High‐Throughput Method for Determining the Enantioselectivity of Enzyme‐Catalyzed Hydroxylations Based on Mass Spectrometry

Abstract

Up to speed: An accurate, sensitive, high-throughput, and simple method for measuring the product ee value of enzyme-catalyzed hydroxylations (see scheme) is based on the use of enantiopure or enantioenriched deuterated substrates and mass spectrometric detection. Catalytic asymmetric synthesis has attracted great attention because of the importance of enantiopure compounds in the preparation of pharmaceuticals, fine chemicals, and agrochemicals.1 Recently, the discovery of enantioselective catalysts for such syntheses has focused on the generation and screening of libraries of chemical catalysts2 and enzymes.3, 4 While huge catalyst libraries can be quickly created by combinatorial synthesis or molecular biotechnology, such as the error-prone polymerase chain reaction (epPCR) and DNA shuffling,5 the analysis of the enantioselectivities of these catalysts is the main bottleneck. Many high-throughput methods for determining the enantioselectivity of catalysts have been developed.6 While the catalyst enantioselectivity for kinetic resolutions can be quickly estimated based on the different reaction rates of the two enantiomers of the substrates, the determination of catalyst enantioselectivity for asymmetric transformations from an achiral substrate has to rely on the measurement of product ee values except for some reversible transformations, the back reactions of which could be studied as kinetic resolutions.7 As the general and conventional methods for the analysis of product ee values by GC or HPLC with a chiral column suffer from long analysis times and other drawbacks, the quick determination of the ee values of some special products can be achieved by GC/GC with a chiral column,8a HPLC with CD/UV8b,8c or optical rotation/refractive index unit detection,8c chirally modified capillary electrophoresis,8d electrospray ionization tandem mass spectrometry (MS),8e color indicators based on doped liquid crystals,8f or competitive enzyme immunoassays.8g More complicated high-throughput methods requiring further conversion of the product have also been developed: product ee values can be estimated by exploiting kinetic resolution effects on the product,9a by using mass- or fluorescence-tagged quasienantiomeric mixtures of acylating agents with MS9b,9c or fluorescence9d detection; with known product concentration, the product ee value can be determined by using an enzyme to catalyze a further transformation of the product detected by UV spectroscopy10a or IR thermography;10b the product ee value can also be established by using two enantioselective enzymes to modify the product with UV detection of NAD(P)H formation.11 Nevertheless, the application of these relatively complicated methods in catalyst discovery depends on many factors and has to be evaluated on a case-by-case basis. Enantioselective hydroxylations are important reactions for the preparation of chiral alcohols, which are useful and valuable pharmaceutical intermediates and fine chemicals. While chemical hydroxylations often suffer from poor chemo-, regio-, and enantioselectivity,12 enzymatic hydroxylations have received great attention.13 Many microorganisms have been discovered and developed for a number of useful biohydroxylations.14 Several monooxygenases have been purified, cloned, and genetically engineered for more efficient biohydroxylations.15 In many cases, the enantioselectivity of enzymatic hydroxylations needs to be further improved, and some success has been made by directed evolution of monooxygenases, such as P450 BM-3 hydroxylase.16 However, none of the known high-throughput enantioselectivity assays has been applied in such evolutions. The screening of enzymes was based on activity-assay and GC or HPLC analysis with a chiral column of the selected active mutants. The lack of practical high-throughput enantioselectivity assays may limit the chance of success in discovering appropriate enzymes with desired enantioselectivity. We are interested in developing a practical high-throughput enantioselectivity assay for biohydroxylations. MS-based assays are known to be simple and to have high throughput, such as the use of isotopically labeled quasienantiomeric or meso substrates17 to determine the catalyst enantioselectivity. However, this approach cannot be extended to asymmetric transformations of a chiral non-meso substrate. Herein, we report a different strategy and new principle for high-throughput ee determination for biohydroxylations by the use of an optically active, isotopically labeled substrate with MS detection. Principle of a high-throughput enantioselectivity assay for biohydroxylation based on the use of enantiomerically enriched deuterated substrates and MS detection. The symbols are explained in the text. To prove the concept, the deuterated substrates (R)-3 and (S)-3 were prepared according to the route shown in Scheme 2, starting with commercially available enantiopure (R)-2 and (S)-2.18 Although the ee values of (R)-3 and (S)-3 cannot be directly determined, we assume that (R)-3 and (S)-3 have the same ee value as (S)-6 and (R)-6 (86 %), as the configuration cannot be changed during deuteration. Synthesis of deuterated (R)-3 and (S)-3. The synthesized enantioenriched (R)-3 and (S)-3 were used for biohydroxylation, and our recently discovered strain Pseudomonas monteilii TA-519 was chosen as the biocatalyst. The biohydroxylations were performed with resting cells (10 g cdw L−1; cdw=cell dry weight) of the TA-5 strain and 10 mM substrate in 100 mM potassium phosphate buffer (pH 7.0) at 30 °C for 15 minutes. The products were extracted with ethyl acetate and analyzed by GC/MS. Typical spectra are shown in Figure 1, and a and b values were easily obtained from the intensities of the signals at m/z 122 (M) and 123 (M+1). The ee value was calculated from Equations (5) and (6) as 82 % (R; Table 1, entry 1). GC/MS analysis of the products from biohydroxylation of (R)-3 and (S)-3 with P. monteilii TA-5. A,C) Gas chromatograms; B,D) mass spectra of the product peak at 2.12 min in GC. Entry Product conc.[a] [mM] n a b x ee[b] [%] ee[c] [%] 1 0.170 13.3 3.050 0.093 0.102 82 (R) 83 (R) 2 0.085 13.3 3.972 0.124 0.103 81 (R) – 3 0.034 13.3 2.913 0.120 0.130 77 (R) – For comparison, the hydroxylation of ethyl benzene (1; 10 mM) was carried out with P. monteilii TA-5 under the same conditions for 15 minutes, and the product was extracted with ethyl acetate. The product was further concentrated for chiral HPLC analysis, which gave an ee value of 83 % (R). Thus, our new method gave an accurate ee value, with an error of only about 1 % ee. The high sensitivity of the method was also demonstrated: analysis of diluted samples with product concentrations of 0.085 and 0.034 mM gave product ee values of 81 % (R) and 77 % (R), respectively, with only a slight difference from the real value. The GC/MS analysis of each sample took about 2.2 minutes at 160 °C (Figure 1). The analysis time was further reduced to 1.5 minutes at 180 °C, with the same analytical accuracy. Theoretically, 960 samples could be determined per day. To further demonstrate the generality of this method, enantioselective biohydroxylation of N-benzylpyrrolidine (7) at the 3-position was selected as another example. This represents an enantioselective hydroxylation of a nonactivated carbon atom, a symmetric substrate with equal CH bonds at the 3- and 4-positions, and a useful transformation to prepare the corresponding product (S)-8 or (R)-8, which is a key synthetic intermediate for several pharmaceuticals.20 Principle of a high-throughput enantioselectivity assay for biohydroxylation of a symmetric substrate based on the use of enantiopure deuterated substrates and MS detection. To prove the concept, (S)-9 and (R)-9 were prepared for the first time from the commercially available enantiopure (S)-8 and (R)-8 according to the routes shown in Scheme 4. The ee value of 9 is expected to be the same as that of the corresponding precursor 13: (S)-9 in 99 % ee and (R)-9 in 95 % ee. Synthesis of [3-D](R)- and -(S)-1-benzylpyrrolidine (9). DME=dimethyl ether. To evaluate the new method, (S)-9 and (R)-9 were used for the hydroxylation, and the potential of using LC/MS for the analysis was explored. At the beginning, Sphingomonas sp. HXN-200, a known biocatalyst for the hydroxylation of non-activated carbon atoms,20 was selected as the catalyst. Biohydroxylation was performed with 5 mM (R)-9 or (S)-9 in a cell suspension (5.5 g cdw L−1) in 50 mM potassium phosphate buffer (5 mL, pH 7.5) containing glucose 2 % (w/v) at 30 °C for 1 h. Samples were taken directly from the aqueous biotransformation mixtures and analyzed by LC/MS; typical chromatograms are shown in Figure 2. No column separation was required, and the a and b values were easily obtained from the intensities of the signals at m/z 178 (M) and 179 (M+1). Thus, the y and ee values were calculated based on Equations (11) and (12). LC/MS analysis of the products from biohydroxylation of (S)-9 and (R)-9 with Sphingomonas sp. HXN-200. A,C) Liquid chromatograms; B,D) mass spectra of the peak between 0.2 and 0.3 min in LC. As summarized in Table 2, entry 1, the product ee value for biohydroxylation of 7 with Sphingomonas sp. HXN-200 was estimated as 53 % (S) from this method. The real product ee value for the biohydroxylation of 7 with this catalyst under the same conditions was established as 54 % (S) by chiral HPLC analysis. This result demonstrates the high accuracy of the new method. Moreover, the use of LC/MS analysis does not require the extraction of product into organic solvent, thus allowing the direct use of aqueous samples taken from the biotransformation mixtures. It also provides high sensitivity, with no problem of using a 20 times diluted sample containing 0.023 mM product. In comparison, chiral HPLC analysis required the extraction of the product from aqueous medium (5 mL) with ethyl acetate (5 mL) followed by concentration of the sample by evaporation to 30 μL. In fact, chiral HPLC analysis failed with a nonconcentrated sample because of the sensitivity limitation. In addition, LC/MS analysis took only 1.0 minute, which provided a theoretical analytical throughput of 1440 samples per day. In comparison, chiral HPLC analysis required about 30 minutes for each sample. Entry Biocatalyst[a] t [h] Product conc.[b] [mM] a b y ee[c] [%] ee[d] [%] 1 Sphingomonas sp. HXN-200 1.0 0.023 0.080 0.017 3.25 53 (S) 54 (S) 2 E. coli BL21(DE3) 1AF4 0.5 0.005 0.049 0.179 0.372 46 (R) 42 (R) 3 P. oleovorans GPo1 12 0.009 0.014 0.061 0.324 51 (R) 57 (R) The method was further examined with other available biocatalysts for this hydroxylation. Recombinant Escherichia coli BL21 expressing the 1AF4 mutant of the P450 monooxygenase of Sphingomonas sp. HXN-200 catalyzed the biohydroxylation of 7 to give the corresponding (R)-8 in 42 % ee (Table 2, entry 2). This is the opposite enantioselectivity to that obtained with Sphingomonas sp. HXN-200. To obtain the product ee value by our new method, biotransformation of (S)-9 and (R)-9 with BL21 (DE3) 1AF4 was performed under the same conditions, and samples from aqueous buffer were analyzed by LC/MS without column separation. As shown in Table 2, entry 2, the product ee value was established as 46 % (R) by this method, which has an error of only 4 % ee. Also in this case, the ee determination is sensitive, with a product concentration of 0.005 mM, and fast, with an analysis time of 1.0 minute. P. oleovorans GPo1 contains a well-known membrane-bound alkB hydroxylase system and catalyzes the hydroxylation of a number of aliphatic compounds.21 Biohydroxylation of 7 with this strain gave (R)-8 in 57 % ee (Table 2, entry 3). Similarly, the ee value was also established by using the new method through the separate biotransformation of (S)-9 and (R)-9 with P. oleovorans GPo1 and LC/MS analysis of the aqueous samples. As shown in Table 2, entry 3, the product ee value was established as 51 % (R) by this method with an error of 6 % ee. Once again, the analysis is sensitive and fast, being able to determine samples containing 0.009 mM product within 1.0 minute. In summary, we have developed a new method for measuring the product ee value for enantioselective hydroxylation based on the use of enantiopure or -enriched deuterated substrates and MS detection. Our method has several distinctive features. 1) The ee value can be determined with satisfactory accuracy, independent of the type of hydroxylation and the nature of the biocatalyst. 2) The analysis method is very sensitive. It can analyze samples with product concentration as low as 0.005 mM, thus being suitable for the screening of enantioselective enzymes for biohydroxylation, for which the product concentration is often low. 3) The analysis is based on MS and does not require separation. It takes only 1.0 minute for LC/MS analysis, thus providing high-throughput analysis of 1440 samples per day. 4) The analysis method is simple. It allows direct analysis of aqueous samples from biotransformation and does not require further reactions of the bioproduct, which is often necessary for many other assays. 5) The deuterated substrate can be easily prepared from the corresponding alcohol, and does not need to be enantiopure. Currently we are applying this method to the discovery of enantioselective enzymes for biohydroxylations through direct evolution of monooxygenases. General procedure for determining the ee value of bioproduct 2 from biohydroxylation of 1: In two parallel experiments, enantiomerically enriched (R)-3 and (S)-3 were added to a cell suspension (10 g cdw L−1) of P. monteilii TA-5 in 100 mM potassium phosphate buffer (1 mL, pH 7.0) to a substrate concentration of 10 mM. The mixture was shaken at 1000 rpm at 30 °C for 15 min. The cells were removed by centrifugation, and the aqueous solution (0.4 mL) was mixed with ethyl acetate (0.4 mL). The organic phase was separated and analyzed by GC/MS. The ratio of the signals at M and M+1 in the mass spectra was determined as 3.050 for a and 0.093 for b, respectively, and the values were used to calculate the ee value as 82 % (R) for product 2 from the biohydroxylation of 1 with the same enzyme. Upon comparison of the product ee value, 83 % (R), from the real biohydroxylation of 1 with the same enzyme and determined by chiral HPLC analysis, the GC/MS-based method gave an accurate ee value with an error of 1 % ee. The analysis took 2.2 min for samples with a product concentration from 0.034 to 0.170 mM. General procedure for determining the ee value of bioproduct 8 from biohydroxylation of 7: In two parallel experiments, (S)-9 or (R)-9 was added, to a final concentration of 5.0 mM, to a suspension of cells of Sphingomonas sp. HXN-200 (5.5 g cdw L−1) in 50 mM potassium phosphate buffer (5 mL, pH 7.5) containing 2 % (w/v) glucose. The mixture was shaken at 300 rpm at 30 °C for 1 h, and the cells were removed by centrifugation. The supernatant (50 μL) was diluted 20 times with methanol and used as the sample for LC/MS analysis. From the mass spectra, a and b values were obtained as 0.080 and 0.017, respectively, thus establishing the ee value of product 8 as 53 % (S) from the biohydroxylation of 7 with Sphingomonas sp. HXN-200. As biohydroxylation of 7 with the same strain gave product 8 in 54 % ee (S) determined by chiral HPLC analysis, the LC/MS based method is accurate, with an error of only 1 %. The analysis took 1.0 min for samples with a product concentration of 0.023 mM. Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer-reviewed, but not copy-edited or typeset. They are made available as submitted by the authors. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.