Leukotriene A₄ hydrolase: Selective abrogation of leukotriene B₄ formation by mutation of aspartic acid 375

Abstract

Leukotriene A₄ (LTA₄, 5S-trans-5,6-oxido-7,9-trans-11,14-cis-eicosatetraenoic acid) hydrolase (LTA4H)/aminopeptidase is a bifunctional zinc metalloenzyme that catalyzes the final and rate-limiting step in the biosynthesis of leukotriene B₄ (LTB₄, 5S,12R-dihydroxy-6,14-cis-8,10-trans-eicosatetraenoic acid), a classical chemoattractant and immune modulating lipid mediator. Two chemical features are key to the bioactivity of LTB₄, namely, the chirality of the 12R-hydroxyl group and the cis-trans-trans geometry of the conjugated triene structure. From the crystal structure of LTA4H, a hydrophilic patch composed of Gln-134, Tyr-267, and Asp-375 was identified in a narrow and otherwise hydrophobic pocket, believed to bind LTA₄. In addition, Asp-375 belongs to peptide K21, a previously characterized 21-residue active site-peptide to which LTA₄ binds during suicide inactivation. In the present report we used site-directed mutagenesis and x-ray crystallography to show that Asp-375, but none of the other candidate residues, is specifically required for the epoxide hydrolase activity of LTA4H. Thus, mutation of Asp-375 leads to a selective loss of the enzyme's ability to generate LTB₄ whereas the aminopeptidase activity is preserved. We propose that Asp-375, possibly assisted by Gln-134, acts as a critical determinant for the stereoselective introduction of the 12R-hydroxyl group and thus the biological activity of LTB₄.

The development and maintenance of inflammation are governed by a complex network of cellular and humoral factors to which belong the leukotrienes, a class of structurally related paracrine hormones derived from the oxidative metabolism of arachidonic acid (1, 2). In the biosynthesis of leukotrienes, 5-lipoxygenase converts arachidonic acid into the unstable epoxide leukotriene A₄ (LTA₄, 5S-trans-5,6-oxido-7,9-trans-11,14-cis-eicosatetraenoic). This intermediate may in turn be conjugated with glutathione by a specific leukotriene C₄, LTC₄ (5S-hydroxy-6R-S-glutathionyl-7,9-trans-11,14-cis-eicosatetraenoic acid) synthase to form the spasmogenic LTC₄ or hydrolyzed into leukotriene B₄ (LTB₄, 5S,12R-dihydroxy-6,14-cis-8,10-trans-eicosatetraenoic acid), in a reaction catalyzed by LTA₄ hydrolase (LTA4H). LTB₄ is a classical chemoattractant of human neutrophils and triggers adherence and aggregation of leukocytes to the endothelium at only nM concentrations (3). These effects are signaled by means of a specific, high-affinity, G protein-coupled receptor for LTB₄ (BLT₁) (4). In addition, an ORF encoding a second receptor for LTB₄ (BLT₂) was recently discovered in the promoter of the gene for BLT₁ (5). The functional role of BLT₂ is presently not known.

LTA4H (EC 3.3.2.6) is a bifunctional zinc metalloenzyme, exhibiting an aminopeptidase activity in addition to its epoxide hydrolase activity, i.e., the hydrolysis of the unstable allylic epoxide LTA₄ into LTB₄ (for a review see ref. 6). A physiological substrate has not yet been discovered for the aminopeptidase activity but certain arginyl dipeptides and tripeptides as well as p-nitroanilide derivatives of Ala and Arg are hydrolyzed with high efficiency (7). Several lines of evidence indicate that the aminopeptidase activity follows a general base mechanism with Glu-296 and Tyr-383 acting as the base and proton donor, respectively. Moreover, in a recent study, Glu-271 was identified as the recognition site for the N-terminal amino group of the peptidase substrate (8).

Unlike the aminopeptidase activity, the epoxide hydrolase activity is restrained by suicide inactivation that involves binding of LTA₄ to Tyr-378, which is located within a 21-residue active site-peptide denoted K21 (9, 10). The epoxide hydrolase reaction of LTA4H (i.e., the conversion of LTA₄ to LTB₄) is also unique in the sense that the stereoselective introduction of water to the carbon backbone of LTA₄ occurs several methylene units away from the epoxide moiety, presumably proceeding by means of a carbocation intermediate (11). The structural requirements for controlling such a reactive species in the active site and for obtaining sufficient stereochemical specificity are of fundamental biochemical as well as enzymological interest. However, only very few residues have been shown to be involved in the epoxide hydrolase activity of LTA4H. In fact, no single amino acid residue specifically required for the epoxide hydrolase activity has yet been identified.

We recently determined the x-ray crystal structure of LTA4H at 1.95-Å resolution in complex with bestatin ([(2S,3R)-3-amino-2-hydroxy-4-phenylbutanoyl]-l-leucine) (12). At the active center, an extended hydrophobic pocket was found that potentially could bind LTA₄. In this cavity, a hydrophilic patch comprising the residues Gln-134, the phenolic hydroxyl group of Tyr-267, and Asp-375 was identified, potentially contributing polar residues involved in catalysis in an otherwise lipophilic environment (Fig. 1). Interestingly, Asp-375 also belongs to peptide K21, in which it is located within a stretch of charged Asp/Glu residues, any one of which could act as a nucleophile in enzyme catalysis.

Figure 1.

Structure of wild-type LTA4H, spatial relationships at the catalytic center. At the active center, a narrow (6–7 Å wide) L-shaped, hydrophobic pocket extends about 15 Å into the protein. At the angle of the cavity, a hydrophilic patch composed of Gln-134, Tyr-267, and Asp-375 is located. In addition, the positions of the other acidic residues along peptide K21, Asp-368, Asp-371, Asp-373, and Glu-384 are indicated. Amino acids are depicted in stick-and-ball representation. The substrate LTA₄ (yellow) does not belong to the structure and has been modeled into its tentative binding mode. The catalytic zinc is depicted in gold, adjacent to the epoxide moiety of LTA₄. The figure was generated with swiss-pdbviewer (33) and pov-ray (http://www.povray.org).

In the present article we used site-directed mutagenesis, complemented with x-ray crystallography, to analyze the function of all residues within the hydrophobic pocket or peptide K21 that could potentially participate in the epoxide hydrolase and/or aminopeptidase reaction. We show that mutation of Asp-375 selectively eliminates the epoxide hydrolase activity, suggesting that this residue plays a key role in this enzyme mechanism. On the other hand, exchange of Gln-134 for a nonpolar residue selectively enhances the epoxide hydrolase activity, suggesting a regulatory function for this residue. Based on the present data and previous conclusions regarding the active site structure and catalytic mechanisms, we propose that Asp-375, possibly assisted by Gln-134, is responsible for the stereospecific introduction of the 12R-hydroxyl group in LTB₄. In this role, Asp-375 will also be an important functional element that governs the generation of bioactivity in the product LTB₄.

Experimental Procedures

Materials.

Oligonucleotides were synthesized by Cybergene, Stockholm. LTA₄ methyl ester was synthesized as described (13) or purchased from Biomol, Plymouth Meeting, PA. LTA₄ methyl ester was saponified in tetrahydrofurane with 1 M LiOH (6% vol/vol) for 48 h at 4°C. Alanine-p-nitroanilide, isopropyl-β-d-thiogalactopyranoside, PMSF, soybean trypsin inhibitor, and streptomycin sulfate were purchased from Sigma. Nickel-nitrilo triacetic acid resin was from Qiagen, Chatsworth, CA.

Mutagenesis of Human LTA4H cDNA.

Site-directed mutagenesis was carried out by PCR on the recombinant plasmid pT3MB4 designed for expression of (His)₆-tagged LTA4H in Escherichia coli (11). Briefly, it involves two consecutive steps of PCR comprising four different primers, according to the “megaprimer” method (14). PCRs were carried out in a total volume of 50 μl, using 1× Pfu polymerase buffer, 125–150 ng each of primers A and B, 1 unit Pfu polymerase, 10 nmol each dNTPs, and 100 ng template. In the second reaction, 15–20 μl of the first PCR mix was used to supply the reaction 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 cycles of denaturation (45 s), annealing (30 s), and elongation (60 s) on a Perkin–Elmer GeneAmp PCR System 2400.

For generation of [Q134A/L/N]LTA4H, and [H139Q]LTA4H, a SalI and BglII site were used whereas a BglII and a BfrI site were used for [Y267F]LTA4H, [D368N]LTA4H, [D371N]LTA4H, [D373N]LTA4H, [D375A/E/N]LTA4H, and [E384Q]LTA4H. DNA fragments were cleaved by SalI/BglII or BglII/BfrI and purified by agarose gel electrophoresis (1.5%) followed by extraction (QIAEX II Gel Extraction Kit). Mutated fragments were ligated into pT3MB4 (T4 DNA Ligase Protocol), opened with the respective pair of restriction enzymes. Competent E. coli cells (JM101) were transformed with mutated recombinant plasmid and grown in LB medium containing ampicillin (100 μg/ml). Stock cultures were kept at −70°C in a 1:1 mixture of culture medium and 40% (vol/vol) glycerol/0.75% (wt/vol) NaCl, respectively. Recombinant plasmids were purified by using Wizard Minipreps Plus, and all of the mutated inserts were sequenced by using a Dyenamic ET terminator cycle sequencing kit (Amersham Pharmacia) to confirm that no nucleotide alterations had occurred in addition to the desired mutation.

Protein Expression and Purification.

Mutated enzymes were expressed as N-terminal (His)₆-tag fusion proteins in E. coli JM101 cells grown at 37°C in M9 medium (50 mM Na₂HPO₄, 22 mM KH₂PO₄, 20 mM NH₄Cl, 8.5 mM NaCl), pH 7.4, containing 0.4% (wt/vol) glucose, 0.2% (wt/vol) casamino acids, 2 mM MgSO₄, and 0.1 mM CaCl₂. At OD₆₂₀ ≈0.2, isopropyl-β-d-thiogalactopyranoside was added to a final concentration of 500 μM. Cells were harvested at OD₆₂₀ ≈1.8, pelleted at 1,000 × g, and resuspended in 30 ml homogenization buffer (50 mM Tris⋅HCl, pH 8.0, containing soybean trypsin inhibitor) supplemented with 1 mM PMSF. Nucleic acids were removed by streptomycin sulfate precipitation. After centrifugation (10,000 × g for 15 min), the supernatant was filtered (0.22 μm) and applied to a nickel-nitrilo triacetic acid resin. The column was washed with 1 bed volume of 50 mM Tris⋅HCl (pH 8.0), 50 mM sodium phosphate buffer (pH 6.8), 0.5 M NaCl, and 50 mM Tris⋅HCl (pH 8.0) with each solution supplemented with 10 mM imidazole. The final wash consisted of 0.8 bed volumes of 50 mM Tris⋅HCl, pH 8.0, containing 100 mM imidazole. The (His)₆-tagged protein was eluted with 0.6 bed volumes of 50 mM Tris⋅HCl, pH 8.0, containing 100 mM imidazole. If required, a final step of anion-exchange chromatography on a Mono Q HR5/5 column was used. The resin was equilibrated with 10 mM Tris⋅HCl, pH 8.0 and eluted with a linear gradient of KCl (0–0.5 M) in 10 mM Tris⋅HCl, pH 8.0. The purity of the final preparation was assessed by SDS/PAGE on an Amersham Pharmacia Phast system by using 10–15% gradient gels, subsequently stained with Coomassie brilliant blue. Protein concentrations were determined by the Bradford method (BioRad Coomassie brilliant blue, G-250) with a MCC/340 96-well multiscan spectrophotometer and BSA as standard (15).

Aminopeptidase Activity Assays.

The aminopeptidase activity was determined in a spectrophotometric assay at 405 nm, using a MCC/340 multiscan spectrophotometer, essentially as described (16). Briefly, the enzyme (1–20 μg) was incubated at room temperature in 96-well microtiter plates with alanine-p-nitroanilide as substrate in 50 mM Tris⋅HCl, pH 8.0, containing 100 mM KCl. The absorbance at 405 nm was measured at 10-min intervals. Different concentrations of substrate (0.125, 0.25, 0.5, 1, 2, 4, and 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.

The epoxide hydrolase activity was determined from incubations of enzyme (1–20 μg) in 100 μl of 10 mM Tris⋅HCl, pH 8.0, with LTA₄ (2.5–125 μM) for 30 s on ice. Reactions were quenched with 200 μl of MeOH, followed by the addition of 0.4 nmol of prostaglandin (PG) B 1 or PGB₂ as internal standard. Samples were acidified with 5 μl of acetic acid (10%), and metabolites were extracted on solid-phase Chromabond C18 columns. Enzymatic metabolites of LTA₄ were identified and quantified by reverse-phase HPLC.

Reverse-Phase HPLC.

Metabolites of LTA₄ 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 vol) at a flow rate of 1.2 ml/min. The UV detector was set at 270 nm, and metabolites were quantified by using chromatography station for Windows version 1.7 computer software. Known amounts of PGB₁ or PGB₂ were used as internal standard. Calculations were based on peak area measurements by using an extinction coefficients of 30,000 M⁻¹ × cm⁻¹ for the internal standards PGB₁ and PGB₂ and 50,000 M⁻¹ × cm⁻¹ for LTB₄.

Crystallization.

Plate-like crystals of [D375N]LTA4H were obtained by liquid-liquid diffusion in capillaries, as described (12). Briefly, 5 μl of precipitation solution [28% (vol/vol) PEG8000, 0.1 mM Na acetate, 0.1 mM imidazole buffer, pH 6.8, and 5 mM YbCl₃] was injected into the bottom of a melting point capillary and an equal volume of LTA4H (5 mg/ml) in 10 mM Tris⋅HCl, pH 8, was layered on top. The crystals belonged to space-group P2₁2₁2₁ with cell dimensions a = 78.430, b = 87.060, c = 99.070, α = β = γ = 90° at 100 K.

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 YbCl₃, and 25% glycerol. Data were collected at the beamline I711 of the Max-Lab, Lund, Sweden. A complete set of data were collected from a single crystal. Statistics on data collection and quality are given in Table 1. Processing, scaling, and merging of data were carried out by using the program mosflm (34) and programs from CCP4 (17). As a starting structure for refinement, the [E271Q]LTA4H mutant was used (8). All refinement was done by using the cns package (18). Manual model building and interpretation of electron density maps were performed by using the program xtalview (19). During the refinement 440 water molecules, one Zn²⁺ ion, one imidazole molecule, one acetate molecule, and one Yb³⁺ ion were identified. The final R factor was 17.4% and the R_free factor was 22.04%. Most of the model of [D375N]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 97.7% 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 0.010 Å and 1.51°, respectively.

Table 1.

Data collection and refinement statistics

Data collection statistics
Diffraction limit, Å	2.2
Wavelength, Å	0.991
Completeness, %	97.2
Mean I/σ (I)	7.3
Multiplicity of observation	3.4
Rmerge*, %	8.3
Refinement statistics
R factor†, %	17.4
Rfree‡, %	22.04
rms deviation in bond distance, Å	0.010
rms deviation in bond angle, °	1.51

^*R_merge = Σ_hΣ_i|I_i(h) − I(h)|/Σ_hΣ_iI_i(h), where I_i(h) is the ith measurement of reflection h and I(h) is the weighted mean of all measurements of h. 

^† R factor = Σ_h|F_obs(h) − F_calc(h)|/Σ_hF_obs(h), where F_obs and F_calc are the observed and calculated factor amplitudes, respectively. 

^‡ R_free is the R factor calculated for the test set of reflections that are omitted during the refinement process. 

Results

Mutagenetic Replacements, Expression, and Purification of Recombinant Proteins.

To detail their function we exchanged Asp-375 for an Ala, Asn, or Glu residue, Gln-134 for an Asn, Leu, and Ala, and Tyr-267 for a Phe by site-directed mutagenesis (Table 2). This procedure generated the mutants [D375A/E/N]LTA4H, [Q134A/L/N]LTA4H, and [Y267F]LTA4H. As a control for mutants in position 134, His-139 was exchanged for Gln to produce [H139Q]LTA4H. In the wild-type structure, His-139 is also hydrogen-bonded to Asp-375. In addition, the remaining acidic residues within peptide K21, Asp-368, Asp-371, Asp-373, and Glu-384 were exchanged for an Asn, Asn, Asn, and Gln, respectively, generating mutants [D368N]LTA4H, [D371N]LTA4H, [D373N]LTA4H, and [E384Q]LTA4H. The resulting 12 mutants were all expressed as (His)₆-tagged fusion proteins in E. coli, to allow rapid purification on nickel-nitrilo triacetic acid resins. The level of expression was similar for wild-type enzyme and all mutants, with a final yield of about 1–5 mg purified protein per liter cell culture. It may be noted that expression of [Y267F]LTA4H resulted in generation of varying amounts of a protein fragment of about 30 kDa, in addition to the regular 69-kDa LTA4H. Therefore, this mutant was more extensively purified than others, carefully selecting the active 69-kDa subfractions. The phenomenon might indicate the introduction of a structural weakness upon mutation at residue 267.

Table 2.

Mutagenetic replacements in LTA4H, effects on enzyme activities, and kinetic constants

Mutant	Epoxide hydrolase activity			Aminopeptidase activity
	kcat	Km	kcat/Km	kcat	Km	kcat/Km
Q134A	>1,000*	>1,000*	—	18  ± 1	157  ± 17	12
Q134L	721  ± 52	364  ± 70	198	13  ± 0.4	87  ± 6	15
Q134N	89  ± 6	148  ± 27	60	206  ± 4	108  ± 5	190
H139Q	68  ± 6	167  ± 51	40	7  ± 0.6	359  ± 42	2
Y267F	165  ± 8	58  ± 17	285	61  ± 3	64  ± 7	96
D368N	83  ± 6	72  ± 16	116	99  ± 4	107  ± 10	93
D371N	57  ± 5	51  ± 14	112	66  ± 4	77  ± 11	86
D373N	45  ± 4	43  ± 13	104	119  ± 6	101  ± 12	118
D375A	4  ± 1	40  ± 25	9	421  ± 145	1,436  ± 574	29
D375E	ND	ND	—	2  ± 0.1	152  ± 22	1
D375N	ND	ND	—	>1,000*	>1,000*	—
E384Q†	∼25	<3	>80	11  ± 0.7	71  ± 12	16

Mutated recombinant enzymes were expressed in E. coli, purified by affinity chromatography, and assayed for epoxide hydrolase and aminopeptidase activity. Apparent kinetic constants are expressed in % of those of wild-type enzyme. Each data point was calculated as mean ± SD, triplicate determinations. ND, not detectable. For the epoxide hydrolase activity, the values of k_cat and K_m for the wild-type enzyme varied between 0.21 and 0.47 s⁻¹ and 6 and 28 μM, respectively. For the aminopeptidase activity, the corresponding values of k_cat and K_m for the wild-type enzyme were 1.16–1.81 s⁻¹ and 2.3–3.8 mM, respectively. 

^*The enzyme activity of this mutant exhibited unsaturable enzyme kinetics within the range observed. These values are estimations. 

^† Mutant E384Q exhibited substrate inhibition of the epoxide hydrolase activity at higher substrate concentrations (>21 μM). The value of V_max is estimated to be equal to the calculated plateau of V between 5 and 21 μM LTA₄. 

Effects of Mutation of Asp-375.

Exchange of Asp-375 for an Ala, Asn, or Glu resulted in a complete to near-complete loss of epoxide hydrolase activity (Table 2). In contrast, the aminopeptidase reaction was diminished only in [D375E]LTA4H. For the epoxide hydrolase activity, exchange of Asp-375 for an Ala resulted in an enzyme with a low, but detectable, catalytic activity with V_max (k_cat) and K_m of 4% and 40% of wild-type control, respectively. Exchange of Asp-375 for a Glu or Asn resulted in inactive enzymes for which kinetics could not be determined. For the aminopeptidase reaction, exchange of Asp-375 for a Glu resulted in an enzyme with V_max (k_cat) and K_m of 2% and 152% of wild-type enzyme, respectively. Exchange of Asp-375 for either an Ala or an Asn gave rise to active enzymes that, however, did not exhibit saturation kinetics within the substrate intervals used (0.125–8 mM). Mutation of His-139, Tyr-267, or any of the acidic residues, Asp-368, Asp-371, Asp-373, or Glu-384, along peptide K21, were all compatible with a significant, albeit variable, epoxide hydrolase activity (Table 2). For these mutants, values of V_max (k_cat) and K_m ranged between 25% and 165% and 3% and 167%, respectively. Likewise, [H139Q]LTA4H, [Y267F]LTA4H, [D368N]LTA4H, [D371N]LTA4H, [D373N]LTA4H, and [E384Q]LTA4H maintained a variable aminopeptidase activity with values of V_max (k_cat) and K_m between 7% and 119% and 64% and 359%, respectively.

Effects of Mutation of Gln-134.

Exchange of Gln-134 for an unpolar Leu residue resulted in significantly increased epoxide hydrolase activity with V_max (k_cat) and K_m of 721% and 364% of wild-type enzyme, respectively. Exchange of Gln-134 for an Ala also led to a hyperactive enzyme although it did not display saturation kinetics within the concentrations of substrate used (6.3–101 μM). On the other hand, a conservative replacement of Gln-134 for Asn resulted in a V_max (k_cat) and K_m of 89% and 148% of wild-type enzyme, respectively. With respect to the aminopeptidase activity, all of the mutants [Q134L/A/N]LTA4H displayed significant aminopeptidase activities with values of V_max (k_cat) and K_m ranging between 13% and 206% and 87% and 157% of wild-type enzyme, respectively (Table 2).

Structure of [D375N]LTA4H.

The structure of the [D375N]LTA4H mutant is very similar to that of the native LTA4H (12). Overall rms difference of all Cα atoms between the two structures is 0.45 Å. The only loop that has changed its conformation is loop 175–185, which is located near the surface and far away from the active site. This is probably caused by a difference in crystal contacts because the native protein crystallizes in a different space group. Like in the native structure, the aminopeptidase inhibitor bestatin is bound to the Zn²⁺ ion in the active site. The conformation and the interactions this inhibitor makes with the mutated protein are nearly identical to the complex with wild-type LTA4H.

Only very small changes in conformation are observed near the active site (Fig. 2). Near the mutated D375N residue, the hydrogen-bonding network has been slightly altered. This alteration is caused mainly by a rotation around the Cα-Cβ bond of the Asn-375 side chain, which causes the formation of a direct weak hydrogen bond between the Nδ2 of Asn-375 and the OH of Tyr-267. In the native protein there was no direct contact between these two residues, instead a bridging water molecule was observed. This water molecule is still present in the structure of the mutant but has a slightly altered position. The distance between the Oδ1 of Asn-375 and the putative nucleophilic water molecule has been somewhat increased in comparison to the native structure (from 2.46 Å in native to 3.31 Å in the mutant structure) and the angle is now unfavorable for the formation of a proper hydrogen bond.

Figure 2.

Structure of the mutant [D375N]LTA4H. (A) Superpositioning of wild-type LTA4H and [D375N]LTA4H. Stereo image of selected residues of the active site of wild-type LTA4H complexed with bestatin superimposed onto [D375N]LTA4H. The wild-type and [D375N]LTA4H structures are shown in pale blue and red, respectively. The Zn²⁺ ions, the bestatin, and water molecules of wild type (pale blue) and [D375N]LTA4H (red) are also shown. The image shows very subtle conformational changes and thus a high degree of structural preservation after mutation of Asp-375 to Asn-375. See text for further discussion. The figure was generated with swiss-pdbviewer (33) and pov-ray (http://www.povray.org). (B) Comparison of the hydrogen-bonding networks around Asp-375 and Asn-375. The side chain of Asn-375 is indicated in red. Likewise the interactions with His-139, Tyr-267, and two water molecules of Asp-375 and Asn-375 are depicted in dotted black and red lines, respectively. Distances are given in Å.

Discussion

LTB₄ is a potent chemoattractant and immune modulating lipid mediator. To elicit these biological effects, LTB₄ signals by means of a high-affinity, G protein-coupled, seven-transmembrane receptor present on the surface of a variety of immunocompetent cells, e.g., peripheral leukocytes, including neutrophils and eosinophils, as well as peritoneal macrophages (20, 21). Studies of structure activity relationships for LTB₄ and related compounds have revealed that two chemical properties are key to the bioactivity of this molecule, namely the cis-trans-trans geometry of the conjugated triene system and the R stereochemistry of the hydroxyl group at C12 (22–24). It is therefore of particular interest to learn the mechanisms by which LTA4H manages to incorporate these structural features into LTB₄ during enzymatic hydrolysis of LTA₄.

Recently, we determined the x-ray crystal structure of LTA4H in complex with bestatin at 1.95-Å resolution (12). Behind the inhibitor a narrow extended hydrophobic cavity was found that potentially could bind LTA₄. In addition, a hydrophilic patch, composed of Gln-134, the hydroxyl group of Tyr-267, and Asp-375, is present in that cavity, each of which could potentially act as a general base catalyst in the epoxide hydrolase reaction. Moreover, Asp-375 is located within an acidic stretch of five Asp/Glu residues (Asp-368, Asp-371, Asp-373, Asp-375, and Glu-384) in peptide K21, a previously identified active site peptide. In fact, peptide K21 carries Tyr-378, the binding site for LTA₄ during suicide inactivation/covalent modification of LTA4H (9, 10). To possibly identify a nucleophilic amino acid involved in the epoxide hydrolase reaction we mutated all of the candidate residues of the hydrophilic patch of the putative LTA₄ binding pocket as well as those of peptide K21.

Asp-375 Is a Catalytic Residue Specifically Required for the Epoxide Hydrolase Reaction.

Exchange of Asp-375 for an Asn, Ala, or Glu residue essentially eliminated the ability of LTA4H to convert LTA₄ into LTB₄ (Table 2). The only mutant with detectable epoxide hydrolase activity was [D375A]LTA4H, which may reflect that an unpolarized water molecule may slip into the extra space near the substrate, gained by the reduced side-chain length caused by the mutation. In contrast, a significant aminopeptidase activity was observed in [D375A]LTA4H and [D375N]LTA4H whereas the activity of [D375E]LTA4H was greatly reduced, although clearly detectable. Mutations of Tyr-267, Gln-134, and the remaining acidic residues in peptide K21 were all compatible with significant, albeit variable, epoxide hydrolase and aminopeptidase activities (Table 2). To explore possible structural alterations induced by the mutations of Asp-375, we determined the crystal structure of the mutant with the most conservative replacement, namely [D375N]LTA4H. This structure revealed the presence of the catalytic metal, as expected from the preserved aminopeptidase activity, and an almost completely conserved overall architecture, particularly at the active site (Fig. 2). Apparently, the loss of epoxide hydrolase activity is site-specific and can be attributed to the removal of the carboxylate from the side chain of Asp-375. We conclude that Asp-375 is involved in a key step of the epoxide hydrolase reaction. Hence, by a single point mutation at position 375 we can dissociate the two catalytic activities of LTA4H with preservation of the aminopeptidase activity.

LTA4H is a member of the M1 family of the MA clan of metallopeptidases, which includes enzymes such as aminopeptidase A, aminopeptidase B, aminopeptidase N, and angiotensin converting enzyme. Members of this family share a common zinc binding signature and a proteolytic activity, which seems to proceed according to a zinc-assisted general base mechanism (25, 26). In LTA4H, Glu-296 and Tyr-383 are catalytic in the aminopeptidase reaction in which they act as a general base catalyst and proton donor, respectively (27, 28). In addition, Glu-271 is involved in the reaction and appears to bind the N-terminal amino group of the peptidase substrate (8). This residue is also catalytic for the epoxide hydrolase reaction in which it seems to assist in the initial activation and opening of the epoxide ring. However, besides the catalytic zinc and Glu-271, very little is known about the functional elements that are required for the conversion of LTA₄ into LTB₄. In fact, Asp-375 is the first example of an amino acid that is specifically involved in the epoxide hydrolase reaction. Yet, Asp-375 is not unique for LTA4H and is present in related aminopeptidases (Fig. 3). One such example is aminopeptidase B, which seems to lack an epoxide hydrolase activity, although conflicting data are present in the literature (29, 30). Apparently, from a noncatalytic function in certain early aminopeptidases, evolution has gradually brought Asp-375 into a key role for the epoxide hydrolase reaction in LTA4H. In this context, it is interesting to note that Asp-375 is indeed present in Saccharomyces cerevisiae LTA4H, a product of an ancestral gene for mammalian LTA4H that exhibits both a leucyl aminopeptidase and a primitive epoxide hydrolase activity against LTA₄ (31, 32).

Figure 3.

Sequence alignment of members of the M1 family of the MA clan of metallopeptidases. Shown are residues subjected to mutational analysis and vicinal stretches of amino acids. Asp-375 (boxed) is the only residue specifically involved in the epoxide hydrolase reaction.

Exchange of Gln-134 for a nonpolar residue such as Leu or Ala led to an unexpected increase in epoxide hydrolase efficiency, whereas a conservative replacement, Gln → Asn, had minimal effects. The proper interpretation of these results is not obvious but the data may be taken as evidence that the polar side chain of Gln-134 plays a regulatory role in the epoxide hydrolase mechanism. The location of Gln-134 in the hydrophilic patch is in line with this hypothesis.

Potential Role of Asp-375 in the Epoxide Hydrolase Reaction.

The bioactivity of LTB₄ depends on the presence of a 12R-hydroxyl group and a cis-trans-trans-conjugated triene system. After the determination of the crystal structure of LTA4H we are now beginning to understand the mechanisms by which the enzyme controls the generation of this chemistry during conversion of LTA₄ into LTB₄. A narrow, L-shaped hydrophobic pocket has been found at the active center, which may be the binding site for LTA₄, and into which LTA₄ has been modeled (Fig. 1). Indeed, several residues within the active site-peptide K21 are located along this cavity, including Asp-375, which is also a component of a hydrophilic patch positioned close to the point where the L-shaped cavity bends. If LTA₄ is modeled into this hydrophobic pocket with a head-to-tail orientation that positions the allylic epoxide close to the zinc, the C7-C20 olefinic tail fits snugly and occupies almost the entire cavity. With this orientation of the substrate, C12 gets close to the hydrophilic patch and Asp-375, suggesting an involvement of this residue in the introduction of the 12R-hydroxyl group (Fig. 4). In line with this hypothesis, the structure of LTA4H in complex with bestatin revealed the presence of a water molecule hydrogen-bonded to Asp-375 that could be the origin of the hydroxyl moiety. Hence, assuming that LTA4H operates according to an S_N1 mechanism, we propose that Asp-375 is a major determinant for the stereospecific nucleophilic attack at C12 of a carbocation intermediate to generate the 12R-hydroxyl group of LTB₄. In this capacity, Asp-375 is also of fundamental importance for the ability of LTA4H to catalyze the formation of a biologically active product.

Figure 4.

Role of Asp-375 in the epoxide hydrolase reaction. LTA₄ is modeled into the L-shaped hydrophobic pocket with its ω-end pointing to the bottom of the cavity. After activation and opening of the epoxide by the Zn²⁺ and Glu-271 according to an S_N1 mechanism, a carbocation is formed whose charge will be delocalized over the conjugated triene system. At C12, a nucleophilic attack can be mediated by means of Asp-375 to produce the 12R-hydroxyl group of LTB₄ (for further details see text). For an alternative role of Glu-271 involving an ester intermediate, see ref. 8.

Because exchange of the polar side chain of Gln-134 for an alkyl moiety resulted in a greatly enhanced epoxide hydrolase activity, it is also possible that this residue has a regulatory, albeit not critical, role in this reaction. Further studies are needed to elucidate the detailed molecular mechanisms by which Asp-375 governs the stereoselective chemistry catalyzed by LTA4H.

Abbreviations

LTA₄

leukotriene A₄ (5S-trans-5,6-oxido-7,9-trans-11,14-cis-eicosatetraenoic acid)

LTB₄

leukotriene B₄ (5S,12R-dihydroxy-6,14-cis-8,10-trans-eicosatetraenoic acid)

LTA4H

LTA₄ hydrolase

PGB

prostaglandin B