Published in J . Chem. Soc., Perkin Trans 2, 1994, 3.
(c) Royal Society of Chemistry

The Stereoelectronic Influence of Fluorine in Enzyme Resolutions of a-Fluoroesters

David O'Hagan[a] and Henry S Rzepa[b]

[a]Department of Chemistry, University of Durham, South Road, Durham, DH1 3LE, UK.[b] Department of Chemistry, Imperial College of Science, Technology and Medicine, London, UK SW7 2AY, E-mail (Internet): rzepa@ic.ac.uk
The very high enantioselectivity shown by two different lipase enzymes towards [alpha]-fluoroesters is shown from quantitative PM3 and ab initio 6-31G* SCF-MO calculations to arise from stereoelectronic factors, which are predicted to be reduced when the transition states are modelled in the condensed aqueous phase.
Selectively fluorinated compounds are increasingly being deployed for pharmaceuticals and fine chemicals applications and the development of new methods towards the generation homochiral fluorinated compounds is a current challenge.[1] Synthetic access to such compounds, particularly with -F or -CF3 at the stereogenic centre, is limited[2] and in this regard enzymic resolutions[3] and microbial reductions[4] have offered a versatile and successful approach to such systems. In the arena of biotransformations two hydrolytic resolutions of a-fluoroesters have been reported,[5,6] where different lipase enzymes have displayed an almost complete distinction (>99%ee) between hydrogen and fluorine atoms (Scheme). Although the steric influence of fluorine has been detected[7] in certain enzyme systems, the perturbation is small and it is difficult to rationalise the selectivity of the lipase resolutions on the basis of steric effects alone. We have recently suggested[7] that the stereoelectronic influence of fluorine in these two bio-transformations is a major controlling factor where the serine hydroxide nucleophile will prefer to attack anti to the fluorine as illustrated in the Newman projection (Scheme). We report here quantitative gas phase and solution theoretical models for this reaction which demonstrate that the enantioselectivities observed are entirely consistent with such an effect.

The model reaction between methyl 2-fluoropropionate and methoxide or hydroxide anion was studied at the PM3 or ab initio 6-31G* SCF-MO levels respectively, methods which have been successfully applied to stereoelectronic rationalisation of several other reactions.[8] It can be envisaged that in the anti transition state (1, Figure 1a) there is an n-[sigma]* stabilising interaction of the Ahn-Eisenstein type[9], between the active site serine hydroxide nucleophile of the enzyme and the low energy [sigma]* orbital anti-periplanar to the C-F bond of the substrate. This interaction is absent in the syn conformation (2, Figure 1b).

The calculated discrimination in favour of 1 (Figure 1a) is 2.4 or 2.8 kcal mol[-1] at the PM3 or 6-31G* levels respectively (Table). The calculated PM3 entropy difference between 1 and 2 of 0.4 cal K[-1] mol[-1] results in a free energy difference of 2.5 kcal mol[-1 ]at 300K.[ ]Energy differences of this magnitude correspond to a anti/syn ratio of ~100:1. The enthalpy and free energy differences reduce to 1.3 and 0.7 kcal mol[-1] respectively when the PM3-COSMO aqueous solvation model[10 ]is applied, implying a significant reduction in specificity. The origins of these effects are evident in the PM3 energies of the oxy-anion lone pair HOMO orbitals. Those for 1 ~ 0.17 - 0.21 eV more stable than 2 as a result of stabilisation by interaction with the [sigma]* orbital, which also induces a small contraction in the O-C forming bond. The orbital energy differences are reduced to ~ 0.05 eV when the PM3-COSMO model is applied, largely because the oxy-anion orbital is considerably stabilised by the solvation model (Table). This in turn increases the energy difference between this orbital and the unoccupied C-F [sigma]* orbital from ~ 10 eV to ~ 11.5 eV, thus reducing the perturbational stabilisation. It can be expected that a substrate bound to the enzyme will be essentially desolvated at the active site[11], and therefore such stereoelectronic control will be maximised under these conditions.

Models of the common hydrolytic enzymes e.g.(C. cylindracea,[12] pig liver[13] etc.) have been proposed on the basis of substrate/activity profiles which attempt to map the steric constraints of the active sites. In the light of the current study, the efficacy by which individual lipases resolve a-fluorinated esters should afford additional spatial information deduced from the stereoelectronic interactions allowed at that active site.

Theoretical calculations were carried out at the restricted Hartree-Fock level (RHF) PM3 semi-empirical method, as implemented in the MOPAC 93 program,[8 ]using a relative permittivity of 78.4 for water and 60 surface segments per atom for the COSMO model. Ab initio calculations were performed using the Gaussian-92 program. All structures were optimised using the eigenvector following algorithm, followed by a vibrational analysis to characterise the transition state. Use of the COSMO option approximately doubles the computation time compared with a gas phase calculation. Computer readable files for Apple Macintosh and Microsoft Windows systems in Quicktime(TM) video animation format illustrating the three dimensional properties of 1 and 2 are available for general access from the Gopher+ server gopher.ch.ic.ac.uk. These files will reside in the rzepa/Royal_Society_of_Chemistry/Perkin_Trans_2/3_ZZZZZZ directory for a period of at least two years from the publication of this paper. A description of how to visualise such material, together with appropriate programs is available from the same source.

References

1. D. Seebach, Angew. Chem. Int. Ed. Engl., 1990, 29, 1320; J. Mann, Chem. Soc. Rev., 1987, 16, 361; J. T. Welch, Tetrahedron, 1987, 43, 3123.

2. K. Iseki, T. Nagai and Y. Kobayashi, Tetrahedron Letts., 1993, 34, 2169; R. W. Lang, Drug News and Perspectives, 1991, 4, 13; G. L. Hann and P. Sampson, J. Chem. Soc. Chem. Commun., 1989. 1650; E. Differding and R. W. Lang, Tetrahedron Letts., 1988, 29, 6087.

3. D. O'Hagan, N. A. Zaidi and R. B. Lamont, Tetrahedron Asymmetry, 1993, 4, 1703; D. O'Hagan and N. A. Zaidi, J. Chem. Soc. Perkin Trans I, 1992, 947; T. Yamazaki, N. Okamura and T. Kitazume, Tetrahedron Asymmetry, 1990, 1, 521; T. Yamazaki, T. Ohnogi and T. Kitazume, Tetrahedron Asymmetry, 1990, 1, 215; J. T. Lin, T. Yamazaki and T. Kitazume, J. Org. Chem., 1987, 52, 3211; T. Kitazume and J. T. Lin, J. Fluorine Chem., 1987, 34, 461.

4. T. Fujisawa, I. Ichikawa and M. Shimizu, Tetrahedron Asymmetry, 1993, 4, 1237.

5. P. Kalaritis, R. W. Regenyne, J.J. Partidge and D. L. Coffen, J. Org. Chem., 1990, 55, 812.

6. T. Kitazume, K. Murata and T. Ikeya, J. Fluorine Chem., 1986, 31, 143.

7. L. Dasaradhi and D. O'Hagan, Bio. Med. Chem. Letts., 1993, 3, 1655.

8. O. Casher, D. O'Hagan, C. A. Rosenkranz, H. S. Rzepa and N. A. Zaidi, J. Chem. Soc., Chem. Commun, 1993, 1337; M. S. Baird, J. R. Al Dulayymi, H. S. Rzepa and V. Thoss, J. Chem. Soc., Chem. Commun, 1992, 1323; B. Halton, R. Boese and H. S. Rzepa, J. Chem. Soc., Perkin Transactions 2, 1992, 447

9. S.S. Wong and M. N. Paddon-Row, J. Chem. Soc. Chem. Commun., 1990, 456; N. T. Anh and O. Eisenstein, Nouv. J. Chem, 1977, 1, 61.

10. A. Klamt and G. Shürmann, Perkin Transactions, 1993, 799; H. S. Rzepa, and G. Suñer, J. Chem. Soc., Perkin Trans 2, 1994, 1397; H. S. Rzepa, and G. Suñer, unpublished results

11. M. J. S. Dewar, Enzyme, 1986, 36, 8; W. N. Lipscomb, Acc. Chem. Res., 1982, 15, 232.

12. T. Oberhauser, K. Faber and H. Griengl, Tetrahedron, 1989, 45, 1679; R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport and L. A. Cuccia, J. Org. Chem., 1991, 56, 2656; K. Hult and T. Norin, Pure & Appl. Chem., 1992, 64, 1129.

13. P. Mohr, N. Waespe-Sarcevic, C. Tamm, K. Gawronska, J. K. Gawronski, Helv. Chim. Acta., 1983, 66, 2501.

Table. Calculated Properties of Transition States 1 and 2.

Figure 1. Calculated PM3 geometry of transition state 1 (a) and 2 (b).