Effect of coenzyme analogues on enantioselectivity of alcohol dehydrogenase

摘要

1083J. CHEM. SOC. PERKIN TRANS. 1 1992 Effect of Coenzyme Analogues on Enantioselectivity of Alcohol Dehydrogenase Changsheng Zheng and Robert S. Phillips* Departments of Chemistry and Biochemistry, University of Georgia, Athens, GA 30602, USA Secondary alcohol dehydrogenase from Thermoanaerobacter ethanolicus catalyzes the reduction of butan-2-one with much higher enantioselectivity when NADP is replaced by APADP, SNADP or NAD; as expected, the enantiomeric ratios (~=~~/~~)~/(~~~~/~~)~of the reaction of SADH with (R)-and (S)-butan-2-01 increase with the coenzyme analogues. Although enzyme use in asymmetric organic synthesis is now widespread as a result of the high stereoselectivity and specificity which may be achieved, the relative inflexibility of enzyme stereoselectivity (e.g.high stereochemical purity with one substrate, but not with another) often limits wider synthetic exploitation. Recent, novel approaches to expand the range of useful enzyme stereoselectivity involve changes in reaction conditions. We have reported our studies on the temperature- dependent enantiospecificity of secondary alcohol dehydroge- nase (SADH) from Thermoanaerobacter ethanolicus. Klibanov and co-workers have found a solvent dependent enantioselectiv- ity of the protease, subtilisin Carlsberg.2 Here, we report that coenzyme analogues can enhance the stereoselectivity of a dehy- drogenase. Asymmetric reduction of small aliphatic ketones using enzymes has enjoyed only limited SUCC~SS;~ e.g. reduction of aliphatic ketones with yeast alcohol dehydrogenase gave secondary alcohols with only moderate optical p~rity,~ and acyclic ketones are poor substrates for horse liver alcohol de- hydrogenase (HLADH).' The secondary alcohol dehydrogen- ases from thermophilic bacteria, Thermoanaerobium brockii (TBADH)6 and T.ethanoiicus (SADH),'*7 exhibit high activity and high stereoselectivity with a wide range of acyclic ketones. However, TBADH and SADH reduce butan-2-one and pentan- 2-one with low enantioselectivity, using NADP as co-factor.'Y6 Since thionicotinamide adenine dinucleotide (SNAD) and 3-acetylpyridine adenine dinucleotide (APAD) showed signifi- cantly higher activity than NAD in HLADH oxidation reactions,' we decided to study the NADP analogues, SNADP and APADP, as well as NAD, to evaluate the effect of coenzyme structure on the stereoselectivity of butan-2-one and pentan-2- one reduction by SADH.Table 1, summarizes the results obtained from these reactionsf For butan-2-one at 37 "C, when NADP was replaced by APADP, SNADP or NAD, the enantiomeric ratio, E (=R/S),of the butan-2-01 product increased from 1.3 to 3.7, 2.9 or 4.6, respectively; at 47 "C, E increased from 1.9 to 9.4,6.6 or 7.0. As the temperature of the reaction was increased, E also increased, e.g. with APADP, from 3.7 to 9.4 when the temperature increased from 37 to 47 "C.Whereas the reduction of butan-2-one by SADH with NADP is not sufficiently stereoselective (30 e.e.) to be of preparative value, the (R)-butan-2-01 obtained under these latter conditions has stereochemical purity 80 e.e.In the reduction of pentan-2- one, we also observed a small but significant increase in E with the cofactor analogues (Table 1); the (9-pentan-2-01 obtained is thus of somewhat lower enantiomeric purity. These results suggest that these cofactor analogues favour the hydride transfer on the pro-(S)face of the ketone, relative to NADP. In contrast to the oxidations reported with HLADH,* the relative rates of pentan-2-one reduction by SADH with these analogues are considerably slower than NADP (Table 1). We have also studied the steadv-state kinetics of the reaction Table 1 Reduction of acyclic ketones with coenzyme analogues catalysed by SADH from T.ethano1icus.f OH ?H 1; X=Et 2; X=Pr Substrate Coenzyme T/ "C Ratio (R/S) Relative rate" 1 NADP 37 1.3 1 APADP 37 3.7 1 SNADP 37 2.9 1 NAD 37 4.6 1 NADP 47 1.9 1 APADP 47 9.4 1 SNADP 47 6.6 1 NAD 47 7.0 2 NADP 37 0.33 100 2 APADP 37 0.36 21.9 2 SNADP 37 0.38 32.1 2 NAD 37 0.42 22.6 2 APADP 47 0.32 2 NAD 47 0.45 The relative rate of pentan-2-01 formation with NADPH at 37 "C was assigned as 100. of SADH and these coenzyme analogues with (R)-and (S)-butan-2-01. Based on the results presented in Table 1, we predicted that (kc,,/Km)R/(kc,,/Km)sfor butan-2-01 would increase when NADP is replaced by APADP or SNADP. The kinetic parameters were measured with purified SADH as previously described.' The results of these kinetics studies agree with those observed from preparative reactions, since APADP and SNADP showed higher E values than NADP at each temperature.A plot of -RTlnE against absolute temperature (Fig. 1) shows that APADP and SNADP possess linear temperature dependencies of stereospecificity similar to NADP.' By changing from NADP to coenzyme analogues, the enantioselectivity of the reduction of butan-2-one catalysed by SADH from Thermoanaerobacter ethanolicus increased signi- ficantly. We believe that in other NAD or NADP-dependent oxidoreductase reactions with poor stereoselectivity, choice of the appropriate coenzyme analogue could be a powerful factor in improving stereoselectivity and extending the utility of available enzymes in asymmetric synthesis.Experimental Reduction of Ketones with SADH.-Reaction mixtures (10 cm3 portions) containing coenzyme (0.05 mmol dm-3) substrate (0.1 cm3), propan-2-01 (1 cm3), (SADH) (12 units; as sonicated cell extract of T. ethanolicus'), in Tris buffer (pH 8; 50 mmol) were kept in a constant temperature bath for 10 h and then 1-1084 400 T0 -g -400 -85 E -800 -1 200 -1 600 r I I I I 1 35 305 315 325 335 TIK Fig. 1 worked up by saturation with (NH4)2S04 and ether extraction. After being dried (Na2S04) and evaporated at reduced pressure, the crude mixtures were analysed for relative rate by gas chromatography. Determination of Enantiomeric Ratio, R/S.-A sample of alcohol was mixed with (S)-N-trifluoroacetylprolylchloride (2 equiv.) in CH2C12 at room temperature for 2 h.The resulting solution was then analysed by gas chromatography on a Chirasil-Val1 capillary column. Kinetics Experiments.-Cuvettes contained the following components: coenzyme (0.1 mmol dm-3), alcohol (0.2 mol dm-3; 10-160 mm3) and Tris-HC1 buffer (pH = 8.9; 100 mrnol d~n-~), J. CHEM. SOC. PERKIN TRANS. I 1992 in a final volume of 0.6 cm3. Because of the high temperature coefficient of Tris, the pH of the Tris-HC1 buffer was adjusted to 8.9 at each temperature. The cuvettes were preincubated at the appropriate temperature from 25-50 "Cbefore the reaction was started by addition of enzyme solution.The rates were measured spectrophotometrically by the increase in absorbance of NADPH, APADPH and SNADPH at 340, 363, and 396 nm, respectively, on a Gilford Response UV-VIS spectro-photometer equipped with a six-cell holder and electronic temperature control. The values of k,,,/Km for each enantiomer of the alcohol were calculated at each temperature. References 1 V. T. Pham, R. S. Phillips and L. G. Ljungdahl, J. Am. Chem. Soc., 1989, 111, 1935; V. Y. Pham and R. S. Phillips, J. Am. Chem. SOC., 1990,112,3629. 2 P. A. Fitzpatrick and A. M. Klibanov, J. Am. Chem. Soc., 1991, 113, 3166. 3 C. J. Suckling and K. E. Suckling, Chem. SOC.Rev.,1974,3,387; S. W. May and S. R. Padgette, Biotechnology, 1983, 677; G. M. Whitesides and C. H. Wong, Aldrichim. Acta, 1983,16,27. 4 R. Macleod, H. Prosser, H. L. Fikentscher, J. Lanyi and H. S. Mosher, Biochemistry, 1964,3,838. 5 T.Takemura and J. B. Jones, J. Org. Chem., 1983,48,791; V. Prelog, Pure Appl. Chem., 1964,9,119. 6 E. Keinan, E. K. Hafeli, K. K. Seth and R. Lamed, J. Am. Chem. Soc., 1986, 108, 162; E. Keinan, K. K. Seth and R. Lamed, J. Am. Chem. SOC.,1986,108,3474. 7 F. 0. Bryant, J. Wiegel and L. G. Ljungdahl, Appl. Enuiron. Microbiology, 1988,54,460. 8 R. J. Kazlauskas, .I.Org. Chem., 1988,53, 4633; N. A. Beijer, H. M. Buck, L. Sluytennan and E. M. Meijer, Biochim. Biophys. Acta, 1990, 1039,277. Paper 2/01 116F Received 2nd March 1992 Accepted 2nd March 1992