Computational Approaches in Peptide and Protein Design: An Overview - Peptide and Protein Design for Biopharmaceutical Applications - page 13

It also includes conformation C 7 eq , located into the 'pass' zone (that with
c values around 60 - 30) extended toward the second main region of
plausible conformations centred around ( 90,0 ). This region spreads
through both left quadrants and contains conformations close to the
right-handed a-helix (the 'alpha R' region). The third region is the smal-
lest one, located in the upper-right quadrant, and contains conformations
close to the left-handed a-helix (the 'alpha L' region). Only a few con-
formations were experimentally found in the lower-right quadrant,
which includes conformation C 7 ax . Interestingly, experimental distribu-
tions in the regions corresponding to the right- and left-handed a-helices
show a somewhat diagonal shape.
Sampling of the Ramachandran map with the AMBER and
CHARMM22 force fields (Figure 2.1(a,b)) resulted in two main regions
located in the left half of the map. For AMBER, population of the alpha R
region was significantly higher than that of the beta region, and only a
few conformations were found in the alpha L region (after performing
additional molecular dynamics simulations; see [1] for details). These
results are not consistent with the experimental data in Figure 2.1(e); one
may expect that energy estimations involving the AMBER force field will
lead to large numbers of false positives in the alpha R region when
sampling conformational possibilities of peptide systems. The results
obtained with CHARMM22 (Figure 2.1(b)) showed a high population
of the beta region; however, the second equally populated region is
shifted down to the lower-left quadrant and possesses only a small over-
lap with the experimentally determined alpha R region.
Much more consistent with experimental data were the
Ramachandran maps obtained with the GROMOS and OPLS-AA ('AA'
stands for 'all atoms', meaning 'including all hydrogens') force fields
(Figure 2.1(c,d)). In both cases, the sampled beta and alpha R regions
were located close to the experimentally determined ones, and popula-
tions of the beta regions clearly exceeded populations of the alpha R
regions. Sampling also found conformations belonging to the pass zone
between two regions, as well as conformations in the upper-right quad-
rant. These latter conformations were more frequent in the GROMOS
map, but more consistent with the experimentally observed alpha L
region in the OPLS-AA map. One may expect that energy estimations
based on the GROMOS or OPLS-AA force fields will lead to more
adequate sampling of peptide systems with less false positives or nega-
tives. In fact, both force fields are currently more frequently used in
conformational

calculations of peptides and proteins than either

AMBER or CHARMM.