Introduction

Many approaches have been developed to produce chiral chemicals for the manufacture of a wide range of intermediates in the pharmaceutical, agrochemical, fine chemical and food industries [1]. In recent years, biocatalysis has emerged as a powerful strategy for the production of enantiomerically pure building blocks that are difficult to be obtained through chemical and fermentation processes. Enantioselective α-hydroxyacid dehydrogenases (α-HADHs) can specifically oxidize one enantiomer of α-hydroxyacid to α-ketoacid, thus leaving the other enantiomer in reaction system. They can be used as catalysts for the production of optically active α-hydroxyacids from their racemic mixtures [2–5]. α-HADHs are promising for industrial application as they can carry out the oxidation at ambient pressure and temperature with high selectivity, activity and broad substrate specificity; most importantly, the isolation of chiral product from the reaction mixture becomes very simple and efficient, which are key characteristics for their development potential as industrial process [6]. Although the theoretical recovery of product is limited to 50 %, a two-step stereoinversion of racemic substrate to one single enantiomer through the asymmetric oxidation with α-HADH to the corresponding ketoacid followed by reduction with a highly stereo-complementary ketoacid reductase can be considered to overcome this disadvantage. The product can be obtained in 100 % yield and with 100 % enantiomeric excess (ee) [7–10].

Several microorganisms with enantioselective α-HADHs, including Alcaligenes bronchisepticus KU1201 [11], Pseudomonas polycolor IFO 3918 [12], Alcaligenes sp. ECU0401 [13] and P. putida ECU1009 [14], have been reported to catalyze the asymmetric oxidation of racemic α-hydroxyacids for the production of optical pure α-hydroxyacids. However, there is still a need to find strains with α-HADHs that are able to oxidize a broad spectrum of substrates with excellent enantiomeric purity and high activity. Generally employed analytical methods for screening enantioselective α-HADHs are HPLC, GC or LC–MS, which are not readily amenable to high-throughput [12, 14]. All these are very time consuming if the traditional methods of screening are employed. Hence, it is necessary to develop a rapid screening method to identify new biocatalysts with desired activity and selectivity. Up to date, there is no high-throughput screening method for the identification of enantioselective α-HADH, which makes the screening step as rate limiting. Therefore, a direct and sensitive readout of α-HADH activity and enantioselectivity has to be considered.

In this study, we describe a simple and rapid method for the rapid screening of a library of enantioselective α-HADHs based on the analysis of ketoacids. Several methods have been developed for the analysis of ketoacids using certain types of derivatization [15, 16]. Among them, the most widely employed is derivatization via 2,4-dinitrophenylhydrazine (DNPH) [15]. The DNPH derivatives of ketoacids are homogeneously red-brown in alkaline condition and can be analyzed by spectroscopic techniques (Scheme 1). When α-hydroxyacid oxidation reaction caused by α-HADH exists, the color of DNPH derivatives of ketoacids in the samples can be easily quantified by spectrophotometric measurements at 458 nm. Use of optically pure (R)- and (S)-α-hydroxyacid concurrently as substrates allows the determination of the apparent enantioselectivity (E app). This colorimetric reaction can be applied to screen for active and enantioselective α-HADHs (Fig. 1). With this method, a large number of samples can be examined and (R)- or (S)-α-HADHs can be obtained much faster and easier than with traditional GC or HPLC-based methods.

Scheme 1
scheme 1

Asymmetric oxidation of α-hydroxyacids with enantioselective α-HADH, yielding optically active α-hydroxy acids. The ketoacids released during the bioconversion can react with DNPH to produce a bright yellow hydrazones, which can be easily quantified by spectrophotometric measurements at 458 nm in alkaline condition. The initial rates of ketoacids formation can thus be determined. Use of optically pure (R)- and (S)-α-hydroxyacid concurrently as substrates allows the determination of the apparent enantioselectivity (E app)

Full size image
Fig. 1
figure 1

Examples of the selective colorimetric screen for enantioselective α-HADH-producing microorganisms. If the positive reaction occurred, the color of DNPH derivatives of ketoacids in the sample will change to red-brown after adding DNPH solution and NaOH. The color change is visible to the naked eye. The strain will be rejected if neither enantiomer reacts or if both enantiomers react at similar rates. If one enantiomer reacts significantly faster than the other, the strain will be tested further

Full size image

Materials and methods

Chemicals

1a–1l, (R)-1a–1l, (S)-1a–1l and 2a–2l (Scheme 1) were purchased from J&K Chemical Co., Ltd. (Shanghai, China). Sinorhizobium sp. CCTCC M 2011391 and Serratia marcescens CCTCC M 2011392, which are known producers of enantioselective α-HADH, were obtained from China Center for Type Culture Collection (CCTCC, Wuhan, China). All other chemicals were of reagent grade and obtained commercially. DNPH solution was prepared by dissolving 198 mg of DNPH and 83.3 mL of hydrochloric acid (38 %) in distilled water to make 1 L.

Microbial cultures and cultivation conditions

The rich medium for fermentation of α-HADH consisted (g/L): glucose 10, yeast extract 10, K2HPO4 2.5, KH2PO4 2.5, MgSO4 0.2, FeSO4 0.03, and (R,S)-1a 2. (pH 7.0). Sinorhizobium sp. CCTCC M 2011391 and S. marcescens CCTCC M 2011392 were grown at 30 °C for 24 h in the rich medium. The preculture (3 mL) was transferred to a 50-mL medium of the same composition and was allowed to grow for 48 h at 30 °C in an incubator shaker (150 rpm).

Potential enantioselective α-HADH-producing microorganisms were isolated using enrichment culture technique. Each isolate was cultivated in rich medium at 30 °C. After 24 h of cultivation, 3 mL of this culture was transferred to 50 mL rich medium in a 250 mL-flask and incubated at 30 °C for 48 h (150 rpm). Cells were harvested by centrifugation at 12,000×g for 8 min at 4 °C. The cell pellets were washed thoroughly with 100 mM phosphate buffer (pH 8.0), and then they were stored at 4 °C for further use. α-HADH enantioselectivity of each strain was measured using (R)- and (S)-α-hydroxyacid concurrently as substrates to evaluate the E app.

Screening procedure

Resting cells (0.2 g) of different microorganisms were resuspended in 10.0 mL phosphate buffer (100 mM, pH 8.0) containing 20 mM (R)-isomer or (S)-isomer. The mixtures were shaken at 30 °C and 150 rpm. Samples (10 μl) were taken at regular intervals and transferred to a 96-well microtiter. 100 μl of DNPH (1.0 mM) was added to the plate for the derivatization with ketoacid produced during the bioconversion. The colorimetric reactions were carried out for 30 min at room temperature and then 890 μl of NaOH (0.4 M) was added. The absorbance of the homogeneous red-brown solution in the plate was determined at 458 nm by an absorbance microplate reader (SpectraMAX puls 384, MD, USA). The initial reaction rates for each enantiomer separately were calculated. The quotient of these rates gives the E app. The enantioselectivity of α-HADH from these selected microorganisms was further confirmed by HPLC in the kinetic resolution of 1. Reactions were carried out at 30 °C and 150 rpm for 3 h in 10 mL phosphate buffer (100 mM, pH 8.0) with a wet cell concentration of 20 g/L and 20 mM substrate.

Asymmetric oxidation of different substrates with the newly isolated strain

Resting cells of P. aeruginosa were resuspended to a cell concentration of 50 g/L in 10 mL phosphate buffer (100 mM, pH 8.0) containing 20 mM 1a–1l. The mixtures were shaken at 30 °C and 150 rpm. Samples were taken at regular intervals and the reaction was terminated through centrifugation (10,000×g, 5 min). The yield and ee (%) were determined by chiral HPLC.

HPLC analysis

The concentrations of 1a–1l and 2a–2l were assayed by HPLC (LC-10AS, Shimadzu, Japan) equipped with a octadecylsilica column (250 × 4.6 mm) (Elite Analytical instruments Co., Ltd, China) at a flow rate of 1.0 mL/min with a solvent system composed of 10 mM NH4H2PO4 and methanol (4:1, v/v). Detection wavelength was set at 228 nm.

The optical purities of products were assayed by HPLC (Dionex UltiMate 3000, USA) equipped with a chiral column (Chirobiotic™ R 250 × 4.6 mm, Sigma, USA) at a flow rate of 1.0 mL/min with a solvent system composed with 0.5 % AcOH-CH3CN (20:80, v/v). The eluate was monitored at 215 nm. Enantioselectivities (E true values) were calculated from conversion and enantiomeric excess (% ee) according to the equation E true = ln[(1 − c) (1 − ee s)]/ln[(1 − c) (1 + ee s)] as described by Chen et al. [17].

Results

Full wavelength scanning for the colored complexes 3

To develop a simple colorimetric assay for rapid identification of active and enantioselective α-HADH based on the reaction of ketoacids with DNPH to form stable hydrazones, it is very important to choose the suitable detection wavelength. We selected 1a–1d as the model substrates, the colored complexes 3a–3d were prepared by reactions of ketoacids 2a–2d with DNPH, respectively [18]. The reactions were carried out for 30 min at room temperature. After adding the NaOH solution, the reaction mixtures become homogeneously red-brown. To find out the ideal detection wavelength, the colored complexes solutions were taken full wavelength scanning with a Beckman DU800 spectrophotometer (Beckman Coulter, USA) from 300 to 700 nm. The results are shown in Fig. 2. The maximum absorbance of DNPH and 3a–3d were 350 and 458 nm, respectively. It was suitable to choose the 458 nm as the detection wavelength, which can accurately determine the amount of the ketoacids 2a–2d and exclude the interference of DNPH.

Fig. 2
figure 2

Absorption spectrums of DNPH and colored complexes 3a–3d

Full size image

Linearity of calibration curves

To quantify the ketoacids 2 formed in the biotransformation by α-HADH, according to the Beer–Lambert equation, it is necessary to find the relationship between the concentrations of 2 and the absorbances. Different concentrations of 2a–2d which were enzymatically synthesized from 1a–1d by Sinorhizobium sp.CCTCC M 2011391 were derivatized with DNPH. The relationship between the concentration of 2a and the absorbance at 458 nm is shown in Fig. 3. The insert showed the linear range of the measurement. The linearity parameters of spectrophotometry between the concentrations of 2b–2d and the absorbance at 458 nm are all shown in Table 1. OD458 were linear with the concentrations of 2b–2d. Thus, it is possible to calculate the conversion for each reaction after determining the standard cures. Lower limit of quantifications of 2a–2d in the solution was 10 μM. Using this method, the ketoacids could be determined sensitively in the μM concentration range by spectrophotometry. Moreover, this assay could be performed either in test tubes with a UV–Vis spectrophotometer or on a 96-well plate with a universal microplate spectrophotometer requiring only a simple testing apparatus at a room temperature, indicating that this method is simple, easy to use and sensitive.

Fig. 3
figure 3

The relationship between the concentration of ketoacid 2a and the absorbance at 458 nm. The insert shows the linear range of the measurement

Full size image
Table 1 Linearity parameters of spectrophotometry
Full size table

Quantitative validation of the assay

To confirm that the color change accurately measured the enzyme activity, the progresses of Sinorhizobium sp. CCTCC M 2011391 [a known producer of (S)-α-HADH] catalyzed asymmetric oxidation of 1a–1d were determined both by the spectrophotometry and HPLC. Typical graphs of 2a–2d concentrations versus reaction time are given in Fig. 4. Comparisons of both data showed that the concentrations of ketoacids 2a–2d based on the spectrophotometry were in close agreement with the values obtained from HPLC. These results indicated that this colorimetric method can be used to determine the α-HADH activity. To demonstrate the efficiency and accuracy of this method for the determination of α-HADH enantioselectivity, we both determined the E app and true enantioselectivity (E true) through colorimetry and HPLC analysis method in the kinetic resolution of 1a–1d by Sinorhizobium sp. CCTCC M 2011391 and S. marcescens CCTCC M 2011392. E app values are the ratio of initial rates determined from the oxidation of optically pure (S) and (R)-enantiomers. E true values are derived from resolutions of racemates. Measuring initial rates of the ketoacids generated from (S)- and (R)-substrates (20 mM) at a wet cell concentration of 20 g/L gives the E app. E true values are determined according to the equation of Chen et al. [17] using racemic substrates (20 mM) for conversion reaction. The bioconversions of (R) and (S)-α-hydroxyacids by Sinorhizobium sp. CCTCC M 2011391 are shown in Fig. 5. The results revealed an almost sole increase in absorption for the (S)-enantiomer reflecting the high enantioselectivity of α-HADH from Sinorhizobium sp. CCTCC M 2011391. The E app determined by the colorimetric method closely matches the value of true enantioselectivity (E true) determined by HPLC analysis (Table 2). These results disclosed that this colorimetric method was sensitive enough to screen the enantioselective α-HADH-producing microorganisms.

Fig. 4
figure 4

Time course of Sinorhizobium sp. CCTCC M 2011391 catalyzed the oxidation of 1a (open triangles, filled triangles), 1b (open circles, filled circles), 1c (open diamonds, filled diamonds) and 1d (open squares, filled squares). The concentrations of ketoacids were determined by the colorimetric method (open symbols) and HPLC (filled symbols). The experiments were carried out in triplicate and the mean values are taken. The standard deviations of the experiments were <15 %

Full size image
Fig. 5
figure 5

Time course of Sinorhizobium sp. CCTCC M 2011391 catalyzed the oxidation of (R)-1a–1d (open symbols) and (S)-1a–1d (filled symbols). open triangles (R)-1a, filled triangles (S)-1a, open circles (R)-1b, filled circles (S)-1b, open triangles (R)-1c, filled diamonds (S)-1c, open squares (R)-1d, filled squares (S)-1d. Reactions were carried out at 30 °C and 150 rpm in 10 mL phosphate buffer (100 mM, pH 8.0) with a cell concentration of 20 g/L and 20 mM (R)-or (S)-1a–1d, respectively. Measuring initial rates of the ketoacids generated from pure enantiomers of (S)- and (R)-substrates as the E app. The experiments were carried out in triplicate and the mean values are taken. The standard deviations of the experiments were <15 %

Full size image
Table 2 Comparison of enantioselectivity determined by the high-throughput screening method (E app) and by HPLC analysis (E true) in the kinetic resolution of 1a–1d by Sinorhizobium sp. CCTCC M 2011391 and S. marcescens CCTCC M 2011392
Full size table

Isolation of enantioselective α-HADHs by the high-throughput screening method

The potential α-HADH-producing microorganisms were evaluated with the high-throughput screening method employing (R) and (S)-α-hydroxyacid as substrates concurrently for the identification of α-HADH enantioselectivity. We selected (R) and (S)-1a as substrates, 34 isolates including bacteria and yeasts, able to produce enantioselective α-HADH, were selected from 526 potential α-HADH-producing microorganisms. Twenty-four strains preferentially produce (S)-α-HADHs. Ten strains preferentially produce (R)-α-HADHs. Microorganisms that prefer the (S)-isomer are widespread in the nature, while those with α-HADHs selective for the (R)-isomer are relatively rare. The enantioselectivity of α-HADHs from these selected microorganisms was confirmed by HPLC in the kinetic resolution of 1a. The results are shown in Fig. 6. HPLC results substantiate the findings from the high-throughput screening method and established beyond doubt that this method can be used for high-throughput enantioselective screening of α-HADH-producing microorganisms. One of the isolated strains ZJB1125 which have the relatively high activity and enantioselectivity of α-HADH was identified as P. aeruginosa by 16S rDNA sequence (GenBank accession no. JQ927220) analysis, chemotaxonomical analysis and biochemical test. This strain was deposited at China Center for Type Culture Collection as CCTCC M 2011394.

Fig. 6
figure 6

Confirmation of activity and enantioselectivity of the α-HADHs from the selected microorganisms and positive control strains by HPLC in the kinetic resolution of 1a. Selected strains: filled circles bacteria, open squares yeasts, filled diamond ZJB1125. Positive control strains: filled triangle Sinorhizobium sp. CCTCC M 2011391, filled inverted triangle Serratia marcescens CCTCC M 2011392. Reactions were carried out at 30 °C and 150 rpm for 3 h in 10 mL phosphate buffer (100 mM, pH 8.0) with a cell concentration of 20 g/L and 20 mM 1a. Minus values of ee stand for (S) absolute configuration of product

Full size image

Asymmetric oxidation of different substrates with the newly isolated strain

Reactions were carried out to investigate the asymmetric oxidation of 1a–1l to prepare (R)-1a–1l. The yields and ees (%) of the products in the reactions were monitored by chiral HPLC. The results are shown in Table 3. It can be seen that P. aeruginosa was highly (S)-selective for 1a–1l. Using the resting cell as a biocatalyst for enantioselective oxidation 1a–1l, the optically pure α-hydroxyacids (R)-1a–1l were obtained in 42.3–49.9 % yield with ee of 92.5–100 % within 3–10 h. The position of the substituent and the pattern of the substitution on the benzene ring had some effect on the enantioselectivity. Comparing the specific activity and substrate specificity, the newly isolated P. aeruginosa was superior to all other α-HADH-producing strains reported [11–14]. It can be concluded that the isolated microorganism is a suitable candidate for the production of optically active α-hydroxyacids from their racemic mixtures. This strain appears promising for potential applications in industry. Because of the encouraging results described above, work is in progress in our laboratory to optimize the different physico-chemical parameters for the higher enzyme productivity and characterization of the enantioselective α-HADH produced by P. aeruginosa.

Table 3 Preparations of (R)-1a–1l by asymmetric oxidation of 1a–1l with newly isolated P. aeruginosa
Full size table

Discussion

The application of biocatalysis for the production of key chiral intermediates in the pharmaceutical and fine chemicals industries is rapidly growing [19]. A major problem in the development of a specific biocatalysis is to find the appropriate biocatalyst. If there are no commercially available enzyme preparations, the desired activities will be found either by screening of strains from culture collections or by isolation of new microorganisms. Screening these biocatalysts for activity and enantioselectivity against target molecules becomes a major bottleneck in the process-development timeline. Implementation of a rapid enantioselective screening method is very important [20–22]. One of the most convenient ways to assay an enzyme is through a method that allows the development of color and thus can be used in qualitative as well as quantitative measurements [23, 24]. In this work, a novel and simple colorimetric assay for the identification of active and enantioselective α-HADH based on the reaction of ketoacids with DNPH to form stable hydrazones was established. Using the convenient and rapid screening method described here, one can analyze large numbers of enantioselective α-HADH-producing microorganisms or mutants simultaneously in a short period, thus reducing the number of samples to be analyzed in a more quantitative detail (HPLC analysis). Although the use of DNPH to precolumn derivatization reaction for HPLC determination has been reported [16, 25, 26], this is the first report on a rapid colorimetric assay method for the identification of enantioselective α-HADH using whole cells.

This new method offers some advantages over the conventional screening method. Firstly, the assay format contained a very broader substrate spectrum. Besides model compounds, other hydroxyacids including the different substituent groups on benzene or different length of side chain are also suitable for this screening strategy. Therefore, we are able to screen satisfactory enantioselective α-HADHs by applying different substrates for specific biotransformation. Secondly, the color change from light yellow to red-brown is readily apparent and sensitive when a positive reaction occurs. This characteristic is very useful when a primary screening is needed. Moreover, this method is quantitative. Use of optically pure (R)- or (S)-enantiomers allows the determination of the E app. Thirdly, this assay can be both performed in microplates and cuvettes with a conventional UV–vis spectrophotometer under room temperature. As GC, HPLC, or NMRS is avoided, the screening process is dramatically faster and the requirement for equipment is very low. Lastly, the DNPH reagent/solution is stable under ambient conditions and does not need to be freshly prepared before each use. Therefore, this method is particularly useful for the screening of large numbers of microorganisms or mutants, allowing quick identification of organisms with desired enantioselectivity, discarding poor to moderate enantioselective organisms or mutants that will not be acceptable for the development of a biocatalytic resolution process. This fits the strategy of a hierarchical screen for the identification of best biocatalyst as one of the earliest steps and eliminates the weakest candidates for a more streamlined process viability study [24, 27, 28]. The screen was designed in such a way as to obtain α-HADHs with high activity and enantioselectivity for the production of optically active hydroxyacids from their racemic mixtures, which will have practical applications.