In a series of four racemic phenoxyalkyl-alkyl carbinols, 1-phenoxy-2-hydroxybutane (1) is enantioselectively acetylated by Burkholderia cepacia (formerly Pseudomonas cepacia) lipase with an E value $ 200, whereas for the other three racemates E was found to be # 4. To explain the high preference of B. cepacia lipase for (R)-(1)-1, a precursor of its transition state analogue with a tetrahedral P-atom, (R P ,S P )-O-(2R)-(1-phenoxybut-2-yl)-methylphosphonic acid chloride was prepared and crystallized in complex with B. cepacia lipase. The X-ray structure of the complex was determined, allowing to compare the conformation of the inhibitor with results of molecular modelling.Keywords: Burkholderia/Pseudomonas cepacia lipase, racemic sec alcohols, transition state (TS) analogue, crystal structure, molecular modelling.Among more than 30 commercially available lipases [1][2][3][4] that are frequently used in enantioselective acylation of alcohols and amines [5,6] or in esterification of carboxylic acids and hydrolysis of their esters [7,8], Burkholderia cepacia (formerly Pseudomonas cepacia) lipase is one of the most thoroughly studied. The X-ray crystal structures of the native [9,10], and inhibited [11,12] enzyme in open conformations were reported. These structural data have been used in computer simulations of enantioselective ester hydrolysis catalysed by B. cepacia lipase [13][14][15]. Other relevant reports comprise the study of electronic effects of substituents on the enantioselectivity of B. cepacia lipase [16] and flexible docking based on structural information of inhibited B. cepacia lipase [17].Models for predicting enantiopreference in lipasecatalysed acylations of sec alcohols have been proposed based on experimental results. Kazlauskas et al. [18] proposed an empirical rule that predicts which enantiomer reacts faster. This rule relates the relative size of the substituents on the stereogenic centre with enantioselectivity. Other authors reported results related to this rule, in particular the effect of large- [19] and medium-sized substituents [20]. However, these rules do not clarify the detailed mechanism of the enantiorecognition.For some time we have been studying lipase catalysed stereoselective acylations [21-24] of conformationally flexible sec alcohols as substrates. Two important results emerged. First, not all lipases acylate macrocyclic, sterically more constrained sec alcohols with higher stereoselectivity compared to their open-chain counterparts [24]. Second, some lipases acylate with increased enantioselectivity acyclic sec alcohols with perturbing L (large) and M (medium) groups at larger distance from the stereogenic centre [23]. In this latter case, a nonmonotonous correlation between E value and the distance (n) of the perturbing groups in 1-4 was observed (Fig. 1).Continuing this project, we have undertaken biocatalytic, structural, and modelling studies to get more defined information on the mechanism that provides a high degree or bias of enantioselectivity in acetylation of ...

To understand the origin of high enantioselectivity of Burkholderia cepacia lipase (BCL) toward secondary alcohol, (R,S)-1-phenoxy-2-hydroxybutane (1), and its ester (E1), we determined the crystal structure of BCL complexed with phosphonate analogue of S-E1 and accomplished a series of MM, MC, and QM/MM studies. We have found that the inhibitor in the S configuration binds into the BCL active site in the same manner as the R isomer, with an important difference: while in case of the R-inhibitor the H-bond between its alcohol oxygen and catalytic His286 can be formed, in the case of the S-inhibitor this is not possible. Molecular modeling for both E1 enantiomers revealed orientations in which all hydrogen bonds characteristic of productive binding are formed. To check the possibility of chemical transformation, four different orientations of the substrate (two for each enantiomer) were chosen, and a series of ab initio QM/MM calculations were accomplished. Starting from the covalent complex, we modeled the ester (E1) hydrolysis and the alcohol (1) esterification. The calculations revealed that ester release is possible starting with all four covalent complexes. Alcohol release from the BCL-E1 complex in which the S-substrate is bound in the same manner as the S-inhibitor in the crystal structure however is not possible. These results show that the crystallographically determined binding modes should be taken with caution when modeling chemical reactions.

No abstract