition of carbonic anhydrase by sulfonamides appears to involve: stabilization of an appropriately oriented initial complex by hydrophobic binding of the aromatic ring of the inhibitor to residues of the cavity forming the active site; ionization of the sulfonamido group, facilitated by its proximity to zinc; protonation and displacement of the high pH ligand to the metal controlling catalytic activity, thought here to be a histidine residue; and formation by the sulfonamido group of an ionic bond to zinc and a hydrogen bond to the hydroxyl group of serine or threonine. Diversity of spectra produced with various sulfonamides suggests that substituents on the ring and heteroatoms within the ring interact with additional groups at the active site. Increase in inhibitory potency appears to involve optimizing the number as well as the strength of these interactions. An upper limit for the dissociation rate of these complexes of 10 sec-1 was obtained.     

A proposal for the mechanism of carbonic anhydrase action

A new catalytic mechanism is proposed for the hydration of CO₂ by the zinc metalloenzyme carbonic anhydrase. This mechanism identifies the group controlling catalytic activity as an active site histidine, in which the protonated imidazole ring coordinates to zinc, losing a proton. Geometric constraints on the histidine unit make the metal-ligand bond a strained and, therefore, labile one. In the hydration reaction, the metal-bound neutral histidine moiety serves as a proton acceptor for the transient ionization of metal-bound water. Zinc-bound hydroxide attacks the carbon of the substrate to generate metal-bound bicarbonate, and the system regenerates itself by losing the elements of carbonic acid.     

Effects of two ionizing groups on the active site of human carbonic anhydrase B.

Investigation of some pH-dependent properties of human erythrocyte carbonic anhydrase B indicate that the active site is influenced by at least two charged groups. The properties studied include the pH dependence of inhibition of native, monocarboxamidomethyl, and monocarboxymethyl enzymes by iodide ion and the pH dependence of the visible spectra of the cobalt derivatives of these enzymes. One ionizing group has a pKa of about 7.3 in the native enzyme, 8.2 in the carboxyamidomethyl enzyme, and 9.0 in the carboxymethyl enzyme. It has a major influence on activity and anion inhibition, and on the visible spectra of the cobalt enzymes. A second group has a pKa of about 6.1 in native and modified enzymes. When zinc is at the active site, the secondary group in its acidic form decreases the Ki for I-. With the carboxyamidomethyl and carboxymethyl enzymes, the Ki decreases by about an order of magnitude. However, if cobalt is substituted for zinc in the modified enzymes, this group does not influence the Ki for I- and the binding of I- does not influence the pKa of the spectral transitions caused by ionization of this secondary group. In the case of nonalkylated Co2+-enzyme, another ionizing group with a pK of about 6.2 prevents the binging of I- at low pH. These results show that the active site is altered when cobalt is substituted for zinc in carbonic anhydrase B.     

The esterase activity of bovine carbonic anhydrase B above pH 9. Reversible and cooalent inhibition by acetozolamide.

The reversible complex between the metalloenzyme bovine carbonic anhydrase B and the sulfonamide inhibitor acetazolamide can be "frozen" irreversibly by the addition of a covalent bond between the methyl group of the inhibitor and the tau-nitrogen of histidine-64. In both cases the inhibited enzyme is inactive as an esterase toward p-nitrophenyl propionate at physiological pH but retains activity controlled by an ionization in the protein exhibiting a pK-a greater than 10. Similarly, both the covalently and reversibly inhibited enzymes in which the catalytically essential Zn(II) ion has been replaced with Co(II) display the same visible absorption spectrum which is invariant over the pH range from 5 to 12. The evidence therefore indicates that the position of the acetazolamide moiety in the active site is independent of both pH and the presence of the covalent bond to histidine-64. Moreover, when reversibly bound, this inhibitor has been shown to replace the water molecule (or hydroxide ion) known to occupy the fourth coordination position of the metal ion and frequently implicated in the catalytic mechanism of carbonic anhydrases. Thus, the activity exhibited by the inhibited enzymes and consequently the second rise observed in the pH rate profile of the native enzyme above pH 0 cannot reflect the ionization of such a water molecule in contrast to what has been postulated previously (Pocker, Y., and Storm, D. R. (1968) Biochemistry 7, 1202-1214). Displacement of the zinc-bound solvent molecule rather than the alkylation of histidine-64 is suggested, however, as the cause of the inactivation of the alkylated enzyme round neutrality. Taken together, the biphasic pH rate profile of native bovine carbonic anhydrase B as well as the activity retained by the alkylated enzyme above pH 9 are best described by a model in which two groups in the enzyme ionize independently, thereby raising the possibility that the high pH activity is controlled by an ionization outside the active site region of the enzyme. Above pH 9.5 the pK; for the reversible interaction between native carbonic anhydrase and acetazolamide falls off linearly with increasing pH. The slope of --1.56 suggests that, among other factors, more than one ionization is responsible for the descending limb of the pH-i-pH profile.     

Magnetic resonance study of exchangeable protons in human carbonic anhydrases.

A titratable exchangeable proton resonance assignable to a histidine imidazole ring N--H proton is observed approximately minus 15 ppm downfield from tetramethylsilane. The chemical shift of this resonance is affected by sulfonamide and anion inhibitors, and by removal of zinc or replacement of zinc by cobalt, indicating that the proton is located at or near the active site. The pH dependence of the chemical shift of this resonance, which is abolished by inhibitors, reflects the titration of a group with a pK-a of 7.3 in human carbonic anhydrase B and smaller than or equal to 7.1 in human carbonic anhydrase C. These pK-a values are interpreted to be due to the ionization of a neutral imidazole to form the imidazolate anion coordinated to zinc. A mechanism for enzymatic catalysis involving reversible deprotonation and coordination of a histidine to the metal is consistent with these studies.     

Differential modification of specificity in carbonic anhydrase catalysis

Incubation of carbonic anhydrase II with acrolein results in a rapid, time-dependent loss of all but approximately 3-6% of the original catalytic activity toward CO2 hydration and HCO3- dehydration, with the inactivation rate being first-order in both acrolein and the enzyme. The pH dependence of the inactivation rate constant can be adequately described with a function incorporating a pK alpha of 7.15 and a maximal value for kinact [corrected] of 26.2 M-1 min-1, indicating that at least one of the catalytically essential residues that ionizes at this pH is involved in the modification scheme. The amount of residual CO2 hydratase activity is proportional to the molar excess of acrolein over carbonic anhydrase II with 5 histidyl and 3 lysyl residues being subject to alkylation under conditions where [acrolein] to [carbonic anhydrase II] ratio is greater than 100. Because all lysyl residues were shown previously to be amidinated without detectable loss of activity, it was assumed that the modification of one (or more) of the histidines was primarily responsible for the observed inactivation. The number of modified histidyl residues could be related to residual activity by using the statistical analysis of Tsou (Tsou, C.-L. (1962) Sci. Sin. (Engl. Ed.) 11, 1535-1558) which indicates that one essential histidine reacts approximately four times faster than the other (histidyl) residues. In sharp contrast with the phenomenon observed in connection with CO2 hydration and HCO3- dehydration, acrolein improves the catalytic efficiency of the enzyme toward p-nitrophenyl acetate hydrolysis and acetaldehyde hydration, with the relative activity increasing by approximately 12 and 34%, respectively. The widely differing effects imparted by the same reagent represent the first step toward differential control of the specificity of carbonic anhydrase II.     

Proton transfer to residues of basic pK(a) during catalysis by carbonic anhydrase.

The maximal velocity in the hydration of CO(2) catalyzed by the carbonic anhydrases in well-buffered solutions is limited by an intramolecular proton transfer from zinc-bound water to acceptor groups of the enzyme and hence to buffer in solution. Stopped-flow spectrophotometry was used to accumulate evidence that this maximal velocity is affected by residues of basic pK(a), near 8 to above 9, in catalysis of the hydration of CO(2) by carbonic anhydrases III, IV, V, and VII. A mutant of carbonic anhydrase II containing the replacement His-64-->Ala, which removes the prominent histidine proton shuttle (with pK(a) near 7), allows better observation of these basic groups. We suggest this feature of catalysis is general for the human and animal carbonic anhydrases and is due to residues of basic pK(a), predominantly lysines and tyrosines more distant from the zinc than His-64, that act as proton acceptors. These groups supplement the well-studied proton transfer from zinc-bound water to His-64 in the most efficient of the carbonic anhydrases, isozymes II, IV, and VII.     

Enhancement of the catalytic properties of human carbonic anhydrase III by site-directed mutagenesis

Among the seven known isozymes of carbonic anhydrase in higher vertebrates, isozyme III is the least efficient in catalytic hydration of CO2 and the least susceptible to inhibition by sulfonamides. We have investigated the role of two basic residues near the active site of human carbonic anhydrase III (HCA III), lysine 64 and arginine 67, to determine whether they can account for some of the unique properties of this isozyme. Site-directed mutagenesis was used to replace these residues with histidine 64 and asparagine 67, the amino acids present at the corresponding positions of HCA II, the most efficient of the carbonic anhydrase isozymes. Catalysis by wild-type HCA III and mutants was determined from the initial velocity of hydration of CO2 at steady state by stopped-flow spectrophotometry and from the exchange of 18O between CO2 and water at chemical equilibrium by mass spectrometry. We have shown that histidine 64 functions as a proton shuttle in carbonic anhydrase by substituting histidine for lysine 64 in HCA III. The enhanced CO2 hydration activity and pH profile of the resulting mutant support this role for histidine 64 in the catalytic mechanism and suggest an approach that may be useful in investigating the mechanistic roles of active-site residues in other isozyme groups. Replacing arginine 67 in HCA III by asparagine enhanced catalysis of CO2 hydration 3-fold compared with that of wild-type HCA III, and the pH profile of the resulting mutant was consistent with a proton transfer role for lysine 64. Neither replacement enhanced the weak inhibition of HCA III by acetazolamide or the catalytic hydrolysis of 4-nitrophenyl acetate.     

Structure of native and apo carbonic anhydrase II and structure of some of its anion-ligand complexes.

In order to obtain a better structural framework for understanding the catalytic mechanism of carbonic anhydrase, a number of inhibitor complexes of the enzyme were investigated crystallographically. The three-dimensional structure of free human carbonic anhydrase II was refined at pH 7.8 (1.54 A resolution) and at pH 6.0 (1.67 A resolution). The structure around the zinc ion was identical at both pH values. The structure of the zinc-free enzyme was virtually identical with that of the native enzyme, apart from a water molecule that had moved 0.9 A to fill the space that would be occupied by the zinc ion. The complexes with the anionic inhibitors bisulfite and formate were also studied at neutral pH. Bisulfite binds with one of its oxygen atoms, presumably protonized, to the zinc ion and replaces the zinc water. Formate, lacking a hydroxyl group, is bound with its oxygen atoms not far away from the position of the non-protonized oxygen atoms of the bisulfite complex, i.e. at hydrogen bond distance from Thr199 N and at a position between the zinc ion and the hydrophobic part of the active site. The result of these and other studies have implications for our view of the catalytic function of the enzyme, since virtually all inhibitors share some features with substrate, product or expected transition states. A reaction scheme where electrophilic activation of carbon dioxide plays an important role in the hydration reaction is presented. In the reverse direction, the protonized oxygen of the bicarbonate is forced upon the zinc ion, thereby facilitating cleavage of the carbon-oxygen bond. This is achieved by the combined action of the anionic binding site, which binds carboxyl groups, the side-chain of threonine 199, which discriminates between hydrogen bond donors and acceptors, and hydrophobic interaction between substrate and the active site cavity. The required proton transfer between the zinc water and His64 can take place through water molecules 292 and 318.     

Similarities in active center geometries of zinc-containing enzymes, proteases and dehydrogenases.

The spatial environment of the active centers for the four zinc-containing enzymes carbonic anhydrase, liver alcohol dehydrogenase, thermolysin and carboxypeptidase were compared and contrasted. The zinc is co-ordinated by three protein groups. In addition, a water molecule and substrate carbonyl may assume a fourth or fifth position. A group whose function is to abstract a proton from water during catalysis was found to have a constant spatial arrangement with respect to the zinc atom. The co-ordination sphere around the zinc is systematically distorted from a regular tetrahedral geometry with one specific ligand position being invariably occupied by a histidine residue. The orientation of the imidazole ring is moderately constant with respect to the Zn pyramid, a constraint possibly imposed by the adjacent substrate to permit its positioning suitable for catalysis.The comparison of carboxypeptidase and thermolysin was previously reported (Kester and Matthews, 1977a). The position of the water molecule as found in liver alcohol dehydrogenase when placed in thermolysin or carboxypeptidase would be consistent with a transient pentagonal Zn co-ordination during catalysis.Comparison of carboxypeptidase and carbonic anhydrase showed that the specificity pocket of carboxypeptidase superimposed onto a hydrophobic cavity of unknown function in carbonic anhydrase. The glycyl-l-tyrosine pseudo-substrate of carboxypeptidase fits well into the cavity, suggesting a probable binding site for esters in carbonic anhydrase. The excellent esterase activity of both these enzymes can thus be explained by a common binding mode and arrangement of catalytic groups.A comparison of trypsin and thermolysin demonstrates that, although their functional groups differ in character, the peptidase activity could be catalyzed in a similar manner. The proton-abstracting function of His57 in trypsin is generated by Glul43 acting on the Zn co-ordinated water, while the proton donor function of His57 in trypsin is generated by His231 in thermolysin.A comparison of liver alcohol dehydrogenase with other dehydrogenases suggests that His51 is not only a proton sink but also electronically provides an essential positive charge at crucial moments during catalysis. In contrast Arg 109 of lactafce dehydrogenase performs the same function by virtue of a conformational change. The superposition indicates that the zinc co-ordinated water oxygen has the proton acceptor function in liver alcohol dehydrogenase corresponding to the essential histidine groups in lactate dehydrogenase and glyceraldehyde-3-phosphate dehydrogenase.A compendium of the handedness of the catalytic configuration about the reactive atoms for ten different enzymes has been tabulated.     

The active site structure of Thalassiosira weissflogii carbonic anhydrase 1

X-ray absorption spectroscopy at the Zn K-edge indicates that the active site of the marine diatom Thalassiosira weissflogii carbonic anhydrase is strikingly similar to that of mammalian alpha-carbonic anhydrase enzymes. The zinc has three histidine ligands and a single water at 1.98 A. This is quite different from the beta-carbonic anhydrases of higher plants in which zinc is coordinated by two cysteine thiolates, one histidine, and a water molecule. The diatom carbonic anhydrase shows no significant sequence similarity with other carbonic anhydrases and may represent an example of convergent evolution at the molecular level.     

Proton nuclear magnetic resonance studies of histidines in horse carbonic anhydrase I

The 250 MHz 1H-NMR spectrum of horse carbonic anhydrase I (or B) (carbonate hydro-lyase, EC 4.2.1.1) was measured as a function of pH under various conditions. Eight resonances corresponding to histidine C-2 protons and four resonances corresponding to histidine C-4 protons were identified and assigned to individual histidine residues in the enzyme molecule. Substantial similarities between horse and human carbonic anhydrases I were demonstrated. While the human enzyme has three titratable histidine residues in its active site, the horse enzyme has only two, His-67 in the human enzyme being replaced by Gln in the horse enzyme (Jabusch, J.R., Bray, R.P. and Deutsch, H.F. (1980) J. Biol. Chem. 255, 9196-9204). This substitution has small but significant effects on the behaviour of the other active-site histidines. His-64 and His-200. However, His-64 has an anomalously low pKa value also in horse isoenzyme I, as previously observed in human isoenzyme I (Campbell, I.D., Lindskog, S. and White, A.I. (1974) J. Mol. Biol. 90, 469-489).     

Hydrolysis of 4-nitrophenyl acetate catalyzed by carbonic anhydrase III from bovine skeletal muscle

We report three experiments which show that the hydrolysis of 4-nitrophenyl acetate catalyzed by carbonic anhydrase III from bovine skeletal muscle occurs at a site on the enzyme different than the active site for CO2 hydration. This is in contrast with isozymes I and II of carbonic anhydrase for which the sites of 4-nitrophenyl acetate hydrolysis and CO2 hydration are the same. The pH profile of kcat/Km for hydrolysis of 4-nitrophenyl acetate was roughly described by the ionization of a group with pKa 6.5, whereas kcat/Km for CO2 hydration catalyzed by isozyme III was independent of pH in the range of pH 6.0-8.5. The apoenzyme of carbonic anhydrase III, which is inactive in the catalytic hydration of CO2, was found to be as active in the hydrolysis of 4-nitrophenyl acetate as native isozyme III. Concentrations of N-3 and OCN- and the sulfonamides methazolamide and chlorzolamide which inhibited CO2 hydration did not affect catalytic hydrolysis of 4-nitrophenyl acetate by carbonic anhydrase III.     

Metal substitutions incarbonic anhydrase: a halide ion probe study.

³⁵Cl nmr line width and activity measurements are reported for the zinc, mercury and cadmium forms of bovine and human carbonic anhydrase B. The zinc data agree well with previous reports; however, there is no ³⁵Cl nmr line broadening observed for the cadmium and mercury derivatives of the enzyme which are inactive in the presence of excess zinc. The results suggest altered coordination geometry or first coordination sphere saturation by protein ligands for the cadmium and mercury derivatives of the enzyme.     

Proton magnetic resonance studies of histidines in human, rhesus monkey, and bovine carbonic anhydrases.

Histidine C-2 proton resonances in rhesus monkey carbonic anhydrase B (carbonate hydro-lyase, EC 4.2.1.1) and bovine carbonic anhydrase were investigated using 270-MHz proton magnetic resonance. The results suggest that there are extensive three-dimensional homologies between the human B and rhesus B enzymes and between the human C and bovine enzymes. Resonances from solvent exchangeable protons have been observed in the 11-16 ppm range in the NMR spectra of human carbonic anhydrases B and C and bovine carbonic anhydrase. Up to five of these are sensitive to changes of pH and the presence of inhibitors. Three of these resonances are assigned to NH protons of the metal coordinated imidazole groups. These results are discussed in relation to various models for the catalytic mechanism of carbonic anhydrase.     

Purification and Properties of Three Pig Erythrocyte Carronic Anhydrases

Three distinct forms of the zinc containing enzyme carbonic anhydrase were isolated from pig erythrocytes. One low activity type enzyme and two genetic variants of the high activity type enzyme with identical CO2 hydratase activities which were 8 times as high were isolated from Danish Black and White Swine. In the isolation procedure described, the hemoglobin was eliminated by precipitation with chloroform-ethanol, and the isoenzymes were separated by DEAE-Sephadex chromatography. A number of enzymatically active minor components were separated. They were apparently all genetically linked to one of the three major components. The three purified isoenzymes behaved as homogeneous components during isoelectric focusing and electrophoresis at different pH values. They were characterized in terms of molecular weight, isoelectric pH, zinc content, amino acid composition, and enzymatic activity against CO2, p-nitrophenyl acetate, and β-napthyl acetate. The circular dichroism of the enzymes in the ultraviolet region was studied. The properties of the enzymes were similar to those of carbonic anhydrases of corresponding types isolated from other mammalian species. Sulphur containing amino acid residues were absent in the low activity type enzyme. The amino acid composition of the two high activity mutants deviated only in that an arginine residue in the most widespread genetic variant was replaced by a histidine residue in the less frequent variant. Otherwise the two mutants showed identical properties.     

Temperature and requency dependence of solvent proton relaxation rates in solutions of manganese(II) carbonic anhydrase.

Longitudinal and transverse proton relaxation rates of water in solutions of manganese(II) bovine carbonic anhydrase have been measured by pulsed nuclear magnetic resonance spectrometry as a function of temperature (2-35 degrees), frequently (5-100 MHz) and pH. The pH dependence of the longitudinal relaxation rate was fitted to a sigmoidal curve with a pK value at 7.8, while the esterase activity of the manganese(II) enzyme in the hydrolysis of p-nitrophenyl acetate revealed an inflection point at pK = 8.2. The hydration number of manganese(II) carbonic anhydrase could be derived using either the frequency dependence of T1p or the T1p/T2p ratio at only one (high) frequency. Both treatments are in agreement with a model in which one water molecule is bound to the metal at high pH. At low pH the relaxation data imply that no-H20 exists in the first coordination sphere of the manganese ion. The various parameters which are responsible for the proton relaxation mechanisms have been evaluated and are compared to other manganese(II) enzyme systems. The pH dependence of the binding constant of manganese to apocarbonic anhydrase is also reported.     

A study of the histidine residues of human carbonic anhydrase B using 270 MHz proton magnetic resonance

Nine resonances in the 270 MHz proton magnetic resonance spectrum of human carbonic anhydrase B have been identified with imidazole C₍₂₎ protons of histidine residues, six of which are observed to titrate with pK_a values in the range 4.7 to 7.4. The behaviour of the nine resonances has been studied in the presence of the inhibitors, iodide, cyanide, acetate, hexacyanochromate, and imidazole. Measurements have also been made of the enzyme in its apo, cobalt, and mono-alkylated forms. Used in conjunction with the crystal structure, these results have enabled the tentative assignment of all nine resonances to particular histidine residues in the amino-acid sequence. Three of the active-site histidines at positions 64, 67, and 200 have low pK_a values and cannot be directly linked to the activity of the enzyme. However, the resonances assigned to the three metal-liganding histidines do exhibit changes on anion binding and with pH, which parallel changes in the esterase activity. These results are consistent with the model of an ionizable water molecule bound to the zinc ion.Linewidth measurements of the resonances of the histidine residues on the enzyme surface are used to estimate pseudo-first-order rate constants of the order of 4 × 10³ s^?1 for D⁺ exchange between imidazole N and solvent in the absence of buffer. These rates are observed to increase in the presence of small amounts of the buffers Tris and imidazole.     

Essential histidine residue in 3-ketosteroid-delta 1-dehydrogenase.

The variation with pH of kinetic parameters was examined for 3-ketosteroid-delta 1-dehydrogenase from Nocardia corallina. The Vmax/Km profile for 4-androstenedione indicates that activity is lost upon protonation of a cationic acid-type group with a pK value of 7.7. The enzyme was inactivated by diethylpyrocarbonate at pH 7.4 and the inactivation was substantially prevented by androstadienedione. Analyses of reactivation with neutral hydroxylamine, pH variation, and spectral changes of the inactivated enzyme revealed that the inactivation arises from modification of a histidine residue. Studies with [14C]diethylpyrocarbonate provided support for the idea that the 1-2 essential histidine residues are essential for the catalytic activity of the enzyme. Dye-sensitized photooxidation led to 50% inactivation of the enzyme with the decomposition of two histidine residues. This inactivation was also prevented by androstadienedione. Dancyl chloride caused a loss of the enzyme activity. Modifiers of glutamic acid, aspartic acid, cysteine, and lysine did not affect the enzyme activity. Butanedione and phenylglyoxal in the presence of borate rapidly inactivated the enzyme, indicating that arginine residues also have a crucial function in the active site. The data described support the previously proposed mechanism of beta-oxidation of 3-ketosteroid.