Authors performed detailed analysis of the contribution
of electrostatic interactions on the protein-protein binding.

Systems considered:
They considered 4 protein cplxes: BN-BS, Ras-Raf, NC41 -N9 (neuromidase),
human growth hormone and its receptor.

Method:
They used FDPB implemented in Delphy for the electrostatic calc, and SURFACV program
for SASA calc.

Findings:
1) For the complexes considered they determined different modes of binding...
2) The total electr. contribution to bind is found to be inversly correlated with
burried total and non-polar surface area.
3) The largest contributions to the electrostatic free energy of binding comes from
desolvation and intermolecular interactions
4) In the 4 complexes examined the total el contribution to binding is similar to that calculated
for the interfacial residues alone
5) The contribution of charged residues alone correlates poorly with the total electrostatic
contribution to binding (i.e. signifficant is contribution from polar - charged res. interactions ...)
6) The change in intramolecular interaction make relatively small contribution to
Exception is E73 in BN . Stabilizing interactions of this residue with R72, R83, R87 and K27
is enhanced upon binding.

Findings related to BN-BS (complex betwen the BN (wt) and BS ( double mutant Cys-> Ala (40,82)), we have not considered this cplx)
1) In the case of BN-BS they found that the net electrostatic effect is almost neutral (+3.6 kcal/mol)
(For Ras-raf it is stabilizing 9-11.5), and for the other two cplxes destabilizing.
2) The desolvation penalty large despite the smaller surface area
(i.e. the expectation that large interface bury more charged and polar groups and
causes the desolvation penalty does not exactely hold for BN-BS)
3) Stabilizing interactions of the E73 residue in BN with R72, R83, R87 and K27
is enhanced upon binding.

4) The interfacial amino acid residues that have the largest electrostatic
contribution to binding

D39, E76, R87, Thr42, H102 enhence binding

K62, E60,K27, D35 decrease stability because of the desolvation penalty being larger than
the favorable intermolec int.

E73 desolvation penalty + unfavorable intermolecular int , BUT its stabilizing interactions
with R72, R83, R87 and K27 is enhanced upon binding.

How this corelate with the COMBINE model:

All these residues appear in the model as the most important, but since we consider residue to residue interaction
in MOSTcases : favorable interaction ( electrostatic: 35-59, 39-83, 39-83, 39-87, 39-62, 80-27)
(vdw: 29-102, 3-102, 35-60, 36-102, 38-27, 39-83) are proportial with binding affinty,
only exception (among the 13 HIGHESTweighted is the vdw 76-59 interaction).

For better comaprison influence od SASA terms to binding affinity should be considered + TOTAL intermolecular interaction of each AA )

~
~