MARCH/APRIL1994 VOLUMfi 7, NUMBER 2 Q Copyright 1994 by the American Chemical Society
Communications Microsomal and Soluble Epoxide Hydrolases Are Members of the Same Family of C-X Bond Hydrolase Enzymes Gerard
M. Lacourciere and Richard N. Armstrong'
Department of Chemistry and Biochemistry, University of Maryland, College Park, Maryland 20742 Received October 29, 1993"
Sequence alignments of mammalian microsomal (MEH) and soluble epoxide hydrolases (SEH) with bacterial haloalkane dehalogenase (HAD) and haloacetate dehalogenase fHAcD) together with structural and functional evidence suggest that these four enzymes are structurally and mechanistically related. The catalytic mechanism of HAD and MEH have been recently shown to involve an ester intermediate formed by alkylation of an active site carboxyl group. Very pronounced sequence similarities of regions of MEH, SEH, and HAcD with the active site of HAD suggest that all four enzymes belong to the same family of C-X bond hydrolases which involve an alkyl-enzyme intermediate. The catalytic triads (nucleophile-base-acid) of MEH and SEH are proposed to be Asp226-His431-Asp352 and Asp333-His523-Asp495, respectively, on the basis of sequence alignments with HAD. Although compelling arguments, through sequence alignments, can be made for the assignment of the nucleophile-base pair of the triad, the identity of the acid residue (e.g., Asp352 and Asp495) is more speculative. The threedimensional structures of both MEH and SEH are suggested to contain structural elements af the alp hydrolase fold.
Introduction Epoxide hydrolases catalyze the addition of water to epoxides and mene oxides. The mammalian epoxide hydrolases include two enzymes, microsomal epoxide hydrolase (MEHI' and soluble epoxide hydrolase (SEH), which have rather broad substrate specificities and two epoxide enzymes, leukotriene & hydrolase and hydrolase, which are specific for their title substrates (1-
* To whom correspondence should be addressed. Phone: (301) 4061812; FAX: (301) 406-7966. Abstract published in Aduance ACS Abstracts, January 16, 1994. 1 Abbreviations: MEH, microsomal epoxide hydrolase; SEH, soluble epoxide hydrolase; HAD, haloalkane dehalogenase; HAcD, haloacetate dehalogenase.
5). Mammalian MEH and SEH have different physical
properties and substrate preferences. The soluble (cytosolic or peroxisomal) enzyme is a homodimeric protein with a subunit molecular m-8 of about 62.3 m a . The microsomal enzyme has a molecular mass of 52.5 kDa (6). Little is known about the aggregation state of MEH in the microsomalmembraneor when solubilized. &thoue;h both MEH and SEH have broad, overlapping specificities, their individual substrate preferences me quite distinct (7). The catalytic mechanism of the microsomal enzyme was long thought to involve a direct, general-base-assieted attack Of water On the (89'), an mechanism involving an alkyl-enzyme intermediate was
0893-228x/94/?707-0121$04.50/0Q 1994 American Chemical Society
122 Chem. Res. Toxicol., Vol. 7, No. 2, 1994
Scheme 1
proposed as early as 1980 (10).We recently reported chemical evidence that the catalytic mechanism of MEH does indeed involve an ester (or alkyl-enzyme) intermediate, as illustrated in Scheme 1A (11). This conclusion is buttressed by the observation that portions of the sequence of MEH bear an unmistakable resemblance to regions of sequence from the active site of a bacterial haloalkane dehalogenase (HAD) and haloacetate dehalogenase (HAcD) (12,13). The former enzyme was recently shown by X-ray crystallographic analysis to catalyze the dehalogenation of 1,2-dichloroethane by way of an ester intermediate (Scheme 1B) (14). Thus, mammalian MEH is a member of a previously unrecognized class of C-X bond hydrolases that catalyze their reactions by way of an alkyl-enzyme intermediate" It seems reasonable to expect that other epoxide hydrolases such as SEH might also be members of the same mechanistic class of enzymes. The primary structures of three mammalian SEH enzymes, deduced from cDNA clones, were reported recently (15-1 7). Somewhat surprisiily, homology searches conducted in these investigations at both the DNA and protein level were reported to have revealed no significant similarity between the sequences of SEH and MEH or any other protein in the data banks. We have compared the sequences of SEH with MEH and two bacterial dehalogenasesand come to the opposite conclusion. The mammalian microsomal and soluble epoxide hydrolases differ in many respects, but as we report here, appear to share similar structural motifs in their active sites, which suggests that they have a common, two-step catalytic mechanism that involves an ester (alkyl-enzyme) intermediate.
Methods Homologous regions in the protein sequences were identified using the Lawrence algorithm (18) and the Dayhoff cost matrix (19)imbedded in the EuGene program package (Lark Sequencing Technologies,Houston, TX). The alignments were considered significant if the standard deviation alignment score was 12.5 and/or if the region corresponded to part of the active site of HAD as defined by crystallographic analysis and occurred in the correct linear arrangement of at least three of the four polypeptides. The soluble (15) and microsomal (6)epoxide hydrolases from rat were chosen for sequence alignments though similar resultscan be obtainedwith correspondingsequencesfrom other species. Comparisons were also made to the mouse leukotriene & hydrolase (20). ~~
~
~~
~
*We use the term C-X bond hydrolase to refer only to hydrolase enzymes that proceed by way of an alkyl-enzyme (eater) intermediate. The C-X bond hydrohea are dietinguished from the more common hydrolases involving acyl-enzyme (ester) intermediates by the fact that the carbonyl portion of the ester is contributed by the enzyme, not the substrate,and that the oxygen atom in the hydrolysis product ia derived directlyfrom the enzyme and only indirectlyfrom the solvent pool. That is, the enzyme acta as a covalenthydroxideshuttle from solventto product.
Communications
Results and Discussion The homologous regions of mammalian SEH and MEH and the bacterial dehalogenases (HAD and HAcD) are illustrated in Figure 1. Four homology regions, numbered as they occur in the linear sequence, were identified by the above criteria. Regions 1 and 2 exhibit the most significant homology both from a statistical and functional standpoint. The carboxylate side chain of Asp124, which is the nucleophile in the formation of the 8-chloro ester intermediate in the HAD-catalyzed reaction (Scheme lB), is located in homology region 2. This particular region has the highest sequence alignment score in four out of the six possible pairwise comparisons, as summarized in Table 1. In only one case, the comparison of MEH and HAcD, did the pairwise sequence alignment fail to identify the active site aspartate residue. Although there is no direct evidence that Asp226 is the active site nucleophile in MEH, we have postulated that this is the case on the basis of single-turnover experiments that indicate an ester intermediate intervenes in this reaction and on the basis of sequence alignments with HAD (12 ) . Single-turnover isotopic labeling experiments have not been performed on SEH. However, the pronounced homology domain found between SEH and HAD, which contains 99 residues and includes both regions 1 and 2, suggests that the catalytic mechanism of SEH also involves an ester intermediate with Asp333 as the nucleophile. Five of six pairwise alignments of homology region 1 show significant sequence similarity (standard deviation alignment scores 22.5). This region contains a His-GlyX-Pro sequence, the X residue of which is postulated to help form the oxyanion hole to stabilize the tetrahedral intermediate in the hydrolysis of the ester. The crystal structure of HAD suggests that the main-chain amide NH of the X residue (Glu56 of HAD) has such a function. This group of residues forms a cis-proline reverse turn with the side chain of the X residue pointing away from the active site cavity and the amide NH pointing into the cavity (21). It is very possible that the amide NH of the X residue serves the same function in all four enzymes. That the X residues (TrplBO in MEH and Phe263 of SEH) imbedded in the sequence precedingthe cis-proline turn are different is of little consequence since it is the amide NH of the backbone that presumably participates in catalysis. However, it is interesting to note that mutation (His148 Asnl48) of one of the conserved residues in the His-GlyX-Pro motif of MEH has a deleterious but not fatal effect on the catalytic activity of the enzyme (22). Homology region 4 contains the residue (His289) that acts as a general base to catalyze the addition of water to the carbonyl group of the 8-chloro ester in the second half-reaction of HAD (Scheme 1B). Although the sequence alignment scores in this region are not statistically significant,the recent identification (22)of His431 of MEH as an essential residue in catalysis supports a meaningful functional alignment of this region. The fact that all four enzymes have a Gly-His-aromatic residue sequence near (within 30 residues of) their C-terminus suggests that His523 probably serves as the general base in the hydrolysis of the ester in SEH-catalyzed reactions. We also note that it is common for the His residue (general base) of the catalytic triad in enzymes of the cul8 hydrolase fold family to be proceeded by a residue with a small side chain (Gly or Ala).
-
Chem. Res. Toxicol., Vol. 7, No. 2, 1994 123
Communications Region 1
* MEH
145 LMVHGWPGSFYEFYKIIPLLTDPKSHGLSDE
SEH
260 CLCHGFPESWFSWRYQIPAAQAGFRVLAIDMKGYGDSSSPPEIEEYAMELLCEE
HAD
51 LCLHGEPTWSYLYRKMIPVFAESGARVIAPDFFGFGKSDKPVDEEDYTFEFHRNF I I l l I : :: I : I I I I I :I I 74 DRSNYSFRTFAHD 30 LMLHGFPQNRAMWARVAPQLAEHHTWCA
IIIi i l i
iIIIIi II I HACD
:
I I :I:
I
I II:I I
I:) I
I : I:I
I:
Region 2
**
MEH
SEH HAD HACD
208 FYKLMTRLGFQKFYIQGGDSLICTNMAQMVPNHVKGLHLNMA :: : I I I I I :: I l l I :I: : 315 MVTFLNKLGIPQAVFIGHDWAGVLVWNMALFHPERVRAVASLNTPLM : : :: : I : :: : / I I I : : I I : : :I II 106 LLALIERLDLRNITLVVNDWGGFLGLTLPMADPSRFKRLIIMNACLM I :: : I : II :I:II I : : I I :I: 87 QLCVMRHLGFERFHLVGHDRGGRTGHRMALDHPEAVLSLT Region 3
*
MEH
SEH HAD
352
DDLL
I :I 488 LMVTAEKDIVLRPEM :I I I : I I:: 253 FMAIGMKDKLLGPDV
Region 4
*
428 RGGHFAA I1 520 DCGHWTQ I I1 I 286 DAGHFVQ :Ill
269 PGGHFFV Figure 1. Sequence alignments of sections of microsomal epoxide hydrolase (MEH),soluble epoxide hydrolase (SEH),haloalkane dehalogenase (HAD),and haloacetate dehalogenase (HAcD).The sequences are from refs 6,15,12, and 13,respectively. Three of the six possible pairwise alignments are shown. The proposed active site residues of SEH and MEH deduced from the sequence alignments and the three-dimensionalstructure of HAD are identified by asterisks, and their possible functions are discussed in the text. Sequence identities and similarities are indicated by uln and ":", respectively. HACD
Table 1. Sequence Similarities for Homology Region 2 in Six Pairwise Comparisons of Four C-X Bond Hydrolases. SEH HAD MEH SD no. SD no. SD no. protein res % Id score res % Id score rea % Id score HAcD HAD
5 40.0
