Theoretical conformational analysis of peptides. Evolution of a strategy and its application to cholecystokinin analogs

Abstract

Scheraga's ECEPP program has been used to determine the relationship between one-point energies and the energies minimized by a quadratic procedure, for different sized [Formula: see text] dihedral angle grid searches of simple peptides. Based on these trials, a new subroutine, INIT, has been written and incorporated into the program. This subroutine calculates the one-point ECEPP energies of up to 200,000 random permutations of [Formula: see text], 180 and (χ1) = −60, 180 with all ω = χ≠1 = 180; after each 40,000 permutations, the structures having the 10 lowest energies are ordered and selected for minimization. The ECEPP program has been modified further to provide an MMP2(85) input file as part of its normal output. This allows structures located by INIT to be minimized automatically, and then refined by the Newton–Raphson minimization and MMPEP parameters of MMP2(85).During the development of this protocol, it has been found that MMPEP correctly reproduces the C7 conformational preference of the alanyl dipeptide in nonpolar media, and also the experimentally observed shift to αR and PII conformations in polar solvents, when the dielectric constant ε = 78.5 D. This dielectric constant has, therefore, been selected for a conformational analysis, using INIT/MMPEP, of the peptides Gly-Trp-Met-Asp-Phe-NH2 (GLMAP), Gly-D-Trp-Met-Asp-Phe-NH2 (GDMAP), and Gly-Gly-Met-Asp-Phe-NH2 (GGMAP); GLMAP is CCK-5, a cholecystokinin fragment possessing anticonvulsant activity, which is found in the brain. The analogs GDMAP and GGMAP are not active, or much less so than GLMAP. One strongly preferred structure has been found, in each case, for GLMAP (GLMAP3) and GDMAP (GDMAP39). These structures are very similar, in the nature of the edge-to-face stacking of their Trp and Phe aromatic rings, but the C-terminal regions differ. The conformation of the C-terminal tetrapeptide of GLMAP3 is identical to a previously calculated structure of CCK-4. A comparison of the C-terminal regions of GLMAP3 and GDMAP39 with the structure of 5,5-diphenylhydantoin suggests that the mechanisms of anticonvulsant action of phenytoin and CCK-5 are not the same. In the absence of an aromatic–aromatic interaction, no clear-cut conformational preference is found for GGMAP. The structure of GGMAP whose C-terminal region is the same as that of GLMAP3 (GGMAP20) is 1.51 kcal/mol higher in energy than the lowest energy structure, and lacks the hydrophobic wall provided by the Trp residue of GLMAP3.