ABSTRACT
The synthesis of diverse dl-configuration dipeptides in a one-pot reaction was demonstrated by using a function of the aminolysis reaction of a d-stereospecific amidohydrolase from Streptomyces sp., a clan SE, S12 family peptidase categorized as a peptidase with d-stereospecificity. The enzyme was able to use various aminoacyl derivatives, including l-aminoacyl derivatives, as acyl donors and acceptors. Investigations of the specificity of the peptide synthetic activity revealed that the enzyme preferentially used d-aminoacyl derivatives as acyl donors. In contrast, l-amino acids and their derivatives were preferentially used as acyl acceptors. Consequently, the synthesized dipeptides had a dl-configuration when d- and l-aminoacyl derivatives were mixed in a one-pot reaction. This report also describes that the enzyme produced cyclo(d–Pro-l-Arg), a specific inhibitor of family 18 chitinase, with a conversion rate for d-Pro benzyl ester and l-Arg methyl ester to cyclo(d-Pro–l-Arg) of greater than 65%. Furthermore, based on results of cyclo(d-Pro–l-Arg) synthesis, we propose a reaction mechanism for cyclo(d-Pro–l-Arg) production.
INTRODUCTION
Peptides incorporating d-amino acids in nature were found to be antimicrobial compounds and signal molecules. They are expected to increase the functional variety and applications of peptides (11, 12, 33, 38, 42). For example, alitame, l–Asp-d-Ala fenchyl ester, is known to be over a thousand times sweeter than sucrose (57). Cyclo(d-Pro–l-Arg) is known to act as a specific inhibitor of family 18 chitinase. It is also regarded as the ideal lead compound for the development of antifungal reagents and insecticides (25). In fact, d-amino-acid-containing peptides have been produced through organic synthesis (21, 44), and the complicated reaction steps and the racemization of amino acids present difficulties for organic synthesis. Therefore, enzyme-catalyzed peptide synthesis is well recognized as an alternative method for chemical synthesis. Recent reports have described enzymatic synthesis using peptidase (30, 32), esterase (52), aminoacyltransferase (53), and d-alanine–d-alanine ligase (48). Among them, aminolysis reactions of serine peptidases have increasingly attracted attention as tools for the synthesis of peptides because of their obviation of expensive additives such as ATP, their simple kinetics, and their ease of handling (1, 36).
In a previous study, we obtained two d-stereospecific amidohydrolases from Streptomyces spp. and confirmed that they have the function of an aminolysis reaction (7). The enzymes catalyze d-Phe–d-Phe synthetic activity when d-Phe methyl ester (d-Phe–OMe) is used as the substrate. In the predicted mechanism for d-Phe–d-Phe production, the enzyme uses d-Phe–OMe as an acyl donor and acceptor. Also, d-Phe–d-Phe–OMe emerged as an initial product. The esterified position of d-Phe–d-Phe–OMe was then hydrolyzed to d-Phe–d-Phe (Fig. 1). Therefore, the enzyme is of particular interest as a useful tool for the synthesis of various peptides incorporating d-amino acid. The primary structures of the enzymes show 30 to 60% identity with those of peptidases belonging to the clan SE S12 peptidase family in the MEROPS peptidase database (45). This family was categorized as serine peptidase with d-stereospecificity, the representative enzyme of which is d-Ala–d-Ala carboxypeptidase B from Streptomyces, which was used previously as a model for the study of penicillin-recognizing enzymes (19, 26). In addition, the enzyme has transpeptidase activity, and its properties are well known (27).
Possible mechanism for d-Phe-d-Phe synthesis from d-Phe–OMe catalyzed by a d-stereospecific amidohydrolase from Streptomyces sp. 82F2.
For this study, we constructed a recombinant enzyme of a d-stereospecific amidohydrolase from Streptomyces sp. 82F2 and evaluated its peptide bond formation activity, particularly its specificity toward acyl donors and acceptors. Although the function and structure of homologous enzymes have already been reported, the enzyme described herein possesses novel properties, such as the synthetic activity of a dl-configuration peptide. This study also demonstrates that the enzyme is applicable to the synthesis of biologically active dipeptides, cyclo(d-Pro–l-Arg).
MATERIALS AND METHODS
Materials, bacterial strain, and plasmids.Porcine kidney d-amino acid oxidase and snake venom l-amino acid oxidase were purchased from Sigma Chemical Co. Horseradish peroxidase was purchased from Wako Pure Chemical Industries Ltd. Aminoacyl derivatives were purchased from Bachem AG, Aldrich Chemical Co. Inc., Sigma Chemical Co., Novabiochem Corp., and Wako Pure Chemical Industries Ltd. The plasmids PCR-Blunt II-TOPO (Invitrogen Corp.) and pET-22b (Novagen Inc.) were used as cloning and expression vectors, respectively. As a host strain for general cloning procedures, Escherichia coli JM109 was used; E. coli Rosetta(DE3) was used as a host strain for gene expression.
Enzyme assay.For routine assays, the enzyme activity was determined by using a continuous spectrophotometric assay with d-Phe–p-nitroanilide (pNA) as the substrate. In the assays, 10 μl of substrate solution (10 mM) was added to 90 μl of a mixture containing 200 mM Tris-malate (pH 6.5) and 5 μg/ml of enzyme at 25�C. The increased absorption at 405 nm caused by the release of p-nitroaniline each minute was monitored continuously by use of a microtiter plate reader (680; Bio-Rad Laboratories Inc.). The initial activity rate was determined from a linear part of the optical density profile of p-nitroaniline measured using the same instrument.
Construction of expression vectors.The gene encoding the enzyme without a signal sequence was amplified from chromosomal DNA of Streptomyces sp. 82F2 by PCR using primers 5′-TGGCCATGGCGCCCGCGAAGCCGGACCACG-3′ and 5′-AAGCTTCTACTTCTTCGGGGCGGTGCCGCA-3′ (the underlined regions are the MscI or HindIII site). The PCR product was cloned into PCR-Blunt II-TOPO. The correct cloning was confirmed by sequencing. The gene encoding d-stereospecific amidohydrolases was then subcloned into the MscI-HindIII gap of pET-22b. The gene was fused to the pelB signal sequence, a sequence for directing the protein to the bacterial periplasm at the 5′ end (yielding pET-82F2DAP).
Expression and purification of recombinant enzymes.We cultivated E. coli Rosetta(DE3) harboring pET-82F2DAP at 25�C for 48 h in 50 ml of Overnight Expression Instant TB medium (Novagen Inc.). After the culture was centrifuged to remove the cells, the recombinant enzyme was purified by using the following procedures. The culture supernatant was dialyzed against 20 mM sodium acetate (pH 5.5). The dialysate was loaded onto a Vivapure-S spin column (Sartorius AG) equilibrated with 20 mM sodium acetate (pH 5.5). After washing with the same buffer containing 0.1 M NaCl, the bound protein was eluted by using the same buffer containing 0.4 M NaCl. The homogeneity of purified proteins was confirmed by 12% SDS-PAGE under denaturing conditions (35).
Peptide synthesis by an aminolysis reaction.Peptide synthesis by an aminolysis reaction was performed as follows. Two microliters of acyl donor substrate and 2 μl of acyl acceptor substrate (aminoacyl derivative solution [in dimethyl sulfoxide {DMSO}] at an appropriate concentration [ca. 0.5 M]) were added to 44 μl of 0.25 M Tris-HCl (pH 8.5). The reaction was initiated by the addition of 2 μl of a 0.2-mg/ml or 1-mg/ml [only for cyclo(d-Pro–l-Arg) synthesis] enzyme solution to the mixture. The reaction was then continued at 25�C for an appropriate time (2 min to 40 h). The reaction was terminated by the addition of 0.05 ml of 3% formic acid to the mixture. The reaction mixture was then analyzed by mass spectrometry (MS).
MS analysis.The molecular mass of peptides synthesized by the enzyme was determined by using electrospray ionization-time of flight (ESI-TOF) MS. For ESI-TOF MS analysis, the reaction mixture was diluted with a 200-fold volume of 0.1% formic acid. After the solution was filtered, 5 μl of each sample was analyzed by using an ESI-TOF MS system (LCT Premier XE; Waters Corp.). The data were processed by using a computer program (MassLynx; Waters Corp.). The ESI MS data are summarized in the supplemental material.
Quantification of synthesized dipeptides.For the calculation of the amounts of d-Phe–d-Phe and hydrolysates of aminoacyl derivatives, we measured their concentrations using the standard curves described below. The concentrations of d-Phe–l-Trp and d-Phe–l-Trp-OMe (cFW) (in mM) were calculated from the difference in the amounts of substrate (d-Phe benzyl ester [d-Phe–OBzl]) (cF1) (in mM) and the concentrations of d-Phe (a hydrolysate of d-Phe-OBzl) (cF2) (in mM) and d-Phe–d-Phe (a by-product of the aminolysis reaction) (cFF) (in mM) according to the formula cFW = cF1 − (cF2 + 2cFF). For the calculation of the concentration of synthesized cyclo(d-Pro–l-Arg) (cCPR) (in mM), we measured the concentration of d-Pro (a hydrolysate of d-Phe–OBzl) (cP2) (in mM) and then calculated the difference in the amounts of substrate (d-Pro–OBzl) (cP1) (in mM) and the concentrations of d-Pro according to the formula cCPR = cP1 − cP2. The method for kinetic analysis to determine the Km values for d-Phe–OBzl (acyl donor), l-Trp (acyl acceptor), and l-Trp-OMe (acyl acceptor) for peptide bond formation activity is described in the supplemental material.
Using an ultraperformance liquid chromatography (UPLC)-ESI-TOF MS instrument equipped with a C18 reverse-phase system (Acquity UPLC; Waters Corp.), we quantified the concentrations of the respective chemicals. The reaction mixture was diluted with a 4,000-fold volume of 0.1% formic acid and filtered. Five microliters of each sample was then subjected to chromatography. Each sample was eluted with solvent A-solvent B at a 95:5 dilution for 2 min, solvent A-solvent B at an 80:20 dilution for 1 min, a solvent A-solvent B gradient of 80:20 to 50:50 dilutions for 2 min, and solvent A-solvent B at a 20:80 dilution for 2 min, where solvent A was Milli-Q water containing 0.1% formic acid and solvent B was acetonitrile containing 0.1% formic acid. The data were processed by using a computer program (MassLynx; Waters Corp.). The concentration was calculated from the area of ion intensity using the standard curves for the respective chemicals.
Hydrolytic activity toward d-Phe–OBzl.The hydrolytic activity toward d-Phe–OBzl was determined by quantitative assays of amino acids derived from the hydrolyzed substrates. Two microliters of an enzyme solution (0.2 mg�ml−1) and 2 μl of d-Phe–OBzl (0.5 M in DMSO) were added to 46 μl of 200 mM Tris-maleic acid (pH 8.5). The reaction mixture was incubated at room temperature for 5 min. The reaction was then stopped by the addition of 50 μl of 3% formic acid to the mixture. The liberated d-Phe was quantified by using a UPLC-ESI-TOF MS instrument equipped with a C18 reverse-phase system. The method used for kinetic analysis to determine the Km value for d-Phe–OBzl for hydrolytic activity is described in the supplemental material.
Preference of acyl donor and acyl acceptor substrates.For investigations of the preferences of acyl donor and acceptor substrates, we compared the amounts of the substrate consumed by the aminolysis reaction. For investigations of the acyl acceptor preferences, 2 μl of 0.5 M d-Phe benzyl ester and 4 μl of 0.5 M acyl acceptor (l-amino acid or its derivative) were added to 42 μl of 0.25 M Tris-HCl (pH 8.5). The reaction was initiated by the addition of 2 μl of a 0.2-mg/ml enzyme solution to the mixture. Because d-Phe–OBzl was consumed completely after a 3-h reaction at 25�C, the reaction was performed for 3 h. After the reaction, the d-Phe liberated by the hydrolytic activity was detected by using the 4-aminoantipyrine phenol method (2) coupled with a d-amino acid oxidase reaction. The reaction solution (10 μl) was added to 90 μl of a mixture containing 200 mM Tris-HCl (pH 8.0), 0.5 mM 4-aminoantipyrine, 1.7 mM phenol, 50 μg/ml horseradish peroxidase, and 0.2 mg/ml porcine kidney d-amino acid oxidase. After incubation for 2 h at 37�C, the absorbance at 505 nm was determined. The concentration of d-Phe liberated by using the hydrolysis reaction was determined from a linear part of the optical density profile of d-Phe measured by using a microtiter plate reader (680; Bio-Rad Laboratories Inc.). The acyl acceptor preference was determined by calculations of the amount of d-Phe–OBzl consumed in aminolysis by deducing the amount of free d-Phe from the quantity of d-Phe–OBzl added to the reaction mixture.
For investigations of the acyl donor preference, we used l-Trp-OMe as a substrate for the acyl acceptor. First, 2 μl of 0.5 M acyl donor substrate (d-aminoacyl derivative) and 2 μl of 0.5 M l-Trp-OMe were added to 44 μl of 0.25 M Tris-HCl (pH 8.5). The reaction was initiated by the addition of 2 μl of a 0.2-mg/ml enzyme solution to the mixture. After reaction for 24 h at 25�C, 50 μl of 0.5 M NaOH was added to the reaction mixture for the deesterification of the remaining acyl acceptor substrate. The liberated amino acids were assayed by using l-amino acid oxidase according to the procedures described above.
Validation of enzymatic or nonenzymatic cyclization of d-Pro–l-Arg-OMe.To determine whether the cyclization of d-Pro–l-Arg-OMe, the product of the enzymatic reaction from d-Pro-OBzl and l-Arg-OMe, occurred enzymatically or nonenzymatically, we first synthesized d-Pro–l-Arg-OMe enzymatically as follows. Two microliters of d-Pro-OBzl (0.5 M dissolved in DMSO) and l-Arg-OMe (0.5 M dissolved in DMSO) was added to 44 μl of 0.25 M Tris-HCl (pH 8.5). The reaction was initiated by the addition of 2 μl of a 0.2-mg/ml enzyme solution to the mixture, continued at 25�C for 1 h, and then terminated by the addition of 0.05 ml of 3% formic acid to the mixture. Subsequently, the product d-Pro–l-Arg-OMe was separated from protein by ultrafiltration. From the obtained solution, 10 μl containing d-Pro–l-Arg-OMe was then added to 40 μl of 0.25 M Tris-HCl (pH 8.5). After 1 h and 24 h of incubation at 25�C, cyclo(d-Pro–l-Arg) production was assayed according to the procedures described above in “MS analysis.”
RESULTS
Overproduction of the recombinant d-stereospecific amidohydrolase.We constructed an expression plasmid (pET-82F2DAP) to overexpress the d-stereospecific amidohydrolase from Streptomyces sp. 82F2 by E. coli. The gene was expressed under the control of the T7 promoter. The recombinant enzyme was secreted extracellularly to up to 70% of the total extracellular protein amount (see Fig. S1 and Table S1 in the supplemental material). After the enzyme was purified to homogeneity, we confirmed the properties of the recombinant enzyme, such as its thermal stability, pH optimum, pH stability, and substrate specificity for hydrolytic activity. The results show that it had the same properties as those of enzymes purified from Streptomyces sp. 82F2 reported previously (7; data not shown).
The enzyme can use l-amino acyl derivatives as a substrate for the aminolysis reaction.Through investigations of homopeptide synthetic activity using 43 aminoacyl derivatives as substrates, a homopeptide was detected when d- and l-Leu derivatives, l-Met-OMe, d- and l-Phe derivatives, d- and l-Trp derivatives, and d- and l-Tyr derivatives were used (see Table 2). It is particularly interesting that dipeptidyl derivatives were detected when an l-aminoacyl derivative was used as a substrate, although nonderivatized dipeptides were synthesized from d-aminoacyl derivatives similarly when d-Phe–OMe was used as the substrate for aminolysis (Fig. 1). Free d-amino acid cannot act as an acyl acceptor for peptide synthetic activity (7). Therefore, the enzyme recognizes only d, d-configuration dipeptide esters as a substrate for hydrolysis.
Arrangement of the synthesized dipeptide.This investigation revealed that free l-amino acid and its derivatives are beneficial for use as acyl acceptors for peptide synthesis. Free l-Trp added to the reaction mixture for d-Phe–d-Phe production (d-Phe–OBzl was maintained at a steady level of 20 mM) decreased the productivity of d-Phe–d-Phe and free d-Phe, a hydrolysate of d-Phe–OBzl. The level of productivity of d-Phe–l-Trp was higher at higher free l-Trp concentrations (Fig. 2A). The roughly calculated concentration of d-Phe–l-Trp in the reaction mixture was 11.8 mM when 40 mM free l-Trp was added to the reaction mixture for d-Phe–d-Phe production, which indicates that the rate of conversion of 20 mM substrate d-Phe–OBzl to d-Phe–l-Trp can be estimated to be approximately 59%. The initial rate (ca. 5 min) for d-Phe–l-Trp synthesis was 40.5 � 2.4 μmol�min−1�mg−1 when 20 mM d-Phe–OBzl and 40 mM free l-Trp were used as substrates (Table 1).
Effect of substrate concentration on peptide synthesis. (A) Effect of l-Trp concentration. (B) Effect of l-Trp-OMe concentration. d-Phe–OBzl at 20 mM and l-Trp or l-Trp-OMe at 0 to 40 mM were used as substrates. For investigations, the reaction was performed at 25�C for 1 h. The y axis of each panel shows the area of ion intensity. Each value represents an average of values from three independent experiments � the standard deviation.
Enzyme kinetics for peptide bond formation activity
A similar result was obtained from the investigation of the addition of l-Trp-OMe to the reaction mixture containing a steady level of 20 mM d-Phe–OBzl (Fig. 2B). In this reaction, 94% of 20 mM d-Phe–OBzl was converted to d-Phe–l-Trp-OMe; no l-Trp–d-Phe–OBzl was detected when 40 mM l-Trp-OMe was added to the reaction mixture. The initial rate (ca. 5 min) for d-Phe–l-Trp-OMe synthesis was 64.7 � 0.9 μmol�min−1�mg−1 when 20 mM d-Phe–OBzl and 40 mM l-Trp-OMe were used as substrates (Table 1). The value is almost identical to the rate of d-Phe–OBzl hydrolysis at pH 8.5. The results indicate that the enzyme preferentially uses l-aminoacyl derivatives as acyl acceptors unless no d-aminoacyl derivative exists in the reaction mixture. Furthermore, the enzyme does not hydrolyze the ester bond of d-l-formed dipeptidyl esters, judging from the results obtained when d-Phe–OBzl and l-Trp-OMe were used. For this investigation, we additionally confirmed that no racemization of the product d-Phe–l-Trp-OMe occurred (see Fig. S2 in the supplemental material).
We further performed a kinetic characterization of the enzyme. The Km values for free l-Trp and l-Trp-OMe as acyl acceptors for peptide synthetic activity were 316 and 8.48 mM, respectively. This investigation revealed remarkable differences in Km values for d-Phe–OBzl in terms of its hydrolytic activity, d-Phe–l-Trp synthesis, and d-Phe–l-Trp-OMe synthesis. As presented in Table 1, the Km value for d-Phe–OBzl for d-Phe–l-Trp-OMe synthesis was 0.37 mM when l-Trp-OMe was used as an acyl acceptor. In contrast, the Km value for d-Phe–OBzl for hydrolytic activity was approximately 10-fold lower than that for d-Phe–l-Trp-OMe synthesis. In addition, when free l-Trp was used as an acyl acceptor, the Km value for d-Phe–OBzl for d-Phe–l-Trp synthesis was 6-fold higher than that for d-Phe–l-Trp-OMe synthesis (double-reciprocal plots of the Michaelis-Menten equation for the determination of all Km values are shown in Fig. S3 in the supplemental material). The difference in Km values was regarded as being associated with a change in the space for acyl donor binding or the electrostatic environment by the change in the acyl acceptor.
Specificity toward acyl donors and acceptors.We evaluated the synthesis of heteropeptides by reacting d-Phe–OBzl or l-Leu ethyl ester with various aminoacyl derivatives as substrates at pH 8.5 for 1 h. Although the substrates used for homopeptide synthesis were limited, the results of this investigation indicate that the enzyme can use various free l-amino acids and aminoacyl derivatives as acyl acceptors and donors (Table 2). Almost all chemicals can be categorized as an acyl acceptor, an acyl donor, or both a donor and acceptor. Dipeptidyl derivatives were detected as one product obtained when l-aminoacyl derivatives were used as a substrate for the reaction.
Tested chemicals and peptides synthesized by use of recombinant enzymesaa
We next examined acyl acceptor and donor preferences by the calculation of the rate of conversion of acyl substrates to dipeptides, as described in Materials and Methods. As depicted in Fig. 3, a high conversion rate is biased toward synthesis using acyl donors and acceptors that have hydrophobic side chains. In addition, the results of acyl acceptor preferences indicated that the conversion rate was affected by the flanking moiety. Some examples are that the rate of conversion of d-Phe–OBzl to dipeptide when l-Leu-OBzl was used as an acyl acceptor was the highest among those investigated when l-Leu derivatives were used as acyl acceptors. A similar result was obtained when other l-aminoacyl derivatives were used.
Substrate specificity of the aminolysis reaction catalyzed by a d-stereospecific amidohydrolase. (A) Specificity of the enzyme toward the acyl acceptor. In the assay, 20 mM d-Phe–OBzl and 20 mM l-amino acid or l-aminoacyl derivatives were used as the acyl donor and acyl acceptor. (B) Specificity of the enzyme toward the acyl donor. In the assay, 20 mM d-aminoacyl derivatives and 40 mM l-Trp-OMe were used as the acyl donor and acyl acceptor. In all cases, the reaction was performed by using 0.4 μg of the enzyme (in a 50-μl reaction mixture) with vigorous shaking at 25�C at pH 8.5 for 1 h. Each value represents an average of values from three independent experiments � the standard deviation.
Cyclo(d-Pro–l-Arg) production.The enzyme uses d-Pro-OBzl and l-Arg-OMe as an acyl donor and an acyl acceptor, respectively. Therefore, we attempted to synthesize cyclo(d-Pro–l-Arg), which shows an inhibition of family 18 chitinase (25). As depicted in Fig. 4, when 20 mM d-Pro-OBzl and 20 mM l-Arg-OMe were used as substrates, the two products d-Pro–l-Arg-OMe and cyclo(d-Pro–l-Arg) were detected after a 30-min reaction. In addition, the concentration of d-Pro–l-Arg-OMe was decreased and that of cyclo(d-Pro–l-Arg) was increased after a 3-h reaction (Fig. 4). The results indicate that the cyclization of d-Pro–l-Arg-OMe occurred concomitantly with peptide bond formation in a one-pot reaction. For this investigation, we additionally confirmed that no racemization of the product cyclo(d-Pro–l-Arg) occurred (see Fig. S2 in the supplemental material).
Cyclo(d-Pro–l-Arg) synthesis catalyzed by a d-stereospecific amidohydrolase. (A) Extracted ion chromatograms of substrates (d-Pro-OBzl and l-Arg-OMe) and products (d-Pro and peptides) and total ion chromatograms (T.I.C.) of samples from 30-min and 3-h reactions. The m/z values for monitoring the substrates and products are shown in each panel. c.p.s., counts per second. (B) MS of peptide products synthesized with d-Pro-OBzl and l-Arg-OMe as substrates. (Top) MS chart for peak a in the extracted ion chromatograms of m/z 286.19 (d-Pro–l-Arg-OMe). (Bottom) MS chart for peak b in the extracted ion chromatograms of m/z 254.15 [cyclo(d-Pro–l-Arg)].
As portrayed in Fig. 5, when 20 mM d-Pro-OBzl and 20 mM l-Arg-OMe were used as substrates, d-Pro–l-Arg-OMe was synthesized efficiently up to 40 min. The concentration of the product d-Pro–l-Arg-OMe then decreased gradually. It was converted completely into cyclo(d-Pro–l-Arg) after 8 h. Cyclo(d-Pro–l-Arg) was synthesized efficiently up to 4 h. The quantity of cyclo(d-Pro–l-Arg) increased only slightly thereafter. In terms of substrate consumption, d-Pro-OBzl was consumed completely and converted into cyclo(d-Pro–l-Arg) or free d-Pro after 3 h. In contrast, free l-Arg emerged only slightly up to 8 h (data not shown), indicating that the enzyme uses l-Arg-OMe only as an acyl acceptor and that it has no hydrolytic activity toward l-Arg-OMe. The initial rate (ca. 5 min) of 17.4 � 3.19 μmol�min−1�mg−1 for d-Pro–l-Arg-OMe synthesis was calculated from the consumption of l-Arg-OMe. The calculated concentration of cyclo(d-Pro–l-Arg) in the reaction mixture was 14.3 mM. The result shows that the rate of conversion of 20 mM substrate d-Pro-OBzl to cyclo(d-Pro-l-Arg) was approximately 71%.
Time dependence of d-Pro–l-Arg-OMe and cyclo(d-Pro–l-Arg) syntheses. (A) Profiles of concentrations of the substrates and Pro, a hydrolysate of d-Pro-OBzl. (B) Profiles of the products d-Pro–l-Arg-OMe and cyclo(d-Pro–l-Arg). The reaction was continued up to 8 h. l-Arg-OMe at 20 mM and d-Pro-OBzl at 20 mM were used as the acyl acceptor and donor, respectively. The reaction was performed at 25�C. Each value represents an average of data from three independent experiments � the standard deviation.
Cyclization of d-Pro–l-Arg-OMe.To verify whether the cyclization of d-Pro–l-Arg-OMe occurred enzymatically or nonenzymatically, the product d-Pro–l-Arg-OMe was separated and exposed at pH 8.5. As presented in Fig. 6, cyclo(d-Pro–l-Arg) production was continued by the exposure of d-Pro–l-Arg-OMe to pH 8.5, although no production of cyclo(d-Pro–l-Arg) was observed for pHs lower than 6.0 (data not shown). The reaction rate of cyclization was not increased by the addition of the enzyme (see Fig. S4 in the supplemental material). Therefore, the result indicates that the cyclization of d-Pro–l-Arg-OMe occurred nonenzymatically. The proposed mechanism for cyclo(d-Pro–l-Arg) production is presented in Fig. 7.
Cyclization of d-Pro–l-Arg-OMe. Shown are extracted ion chromatograms (m/z 286.19 and m/z 254.15) and MS of separated d-Pro–l-Arg-OMe synthesized from d-Pro-OBzl and l-Arg-OMe (top) and separated d-Pro–l-Arg-OMe exposed at pH 8.5 for 3 h (middle) and for 8 h (bottom).
Proposed mechanism for cyclo(d-Pro–l-Arg) [c(d-Pro–l-Arg)] production by a d-stereospecific amidohydrolase.
DISCUSSION
Several stereospecific peptidases that recognize an amide bond involving d-amino acid residues have been reported (4, 8–10, 15, 17, 22, 24, 46, 47, 50, 51). They are regarded as being associated mainly with the biosynthesis and remodeling of peptidoglycan. Among them, peptidases such as d-Ala–d-Ala carboxypeptidase B from Streptomyces sp. R61 (34), d-aminopeptidase from Ochrobactrum anthropi (30), and d-peptidase from Bacillus cereus (32) have the function of an aminolysis reaction. The enzymes belong to clan SE, the enzymes of the S11, S12, and S13 peptidase families, which are specialized for roles in bacterial cell wall metabolism. The enzyme used for this study also showed 30 to 60% identity with those of peptidases of the S12 peptidase family (7). The peptidases of family S12 exhibit various aminolysis reactions (http://merops.sanger.ac.uk/). Among the enzymes of the S12 peptidase family, d-Ala–d-Ala carboxypeptidase B from Streptomyces sp. R61 has been widely studied as the prototype of the enzymes of this family (28, 43). The enzyme catalyzes two reactions: dd-carboxypeptidase activity, in which there is a transfer of the C-terminal d-Ala to water, and dd-transpeptidase activity, in which the peptidoglycan monomer is transferred to an exogenous receptor after the removal of the C-terminal d-Ala (20). Representative substrates of this enzyme are Nα,Nα-Ac2-l-Lys–d-Ala–d-Ala. In contrast, the enzyme in this study exhibited only slight activity toward the peptide substrate and high hydrolytic activity toward d-aminoacyl esters with a hydrophobic side chain (7). The peptidases of family S12 exhibit widely various activities. Therefore, the enzyme in this study is one enzyme with a substrate specificity that differs from that of d-Ala–d-Ala carboxypeptidase B.
As an enzyme with a function similar to that of the enzyme described as a result of this study, d-aminopeptidase from Ochrobactrum anthropi recognizes d-Ala–pNA as a substrate for hydrolytic activity. This enzyme can remove N-terminal d-Ala from the peptide substrate d-Ala–l-Ala–l-Ala. Although the enzyme in this study was unable to hydrolyze a peptide substrate, it can recognize d-Phe–pNA as a substrate for hydrolytic activity. Therefore, the d-aminopeptidase from Ochrobactrum anthropi is the most functionally similar enzyme to the d-stereospecific amidohydrolase from Streptomyces sp. 82F2. The three-dimensional structures of d-Ala–d-Ala carboxypeptidase B from Streptomyces sp. R61 and d-aminopeptidase from Ochrobactrum anthropi have been described in the literature (13, 31). A previous study described that the loop in the C-terminal domain (L3 loop) was responsible for their different substrate specificities (16). Clarification of the structure associated with the substrate specificity of the enzyme described as a result of this study remains an important problem to be solved to elucidate the mechanism for substrate recognition.
The enzyme examined in this study showed a high specific activity for aminolysis. The initial rates for peptide synthetic activities were 17.4, 40.5, and 64.7 μmol�min−1�mg−1 for d-Pro–l-Arg-OMe, d-Phe–l-Trp, and d-Phe–l-Trp-OMe syntheses, respectively. According to a previous report by Kato et al. (29), the kcat value for d-alanyl N-alkylamide synthesis by aminolysis, as catalyzed by the d-aminopeptidase from Ochrobactrum anthropi, was 7,700 min−1. The calculated kcat values for the d-Pro–l-Arg-OMe, d-Phe–l-Trp, and d-Phe–l-Trp-OMe synthetic activities of the enzyme in this study were 750, 1,540, and 2,460 min−1, respectively, which are lower values than those of the d-aminopeptidase from Ochrobactrum anthropi. There is the possibility of obtaining higher kcat values than those obtained in this study by an investigation of substrate specificity using various amine chemicals as acyl acceptors.
Except for the peptidases of family S12, some serine peptidases exhibit peptide bond formation from aminoacyl derivatives. They include the aminolysis of esters, thioesters, and amides, in accordance with their hydrolytic activities (14, 56). In enzymatic peptide synthesis by a reverse reaction, not by the aminolysis reaction, peptides that act as a good substrate for hydrolysis are appropriate targets of synthesis (3, 39, 41). In contrast, the enzyme examined in this study showed that the products were not hydrolyzed by the enzyme itself (Fig. 2). A similar phenomenon was observed previously for an S9 aminopeptidase from Streptomyces. Although the S9 aminopeptidase has a catalytic triad, Ser-His-Asp (40, 54), and although its catalytic mechanism differs completely from that of an enzyme belonging to the S12 family, the primal reaction steps for aminolysis and the hydrolysis of both enzymes are almost identical. S9 aminopeptidase can synthesize various β-Ala-containing peptides, including carnosine (β-Ala–His) (6), by its aminolysis reaction. Because S9 aminopeptidase cannot hydrolyze carnosine, the synthesized carnosine accumulates in the reaction mixture. In fact, the level of hydrolytic activity of the enzyme reported in this study toward the peptide substrate was extremely low (7). Therefore, the products are not always recognized as substrates for the hydrolytic activity of the enzyme from Streptomyces sp. 82F2 as for that of the S9 aminopeptidase. The enzyme properties reported in this study differ from those of the S9 aminopeptidase by the recognition of various amine chemicals as acyl acceptors instead of water molecules.
This report also shows that the enzyme from Streptomyces sp. 82F2 is applicable for the synthesis of biologically active dipeptides, cyclo(d-Pro–l-Arg), with the conversion rate of d-Pro-OBzl and l-Arg-OMe to cyclo(d-Pro–l-Arg) being greater than 65%. The cyclization of d-Pro–l-Arg-OMe occurred nonenzymatically (Fig. 6). The same phenomenon was described in a previous report of cyclic peptide synthesis using S9 aminopeptidase from Streptomyces (5, 55). Several molecules containing a 2,5-diketopiperazine moiety, a cyclic dipeptide, show promising biological activities that are expected to be helpful in treating human diseases (18, 37). In addition to the use of peptidases, an enzymatic method using cyclodipeptide synthase for the synthesis of such cyclic dipeptides was reported previously (23, 49). Nevertheless, the enzyme described above required expensive substrates, such as aminoacyl-tRNA. Therefore, the synthesis of cyclic dipeptides using the enzyme from Streptomyces sp. 82F2 is a more cost-effective method than the method described above. In addition, the enzyme can use free l-amino acids as acyl acceptors, indicating that the structure of peptides (linear or cyclic) synthesized by the enzyme from Streptomyces sp. 82F2 can be designed easily by selecting the acyl acceptor.
The results of this study demonstrated that the d-stereospecific amidohydrolase from Streptomyces sp. 82F2 is applicable to the synthesis of various dl-configuration peptides. The enzyme uses d-aminoacyl derivatives preferentially as acyl donors and uses l-amino acids and their derivatives preferentially as acyl acceptors. Therefore, the arrangement of synthesized peptides can be highly controlled. This report also describes that the enzyme is applicable for the synthesis of biologically active dipeptides, cyclo(d-Pro–l-Arg), with a high conversion rate. The next task for us will be to screen biologically active dl-configuration peptides from a peptide library constructed by using an aminolytic reaction using various racemic aminoacyl derivatives.
ACKNOWLEDGMENT
This work was supported by a grant-in-aid for scientific research from the Japan Science and Technology Agency.
FOOTNOTES
- Received 19 May 2011.
- Accepted 19 September 2011.
- Accepted manuscript posted online 23 September 2011.
↵† Supplemental material for this article may be found at http://dx.doi.org/10.1128/AEM.05543-11.
- Copyright � 2011, American Society for Microbiology. All Rights Reserved.
