Protein Engineering vol. 16 no. 5 pp. 319-322, 2003
© 2003 Oxford University Press
Structural and energetic determinants for enantiopreferences in kinetic resolution of lipases
1Institute of Pharmaceutical Chemistry, University of Marburg, Marbacher Weg 6, D-35032 Marburg and 2BASF AG, Section GVF/E Biocatalysis, D-67056 Ludwigshafen, Germany
3 To whom correspondence should be addressed. e-mail: klebe{at}mailer.uni-marburg.de
Keywords: crystal structure/enantiopreference/kinetic resolution/lipase/transition-state analogs
 |
Introduction |
|---|
 Top Introduction ResultsDiscussionReferences |
|---|
Lipases can catalyze the hydrolysis, transesterification and amidation of a broad range of esters and amides with distinct stereopreference in addition to their physiological function of cleaving triacylglycerols (Schmid and Verger, 1998
). As a consequence of their stability and large-scale availability, they have found widespread applications in the enantioselective synthesis of precursors to pharmaceuticals and in the kinetic resolution of racemic mixtures (Reetz, 2002). From a thermodynamic point of view, the latter ability to accomplish chiral discrimination results from a difference in Gibbs free energy of activation, 
RSG
(Eyring and Polanyi, 1931; Phillips, 1992, 1996). It is experienced by both stereoisomeric transition states formed with the two enantiomers in the rate-limiting step with respect to the ground state. In order to predict and subsequently design enantioselectivity (Kazlauskas, 2000) towards different substrates, it is necessary to understand the structural and energetic determinants by which lipases distinguish enantiomers at a molecular level.
Analyzing the enantioselectivity of lipases towards a large range of secondary alcohols, Kazlauskas formulated an empirical rule to predict the enantiopreference in hydrolysis and transesterification reactions (Kazlauskas, 1993). Comp arison of the crystal structures of Candida rugosa lipase inhibited by both enantiomeric forms of an (RC)- and (SC)-menthylphosphonate transition-state analog (TSA) revealed geometric differences between the two structures (Cygler et al., 1994). In the TSA complex corresponding to the slow-reacting substrate, the geometry of the catalytic triad is perturbed, with the imidazole ring plane of the catalytic histidine rotated away from the alcohol moiety, thus leading to disruption of a hydrogen bond between the menthyl oxygen and N
of histidine. This bond is essential for the catalytic reaction; in contrast, it is correctly formed in the TSA of the fast-reacting substrate.
Candida antarctica lipase B (CaL B) is frequently used in kinetic resolution (Anderson et al., 1998). Crystal structure analysis reveals a rather narrow and deep active-site cleft with a small stereospecificity pocket flanked by residues Thr42, Ser47 and Trp104 (Uppenberg et al., 1995; Orrenius et al., 1998). This structurally constrained binding site provides an ideal platform for enantioselective recognition of substrates.
|
Results |
|---|
|
TopIntroductionResultsDiscussionReferences |
|---|
We decided to study the kinetic resolution of (RC)- and (SC)-1-phenylethylamine using CaL B that has been preacylated with an ethoxyacetyl group (Figure 1). This reaction shows a remarkable stereodiscrimination with an enantiomeric ratio E of >1000 in favour of (RC)-1-phenylethylamide (>99.9% ee) (Balkenhohl et al., 1997
). The amine makes a nucleophilic attack on the acyl enzyme under non-aqueous conditions to produce the corresponding amide. By transient stopped-flow kinetics (Gutfreund and Sturtevant, 1956; Gutfreund, 1996) we could show that neither substrate binding nor deacylation of the acyl enzyme is rate limiting for hydrolysis of p-nitrophenethylacetamide as a model compound. This conclusion is based on the fact that neither an initial burst (excluding the deacylation step of the acyl-enzyme complex as rate limiting) nor an initial lag (ruling out that substrate binding to the enzyme is rate limiting) could be observed. This implies that the formation of the tetrahedral transition state is most likely rate limiting and accordingly involved in the enantioselectivity discriminating step. Studying the temperature dependence of the enantiomeric ratio E between 20 and 90°C reveals a linear van’t Hoff diagram (lnE = –R – SG/RT = –R – SH/RT + R – SS/R). The differential activation free energy (Overbeeke et al., 1998) shows that the fast-reacting (Rc)-amine is discriminated with respect to the slow-reacting (Sc)-amine by a free energy difference of 19.4 ± 6 kJ/mol (Figure 2). Factorizing R – SG into enthalpic and entropic contributions reveals that the fast-reacting substrate is enthalpically favored by 33.1 ± 3 kJ/mol. Although the enthalpic contribution dominates, the entropic contributions are opposite: the slow-reacting enantiomer is entropically favored by TR – SS of –13.7 ± 3 kJ/mol at 298 K. This is consistent with the observed overall reduction in enantiopreference with increasing temperature. Similar experimental results have been reported for CaL B with respect to ester substrates (Ottosson and Hult, 2001; Ottosson et al., 2002).
|
|
To obtain a structural explanation for these remarkable energy differences, we synthesized precursors of the enantiopure phospho-analog TSA inhibitors with a p-nitrophenyl substituent as an appropriate leaving group. The stereochemistry of the four stereoisomers revealed was confirmed by small molecule crystal structure analysis. Only the TSA precursors with the S-configuration at phosphorus react with the lipase through nucleophilic attack of Ser105O
at the central phosphorus. The rate of inhibition observed by measuring the residual enzymatic activity is faster for (RCSP)-TSA in agreement with the stereopreference of the kinetic resolution. Selective reactivity of the (RC)-TSA and (SC)-TSA stereoisomers can be explained by the fact that only these stereoisomers place the phenylethylamine moiety correctly into the stereospecificity pocket simultaneously forming hydrogen bonds to His224 and via one of the oxygens at phosphorus to the hydrogen-bond acceptors in the oxy-anion hole. Product formation with the enzyme was confirmed for both the (RC)- and (SC)-phospho-TSAs by MALDI-TOF mass spectrometry. The crystal structures of the (RC)- and (SC)-complex (Figure 3) could be determined to 2.0 and 2.5 Å resolution, respectively (Table I).
|
|
The electron density for the fast-reacting (RC)-phospho-TSA unambiguously delineates the binding mode of the inhibitor covalently bound to Ser105O
(Figure 3a). A gauche conformation of the ethoxymethyl P-substituent facilitates occupation of the acid-binding pocket. The oxyanion hole is filled by one of the terminal P-oxygen atoms, which makes hydrogen bonds with Gln106NH, Thr40NH and Thr40O. The amide NH of the inhibitor hydrogen-bonds to His224N, while the methyl group at the stereogenic centre is directed into the stereospecificity pocket. The terminal phenyl moiety is well defined and orients perpendicular to the side chain of Trp104. The fast-reacting (RC)-phospho-TSA appears to exhibit perfect surface and shape complementarity to the CaL B binding pocket (Figure 3b). In contrast, electron density for the slow-reacting (SC)-phospho-TSA allows only partial assignment of the molecular skeleton, despite comparable resolution of the dataset (Figure 3c). Although the positions of the covalently attached P and the ethoxymethyl substituents are clearly assigned, the remaining amide and ethylphenyl moieties appear to be disordered.
To investigate the possibility that this lack of density might be caused by increased residual mobility of the slow-reacting (SC)-TS, we performed computer simulations to map the accessible configuration space of the (RC)- and (SC)-phospho-TSA in the binding pocket. Either systematic searches for allowed conformers of the two inhibitors using Sybyl 6.4 (Tripos, St. Louis, MO) and molecular dynamics simulations based on the MAB-force field in MOLOC (Gerber and Muller, 1995; Gerber, 1998) indicated significant dynamic fluctuations of the amide portion in the slow-reacting (SC)-TS (Figure 3c). Whereas the molecular arrangements sampled for the fast-reacting (RC)-TS all cluster closely around the geometry found in the crystal structure (Figure 3a), simultaneously keeping the hydrogen bond between the amide NH and His224 intact, individual frames for the (SC)-TS indicate virtually unrestricted tumbling of the terminal phenyl portion about its central axis. In addition, the position of the inhibitor NH fluctuates widely, resulting in frequent rupture of the hydrogen bond to His224N. Figure 4 shows the statistics about the length of the NH...His224N distance. For the (RC)-TS, a value of 2.9 Å with small deviations is found, whereas this distance shows pronounced fluctuations in the corresponding (SC)-TS. The histographic distribution indicates that the hydrogen bond is only occasionally formed along the trajectory in the latter case. As this hydrogen bond is assumed to be essential for the enzyme reaction, we suggest that the slow-reacting substrate achieves a transition state productive for reaction less frequently than the fast-reacting substrate.
|
|
Discussion |
|---|
|
TopIntroductionResultsDiscussionReferences |
|---|
The kinetic experiments show that the fast-reacting enantiomer is favored by enthalpy, whereas the slow-reacting substrate compensates some of the discriminating
R – SG advantage by a beneficial entropic contribution. The enthalpic advantage of the fast-reacting (R)-amine supposedly results from its nearly perfect fit to the binding pocket of CaL B (Figure 3b). Numerous favorable van der Waals contacts and the formation of ideal hydrogen bonds immobilize and stabilize the TS in an orientation productive for the reaction sequence. On the other hand, firm immobilization is synonymous with a pronounced decrease in motional degrees of freedom, so that binding of this TS would be entropically disfavored. The slow-reacting substrate shows significant residual mobility in its transition state so that less entropy is lost upon binding. However, the residual mobility in the binding pocket corresponds to less stabilizing contacts in the transition state. Accordingly, the substrate resides less frequently in an orientation productive for the enzyme reaction and profits less from favorable enthalpic interactions. Hence for the substrates presented in this study, enantioselectivity is achieved through a difference in residual mobility of the enantiomeric inhibitors within the binding site of the protein. Clearly, an increase in the reaction temperature will result in higher residual mobility, reducing the enthalpic advantage of the fast-reacting substrate and yielding a reduced enantiopreference. Similarly, increasing the available space in the binding pocket, as observed in the Trp104His mutant, also leads to reduced discrimination (Patkar et al., 1998). Recently, Ottosson et al. reported temperature-dependent kinetic data demonstrating the enantioselectivity of CaL B towards secondary alcohols (Ottosson et al., 2002). Through molecular dynamics calculations they collected evidence that differences in the activation entropy discriminate between the two enantiomers. Similarly to our crystallographic results, in their computer simulations both substrates experience different spatial degrees of freedom and mobility in the active site. The substrate accessible volume is indicated as larger for the enantiomer preferred by entropy. This agrees with our experimental results found for the amines and thus supports their simulations with the alcohols. Conclusion
The present case study of kinetic resolution using Candida antarctica lipase shows that the combination of kinetic and crystallographic investigations together with computer simulations helps to elucidate the factors responsible for chiral discrimination. The fast-reacting enantiomer is enthalpically favored through a virtually perfect active-site complementarity, whereas the slow-reacting substrate compensates some of the discriminating free enthalpy advantage by a beneficial entropic contribution. This is due to a higher residual mobility and thus a smaller loss in entropy upon binding. It resides less frequently in an orientation productive for the enzyme reaction. With this knowledge in hand, we should now be able to understand enantiopreference towards other substrates and to tailor the lipase active site towards alternative substrates by enzyme engineering.
|
Acknowledgements |
|---|
We thank G.Bourenkov and H.D.Bartunik at the MPG outstation DESY-Hamburg for support and the MPG for beam time at the HASYLAB-DORIS. Parts of this work were presented at the 5th International Symposium on Biocatalysis and Biotransformations (BioTrans 2001), Darmstadt, Germany.
|
References |
|---|
|
TopIntroductionResultsDiscussionReferences |
|---|
Anderson,E.M., Karin,M. and Kirk,O. (1998) Biocatal. Biotransfor., 16, 181–204.
Balkenhohl,F., Ditrich,K., Hauer,B. and Ladner,W. (1997) J. Prakt. Chem., 339, 381–384.[CrossRef]
Cygler,M., Grochulski,P., Kazlauskas,R.J., Schrag,J.D., Bouthillier,F., Rubin,B., Serreqi,A.N. and Gupta,A.K. (1994) J. Am. Chem. Soc., 116, 3180–3186.[CrossRef]
Eyring,H. and Polanyi,M. (1931) Z. Phys. Chem. B, 12, 279.
Gerber,P.R. (1998) J. Comput.-Aided. Mol. Des., 12, 37–51.
Gerber,P.R. and Muller,K. (1995) J. Comput-Aided Mol. Des., 9, 251–268.
Gutfreund,H. (1995) Kinetics for the Life Sciences: Receptors, Transmitters and Catalysts. Cambridge University Press, New York; ISBN 052148586X.
Gutfreund,H. and Sturtevant,J.M. (1956) Proc. Natl Acad. Sci. USA, 42, 719–728.
Kazlauskas,R.J. (1994) TIBTECH, 12, 464–472.
Kazlauskas,R.J. (2000) Curr. Opin. Chem. Biol., 4, 81–88.[CrossRef][Web of Science][Medline]
Orrenius,C., Haeffner,F., Rotticci,D., Norin,T. and Hult,K. (1998) Biocatal. Biotransform., 16, 1–15.
Ottosson,J. and Hult,K. (2001) J. Mol. Catal. B, 11, 1025–1028.[CrossRef]
Ottosson,J., Fransson,L. and Hult,K. (2002) Protein Sci., 11, 1462–1471.[CrossRef][Web of Science][Medline]
Overbeeke,P.L.A., Orrenius,S.C., Jongejan,J.A. and Duine,J.A. (1998) Chem. Phys. Lipids, 93, 81–93.[CrossRef]
Patkar,S., Vind,J., Kelstrup,E., Christensen,M.W., Svendsen,A., Borch,K. and Kirk,O. (1998) Chem. Phys. Lipids, 93, 95–101.[CrossRef][Web of Science][Medline]
Phillips,R.S. (1992) Enzyme Microb. Technol., 14, 417–419.[CrossRef]
Phillips,R.S. (1996) TIBTECH, 14, 13–16.
Reetz,M.T. (2002) Curr. Opin. Chem. Biol., 6, 145–150.[CrossRef][Web of Science][Medline]
Schmid,H. and Verger,R. (1998) Angew. Chem., Int. Ed., 37, 1608–1633.[CrossRef]
Uppenberg,J., Ohrner,N., Norin,M., Hult,K., Kleywegt,G.J., Patkar,S., Waagen,V., Anthonsen,T. and Jones,T.A. (1995) Biochemistry, 34, 16838–16851.[CrossRef][Medline]
Received September 30, 2002; revised March 25, 2003; accepted April 8, 2003.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Y. Nakagawa, A. Hasegawa, J. Hiratake, and K. Sakata Engineering of Pseudomonas aeruginosa lipase by directed evolution for enhanced amidase activity: mechanistic implication for amide hydrolysis by serine hydrolases Protein Eng. Des. Sel., July 6, 2007; (2007) gzm025v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||