(1) College of Pharmacy, Chungnam University, Daejeon 305-764, Korea
(2) Laboratorium fuer Physikalische Chemie, ETH-Zentrum, 8092 Zurich,
Switzerland
Currrently the solution structure of small proteins can be most efficiently refined by the application of NOE and J value restraining in conjunction with molecular dynamics simulatios. NMR experiments, however, yield another important parameter, the chemical shift of each resonance. This parameter is in principle the primary observable. It contains uniform structural information and is sensitive to conformation. To enable the effective use of chemical shift data in structure refinement there has been continuous efforts to improve the calculation algorithm (Cross et al.1985) and to optimise the structural parameters (Osapay et al. 1991, Williamson et al. 1992) in order to better correlate the predicted chemical shift with the 3D structure of protein. In this work it is shown how the observed chemical shift values can be used to restrain the structure during a molecular dynamics simulation.
The observable chemical shift can be approximated generally by a conformation- independent local term and conformation-dependent non-local term. For the local effect, mainly arising from diamagnetic and paramagnetic shielding around the nucleus, we used new reference values derived from the peptide modeling instead of 'random coil' values (Bundi et al. 1979). The important components of the non-local effect are ring current effects from the nearby aromatic rings, magnetic anisotrpy from the peptide groups and electric field effects of the polar groups. All these terms were modelled classically in order to calculate the structure dependent chemical shift values of Ca Hydrogens of protein. The discrepancy between the observed and calculated values was then expressed as a harmonic potential function and implemented to new GROMOS force field (van Gunsteren et al. 1996).
Normal free MD and chemical shift restrained MD calculations were performed using the protein BPTI and show that chemical shift restraining MD can give complementary information to the conventional methods.
References
Bundi, A. and Wuthrich, K. (1979) Biopolymers 18, 285-297.
Cross, K.J. and Wright, P.E. (1985) J. Magn. Reson. 64, 220-231.
Osapay, K. Case, D.A. (1991) J. Am. Chem. Soc. 113, 9436-9444.
van Gunsteren, W.F. (1996) GROMOS96 Reference Manual, Zurich.
Williamson, M.P. Asakura, T. Nakamura, E. and Demura, M. (1992) J. Biomol. NMR 2, 83-98