Online capillary solid-phase microextraction coupled liquid chromatography-mass spectrometry for analysis of chiral secondary alcohol products in yeast catalyzed stereoselective reduction cell culture

•An online SPME-LC-MS analysis method has been developed.•Dual switching valves were used for online salt elimination.•R- and S-4-phenyl-2-butanol in complicated yeast cell culture were analyzed.•The method has been proved to be sensitive, precise, and accurate.

An online solid-phase microextraction coupled liquid chromatography-electrospray ionization-ion trap mass spectrometry was developed for the analysis of trace R- and S-4-phenyl-2-butanol (R- and S-pbol) in salt rich cell culture of Saccharomyces cerevisiae catalyzed stereoselective reduction of 4-pheny-2-butanone (pbone). A Supel-Q PLOT capillary column was used for the extraction and deionized distilled water was used as the extraction mobile phase. The extraction flow rate and extraction time were at 0.1 mL min−1 and 0.95 min, respectively. The three target analytes, pbone, R-pbol, and S-4-pbol, were desorbed and eluted by the mobile phase of water/methanol/isopropanol (55/25/20, v/v/v) with a flow rate of 0.5 mL min−1 and analyzed by a chiral column. The mass spectrometric detection of the three target analytes was in positive ion mode with the signal [M+Na]+. The matrix-matched external standard calibration curves with linear concentration range between 0 and 50 μg mL−1 were used for quantitative analysis. The linear regression correlation coefficients (r2) of the standard calibration curves were between 0.9950 and 0.9961. The yeast mediated reduction was performed with a recation culture of yeast incubation culture/glycerol (70/30, v/v) for 4 days. This biotransformation possessed 82.3% yield and 92.9% S-enantomeric excess. The limit of detection (LOD)/limit of quantification (LOQ) for pbone, R-pbol, and S-pbol was 0.02/0.067, 0.01/0.033, and 0.01/0.033 μg mL−1, respectively. The intra-day and inter-day precisions from repeated measurements were 10.8–21.1% and 11.6–18.7%, respectively. The analysis accuracy from spike recovery was 84–91%.