PEDS Advance Access published online on August 10, 2007
Protein Engineering Design and Selection, doi:10.1093/protein/gzm032
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Influence of different carboxy-terminal mutations on the substrate-, reaction- and enantiospecificity of the arylacetonitrilase from Pseudomonas fluorescens EBC191
1 Institut für Mikrobiologie 2 Institut für Industrielle Genetik, Universität Stuttgart, Allmandring 31, 70569 Stuttgart, Germany
4 To whom correspondence should be addressed. E-mail: andreas.stolz{at}imb.uni-stuttgart.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Different members of the nitrilase superfamily (D-carbamoylases, Nit-Fhit proteins, amidases, cyanide dihydratases and nitrilases) were compared by multiple sequence alignments and a long carboxy-terminal extension (about 50 amino acids) identified in all nitrilases and cyanide dihydratases which was not present in other members of the nitrilase superfamily. The function of this C-terminal part was experimentally analysed in the arylacetonitrilase of Pseudomonas fluorescens EBC191 by the construction of various deletion mutants, chimeric enzymes with other bacterial nitrilases and site-specific mutagenesis. The enzyme variants were tested with the substrates 2-phenylpropionitrile and mandelonitrile and compared regarding specific activities, degree of amide formation and enantioselectivity. The enzyme variants containing deletions up to 32 amino acids did not show significant differences in comparison with the wild-type enzyme. Deletion mutants with 47–67 amino acids missing generally demonstrated reduced enzyme activities, increased amounts of amide formation and increased proportions of the (R)-enantiomers of the amides and acids formed. Also certain exchanges of H296 in the C-terminal motif DpvGHY led to enzyme variants with a similar phenotype. Chimeric enzymes which contained up to 59 amino acids deriving from the nitrilases of Rhodococcus rhodochrous NCIMB11216 or Alcaligenes faecalis ATCC8750 were active and resembled, with respect to the enantioselectivity and degree of amide formation, the wild-type enzyme of P.fluorescens.
Keywords: enzyme catalysis/mandelic acid/nitrilase
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Nitrilases are a commercially very interesting group of enzymes which hydrolyse organic nitriles to carboxylic acids and ammonia. During the last years, several biotransformations have been described which utilise the chemo-, regio- or enantioselectivity of nitrilases (Matthew et al., 1988; Kobayashi and Shimizu, 1994
; Bunch, 1998; Gavagan et al., 1998; Martinková and K
en, 2002; Schulze, 2002). The ability of nitrilases to perform an enantioselective hydrolysis of 
-substituted racemic nitriles to optical active carboxylic acids has been described in some publications and various patent applications and may be useful for the production of a wide range of products, such as -amino-, -hydroxy- or -methylcarboxylic acids. From the scientific literature, it appears that nitrilases show with most substrates only a low degree of enantioselectivity and that industrial relevant enantioselectivities have mainly been observed during the conversion of (substituted) mandelonitrile(s) (MNs) (Yamamoto et al., 1991, 1992; Layh et al., 1992; Gradley et al., 1994).
In the previous work from our group, several bacterial strains were described which hydrolyse a wide range of different nitriles (Layh et al., 1992, 1997). One of these strains, Pseudomonas fluorescens EBC191, produced a nitrilase that converted different arylacetonitriles (Layh et al., 1992). Recently, the encoding gene was cloned and sequenced and the enzyme heterologously expressed in Escherichia coli. The recombinant enzyme converted various phenylacetonitriles [e.g. 2-phenylpropionitrile (2-PPN), MN, phenylglycinonitrile or O-acetoxymandelonitrile] and also aliphatic 2-acetoxynitriles with moderate enantioselectivities to the corresponding -substituted carboxylic acids. Furthermore, with some substrates also significant amounts of the corresponding amides were formed (Heinemann et al., 2003; Kiziak et al., 2005; Fernandes et al., 2006; Mateo et al., 2006; Rustler et al., 2007).
The directed modification of nitrilases by molecular techniques is still severely hampered because currently no crystal structure of a nitrilase is available. Therefore, it was attempted in the present study by sequence comparisons to make some structural predictions and to identify regions within the proteins that are specific for nitrilases and thus might be interesting targets for the modification of nitrilases. In the present study, we describe the influence of various modifications in the carboxy-terminal part of the nitrilase from P.fluorescens EBC191 on the reaction rates, enantioselectivity and amide formation.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Bacterial strains and culture conditions
Plasmid pIK9, which encodes for a His-tagged variant of the nitrilase from P.fluorescens EBC191, was used for the construction of all enzyme variants (Kiziak et al., 2005). All cloning experiments and plasmid preparations were carried out in E.coli JM109 (Vieira and Messing, 1982). Escherichia coli strains were grown at 37°C in 2 x YT liquid medium or on 2 x YT agar plates (Sambrook et al., 1989) supplemented with 100 µg/ml ampicillin. The induction of the nitrilase was achieved by the addition of rhamnose (0.1% w/v) to the growth medium (Kiziak et al., 2005). The cells were harvested by centrifugation in the late exponential or early stationary growth phase.
DNA preparation, DNA manipulation and transformation
The GFX Micro plasmid kit (Amersham Pharmacia, Piscataway, NJ, USA) was used for the preparation of plasmid DNA. The isolation of genomic DNA and all DNA manipulations were carried out as described by Sambrook et al. (1989). Cells of E.coli were transformed with plasmid DNA by the method of Chung et al. (1989).
The standard reaction mixtures contained in a total volume of 40 µl:2.5 µM of the oligonucleotide primers, 0.23 mM each of the four dNTPs, 10–100 ng DNA, DNA polymerase (Pwo polymerase from Roche or platinum Pfx polymerase from Gibco) in the reaction buffer supplied with the DNA-polymerase. The reaction mixtures were prepared on ice. The reactions were started by the addition of the DNA-polymerase (0.75–1.2 U) followed by a short centrifugation step. The reaction mixtures were subsequently transferred to the thermocycler which was previously cooled to 4°C and the following polymerase chain reaction (PCR) program was immediately started: an initial denaturation (92°C, 135 s) was followed by 25 cycles consisting of annealing at 55°C (30 s), a polymerisation step (68°C, 30–60 s, usually 45 s) and denaturation (92°C, 15 s). The final polymerisation step was extended to 300 s. If this programme did not result in the amplification of sufficient amounts of DNA, the PCR reactions were repeated using a decreased annealing temperature (50°C) and 30 reaction cycles.
Construction of mutants and chimeric enzymes
The deletion mutants of the nitrilase from P.fluorescens EBC191 were constructed using appropriate PCR-primers (Table I) which resulted in the amplification of C-terminally truncated forms of the nitrilase. In order to obtain different deletion mutants, the 5'-primer S3183_2 was combined with the 3'-primers Del-C20-rev–Del-C75-rev (Table I). This allowed after restriction of the amplified DNA-fragments the cloning of the gene fragments into plasmid pJOE2702 (Stumpp et al., 2000).
|
The complete nitrilase gene from Rhodococcus rhodochrous NCIMB 11216 was amplified using the sequence provided by the NCBI database under the accession number CQ874192 [GenBank] .
The nitrilase gene from Alcaligenes faecalis ATCC 8750 was amplified using primers derived from the sequence of the nitrilase gene of A.faecalis JM3 (Kobayashi et al., 1993). The nitrilase gene of A.faecalis ATCC 8750 was subsequently sequenced and shown to be identical to the nitrilase gene from A.faecalis JM3.
The chimeric enzymes were constructed via gene splicing by overlap extension (SOE) (Horton et al., 1989). In the first step of the SOE reactions for the construction of the chimeric enzymes Nit(pa-C47), Nit(pa-C59), Nit(pa-C79), Nit(pr-C59) and Nit(pr-C80), the ‘inner primers’ pa-C47-fwd and pa-C47-rev, pa-C59-fwd and pa-C59-rev, pa-C79-fwd and pa-C79-rev, pr-C59-fwd and pr-C59-re, and pr-C80-fwd and pr-C80-rev, respectively, were used (Table I). These primers were used first separately in two different PCR experiments together with the ‘general flanking 5'- and 3'-primers’ S3183_2 and S2949_2-rev (Table I) for the amplification of the relevant parts of the nitrilases from P.fluorescens, R.rhodochrous or A.faecalis. The two PCR experiments were performed as described earlier and the resulting DNA fragments purified by agarose gel electrophoresis and elution of the DNA-fragments from the gels (Easy Pure Kit, Biozym, Hessisch-Oldendorf, Germany). The two DNA-fragments were joined in a final PCR experiment using the ‘general flanking 5'- and 3'-primers’ S3183_2 and S2949_2-rev. The resulting amplified DNA-fragments with the expected sizes of the chimeric genes were cut with NdeI and BamHI and inserted into NdeI-BamHI cleaved pJOE2702 (Table II).
|
The amino acid exchanges in position 296 of the nitrilase were also introduced by SOE using the ‘general flanking 5'- and 3'-primers’ S3183_2 and S2949_2-rev and the ‘specific’ primers H296A-fwd and H296A-rev, H296F-fwd and H296F-rev, H296K-fwd and H296K-rev, H296R-fwd and H296R-rev, respectively (Table I). The abbreviations and characteristics of all constructed plasmids are given in Table II.
DNA sequences were determined by the dideoxy chain termination method using an automated DNA sequencer (ALF-Sequencer, Pharmacia Biotech). Database searches were performed online with the programs blastx, blastp and blastn provided by the BLAST E-mail server (Altschul et al., 1990; Gish and States, 1993). CLUSTALX (version 1.83) (Thomson et al., 1994) was used to align the amino acid sequences. All parameters were set at their default values. The sequence alignments were edited and analysed using the multiple sequence alignment editor and shading utility GeneDoc (version 2.7.000) (Nicholas and Nicholas, 1996).
The nitrilase activity of resting cells was routinely determined in reaction mixtures (1 ml each) containing 50 µmol Tris–HCl buffer (pH 7.5), 10 µmol nitrile and an appropriate amount of cells. All nitrile stock solutions (200 mM each) were prepared in methanol. The reaction mixtures were incubated at 30°C. After different time intervals, samples (200 µl each) were taken and the reactions were stopped by the addition of 1 M HCl (20 µl). The samples were centrifuged at 15 000g for 10 min and the supernatants were analysed using high-pressure liquid chromatography (HPLC).
Protein concentrations were determined using the Bio-Rad protein assay (Bio-Rad Laboratories, München, Germany).
The proteins were separated on a 9% (w/v) denaturing SDS–polyacrylamide gel using the discontinuous buffer system of Laemmli (1970) and stained with Coomassie Blue R250.
2-PPN, MN and their corresponding amides and acids were analysed by HPLC. A S1121 solvent delivery system equipped with a S3205 UV–VIS detector from Sykam GmbH (Gilching, Germany) was used. For the achiral analysis, a reversed-phase column [250 by 4 mm (internal diameter); GROM, Herrenberg, Germany] filled with 5 µm diameter particles of Lichrospher RP18 (E. Merck AG, Darmstadt, Germany) was used to identify individual compounds which were detected spectrophotometrically at 210 nm. Methanol (50% v/v) and H3PO4 (0.1% v/v) in H2O was used as mobile phase to analyse the conversion of the nitriles.
Separation of the enantiomers of mandelic acid, 2-phenylpropionic acid (2-PPA) and the corresponding amides was achieved on a Chiral-HSA column (ChromTech AB, Hägersten, Sweden). The mobile phases consisted of sodium phosphate buffer (100 mM, pH 7.0) containing 4.5% (v/v) acetonitrile or sodium phosphate buffer (10 mM, pH 6.0) plus 0.5% (v/v) 2-propanol, respectively.
All chemicals were obtained from Sigma-Aldrich Chemie GmbH (Deisenhofen, Germany) or E. Merck AG (Darmstadt, Germany). The pure enantiomers of MN were prepared as previously described (Rustler et al., 2007). The biochemicals were supplied from Roche Diagnostics GmbH (Mannheim, Germany). The enzymes for DNA-manipulation were purchased from Roche Diagnostics GmbH and NEB and used according to the manufacturers suggestions.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Sequence alignments of different nitrilases, D-carbamoylases and a NitFhit protein
Nitrilases have been grouped together with various other enzymes and proteins of unknown function, such as cyanide (di)hydratases, D-carbamoylases (=N-carbamyl-D-amino acid hydrolases), certain amidases or the NitFhit proteins, as members of the nitrilase superfamily (Pace and Brenner, 2001; Brenner, 2002). There are currently no crystal structures available for nitrilases, but for other members of the nitrilase superfamily (D-carbamoylases from Agrobacterium sp. KNK712 and A.radiobacter, a NitFhit protein from Caenorhabditis elegans and a homologous protein of unknown function from Saccharomyces cerevisiae (Nakai et al., 2000; Pace et al., 2000; Wang et al., 2001; PDB 1erz, PDB 1F89). Furthermore, a catalytical triad consisting of cysteine, lysine and glutamate residues has been identified in the D-carbamoylase from Agrobacterium sp. KNK712 (Nakai et al., 2000). Therefore, sequence alignments were performed with various nitrilases of bacterial and plant origin, D-carbamoylases and a NitFhit protein in order to identify conserved structural elements and the catalytical centre of the nitrilase from P. fluorescens EBC191. Initial sequence alignments using the program Clustal X showed that the degree of sequence identity was >45% among the bacterial nitrilases. In contrast, only 12–14% sequence identity was observed between the nitrilase and the D-carbamoylase or the ‘Nit-domain’ of the NitFhit protein, respectively. The alignments of these rather divergent proteins were further refined by separate alignments of conserved sequences among the different enzyme families (Fig. 1).
|
Prediction of specific structural motifs and the catalytical triad in the nitrilase from P.fluorescens EBC191
The amino acid alignment was verified by structural predictions using the program Predict Protein (EMBL-EBI, http://www.embl-heidelberg.de/predictprotein). This demonstrated that the program was able to predict the experimentally determined structural features of the D-carbamoylase and the NitFhit protein (such as -helices and ß-folds) from the amino acid sequences. Furthermore, the program also predicted the same secondary structures at similar positions in the nitrilase of P.fluorescens EBC191 (Fig. 1). This suggested that (the) nitrilase(s) possess a similar general protein-fold (‘ßß-sandwich’) as D-carbamyolases and NitFhit proteins. Nevertheless, several differences were observed. Thus, two structures were identified (‘loop A and B’) which were found in the reference proteins but not in nitrilases of plant or bacterial origin. In addition, three sequence motifs (‘loops 1,2,3’) were identified which were present in bacterial or plant nitrilases but not in the reference proteins. The most pronounced differences between nitrilases and the reference proteins were found in the C-terminal regions of the proteins, which began after the fifth helix in the Nit-domain of the NitFhit protein and the nitrilase or after the seventh helix in the D-carbamoylase. These parts were rather small in the NitFhit protein and the D-carbamoylase (19 and 21 amino acids, respectively), but rather large (about 40 amino acids longer) in the nitrilases.
Importance of the C-terminal region of the nitrilase on the catalytical activity, enantioselectivity and amide formation
The alignments suggested that the nitrilases contained in contrast to the other members of the nitrilase superfamily a C-terminal extension behind helix 5. This structure was in the nitrilase from P.fluorescens about 60 amino acids large and contained according to the structure prediction two short ß-folds, which were followed by a short -helix. The alignment of the C-terminal parts of different bacterial and plant nitrilases demonstrated the conserved motif D-(P/S/F)-X-G-H-Y (with X being almost always a small amino acid). This motif could not be detected in any other member of the nitrilase superfamily and therefore suggested some functional significance. It was, therefore, decided to experimentally analyse the influence of the C-terminal part of the nitrilase from P.fluorescens EBC191 on the conversion of the chiral model substrates 2-PPN and MN by site-specific mutagenesis, forming deletion mutants and formation of chimeric enzymes.
Conversion of 2-phenylpropionitrile by the wild-type nitrilase from P.fluorescens expressed in E.coli JM109(pIK9)
It was previously found that the nitrilase from P.fluorescens EBC191 converted several -substituted phenylacetonitriles with moderate enantioselectivities to the corresponding carboxylic acids and that with certain substrates also considerable amounts of the corresponding amides were formed (Layh et al., 1992; Kiziak et al., 2005). In these experiments, the substrate conversions and enantioselectivities of the reactions were usually compared after about 30% substrate turnover. In order to allow a more detailed comparison of the reactions catalysed by the wild-type enzyme and mutant variants of the enzyme, the time-courses for the turnover of 2-PPN and MN by E.coli JM109(pIK9) were analysed which heterologously expressed high quantities of the nitrilase from P.fluorescens EBC191 (Kiziak et al., 2005).
Resting cells of E.coli JM109(pIK9) converted 2-PPN with a specific activity of about 1.2 U/mg of protein. Initially, an almost linear increase in the concentrations of (R)- and (S)-2-PPA were observed and (S)-2-PPA was formed at 30% substrate turnover with a slight enantiomeric excess (ee) of about 65% (Fig. 2A). In addition to the acid, also some 2-PPA amide (2-PPAA) was formed (about 0.2% of the amount of the 2-PPA formed). The increase in the concentration of the amide during the initial phase of the reactions was almost linear and (R)-2-PPAA was preferentially formed (Fig. 2B).
|
Conversion of mandelonitrile by E.coli JM109(pIK9)
The enzymatic conversion of MN differed from the reaction observed with 2-PPN, because MN exists at pH 7 under the reaction conditions in a reversible equilibrium with benzaldehyde and HCN. This allows a dynamic enzymatic resolution of MN, because the chemical reaction always generates racemic mixtures of MN (even if only one enantiomer of MN is initially present). Enantioselective nitrilases can therefore transform a racemic mixture of MN to stoichiometric amounts of (R)-mandelic acid (Yamamoto et al., 1991).
For a valid comparison of the enantioselectivities of different generated nitrilase mutants, it was necessary to ensure that always both enantiomers of MN were present during the biotransformation process. Therefore, the cell densities of the recombinant E.coli strains had to be carefully adjusted so that the product formation was always 
0.1 mM/min. It was assumed that under these conditions during the whole experiment MN was available for the cells as an almost racemic mixture. The correctness of this assumption was verified in the biotransformation of (R,S)-MN by E.coli JM109(pIK9) producing the wild-type nitrilase from P.fluorescens. The nitrilase hydrolysed (R,S)-MN about eight times quicker than 2-PPN and demonstrated a slight preference for the formation of (R)-mandelic acid. Furthermore, significant amounts (about 20 %) of mandelic acid amide (MAA) were formed. The cells formed preferentially (S)-MAA. In the time-course of this reaction, the ees of mandelic acid (MA) and MAA did not change during the whole reaction (Fig. 3). This could only occur if the nitrilase was supplied during the whole reaction with almost racemic MN.
|
Deletion of various parts from the C-terminal region of the nitrilase from P.fluorescens EBC191
Deletion mutants were constructed by amplification of the nitrilase gene from plasmid pIK9 using PCR primers which introduced stop codons at different positions at the 3' end of the nitrilase gene. This resulted in the formation of plasmids pCK101–pCK107 (Tables I and II). Thus, deletion mutants were obtained, which were shortened at the carboxy-terminus for 20, 32, 47, 55, 60, 67 or 75 amino acids [=Nit(del-C-20)–Nit(del-C-75)]. From the structural predictions, it was expected that the deletions in the mutants Nit(del-C-67) and Nit(del-C-75) extended into the structural conserved helix E.
The individual recombinant E.coli strains were cultivated, the nitrilase-variants induced by the addition of rhamnose and cell extracts prepared. The cell extracts were analysed by SDS–gel electrophoresis and the intensity of the nitrilase bands compared. Thus, in E.coli JM109(pIK9), which expressed the wild-type nitrilase, the formation of a very pronounced nitrilase band was observed which was estimated to contain about 20% of the total soluble protein (Kiziak et al., 2005). In the cell extracts of the recombinant strains E.coli JM109 (pCK101–pCK106) synthesizing Nit(del-C-20)–Nit(del-C-67), the respective protein bands demonstrated as expected decreasing sizes. The band intensities were estimated to be at least 50–70% of the band intensities caused by the wild-type nitrilase. It was therefore deduced that the truncated enzyme variants were presumably expressed in similar amounts as the wild-type enzyme, because the smaller proteins should result in a less pronounced staining of the truncated proteins.
Resting cells of E.coli JM109(pCK101–pCK106) which produced the deletion mutants Nit(del-C-20)–Nit(del-C-67) showed nitrilase activity with 2-PPN and only the shortest enzyme variant [Nit(del-C-75)] from E.coli JM109(pCK107) was enzymatically inactive. Those cells which formed the active enzyme variants Nit(del-C-20) and Nit(del-C-32) strongly resembled the wild-type enzyme. In contrast, the variants Nit(del-C-47)–Nit(del-C-67) demonstrated reduced nitrilase activities towards 2-PPN and also a reduced degree of enantioselectivity for the formation of (S)-2-PPA (Table III). Surprisingly, these mutants also showed an increased formation of amide in relation to the acid formation (up to almost 10% compared to 0.2% with the wild-type). Furthermore, with these variant enzymes also increased ee-values for the formation of (R)-2-PPAA (ee > 75% at 30% nitrile conversion) were found—compared with an ee-value of only 7% observed with the wild-type enzyme (Table III).
|
With racemic MN as substrate, it was found that the variants Nit(del-C-47)–Nit(del-C-67) demonstrated an increased enantioselectivity for the formation of (R)-mandelic acid (Table III). The deletion mutants converted also MN (as observed with 2-PPN) to increased amounts of the corresponding amide. The mutants Nit(del-C-60) and Nit(del-C-67) even produced more amide than acid as product and acted, therefore, formally more as nitrile hydratases than as nitrilases. Furthermore, the deletion variants showed in relation to the wild-type a decreased ee for the formation of (S)-mandeloamide (Table III).
Decreased stability of the deletion mutants during repeated freeze-thawing cycles
It was assumed that the extended C-terminus of the nitrilases could have a function in the formation or stabilization of the active (presumably multimeric spiral-shaped) enzymes as previously suggested for cyanide dihydratases (Sewell et al., 2003, 2005). Therefore, the stability of the truncated nitrilase-variants in the recombinant E.coli was compared during up to six freeze-thaw cycles. Thus, it was observed that the strains producing the shortest enzyme variants [=Nit(del-C-47)–Nit(del-C-67)] showed a significant decreased stability of the nitrilase activity compared with the wild-type enzyme (Fig. 4). In contrast, no significant differences in the stability of the nitrilase activities were observed when the cells were incubated for 72 h at 0°C.
|
Conversion of the pure enantiomers of mandelonitrile by the deletion mutant Nit(del-C-60)
It was previously demonstrated for the conversion of (R)- and (S)-MN that the individual enantiomers of MN were converted by the nitrilase from P.fluorescens EBC191 only to the corresponding enantiomers of the amide and acid (Mateo et al., 2006). Thus, (S)-MN was only converted to (S)-mandeloamide and (S)-mandelic acid. Furthermore, it was also shown that (S)-MN was converted to approximately equimolar amounts of (S)-mandelic acid and (S)-mandeloamide, whereas (R)-MN was converted to the (R)-acid and the (R)-amide in a proportion of about 8:1 (Mateo et al., 2006; Rustler et al., 2007). It was therefore analysed if this was also valid for the truncated enzyme variants. The recombinant E.coli strains synthesising either the wild-type enzyme or the Nit(del-C-60) mutant were incubated with 5 or 10 mM of either the (S)- or the (R)-enantiomer of MN. These reactions were performed in Na-citrate buffer pH 5.0 in order to reduce the spontaneous racemisation of MN observed at higher pH values (Rustler et al., 2007). Thus, it was found that also the truncated enzyme variant converted (S)-MN to much higher concentrations of MAA than the respective (R)-enantiomer (Fig. 5). Furthermore, it was found that the mutant form of the enzyme formed with both (S)- and (R)-MN a higher degree of amide than the wild-type enzyme. This resulted for the turnover of (S)-MN by the mutant form in the production of more than 80% (S)-MAA in comparison with the formation of the (S)-acid (Fig. 5).
|
Chimeric enzyme variants
The experiments with the deletion mutants demonstrated that the C-terminal part of the nitrilases possessed some influence on stability, catalytical activity, amide formation and enantioselectivity of the enzyme. It was therefore decided to produce chimeric enzymes between the investigated nitrilase from P.fluorescens EBC191 (NitP) and some other nitrilases with different properties. For these experiments, the benzonitrile-converting ‘aromatic’ nitrilase from R.rhodochrous NCIMB 11216 (NitR) (Harper, 1977) and the arylacetonitrilase from A.faecalis ATCC 8750 (NitA) were chosen. The nitrilase from A.faecalis ATCC 8750 had been previously shown to convert MN with a rather high enantioselectivity into (R)-mandelic acid (Yamamoto et al., 1992). Chimeric nitrilases were produced using the technique of ‘SOE’. The chimeric nitrilases obtained were abbreviated according to the origin of the globular (first letter) and C-terminal part (second letter) of the chimeric enzyme and the number of the amino acid (counting from the C-terminus) at which the fusion was performed. Thus, the chimeric variants Nit(pa-C47), Nit(pa-C59), Nit(pa-C79), Nit(pr-C59) and Nit(pr-C80) were obtained.
In contrast to the situation with the deletion mutants, all chimeric nitrilases were active (Table I). The enzymatic activities of the chimeric enzyme with 2-PPN were 
29–85% of the activity found in the wild-type nitrilase of P.fluorescens EBC191. Thus, the examples of Nit(pa-C79) and Nit(pr-C80) demonstrated that the C-terminal parts of different nitrilases could replace each other. The chimeric enzymes formed (as the unmodified nitrilase from P.fluorescens and the C-terminally truncated variants) preferentially (S)-2-PPA from racemic 2-PPN. Also the chimeric nitrilases with a large C-terminal part deriving from A.faecalis showed a pronounced (S)-selectivity for the formation of (S)-2-PPA, although the nitrilase from A. faecalis demonstrated a preference for the formation of the (R)-acid (58% ee at 30% conversion). This demonstrated that the globular part of the nitrilase mainly determined the degree of enantioselectivity exerted by the enzyme.
The chimeric enzymes formed slightly increased amounts of 2-PPAA (in relationship to the amounts of 2-PPA formed) from 2-PPN. The chimeric enymes formed compared with the wild-type enzyme 2-PPAA with an increased (R)-selectivity as observed before for the truncated enzyme variants (Table I).
The chimeric enzymes showed with MN as substrate only slight deviations from the wild-type enzyme from P.fluorescens EBC191. The most significant change was an increased formation of MAA by the mutants Nit(pa-C79) and Nit(pr-C80) which contained the largest changes in the amino acid sequence.
In contrast to the deletion mutants, almost all [with the exception of Nit(pa-C47)] chimeric nitrilases were stable during repeated freeze-thawing cycles and resembled in this respect the wild-type enzyme from P.fluorescens EBC191 (Fig. 4).
Specific exchanges of single amino acids in the C-terminal region of the nitrilase from P.fluorescens EBC191
The sequence comparisons among different nitrilases suggested the conservation of the motif D-(P/S/F)-X-G-H-Y (discussed earlier). Therefore, it was decided to investigate the effects of mutations in this part of the enzyme on the enzyme activity by exchanging the histidine-residue (in the position 296 of the nitrilase) against an alanine, phenylalanine, lysine or arginine residue. This resulted in the construction of plasmids pCK131, pCK132, pCK133 and pCK134 (Table I). The four mutants showed significant decreases in their enzymatic activities (14–48% of the activities of the wild-type enzyme) (Table III). The enantioselectivities of mutant enzymes Nit(H296A) and Nit(H296F) for the formation of (S)-2-PPA were only slightly decreased (about 50% ee), whereas Nit(H296K) and Nit(H296R) showed only low preference for the formation of (S)-2-PPA (about 15% ee). All four mutants excreted increased amounts of 2-PPAA (0.7–2.4%).
All four mutants with amino acid exchanges at the position H296 showed with MN as substrate increased ee-values for the formation of (R)-MA and also formed increased amounts of mandeloamide. The mutant enzyme Nit(H296F) formed almost equimolar amounts of MAA and MA from MN. Furthermore, all four mutants were more sensitive against freeze-thaw cycles than the wild-type enzyme (Fig. 4).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
The C-terminally truncated nitrilase variants demonstrated that deletions up to about 30 amino acids did not significantly change the relevant characteristics of the enzyme. In contrast, longer C-terminal deletions of 47–67 amino acids resulted in significant decreases in enzyme activities and only
10% of the activity was found compared with the wild-type enzyme. Further, more extensive deletions completely abolished the enzyme activity. Thus, it appears that a region at the C-terminus between the amino acids 302 and 319 is important for enzyme activity, although these amino acids do not take part in the presumed catalytical triad. This was an interesting observation because it was previously suggested for the cyanide dihydratases from Bacillus pumilus C1 and Pseudomonas stutzeri AK61 (which are cyanide-converting members of the nitrilase enzyme family) that the homologous protein region (‘D-surface’) participates in the formation of the spiral-shaped holoenzymes. Deletions in the C-terminal parts of these enzymes led to inactive enzymes (Sewell et al., 2005). Similarly, it was shown for the D-carbamoylase from Agrobacterium sp. KNK712 that the C-terminus is involved in the formation of the substrate-binding pocket (Nakai et al., 2000).
All four deletion mutants with clearly measurable but reduced nitrilase activity [Nit(del-C-47, -55, -60 and -67)] also showed an increased formation of the amides (in relationship to the acid formation). Thus, the relative amounts of MAA increased by about a factor of two and the 2-PPAA formation increased even for a factor of about 35 (from 0.2% to 9.7 %) during the conversion of 2-PPN by the mutant Nit(del-C60). It was previously shown that nitrilases contain a cysteine residue in the active centre and form during catalysis a covalently bound intermediate (Stevenson et al., 1990; Kobayashi et al., 1992). The formation of either the amide or the acid from the tetrahedral intermediate in which the substrate is covalently bound to the catalytically active cysteine residue of the enzyme can be chemically explained via a formal thiol elimination (resulting in amide formation) or the competing release of ammonia (resulting in acid formation). The release of ammonia (and thus the formation of the acid) would chemically require a protonation of the nitrogen atom originating from the nitrile group, in order to obtain a good leaving group. In this model, an increased formation of the amide can be caused either by the structure of the substrate (e.g. by electron-withdrawing substituents which destabilise the positive charge on the nitrogen group) or by sterical factors which hamper the transfer or the stabilisation of the positive charge on the amino group (Fernandes et al., 2006). The effect of (increased) electronegativities of substituents at the -position of the substrates for the increased formation of amides is substantiated by the data obtained during the present study because it was found that all generated mutants, such as the wild-type enzyme, formed with MN as substrate higher amounts of amides than with 2-PPN. In addition, the constructed mutant variants of the P.fluorescens nitrilase demonstrated for the first time that it is possible to increase the relative amounts of amides formed by changing the enzyme structure.
The increase in the rates of amide formation found for the deletion mutants Nit(del-C-47)–Nit(del-C-75) correlated with changes in the enantioselectivities. In the case of the conversion of MN, the ee for the production of (R)-MA was much higher for the deletion mutants in comparison with the wild-type enzyme. This was accompanied by a higher rate of amide formation, which was preferably liberated as (S)-MAA. The total amount of the (S)-enantiomers formed was calculated as sum of the concentrations of the (S)-acid and (S)-amide. The combination of the decreased acid and increased amide formation leads to an almost complete loss of the enantioselectivty for the conversion of MN for the deletion mutants.
The same kind of calculation demonstrated for 2-PPN that the wild-type nitrilase from P.fluorescens preferentially accepted (S)-2-PPN. In contrast, it was found that the C-terminal deletion mutants also showed almost no enantiomeric preference with this substrate. This indicated that in contrast to the wild-type enzyme both enantiomers of 2-PPN were bound with about the same affinity to the active site in mutants lacking the C-terminal part. Therefore, it appears that this part of the enzyme is involved in the discrimination of the two nitrile enantiomers.
The comparison of the deletion mutants with the chimeric enzymes demonstrated that chimeric enzymes which contained 59 amino acids from the nitrilases of R.rhodochrous or A.faecalis (and amino acids 1–298 from P.fluorescens) resembled the wild-type enzyme according to the relative amounts of amides formed and the enantioselectivity (for the formation of amide plus acid) during the turnover of MN. In contrast, it was found that the chimeric enzymes which contained 79 or 80 amino acids from the homologous enzymes from R.rhodochrous or A.faecalis (and amino acids 1–273 from P.fluorescens) behaved more similar to the deletion mutants. The relevant part of the Pseudomonas nitrilase between amino acids 273–298 contained the motif D-(P/S)-X-G-H-Y that was specifically conserved among bacterial nitrilases. The analysis of the mutations in the relevant histidine residue (His296) suggested that two different types could be distinguished, the mutants which carried an exchange of the histidine residue against the basic residues arginine or lysine resembled the deletion mutants, because they demonstrated significantly decreased enantioselectivities for the formation of (S)-PPA and an increased formation of the amide. The exchange of the histidine against an alanine and/or phenylalanine residue also resulted in an increased amide formation but without significant changes in the formation of (S)-PPA.
These results suggested that the C-terminal part of the nitrilases seems to fulfil two basic functions. The comparison of the wild-type enzyme, the deletion mutants and the chimeric enzymes demonstrated that a missing C-terminus leads to a decreased enzyme activity and stability, the loss of the preferred binding of one of the nitrile enantiomers and an increased degree of amide formation. A possible explanation for the reduced activity and stability of the mutants might be that the C-terminus is involved in the interaction of the individual subunits of the enzyme in such a way that the C-terminal part of one subunit covers the active site of the neighbouring monomer. It is possible that the truncated modified nitrilases, which show only about 10% enzyme activity and a highly increased instability, have lost their ability to form the more active spiral-shaped holoenzymes. The open conformation of the monomeric enzyme without coverage by the C-terminal part would also explain the loss of enantioselectivity during nitrile conversion.
Furthermore, the mutations of the His296 residue suggested that the C-terminal part of the enzyme is not only involved in the overall multimeric structure of the enzyme, but also that specific side chains are directly involved in the catalytical properties of the nitrilase. This might explain why the introduction of basic residues at position 296 had a significantly stronger effect on the enzyme activity than that of neutral amino acid residues.
|
Footnotes |
|---|
3 Present address: Lonza AG, Biotechnology Research and Development, CH-3930 Visp, Switzerland
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionReferences |
|---|
Altschul S.F., Gish W., Miller W., Myers E.W., Lipmann D.J. J. Mol. Biol. (1990) 215:403–410.[CrossRef][Web of Science][Medline]
Bradford M.M. Anal. Biochem. (1976) 72:248–254.[CrossRef][Web of Science][Medline]
Brenner C. Curr. Opin. Struct. Biol. (2002) 12:775–782.[CrossRef][Web of Science][Medline]
Bunch A.W. Biotechnology Vol. 8a Biotransformations I.—Rehm H.J., Reed G., eds. (1998) Weinheim: Wiley-VCH. 277–324.
Chung C.T., Niemela S.L., Miller R.H. Proc. Natl Acad. Sci. USA (1989) 86:2172–2175.
Fernandes B.C.M., Mateo C., Kiziak C., Wacker J., Chmura A., van Rantwijk F., Stolz A., Sheldon R.A. Adv. Synth. Catal. (2006) 348:2597–2603.[CrossRef]
Gavagan J.E., Fager S.K., Fallon R.D., Folsom P.W., Herkes F.E., Eisenberg A., Hann E.C., DiCosimo R. J. Org. Chem. (1998) 63:4792–4801.[CrossRef][Web of Science]
Gish W., States D.J. Nat. Genet. (1993) 3:266–272.[CrossRef][Web of Science][Medline]
Gradley M.L., Deverson C.J.F., Knowles C.J. Arch. Microbiol. (1994) 161:249–251.
Harper D.B. Biochem. J. (1977) 165:309–319.[Web of Science][Medline]
Heinemann U., Kiziak C., Zibek S., Layh N., Schmidt M., Griengl H., Stolz A. Appl. Microbiol. Biotechnol. (2003) 63:274–281.[CrossRef][Web of Science][Medline]
Horton R.M., Hunt H.D., Ho S.N., Pullen J.K., Pease L.R. Gene (1989) 77:61–68.[CrossRef][Web of Science][Medline]
Kiziak C., Conradt D., Stolz A., Mattes R., Klein J. Microbiology (2005) 151:3639–3648.
Kobayashi M., Izui H., Nagasawa T., Yamada H. Proc. Natl Acad. USA (1993) 90:247–251.
Kobayashi M., Komeda H., Yanaka N., Nagasawa T., Yamada H. J. Biol. Chem. (1992) 267:20746–20751.
Kobayashi M., Shimizu S. FEMS Microbiol. Lett. (1994) 120:217–224.[CrossRef][Web of Science]
Laemmli U.K. Nature (1970) 227:680–685.[CrossRef][Medline]
Layh N., Hirrlinger B., Stolz A., Knackmuss H.-J. Appl. Microbiol. Biotechnol. (1997) 47:668–674.[CrossRef][Web of Science]
Layh N., Stolz A., Förster S., Effenberger F., Knackmuss H.-J. Arch. Microbiol. (1992) 158:405–411.[Web of Science]
Martinková L., Ken V. Biocatal. Biotrans. (2002) 20:79–93.
Mateo C., Chmura A., Rustler S., van Rantwijk F., Stolz A., Sheldon R.A. Tetrahedron Asymmetry (2006) 17:320–323.[CrossRef][Web of Science]
Mathew C.D., Nagasawa T., Kobayashi M., Yamada H. Appl. Environ. Microbiol. (1988) 54:1030–1032.
Nakai T., et al. Structure (2000) 8:729–739.[Medline]
Nicholas K.B., Nicholas H.B. Jr. GeneDoc: a tool for editing and annotating multiple sequence alignments. (1996) Distributed by the author per anonymous ftp.
Pace H.C., Brenner C. Genome Biol. (2001) 2:0001.1–0001.9. [Online].
Pace H.C., Hodawadekar S.C., Draganescu A., Huang J., Bieganowski P., Pekarsky Y., Croce C.M., Brenner C. Curr. Biol. (2000) 10:907–917.[CrossRef][Web of Science][Medline]
Rustler S., Müller A., Windeisen V., Chmura A., Fernandes B., Kiziak C., Stolz A. Enzyme Microb. Technol. (2007) 40:598–606.[CrossRef][Web of Science]
Sambrook J., Fritsch E.F., Maniatis T. Molecular Cloning: A Laboratory Manual. (1989) 2nd edn. Cold Spring Harbour Laboratory Press.
Schulze B. Enzyme Catalysis in Organic Synthesis.—Drauz K., Waldmann H., eds. (2002) Weinheim: Wiley-VCH. 699–715.
Sewell B.T., Thuku R.N., Zhang X., Benedik M.J. Ann. NY Acad. Sci. (2005) 1056:153–159.[CrossRef][Web of Science][Medline]
Sewell B.T., Berman M.N., Meyers P.R., Jandhyala D., Benedik M.J. Structure (2003) 11:1413–1422.[Medline]
Stevenson D.E., Feng R., Storer A.C. FEBS Lett. (1990) 277:112–114.[CrossRef][Web of Science][Medline]
Stumpp T., Wilms B., Altenbuchner J. Biospektrum (2000) 6:33–36.
Thomson J.D., Higgins D.G., Gibson T.J. Nucleic Acids Res. (1994) 22:4673–4680.
Vieira J., Messing J. Gene (1982) 19:259–268.[CrossRef][Web of Science][Medline]
Wang W.-C., Hsu W.-H, Chien F.-T., Chen C.-Y. J. Mol. Biol. (2001) 306:251–261.[CrossRef][Web of Science][Medline]
Yamamoto K., Fujimatsu I., Komatsu K.-I. J. Ferment. Bioeng. (1992) 73:425–430.[CrossRef]
Yamamoto K., Oishi K., Fujimatsu I., Komatsu K.-I. Appl. Environ. Microbiol. (1991) 57:3028–3032.
Received April 4, 2007; revised May 30, 2007; accepted May 31, 2007.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||