Characterization of a novel Ser-cisSer-Lys catalytic triad in comparison with the classical Ser-His-Asp triad.

Amidase signature family enzymes, which are widespread in nature, contain a newly identified Ser-cisSer-Lys catalytic triad in which the peptide bond between Ser131 and the preceding residue Gly130 is in a cis configuration. In order to characterize the property of the novel triad, we have determined the structures of five mutant malonamidase E2 enzymes that contain a Cys-cisSer-Lys, Ser-cisAla-Lys, or Ser-cisSer-Ala triad or a substitution of Gly130 with alanine. Cysteine cannot replace the role of Ser155 due to a hyper-reactivity of the residue, which results in the modification of the cysteine to cysteinyl sulfinic acid, most likely inside the expression host cells. The lysine residue plays a structural as well as a catalytic role, since the substitution of the residue with alanine disrupts the active site structure completely. The two observations are in sharp contrast with the consequences of the corresponding substitutions in the classical Ser-His-Asp triad. Structural data on the mutant containing the Ser-cisAla-Lys triad convincingly suggest that Ser131 plays an analogous catalytic role as the histidine of the Ser-His-Asp triad. The unusual cis configuration of Ser131 appears essential for the precise contacts of this residue with the other triad residues, as indicated by the near invariance of the preceding glycine residue (Gly130), structural data on the G130A mutant, and by a modeling experiment. The data provide a deep understanding of the role of each residue of the new triad at the atomic level and demonstrate that the new triad is a catalytic device distinctively different from the classical triad or its variants.

A large group of enzymes, designated as amidase signature (AS) 1 family enzymes, are widely distributed throughout the hierarchy of living organisms. A search against the Swiss-Prot data base with the BLAST algorithm (1) identifies over 200 different AS family proteins from 90 different organisms. These enzymes are characterized by the AS sequence, which is a conserved stretch of ϳ130 amino acids (2,3). The identified biochemical function of the AS family enzymes has commonly been the hydrolysis of the amide bond (CO-NH 2 ). The known biological functions of the AS family enzymes vary widely, including the formation of Gln-tRNA Gln by the transamidation of misacylated Glu-tRNA Gln via amidolysis of glutamine in many different bacteria (4,5), catabolism of neuromodulatory fatty acid amides in mammals (6 -9), the formation of indole-3-acetic acid in pathogenic plant bacteria (10), and the metabolic turnover of carbon/nitrogen-containing compounds in both prokaryotes and eukaryotes (2,11). One member of this family, nitrile hydratase, is used in the industrial scale production of acrylamide and nicotinamide (12).
Despite the functional variety and the wide evolutionary distribution of the AS family enzymes, which rival that of the other classical serine hydrolases, e.g. trypsin-like serine protease family containing about 270 different members in the Swiss-Prot data base, little is known about the structure-function relationship of the AS family enzymes. Recently, we reported the structure of malonamidase E2 (MAE2) from Bradyrhizobium japonicum, the first structure of an AS family enzyme (13). The structure revealed a novel catalytic triad Ser-cisSer-Lys. The peptide bond between the second serine, Ser 131 , and the preceding residue, Gly 130 , is in an unusual cis configuration. The novel catalytic triad had not been observed in any other known hydrolytic enzymes. Subsequently, the structures of two other AS family enzymes have been reported: peptide amidase of Stenotrophomonas maltophilia (14) and rat fatty acid amide hydrolase (15), both of which revealed the cis configuration of Ser 131 . The triad residues, found exclusively on the AS sequences, are absolutely conserved, and no variation of the chemical makeup of the triad has been observed. This is in a sharp contrast with the classical Ser-His-Asp triad, which is often found in several completely dissimilar protein folds (16), and whose variations are found in similar protein folds, such as the Cys-His-Glu triad in human rhinovirus 3C protease, an ␣-chymotrypsin-like protease (17) and Ser-His-Glu triad in a fungal lipase (18). While the classical triad has been characterized in minute detail, little is known about the characteristics of the Ser-cisSer-Lys triad. In this report, we provide a structural basis for understanding the invariability of the chemical makeup of the novel triad and delineate the role of each constituent amino acid in detail. This study led to the conclusions that the novel triad is a catalytic device distinctively different from the classical triad or its variants, and for this reason, that the AS family is evolutionary distinct from any other known hydrolase families.

Site-directed Mutagenesis and Protein
Purification-S155C, S131A, K62A, G130A, and T150A MAE2 mutants were created using the QuikChange kit (Stratagene). The plasmids containing each of the mutant MAE2 genes was transformed into Escherichia coli strain BL21(DE3). The expression of each of the MAE2 mutants was induced by 1 mM isopropyl-␤-D-thiogalactopyranoside at an optical density of 0.6 -0.8 at 37°C for 16 h except for the G130A mutant, which required the induction at 24°C. Bacterial lysates were prepared by sonication in buffer A, 20 mM Tris-HCl solution (pH 7.4) containing 15% glycerol. After centrifugation, the supernatant was applied to a HiTrap TM Blue HP column (Amersham Biosciences) and eluted with a linear gradient from 0 to 2 M of NaCl in buffer A. The fractions containing MAE2 were dialyzed against buffer B, 20 mM Tris-HCl solution (pH 7.4) containing 2 M NaCl. The dialyzed protein sample was loaded onto the RE-SOURCE TM PHE column (Amersham Biosciences) pre-equilibrated with buffer B containing 2 mM ZnSO 4 . The unbound fraction containing MAE2 was concentrated and subsequently loaded on a Superdex 200 HR 20/60 column (Amersham Biosciences) pre-equilibrated with 20 mM Tris-HCl (pH 7.4). The peak fractions were concentrated to 10 mg/ml using centriprep-10K (Amicon). For the purification of the S155C mutant, we included 1 mM dithiothreitol in all the buffers used.
Crystallization, X-ray Data Collection, and Structure Determination-The crystals of the MAE2 mutants were obtained in droplets containing 1.5 l of protein sample (10 mg/ml) and an equal volume of precipitant solution containing 20% polyethylene glycol 1000 and 0.1 M Tris-HCl (pH 7.0). The crystals of the S131A MAE2 mutant in complex with malonamate or with malonate were obtained by the cocrystallization method in the presence of 5 mM malonamate or 15 mM malonate using the same precipitant solution. All crystals were flash-cooled in the stream of cooled nitrogen gas (100 K) prior to x-ray data collection. All data sets were collected on the Rigaku RU300 rotating anode x-ray generator equipped with focusing x-ray mirrors and operated at 50 kV/100 mA using an R-AXIS IV 2ϩ area detector (Molecular Structure Corporation). The data were integrated with DENZO and scaled with SCALEPACK (19). All crystals belong to the space group P2 1 2 1 2, and have almost identical unit cell parameters to those of the wild-type protein (a ϭ 104.29, b ϭ 95.58, and c ϭ 74.90 Å). The asymmetric unit contains one dimer of the enzyme. The structures of the MAE2 mutants were determined by direct refinement of the structure of the wild type against diffraction data using the program CNS (20). The model contained the omission of a mutated residue. The refinement statistics for the final structures are shown in Table I.
Measurement of the Enzymatic Activity-The catalytic activities of the wild-type and the mutant MAE2 enzymes were assayed by measuring the formation of malonohydroxamate using the method described previously (21). The enzymatic reactions for kinetic measurement were started by the addition of protein to 500 l of the assay mixtures, which contains varying concentration of malonamate (4, 6, 10, 12, 14, 16 mM), neutralized 20 mM NH 2 OH, and 0.1 M Tris-HCl (pH 8.0). The reaction mixture was incubated at 37°C for 30 min, and the reaction was stopped by adding 750 l of the coloring reagent solution containing 10% trichloroacetic acid, 0.66 N HCl, and 15% FeCl 3 . When enzyme activity was not detected, the reaction time was extended to 90 min. Any cloudiness was removed by centrifugation prior to the absorbance measurement at 540 nm.

RESULTS AND DISCUSSION
Overall Structure and Catalytic Mechanism-MAE2 is a dimeric protein. The structures of the two subunits are essentially the same. Previously, we showed that the absolutely conserved Ser 155 , Ser 131 , and Lys 62 constitute a catalytic triad, and Ser 155 is the nucleophile initiating the catalysis (13). The conserved AS sequence forms roughly the core of the molecule and interacts with a substantial part of the remainder of the sequence (Fig. 1a). The two most highly conserved segments on the AS sequence are Gly 129 -Gly-Ser-Ser-Ser-Gly 134 and Thr 150 -Gln-Thr-Gly-Gly-Ser 155 (Fig. 1b). The first segment is the loop containing the cis peptide bond between Gly 130 and Ser 131 . The unusual cis configuration allows two direct hydrogen-bonding interactions between Ser 131 and Ser 155 O␥ (Fig. 1a). The second conserved segment contains the oxyanion hole composed of the backbone -NH groups of Thr 152 , Gly 153 , Gly 154 , and Ser 155 . Lys 62 is in direct contact with the two loops via the interactions with the side chains of Ser 131 , Ser 132 , and Thr 150 .
The catalytic mechanism of the classical Ser-His-Asp triad is known to begin with the polarization of the serine residue by the histidine residue, which is initially in the deprotonated form that is able to abstract the hydroxyl proton of the serine residue (22). The aspartate promotes the base role of the histidine by stabilizing the charged imidazole of the histidine. The same mechanism is not conceivable for the Ser-cisSer-Lys triad because the chemical property of the Ser-Lys pair is very different from that of the His-Asp pair. Since the pK a of serine is around 13, cisSer 131 of the Ser-cisSer-Lys triad is most likely in the protonated form at physiological pH and is unlikely to function as a general base to abstract a proton from Ser 155 . The chemical environment around Ser 131 O␥ is not unusually polar to promote deprotonation of the residue. Rather, C␤ of Ala 111 , which is only 3.48 Å apart from Ser 131 O␥, renders the environment partly hydrophobic. We proposed that Ser 155 is in the deprotonated state per se mainly by virtue of the two direct hydrogen-bonding interactions provided by the backbone -NH and the side chain -OH groups of cisSer 131 (Fig. 1a) and by the additional dipolar interactions provided by the oxyanion hole and the guanidine group of Arg 158 (13). Upon the nucleophilic attack of Ser 155 on the carbonyl carbon of the substrate, Ser 131 is proposed to function as the catalytic acid that provides a proton to the leaving amino group, which is firmly supported by where ͉F o ͉ and ͉F c ͉ are the observed and calculated structure factor amplitudes, respectively. R free was calculated with 5% of the data. the present work described below. The proton transfer should be facilitated by the amino group of Lys 62 that can stabilize the deprotonated form of Ser 131 . The proposed mechanism is depicted in Fig. 2. Although we will not rule out other possibilities of polarizing Ser 155 of the triad, the proposed mechanism may best explain the role of the cis peptide bond; allowing the two groups of Ser 131 to simultaneously polarize Ser 155 .
Cysteine Cannot Replace the Role of Ser 155 -In cysteinyl proteases, the serine nucleophile is substituted with cysteine. Several hydrolases of this family, including human ubiquitin C-terminal hydrolase (23), tobacco etch virus protease (24), and turnip mosaic potyvirus NIa (25), are known to contain a Cys-His-Asp triad. Many papain-like cysteine protease family members utilize a Cys-His-Asn catalytic triad. The asparagine residue of the triad was shown to be non-essential for catalysis at least in papain (26), which is probably a reflection of the higher reactivity of cysteine than serine. We tested the suitability of sulfur to function as a nucleophile in the active site environment of MAE2 by mutating Ser 155 to cysteine. Although the cysteinyl thiol is chemically more reactive than the serine hydroxyl in general, the S155C MAE2 mutant was inactive. The structure of the mutant enzyme provides a clear explana-tion to this observation. Cys 155 had extra density extending from the sulfur atom in the initial electron density map at 1.8 Å resolution (Fig. 3). The chemical environment in the vicinity of the extra density immediately indicated that the cysteine residue was modified as cysteine sulfinic acid, Cys-SO 2 H. The side chain of this residue directly interacts with the oxyanion hole and the backbone -NH of Ser 131 . Cysteine sulfinic acid is an oxidation product of cysteine sulfenic acid, Cys-SOH, which is produced from the reaction of a cysteine residue with a peroxide molecule (27). While cysteine sulfenic acid is generally unstable and could be reduced to back to cysteine in the presence of dithiothreitol, cysteine sulfinic acid is not reactive to the reducing agent (28,29). Consistently, the enzyme activity of this mutant was not restored in the presence of 10 mM dithiothreitol. Although we included 1 mM dithiothreitol in all buffer solutions used for the purification of the mutant enzyme, the bacterial cell lysate as well as the purified mutant enzyme were enzymatically inactive. Therefore, Cys 155 must have been modified to Cys-SOH inside the cells used for the production of the mutant protein, and the further oxidation to Cys-SO 2 H might have also taken place during the expression of the protein.
Since the active site is composed of the most highly conserved peptide segments, both the loop containing cisSer 131 and the segment containing the oxyanion hole should adopt very similar conformations in all AS family members. Presumably, cysteine has not been selected as the catalysis-initiating nucleophile by any AS family enzymes because the active site milieu renders the cysteine residue at this position highly reactive, which is unwanted and uncorrectable.
The substitution of the serine of the Ser-His-Asp triad with cysteine results in a 10 6 -fold decrease in the catalytic activity of trypsin. A crystallographic study showed that the cysteine is in the reduced state, and the bulkiness of the cysteinyl sulfur atom obstructs the oxyanion hole (30). This obstruction was suggested to be the main structural ground for the reduced activity of the mutant. The oxyanion hole of MAE2 consists of four backbone-NH groups (Fig. 3), which appears to readily accommodate an oxygen atom of a peroxide molecule to provide a low energy barrier for the reaction of Cys 155 or Cys 155 -OH with the other oxygen atom of the molecule. Two subsequent reactions of Cys 155 with peroxide molecules can lead to the formation of Cys 155 -SO 2 H. In contrast, the oxyanion hole of trypsin consists of two backbone -NH groups and apparently does not support the oxidation of the substituted cysteine to cysteine sulfenic acid or the further oxidation.
The Catalytic Roles of Ser 131 -The structures of the S131A mutant were determined as a complex with the substrate malonamate ( Ϫ OOC-CH 2 -CO-NH 2 ) and with the product malonate ( Ϫ OOC-CH 2 -COO Ϫ ). The shape of the electron density of the bound substrate is readily distinguishable from that of the bound product (Fig. 4). Therefore, the hydrolysis of the substrate by the mutant enzyme did not take place during several days of the substrate soaking into the crystals of the mutant enzyme. The mutant exhibits about 2000-fold lower catalytic efficiency, k cat , compared with the wild-type enzyme. The structure of the mutant shows that Wat 279 is 3.86 Å apart from the amide nitrogen of the bound substrate, which may allow the mutant to have the marginal activity. The crystallization conditions might have reduced the activity significantly further. The amide oxygen of the substrate is involved in the hydrogen bonds with the oxyanion hole, and its amino group is involved in a hydrogen bond with the carbonyl oxygen of Ala 111 (Fig. 4a). In comparison, the carboxyl oxygen at one end of the product, corresponding to the amide nitrogen of the substrate, is rotated away and beyond hydrogen-bonding distance from the carbonyl oxygen of Ala 111 (4.67 Å). The conformations of the active site residues are the same in the two structures. Moreover, the alanine substitution does not alter the conformation of any of the active site residues relative to these in the structure of the wild-type enzyme in complex with the product (PDB code: 1OCL). Therefore, the electrostatic interactions in the binding of the substrate to the S131A mutant and to the wild-type enzyme would be the same. In this binding mode, the substrate amide carbon is 2.90 Å apart from Ser 155 O␥, and the substrate amide nitrogen would be within 3.45 Å of Ser 131 O␥, which can be estimated on the basis of the superposition of the structures of the wild-type and the mutant enzyme in complex with the substrate (Fig. 4b). The amide nitrogen must be pulled closer to Ser 131 O␥ when Ser 155 attacks the substrate and forms the covalent bond. These structural data clearly suggest that the hydroxyl group of Ser 131 is the catalytic acid providing a proton to the leaving amino group of the substrate. It is concluded that the drastically lower activity of the S131A mutant is solely due to the loss of the catalytic roles of the Ser 131 -OH group: the polarization of Ser 155 and the protonation of the leaving group. These are the same catalytic roles performed by the histidine of the Ser-His-Asp triad, although the initial protonation states of the two residues may be different, as we proposed.
We modeled the threonine residue on the basis of the conformation of Ser 131 in the wild-type structure to see if threonine could replace Ser 131 . The methyl group of the modeled threonine residue exhibits multiple steric crashes with the backbone oxygen atoms of Ala 111 and Ser 112 and the side chain oxygen atom of Thr 108 , all situated at distances less then 2.8 Å of the methyl carbon (data not shown). These atoms line the wall of the bottom part of the active site, indicating that the substitution of Ser 131 with threonine would disrupt at least the structure of the active site. These observations provide an explanation for the absolute conservation of the serine residue at this position.
Role of the Conserved Glycine Preceding Ser 131 -Gly 130 is strictly conserved in all but two AS family proteins, which contain the substitution of the glycine residue with alanine (access code: gi13507185) or valine (access code: gi15610511), respectively. Since the peptide bond between Gly 130 and Ser 131 is in the cis configuration, we investigated whether substitution of Gly 130 with alanine may result in the trans peptide bond between the two residues with a concurrent defect in the catalytic activity. The structure of the G130A mutant was determined at 2.2 Å. From the beginning of the refinement of the structure, the electron density for Ala 130 was barely visible and that for Gly 129 and Ser 131 was fairly weak, indicating that the substitution results in the unstable conformation of this segment, in contrast with the well-defined conformation of this segment in the wild-type enzyme. Although weak or fragmented, the electron density could be explained by a conformation of the peptide bond between Ala 130 and Ser 131 in the cis configuration (Fig. 5). This is true for the both MAE2 molecules in the asymmetric unit of the crystal. Alanine modeled at the position of residue 130 in the structure of the wild-type enzyme shows that its C␤ atom is too close to the C␣ atom of cisSer 131 (2.9 Å, data not shown). This indicates that Ala 130 has to adopt backbone torsion angles different from those of Gly 130 to avoid steric crash. Apparently, the Ala 130 -containing loop cannot form a rigid structure by adopting an alternative conformation. The "shaking" loop conformation in the mutant enzyme results in 6-fold decrease in k cat and 2-fold increase in the Michaelis constant K m . Most likely, the ability of the mutant to adopt the cis configuration avoids a chaotic failure in the enzyme activity. The data explain why Gly 130 is virtually invariable and also why the two exceptions to the conservation of the glycine residue are allowed. When the trans peptide bond was modeled between Gly 130 and Ser 131 , the -NH group of Ser 131 cannot form a hydrogen bond with Ser 155 O␥ and the -OH group of Ser 131 cannot make the precise and simultaneous contacts with Ser 155 and Lys 62 that are observed for the cis configuration. All these observations suggest that the cis configuration of Ser 131 is essential for the catalytic function of the triad.
Lys 62 Plays Catalytic and Structural Roles-The direct interaction of the amino group of Lys 62 and the hydroxyl group of Ser 131 suggests that Lys 62 plays the catalytic role of stabilizing the Ser 131 alkoxide in the course of the catalysis. The structure of the K62A mutant reveals that the lysine residue of the triad plays a critical structural role as well as the proposed catalytic role. In the structure of the wild-type enzyme, the amino group of Lys 62 is in direct contact with Ser 132 and Thr 150 (Fig. 1a), which are conserved as either threonine or serine. Thr 150 O␥ interacts indirectly with Thr 152 O␥ via a bound water molecule (Wat 215 ), forming a hydrogen-bonded network of Lys 62 -Thr 150 -Wat 215 -Thr 152 . The substitution of the serine residue with alanine in fatty acid amide hydrolase corresponding to Ser 132 decreased the catalytic activity by ϳ100-fold (31), while the substitution of the threonine residue with alanine in the peptide amidase corresponding to Thr 150 decreased the catalytic activity by ϳ3-fold (14). The absence of these two interactions in the K62A mutant prompts a significant rearrangement of the segment containing the oxyanion hole, which forms several new interactions with surrounding residues and water molecules, as deduced from the unambiguous electron density of the active site (Fig. 6a). Notably, Ser 155 is completely out of place from its correct position and interacts with the -NH group of Ile 157 . Furthermore, the loop comprising the oxyanion hole is also distorted compared with that in the wild-type enzyme. It is hard to imagine that the binding of the substrate could trigger a restoration of the catalytically competent active site structure. The complete spoiling of the active site clearly explains why the K62A mutant exhibits no catalytic activity. In order to probe the importance of the hydrogen bond between Lys 62 and Thr 150 in maintaining the structure of the segment containing the oxyanion hole, we determined the structure of the T150A mutant enzyme. The structure reveals that two water molecules (Wat 215 and Wat 386 ) relocate to fill in the space vacated by the mutation and maintain the indirect interaction between Lys 62 and Thr 152 (data not shown). The loop structure of the oxyanion-containing segment of this mutant is nearly the same FIG. 4. Binding of substrate and product to the S131A mutant. a, 2F o ϪF c electron density maps for the substrate malonamate and product malonate bound to the S131A mutant. The maps were calculated to 2.3 and 1.8 Å for the malonamate and the malonate complex, respectively, and contoured at 1.5 . The hydrogen bonds (Ͻ2.9 Å) are shown in dotted lines. The amide group of the bound malonamate is tilted toward the Ser 131 -OH in comparison with the corresponding carboxyl group of the bound malonate. The conformations of the active site residues are the same in the substrate-bound and the product-bound structures. b, superposition of the structures of the wild-type and the S131A mutant enzymes. The backbone atoms of the structures were superimposed with root mean square deviation of 0.13 Å. The wild-type and the mutant enzymes are shown in coral and blue, respectively. The distances between the substrate bound to the mutant and the catalytic residues in the structure of the wild-type enzyme are shown in Å. as that of the wild-type enzyme. Consistently, the mutant enzyme exhibits about one-fourth of the catalytic activity of the wild-type enzyme (Fig. 7), which is similar to the consequence of the corresponding mutation in the peptide amidase (14). We speculate that the hydrogen-bonded network that links Lys 62 and the oxyanion-containing segment via Thr 150 is important for maintaining the active site structure, but the importance is masked by the salvaging water-mediated hydrogen bonds in this mutant. In the K62A mutant, the link is lost by the absence of the long side chain of Lys 62 , which allows the direct interaction between Ser 132 and Thr 150 and the penetration of two water molecules into the vacated site (Fig. 6a). The loss of the role of Lys 62 in propping the active site structure consequently prompts the derangement of the oxyanion hole-containing segment by forming new hydrogen-bonding interactions (Fig. 6).
In contrast with the consequence of the K62A mutation, the substitution of the aspartic acid with alanine of the Ser-His-Asp triad does not result in a complete loss of the enzyme activity of subtilisin (32). The mutation affects the general acid/base catalytic function, but it does not affect the integrity of the active site structure, as the mutant exhibited only a small increase in K m (ϳ2-fold). Cytomegalovirus protease contains a Ser-His-His triad, and the stereochemical positions of the catalytic components of the enzyme are essentially the same as those of trypsin-like proteases (33). These observations indicate that the aspartic acid in the Ser-His-Asp triad plays the catalytic role only, while the lysine residue in the Ser-cisSer-Lys triad plays the structural role as well as the catalytic role. Probably, the dual function, which the third residue of the novel triad serves, has suppressed selection of other amino acids by the AS family enzymes during evolution.

CONCLUSIONS
The study presented here delineates the role of each residue of the Ser-cisSer-Lys triad in detail and addresses the question of why the chemical make-up of the triad is absolutely conserved. We showed that the Cys-cisSer-Lys triad is not functional because the cysteine residue is modified to cysteine sulfinic acid. The active site environment of MAE2 appears to readily promote the oxidation of Cys 155 to cysteine sulfinic acid. The structures of the S131A mutant in complex with the product and with the substrate demonstrate that Ser 131 is the catalytic acid protonating the leaving amino group of the substrate. The consequent ion-pair formation between the serine alkoxide and the amino group of the lysine would facilitate the proton transfer reaction. The cis configuration of Ser 131 appears essential for the formation of the two direct hydrogen bonds with Ser 155 and another with Lys 62 . A simple modeling experiment suggests that any residue bulkier than serine at this position would disrupt at least the structure of the active site. The third lysine residue was shown to play a critical structural role of propping the active site structure as well as the proposed catalytic role of promoting the protonation of the leaving group by Ser 131 of the triad. The dual role appears to be supported only by lysine at this position that has to make precise contacts with the three different conserved residues. Currently, the structures of three AS family enzymes are available. The AS sequences of the three enzymes comprise the catalytic core scaffolds that are very similar to each other. The orientations of the catalytic components, the triad, and the oxyanion hole-containing segment of the three enzymes are superimposable with root mean square deviations less than 0.32 Å. We, therefore, suggest that the conclusions drawn from this mutational-structural study would be generally true for the whole AS family enzymes. The characterization of the Ser-cisSer-Lys triad described above indicates that the triad is not a variation of the Ser-His-Asp triad or any other known catalytic devices, such as catalytic dyads, and that the AS family enzymes are evolutionary distinct from all the other known hydrolases.
Acknowledgment-We thank the Pohang Accelerator Laboratory for the use of their x-ray facility. The k cat values of the wild-type MAE2 and mutant enzymes are shown in relative scale. The measured k cat value of the wild-type enzyme is 1844 Ϯ 178 s Ϫ1 and that of the S131A mutant is 1.07 Ϯ 0.04 s Ϫ1 . The activities of the S155C and K62A mutants were undetectable.