d-2-Hydroxyglutarate dehydrogenase plays a dual role in l-serine biosynthesis and d-malate utilization in the bacterium Pseudomonas stutzeri

Pseudomonas is a very large bacterial genus in which several species can use d-malate for growth. However, the enzymes that can metabolize d-malate, such as d-malate dehydrogenase, appear to be absent in most Pseudomonas species. d-3-Phosphoglycerate dehydrogenase (SerA) can catalyze the production of d-2-hydroxyglutarate (d-2-HG) from 2-ketoglutarate to support d-3-phosphoglycerate dehydrogenation, which is the initial reaction in bacterial l-serine biosynthesis. In this study, we show that SerA of the Pseudomonas stutzeri strain A1501 reduces oxaloacetate to d-malate and that d-2-HG dehydrogenase (D2HGDH) from P. stutzeri displays d-malate–oxidizing activity. Of note, D2HGDH participates in converting a trace amount of d-malate to oxaloacetate during bacterial l-serine biosynthesis. Moreover, D2HGDH is crucial for the utilization of d-malate as the sole carbon source for growth of P. stutzeri A1501. We also found that the D2HGDH expression is induced by the exogenously added d-2-HG or d-malate and that a flavoprotein functions as a soluble electron carrier between D2HGDH and electron transport chains to support d-malate utilization by P. stutzeri. These results support the idea that D2HGDH evolves as an enzyme for both d-malate and d-2-HG dehydrogenation in P. stutzeri. In summary, D2HGDH from P. stutzeri A1501 participates in both a core metabolic pathway for l-serine biosynthesis and utilization of extracellular d-malate.

As shown in Fig. 1A, when active SerA was added in the reaction mixture containing OAA and NADH, biotransformation resulted in the production of a new compound that had a retention time of 14.22 min, which corresponded to the peak of authentic malate. Malate exists in two stereoisomeric forms: L-malate and D-malate. As shown in Fig. 1B, D-malate was efficiently catalyzed by D-malate dehydrogenase from E. coli, whereas L-malate was not, under the same assay condition. When reaction products of the SerA were added in the assay mixture, high D-malate dehydrogenase activity was detected (Fig. 1B), indicating that the product obtained from OAA by SerA was not L-malate, but D-malate.

D2HGDH has D-malate-oxidizing activity
The D2HGDHs in P. stutzeri, Saccharomyces cerevisiae, and Arabidopsis thaliana catalyze the conversion of D-2-HG to 2-KG (Table 2) (4,21,22). The N-terminal His-tagged D2HGDH in P. stutzeri A1501 was purified by nickel-affinity chromatography as described previously (4). The dehydrogenase activity of D2HGDH in the presence of the artificial electron acceptor 2,6-dichlorophenol-indophenol (DCIP) on dif-ferent 2-hydroxy acids was assayed. D2HGDH showed high activities with D-2-HG and D-malate (Table 2) but no detectable activity on L-2-hydroxyglutarate, D-lactate, and other 2-hydroxy acids. The rates of dehydrogenation of substrates catalyzed by the D2HGDH followed Michaelis-Menten kinetics. Double-reciprocal plots of the initial rates plotted against the concentrations of D-2-HG and D-malate yielded K m values of 0.17 Ϯ 0.02 and 3.61 Ϯ 0.14 mM, respectively. k cat was estimated to be 7.90 Ϯ 1.05 and 11.70 Ϯ 0.39 s Ϫ1 , respectively, with DCIP as the electron acceptor ( Table 2).
The prosthetic group of D2HGDH was previously experimentally confirmed to be FAD (22). The UV-visible absorbance spectra of D2HGDH displayed maxima centered at 378 and 450 nm, consistent with the presence of a flavin cofactor bound to the protein ( Fig. 2A). Robust and persistent decreases in the absorbance at 378 and 450 nm were detected when the reaction mixtures containing D2HGDH, D-2-HG, or D-malate were flushed with nitrogen prior to spectral analysis (Fig. 2, B and C). These observations showed that FAD bound to D2HGDH acts as an intermediate electron acceptor during the oxidation of D-2-HG and D-malate. The effect of metal ions on the activity of the D2HGDH toward D-2-HG and D-malate was tested by adding metal salts (final concentration of 10 M or 10 mM) to the assay buffer. As shown in Fig. 3 (A and B), Zn 2ϩ positively influenced the D2HGDH activity at a concentration of 10 M but inhibited D2HGDH activity at a concentration of 10 mM. Ca 2ϩ also positively affected D2HGDH activity at a concentration of 10 mM. Fe 2ϩ and Co 2ϩ inhibited D2HGDH activity at a concentration of 10 mM. EDTA partially inhibited D2HGDH activity. Thus, 10 M Zn 2ϩ was added in the reaction mixture used for assaying the activity of D2HGDH.

Dual role of D-2-hydroxyglutarate dehydrogenase
4 mM) and 2.5 mM D-malate, and the reaction mixtures were analyzed by HPLC after 40 min. D-Malate consumption correlated with the production of two compounds that had the same retention times as those of OAA and pyruvate (Fig. 4A), respectively. D-Malate dehydrogenase in E. coli catalyzes the ␤-decarboxylating dehydrogenation of D-malate to produce pyruvate, with OAA as a reaction intermediate. To identify whether D2HGDH also has OAA-decarboxylating activity, purified D2HGDH was incubated in the presence of 2.5 mM OAA. The reaction mixture with heat-inactivated D2HGDH was used as the control. As shown in Fig. 4B, pyruvate was detected with an identical amount in the reaction mixtures with both active and heat-inactivated D2HGDH. This result indicated that D2HGDH did not have OAA-decarboxylating activity. The detection of pyruvate in the reaction mixture might be due to the spontaneous decarboxylation of OAA. D2HGDH catalyzes the ␣-dehydrogenation of D-malate to OAA but not the ␤-decarboxylating dehydrogenation of D-malate to pyruvate.

D2HGDH couples with electron transfer flavoprotein (ETF) for efficient oxidization of D-malate
Multiple turnover of D2HGDH requires continuous electron transfer. In a previous work, we confirmed that ETF functions as a soluble electron carrier between D2HGDH and electron transport chains (4). P. stutzeri A1501 ETF was expressed, purified, and added in the reaction mixture containing D2HGDH, D-malate, and DCIP. As shown in Fig. 5, the addition of ETF strongly favored D-malate oxidation. As the ETF concentration increased, the V max values also increased continuously (Table  3). When ETF was considered as the substrate for D2HGDH catalyzed D-malate dehydrogenation, the K m and V max values were measured as 4.71 Ϯ 0.03 M and 2.71 Ϯ 0.03 units mg Ϫ1 , respectively ( Table 2). During the oxidation of D-malate catalyzed by D2HGDH, ETF could also play a role as the mediator of electron transfer.

D2HGDH is required for D-malate utilization of P. stutzeri A1501
The D2HGDH mutant strain showed high accumulation of extracellular D-2-HG, indicating a high flux toward 2-KG reduction in P. stutzeri A1501. Metabolomic analysis also indicated a 4.5-fold change of intracellular malate (4). However, the D2HGDH mutant did not have extracellular accumulation of D-malate. 2-KG and OAA could serve as substrates of SerA, but the intracellular concentration and k cat /K m for 2-KG were much higher than those of OAA (Table 1). That might be the reason why only extracellular D-2-HG can be detected in the D2HGDH mutant strain.
Given the high activity of D2HGDH detected on D-malate, we hypothesized that D2HGDH could be involved in D-malate utilization. P. stutzeri A1501 could use D-malate as the sole carbon source for growth. Compared with the P. stutzeri A1501 and A1501-⌬d2hgdh-d2hgdh ϩ , A1501-⌬d2hgdh exhibited little ability to assimilate D-malate (Fig. 6A), demonstrating that D2HGDH is critical in D-malate metabolism. The same result was obtained in a medium containing D-2-HG as the sole carbon source (Fig. 6B).
We also determined the in vivo role of ETF in D-malate catabolism. An ETF mutant abolished growth in a medium containing D-malate. Complementation of this ETF mutant with a plasmid (pBBR1MCS-5) containing etf (P. stutzeri A1501-⌬etf-etf ϩ ) restored growth in a medium containing D-malate (Fig. 6A). These results indicated that ETF can mediate the electron transfer process during D-malate catabolism.
The growth of P. stutzeri A1501, A1501-⌬d2hgdh-d2hgdh ϩ , A1501-⌬d2hgdh, A1501-⌬etf, and A1501-⌬etf-etf ϩ in media containing OAA or 2-KG was also assayed. As shown in Fig. 6 (C and D), all of the strains grew in the presence of OAA or 2-KG as the sole carbon source. The decreased biomass and growth rate of P. stutzeri A1501-⌬d2hgdh and A1501-⌬etf that was observed might be because the accumulation of

Dual role of D-2-hydroxyglutarate dehydrogenase
D-2-HG that occurs in these mutants could influence L-serine biosynthesis.

D2HGDH in P. stutzeri A1501 and D-malate dehydrogenase in E. coli are not mutually replaceable
D-Malate dehydrogenase of E. coli can catalyze the ␤-decarboxylating dehydrogenation of D-malate to produce pyruvate. Because D-2-HG is a structural homolog to D-malate, the activity of D-malate dehydrogenase in E. coli toward D-2-HG was also assayed. As shown in Fig. 7C, no increase of absorbance at 340 nm was detected when D-2-HG was added in a reaction mixture containing D-malate dehydrogenase and NAD ϩ . When D-malate dehydrogenase was expressed in P. stutzeri A1501-⌬d2hgdh, the strain P. stutzeri A1501-⌬d2hgdh-dmlA was still not able to utilize D-2-HG as the sole carbon source (Fig. 7A), and accumulation of D-2-HG during the utilization of glucose and other carbon sources was detected. Thus, the D-malate dehydrogenase in E. coli cannot compensate the role of D2HGDH in P. stutzeri A1501. D2HGDH in P. stutzeri A1501 catalyzes the ␣-dehydrogenation of D-malate and D-2-HG to produce OAA and 2-KG, respectively (Fig. 4, A and C). As shown in Fig. 7D, decreased absorbance at 600 nm was detected when D-malate or D-2-HG was added in a reaction mixture containing D2HGDH and DCIP, respectively. E. coli K-12 MG1655-⌬dmlA with the deletion of the D-malate dehydrogenase-encoding gene could not grow on a medium containing D-malate as the sole carbon source. When D2HGDH was expressed in E. coli K-12 MG1655-⌬dmlA, the strain P. stutzeri A1501-⌬dmlA-d2hgdh was still not able to utilize D-malate as the sole carbon source (Fig. 7B). This might have been because D2HGDH also requires ETF as its mediator for electron transfer. Thus, the D2HGDH in P. stutzeri A1501 cannot compensate the role of D-malate dehydrogenase in E. coli.

Dual role of D-2-hydroxyglutarate dehydrogenase D2HGDH is induced during D-malate utilization
The expression of D-malate dehydrogenase in E. coli is induced by D-malate (11). It can only complement leucine aux-otrophy in a ⌬leuB strain in media containing D-malate (12). D2HGDH in P. stutzeri A1501 is involved in L-serine biosynthesis, which is a crucial metabolic process. Presently, P. stutzeri A1501 was cultured with OAA, 2-KG, D-2-HG, D-malate, glucose, or pyruvate as the sole carbon source, and the specific activities of D2HGDH were examined. D2HGDH activities were detectable in cells grown in different types of media including glucose, pyruvate, OAA, and 2-KG at any growth period (Table 4). However, the specific activities were much higher in crude cell extract prepared from P. stutzeri A1501 grown in D-2-HG or D-malate than those in other carbon sources.
The effect of D-2-HG or D-malate on D2HGDH expression was also assessed by a ␤-gal assay. A 199-bp fragment upstream translation initiation site of d2hgdh was PCR-amplified from

Dual role of D-2-hydroxyglutarate dehydrogenase
P. stutzeri A1501 genomic DNA. The PCR product was ligated into the pME6522 to generate pME6522-P d2hgdh , which carried the P d2hgdh -lacZ transcriptional fusions. Much higher ␤-gal activities were observed when P. stutzeri A1501-pME6522-P d2hgdh was grown in the presence of D-2-HG or D-malate (Fig.  8). These results further confirmed that D2HGDH is induced during utilization of exogenous D-malate.

Discussion
D-2-HG is a five-carbon dicarboxylic acid. It often accumulates in patients with the rare neurometabolic disorder of D-2hydroxyglutaric aciduria (23)(24)(25). Mutation in isocitrate dehydrogenase with high 2-KG-reducing activity and accumulation of D-2-HG have also been detected in several types of cancer (26 -33). D-2-HG is a key effector that promotes oncogenesis and D-2-hydroxyglutaric aciduria. Therefore, D-2-HG is suspected to be an abnormal metabolite, and it requires D2HGDH, a "metabolite repair enzyme," for maintenance of a low level. In this study, we identified that the D2HGDH plays a dual role in L-serine biosynthesis and D-malate utilization of P. stutzeri. During the L-serine biosynthesis of P. stutzeri, D-2-HG is pro-duced to overcome the thermodynamic barrier of D-3-PG dehydrogenation. D2HGDH is constitutively expressed at a low level to convert D-2-HG to 2-KG, preventing the inhibitory effect of D-2-HG on SerA and maintaining sufficient pools of 2-KG for L-serine biosynthesis. SerA can also catalyze the reduction of OAA to D-malate with low efficiency. Thus, D2HGDH also plays a role in converting a trace amount of D-malate to OAA during L-serine biosynthesis. D-Malate is distributed in different habitats and can also be produced during the bacterial utilization of various chemicals, such as gentisate, maleate, and OAA. D-Malate utilization might be an important metabolic process of different bacteria. When D-malate is used as the sole carbon source, D2HGDH will be induced to a higher level, converting D-malate into OAA. OAA will then be utilized to support the growth of P. stutzeri. Because L-serine is also required during D-malate utilization, D2HGDH also plays a role in supporting L-serine biosynthesis (Fig. 9).
As a generalist enzyme, D-malate dehydrogenase in E. coli is also active in the leucine biosynthesis pathway by catalyzing isopropylmalate to 2-ketoisocaproate. It also has tartrate dehy-

Dual role of D-2-hydroxyglutarate dehydrogenase
drogenase activity, but E. coli could not use tartrate as a sole carbon source for growth, likely because of a lack of membrane transporters for tartrate (12). D2HGDH in P. stutzeri could not catalyze the dehydrogenation of tartrate or isopropylmalate. Actually, there is an isopropylmalate dehydrogenase-encoding gene in P. stutzeri A1501, and this strain could not use tartrate as a sole carbon source. D2HGDH will not participate in tartrate or isopropylmalate metabolism in P. stutzeri A1501. Tartrate dehydrogenase has been reported in some Pseudomonas strains, such as Pseudomonas putida (34). This enzyme also has D-malate dehydrogenation activity (35,36). The specific functions of tartrate dehydrogenase and D2HGDH in metabolism (especially D-malate utilization) of these strains with the two enzymes warrant detailed study.
All the reported D2HGDHs have both D-malate and D-2-HG dehydrogenation activities. This might be because D-malate and D-2-HG may be simultaneously produced by the same enzyme, such as SerA and PdxB (4, 7). The major task of D2HGDH is the conversion of D-malate and D-2-HG to OAA and 2-KG, respectively. Thus, D2HGDH is selected for D-malate and D-2-HG dehydrogenation and has been maintained as a generalist during evolution. Another piece of evidence supporting the selective evolution of the dual role of D2HGDH is that the expression of D2HGDH is induced by D-malate or D-2-HG.
In addition to SerA, PdxB in E. coli (a key enzyme in pyridoxine biosynthesis) also has 2-KG and OAA reduction activities (7). Thus, there should be a high carbon flux flowing through D-2-HG and D-malate during metabolism of E. coli. D-Malate  Table 4 The specific activities of D2HGDH in P. stutzeri A1501 during growth in medium containing different carbon sources The values are the means Ϯ S.D. (n ϭ 3). The initial concentration of carbon source is ϳ4 g liter Ϫ1 . a Activities were examined with 10 mM D-malate. b Activities were examined with 1 mM D-2-HG. DCIP was used as the electron acceptor.

Dual role of D-2-hydroxyglutarate dehydrogenase
dehydrogenase in E. coli can catalyze the dehydrogenation of D-malate, but not the dehydrogenation of D-2-HG. This is understandable because D-malate dehydrogenase is a ␤-decarboxylating dehydrogenase, whereas there is no ␤-hydroxy group in D-2-HG. Actually, there is no D2HGDH homolog in E. coli, indicating the existence of other unknown enzymes involved in the D-2-HG catabolism. D-2-HG synthase activity, which produces D-2-HG from propionyl-CoA and glyoxylate, has been described in E. coli (37,38). The splitting reaction catalyzed by D-2-HG synthase might participate in the D-2-HG metabolism of E. coli. This speculation could be validated when the genes associated with the reported bacterial D-2-HG synthase have been identified. In summary, the present investigation of D2HGDH provides a better understanding of this novel enzyme. It can accommodate D-malate and D-2-HG as substrates and utilize ETF as its electron mediator. D-2-HG has been viewed as an abnormal metabolite. D2HGDH-catalyzed D-2-HG conversion to 2-KG seems to be a process of metabolite repair, and D2HGDH is a metabolite repair enzyme, which maintains D-2-HG at low levels. Compared with other typical metabolite repair enzymes, the most distinctive feature of D2HGDH is the additional D-malate dehydrogenation activity, which is crucial for Dmalate utilization. To our knowledge, the present study provides the first experimental example of a metabolite repair enzyme that supports a core metabolic pathway (L-serine biosynthesis) and extracellular compound (D-malate) utilization in addition to its metabolite repair function (Fig. 9). To further understand the roles of metabolite repair enzymes, more information, especially their possible multiple functions, are required. MTT, isopropyl-␤-D-1-thiogalactopyranoside, and phenylmethanesulfonyl fluoride were purchased from Sigma. Yeast extract powder and tryptone were obtained from Oxoid Limited (Basingstoke, United Kingdom). Other chemicals were of analytical reagent grade.

Bacteria and culture conditions
Bacterial strains used in this study are listed in Table S1. The minimal medium for Pseudomonas was supplemented with different carbon sources as described previously (4). The P. stutzeri A1501 was cultured in minimal medium at 30°C, with shaking at 200 rpm, and the cell density (A 600 ) of the cultures was measured at various time points. E. coli strains were grown in LB medium at 37°C. Agarose (1.5%) was added to obtain solid medium. Antibiotics were used, when appropriate, at the following concentrations: ampicillin, 100 g ml Ϫ1 ; kanamycin, 50 g ml Ϫ1 ; and gentamycinsulfate, 30 g ml Ϫ1 .

Expression and purification of recombinant enzymes
The plasmids and primers used in this study are listed in Table S1 and S2. The D-malate dehydrogenase-encoding gene dmlA was PCR-amplified from E. coli K-12 MG1655 genome with the high-fidelity DNA polymerase TransStart FastPfu (TransGen) using primers PD1 and PD2 (Table S2). The PCR product was double-digested with NcoI and BamHI and then ligated into similarly treated pET-28a(ϩ) to form the expression construct pET-dmlA, in which a His 6 tag-encoding sequence was fused into dmlA. Its sequence was verified by DNA sequencing. pET-d2hgdh, pET-etfAB, and pET-serA plasmids were created as described previously (4). The obtained plasmids were transformed into BL21 (DE3) cells for protein expression.
The recombinant E. coli strains were cultured in LB medium (50 g ml Ϫ1 kanamycin) at 37°C to an A 600 of 0.5-0.7 and induced at 16°C with 1 mM isopropyl-␤-D-1-thiogalactopyranoside for 10 h. The cells were harvested and washed twice with buffer A (20 mM sodium phosphate, 500 mM NaCl, pH 7.4) by centrifugation at 6,000 rpm for 10 min at 4°C and then resuspended in the same buffer containing 1 mM phenylmethanesulfonyl fluoride and 10% glycerol (v/v). The cells were broken by ultrasonic (Sonics 500 W; 20 kHz) on ice and centrifuged at 12,000 rpm for 40 min at 4°C. The supernatant (the crude extract) was loaded onto a 5-ml HisTrap HP column (GE Healthcare Life Sciences) equilibrated with buffer A and washed with 10% elution buffer B (20 mM sodium phosphate, 500 mM imidazole, 500 mM NaCl, 10% glycerol, pH 7.4) to remove any weakly bound protein and then eluted with different gradients of elution buffer at a flow rate of 5 ml min Ϫ1 . The fractions containing target enzymes were concentrated by ultrafiltration and then stored at Ϫ80°C.

Spectral characterization of D2HGDH
The UV-visible absorbance spectra of the purified recombinant D2HGDH proteins were recorded in a UV-visible Ultrospec 2100 pro spectrophotometer (GE Healthcare) in 2-nm steps and with a slit width of 1 nm. The experiments were performed at 30°C in 1-ml anaerobic cuvettes closed

Dual role of D-2-hydroxyglutarate dehydrogenase
were kept under a constant flow of nitrogen for at least 5 min.

Enzymic assays
The D2HGDH activity was assayed by recording the reduction of DCIP spectrophotometrically at 600 nm in a reaction mixture containing 50 mM Tris-HCl buffer (pH 7.4), phenazinemethosulfate (200 M), 100 M DCIP, and variable concentrations of substrate at 30°C. The double-reciprocal plot method was used to estimate the kinetic constants of D2HGDH. One unit of D2HGDH activity was defined as the amount of enzyme that reduced 1 mol of DCIP/min.
The SerA activity was assayed by recording the decrease of NADH spectrophotometrically at 340 nm (⑀ ϭ 6,220 M Ϫ1 cm Ϫ1 ) in a reaction mixture containing 50 mM Tris-HCl buffer (pH 7.4), 0.2 mM NADH, and variable concentrations of substrate incubated at 30°C. One unit of enzyme activity was defined as the amount that oxidizes 1 mol of NADH/min. All the assays were carried out in quartz cuvettes (1-cm light path) filled with 0.8 ml of reaction mixture at using a UV-visible spectrophotometer (Ultrospec 2100 pro; Amersham Biosciences).

Reporter plasmid construction and ␤-gal assay
DNA fragments of 199 bp covering the intergenic regions upstream of the d2hgdh were synthesized by PCR using the following pairs of oligonucleotides 6522D1 and 6522D2 to generate pME6522-P d2hgdh (Table S2). The reporter plasmid pME6522-P d2hgdh was transferred into the WT strain P. stutzeri A1501 to generate P. stutzeri A1501-pME6522-P d2hgdh .
For measurement of ␤-gal activities, P. stutzeri A1501-pME6522-P d2hgdh was grown in the medium containing different carbon sources. The initial concentration of carbon source is ϳ4 g liter Ϫ1 . The recombinant strains were cultured to an A 600 of 1.5-1.8 and harvested 1ϳ2 ml by centrifugation at 8,000 rpm for 5 min. Bacteria were resuspended in 1 ml of Z buffer, and then 100 l of chloroform and 50 l of 0.1% (w/v) SDS were added. After mixing, the mixture was equilibrated at 30°C for 10 min. The lysates were centrifuged (4°C), and the super-natants were immediately used for ␤-gal activity measurements, basically as described by Miller (39) using onitrophenyl-␤-D-galactopyranoside as the substrate subsequent quenching of the reaction with Na 2 CO 3 .

The conformation identification of SerA-produced malate
The reaction mixture (1 ml) contained 50 mM Tris-HCl (pH 7.4), 10 mM NADH, 40 mM OAA, and 100 M SerA incubated at 30°C for 4 h. The purified DmlA was used to define the configuration of the SerA-produced malate. The reaction mixtures (0.8 ml) contained 0.2 mM NAD ϩ and 4 mg of DmlA in 50 mM Tris-HCl (pH 7.4) at 30°C. To initiate the reactions, 40 l of 1 mM authentic D-malate, 1 mM authentic L-malate, or SerA-produced malate was added, and the reactions were monitored at 340 nm using a UV-visible spectrophotometer.
Author contributions-X. G. and C. G. writing-original draft; M. Z. data curation; M. C. and Z. K. investigation; W. Z. methodology; P. X., C. M., and C. G. writing-review and editing.