Previous Index Next
Page 24
(previous) (Page 000024) (next)
extended to include conformations, vibrations and heats of hydro-
genations of nonwconjugated olefins25; conformations and vibrations
of amide—alkane rings (lactams)26; heats of sublimation and crystal
structures of hydrogen bonded crystals of amides15 and carboxylic
acidsll, as well as some of their dipole momentslsu11 and enthalpies
of dimerization in the gas phase ?.

The families of alkanes, amides and carboxylic acids contain
the constituents which comprise most of the amino acid residues of
protein chains, and indeed the major motivation for deriving a
consistent force field for these families was its potential
application for biopolymers. However, a comprehensive consistent
force field for biological molecules should include more groups
e.g. aromatic and heterocyclic rings, carbohydrate, phosphate and
sulfur groups; it should also extend considerably the consistent
analysis of intra—molecular energy parameters pertaining to poly—
peptide and polynucleotide chains. This could be done fully only
if more experimental data could be accumulated to supply the specific
needs of the consistent force field analysis.

Among the past accomplishments of the consistent force field
method, perhaps the most interesting are those related to the non“
bonded interactions, namely the Van der Heals repulsions and
attractions, as well as the Coulomb forces between partial charges
and the hydrogen bond. '

The Lennard—Jones (L3) potential is commonly considered to be
a "12—6" potential, that is, the exponent n in the repulsive term
Ar'n is taken to be n=12 (see above). Examining the LJ potential
for alkanes, we found that the LJ parameters optimized for intramole—
cular interactions were too low for intermolecular interactions, while
the LJ parameters optimized for intermolecular interactions were
too high fornintramolecular interactions. These trends indicated
that the r_14 dependence is too steep. Trying various values of n
we found that the “9—6” potential (i.e. n=9) was the best for alkanegl.
The same trend was observed in carb0xylic acids, and to a lesser
extent in amides.

Extending the consistent force field to hydrogen bonded amide
crystalsls; we put all hydrogen—bond potentials available in the
literature to the least squares test. The unequivocal, at that time
surprising, result was that there is no need for a Special potential
for the hydrogen bond. The objective parameter—fitting algorithm
of the least—squares led to reasonable Lennard—Jones and partial
charge parameters, while the special functions, representing the
expected covalent character of the hydrogen bond, received neglibibly
Small or ill—determined parameters (i.e. having large standard-
deviations). Whether such functions were included or omitted did

Previous Index Next