DOI: 10.1002/cbic.201402632
Full Papers
Structural Analysis Reveals the Substrate-Binding Mechanism for the Expanded Substrate Specificity of Mutant meso-Diaminopimelate Dehydrogenase Weidong Liu, Rey-Ting Guo, Xi Chen, Zhe Li, Xiuzhen Gao, Chun-Hsiang Huang, Qiaqing Wu, Jinhui Feng, and Dunming Zhu*[a] late (meso-DAP), d-leucine (d-leu), and 4-methyl-2-oxopentanoic acid (MOPA) were solved. meso-DAP was found in an area outside the catalytic cavity; this suggested a possible two-step substrate-binding mechanism for meso-DAP. d-leu and MOPA each bound both to Leu154 and to Gly155 in the open form of CtDAPDH, and structural analysis revealed the molecular basis for the expanded substrate specificity of the mutant meso-diaminopimelate dehydrogenases.
A meso-diaminopimelate dehydrogenase (DAPDH) from Clostridium tetani E88 (CtDAPDH) was found to have low activity toward the d-amino acids other than its native substrate. Sitedirected mutagenesis similar to that carried out on the residues mutated by Vedha-Peters et al. resulted in a mutant enzyme with highly improved catalytic ability for the synthesis of d-amino acids. The crystal structures of the CtDAPDH mutant in apo form and in complex with meso-diaminopime-
Introduction Meso-diaminopimelate dehydrogenase (DAPDH, EC.1.4.1.16) catalyzes reversible oxidative deamination/reductive amination reactions at the d-center of meso-2,6-diaminopimelate (mesoDAP, Scheme 1), and most of the known wild-type DAPDHs are
d-amino acids, thus a change in the l-center-binding site of DAPDH should expand its substrate spectrum towards other damino acids. Working on the basis of the structure information of CgDAPDH, Vedha-Peters et al.[2a] first successfully improved the activity of CgDAPDH to other d-amino acids through three rounds of directed evolution, with the first round of mutagenesis targeting those residues that interact with the substituents on the l-carbon in the native substrate meso-DAP and the other two rounds targeting the whole enzyme. Recently, Akita et al.[2b, c] reported similar results for the mutation of the corresponding amino acid residues of a DAPDH from Ureibacillus thermosphaericus (UtDAPDH). In these studies, the mutated positions of DAPDH are equivalent and include residues in the binding sites both of the l-center and of the d-center. The l-center residues are Thr169 (Thr173), Arg195 (Arg199), and His245 (His249) of CgDAADH (UtDAPDH), whereas the d-center residue is Gln150 (Gln154) of CgDAPDH (UtDAPDH). We recently discovered a thermostable DAPDH from Symbiobacterium thermophilum (StDAPDH) with significant activity towards other d-amino acids as well as meso-DAP,[7] and its substrate specificity was expanded to include larger amino acids such as phenylalanine by mutagenesis on its l-center residues.[8] The structural and mutagenesis studies[9] revealed that a new substrate entrance tunnel, along with residue Met152, was important for the binding of small substrates such as pyruvate/d-alanine. Structural comparison showed that the major difference between the catalytic cavity residues of StDAPDH and of CgDAPDH lies in the d-center residues Met152 (StDAPDH) and Gln150 (CgDAPDH), respectively. Gln150 in CgDAPDH was mutated to leucine in Vedha-Peters’ second round of random mutagenesis, and the mutation was reported to be important for acceptance of smaller amino acid substrates.[2a] Our mutagenesis results also showed that the effect
Scheme 1. Reaction catalyzed by DAPDH.
specific to meso-DAP.[1] DAPDHs are the only class of d-amino acid dehydrogenases to have been used as biocatalysts for the reductive amination of a-keto acids to synthesize d-amino acids,[2] which are important building blocks of various fine chemicals such as pharmaceuticals and food ingredients.[3] Crystal structures of the DAPDH from Corynebacterium glutamicum (CgDAPDH; PDB IDs: 1DAP,[4] 2DAP, 3DAP,[5] and 1F06[6]) reveal the binding mode of meso-DAP, in which the l-center and d-center of meso-DAP can be distinguished by DAPDH and the enzyme catalyzes reaction only at the d-center. The l-center of meso-DAP corresponds to the R group of other [a] Dr. W. Liu, Prof. Dr. R.-T. Guo, Dr. X. Chen, Z. Li, Dr. X. Gao, Dr. C.-H. Huang, Prof. Dr. Q. Wu, Dr. J. Feng, Prof. Dr. D. Zhu Industrial Enzymes National Engineering Laboratory Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences 32 Xi Qi Dao, Tianjin Airport Economic Area, Tianjin 300308 (China) E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201402632.
ChemBioChem 2015, 16, 924 – 929
924
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers of methionine in this position was the same as that of leucine.[9] All this information implies that the mode of binding of small amino acid substrates by CgDAPDH/UtDAPDH mutants might be similar to that by StDAPDH. However, there has been no direct evidence to confirm this, because no structural information for these CgDAPDH or UtDAPDH mutants is available. Here we report a new DAPDH from Clostridium tetani E88 (CtDAPDH). Like most DAPDH family members, it exhibits high activity towards meso-DAP, but low activity to other d-amino acids. Mutagenesis of this enzyme in a similar way to that reported for CgDAPDH and UtDAPDH[2a, b] also resulted in a similar enhancement of activity for the synthesis of other d-amino acids and a decrease in activity toward the native substrate. Protein crystals of the CtDAPDH mutant were obtained, and the structures in apo form and in complex with meso-DAP, with d-leucine (d-leu), and with 4-methyl-2-oxopentanoic acid (MOPA) were solved. The new substrate binding information from the mutant CtDAPDH gave some insights into how these mutations affect the substrate specificities of DAPDHs.
Table 1. Specific activities of CtDAPDH and its mutant.
Reductive amination phenylpyruvic acid pyruvic acid 4-methyl-2-oxopentanoic acid Oxidative deamination meso-DAP
0.112 0.009 0.705 0.035 0.211 0.015
1.036 0.026
0.0018 0.0001
towards the native substrate meso-DAP, and improved the reductive amination activity towards a-keto acids. The kinetic parameters of CtDAPDH and its mutant toward d-alanine (oxidative deamination) and pyruvate (reductive amination) were determined as described before[7] (Table S2). The mutant showed higher Km and kcat values, so the improved activity was the result of the increase in the latter. The kinetic parameters toward phenylpyruvic acid, 4-methyl-2-oxopentanoic acid, and meso-DAP were also measured, but those for their reverse reactions could not be determined, due to low activity. Phenylpyruvic acid, pyruvic acid, and 4-methyl-2-oxopentanoic acid were reductively aminated to produce d-phenylalanine, d-alanine, and d-leucine, respectively, each with 99 ee (chiral HPLC analyses of these are presented in Figure S2, S3, and S4). Therefore, the obtained mutant enzyme of CtDAPDH exhibited a substrate scope change similar to that displayed by CgDAPDH[2a] and UtDAPDH mutants,[2b, c] and so could be a useful biocatalyst for the synthesis of d-amino acids. The CtDAPDH mutant maintained most of its activity after 2 h at 40 8C, and had lost about half of its activity after 1 h at 55 8C. The activity was completely lost after incubation at 60 8C for 30 min (Figure S5). As such, its thermostability is better than that of the reported CgDAPDH mutant,[2a] but worse than that of UtDAPDH mutant.[2b,c]
Mutagenesis of CtDAPDH The putative meso-diaminopimelate dehydrogenase from C. tetani E88 (GenBank accession number AAO36992.1) shows 46 and 66 % protein sequence identity with CgDAPDH and with UtDAPDH, respectively. The mutagenesis of CgDAPDH[2a] (Gln150Leu, Asp154Gly, Thr169Ile, Arg195Met, and His245Asn) and of UtDAPDH[2b, c] (Gln154Leu, Asp158Gly, Thr173Ile, Arg199Met, and His249Asn) has been reported to broaden the substrate ranges of these enzymes. The corresponding residues in CtDAPDH are Gln154, Asp158, Thr173, Arg199, and His249. In this study, we focused on the activity-center-related residues of CtDAPDH. Asp158 was thus omitted in the mutagenesis because it was assumed to be located far away from the activity center, and two additional mutations—Pro248Ser and Asn276Ser—within the CtDAPDH activity center were added; they correspond to Pro244 and Asn271 in CgDAPDH, respectively, which have been reported to be important for the enzyme activity.[10] A mutant CtDAPDH with the mutations Gln154Leu, Thr173Ile, Arg199Met, Pro248Ser, His249Asn, and Asn276Ser was therefore created. The most frequently observed residues at these positions are the same as those before mutagenesis (hot-spot analysis results are presented in Table S1 in the Supporting Information), and the protein sequence alignment of the three DAPDHs with the mutation sites are shown in Figure S1.
Structures of the CtDAPDH mutant in general In an effort to gain understanding of the structural basis for the substrate scope change in these DAPDH mutant enzymes, the protein crystals of CtDAPDH mutant in apo form and in complex with the wild-type substrate meso-DAP and the new substrates MOPA and d-leu were diffracted to 2.05, 2.0, 2.0, and 2.25 æ, respectively. Their structures were all solved by molecular replacement; the Rwork/Rfree ratios of the final models were 19.7/25.7, 19.9/24.6, 20.5/26.4, and 20.6 %/26.1 %, respectively. The PDB IDs for CtDAPDH mutant in apo form and in complex with meso-DAP, MOPA, and d-leu are 3WGO, 3WGQ, 3WGY, and 3WGZ, respectively, and the statistics of data collection and refinements are presented in Table S3. For all the four structures, there are two monomers (chains A and B) in the asymmetric unit. PISA analysis[11] shows that all CtDAPDH mutant structures should assemble into dimers in solution. Size-exclusion chromatography showed that the molecular weight was about 73.5 kDa; the calculated MW of CtDAPDH monomer is 36.3 kDa, thus confirming the dimeric
Characterization of CtDAPDH and its mutant The genes for CtDAPDH and for its mutant were expressed in Escherichia coli, and the recombinant proteins were purified by the procedure presented in the Experimental Section. Their specific activities towards meso-DAP and some a-keto acids are listed in Table 1. The activity results showed that the mutagenesis significantly lowered the oxidative deamination activity www.chembiochem.org
n.d.[a] 0.219 0.013 0.022 0.001
[a] n.d.: not detected.
Results and Discussion
ChemBioChem 2015, 16, 924 – 929
Specific activity [U mg¢1] WT CtDAPDH CtDAPDH mutant
Substrate
925
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers How mutagenesis of DAPDH affect the substrate specificity The domains of DAPDH family members are observed to move significantly on binding the co-enzyme or substrate, or both co-enzyme and substrate, leading to three different structural states: apo form, open form, and closed form.[4–6, 9] In the case of StDAPDH, the apo form is the state in which either nothing or only substrate binds in the structure, and the domain angle is around 348. When only co-enzyme binds, the domain angle opens up to around 398 to the open form, and when both coenzyme and substrate bind together, the structure changes to the closed form, with a domain angle of around 278. CtDAPDH is more similar to CgDAPDH than to StDAPDH (the protein sequence identities of CtDAPDH to CgDAPDH and to StDAPDH are 46 and 28.9 %, respectively). Because the overall structures of all four CtDAPDH mutant structures are similar, the structure of CtDAPDH mutant in complex with MOPA (3WGY) was used for comparison. Its chain A fitted best with the apo-form CgDAPDH (2DAP: A; the RMSD of the two structures was 0.927 æ, Figure 2 A) whereas its chain B superimposed well with the open-form CgDAPDH (1DAP: A; the RMSD of the two structures was 0.826 æ, Figure 2 B). This indicates that chain A of the CtDAPDH mutant should be the apo form and chain B the open form. The catalytic cavity of wild-type CtDAPDH should be similar to that of CgDAPDH, on the basis of their sequence similarity and catalytic characters. In order to explore how the mutagenesis affected the catalytic cavity, we compared the CtDAPDH mutant structure with that of CgDAPDH. The catalytic cavity of CtDAPDH mutant (3WGY: A) is composed of the following residues, with their counterparts in CgDAPDH (2DAP: A) in parentheses afterwards: Trp123 (Trp119), Trp148 (Trp144), Leu154 (Gln150), His156 (His152), Ile173 (Thr169), Met199 (Arg195),
Figure 1. Comparison of CtDAPDH mutant structures. A) Monomer of CtDAPDH mutant and its domains, B) assembly of CtDAPDH mutant monomer into dimer, C) dimer of CtDAPDH mutant bound with MOPA, D) comparison of chain A and chain B of CtDAPDH mutant bound with MOPA, E) comparison of chains A of all four CtDAPDH mutant structures, and F) comparison of chains B of all four CtDAPDH mutant structures.
aggregation of CtDAPDH in solution. Like CgDAPDH and StDAPDH, the monomer can be divided into the following three domains (Figure 1 A): the dinucleotide-binding domain (residues 1–122), the substrate-binding domain (residues 123– 247), and the C-terminal domain (residues 248–326) containing some antiparallel b-sheets and contributing to the linking of two adjacent monomers. The dimeric assembly is presented in Figure 1 B. Superimposition of the two monomers in the same asymmetric unit showed that their domain angles differed from one another. The monomer with the smaller domain angle was assigned as chain A, and the other one, with the larger domain angle, as chain B (Figure 1 C and D; the RMSD of apo-form chain A vs. chain B is 0.371 æ). The domain angles of chain A from the four CtDAPDH structures are similar to each other (Figure 1 E; the RMSDs of the chains A of CtDAPDH in complex with meso-DAP, MOPA, and d-leu vs. the apo-form structure are 0.080, 0.121, and 0.158 æ, respectively). Similar domain angles were found with the chains B of the four CtDAPDH structures (Figure 1 F; the RMSDs of the chains B of CtDAPDH in complex with meso-DAP, MOPA, and d-leu vs. the apo-form structure are 0.096, 0.104, and 0.163 æ, respectively). As such, the overall appearances of all four structures are similar; the differences between the pairs of subunits are similar to that in the reported CgDAPDH structure (PDB ID: 1DAP).[4] ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
Figure 2. Comparison of CtDAPDH mutant structures with CgDAPDH structures. A) Comparison of CtDAPDH mutant chain A (3WGY: A in green) with its counterpart in CgDAPDH (2DAP: A in brown), B) Comparison of CtDAPDH mutant chain B (3WGY: B in light blue) with its counterpart in CgDAPDH (1DAP: A in purple). C) Stereoview of activity centers of CtDAPDH mutant and CgDAPDH: the apo-form CtDAPDH mutant (3WGY: A) is in green and the apo-form CgDAPDH (2DAP: A) with bound meso-DAP is in brown.
926
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Asn249 (His244), and Ser276 (Asn270; stereoview in Figure 2 C). The comparison results show that the overall folding of the activity center is not changed much by the mutagenesis and that the size of its catalytic cavity is large enough for the binding of meso-DAP, although we do not find meso-DAP binding in the activity cavity of the CtDAPDH mutant. Of the residues mutated, it seems that the l-center-related mutations Thr173Ile and Arg199Met make major contributions to the lack of binding of meso-DAP in the catalytic cavity. meso-DAP, the native substrate for the wild-type DAPDH, was found in the catalytic cavities of both the CgDAPDH and the StDAPDH crystal structures. In the mutant CtDAPDH structures, meso-DAP was not found in the catalytic cavity, but in another area far away from the co-enzyme binding site, in which no reaction should be able to take place (Figure 3 A). The major neighboring residues of bound meso-DAP in the CtDAPDH mutant are His96, Asp124, Arg132, Thr144, His145, and Thr146. These residues are not mutated, so this binding mode should have little relation with the mutagenesis and could exist in wild-type CtDAPDH. Structural comparison with CgDAPDH and StDAPDH structures show that the corresponding binding area might also exist in those structures: the corresponding neighboring residues in CgDAPDH (StDAPDH) are His92 (His94), Asp120 (Asp122), Arg128 (Asp130), Thr140 (Thr142), His141 (Tyr142), and Thr142 (Thr144; Figure 3 B). The size and hydrophobicity similarity of these cavities imply that meso-DAP might also be able to bind in these areas in those structures. The binding areas are relatively peripheral in DAPDH and more exposed to solution during the catalytic procedure. As such, it is reasonable to expect that meso-DAP from solution might bind here first and then move deeper into the catalytic cavity for catalysis, resulting in the high activity of the wild-type DAPDHs toward meso-DAP. The mutant CtDAPDH has almost lost its activity toward meso-DAP because the mutations have changed the catalytic cavity from hydrophilic to hydrophobic, preventing the movement of the substrate from the outer binding site to the catalytic cavity, as evidenced by the mutant enzyme structure with meso-DAP. Similarly, in the first round of Vedha-Peters’ mutagenesis on l-center-related sites of CgDAPDH,[2a] a mutant (Thr170Ile, Arg196Met, His245Asn) active toward d-lysine was obtained, but it no
longer exhibited significant activity to native substrate mesoDAP. It appears likely that these mutated residues act together to form an environment that prohibits the binding of native substrate meso-DAP and its entering the catalytic cavity. Both MOPA and d-leu are bound at similar positions in the CtDAPDH mutant; their relative positions in the catalytic cavity are presented in Figure 4 A and B. An example of bound d-leu, together with its electron density map, is presented in Figure 4 C. The binding modes of MOPA and d-leu are similar, with the carboxyl group being bound to the main-chain N atoms of Leu154 and Gly155, and the hydrophobic R group bound with the side chain of Leu154. A comparison of this binding to that in open-form CgDAPDH and StDAPDH is presented in Figure 4 D.
Figure 4. Bound MOPA and d-leu in the CtDAPDH mutant. A) The relative positions of bound MOPA/d-leu in the CtDAPDH mutant. B) Enlarged view of bound MOPA/d-leu in the CtDAPDH mutant superimposed on the catalytic cavity with modeled meso-DAP and NADP; the modeled meso-DAP and NADP are in yellow, whereas MOPA, d-leu, and related residues are in purple, blue, and green respectively. C) Bound d-leu with electron density map in the CtDAPDH mutant (3WGZ: B); the Fo¢Fc omit map is contoured at 2 s (red) and 1.5 s (gray), the main chain and electron density map are in gray, d-leu is in blue, and related residues are in green. D) Comparison of dleu-binding-related area in open-form CtDAPDH (3WGZ: B-related residues in green, the ligand is in blue), in apo-form CgDAPDH (1DAP: A-related residues in brown), and in apo-form StDAPDH (3WBB: A-related residues in carmine).
Although the actual binding positions of MOPA/d-leu are not exactly the same as our previously observed binding positions of pyruvate/d-alanine in the StDAPDH structure,[9] they follow the same binding mode. The substrate carboxyl group is bound to the main-chain nitrogen in the same relative position, and the R group is bound through hydrophobic interactions with the side chain of leucine or methionine. Although Leu154 of the CtDAPDH mutant is buried inside the protein in the apo form or the closed form, its exposure to solution in the open form facilitates the binding of MOPA or d-leu and their entrance into the catalytic cavity. This position corre-
Figure 3. The relative location of the meso-DAP binding position in DAPDH. A) Actual binding of meso-DAP in the CtDAPDH mutant (3WGQ: A), the Fo¢Fc omit map is contoured at 2 s (gray). B) Comparison of the meso-DAP binding area in the apo-form CtDAPDH mutant (3WGQ:A, green) with the corresponding area in the apo-form CgDAPDH (2DAP:A, deep yellow) and the apo-form StDAPDH(3WB9: A, carmine).
ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
927
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Experimental Section
sponds to Met152 in StDAPDH, which moves significantly during domain movement and is believed to be the gate for entry of small substrates.[9] In the second round of VedhaPeters’ mutagenesis of CgDAPDH,[2a] a mutant with the additional mutation Gln150Leu was obtained by screening, and this mutant was reported to accept other amino acids as substrates. This residue of CgDAPDH corresponds to Leu154 in the CtDAPDH mutant and to Met152 in StDAPDH. The structural analysis of StDAPDH[9] and the CtDAPDH mutant deciphered the effect of this position in binding and the entrance of small amino acid substrates, thus providing some insights into the molecular basis of these mutant enzymes exhibiting expanded substrate specificity. Vedha-Peters et al. also reported a further enhancement of activity toward d-amino acid substrates by Asp155Gly in their third round of mutagenesis.[2a] This aspartic acid is far away from the activity center and is closer to the d-center than the l-center; it should therefore have no direct effect on catalysis. Its lies at the outer side of the substrate entrance tunnel for small d-amino acids, and its mutagenesis to glycine might open up the substrate entrance tunnel size, facilitate the entrance of amino acid substrates, and thus improve the enzyme activity. As well as the positions mentioned above, two other mutations—Pro248Ser and Asn276Ser—were also introduced in this research. Although they both lie close to the catalytic cavity, no direct evidence linking them and catalysis or substrate binding was discovered. They might still affect the protein folding or domain motion in some way.
Expression and purification of CtDAPDH and its mutant: The expression, purification, and activity assay procedures both for CtDAPDH and for its Q154L/T173I/R199M/P248S/H249N/N276S mutant were as follows: Ctdapdh and Ctdapdh-mutant genes were codon-optimized against E. coli and synthesized by Shanghai Xuguan Biotechnological Development Co. Both genes were inserted into the NcoI/XhoI restriction sites in pET32a and expressed in E. coli BL21(DE3) with His6-tag, S-tag, and enterokinase sites at its N terminus. The expression strains harboring the expression plasmid were inoculated in lysogeny broth (1000 mL) at 37 8C and 200 rpm. When the OD600 reached around 0.7, the temperature was lowered to 25 8C, and IPTG (0.5 mm) was added. The culture was induced for another 10 h. Harvested cells were resuspended in Tris·Cl buffer (20 mm, H 8.5) containing NaCl (500 mm), and homogenized in a French press. The crude extract was filtered through a 0.22 mm filter and loaded on a Ni-NTA column pre-equilibrated with the same buffer. The column was washed with Tris·Cl buffer (20 mm, pH 8.5) containing NaCl (500 mm) and imidazole (50 mm), and the target protein was eluted with Tris·Cl buffer (20 mm, pH 8.5) containing NaCl (500 mm) and imidazole (250 mm). The eluted protein was concentrated and transferred to Tris·Cl buffer (10 mm, pH 8.5) containing NaCl (50 mm) with a Millipore Ultra centrifugal Filter (cut-off molecular weight 10 kDa). N-terminal tags were removed by digesting with enterokinase (NEB), and the target NADPH protein was obtained by passage through a second Ni-NTA column. The final protein was concentrated to around 30 mg mL¢1 for activity assay, and transferred to Tris·Cl buffer (20 mm, pH 8.5) containing NaCl (50 mm) and glycerol (5 %) for further crystallization experiments. SDS-PAGE of purification procedures of CtDAPDH and its mutant is presented in Figure S6. Activity assay: Activity assays were carried out at 30 8C with 96well microplates, by spectrophotometric monitoring of the increase (oxidative deamination direction) or decrease (reductive amination direction) in absorbance of NADP(H) at 340 nm (the extinction coefficient of NADPH was 6.22 mm¢1 cm¢1). For reductive amination, the reaction mixture (200 mL) contained NH4Cl (200 mm), a-keto acid (25 mm), and NADPH (0.2 mm) in sodium carbonate buffer (100 mm, pH 9.0), and an appropriate amount of enzyme was added to initiate the reaction. For oxidative deamination, the reaction mixture (200 mL) contained d-amino acid (25 mm) and NADP + (0.5 mm) in sodium carbonate buffer (100 mm, pH 9.5), and an appropriate amount of enzyme was added to initiate the reaction. One unit (U) of enzyme was defined as the amount of enzyme producing or consuming 1 mmol NADPH per minute, and an optical pathway length of 0.5 cm was used for activity calculation.
Conclusions A meso-diaminopimelate dehydrogenase (DAPDH) from C. tetani E88 (CtDAPDH) was expressed in E. coli, and the recombinant enzyme showed little activity toward d-amino acids other than its native substrate, meso-DAP. Site-directed mutagenesis on the residues, similar to that previously reported for CgDAPDH and UtDAPDH, resulted in a mutant enzyme with almost no activity toward its native substrate. The mutant did, however, show high activity for the reductive amination of some a-keto acids to d-amino acids. The crystal structure of the CtDAPDH mutant with meso-DAP revealed an new substrate-binding area close to the protein surface; this bound meso-DAP prior to its entering into the catalytic binding site, thus suggesting a possible two-step substrate-binding mechanism. The carboxyl groups both of MOPA and of d-leu are bound to the main-chain N atoms of Leu154 and Gly155, whereas their hydrophobic R groups both interact with the side chain of Leu154. The special binding mode of d-leu and MOPA with the CtDAPDH mutant is similar to the previously observed binding mode of pyruvate/d-alanine with StDAPDH. These results reveal the structural basis for the substrate specificity change of these DAPDH mutant enzymes and should provide useful guidance for the development of highly active DADPHs for the synthesis of target d-amino acids by rational protein engineering. ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
Reductive amination of a-keto acids: a-Keto acid substrate (0.1 mmol, phenylpyruvic acid, pyruvic acid, or 4-methyl-2-oxopentanoic acid), NH4Cl (0.3 mmol), glucose (80 mg), GDH (0.2 mg), NADP + (0.5 mg), and a suitable amount of purified mutant CtDAPDH enzyme were mixed in Na2CO3/NaHCO3 buffer (200 mm, pH 9.0, 1 mL). The reaction mixture was incubated at 37 8C for 24 h, and the reaction was stopped by adding HClO4 solution (30 mL). After centrifugation, the supernatants were filtered through a 0.22 mm filter and analyzed by chiral HPLC. For phenyl+) alanine, the analysis was performed with a Crownpak CR(+ column (Daicel Chemical Industries, Japan), isocratic elution (0.5 mL min¢1) of perchloric acid solution (pH 1.5), and UV detection at 200 nm; the retention times of d-phenylalanine and l-phenylalanine were 10.6 min and 13.1 min, respectively. For leucine and alanine, the analysis was performed on a C18 column with iso-
928
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Acknowledgements
cratic elution of phosphoric acid/triethylamine (50 mm, pH 3.0, solvent A) and with acetonitrile (solvent B) and UV detection at 340 nm after derivatization with N-a-(5-fluoro-2,4-dinitrophenyl)-lalaninamide (FDAA). FDAA derivatizations were performed according to the FDAA instruction manual (Thermo Fisher Scientific). For leucine, the mobile phase was A/B 60:40 at 0.5 mL min¢1, and the retention times of FDAA derivatives of d-leucine and l-leucine were 8.8 min and 8.4 min, respectively; For alanine, the mobile phase was A/B 70:30 at 0.6 mL min¢1, and the retention times of FDAA derivatives of d-alanine and l-alanine were 14.8 min and 9.5 min, respectively.
This work was financially supported by the National Natural Science Foundation of China (No. 21072151 and No. 21102100) and the National Basic Research Program of China (973 Program, No. 2011CB710800). We thank the National Synchrotron Radiation Research Center of Taiwan for allocating beam time and assistance with the data collection. Keywords: amination · dehydrogenases · enzyme catalysis · protein structures · structural biology
Crystallization and data collection: Crystallization conditions for the CtDAPDH mutant were screened by the sitting-drop vapor diffusion method. The reservoir solution (1 mL) was mixed with protein solution (1 mL) and evaporated against reservoir solution of Hampton Research Crystal Screen Kits (100 mL). meso-DAP, d-leucine, or 4-methyl-2-oxopentanoic acid (MOPA) were each added to a final concentration of 10 mm in co-crystallization experiments. All CtDAPDH mutant crystals were grown under the same conditions at 298 K with a protein concentration of 10 mg mL¢1; the best crystallization conditions were with 2-(morpholin-4-yl)ethanesulfonic acid (MES; 0.1 m, pH 6.0), PEG 20 000 (15 %, w/v), and glycerol (5 %, v/v). The crystals were cryo-protected by increasing the glycerol concentration to a final concentration of 20 % before data collection. All diffraction data were collected at beamline BL15A1 of the National Synchrotron Radiation Research Center (NSRRC; Hsinchu, Taiwan), and the diffraction data were processed by use of the program HKL3000.[12]
[1] a) S. G. Reddy, G. Scapin, J. S. Blanchard, Proteins Struct. Funct. Bioinf. 1996, 25, 514 – 516; b) H. Misono, H. Togawa, T. Yamamoto, K. Soda, J. Bacteriol. 1979, 137, 22 – 27; c) H. Misono, K. Soda, J. Biol. Chem. 1980, 255, 599 – 605; d) S. Ishino, T. Mizukami, K. Yamaguchi, R. Katsumata, K. Araki, Agric. Biol. Chem. 1988, 52, 2903 – 2909. [2] a) K. Vedha-Peters, M. Gunawardana, J. D. Rozzell, S. J. Novick, J. Am. Chem. Soc. 2006, 128, 10923 – 10929; b) H. Akita, H. Suzuki, K. Doi, T. Ohshima, Appl. Microbiol. Biotechnol. 2014, 98, 1135 – 1143; c) H. Akita, K. Doi, Y. Kawarabayasi, T. Ohshima, Biotechnol. Lett. 2012, 34, 1693 – 1699. [3] a) M. Friedman, C. E. Levin, Amino Acids 2012, 42, 1553 – 1582; b) M. Friedman, Chem. Biodiversity 2010, 7, 1491 – 1530; c) M. Friedman, J. Agric. Food Chem. 1999, 47, 3457 – 3479. [4] G. Scapin, S. G. Reddy, J. S. Blanchard, Biochemistry 1996, 35, 13540 – 13551. [5] G. Scapin, M. Cirilli, S. G. Reddy, Y. Gao, J. C. Vederas, J. S. Blanchard, Biochemistry 1998, 37, 3278 – 3285. [6] M. Cirilli, G. Scapin, A. Sutherland, J. C. Vederas, J. S. Blanchard, Protein Sci. 2000, 9, 2034 – 2037. [7] X. Gao, X. Chen, W. Liu, J. Feng, Q. Wu, L. Hua, D. Zhu, Appl. Environ. Microbiol. 2012, 78, 8595 – 8600. [8] X. Gao, F. Huang, J. Feng, X. Chen, H. Zhang, Z. Wang, Q. Wu, D. Zhu, Appl. Environ. Microbiol. 2013, 79, 5078 – 5081. [9] W. Liu, Z. Li, C. H. Huang, R. T. Guo, L. Zhao, D. Zhang, X. Chen, Q. Wu, D. Zhu, ChemBioChem 2014, 15, 217 – 222. [10] D. J. Rozzell, S. J. Novick (Biocatalytics, Inc.), US 7550277, 2009. [11] E. Krissinel, K. Henrick, J. Mol. Biol. 2007, 372, 774 – 797. [12] W. Minor, M. Cymborowski, Z. Otwinowski, M. Chruszcz, Acta Crystallogr. Sect. D Biol. Crystallogr. 2006, 62, 859 – 866. [13] I. J. Tickle, R. A. Laskowski, D. S. Moss, Acta Crystallogr. Sect. D Biol. Crystallogr. 1998, 54, 547 – 557. [14] F. Long, A. A. Vagin, P. Young, G. N. Murshudov, Acta Crystallogr. Sect. D Biol. Crystallogr. 2008, 64, 125 – 132. [15] G. N. Murshudov, P. Skubak, A. A. Lebedev, N. S. Pannu, R. A. Steiner, R. A. Nicholls, M. D. Winn, F. Long, A. A. Vagin, Acta Crystallogr. Sect. D Biol. Crystallogr. 2011, 67, 355 – 367. [16] G. Langer, S. X. Cohen, V. S. Lamzin, A. Perrakis, Nat. Protoc. 2008, 3, 1171 – 1179. [17] J. E. Debreczeni, P. Emsley, Acta Crystallogr. Sect. D Biol. Crystallogr. 2012, 68, 425 – 430. [18] A. W. Schìttelkopf, D. M. van Aalten, Acta Crystallogr. Sect. D Biol. Crystallogr. 2004, 60, 1355 – 1363. [19] A. T. Brìnger, P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, G. L. Warren, Acta Crystallogr. Sect. D Biol. Crystallogr. 1998, 54, 905 – 921. [20] A. Pavelka, E. Chovancova, J. Damborsky, Nucleic Acids Res. 2009, 37, W376—383. [21] The UniProt Consortium, Nucleic Acids Res. 2012, 40, D71 – D75.
Structure determination: The crystals all belong to the P21 space group with two monomers per asymmetric unit, and 5 % reflections were randomly selected for calculating Rfree as a monitor[13] prior to structure refinements. The crystal structures of the apo form and in complex with meso-DAP, MOPA, or d-leu were all determined by molecular replacement with BALBES,[14] and the best results were obtained with the structure of CgDAPDH (PDB ID: 1F06)—which has 46 % sequence identity with CtDAPDH—as template. The solution was then refined with REFMAC5.[15] The resulting phases were input into ARP/wARP[16] for refinement and initial model building. Further manual model adjustment and addition of water and ligands were carried out with the 2Fo¢Fc map contoured at 1.0 s in COOT.[17] The topologies of the ligands and small molecules were generated from the PRODRG server,[18] and the subsequent refinements were carried out by use of CNS.[19] The biological assembly of CtDAPDH structures was analyzed by PISA server with final structures[11] (Protein Interfaces, Surfaces and Assemblies). All figures were prepared with PyMOL (http://www.pymol.org/). Structure comparisons: Structural comparisons were performed by superimposing each CtDAPDH mutant structure monomer against each monomer in the CgDAPDH and StDAPDH structures. The structures with the lowest RMSD values were considered to be counterparts. Catalytic cavity comparison was done by comparing chain A of apo-form CtDAPDH with chain A of the apo-form mesoDAP-bound CgDAPDH structure, and the residues less than 5 æ from the meso-DAP substrate were taken into consideration. The CtDAPDH mutant structure was submitted into the Hot-Spot Wizard server[20] for the residue conservation analysis of related sites, and the variance data were from Uniprot[21] (http://www. uniprot.org).
ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
Received: October 30, 2014 Published online on March 6, 2015
929
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers
Structural Analysis Reveals the Substrate-Binding Mechanism for the Expanded Substrate Specificity of Mutant meso-Diaminopimelate Dehydrogenase Weidong Liu, Rey-Ting Guo, Xi Chen, Zhe Li, Xiuzhen Gao, Chun-Hsiang Huang, Qiaqing Wu, Jinhui Feng, and Dunming Zhu*[a] late (meso-DAP), d-leucine (d-leu), and 4-methyl-2-oxopentanoic acid (MOPA) were solved. meso-DAP was found in an area outside the catalytic cavity; this suggested a possible two-step substrate-binding mechanism for meso-DAP. d-leu and MOPA each bound both to Leu154 and to Gly155 in the open form of CtDAPDH, and structural analysis revealed the molecular basis for the expanded substrate specificity of the mutant meso-diaminopimelate dehydrogenases.
A meso-diaminopimelate dehydrogenase (DAPDH) from Clostridium tetani E88 (CtDAPDH) was found to have low activity toward the d-amino acids other than its native substrate. Sitedirected mutagenesis similar to that carried out on the residues mutated by Vedha-Peters et al. resulted in a mutant enzyme with highly improved catalytic ability for the synthesis of d-amino acids. The crystal structures of the CtDAPDH mutant in apo form and in complex with meso-diaminopime-
Introduction Meso-diaminopimelate dehydrogenase (DAPDH, EC.1.4.1.16) catalyzes reversible oxidative deamination/reductive amination reactions at the d-center of meso-2,6-diaminopimelate (mesoDAP, Scheme 1), and most of the known wild-type DAPDHs are
d-amino acids, thus a change in the l-center-binding site of DAPDH should expand its substrate spectrum towards other damino acids. Working on the basis of the structure information of CgDAPDH, Vedha-Peters et al.[2a] first successfully improved the activity of CgDAPDH to other d-amino acids through three rounds of directed evolution, with the first round of mutagenesis targeting those residues that interact with the substituents on the l-carbon in the native substrate meso-DAP and the other two rounds targeting the whole enzyme. Recently, Akita et al.[2b, c] reported similar results for the mutation of the corresponding amino acid residues of a DAPDH from Ureibacillus thermosphaericus (UtDAPDH). In these studies, the mutated positions of DAPDH are equivalent and include residues in the binding sites both of the l-center and of the d-center. The l-center residues are Thr169 (Thr173), Arg195 (Arg199), and His245 (His249) of CgDAADH (UtDAPDH), whereas the d-center residue is Gln150 (Gln154) of CgDAPDH (UtDAPDH). We recently discovered a thermostable DAPDH from Symbiobacterium thermophilum (StDAPDH) with significant activity towards other d-amino acids as well as meso-DAP,[7] and its substrate specificity was expanded to include larger amino acids such as phenylalanine by mutagenesis on its l-center residues.[8] The structural and mutagenesis studies[9] revealed that a new substrate entrance tunnel, along with residue Met152, was important for the binding of small substrates such as pyruvate/d-alanine. Structural comparison showed that the major difference between the catalytic cavity residues of StDAPDH and of CgDAPDH lies in the d-center residues Met152 (StDAPDH) and Gln150 (CgDAPDH), respectively. Gln150 in CgDAPDH was mutated to leucine in Vedha-Peters’ second round of random mutagenesis, and the mutation was reported to be important for acceptance of smaller amino acid substrates.[2a] Our mutagenesis results also showed that the effect
Scheme 1. Reaction catalyzed by DAPDH.
specific to meso-DAP.[1] DAPDHs are the only class of d-amino acid dehydrogenases to have been used as biocatalysts for the reductive amination of a-keto acids to synthesize d-amino acids,[2] which are important building blocks of various fine chemicals such as pharmaceuticals and food ingredients.[3] Crystal structures of the DAPDH from Corynebacterium glutamicum (CgDAPDH; PDB IDs: 1DAP,[4] 2DAP, 3DAP,[5] and 1F06[6]) reveal the binding mode of meso-DAP, in which the l-center and d-center of meso-DAP can be distinguished by DAPDH and the enzyme catalyzes reaction only at the d-center. The l-center of meso-DAP corresponds to the R group of other [a] Dr. W. Liu, Prof. Dr. R.-T. Guo, Dr. X. Chen, Z. Li, Dr. X. Gao, Dr. C.-H. Huang, Prof. Dr. Q. Wu, Dr. J. Feng, Prof. Dr. D. Zhu Industrial Enzymes National Engineering Laboratory Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences 32 Xi Qi Dao, Tianjin Airport Economic Area, Tianjin 300308 (China) E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cbic.201402632.
ChemBioChem 2015, 16, 924 – 929
924
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers of methionine in this position was the same as that of leucine.[9] All this information implies that the mode of binding of small amino acid substrates by CgDAPDH/UtDAPDH mutants might be similar to that by StDAPDH. However, there has been no direct evidence to confirm this, because no structural information for these CgDAPDH or UtDAPDH mutants is available. Here we report a new DAPDH from Clostridium tetani E88 (CtDAPDH). Like most DAPDH family members, it exhibits high activity towards meso-DAP, but low activity to other d-amino acids. Mutagenesis of this enzyme in a similar way to that reported for CgDAPDH and UtDAPDH[2a, b] also resulted in a similar enhancement of activity for the synthesis of other d-amino acids and a decrease in activity toward the native substrate. Protein crystals of the CtDAPDH mutant were obtained, and the structures in apo form and in complex with meso-DAP, with d-leucine (d-leu), and with 4-methyl-2-oxopentanoic acid (MOPA) were solved. The new substrate binding information from the mutant CtDAPDH gave some insights into how these mutations affect the substrate specificities of DAPDHs.
Table 1. Specific activities of CtDAPDH and its mutant.
Reductive amination phenylpyruvic acid pyruvic acid 4-methyl-2-oxopentanoic acid Oxidative deamination meso-DAP
0.112 0.009 0.705 0.035 0.211 0.015
1.036 0.026
0.0018 0.0001
towards the native substrate meso-DAP, and improved the reductive amination activity towards a-keto acids. The kinetic parameters of CtDAPDH and its mutant toward d-alanine (oxidative deamination) and pyruvate (reductive amination) were determined as described before[7] (Table S2). The mutant showed higher Km and kcat values, so the improved activity was the result of the increase in the latter. The kinetic parameters toward phenylpyruvic acid, 4-methyl-2-oxopentanoic acid, and meso-DAP were also measured, but those for their reverse reactions could not be determined, due to low activity. Phenylpyruvic acid, pyruvic acid, and 4-methyl-2-oxopentanoic acid were reductively aminated to produce d-phenylalanine, d-alanine, and d-leucine, respectively, each with 99 ee (chiral HPLC analyses of these are presented in Figure S2, S3, and S4). Therefore, the obtained mutant enzyme of CtDAPDH exhibited a substrate scope change similar to that displayed by CgDAPDH[2a] and UtDAPDH mutants,[2b, c] and so could be a useful biocatalyst for the synthesis of d-amino acids. The CtDAPDH mutant maintained most of its activity after 2 h at 40 8C, and had lost about half of its activity after 1 h at 55 8C. The activity was completely lost after incubation at 60 8C for 30 min (Figure S5). As such, its thermostability is better than that of the reported CgDAPDH mutant,[2a] but worse than that of UtDAPDH mutant.[2b,c]
Mutagenesis of CtDAPDH The putative meso-diaminopimelate dehydrogenase from C. tetani E88 (GenBank accession number AAO36992.1) shows 46 and 66 % protein sequence identity with CgDAPDH and with UtDAPDH, respectively. The mutagenesis of CgDAPDH[2a] (Gln150Leu, Asp154Gly, Thr169Ile, Arg195Met, and His245Asn) and of UtDAPDH[2b, c] (Gln154Leu, Asp158Gly, Thr173Ile, Arg199Met, and His249Asn) has been reported to broaden the substrate ranges of these enzymes. The corresponding residues in CtDAPDH are Gln154, Asp158, Thr173, Arg199, and His249. In this study, we focused on the activity-center-related residues of CtDAPDH. Asp158 was thus omitted in the mutagenesis because it was assumed to be located far away from the activity center, and two additional mutations—Pro248Ser and Asn276Ser—within the CtDAPDH activity center were added; they correspond to Pro244 and Asn271 in CgDAPDH, respectively, which have been reported to be important for the enzyme activity.[10] A mutant CtDAPDH with the mutations Gln154Leu, Thr173Ile, Arg199Met, Pro248Ser, His249Asn, and Asn276Ser was therefore created. The most frequently observed residues at these positions are the same as those before mutagenesis (hot-spot analysis results are presented in Table S1 in the Supporting Information), and the protein sequence alignment of the three DAPDHs with the mutation sites are shown in Figure S1.
Structures of the CtDAPDH mutant in general In an effort to gain understanding of the structural basis for the substrate scope change in these DAPDH mutant enzymes, the protein crystals of CtDAPDH mutant in apo form and in complex with the wild-type substrate meso-DAP and the new substrates MOPA and d-leu were diffracted to 2.05, 2.0, 2.0, and 2.25 æ, respectively. Their structures were all solved by molecular replacement; the Rwork/Rfree ratios of the final models were 19.7/25.7, 19.9/24.6, 20.5/26.4, and 20.6 %/26.1 %, respectively. The PDB IDs for CtDAPDH mutant in apo form and in complex with meso-DAP, MOPA, and d-leu are 3WGO, 3WGQ, 3WGY, and 3WGZ, respectively, and the statistics of data collection and refinements are presented in Table S3. For all the four structures, there are two monomers (chains A and B) in the asymmetric unit. PISA analysis[11] shows that all CtDAPDH mutant structures should assemble into dimers in solution. Size-exclusion chromatography showed that the molecular weight was about 73.5 kDa; the calculated MW of CtDAPDH monomer is 36.3 kDa, thus confirming the dimeric
Characterization of CtDAPDH and its mutant The genes for CtDAPDH and for its mutant were expressed in Escherichia coli, and the recombinant proteins were purified by the procedure presented in the Experimental Section. Their specific activities towards meso-DAP and some a-keto acids are listed in Table 1. The activity results showed that the mutagenesis significantly lowered the oxidative deamination activity www.chembiochem.org
n.d.[a] 0.219 0.013 0.022 0.001
[a] n.d.: not detected.
Results and Discussion
ChemBioChem 2015, 16, 924 – 929
Specific activity [U mg¢1] WT CtDAPDH CtDAPDH mutant
Substrate
925
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers How mutagenesis of DAPDH affect the substrate specificity The domains of DAPDH family members are observed to move significantly on binding the co-enzyme or substrate, or both co-enzyme and substrate, leading to three different structural states: apo form, open form, and closed form.[4–6, 9] In the case of StDAPDH, the apo form is the state in which either nothing or only substrate binds in the structure, and the domain angle is around 348. When only co-enzyme binds, the domain angle opens up to around 398 to the open form, and when both coenzyme and substrate bind together, the structure changes to the closed form, with a domain angle of around 278. CtDAPDH is more similar to CgDAPDH than to StDAPDH (the protein sequence identities of CtDAPDH to CgDAPDH and to StDAPDH are 46 and 28.9 %, respectively). Because the overall structures of all four CtDAPDH mutant structures are similar, the structure of CtDAPDH mutant in complex with MOPA (3WGY) was used for comparison. Its chain A fitted best with the apo-form CgDAPDH (2DAP: A; the RMSD of the two structures was 0.927 æ, Figure 2 A) whereas its chain B superimposed well with the open-form CgDAPDH (1DAP: A; the RMSD of the two structures was 0.826 æ, Figure 2 B). This indicates that chain A of the CtDAPDH mutant should be the apo form and chain B the open form. The catalytic cavity of wild-type CtDAPDH should be similar to that of CgDAPDH, on the basis of their sequence similarity and catalytic characters. In order to explore how the mutagenesis affected the catalytic cavity, we compared the CtDAPDH mutant structure with that of CgDAPDH. The catalytic cavity of CtDAPDH mutant (3WGY: A) is composed of the following residues, with their counterparts in CgDAPDH (2DAP: A) in parentheses afterwards: Trp123 (Trp119), Trp148 (Trp144), Leu154 (Gln150), His156 (His152), Ile173 (Thr169), Met199 (Arg195),
Figure 1. Comparison of CtDAPDH mutant structures. A) Monomer of CtDAPDH mutant and its domains, B) assembly of CtDAPDH mutant monomer into dimer, C) dimer of CtDAPDH mutant bound with MOPA, D) comparison of chain A and chain B of CtDAPDH mutant bound with MOPA, E) comparison of chains A of all four CtDAPDH mutant structures, and F) comparison of chains B of all four CtDAPDH mutant structures.
aggregation of CtDAPDH in solution. Like CgDAPDH and StDAPDH, the monomer can be divided into the following three domains (Figure 1 A): the dinucleotide-binding domain (residues 1–122), the substrate-binding domain (residues 123– 247), and the C-terminal domain (residues 248–326) containing some antiparallel b-sheets and contributing to the linking of two adjacent monomers. The dimeric assembly is presented in Figure 1 B. Superimposition of the two monomers in the same asymmetric unit showed that their domain angles differed from one another. The monomer with the smaller domain angle was assigned as chain A, and the other one, with the larger domain angle, as chain B (Figure 1 C and D; the RMSD of apo-form chain A vs. chain B is 0.371 æ). The domain angles of chain A from the four CtDAPDH structures are similar to each other (Figure 1 E; the RMSDs of the chains A of CtDAPDH in complex with meso-DAP, MOPA, and d-leu vs. the apo-form structure are 0.080, 0.121, and 0.158 æ, respectively). Similar domain angles were found with the chains B of the four CtDAPDH structures (Figure 1 F; the RMSDs of the chains B of CtDAPDH in complex with meso-DAP, MOPA, and d-leu vs. the apo-form structure are 0.096, 0.104, and 0.163 æ, respectively). As such, the overall appearances of all four structures are similar; the differences between the pairs of subunits are similar to that in the reported CgDAPDH structure (PDB ID: 1DAP).[4] ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
Figure 2. Comparison of CtDAPDH mutant structures with CgDAPDH structures. A) Comparison of CtDAPDH mutant chain A (3WGY: A in green) with its counterpart in CgDAPDH (2DAP: A in brown), B) Comparison of CtDAPDH mutant chain B (3WGY: B in light blue) with its counterpart in CgDAPDH (1DAP: A in purple). C) Stereoview of activity centers of CtDAPDH mutant and CgDAPDH: the apo-form CtDAPDH mutant (3WGY: A) is in green and the apo-form CgDAPDH (2DAP: A) with bound meso-DAP is in brown.
926
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Asn249 (His244), and Ser276 (Asn270; stereoview in Figure 2 C). The comparison results show that the overall folding of the activity center is not changed much by the mutagenesis and that the size of its catalytic cavity is large enough for the binding of meso-DAP, although we do not find meso-DAP binding in the activity cavity of the CtDAPDH mutant. Of the residues mutated, it seems that the l-center-related mutations Thr173Ile and Arg199Met make major contributions to the lack of binding of meso-DAP in the catalytic cavity. meso-DAP, the native substrate for the wild-type DAPDH, was found in the catalytic cavities of both the CgDAPDH and the StDAPDH crystal structures. In the mutant CtDAPDH structures, meso-DAP was not found in the catalytic cavity, but in another area far away from the co-enzyme binding site, in which no reaction should be able to take place (Figure 3 A). The major neighboring residues of bound meso-DAP in the CtDAPDH mutant are His96, Asp124, Arg132, Thr144, His145, and Thr146. These residues are not mutated, so this binding mode should have little relation with the mutagenesis and could exist in wild-type CtDAPDH. Structural comparison with CgDAPDH and StDAPDH structures show that the corresponding binding area might also exist in those structures: the corresponding neighboring residues in CgDAPDH (StDAPDH) are His92 (His94), Asp120 (Asp122), Arg128 (Asp130), Thr140 (Thr142), His141 (Tyr142), and Thr142 (Thr144; Figure 3 B). The size and hydrophobicity similarity of these cavities imply that meso-DAP might also be able to bind in these areas in those structures. The binding areas are relatively peripheral in DAPDH and more exposed to solution during the catalytic procedure. As such, it is reasonable to expect that meso-DAP from solution might bind here first and then move deeper into the catalytic cavity for catalysis, resulting in the high activity of the wild-type DAPDHs toward meso-DAP. The mutant CtDAPDH has almost lost its activity toward meso-DAP because the mutations have changed the catalytic cavity from hydrophilic to hydrophobic, preventing the movement of the substrate from the outer binding site to the catalytic cavity, as evidenced by the mutant enzyme structure with meso-DAP. Similarly, in the first round of Vedha-Peters’ mutagenesis on l-center-related sites of CgDAPDH,[2a] a mutant (Thr170Ile, Arg196Met, His245Asn) active toward d-lysine was obtained, but it no
longer exhibited significant activity to native substrate mesoDAP. It appears likely that these mutated residues act together to form an environment that prohibits the binding of native substrate meso-DAP and its entering the catalytic cavity. Both MOPA and d-leu are bound at similar positions in the CtDAPDH mutant; their relative positions in the catalytic cavity are presented in Figure 4 A and B. An example of bound d-leu, together with its electron density map, is presented in Figure 4 C. The binding modes of MOPA and d-leu are similar, with the carboxyl group being bound to the main-chain N atoms of Leu154 and Gly155, and the hydrophobic R group bound with the side chain of Leu154. A comparison of this binding to that in open-form CgDAPDH and StDAPDH is presented in Figure 4 D.
Figure 4. Bound MOPA and d-leu in the CtDAPDH mutant. A) The relative positions of bound MOPA/d-leu in the CtDAPDH mutant. B) Enlarged view of bound MOPA/d-leu in the CtDAPDH mutant superimposed on the catalytic cavity with modeled meso-DAP and NADP; the modeled meso-DAP and NADP are in yellow, whereas MOPA, d-leu, and related residues are in purple, blue, and green respectively. C) Bound d-leu with electron density map in the CtDAPDH mutant (3WGZ: B); the Fo¢Fc omit map is contoured at 2 s (red) and 1.5 s (gray), the main chain and electron density map are in gray, d-leu is in blue, and related residues are in green. D) Comparison of dleu-binding-related area in open-form CtDAPDH (3WGZ: B-related residues in green, the ligand is in blue), in apo-form CgDAPDH (1DAP: A-related residues in brown), and in apo-form StDAPDH (3WBB: A-related residues in carmine).
Although the actual binding positions of MOPA/d-leu are not exactly the same as our previously observed binding positions of pyruvate/d-alanine in the StDAPDH structure,[9] they follow the same binding mode. The substrate carboxyl group is bound to the main-chain nitrogen in the same relative position, and the R group is bound through hydrophobic interactions with the side chain of leucine or methionine. Although Leu154 of the CtDAPDH mutant is buried inside the protein in the apo form or the closed form, its exposure to solution in the open form facilitates the binding of MOPA or d-leu and their entrance into the catalytic cavity. This position corre-
Figure 3. The relative location of the meso-DAP binding position in DAPDH. A) Actual binding of meso-DAP in the CtDAPDH mutant (3WGQ: A), the Fo¢Fc omit map is contoured at 2 s (gray). B) Comparison of the meso-DAP binding area in the apo-form CtDAPDH mutant (3WGQ:A, green) with the corresponding area in the apo-form CgDAPDH (2DAP:A, deep yellow) and the apo-form StDAPDH(3WB9: A, carmine).
ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
927
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Experimental Section
sponds to Met152 in StDAPDH, which moves significantly during domain movement and is believed to be the gate for entry of small substrates.[9] In the second round of VedhaPeters’ mutagenesis of CgDAPDH,[2a] a mutant with the additional mutation Gln150Leu was obtained by screening, and this mutant was reported to accept other amino acids as substrates. This residue of CgDAPDH corresponds to Leu154 in the CtDAPDH mutant and to Met152 in StDAPDH. The structural analysis of StDAPDH[9] and the CtDAPDH mutant deciphered the effect of this position in binding and the entrance of small amino acid substrates, thus providing some insights into the molecular basis of these mutant enzymes exhibiting expanded substrate specificity. Vedha-Peters et al. also reported a further enhancement of activity toward d-amino acid substrates by Asp155Gly in their third round of mutagenesis.[2a] This aspartic acid is far away from the activity center and is closer to the d-center than the l-center; it should therefore have no direct effect on catalysis. Its lies at the outer side of the substrate entrance tunnel for small d-amino acids, and its mutagenesis to glycine might open up the substrate entrance tunnel size, facilitate the entrance of amino acid substrates, and thus improve the enzyme activity. As well as the positions mentioned above, two other mutations—Pro248Ser and Asn276Ser—were also introduced in this research. Although they both lie close to the catalytic cavity, no direct evidence linking them and catalysis or substrate binding was discovered. They might still affect the protein folding or domain motion in some way.
Expression and purification of CtDAPDH and its mutant: The expression, purification, and activity assay procedures both for CtDAPDH and for its Q154L/T173I/R199M/P248S/H249N/N276S mutant were as follows: Ctdapdh and Ctdapdh-mutant genes were codon-optimized against E. coli and synthesized by Shanghai Xuguan Biotechnological Development Co. Both genes were inserted into the NcoI/XhoI restriction sites in pET32a and expressed in E. coli BL21(DE3) with His6-tag, S-tag, and enterokinase sites at its N terminus. The expression strains harboring the expression plasmid were inoculated in lysogeny broth (1000 mL) at 37 8C and 200 rpm. When the OD600 reached around 0.7, the temperature was lowered to 25 8C, and IPTG (0.5 mm) was added. The culture was induced for another 10 h. Harvested cells were resuspended in Tris·Cl buffer (20 mm, H 8.5) containing NaCl (500 mm), and homogenized in a French press. The crude extract was filtered through a 0.22 mm filter and loaded on a Ni-NTA column pre-equilibrated with the same buffer. The column was washed with Tris·Cl buffer (20 mm, pH 8.5) containing NaCl (500 mm) and imidazole (50 mm), and the target protein was eluted with Tris·Cl buffer (20 mm, pH 8.5) containing NaCl (500 mm) and imidazole (250 mm). The eluted protein was concentrated and transferred to Tris·Cl buffer (10 mm, pH 8.5) containing NaCl (50 mm) with a Millipore Ultra centrifugal Filter (cut-off molecular weight 10 kDa). N-terminal tags were removed by digesting with enterokinase (NEB), and the target NADPH protein was obtained by passage through a second Ni-NTA column. The final protein was concentrated to around 30 mg mL¢1 for activity assay, and transferred to Tris·Cl buffer (20 mm, pH 8.5) containing NaCl (50 mm) and glycerol (5 %) for further crystallization experiments. SDS-PAGE of purification procedures of CtDAPDH and its mutant is presented in Figure S6. Activity assay: Activity assays were carried out at 30 8C with 96well microplates, by spectrophotometric monitoring of the increase (oxidative deamination direction) or decrease (reductive amination direction) in absorbance of NADP(H) at 340 nm (the extinction coefficient of NADPH was 6.22 mm¢1 cm¢1). For reductive amination, the reaction mixture (200 mL) contained NH4Cl (200 mm), a-keto acid (25 mm), and NADPH (0.2 mm) in sodium carbonate buffer (100 mm, pH 9.0), and an appropriate amount of enzyme was added to initiate the reaction. For oxidative deamination, the reaction mixture (200 mL) contained d-amino acid (25 mm) and NADP + (0.5 mm) in sodium carbonate buffer (100 mm, pH 9.5), and an appropriate amount of enzyme was added to initiate the reaction. One unit (U) of enzyme was defined as the amount of enzyme producing or consuming 1 mmol NADPH per minute, and an optical pathway length of 0.5 cm was used for activity calculation.
Conclusions A meso-diaminopimelate dehydrogenase (DAPDH) from C. tetani E88 (CtDAPDH) was expressed in E. coli, and the recombinant enzyme showed little activity toward d-amino acids other than its native substrate, meso-DAP. Site-directed mutagenesis on the residues, similar to that previously reported for CgDAPDH and UtDAPDH, resulted in a mutant enzyme with almost no activity toward its native substrate. The mutant did, however, show high activity for the reductive amination of some a-keto acids to d-amino acids. The crystal structure of the CtDAPDH mutant with meso-DAP revealed an new substrate-binding area close to the protein surface; this bound meso-DAP prior to its entering into the catalytic binding site, thus suggesting a possible two-step substrate-binding mechanism. The carboxyl groups both of MOPA and of d-leu are bound to the main-chain N atoms of Leu154 and Gly155, whereas their hydrophobic R groups both interact with the side chain of Leu154. The special binding mode of d-leu and MOPA with the CtDAPDH mutant is similar to the previously observed binding mode of pyruvate/d-alanine with StDAPDH. These results reveal the structural basis for the substrate specificity change of these DAPDH mutant enzymes and should provide useful guidance for the development of highly active DADPHs for the synthesis of target d-amino acids by rational protein engineering. ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
Reductive amination of a-keto acids: a-Keto acid substrate (0.1 mmol, phenylpyruvic acid, pyruvic acid, or 4-methyl-2-oxopentanoic acid), NH4Cl (0.3 mmol), glucose (80 mg), GDH (0.2 mg), NADP + (0.5 mg), and a suitable amount of purified mutant CtDAPDH enzyme were mixed in Na2CO3/NaHCO3 buffer (200 mm, pH 9.0, 1 mL). The reaction mixture was incubated at 37 8C for 24 h, and the reaction was stopped by adding HClO4 solution (30 mL). After centrifugation, the supernatants were filtered through a 0.22 mm filter and analyzed by chiral HPLC. For phenyl+) alanine, the analysis was performed with a Crownpak CR(+ column (Daicel Chemical Industries, Japan), isocratic elution (0.5 mL min¢1) of perchloric acid solution (pH 1.5), and UV detection at 200 nm; the retention times of d-phenylalanine and l-phenylalanine were 10.6 min and 13.1 min, respectively. For leucine and alanine, the analysis was performed on a C18 column with iso-
928
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Full Papers Acknowledgements
cratic elution of phosphoric acid/triethylamine (50 mm, pH 3.0, solvent A) and with acetonitrile (solvent B) and UV detection at 340 nm after derivatization with N-a-(5-fluoro-2,4-dinitrophenyl)-lalaninamide (FDAA). FDAA derivatizations were performed according to the FDAA instruction manual (Thermo Fisher Scientific). For leucine, the mobile phase was A/B 60:40 at 0.5 mL min¢1, and the retention times of FDAA derivatives of d-leucine and l-leucine were 8.8 min and 8.4 min, respectively; For alanine, the mobile phase was A/B 70:30 at 0.6 mL min¢1, and the retention times of FDAA derivatives of d-alanine and l-alanine were 14.8 min and 9.5 min, respectively.
This work was financially supported by the National Natural Science Foundation of China (No. 21072151 and No. 21102100) and the National Basic Research Program of China (973 Program, No. 2011CB710800). We thank the National Synchrotron Radiation Research Center of Taiwan for allocating beam time and assistance with the data collection. Keywords: amination · dehydrogenases · enzyme catalysis · protein structures · structural biology
Crystallization and data collection: Crystallization conditions for the CtDAPDH mutant were screened by the sitting-drop vapor diffusion method. The reservoir solution (1 mL) was mixed with protein solution (1 mL) and evaporated against reservoir solution of Hampton Research Crystal Screen Kits (100 mL). meso-DAP, d-leucine, or 4-methyl-2-oxopentanoic acid (MOPA) were each added to a final concentration of 10 mm in co-crystallization experiments. All CtDAPDH mutant crystals were grown under the same conditions at 298 K with a protein concentration of 10 mg mL¢1; the best crystallization conditions were with 2-(morpholin-4-yl)ethanesulfonic acid (MES; 0.1 m, pH 6.0), PEG 20 000 (15 %, w/v), and glycerol (5 %, v/v). The crystals were cryo-protected by increasing the glycerol concentration to a final concentration of 20 % before data collection. All diffraction data were collected at beamline BL15A1 of the National Synchrotron Radiation Research Center (NSRRC; Hsinchu, Taiwan), and the diffraction data were processed by use of the program HKL3000.[12]
[1] a) S. G. Reddy, G. Scapin, J. S. Blanchard, Proteins Struct. Funct. Bioinf. 1996, 25, 514 – 516; b) H. Misono, H. Togawa, T. Yamamoto, K. Soda, J. Bacteriol. 1979, 137, 22 – 27; c) H. Misono, K. Soda, J. Biol. Chem. 1980, 255, 599 – 605; d) S. Ishino, T. Mizukami, K. Yamaguchi, R. Katsumata, K. Araki, Agric. Biol. Chem. 1988, 52, 2903 – 2909. [2] a) K. Vedha-Peters, M. Gunawardana, J. D. Rozzell, S. J. Novick, J. Am. Chem. Soc. 2006, 128, 10923 – 10929; b) H. Akita, H. Suzuki, K. Doi, T. Ohshima, Appl. Microbiol. Biotechnol. 2014, 98, 1135 – 1143; c) H. Akita, K. Doi, Y. Kawarabayasi, T. Ohshima, Biotechnol. Lett. 2012, 34, 1693 – 1699. [3] a) M. Friedman, C. E. Levin, Amino Acids 2012, 42, 1553 – 1582; b) M. Friedman, Chem. Biodiversity 2010, 7, 1491 – 1530; c) M. Friedman, J. Agric. Food Chem. 1999, 47, 3457 – 3479. [4] G. Scapin, S. G. Reddy, J. S. Blanchard, Biochemistry 1996, 35, 13540 – 13551. [5] G. Scapin, M. Cirilli, S. G. Reddy, Y. Gao, J. C. Vederas, J. S. Blanchard, Biochemistry 1998, 37, 3278 – 3285. [6] M. Cirilli, G. Scapin, A. Sutherland, J. C. Vederas, J. S. Blanchard, Protein Sci. 2000, 9, 2034 – 2037. [7] X. Gao, X. Chen, W. Liu, J. Feng, Q. Wu, L. Hua, D. Zhu, Appl. Environ. Microbiol. 2012, 78, 8595 – 8600. [8] X. Gao, F. Huang, J. Feng, X. Chen, H. Zhang, Z. Wang, Q. Wu, D. Zhu, Appl. Environ. Microbiol. 2013, 79, 5078 – 5081. [9] W. Liu, Z. Li, C. H. Huang, R. T. Guo, L. Zhao, D. Zhang, X. Chen, Q. Wu, D. Zhu, ChemBioChem 2014, 15, 217 – 222. [10] D. J. Rozzell, S. J. Novick (Biocatalytics, Inc.), US 7550277, 2009. [11] E. Krissinel, K. Henrick, J. Mol. Biol. 2007, 372, 774 – 797. [12] W. Minor, M. Cymborowski, Z. Otwinowski, M. Chruszcz, Acta Crystallogr. Sect. D Biol. Crystallogr. 2006, 62, 859 – 866. [13] I. J. Tickle, R. A. Laskowski, D. S. Moss, Acta Crystallogr. Sect. D Biol. Crystallogr. 1998, 54, 547 – 557. [14] F. Long, A. A. Vagin, P. Young, G. N. Murshudov, Acta Crystallogr. Sect. D Biol. Crystallogr. 2008, 64, 125 – 132. [15] G. N. Murshudov, P. Skubak, A. A. Lebedev, N. S. Pannu, R. A. Steiner, R. A. Nicholls, M. D. Winn, F. Long, A. A. Vagin, Acta Crystallogr. Sect. D Biol. Crystallogr. 2011, 67, 355 – 367. [16] G. Langer, S. X. Cohen, V. S. Lamzin, A. Perrakis, Nat. Protoc. 2008, 3, 1171 – 1179. [17] J. E. Debreczeni, P. Emsley, Acta Crystallogr. Sect. D Biol. Crystallogr. 2012, 68, 425 – 430. [18] A. W. Schìttelkopf, D. M. van Aalten, Acta Crystallogr. Sect. D Biol. Crystallogr. 2004, 60, 1355 – 1363. [19] A. T. Brìnger, P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve, J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, G. L. Warren, Acta Crystallogr. Sect. D Biol. Crystallogr. 1998, 54, 905 – 921. [20] A. Pavelka, E. Chovancova, J. Damborsky, Nucleic Acids Res. 2009, 37, W376—383. [21] The UniProt Consortium, Nucleic Acids Res. 2012, 40, D71 – D75.
Structure determination: The crystals all belong to the P21 space group with two monomers per asymmetric unit, and 5 % reflections were randomly selected for calculating Rfree as a monitor[13] prior to structure refinements. The crystal structures of the apo form and in complex with meso-DAP, MOPA, or d-leu were all determined by molecular replacement with BALBES,[14] and the best results were obtained with the structure of CgDAPDH (PDB ID: 1F06)—which has 46 % sequence identity with CtDAPDH—as template. The solution was then refined with REFMAC5.[15] The resulting phases were input into ARP/wARP[16] for refinement and initial model building. Further manual model adjustment and addition of water and ligands were carried out with the 2Fo¢Fc map contoured at 1.0 s in COOT.[17] The topologies of the ligands and small molecules were generated from the PRODRG server,[18] and the subsequent refinements were carried out by use of CNS.[19] The biological assembly of CtDAPDH structures was analyzed by PISA server with final structures[11] (Protein Interfaces, Surfaces and Assemblies). All figures were prepared with PyMOL (http://www.pymol.org/). Structure comparisons: Structural comparisons were performed by superimposing each CtDAPDH mutant structure monomer against each monomer in the CgDAPDH and StDAPDH structures. The structures with the lowest RMSD values were considered to be counterparts. Catalytic cavity comparison was done by comparing chain A of apo-form CtDAPDH with chain A of the apo-form mesoDAP-bound CgDAPDH structure, and the residues less than 5 æ from the meso-DAP substrate were taken into consideration. The CtDAPDH mutant structure was submitted into the Hot-Spot Wizard server[20] for the residue conservation analysis of related sites, and the variance data were from Uniprot[21] (http://www. uniprot.org).
ChemBioChem 2015, 16, 924 – 929
www.chembiochem.org
Received: October 30, 2014 Published online on March 6, 2015
929
Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
