A Comparison Between Crystallographic Data and Molecular Mechanics Calculations on the Side Chain and Backbone Conformations of Steroids
Steroid conformations observed in neutron and X-ray crystal structure determinations have been compared with structures calculated using two molecular mechanics programs, MM2/P and QCFF/MCA. Analysis of the data for a very accurate low temperature (123°K) neutron diffraction determination of 20-methyl-5-pregnene-3ß,20-diol indicate that the maximum differences between observed and calculated nonhydrogen atom bond lengths and angles are greater for QCFF/MCA calculations (0.04Å and 5°) than for MM2/P calculations (0.02Å and 2.5°). While MM2/P calculations adequately reproduce bond lengths and angles, they do not reproduce the observed torsion angles and overall conformations of the steroids examined. The narrow range of side-chain conformations seen in 85 pregnane structures having a 20-one substituent, the predictable substituent influence apparent in the data, and the bond angle deformation observed in the 16ß-substituted structures strongly suggest that crystallographically observed conformers seldom deviate from lowest minimum energy positions regardless of hypothetical broad energy minima, small differences in energy between metastable states, and small barriers to rotation. The crystaollographic data on Steroids having 4-en-3-one composition reveal patterns of torsion angle variation in the A rings that are not found in the structures subjected to energy minimization. For example the C(3)-C(4)=C(5)-C(10) torsion angle is observed to deviate from planarity by an average of -5.8°. Since this pattern is not reproduced in the calculations it seems likely that constraints upon double bonds used in these molecular mechanics programs are unrealistically weighted. Using the crystallographic data as a guide to the development of better parameters governing non-bonding interactions and rotation about single and double bonds could lead to better approximations of the shape and depth of local and global energy minima and the barriers between them.
KeywordsCarbonyl Progesterone Androgen Epoxide Diol
Unable to display preview. Download preview PDF.
- 2.Barton, D. H. R.: 1955, Experientia Suppl. 11, pp. 121–137.Google Scholar
- 3.Duax, W. L., Norton, D. A.: 1975, Atlas of Steroid Structure, Vol. 1, Plenum Press, New York.Google Scholar
- 4.Duax, W. L., Griffin, J. F., Weeks, C. M.: 1982, Atlas of Steroid Structure, Vol. 2, Plenum Press, New York.Google Scholar
- 6.Fronckowiak, M. D., Doctoral Thesis, State University of New York at Buffalo, 1982.Google Scholar
- 8.Duax, W. L., Osawa, Y.: 1980, Cryst. Struct. Comm., 9, pp. 267–270.Google Scholar
- At the time we made these calculations we had copies of MM1, MMP1 (the MM1 program with features for improved handling of π systems), and MM2 (a second generation MM1 program). Since we found the MM2 program to be a definite improvement over the MM1 program and the MMP1 program to do a better job of matching bond lengths and angles involving π systems, we incorporated the π handling subroutines into MM2 and refer to this program as MM2/P. We found the results in better agreement with the X–ray results and are certain that in the hands of the programs author the results could well be better than those we have calculated. MM1: Allinger, N.L., et al.: As amended by D. A. Pensak, 1981, QCPE, 13, p. 400. MM2: Allinger, N. L., Yuh, Y. H.: 1981, QCPE, 13, p. 395.Google Scholar
- 10.Huler, E., Sharon, R., Warshel, A.: 1977, QCPE, 11, p. 325.Google Scholar
- 12.Duax, W. L., Griffin, J. F., Strong, P. D., Ulick, S., Funder, J. W.: J. Amer. Chem. Soc., submitted.Google Scholar