(Received for publication, March 12, 1997)
From the Department of Biochemistry, Groningen
Biomolecular Sciences and Biotechnology Institute, University of
Groningen, Nijenborgh 4, 9747 AG, Groningen, The Netherlands
The epoxide hydrolase gene from
Agrobacterium radiobacter AD1, a bacterium that is able to
grow on epichlorohydrin as the sole carbon source, was cloned by means
of the polymerase chain reaction with two degenerate primers based on
the N-terminal and C-terminal sequences of the enzyme. The epoxide
hydrolase gene coded for a protein of 294 amino acids with a molecular
mass of 34 kDa. An identical epoxide hydrolase gene was cloned from
chromosomal DNA of the closely related strain A. radiobacter CFZ11. The recombinant epoxide hydrolase was
expressed up to 40% of the total cellular protein content in
Escherichia coli BL21(DE3) and the purified enzyme had a
kcat of 21 s1 with
epichlorohydrin. Amino acid sequence similarity of the epoxide hydrolase with eukaryotic epoxide hydrolases, haloalkane dehalogenase from Xanthobacter autotrophicus GJ10, and bromoperoxidase
A2 from Streptomyces aureofaciens indicated that it
belonged to the
/
-hydrolase fold family. This conclusion was
supported by secondary structure predictions and analysis of the
secondary structure with circular dichroism spectroscopy. The catalytic
triad residues of epoxide hydrolase are proposed to be
Asp107, His275, and Asp246.
Replacement of these residues to Ala/Glu, Arg/Gln, and Ala, respectively, resulted in a dramatic loss of activity for
epichlorohydrin. The reaction mechanism of epoxide hydrolase proceeds
via a covalently bound ester intermediate, as was shown by single
turnover experiments with the His275
Arg mutant of
epoxide hydrolase in which the ester intermediate could be trapped.
Epoxide hydrolases can hydrolyze epoxides to their corresponding
diols by addition of a water molecule. There is a strong interest in
these enzymes since they play a key role in the detoxification of
xenobiotic compounds and have great potential in enantioselective chemistry. Most research has focused on mammalian epoxide hydrolases since these enzymes are of toxicological relevance. The epoxide hydrolase genes that have been cloned so far are of mammalian, insect,
and plant origin (1-9), and they can be distinguished into a class of
microsomal enzymes and a class of soluble enzymes based on cellular
localization and biochemical properties like substrate specificity. The
mammalian epoxide hydrolases are believed to belong to the
/
-hydrolase fold family (10-12) since they show a low but
significant sequence similarity with haloalkane dehalogenase from
Xanthobacter autotrophicus GJ10 (13) of which the
three-dimensional structure has been solved by x-ray crystallography
(14). Sequence similarities are low within this
/
-hydrolase fold
family and are restricted to specific areas within the topology, such
as the preserved positions of the catalytic triad residues.
Agrobacterium radiobacter AD1 is a Gram-negative bacterium that was isolated for its ability to use epichlorohydrin as the sole carbon and energy source (15). Epichlorohydrin is hydrolyzed by an epoxide hydrolase to 3-chloro-1,2-propanediol, which is then converted to glycidol by a haloalcohol dehalogenase. Glycidol is further converted to glycerol which then enters the central metabolic pathway (16). The epoxide hydrolase that catalyzes the cofactor-independent hydrolysis of epichlorohydrin was purified to homogeneity and found to be a monomeric globular protein with a molecular mass of 35 kDa. Its similarity to haloalkane dehalogenase was suggested but could not be proven by biochemical analysis (15). The enzyme has a broad substrate range and epichlorohydrin and epibromohydrin were found to be the best substrates.
Not much is known about structure and sequence of microbial epoxide
hydrolases. To obtain insight in the structure and mechanism of
microbial epoxide hydrolases, we decided to characterize the epoxide
hydrolase gene from A. radiobacter AD1. Due to its small size compared with microsomal and soluble epoxide hydrolase and its
bacterial origin, this enzyme has potential in structural and
biocatalytic studies. Based on the sequence and secondary structure
analysis, the epoxide hydrolase is predicted to be an /
-hydrolase
folded enzyme. The residues Asp107, His275, and
Asp246 were identified as the catalytic triad residues by
sequence analysis and site-directed mutagenesis.
All chemicals were purchased from Acros chimica, Merck, or Sigma. Super taq polymerase was purchased from Sphaero Q. Restriction enzymes and other molecular biology enzymes were from Boehringer Mannheim. Oligonucleotide construction and amino acid analysis were done by Eurosequence BV, Groningen. Sequencing was done with the T7 sequencing kit from Pharmacia.
Strains and Growth ConditionsA. radiobacter AD1, formerly Pseudomonas sp. strain AD1 (15), is able to grow on epichlorohydrin and was maintained at 30 °C on sealed MMY plates (17) with 5 µl of epichlorohydrin added to a piece of filter paper in the lid of the Petri dish. Agrobacterium strain CFZ11 was isolated on 1,3-dichloro-2-propanol. Its properties are similar to those of strain AD1.1 For the isolation of chromosomal DNA, A. radiobacter AD1 was cultivated in closed flasks containing one-fifth of the total volume of MMY medium with 5 mM 1,3-dichloro-2-propanol as the growth substrate. For the preparation of chromosomal DNA, strain CFZ11 was cultivated in NB medium. All strains of Escherichia coli were cultivated in liquid LB medium with, when needed, ampicillin added to a final concentration of 100 µg/ml (18). E. coli JM101 was used for the production of single-stranded DNA for sequencing purposes (19). E. coli BW313 was used for the production of uracil containing single-stranded DNA for Kunkel mutagenesis (20). E. coli BL21(DE3) was used for the high-level expression of epoxide hydrolase (21).
Cloning and Sequencing of the Epoxide Hydrolase GeneThe
epoxide hydrolase gene was initially cloned by means of the polymerase
chain reaction, using degenerate primers that were designed on the
amino acid sequence of the N- and C-terminal amino acid sequences of
the protein (15). The N-terminal sequence was determined again and
found to be TIRRPEDFKHYEVQLPDVKIHYVREGAGPTLLL. An ATG start codon
within a NcoI restriction site was present in the forward
primer:
5-CGGGTACCATGGCAATTCGACGTCCAGAYGAXTTXAAYCAXTAXGA-3
(Y = A/G; X = C/T; start codon shown in bold; NcoI
site underlined). A stop codon and a NcoI restriction site
were incorporated in the reverse primer:
5
-CGGGATCCATGGCTAGCGYAAZGCZGTXTTRAT-3
(Z = G/A/T/C; r = T/G/A; stop codon shown in
bold; NcoI site underlined).
Total DNA of A. radiobacter AD1 was isolated from cells that were cultivated on 1,3-dichloro-2-propanol, using standard procedures (22). DNA amplification by the polymerase chain reaction was done with 1 µg of total DNA on a Perkin Elmer PCR2 apparatus using the standard amplification protocol described by Innis and Gelfand (23), with the exception that primer annealing was first done at 37 °C (3 rounds) and then at 40 °C (25 rounds). The amplified DNA was digested with NcoI and ligated into the NcoI site of the expression vector pGEF+ (plasmid pGELAF+ (24) without the NcoI fragment containing the dhlA gene), resulting in a translational fusion. The ligation mixture was transformed to E. coli BL21(DE3) cells by means of electroporation and a colony displaying epoxide hydrolase activity (see under "Enzyme Assays") was selected. Plasmid DNA was isolated (18) and the cloned fragment was sequenced by the dideoxy method (25). The construct pEH20 was used for further study.
Total DNA from strain CFZ11 was isolated by standard methods (22). For Southern blot analysis the DNA was digested with different restriction enzymes. After agarose gel electrophoresis and capillary transfer onto positively charged nylon membrane (Boehringer Mannheim) the DNA was hybridized with digoxygenin-labeled DNA of the epoxide hydrolase gene from the construct pEH2O. For detection of the hybridizing fragments the standard protocol of Boehringer Mannheim was followed. A hybridizing 2.3-kilobase BamHI/HindIII fragment was cloned into pWKS130 (26). This clone was finally sequenced using the Tag DyeDeoxyTM Terminator Cycle Sequencing Kit (Perkin Elmer) and the ABI 373 automated sequencer (Applied Biosystems Division of Perkin Elmer). Potential promoter sites were searched by using the promoter prediction program Promoter Prediction by Neural Network NNPP (27).
Expression and Purification of Epoxide HydrolaseBoth wild
type and mutant epoxide hydrolase were expressed in E. coli
BL21(DE3). Plasmid DNA was transformed by electroporation to E. coli cells, which were then plated out on LB plates containing ampicillin and incubated overnight at 30 °C. A preculture was started by inoculating 100 ml of LB containing ampicillin with the
transformants from a plate to a starting OD600 of 0.1 and was incubated at 30 °C until an OD600 of 1 to 2 was
reached. The preculture was diluted in 1 liter of LB, containing
ampicillin, and the culture was incubated overnight at 20 °C. The
cells were centrifuged, washed, and resuspended in 50 ml of TEMAG
buffer (10 mM Tris-SO4, 1 mM EDTA,
1 mM -mercaptoethanol, 0.02% sodium azide, and 10%
glycerol, pH 7.5). Cells were broken by continuous sonication and the
extract was centrifuged (200,000 × g, 90 min, 4 °C). The supernatant was applied on a 30-ml DE52 anion exchange column and elution was carried out with a gradient of 0 to 1 M ammonium sulfate in TEMAG. The collected fractions that
displayed epichlorohydrin activity and had the highest protein content
were pooled and dialyzed against PEMAG buffer (5 mM
potassium phosphate, 1 mM EDTA, 1 mM
-mercaptoethanol, 0.02% sodium azide, and 10% glycerol, pH 6.8). A
30-ml hydroxylapatite column was used for further purification, using a
gradient of 5 to 100 mM of phosphate in PEMAG. The
fractions that contained purified epoxide hydrolase were pooled and
dialyzed against TEMAG buffer. The enzyme was stored at 4 °C or
20 °C. Glycerol (10% v/v) was used for its stabilizing effect on
epoxide hydrolase activity upon storage.
Site-directed
mutagenesis was done as described by Kunkel (20) with uracil-containing
single-stranded plasmid pEH20 as template. Asp107 was
mutated to Glu with primer D107E
(5-CGTTGGCCATG(C/A)ATTCGCGGCCAT-3
, Glu codon
in bold, EcoRI restriction site underlined), and to Ala with
primer D107A (5
-GGCCATGCGTTCGCG-3
, Ala codon in bold).
Confirmation of a mutated codon was obtained by PCR with a control
primer (5
-CGTTGGCCATGCG-3
, mutation underlined) and a
downstream primer (5
-TGGCAGCAGCCAACTCAGCT-3
). His275 was
mutated to Arg and Gln with primer H275RQ
(5
-GACGATTGAAGACTGCGGT(A/C)(AG)GTTCTTGATGGTC-3
, mutated codon shown in bold, removed PvuI restriction site
underlined). Asp246 was mutated to Ala with primer D246A
(5
-TTGGGAGCTACTTGC-3
, Ala codon shown in bold).
Confirmation of a mutated codon was obtained by PCR with a control
primer (5
-GGGCACGCAAGTAG-3
, mutation underlined) and an
upstream primer (5
-TAATACGACTCACTATAGGG-3
). The Kunkel mixture was
transformed to E. coli BL21(DE3) cells by electroporation.
Recombinants were screened for epoxide hydrolase activity and plasmid
DNA was isolated. Mutations were confirmed either by restriction
analysis or by PCR with primers that could only amplify DNA with the
desired mutation. Finally, the mutations were checked by dideoxy
sequencing (25).
Epoxide hydrolase activities in whole cells and column fractions were determined by a microtiter plate assay, based on a chromogenic reaction of an epoxide with 4-nitrobenzylpyridine (28). Small amounts of cells or 10-µl samples of column fractions were added to a 96-well microtiter plate together with 100 µl of TE buffer (50 mM Tris-SO4, 1 mM EDTA, pH 9.0) containing 10 mM epichlorohydrin. After incubation for 1-10 min at room temperature, 50 µl of reagent A (100 mM 4-nitrobenzylpyridine in 80% ethylene glycol and 20% acetone (v/v)) was added. The microtiter plate was tightly sealed with silicone rubber and was incubated for 10 min at 80 °C. After cooling to room temperature, 50 µl of reagent B (50% triethylamine and 50% acetone (v/v)) was added. A blue color appeared when epichlorohydrin was not degraded, else the mixture stayed colorless.
Epoxide hydrolase activities were determined quantitatively by following substrate depletion using gas chromatography of ethereal extracts (15) or by following substrate depletion and diol production by gas chromatography of reaction mixtures quenched in acetone. A suitable amount of epoxide hydrolase was incubated in TE buffer with 5 mM substrate. At various time points, 100 µl of sample was added to 1 ml of ice-cold acetone containing 1-nonanol as the internal standard. Protein and salts were removed by centrifugation (15 min, 4000 × g) and the extract was analyzed by GC using a 0.2 mm × 25-m CP-Wax57-CB column (Chrompack, Middelburg, The Netherlands) and a flame-ionization detector. 1 Unit of enzyme activity is defined as the amount of enzyme that catalyzes the production of 1 µmol of diol/min. Protein determination was carried out with Coomassie Brilliant Blue with bovine serum albumin as a standard or by measuring the absorbance of purified enzyme at 280 nm. One OD280 unit corresponded with 0.42 mg of epoxide hydrolase/ml as was determined by the biuret method (29), amino acid analysis, and by dissolving a known amount of freeze-dried epoxide hydrolase in water.
The specific activities of the mutant enzymes and wild type enzyme were
determined in concentrated cell free extracts. A pellet of cells that
was washed with TEM buffer (50 mM Tris-SO4,
mM EDTA, and 5 mM -mercaptoethanol, pH 7.5),
was resuspended in one pellet volume of buffer and disrupted by
sonication. The suspension was centrifuged for 90 min (200,000 × g, 4 °C). The cell-free extract contained protein
concentrations of 80-150 mg/ml and the epoxide hydrolase content was
30-40% of the total protein content, as confirmed by
SDS-polyacrylamide gel electrophoresis and density scanning. 200 µl
of protein extract or an adequate dilution thereof was added to 1.5 ml
of TE buffer (Tris-SO4, 5 mM EDTA, pH 9.0) with
5 mM epichlorohydrin and incubated at 30 °C. Samples of
200 µl were taken in time and quenched in 1.8 ml of ice-cold acetone with 1-nonanol as the internal standard and analyzed by GC.
Far-UV CD-spectra were recorded on a AVIV circular dichroism spectrometer (62A DS) by measuring the change in ellipticity in millidegrees. Enzyme was dialyzed against a 5 mM phosphate buffer, pH 6.8, and spectra were recorded in a 1-mm cuvette at 25 °C. The CD-spectra were corrected for buffer absorbance. Secondary structure elements were extracted from the spectra by using the programs CONTIN (30), SELCON (31), and K2D (32).
Homology Search and Secondary Structure PredictionsThe BLAST program (33) was used to screen protein and DNA data bases for proteins that shared sequence similarity. Multiple sequence alignments were made in ClustalW v1.6 (34). Secondary structure predictions were carried out with the programs SopM (35), ssp (36), Sspred (37), and Predict-Protein (38) that were offered as services on the World Wide Web. The results were compared and similar predictions were taken as a consensus prediction.
Single Turnover ExperimentsSingle turnover experiments
were performed at 30 °C with wild type enzyme and the
His275 Arg mutant on a rapid quench-flow apparatus
(RQF-63) from Kintek Instruments using 1 mM purified enzyme
and 0.5 mM epichlorohydrin in TEMAG buffer, pH 7.5. All
concentrations given represent values after one-to-one mixing of
protein and substrate. The enzyme was concentrated with an Amicon
ultrafiltration cell using a PM10 filter. After mixing, the reaction
mixture with a total volume of 100 µl was quenched with 117 µl of
acetone, and directly injected into 800 µl of ice-cold acetone with
1-nonanol as the internal standard. The concentrations of
epichlorohydrin and 3-chloro-1,2-propanediol were determined by GC. The
experiment was repeated with 0.6 mM His275
Arg mutant of epoxide hydrolase, using 0.6 or 0.3 mM
epichlorohydrin, all in TEM buffer, pH 7.5.
The
epichlorohydrin epoxide hydrolase echA gene was originally
cloned by amplifying the gene by PCR by using two degenerate primers
that were designed on basis of the N- and C-terminal amino acid
sequence. The amplified gene was fused into the start codon of the
expression vector pGEF+, resulting in the construct pEH20. Upon
cloning, the second amino acid was changed from a threonine to an
alanine residue, but it did not have a noticeable effect on the
activity or stability of the enzyme. The cloning procedure was repeated
twice for separate polymerase chain reactions to identify possible
errors that could be introduced during amplification of the gene. All
three clones were entirely sequenced. When the three constructs were
compared with each other, two silent mutations (G231 A, and C618
T) were detected in the echA gene of the construct
pEH2O. Since the amino acid sequence of epoxide hydrolase was not
effected, plasmid pEH20 was used for further study.
To eliminate the possibility that the sequenced epoxide hydrolase gene
differs from the actual sequence due to PCR errors, we decided to clone
the gene directly from chromosomal DNA. Attempts to clone the epoxide
hydrolase gene from chromosomal DNA of A. radiobacter AD1
failed. However, using A. radiobacter CFZ11, we were able to
clone the epoxide hydrolase gene based on homology with the PCR clone
from strain AD1. An open reading frame was found on a 2.3-kilobase
BamHI/HindIII fragment that was identical to the
one in pEH20, that was cloned from strain AD1 (Fig. 1). The echA gene coded for a polypeptide of 294 amino acids and
had a G + C content of 52%, which was very close to an overall G + C
content of 51%. The epoxide hydrolase has a calculated molecular mass
of 34,064 kDa, which fitted the experimentally determined mass of 35 kDa (15). The N-terminal sequence and the two peptides that were
determined by Edman degradation were fully conserved in the translated
amino acid sequence of the echA gene (15). The internal and
C-terminal peptide were also flanked by methionine residues at the
sites where the protein was originally cleaved (Fig. 1). Southern blot
analysis with an echA probe on double digested chromosomal
DNA of strain AD1 with combinations of BamHI, SalI, SmaI, EcoRI, and
HindIII, resulted in hybridization signals of equal size as
can be deduced from Fig. 1. This indicates that strains AD1 and CFZ11
contain identical stretches of DNA.
A perfect ribosome-binding site was found 7 base pairs upstream of the echA gene and also a possible promoter sequence was predicted by the program NNPP with a score of 0.33 on a scale of 0 to 1. Further upstream of the echA gene, two other promoter sites are predicted with scores of 0.90 and 0.92. Upstream of the echA gene an open reading frame was found coding for 116 amino acids with a ribosome-binding site and two potential promoter sequences (scores 0.93 and 0.77). The hypothetical protein showed 24% sequence similarity with a hypothetical protein of E. coli (YCHN_ECOLI; SwissProt entry code) of 117 amino acids of which the function is not described. Another open reading frame of 315 base pairs, ranging from base 539 to 854, lacked a ribosome-binding site and had no significant sequence similarity with sequences in the DNA and protein libraries. Downstream of the echA gene the beginning of an open reading frame was found that coded for 34 amino acids which had 41% sequence similarity with the N terminus of haloalcohol dehalogenase HheA of Corynebacterium sp. strain N-1074 (39). Upstream of this open reading frame lies a perfect ribosome-binding site and three putative promoter sequences (scores 0.90, 0.93, and 0.96).
Expression and Characterization of Epoxide HydrolaseThe
echA gene in pEH20 is under control of a T7 promoter and
epoxide hydrolase was expressed constitutively in a soluble and active
form up to 40% of the total cellular protein content in E. coli BL21(DE3). For purification of the enzyme, cells were harvested at an OD600 of 4-5 and typically 100 mg of more
than 98% pure protein could be obtained from a 1-liter culture with a
purification factor of 2.5 (Fig. 2). The protein could
be stored for at least 3 months at 4 °C or at 20 °C without
significant loss of activity.
The specific activities of purified epoxide hydrolase for some substrates are listed in Table I. Epichlorohydrin and epibromohydrin are the best substrates. Short and long chain 1,2-epoxyalkanes are good substrates for epoxide hydrolase, and since styrene oxide is also degraded, the active site pocket must be sufficiently large to harbor these substrates. Isoprene monoxide (2-methyl-2-vinyloxirane) was also converted, indicating that branching at the second carbon atom of the epoxide ring is possible. No activity was found for cis-2,3-epoxybutane and since both isomers of stilbene oxide were also not degraded (15), it is essential that the epoxide ring is located at the primary carbon atom.
|
A substrate depletion curve of epoxide hydrolase with epichlorohydrin
as the substrate, followed a straight line to the detection limit of 50 µM, indicating that the Km value for
epichlorohydrin was below 50 µM. Since the
Km value for epichlorohydrin is very small, the
specific activity of 38 units/mg of protein1 at a
substrate concentration of 5 mM can considered to be the Vmax, corresponding to a
kcat of 21 s
1.
A
sequence similarity search with the amino acid sequence of
epichlorohydrin epoxide hydrolase (EchA) was performed in various protein and DNA data bases. A selection of the most similar proteins is
shown in a ClustalW alignment in order of their sequence similarity to
EchA (Fig. 3). All epoxide hydrolase sequences that were
present in the data banks were scored in the search. The soluble
epoxide hydrolases from mammalian and plant origin (1-3, 7, 8) were
found to be more similar to EchA than the microsomal epoxide hydrolases
from mammalian and insect origin (4-6, 9), which are
membrane-associated enzymes. Sequence similarity was also found with
putative hydrolases from Caenorhabditis elegans (40) and
Stigmatella aurantiaca (41), but they were omitted from the
alignment because the proteins have not been studied. The highest
similarity was found between EchA and the fluoroacetic acid
dehalogenase (DehH1) from Moraxella sp. strain B. Based on sequence similarity with haloalkane dehalogenase (DhlA) from X. autotrophicus GJ10 (13), DehH1 is believed to be an
/
-hydrolase fold enzyme (42). Other hydrolases that had
significant sequence similarity with EchA are two 2-hydroxymuconic
semialdehyde hydrolases (SwissProt entry codes: DMPD_PSEPU and
XYLF_PSEPU), 2-hydroxy-6-oxo-2,4-heptadienoate hydrolase (TODF_PSEPU),
2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoate hydrolase (BPHD_PSES1),
magnesium-chelatase 30-kDa subunit (BCHO_RHOCA), and
dehydrolipoamide S-acetyltransferase (ACOC_ALCEU).
It has been proposed that the soluble and microsomal epoxide hydrolases
of mammalian origin belong to the /
-hydrolase fold family
(10-12, 43). The sequence similarity of EchA with bromoperoxidase A2
(BpA2) from Streptomyces aureofaciens (44) and DhlA, of
which the three-dimensional structures are known, indicates that EchA also belongs to this class of hydrolases. The overall sequence similarity of EchA with homologous proteins is low but significant (between 13 and 23%), and is mainly located in the N-terminal region.
Low sequence similarities are common between the members of the
/
-hydrolase fold family (45). Two N-terminally located motifs in
EchA, i.e. HGX and GarGXS
(X = any amino acid, ar = aromatic residue) that
are often found in
/
-hydrolase fold enzymes, were also found in
other epoxide hydrolases (10, 12). Both motifs are located on loops
that excurse from the
-sheet and are in proximity of the cap domain.
The part of the alignment that corresponds to the cap domain region of
DhlA and BpA2 shows no sequence similarity at all, which is in
agreement with a role in determining the substrate specificity.
The sequence alignment of Fig. 3 suggests catalytic residues for EchA,
which are in the same position in the alignment as the identified
catalytic residues of BpA2 (44), DhlA (14), soluble epoxide hydrolase
(46), and microsomal epoxide hydrolase (47). The nucleophile (Nu), an
aspartic acid or a serine in the case of BpA2, is conserved among the
depicted hydrolases in the nucleophile elbow sequence
Sm-X-Nu-X-Sm-Sm (Sm = small residue), as
described by Ollis et al. (45). The sequence similarity
indicates that Asp107 may be the nucleophile in EchA. The
histidine residue of the catalytic triad is completely conserved among
the hydrolases and is located close to the C terminus at position 275 in EchA. The sequence around the acid residue of the catalytic triad is
not conserved in the /
-hydrolase fold family (45), but
Asp246 of EchA is clearly aligned with the catalytically
active aspartic acids of BpA2, DhlA, and soluble epoxide hydrolase
(Fig. 3). Residue Glu404 of human microsomal epoxide
hydrolase, that was proposed to be the third member of the catalytic
triad (11), and Glu401 of juvenile hormone epoxide
hydrolase from insect (9) also align with Asp246. No acid
residue is present in DehH1 at a corresponding position. The spacing
between the conserved histidine and the aspartic acid (26-34 residues)
is similar for all other enzymes. Based on sequence similarity,
we propose that the catalytic triad of epoxide hydrolase consists of
the residues Asp107, His275, and
Asp246.
Since the
three-dimensional structures of BpA2 and DhlA are known, we studied the
secondary structure elements of epoxide hydrolase in more detail by
circular dichroism spectroscopy and secondary structure prediction. The
experimental secondary structure elements of the main domains of BpA2
and DhlA are conserved in the sequence alignment (Fig. 3). This
suggests that the secondary structure elements of DhlA and BpA2 can be
extrapolated to EchA. The secondary structure predictions on the amino
acid sequence of EchA led to almost similar results as the alignment.
All the -strands of EchA were predicted to be at the same position
in the alignment as the
-strands of DhlA and BpA2, only
-strand 6 was not present. The same holds for some of the predicted
-helices. When secondary structure predictions were done with the amino acid
sequences of DhlA and BpA2, the
-strands were predicted very well
(not shown). This indicates that the residues that form the
-sheet
highly favor
-stranded structure. The secondary structure predictions were most consistent in the N-terminal part of EchA, which
is also the region with the highest sequence similarity. The secondary
structure of the cap domain region of EchA was predicted to be
predominantly
-helical, although
-stranded structures were also
predicted with some programs. The cap domains of DhlA and BpA2 are
completely
-helical, but they differ in their tertiary structure and
the sequence similarity is too low to predict the location of secondary
structure elements in EchA.
Circular dichroism spectroscopy (CD) was performed on purified epoxide
hydrolase to obtain more structural information. A CD spectrum of wild
type epoxide hydrolase and purified DhlA was recorded in the far-UV
region (Fig. 4). The recorded spectra are much alike and
are typical for proteins with -helical and
-stranded structure
(48). With the programs CONTIN, SELCON, and K2D, the ratios of the
secondary structure elements of EchA,
-helix/
-strand/other, were
determined to be 31/28/41, 32/22/46, and 37/12/51, respectively. Clearly, all programs predict an
/
structure for EchA.
Predictions of the secondary structure elements of DhlA were made as a
control with the same programs and resulted in ratios of 22/18/60,
32/22/46, and 33/13/54, respectively. The programs SELCON and K2D gave
almost identical predictions for EchA and DhlA, indicating that both enzymes have similar structures. The ratio of DhlA, 43/14/43, that was
determined by x-ray crystallography (14), compares relatively well with
the predictions made by SELCON and K2D. The
-strand content was
especially predicted very well by K2D.
Characterization of Epoxide Hydrolase Mutants
In the sequence
alignment of Fig. 3, the residues Asp107,
His275, and Asp246 of epoxide hydrolase were
pointed out as the catalytic residues. To test if these residues were
catalytically active, Asp107 was mutated to Ala and Glu,
His275 was mutated to Arg and Gln, and Asp246
was mutated to Ala. The epoxide hydrolase mutants were all expressed at
20 °C as soluble protein, and in quantities similar to wild type
enzyme. The activities of all mutant enzymes measured in a cell-free
extract with epichlorohydrin were drastically reduced compared with
wild type enzyme, indicating that all three residues are involved in
catalysis. The Asp246 Ala mutant still had some
activity for epichlorohydrin (Table II).
|
Single turnover experiments were performed with wild type enzyme and
the His275 Arg mutant, to test if epichlorohydrin is
converted via a covalently bound ester intermediate. Enzyme (1 mM) was incubated with 0.5 mM epichlorohydrin
and the concentrations of epichlorohydrin and 3-chloro-1,2-propanediol
were measured after different incubation times (Fig. 5).
With wild type enzyme the reaction was complete within 100 ms and
product was formed as soon as the reaction started. No significant
difference between the rates of substrate disappearance and product
formation was observed, which would otherwise be an indication for a
covalently bound ester intermediate.
The single turnover experiment with the His275 Arg
mutant clearly showed that the concentration of epichlorohydrin
decreased significantly in the first 10 s while no product was
formed (Fig. 5B). This implicates that the reaction
mechanism of epoxide hydrolase proceeds via a covalently bound ester
intermediate. After 1 min of reaction time, some product could be
detected (detection limit 50 µM) and after 10 min
epichlorohydrin was almost completely converted to
3-chloro-1,2-propanediol. Although epichlorohydrin was covalently
trapped in the His275
Arg mutant, the enzyme was still
able to hydrolyze the covalently bound ester intermediate with a rate
constant of less than 0.001 s
1, which is below the
detection limit of the steady state rate measurements (Table II).
Substrate depletion was relatively fast in the first 2 s after
which conversion proceeded at a slower rate. This is in agreement with
a slow rate of ester hydrolysis. Repetition of this experiment with
another batch of purified enzyme and various enzyme/substrate ratios
gave reproducible results and indicated that 70% of the His275 Arg enzymes is probably not participating in the
reaction, which may be related to enzyme heterogeneity. The results
could not be explained by preference of the mutant enzyme for one of the enantiomers of epichlorohydrin since both were hydrolyzed at the
same rate (data not shown). A circular dichroism spectrum of the
His275
Arg mutant showed no significant structural
distortions, since the spectrum is similar to that of wild type
epoxide hydrolase (Fig. 4).
The epichlorohydrin epoxide hydrolase (echA) gene of A. radiobacter AD1 was cloned and expressed in E. coli BL21(DE3). The identity was confirmed by the high activity for epichlorohydrin and fragments of amino acid sequence. Epichlorohydrin epoxide hydrolase (EchA) was found to be more similar to soluble epoxide hydrolase than to microsomal epoxide hydrolase. No other bacterial epoxide hydrolase gene has been cloned, but a DNA fragment of 112 base pairs, located upstream of the haloalcohol dehalogenase gene hheB in Corynebacterium sp. strain N-1074, codes for a C terminus of 37 amino acids and has 90% sequence identity with the C-terminal sequence of EchA (39). Downstream of the echA gene of A. radiobacter CFZ11/AD1, we found a segment of an open reading frame encoding for 34 amino acids that had 41% sequence similarity with haloalcohol dehalogenase HheA of Corynebacterium sp. strain N-1074 (39). Since both bacteria have a similar degradation route of epichlorohydrin, it is very likely that the open reading frame found in strain CFZ11 codes for the N terminus of a haloalcohol dehalogenase.
Epoxide hydrolase appears to be an /
-hydrolase folded enzyme. One
of the characteristics of this family of enzymes is that the sequence
similarity is mainly found in some parts of the N-terminal region and
around the catalytic triad residues (10, 12, 42, 49). This also holds
for the echA encoded enzyme. Despite the low sequence
similarity, the GarGXS and the HGXP sequences
were strictly conserved. In DhlA, the sequence HGEP forms the oxyanion hole in which the backbone amide hydrogen of Glu56, which
is located on a sharp cis-proline turn, stabilizes the oxyanion that is formed on the side chain carbonyl oxygen of the nucleophilic Asp124 during the hydrolysis of the ester
intermediate. His54 forms a hydrogen bond
(N
1-O) with the backbone carbonyl of Gly55,
causing a sharp turn (14). Mutation of His148 to Asn in
this motif in rat microsomal epoxide hydrolase led to a significant
decrease in activity (47). This indicates that a sharp turn in the
oxyanion hole is essential for enzyme activity and that therefore this
motif is well conserved. The function of the GarGXS motif in
DhlA is not described, but it is part of a large loop with three
-turns, and it is located close to the oxyanion hole and the
nucleophilic Asp124. Since this motif is well conserved, it
may have a function in positioning of the oxyanion hole in regard to
the nucleophile. In DhlA, the side chain of the aromatic residue
Phe85 of the motif GFGDS is in a tilted-T arrangement with
His54 of the oxyanion hole motif, in which the positively
charged His ring stands perpendicular on the slightly negative charged
surface of Phe85. No hydrogen bonds or salt bridges were
found between these motifs, but a His-Phe interaction can have a
considerable stabilizing effect (50, 51).
The /
-hydrolase fold is a conserved topology with a main domain
that consists of a central
-sheet that is alternated with
-helices that cover both sides. In the structures of haloalkane dehalogenase and bromoperoxidase A2, the
-helical cap domain follows
-strand 6 and covers the active site like a cap. The positions of
the secondary structure elements of haloalkane dehalogenase and
bromoperoxidase A2 compared with the secondary structure elements that
were predicted for epoxide hydrolase strongly suggest a specific topology for epoxide hydrolase (Fig. 3). The cap domain of epoxide hydrolase therefore seems to be located between
-strands 6 and 7 (residues 132-209), since this is the only part of the alignment that
shows little similarity in both sequence and secondary structure predictions. In
/
-hydrolase fold enzymes the
-sheet forms a scaffold for the catalytic triad residues, that are located on loops
excursing from the
-sheet, and since these residues are preserved in
the topology, they are also conserved in the sequence. Mutation of the
residues Asp107, His275, and Asp246
of epoxide hydrolase, resulted in a dramatic drop of enzyme activity, which indicates that these residues are involved in catalysis.
The Asp246 Ala mutant still had some residual activity
for epichlorohydrin, indicating that this residue is involved in
enzymatic activity but not essential. Mutation of the acid residue in
haloalkane dehalogenase,3 2-hydroxymuconic
semialdehyde hydrolase (49), and soluble epoxide hydrolase (46)
resulted in an inactivated enzyme. Probably, Asp246 has
some backup in EchA. In triacylglycerol lipase from Pseudomonas glumae, a partially redundant catalytic aspartate was reported (52). In human pancreatic lipase an alternative catalytic triad was
found in which the catalytic aspartate was shifted from
-strand 7 to
-strand 6 at position 176 (53). In the crystal structure of DhlA,
Asn148 is the analog of Asp176 of human
pancreatic lipase and is located directly after
-strand 6 where it
forms a hydrogen bond with the nucleophilic Asp124 (14). In
the sequence alignment, Asp131 of EchA is aligned with
Asn148 of haloalkane dehalogenase (Fig. 3). So,
Asp131 of epoxide hydrolase is probably positioned close to
the nucleophilic Asp107 and the catalytic
His275. The presence of another aspartate may be sufficient
to retain some activity in EchA when Asp246 is replaced by
alanine. The same argument can also explain why no acid residue could
be found in Fig. 3 for DehH1. Residue Asp129 of DehH1,
which is aligned with Asp131 of EchA and Asn148
of DhlA, could well be part of an alternative catalytic triad, as
present in human pancreatic lipase (53).
Based on these results, we propose a reaction mechanism for EchA in
which the catalytic Asp107 performs a nucleophilic attack
on the primary carbon atom of epichlorohydrin, leading to a covalently
bound ester intermediate (Fig. 6). It was shown earlier
using 18O-labeled water that hydrolysis takes place at the
primary carbon atom of the epoxide ring (15). His275,
assisted by Asp246, abstracts a proton from a water
molecule that hydrolyzes the ester at the carbonyl function of
Asp107. Phe108 of epoxide hydrolase, which is
located next to the nucleophile Asp107, is probably
interacting with the epoxide ring. In haloalkane dehalogenase,
Trp125 is involved in halogen and halide binding (54). The
eukaryotic epoxide hydrolases all have a tryptophan at this position,
but a positively charged edge of phenylalanine is also capable of binding the electronegative oxygen atom of the epoxide ring (55, 56). A
phenylalanine residue next to the nucleophilic serine was also
found in 2hydroxy-6-oxo-2,4-heptadienoate hydrolase and two
2-hydroxymuconic semialdehyde hydrolases that had sequence similarity with EchA.
Beetham et al. (11) postulated that a proton donating group (K-H) should be present in the cap domain of soluble epoxide hydrolase and protonates the leaving group, an oxyanion that is formed upon opening of the epoxide ring. They mentioned Lys406 as a possible candidate, since this residue is conserved among soluble epoxide hydrolases and since chemical modification resulted in inactivation of the enzyme. An alignment of EchA with all soluble epoxide hydrolases indicates that this is the only lysine that is conserved in the cap domain (data not shown). The alignment points out Lys173 of EchA as a possible candidate, but also the two nearby positioned Lys residues 174 and 177 could perform the role as proton donor. The role of Lys173, Lys174, and Lys177 is currently studied further by site-directed mutagenesis and fluorescence spectroscopy.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EBI Data Bank with accession number(s) Y12804[GenBank].