PEDS Advance Access originally published online on December 10, 2008
Protein Engineering Design and Selection 2009 22(2):103-110; doi:10.1093/protein/gzn073
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Using directed evolution to probe the substrate specificity of mandelamide hydrolase
College of Pharmacy, University of Michigan, 428 Church St, Ann Arbor, MI 48109, USA
2 To whom correspondence should be addressed. E-mail: mcleish{at}iupui.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
Mandelamide hydrolase (MAH), a member of the amidase signature family, catalyzes the hydrolysis of mandelamide to mandelate and ammonia. X-ray structures of several members of this family, but not that of MAH, have been reported. These reveal nearly superimposable conformations of the unusual Ser-cisSer-Lys catalytic triad. Conversely, the residues involved in substrate recognition are not conserved, implying that the binding pocket could be modified to change the substrate specificity, perhaps by directed evolution. Here we show that MAH is able to hydrolyze small aliphatic substrates such as lactamide, albeit with low efficiency. A selection method to monitor changes in mandelamide/lactamide preference was developed and used to identify several mutations affecting substrate binding. A homology model places some of these mutations close to the catalytic triad, presumably in the MAH active site. In particular, Gly202 appears to control the preference for aromatic substrates as the G202A variant showed three orders of magnitude decrease in kcat/Km for (R)- and (S)-mandelamide. This reduction in activity increased to six orders of magnitude for the G202V variant.
Keywords: enzyme catalysis/homology model/mutagenesis/protein engineering/selection
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
Mandelamide hydrolase (MAH) from Pseudomonas putida (EC 3.5.1.86 [EC] ) catalyzes the conversion of (R,S)-mandelamide to (R,S)-mandelate and ammonia (Fig. 1) (McLeish et al., 2003
; Gopalakrishna et al., 2004). It is a member of a large group of enzymes which has been termed the amidase signature (AS) family (Mayaux et al., 1991; Chebrou et al., 1996). Characterized by the presence of a highly conserved serine- and glycine-rich stretch of approximately 50 amino acids, members of this family have been identified in over 90 different organisms (Shin et al., 2003), including archaea, eubacteria, fungi, nematodes, plants, insects, birds and mammals (Patricelli and Cravatt, 2000). Given that the function of the amidases is to simply hydrolyze an amide bond, the biological consequences of the reaction are many and varied. These include carbon–nitrogen metabolism in prokaryotes and eukaryotes (Gomi et al., 1991; Mazumder et al., 1991), formation of indole-3-acetic acid in plant bacterial pathogens (Gaffney et al., 1990), degradation of neuromodulatory fatty acid amides in mammals (Deutsch and Chin, 1993; Cravatt et al., 1996), metabolism of atrazine in soil bacteria (Shapir et al., 2005) and formation of Gln-tRNAGln by transfer of ammonia from glutamine (Schon et al., 1988; Curnow et al., 1997).
|
X-ray structures are now available for five members of this family: the fatty acid amide hydrolase (FAAH) from rat (Bracey et al., 2002
), malonamidase E2 (MAE2) from Bradyrhizobium japonicum (Shin et al., 2002, Shin et al., 2003), a peptide amidase (PAM) from Stenotrophomonas maltophilia (Labahn et al., 2002) and glutamyl-tRNAGln amidotransferases from Staphylococcus aureus (Nakamura et al., 2006) and Thermotoga maritima (Murzin et al., 1995). The overall folds are all similar and contain a mixed β-sheet comprising 10 strands. Together they form a distinct family in the SCOP (Structural Characterization of Proteins) database (Murzin et al., 1995). Mechanistic studies have shown that the AS family operates by an unusual catalytic mechanism, employing a combination of serine and lysine residues (Patricelli and Cravatt, 2000) which the X-ray structures showed to be an unusual Ser-cisSer-Lys catalytic triad. However, while the positions of the proposed catalytic residues are virtually superimposable, the residues involved in substrate recognition are not conserved, thus emphasizing the functional diversity of this class of enzymes.
In addition to their diverse biological roles, amidases have found a variety of industrial applications, particularly in the synthesis of chiral compounds (Hirrlinger et al., 1996; Wei et al., 2003; Shaw and Naughton, 2004), but also in the degradation of nylon polymers (Negoro, 2000) and even in peptide synthesis (Schwarz et al., 1992; Labahn et al., 2002). The broad range of reactions catalyzed by the amidases show that enzymes of this class are able to adapt their substrate binding pocket according to circumstances, and implies that amidases are likely to be good candidates for directed evolution experiments aimed at enhancing their industrial utility. In preliminary studies, we have shown that MAH is able to hydrolyze lactamide, albeit with low efficiency (Gopalakrishna et al., 2004). This suggested that it may be possible to convert MAH into a lactamide hydrolase (LAH), i.e. the enzyme would go from recognizing a phenyl ring to recognizing a methyl group (Fig. 1). Such an experiment would parallel our studies on the benzoylformate decarboxylase–pyruvate decarboxylase (BFDC–PDC) interconversion (Siegert et al., 2005), and continue our efforts to evolve the mandelamide/mandelate pathway (Petsko et al., 1993; McLeish et al., 2003) into a novel lactamide pathway. Here we report the development of a screen for LAH activity, and describe how the use of random mutagenesis has allowed us to identify residues involved in the binding of MAH substrates.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
Materials
pET17MAH-his, the expression plasmid for the his-tagged wt MAH, was available from the previous study (Gopalakrishna et al., 2004). GeneMorph II Random Mutagenesis Kit and PfuUltra HF DNA polymerase were purchased from Stratagene. All primers were purchased from Integrated DNA Technologies, Inc. The competent cells BL21-Gold(DE3)pLysS and OneShot Mach1-T1 were purchased from Stratagene and Invitrogen, respectively. (R)- and (S)-mandelamide were prepared as described previously (Gopalakrishna et al., 2004). (R)- and (S)-2-aminophenylacetamide were purchased from Sigma and Bachem, respectively. (R)- and (S)-lactamide were purchased from Sigma. Cyclohexanecarboxamide was from Wako Pure Chemicals. All other chemicals and reagents were purchased from Sigma, Alfa Aesar, or TCI and were of the highest purity available.
Construction of MAH mutant library by random mutagenesis
A random mutagenesis library of MAH was constructed in a two-step process. Initially, random mutagenesis on MAH was carried out by error-prone PCR using the GeneMorph II Random Mutagenesis Kit. Primers were designed to flank the MAH fragment and pET17MAH-his was used as the first template. In a typical error-prone PCR, 10 ng template DNA (3.3 ng target) was amplified in a 50-µl reaction containing 0.2 mM each dNTP, 0.9 µM each primer, 2.5 U Mutazyme II polymerase and 5 µl 10x buffer (Stratagene). The reaction mixture was heated at 95°C for 2 min, followed by 30 cycles of incubation at 95°C for 1 min, annealing at 45–55°C for 1 min, and extension at 72°C for 2 min, and a final incubation at 72°C for 10 min. The annealing temperature was increased linearly from 45 to 55°C over the first five cycles. The PCR product containing the randomly mutated MAH fragment was purified with QIAQuick PCR Purification Kit (Qiagen).
Subsequently, these fragments were cloned into the vector pET17b using megaprimer PCR. Once again pET17MAH-his was used as the template and the MAH fragments generated in the first step were now used as megaprimers. The reaction was carried out in a total volume of 50 µl containing 100 ng template DNA, 0.2 mM each dNTP, 400 ng megaprimer, 5 U PfuUltra HF polymerase and 5 µl 10x PfuUltra HF reaction buffer (Stratagene). The reaction mixture was heated at 94°C for 7 min, followed by 20 cycles of incubation at 95°C for 30 s, annealing at 50°C for 1 min, and extension at 68°C for 12 min, and a final incubation at 72°C for 10 min. The PCR product was digested with DpnI restriction enzyme and transformed into competent Escherichia coli cells OneShot Mach1-T1 (Invitrogen), and the cells were plated on LB agar containing 50 µg/ml ampicillin. After overnight growth at 37°C, all colonies on the plate were collected and plasmid DNA was prepared using QIAprep Spin Miniprep Kit (Qiagen). The resulting plasmid DNA was used to transform E.coli strain BL21-Gold(DE3)pLysS for MAH expression.
To screen for LAH activity, cells were grown in minimal media with lactamide as only carbon source. Both liquid and solid (1.5% agar) minimal media were prepared with the following ingredients in 100 ml: 680 mg Na2HPO4, 300 mg KH2PO4, 50 mg NaCl, 24 mg MgSO4, 1.5 mg CaCl2, 0.16 mg FeCl3, 1 mg thiamine, 500 mg NH4Cl, 5 mg ampicillin, 2.5 mg chloramphenicol and variable lactamide concentration. The E.coli cells which had been transformed with the MAH mutant library, were grown overnight at 37°C on LB agar plate containing 50 µg/ml ampicillin and 25 µg/ml chloramphenicol. All colonies on the plate were collected and washed with liquid minimal media, then grown at 30°C in minimal media in the following sequence: (i) in 20 mM lactate for 3 h; (ii) in 5 mM lactate and 20 mM lactamide for 2 h; (iii) in 100 mM lactamide for 2 days; (iv) on minimal plate containing 20 mM lactamide. Colonies were observed after 3–4 days growth on the plate. Individual colonies were selected, and 1-ml scale cultures were grown for each colony in LB media containing ampicillin and chloramphenicol. MAH expression was induced by 0.4 mM IPTG when OD600 reached 0.6. The cells were grown for an additional 4 h prior to harvesting by centrifugation. The cell pellets were washed with 20 mM NaPO4 at pH 7.0 and resuspended in the same buffer with a total volume of 0.2 ml. Cell lysis was achieved by freezing the cells at –20°C followed by thawing at room temperature. DNase I (12 U/ml) and 5 mM MgCl2 were added, followed by incubation at room temperature for 15 min. The cell lysate was centrifuged and the resulting crude cell free extract was used directly in the activity assay described below.
The mutants MAH G202A and MAH G202V were prepared using PfuUltra HF DNA polymerase and QuikChange site-directed mutagenesis protocol (Stratagene) with pET17MAH-his as the template. The forward primers used for the mutagenesis are shown below with the mutated codon underlined and the lowercase letters indicating base change from the wt,
- G202A: 5'-GGGTACCGATACAGcTGGATCTGTTCGAC-3'
- G202V: 5'-GCAGCTTTGGGgACCGATACAGtTGGATCTGTTCG-3'
Preparation of Q207H/S316N/I437N and I437N
A mixture of the plasmids containing MAH Q207H/S316N/Q382E and MAH T31I/I437N was digested with restriction enzymes BspMI and XhoI. This resulted in fragments of MAH containing the mutations T31I, I437N, Q382E or Q207H/S316N. The triple mutant Q207H/S316N/I437N was generated by ligation using the Fast-Link Ligation Kit (Epicenter Biotechnologies) and was identified by DNA sequencing. For I437N, a mixture of plasmids containing wt and T31I/I437N was digested with BspMI and XhoI and religated. Sequencing was used to identify the I437N mutant.
Protein expression and purification
The his-tagged wt MAH was expressed in E.coli strain BL21(DE3)pLysS and purified by affinity chromatography on His-Select Nickel (Sigma) column, followed by size-exclusion chromatography on Sephacryl S-200 (Amersham) column as described previously (Gopalakrishna et al., 2004). All mutants were expressed in either BL21-Gold(DE3)pLysS or BL21(DE3)pLysS, and purified using the method employed for the wt enzyme.
Extinction coefficient determination
The molar extinction coefficient of wt MAH-His was determined using the Edelhoch method as described by Pace et al. (Pace et al., 1995). The UV spectrum was measured for the purified wt MAH-His in 30 mM MOPS buffer at pH 7.0, and in the same buffer containing 6 M GdnHCl at 25°C using a Cary 100 spectrophotometer. The spectra showed a 
max at 278 nm and provided an extinction coefficient of 28 800 M–1cm–1.
MAH activity was measured by a coupled assay as described previously (Gopalakrishna et al., 2004). The reaction of MAH was coupled to the reaction of glutamate dehydrogenase, and was followed by monitoring the decrease in absorbance at 340 nm due to bleaching of NADPH. The velocity of the reaction is calculated using an extinction coefficient for NADPH of 6220 M–1cm–1. All assays were performed at pH 7.8 and 30°C in 100 mM potassium phosphate buffer containing 1 mM EDTA. A typical assay mixture contained 0.8 mM 
-ketoglutarate, 0.25 mM NADPH, 15 U/ml L-glutamate dehydrogenase, appropriate concentration of amide substrate and MAH in a total volume of 1 ml.
For the determination of the steady-state kinetic parameters, the initial velocity data were fitted to the Michaelis–Menten equation using SigmaPlot (SPSS Inc.). The 
278 of 28 800 M–1cm–1, measured as described above, was employed to determine the enzyme concentrations used in these calculations.
The three-dimensional structure of MAH was modeled with Modeller 9v2 (Sánchez and 
ali, 2000). Initially, the three available structures of members of the AS family were tested as potential models. These were PAM (PDB code 1M22), MAE2 (1OCL) and FAAH (1MT5). Three test models were constructed separately using each of these structures as a template. The individual alignments were performed with Modeller, and manually refined to include hydrophobicity and secondary structure prediction data obtained with PSI-PRED (Jones, 1999). The models were evaluated using Verify3D (Luthy et al., 1992) and WHAT IF (Vriend, 1990). According to the quality evaluation of the individual models, MAE2 was chosen as the main template, whereas region 1–62 was modeled based on PAM and region 339–364 was modeled using the FAAH structure. The quality of the final model was evaluated using the same tools as the individual models. Figures were created with SwissPDBViewer (Guex and Peitsch, 1997) (www.expasy.org/spdbv/) and Pov-Ray (Persistence of Vision Raytracer Pty. Ltd., http://povray.org/).
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
MAH substrate specificity
Given that the substrate range of MAH had not been extensively characterized, it seemed appropriate to do this before looking for determinants of substrate specificity. Accordingly, the activity of MAH with a variety (Fig. 2) of aromatic and aliphatic substrates was measured, and the kinetic parameters obtained. These values are shown in Table I.
|
|
From these results, it is clear that MAH prefers aromatic substrates as replacing the phenyl ring by an aliphatic moiety resulted in a decrease in kcat/Km of at least three orders of magnitude. This was accompanied by a significant increase in Km values which showed a clear, inverse correlation to the length of the side chain, e.g. as the chain length of the amide decreased from six to two carbons, the Km value increased almost 5000-fold. Interestingly, Km values were generally affected much more than kcat values. Compared to phenylacetamide (1), acetamide (14) showed only a 30-fold decrease in kcat value but an increase in Km value of more than five orders of magnitude, while (R)-lactamide (12) and (S)-lactamide (13) showed similar increases in Km value but kcat values were largely unaffected. It was also apparent that substituents at the 2-position have only relatively minor effects on the kinetic parameters, and that there is little evidence of enantioselectivity with either aromatic or aliphatic substrates. It should be noted that, originally, MAH was given its name because its gene was isolated from a DNA fragment that contained other genes associated with the mandelate pathway of P. putida (McLeish et al., 2003
). In light of the results obtained in this study, it may well be more appropriate to refer to MAH as a phenylacetamide hydrolase. Overall it would seem that, for MAH, specificity is principally a binding issue and that once bound, the substrate can be turned over with reasonable efficiency. Thus, MAH would appear to be a good candidate for a directed evolution study.
While E.coli K-12 can utilize lactate as sole carbon source it will not grow on lactamide. However, given that wild-type (wt) MAH is able to hydrolyze lactamide at a very low efficiency (Table I), it was thought that an E.coli strain that expressed high levels of MAH may be able to survive on minimal media with lactamide as only carbon source. Accordingly, E.coli BL21-Gold(DE3)pLysS was transformed with pET17MAH-His (Gopalakrishna et al., 2004) and plated on a minimal medium containing 20 mM lactamide and 0.4 mM IPTG. Controls included the same volume of transformed cells plated on LB, on 20 mM glucose, on 20 mM lactate and on 20 mM lactamide in the absence of IPTG.
The initial screening showed that the cells were able to grow well on LB, and on 20 mM glucose. Growth on 20 mM lactate was observed after 2–3 days, while 4 days elapsed before colonies appeared on plates containing 20 mM lactamide. Somewhat surprisingly, the plates containing 20 mM lactamide/0.4 mM IPTG contained only 1–4% of colonies observed in the absence of IPTG. Samples of cells were collected from both lactamide plates and, after the cell density was adjusted to identical levels based on OD600, the level of MAH was measured by SDS–PAGE and by a coupled activity assay. It was found that the cells grown on the plate containing IPTG had little or no MAH, whereas moderate levels of MAH were observed for the cells grown without IPTG induction (data not shown). This result implies that high levels of MAH become toxic to the cells; therefore, overexpression of MAH induced by IPTG makes cells less likely to survive. However, not only can the cells tolerate the low levels of MAH expressed in the absence of IPTG, but there is sufficient LAH activity to allow cell growth, albeit slow.
Based on these experiments, it seemed reasonable to predict that an expression strain carrying a MAH mutant with improved catalytic efficiency for lactamide would show faster growth on minimal medium containing lactamide.
Construction of MAH mutant library by random mutagenesis
Random mutations on MAH were generated by error-prone PCR using the GeneMorph II Random Mutagenesis Kit, followed by cloning into the vector pET17b by megaprimer PCR. The mutation frequency was controlled by the initial concentration of target DNA used in the error-prone PCR. To verify the level of random mutations, plasmid DNA was prepared from randomly selected individual transformants. Sequencing of the MAH gene (1.6 kb) from transformants grown on LB plates provided a mutation frequency of 5.8 ± 2.2 mutations/kb, while transformants grown on minimal media provided a frequency of 2.8 ± 1.1 mutations/kb. In both cases, the mutations were randomly distributed over the entire MAH sequence. All types of mutations were observed including transitions (A
G and TC, 46%), transversions (AC, AT, CG and GT, 46%), insertions (4%) and deletions (4%) (data not shown).
In this study, two rounds of random mutagenesis were completed. In the first round, the wt enzyme was used as the template. After the initial selection on lactamide minimal media, 18 colonies were selected for screening with the coupled enzyme assay. As described in the experimental section, cell-free extracts were prepared from each of the 18 colonies and the activity of the extracts was determined using both (R,S)-lactamide (100 mM) and (R,S)-mandelamide (0.2 mM) as substrates. The ratio of the observed rates (Vlactamide/Vmandelamide) was used as an indicator to identify mutants with improved relative activity. The mandelamide concentration is well above the wt Km value (
20 µM) so the enzyme will be saturated, while the lactamide concentration is well below the wt Km value (500 mM). Consequently, any change in either kcat or Km value for lactamide, or a significant change in Km value for mandelamide, will be reflected in the ratio of Vlactamide/Vmandelamide. Under these conditions, wt MAH provided a ratio of 0.04, and control assays showed that the consumption of NADPH was negligible in the absence of lactamide or mandelamide.
Among the 18 colonies selected in the first round, all exhibited WT MAH activity with (R,S)-mandelamide, and only one showed an increase in Vlactamide/Vmandelamide ratio (Table II). Sequencing revealed this positive colony to be a double mutant, Q207H/Q382E (Supplementary data are available at PEDS online, Table S1). In addition, the mutant contained a silent mutation which proved useful in a subsequent round of mutagenesis (below). Following purification to homogeneity, kinetic parameters were measured for this mutant using the isomers of both mandelamide (2,3) and lactamide (12,13) as substrates. The Q207H/Q382E was found to display Michaelis–Menten parameters for mandelamide almost identical to those of wt MAH. In fact, the only significant change observed for this mutant is the doubling of the kcat value for lactamide. While this result was somewhat disappointing, it did confirm that changes in Michaelis–Menten parameters were reflected in changes in the Vlactamide/Vmandelamide ratio. Full kinetic details for this variant are provided as Supplementary data available at PEDS online, Table S2.
|
Second round of mutagenesis
The mutant Q207H/Q382E was used as the template in the second round of random mutagenesis. Of the 84 colonies that were selected in the second round, extracts from four exhibited increased relative activity for lactamide, with one showing a 150-fold increase in Vlactamide/Vmandelamide ratio (Table II). The double mutant, T31I/I437N, was somewhat surprising in that it had lost the two amino acid changes carried by the template (Supplementary data are available at PEDS online, Table S1). However, it did retain the silent mutation of the template, suggesting that T31I/I437N was likely to have been derived from Q207H/Q382E. It is not clear why these reversions occurred but, for informational purposes, the changes in both DNA and amino acid sequence are detailed in Supplementary data available at PEDS online, Table S1. Perhaps the most intriguing mutant was G202A/Q207H/Q382E which was active with lactamide but, under the standard assay conditions, showed no activity with mandelamide.
The kinetic parameters for each of the purified mutants are reported in Table III. Of these, the Q207H/S316N/Q382E mutant showed the smallest change in Vlactamide/Vmandelamide ratio, and this was consistent with its Michaelis–Menten parameters. It appears that the addition of the S316N mutation had little effect with aromatic substrates, while 2 to 3-fold increases in kcat value accompanied by minor decreases in Km value were observed for both 12 and 13.
|
The T31I/I437N mutant showed a 10-fold increase in Vlactamide/Vmandelamide ratio. Kinetically, this was manifested in a 10-fold increase in Km value for 2 and a 4-fold increase for 12. A much smaller increase was observed for 3, while the Km value for 13 actually decreased. Overall, this mutant showed a 10-fold selectivity for the (S)-enantiomers. Thus, either the T31I or the I437N mutation affected enantioselectivity for both aromatic and aliphatic substrates. This issue was examined in two ways. In the first instance, a triple mutant, Q207H/S316N/I437N, was prepared and characterized. The results shown in Table III confirm that the triple mutant also shows enhanced enantioselectivity. Further, it exhibited the greatest kcat/Km value for the hydrolysis of 13, i.e. a 7-fold increase in overall efficiency over wt MAH. The single mutant, I437N, was also prepared. The properties of this mutant largely mirrored those of the triple mutants, suggesting that Ile437 is likely to be located at or near the active site, and to play a role in enantioselectivity.
By far the most impressive results were provided by the G202A/Q207H/Q382E and G202S/Q207H/R236C/R369M/Q382E mutants. In the preliminary screen, the latter provided an increase in Vlactamide/Vmandelamide ratio of more than 150-fold, while the former was active with (R,S)-lactamide as a substrate but showed no activity with (R,S)-mandelamide. Closer examination with purified enzyme showed that, in both cases, the change in activity resulted from increases in the Km values for the aromatic substrates of almost three orders of magnitude, with only minor decreases in kcat values. Interestingly, the Km values for 12 and 13 also increased marginally but this was accompanied by up to a 4-fold increase in kcat values.
Taken together, the results from these mutants suggested that mutation of Gly202 may have substantially different effects on aromatic vs. aliphatic substrates. To investigate this possibility, two additional mutants were examined. First, the G202A single mutant was prepared and its kinetic characterization was expanded to include a variety of aromatic and aliphatic substrates. As shown in Table IV, dramatic changes to Km values were observed for all aromatic substrates (290–650-fold increase), but only minor changes in Km values were observed for aliphatic substrates (in the range of 0.4–2.6-fold). This provided further evidence that Gly202 is likely to be located at the binding site of the phenyl ring of mandelamide.
|
Subsequently, the G202V variant was also prepared. Given the additional bulk at position 202, it was not surprising that this variant showed even less tolerance for aromatic substrates (Table V). Indeed, the Km values for both (R)- and (S)-mandelamide increased beyond the limits of substrate solubility. Consequently, the kcat/Km values of 4.0 and 1.7 M–1 s–1 observed for (R)- and (S)-mandelamide, respectively, were obtained at substrate concentrations well below Km (i.e. V/K conditions). These values represent an additional three orders of magnitude decrease in kcat/Km values over those decreases observed for the G202A variant. The kcat/Km values of 0.5 and 1.4 M–1 s–1 observed for (R)- and (S)-lactamide are also down by two orders of magnitude showing that the additional bulk was also beginning to have an effect on hydrolysis of the smaller substrates. Unlike the aromatic substrates, it was possible to determine Km values for (R)- and (S)-lactamide, although they must be considered to be estimates as they were extrapolated from initial rate experiments carried out at substrate concentrations marginally below Km. These values showed only 4 to 7-fold increases over the values obtained for the WT and G202A variants suggesting that, for the smaller substrates, the effect of the valine substitution was on both kcat and Km.
|
Homology modeling of MAH
Directed evolution strategies include both random and targeted mutagenesis (Bloom et al., 2005). Of the two, the latter requires structural or biochemical data to choose appropriate targets and, with the exception of the previously identified catalytic residues (Gopalakrishna et al., 2004), such data were lacking for MAH. Undoubtedly, it would have been possible to generate a homology model of MAH at the outset, but the substrates of the available template enzymes were very different to mandelamide (and to each other). Thus, it would have been difficult to conclusively identify those residues important for substrate binding, certainly with the accuracy required for targeted mutagenesis. However, with the random mutagenesis results in hand, it was thought that a homology model would be of value in locating residues having an apparent effect on substrate specificity.
Three templates, FAAH, MAE2 and PAM were used in the generation of the MAH homology model. Of these, only MAE2 used a substrate, malonamide, that was remotely similar to mandelamide. That said, it was the quality of the individual alignments that resulted in MAE2 being chosen as the main template. The N-terminal region (1–62) was based on PAM and a smaller region (339–364) was based FAAH. The model (Fig. 3A) revealed that two residues that were identified in the random mutagenesis experiments, Gly202 and Gln207, were located adjacent to the catalytic triad. In addition, Arg236 and Ile437 were in close proximity. In contrast, Thr31, Ser316, Arg369 and Gln382 were located some distance away on the enzyme surface. At this point, it is still not really feasible to model mandelamide into the active site, at least with any degree of accuracy. However, when the product of the MAE2 reaction, malonate, is superimposed on this model (Fig. 3B), it is clear that Gly202 is the closest residue. Thus, it is not surprising that an increase in steric bulk at that position results in the considerable decrease in binding affinity for aromatic substrates as shown in Tables IV and V.
|
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
Here it has been shown that, while MAH has a strong preference for aromatic substrates, it will accept a variety of substituents at the 2-position. Further, it shows little or no enantioselectivity and may be better described as a phenylacetamide hydrolase. The kcat values for most substrates are quite similar but the Km values vary by five orders of magnitude. We have developed a method that allows us to select for MAH variants with an enhanced ability to utilize lactamide as a substrate. Combining this selection with two rounds of random mutagenesis enabled us to identify two residues that seem to be important for the substrate specificity of MAH. One of these, Ile437, had some influence on enantiospecificity whereas the other, Gly202, appeared to control specificity for aromatic substrates. A homology model of MAH showed that both of these residues are located near the putative active site. Mutation of Gly202 to alanine resulted in a 2000-fold decrease in kcat/Km for (S)-mandelamide along with a 2-fold increase in kcat/Km for (S)-lactamide. Most importantly, these changes have taken place without any great decrease in kcat values. In fact, the G202A variant has kcat values for (R)- and (S)-lactamide that are only marginally (1.5–2.5-fold) lower than the kcat values of the wt enzyme with (R)- and (S)-mandelamide. That said, the results from the G202V variant suggest that there is a limit to the size of the potential replacements for Gly202, and that the G202A variant may be optimal.
As alluded to in the introduction, previously we have attempted to interconvert BFDC, like MAH an enzyme recognizing a phenyl ring, and PDC which recognizes a methyl group (Siegert et al., 2005). That study employed a rational approach based on X-ray structures of both enzymes. The conversion of BFDC to PDC was relatively unsuccessful, essentially resulting in reduced BFDC activity. However, the conversion of PDC to BFDC provided an enzyme that retained a reasonable level of PDC activity (16-fold decrease in kcat/Km) while gaining significant decarboxylase activity with long-chain 2-ketoacids such as 2-ketohexanoic acid (12-fold increase in kcat/Km). It has been suggested that enzymes can acquire new activities via a ‘generalist’ intermediate which has broad substrate specificity and does not exhibit significant loss of activity with its original substrate (Khersonsky et al., 2006). In fact, many enzymes have a range of weak promiscuous activities that can be enhanced with only a few mutations (Aharoni et al., 2005; Bloom et al., 2005; Hult and Berglund, 2007) and PDC seems to fit into that category. Conversely, MAH appears to represent a far smaller group, in which mutations can result in an immediate change in substrate specificity. Certainly, this is true in so far as aromatic substrates are concerned. Such abrupt changes are often observed when dual selection pressure is applied (Varadarajan et al., 2005; Collins et al., 2006; Khersonsky et al., 2006), which is effectively the case when the Vlactamide/Vmandelamide ratio is used as the selection tool.
In our studies, to date we have used both site-directed (Siegert et al., 2005; Yep et al., 2006) and random (this work) approaches in an attempt to change specificity from large, bearing an aromatic ring to small substrates bearing a methyl group. These efforts have met with limited success, partly because contracting an active site appears to be inherently more difficult than expanding an active site. Another contributing factor may be that the techniques have been applied separately. It has been shown that, when engineering enzymes, a semi-rational approach, combining directed evolution and rational design, is more likely to yield positive results (Chica et al., 2005). Here we have identified several residues, particularly Gly202 and Ile437, which are important in MAH substrate recognition. These residues can be placed in context using the homology model which, in turn, can be used to single out other residues likely to be involved in substrate specificity. The selection method developed here can be made more stringent by decreasing the levels of lactamide in the minimal media. With these tools in hand, we are now in a position to use a more structured approach to developing an efficient LAH.
|
Supplementary data |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
Supplementary data are available at PEDS online
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
This work was supported by the National Science Foundation (grant number EF 0425719).
|
Footnotes |
|---|
1 Present address: Department of Chemistry and Chemical Biology, Indiana University-Purdue University Indianapolis, 402 N. Blackford St, Indianapolis, IN 46202, USA.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
For Dr Kenyon this material was based on work supported by the National Science Foundation while he was working there. Any findings or conclusions expressed here are those of the authors and do not necessarily reflect the views of the National Science Foundation.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionConclusionsSupplementary dataFundingAcknowledgementsReferences |
|---|
Aharoni A., Gaidukov L., Khersonsky O., McQ Gould S., Roodveldt C., Tawfik D.S. Nat. Genet. (2005) 37:73–76.[Web of Science][Medline]
Bloom J.D., Meyer M.M., Meinhold P., Otey C.R., MacMillan D., Arnold F.H. Curr. Opin. Struct. Biol. (2005) 15:447–452.[CrossRef][Web of Science][Medline]
Bracey M.H., Hanson M.A., Masuda K.R., Stevens R.C., Cravatt B.F. Science (2002) 298:1793–1796.
Chebrou H., Bigey F., Arnaud A., Galzy P. Biochim. Biophys. Acta (1996) 1298:285–293.[CrossRef][Medline]
Chica R.A., Doucet N., Pelletier J.N. Curr. Opin. Biotechnol. (2005) 16:378–384.[CrossRef][Web of Science][Medline]
Collins C.H., Leadbetter J.R., Arnold F.H. Nat. Biotechnol. (2006) 24:708–712.[CrossRef][Web of Science][Medline]
Cravatt B.F., Giang D.K., Mayfield S.P., Boger D.L., Lerner R.A., Gilula N.B. Nature (1996) 384:83–87.[CrossRef][Medline]
Curnow A.W., Hong K., Yuan R., Kim S., Martins O., Winkler W., Henkin T.M., Soll D. Proc. Natl Acad. Sci. USA (1997) 94:11819–11826.
Deutsch D.G., Chin S.A. Biochem. Pharmacol. (1993) 46:791–796.[CrossRef][Web of Science][Medline]
Gaffney T.D., da Costa e Silva O., Yamada T., Kosuge T. J. Bacteriol. (1990) 172:5593–5601.
Gomi K., Kitamoto K., Kumagai C. Gene (1991) 108:91–98.[CrossRef][Web of Science][Medline]
Gopalakrishna K.N., Stewart B.H., Kneen M.M., Andricopulo A.D., Kenyon G.L., McLeish M.J. Biochemistry (2004) 43:7725–7735.[CrossRef][Web of Science][Medline]
Guex N., Peitsch M.C. Electrophoresis (1997) 18:2714–2723.[CrossRef][Web of Science][Medline]
Hirrlinger B., Stolz A., Knackmuss H.J. J. Bacteriol. (1996) 178:3501–3507.
Hult K., Berglund P. Trends Biotechnol. (2007) 25:231–238.[CrossRef][Web of Science][Medline]
Jones D.T. J. Mol. Biol. (1999) 292:195–202.[CrossRef][Web of Science][Medline]
Khersonsky O., Roodveldt C., Tawfik D.S. Curr. Opin. Chem. Biol. (2006) 10:498–508.[CrossRef][Web of Science][Medline]
Labahn J., Neumann S., Buldt G., Kula M.R., Granzin J. J. Mol. Biol. (2002) 322:1053–1064.[CrossRef][Web of Science][Medline]
Luthy R., Bowie J.U., Eisenberg D. Nature (1992) 356:83–85.[CrossRef][Medline]
Mayaux J.F., Cerbelaud E., Soubrier F., Yeh P., Blanche F., Petre D. J. Bacteriol. (1991) 173:6694–6704.
Mazumder A., Gerlt J.A., Absalon M.J., Stubbe J., Cunningham R.P., Withka J., Bolton P.H. Biochemistry (1991) 30:1119–1126.[CrossRef][Web of Science][Medline]
McLeish M.J., Kneen M.M., Gopalakrishna K.N., Koo C.W., Babbitt P.C., Gerlt J.A., Kenyon G.L. J. Bacteriol. (2003) 185:2451–2456.
Murzin A.G., Brenner S.E., Hubbard T., Chothia C. J. Mol. Biol. (1995) 247:536–540.[CrossRef][Web of Science][Medline]
Nakamura A., Yao M., Chimnaronk S., Sakai N., Tanaka I. Science (2006) 312:1954–1958.
Negoro S. Appl. Microbiol. Biotechnol. (2000) 54:461–466.[CrossRef][Web of Science][Medline]
Pace C.N., Vajdos F., Fee L., Grimsley G., Gray T. Protein Sci. (1995) 4:2411–2423.[Web of Science][Medline]
Patricelli M.P., Cravatt B.F. J. Biol. Chem. (2000) 275:19177–19184.
Petsko G.A., Kenyon G.L., Gerlt J.A., Ringe D., Kozarich J.W. Trends Biochem. Sci. (1993) 18:372–376.[CrossRef][Web of Science][Medline]
Sánchez R., ali A. Protein Structure Prediction: Methods and Protocols—Webster D., ed. (2000) Totawa, NJ: Humana Press Inc. 97–130.
Schon A., Kannangara C.G., Gough S., Soll D. Nature (1988) 331:187–190.[CrossRef][Medline]
Schwarz A., Wandrey C., Steinke D., Kula M.R. Biotechnol. Bioeng. (1992) 39:132–140.[CrossRef][Web of Science][Medline]
Shapir N., Sadowsky M.J., Wackett L.P. J. Bacteriol. (2005) 187:3731–3738.
Shaw N.M., Naughton A.B. Tetrahedron (2004) 60:747–752.[CrossRef][Web of Science]
Shin S., Lee T.H., Ha N.C., Koo H.M., Kim S.Y., Lee H.S., Kim Y.S., Oh B.H. EMBO J. (2002) 21:2509–2516.[CrossRef][Web of Science][Medline]
Shin S., Yun Y.S., Koo H.M., Kim Y.S., Choi K.Y., Oh B.H. J. Biol. Chem. (2003) 278:24937–24943.
Siegert P., McLeish M.J., Baumann M., Iding H., Kneen M.M., Kenyon G.L., Pohl M. Prot. Eng. Des. Sel. (2005) 18:345–357.[CrossRef]
Varadarajan N., Gam J., Olsen M.J., Georgiou G., Iverson B.L. Proc. Natl Acad. Sci. USA (2005) 102:6855–6860.
Vriend G. J. Mol. Graph. (1990) 8:52–56.[CrossRef][Web of Science][Medline]
Wei Y., Kurihara T., Suzuki T., Esaki N. J. Mol. Cat. B Enzymatic (2003) 23:357–365.[CrossRef]
Yep A., Kenyon G.L., McLeish M.J. Bioorg. Chem. (2006) 34:325–336.[CrossRef][Web of Science][Medline]
Received July 3, 2008; revised November 2, 2008; accepted November 7, 2008.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||