Leukotriene A4 hydrolase/aminopeptidase. Glutamate 271 is a catalytic residue with specific roles in two distinct enzyme mechanisms.

Leukotriene A(4) hydrolase/aminopeptidase is a bifunctional zinc metalloenzyme that converts the fatty acid epoxide leukotriene A(4) into leukotriene B(4), a potent chemoattractant and immune-modulating lipid mediator. Recently, the structure of leukotriene A(4) hydrolase revealed that Glu-271, which belongs to a conserved GXMEN motif in the M1 family of zinc peptidases, and Gln-136 are located at the active site. Here we report that mutagenetic replacements of Glu-271, but not Gln-136, abrogate both catalytic activities of leukotriene A(4) hydrolase. Furthermore, the 2.1 A crystal structure of [E271Q]leukotriene A(4) hydrolase revealed minimal conformational changes that could not explain the loss of enzyme function. We propose that the carboxylate of Glu-271 participates in an acid-induced opening of the epoxide moiety of leukotriene A(4) and formation of a carbocation intermediate. Moreover, Glu-271 appears to act as an N-terminal recognition site and may potentially stabilize the transition-state during turnover of peptides, a property that most likely pertains to all members of the M1 family of zinc aminopeptidases. Hence, Glu-271 is a unique example of an amino acid, which has dual and separate functions in two different catalytic reactions, involving lipid and peptide substrates, respectively.

The leukotrienes (LTs) 1 are potent chemical mediators in a variety of allergic and inflammatory reactions (1,2). In the biosynthesis of LTs, 5-lipoxygenase converts arachidonic acid into the unstable epoxide LTA 4 . This intermediate may in turn be conjugated with GSH to form the spasmogenic LTC 4 or hydrolyzed into the proinflammatory lipid mediator LTB 4 , in a reaction catalyzed by LTA 4 hydrolase (LTA4H). The enzyme product LTB 4 is a classical chemoattractant and triggers adherence and aggregation of leukocytes to the endothelium at nanomolar concentrations (3). In addition, LTB 4 modulates immune responses (4), participates in the host-defense against infections (5,6), and is a key mediator of PAF-induced lethal shock (7,8). These effects are signaled via a specific, highaffinity, G protein-coupled receptor for LTB 4 (BLT 1 ) (9). In addition, a second receptor for LTB 4 (BLT 2 ) was recently discovered, the functional role of which is presently not known (10). LTA 4 hydrolase (EC 3.3.2.6) is a bifunctional zinc metalloenzyme, which integrates a sophisticated epoxide hydrolase activity, specific for the fatty acid derivative LTA 4 with an aniondependent aminopeptidase activity in a common active center; for a review see Ref. 11. The enzyme belongs to the M1 family of zinc metallopeptidases (12), which include enzymes such as aminopeptidase A (EC 3.4.11.7, APA), aminopeptidase B (EC 3.4.11.6, APB) and aminopeptidase N (EC 3.4.11.2, APN) (see Fig. 1). A common structural feature in this group of enzymes is the zinc binding motif HEXXH(X 18 )E (13), and in LTA4H the zinc is coordinated by His-295, His-299, and Glu-318 (14). Both enzyme activities of LTA4H require the catalytic zinc.
The aminopeptidase activity accepts a variety of substrates and certain arginyl di-and tripeptides as well as p-nitroanilide derivatives of Ala and Arg are hydrolyzed with high efficiencies (15). Although it has never been proven, it is generally assumed that the aminopeptidase activity is involved in the processing of peptides related to inflammation and host-defense. Furthermore, this enzyme activity can be selectively abolished by mutation of either of the conserved Glu-296 or Tyr-383, presumably acting as a general base and proton donor, respectively (16,17). The members of the M1 family of metallopeptidases also share a conserved motif, GXMEN (Fig. 1), that is located N-terminally to the zinc site, and for APA and APN recent data have suggested that it participates in the exopeptidase function of these enzymes (18,19).
The epoxide hydrolase reaction, i.e. the conversion of LTA 4 into LTB 4 , is unique in that the stereospecific introduction of a hydroxyl group occurs at a site distant from the epoxide moiety, and it has been suggested that it proceeds via a delocalized carbocation intermediate. Unlike the aminopeptidase activity, the epoxide hydrolase activity of LTA4H is restrained by suicide inactivation, which involves binding of LTA 4 to Tyr-378 (20,21). However, besides the zinc binding ligands, no single amino acid residue critical for the epoxide hydrolase activity has yet been identified. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Recently, we determined the x-ray crystal structure of LTA4H at 1.95 Å resolution in complex with bestatin (22). In the structure, Glu-271, which is located within the GXMEN motif, and Gln-136 are positioned at the active site near the catalytic residue Glu-296, as well as the catalytic zinc and its three amino acid ligands (Fig. 2). Furthermore, Glu-271 and Gln-136 make interactions with bestatin suggesting that they might be involved in the aminopeptidase activity of LTA4H.
In the present report, we show that Glu-271, but not Gln-136, is a catalytic residue that is shared between the epoxide hydrolase and aminopeptidase activities of LTA4H. Yet, Glu-271 carries out a separate chemistry in each of the two reaction mechanisms. We propose mechanistic models in which Glu-271 both participates in the initial activation of the epoxide moiety of the substrate LTA 4 and contributes to the binding of the N-terminal amine and the exopeptidase nature of the peptidecleaving activity. Moreover, it is likely that the conclusions regarding the aminopeptidase activity are valid for all members of the M1 family of metallopeptidases and that the GXMEN sequence qualifies as a consensus motif for an Nterminal recognition site.

EXPERIMENTAL PROCEDURES
Mutagenesis of Human LTA4H cDNA-Site-directed mutagenesis was carried out by PCR on the recombinant plasmid pT3MB4 for expression of His 6 -tagged LTA4H in Escherichia coli. Briefly, it involves two consecutive steps of PCR comprising four different primers, according to the megaprimer method (23). The first reaction includes a modified primer (A) carrying the mutation and a second primer (B) comprising a restriction site within the LTA4H cDNA. Together, primers A and B amplify a DNA fragment denominated megaprimer. In the next PCR, the megaprimer is used together with a fourth primer (C) containing another restriction site to amplify a final fragment carrying the mutation and two different restriction sites at its ends. Polymerase chain reactions were carried out in a total volume of 50 l, using 1ϫ Pfu polymerase buffer, 125-150 ng each of primers A and B, 1 unit Pfu polymerase, 10 nmol each of dNTPs and 100 ng of template. In the second reaction, 15-20 l of the first PCR mix was used to supply with the megaprimer, which was used together with a primer C (125-150 ng). The amplification program included an initial round of denaturation at 94°C (60 s), annealing at 55-66°C (60 s), and elongation at 72°C (90 s) followed by 30 where F obs and F calc are the observed and calculated factor amplitudes, respectively. c R free is the R-factor calculated for the test set of reflections, which are omitted during the refinement process. Phast system using 10 -15% gradient gels, subsequently stained with Coomassie Brilliant Blue. Protein concentrations were determined by the Bradford method using a MCC/340 96-well multiscan spectrophotometer and bovine serum albumin as standard (24).
Aminopeptidase Activity Assays-The aminopeptidase activity was determined in a spectrophotometric assay at 405 nm, using a MCC/340 multiscan spectrophotometer, essentially as described (25). Briefly, the enzyme (1-20 g) was incubated at room temperature in 96-well microtiter plates with alanine-p-nitroanilide (Ala-p-NA) as substrate in 50 mM Tris-HCl, pH 8.0, containing 100 mM KCl. Additional aminopeptidase assays were performed similarly with Leu-p-NA and Lys-p-NA as substrates. The absorbance at 405 nm was measured at 10-min intervals. Different concentrations of substrate (0.125, 0.25, 0.5, 1, 2, 4, 8 mM) were used for kinetic determinations. Spontaneous hydrolysis of the substrate was corrected for by subtracting the absorbance of blank incubations without enzyme.
Epoxide Hydrolase Activity Assays and Reverse-phase HPLC-The epoxide hydrolase activity was determined from incubations of enzyme (1-20 g) in 100 l of 10 mM Tris-HCl, pH 8.0, with LTA 4 (2.5-125 M) for 30 s on ice. Reactions were quenched with 200 l of MeOH, followed by the addition of 0.4 nmol of prostaglandin B 1 (PGB 1 ) or PGB 2 as internal standard. Samples were acidified with 5 l of acetic acid (10%), and metabolites were extracted on solid phase Chromabond C18 columns. Enzymatic metabolites of LTA 4 were identified and quantified by reverse-phase HPLC. Metabolites of LTA 4 were separated by isocratic reverse-phase HPLC on a Waters Nova-Pak C18 column eluted with a mixture of methanol/acetonitrile/water/acetic acid (30:30:40:0.01 by volume) at a flow rate of 1.2 ml/min. The UV detector was set at 270 nm, and metabolites were quantified using Chromatography Station for Windows version 1.7 computer software. Calculations were based on peak area measurements and the known extinction coefficients for the internal standards PGB 1 and PGB 2 (30 000 M Ϫ1 ϫ cm Ϫ1 ) as well as LTB 4 (50 000 M Ϫ1 ϫ cm Ϫ1 ).
Data Collection and Structure Determination-For data collection, crystals were soaked in 15% PEG8000, 50 mM imidazole buffer, pH 6.8, 50 mM sodium-acetate, 2.5 mM YbCl 3 , and 25% glycerol and analyzed at the Beamline I711 of the synchrotron station Max-Lab in Lund, Sweden. A complete set of data was collected from a single crystal. The crystals belonged to space-group P2 1 2 1 2 1 with cell dimensions a ϭ 77.962, b ϭ 86.873, c ϭ 98.787, ␣ ϭ ␤ ϭ ␥ ϭ 90°at 100 K. Statistics on data collection and quality are given in Table I. Processing, scaling, and merging of data were carried out using the program package Denzo/ Scalepack (26). The structure of [E271Q]LTA4H was solved with molecular replacement using the program AMoRe (27), and the structure of wild type LTA4H as a search model (22). All subsequent refinement was done using the CNS package (28) and manual model building as well as interpretation of electron density maps with the programs QUANTA (Molecular Simulations, Inc) and XtalView (29). During the refinement 582 water molecules, one Zn 2ϩ ion, one acetate molecule, and one Yb 3ϩ ion were identified. After a final round of refinement, including all data between 25-2.1 Å and all 610 amino acids, the R-factor was 18.2% and the R free factor was 23.3%. Most of the model of [E271Q]LTA4H is in good density except for the His 6 tag and the first four N-terminal residues. The model exhibits a highly restrained but acceptable stereochemistry with 99.6% of the residues lying in the most favored or additionally allowed regions of the Ramachandran plot. Root mean square deviations for bond lengths and angles are 0.01 Å and 1.6°, respectively.
Molecular Modeling-To model the binding of Ala-p-NA, the program AutoDock 3.0 (30) was used. The substrate, assumed to be protonated with a positively charged amino group, was initially modeled manually into the active site of LTA4H using the program XtalView. The binding conformation was then optimized using the local search algorithm of AutoDock 3.0. Prior to the docking, addition of polar hydrogens and assignment of atomic charges to the model of LTA4H was done using the program QUANTA, which was also used to construct the atomic coordinates of Ala-p-NA. The program MOPAC (31) was used to assign atomic partial charges to the ligand and to further optimize its coordinates prior to the docking calculation. In all other aspects of the docking process, programs included in the AutoDock 3.0 program package were used.

Mutagenetic Replacements, Expression, and Purification of
Recombinant Proteins-To detail their function we exchanged Glu-271 for a Gln, Asp, or Ala residue, whereas Gln-136 was replaced with an Asn, Leu, and Ala by site-directed mutagenesis. In addition, the remaining residues in the GXMEN signature, viz. Gly-268, Gly-269, Met-270, and Asn-272 were mutated into Ala, Ala, Leu, and Ala, respectively. The resulting 10 mutants were all expressed as His 6 -tagged fusion proteins in E. coli to allow rapid and convenient purification on nickel-NTA resins. The level of expression was similar for wild type enzyme  ) and K m ranging between 18 -218% and 66 -221%, respectively, relative to the wild type enzyme. The corresponding specificity constants were between 8 and 301% that of the wild type enzyme.
Effects of Mutation of Gln-136 -Exchange of Gln-136 for a Leu, Asn, or Ala residue did not significantly affect the epoxide hydrolase activity of LTA4H (Table II). Values of V max (k cat ) and K m for [Q136L], [Q136N], and [Q136A]LTA4H reached 20 -57% and 18 -179% of the wild type control, respectively. On the other hand, the mutations led to highly variable effects on both the turnover, and the Michaelis constant for the peptidase substrate Ala-p-NA (Table II). Thus, [Q136L]LTA4H did not obey saturation kinetics and displayed a linear increase in reaction velocity in the substrate interval 0.125-8.0 mM Ala-p-NA. In contrast, [Q136N]LTA4H lacked almost completely the ability to hydrolyze Ala-p-NA, whereas the peptidase activity of [Q136A]LTA4H did not differ from that of wild type enzyme, with values of V max (k cat ) and K m mounting to 103 and 108%, respectively. This large variation in the effects of mutation of Gln-136 makes it difficult to draw any conclusions regarding its functional role. However, since the mutant in which the side chain was extensively truncated retained all of its peptidecleaving ability, the most likely interpretation is that Gln-136 does not play any significant role in this catalytic reaction.
Structure of [E271Q]LTA4H-The mutant and wild type enzymes crystallize non-isomorphously and also bind a different number of Yb 3ϩ ions. Thus, the wild type enzyme crystallizes in space-group P2 1 2 1 2 with three Yb 3ϩ ligands per asymmetric unit, whereas the mutant enzyme crystallizes in space-group P2 1 2 1 2 1 with only one Yb 3ϩ ligand per asymmetric unit.
In the wild type structure, the Zn 2ϩ ion is coordinated to His-295, His-299, one carboxylic oxygen of Glu-318, and the hydroxyl and carbonyl oxygens of bestatin in a pentavalent coordination sphere, forming a square-based pyramid (Fig. 2). In the structure of [E271Q]LTA4H, the Zn 2ϩ is bound to its three canonical amino acid ligands and one carboxylic oxygen of an acetate molecule. These four ligands form an almost ideal tetrahedral coordination geometry. The acetate molecule is also hydrogen-bonded to Tyr-383 and Gln-271.
Exchange of Glu-271 for a Gln does not cause any significant conformational change of the active site (Fig. 3). Superpositioning of the mutant and wild type models gives an overall r.m.s. deviation for 610 ␣-carbons of 0.45 Å. Inspection of the superimposed models shows that the differences are evenly spread over the molecules. The only exceptions are two surface loops, which exhibit higher r.m.s. deviations; these differences in ␣-carbon positioning are probably due to crystal packing differences. Of note, residues lining the active site, including the deeper part of the hydrophobic pocket, show a positional shift in relation to the wild type structure, which is comparable with the rest of the molecule. Thus, the major outline of the active site is well conserved in the mutant structure. However, when comparing the structures of the wild type and mutated enzyme in detail, a small rearrangement of spatially neighboring residues can be observed, which may be explained by the mutation and/or the absence of inhibitor in the crystal of [E271Q]LTA4H. Directly affected residues, besides Glu-271, are Glu-296 and Gln-136. The changes are subtle and mainly affect the hydrogen-bonding network.
The orientation of the amide side-chain of Gln-271 cannot be judged from electron density, and it was instead determined based on changes in hydrogen-bonding network. Glu-271 in the wild type structure is hydrogen-bonded to its own backbone amide and to the free amine of the bound inhibitor bestatin. In [E271Q]LTA4H, the orientation of the amide side group does not allow direct interaction with its backbone amide. Instead, new direct hydrogen bonds are observed to the acetate bound in the active site and to a water molecule now occupying the position held by the free amine of bestatin in the wild type structure. The absence of inhibitor and the hydrogen bonding to the amide group of Gln-271 pull Glu-296 toward the backbone amide of Gln-271.
Furthermore, in the mutant structure, Gln-136 is flipped about 90 degrees, which creates a hydrogen bond to Glu-271 and to a new water molecule replacing the free amine of bestatin. Another water molecule, at a position corresponding to the phenyl ring of bestatin, also makes a hydrogen bond to Gln-136. Thus, both the mutation and the absence of inhibitor could explain the positional change of Gln-136. Similarly, the torsional flip observed for Met-270 and the positional shift of Gln-134 are most likely effects caused by the absence of inhibitor.
Modeling of Ala-p-NA into the Active Site of LTA4H-Alap-NA is a good aminopeptidase substrate for LTA4H. The binding conformation with the lowest energy of a docked Ala-p-NA at the active site of LTA4H agreed very well with biochemical data and is presented in Fig. 4. Thus, the free amine of Alap-NA donates one hydrogen bond to each of the carboxyl groups of Glu-271 and Glu-318 and one to the amine carbonyl oxygen of Gln-136. Furthermore, the amide carbonyl oxygen of Alap-NA is bound to the Zn 2ϩ ion, and the carbonyl carbon is at a suitable position for attack from a water molecule activated by Glu-296. In addition, Tyr-383, which is believed to act as a proton donor in the final step of the peptidase reaction, is appropriately hydrogen-bonded to the amide nitrogen of the substrate. This alignment of Ala-p-NA indicates that it is a chemically good mimetic of tripeptides, the preferred aminopeptidase substrates of LTA4H (15). Moreover, this conclusion agrees well with the fact that the amino-and nitro-groups of Ala-p-NA are separated by the same number of covalent bonds as the N and C termini in tripeptide substrates. DISCUSSION Recently, we solved the x-ray crystal structure of LTA4H in complex with bestatin at 1.95 Å resolution (22). At the zinc site, in juxtaposition to the previously identified catalytic residues, Glu-296 and Tyr-383, a Glu and Gln residue were located in seemingly strategic positions, and their interactions with the inhibitor bestatin indicated that they could well be involved in the aminopeptidase reaction (Fig. 2). The carboxylic acid, Glu-271, is conserved among members of the M1 family of metallopeptidases in a GXMEN motif (Fig. 1), whereas the amine, Gln-136, is not well conserved within this enzyme family. To detail the functional role of Glu-271 and Gln-136 for the two catalytic activities of LTA4H, they were subjected to mutational analysis.
Glu-271 Is a Common Catalytic Residue in Both the Epoxide Hydrolase and Peptidase Reactions-Exchange of Glu-271 for a Gln, Ala, or Leu residue completely abolished the ability of LTA4H to convert LTA 4 into LTB 4 (Table II) as well as its amide cleaving activity against Ala-p-NA (Table II). In contrast, mutation of any of the remaining residues within the GXMEN motif preserved a significant epoxide hydrolase and peptidase activity. On the other hand, mutagenetic replacements of Gln-136 did not allow us to ascribe any specific functional role for this residue (Table II).
The finding that mutation of Glu-271 abolishes both catalytic activities of LTA4H was unexpected and warranted further studies to rule out that the recombinant proteins had lost the prosthetic zinc or had undergone major conformational changes at the active site. To this end, we determined the x-ray crystal structure of the mutant with the most conservative replacement, i.e. [E271Q]LTA4H. This structure revealed the presence of the catalytic metal and an almost completely conserved overall architecture, particularly at the active site, which demonstrates that the loss of enzyme function was sitespecific and could be attributed to the removal of the carboxylate from the side chain of residue 271. Hence, we conclude that Glu-271 is involved in key steps of the epoxide hydrolase and peptidase reactions. In fact, Glu-271 is the first amino acid, besides the three zinc-binding ligands, that has been shown to be required for the epoxide hydrolase mechanism. Moreover, to our knowledge, this is the first example of a single amino acid residue that is critical and common for two distinct catalytic reactions, which turn over lipid and peptide substrates, respectively.
Role of Glu-271 in the Epoxide Hydrolase Mechanism, Assistance in an Acid-induced Activation, and Opening of the Epoxide-Based on the structure of LTA4H as well as data obtained by studies of structure activity relationships of tightbinding hydroxamic acid inhibitors, we recently proposed a mechanistic model for enzymatic hydrolysis of LTA 4 into LTB 4 (22,32). It postulates that LTA 4 is bound to the active site with the oxirane ring near the zinc ion and the C7-C20 olefinic tail buried in a narrow, L-shaped, hydrophobic pocket in the protein. In this model, which apparently is incomplete, the zinc acts as a weak Lewis acid to activate and open the epoxide ring. In light of the present mutational data and the fact that Glu-271 is located in the immediate vicinity of the zinc, it seems likely that Glu-271 is required in the initial epoxide activation. Thus, we propose that Glu-271 polarizes a water molecule bound to the zinc and that the resulting proton will catalyze an acid-induced opening of the oxirane ring (Fig. 5A). This S N 1 reaction will generate a carbocation intermediate whose positive charge will be delocalized over the conjugated triene system. In the last step of the reaction, a nucleophilic attack will occur at C12 to introduce the 12R-hydroxyl group of LTB 4 (Fig. 5A).
Even if several lines of evidence support this model, we cannot rule out an alternative mechanism, which includes the formation of an ester intermediate. In this scheme, the zinc alone activates and opens the epoxide, and the carboxylate of Glu-271 attacks LTA 4 at C6 to form an ester intermediate. In a concerted S N 2Ј reaction, this ester can then be attacked by an hydroxyl group or a carboxylate at C12, and the negative charge can move along the conjugated triene system, eventually leading to an alkyl-oxygen cleavage instead of a normal ester cleavage (Fig. 5B). This model is attractive since formation of an alkylenzyme ester intermediate is a key feature in the mechanisms of other epoxide hydrolases, in particular soluble and microsomal epoxide hydrolase (33)(34)(35)(36). In this context it is interesting to note that LTA 4 is an excellent substrate of soluble epoxide hydrolase, which converts the allylic 5S,6Repoxide into a 5S,6R dihydroxy acid (37). On the other hand, the crystal structure of soluble epoxide hydrolase was recently solved and revealed no structural similarity with LTA4H (35). In addition, alkyl-oxygen cleavages of esters are rare reactions although they have been described for certain allylic or benzylic esters (38).
Role of Glu-271 in the Aminopeptidase Reaction, an N-terminal Recognition Site for Peptide Substrates, and a Determinant for the Exopeptidase Specificity of LTA4H-The total loss of aminopeptidase activity in [E271Q]LTA4H points to a critical role of Glu-271 in this catalytic mechanism. Several lines of evidence indicate that this residue acts as an anionic binding site for peptide substrates. Thus, in the structure of LTA4H in complex with bestatin, Glu-271 makes hydrogen bonds to the free amine of the inhibitor, which chemically resembles a peptide substrate (22). Furthermore, Glu-271 is ideally positioned in the structure to bind the N terminus of Ala-p-NA or an arginyl tripeptide, the preferred peptidase substrate of LTA4H. In addition, this residue, or any equivalent functionality, is absent in the structure of thermolysin, a classical zinc endopeptidase that accommodates peptide substrates of any length (39). In agreement with the crystallographic data, modeling of Ala-p-NA, a chemically good mimetic of a tripeptide substrate into the active site of LTA4H, positions Glu-271 within hydrogen bonding distance to the terminal amino group of the substrate (Fig. 4). Together, these data imply that Glu-271 acts as an N-terminal recognition site for peptide substrates, and in this capacity Glu-271 will play a critical role for the exopeptidase nature of the peptide-cleaving activity of LTA4H. The drastic effects of mutations on k cat with abolishment of catalytic activity suggest that Glu-271 may also be involved in the stabilization of the transition-state (Table II). For APN and APA, it has been suggested that such a transition-state involves an interaction between the nitrogen of the N-terminal  4 followed by a concerted nucleophilic attack at C12 and migration of negative charge over the conjugated triene system. In the final step of the reaction sequence, the carboxylate is regenerated via an alkyl-oxygen bond cleavage. For further details, see "Discussion." amino group of the peptide substrate and the catalytic zinc (18,19). However, in our model for the binding of Ala-p-NA, the distance between these two atoms is too long to allow such an interaction (Fig. 4). Instead, the model indicates that the amino group of the substrate is positioned within hydrogen-bonding distance to Glu-318.
The model for binding of Ala-p-NA is in excellent agreement with previous biochemical and mutational data and implies that Glu-271, as well as Glu-296 and Tyr-383, can be fitted into a general base mechanism for the aminopeptidase activity (Fig.  6). Thus, in the energy-minimized binding conformation, the carbonyl oxygen of Ala-p-NA is indeed bound to the Zn 2ϩ , and Glu-296 is ideally positioned to polarize a water molecule for nucleophilic attack at the carbonyl carbon (Fig. 4). Furthermore, in this model Tyr-383 is located close enough to the peptide nitrogen to function as a proton donor.
Sequence alignments have previously been used to identify the conserved GXMEN motif within the M1 family of zinc peptidases without knowing its functional relevance (12). Further studies with site-directed mutagenesis on APN, APA, and oxytocinase, have suggested that the equivalent of Glu-271 is involved in the aminopeptidase reaction, presumably as an anionic binding site for the ␣-amino group of peptidic substrates (18,19,40). However, these studies have been limited by the fact that the mutated enzymes were not purified before kinetic characterization and by the lack of a crystal structure for any member of the enzyme family to be used as a template for mechanistic models. Nevertheless, when comparing these earlier data with the results of the present investigation, they are certainly in good agreement. Considering the fact that our conclusions regarding the functional role of Glu-271 are based on a combination of mutational analysis and x-ray crystallography, we believe that they can be generalized to include all members of the M1 family of metallopeptidases. Hence, we think it is now justified to denote the GXMEN sequence a consensus motif for an N-terminal recognition site in zinc aminopeptidases.