Molecular Surfaces
10. Protein-Protein Interactions

Michael L. Connolly

1259 El Camino Real, #184
Menlo Park, CA 94025

The Author's Web page
The Author's software web page

We now turn our attention to protein-protein and subunit-subunit interfaces. Subunit-subunit interfaces have been studied by (Janin and Chothia, 1988). Protein-protein interfaces of 15 protease-inhibitor complexes and 4 antibody-portein antigen complexes have also been studied by Janin and Chothia (1990). Duquerroy, Cherfils and Janin (1991) have also studied known protein-protein interfaces, and they found that non-native alternative predicted dockings often had as many hydrogen bonds and interfacial area. Susan Jones and Janet Thornton (1996) have also written a review of protein-protein interactions. Milligan (1996) has review interactions between actin and myosin. Jack Johnson (1996) has reviewed capsid monomer interactions in icosahedral viruses. The structure of hyperthermophilic bacteria's glutamate dehydrogenase enzyme "reveals a key role for ion-pair networks in maintaining enzyme stability at extreme temperatures" (Yip, Stillman, Britton, Artymiuk, Baker, Sedelnikova, Engel, Pasquo, Chiaraluce, Consalvi, Scandurra and Rice, 1995). Zachmann, Kast, Sariban and Brickmann (1993) and Morgan, Miller and McAdon (1979) have studied symmetric dimers. Ban, Day, Wang, Ferrone and McPherson (1996) have determined the structure of an anti-anti-idiotype antibody and found it to be self-complementary. Davies and Cohen (1996) have studied protein-antibody interactions using crystallography. Pellegrini and Doniach (1993) have modeled antibody specificity. Wells (1996) has studied the growth hormone receptor complex. Hubbard and Argos (1994) studied cavities and packing at protein interfaces. Lawrence, and Colman (1993) at the CSIRO, Melbourne, Biomolecular Research Institute have developed methods to study shape complementarity at protein/protein interfaces and have shown that antibody-antigen interfaces are generally less well-packed than other protein-protein of interfaces. Varshney, Brooks, Richardson, Wright and Manocha (1995) have computed interfacial surfaces, that is, surfaces that pass between two protein molecules.

Antibody-antigen interactions have been reviewed by Wilson and Stanfield (1993) and protein-peptide interactions have been reviewed by Stanfield and Wilson (1995). The interactions of thrombin and peptides based upon portions of the thrombin-receptor sequence have been studied (Mathews, Padmanabhan, Ganesh, Tulinsky, Ishii, Chen, Turck, Coughlin and Fenton, 1994; Mathews and Tulinsky, 1995).

Molecular surfaces have been applied to the protein-protein docking problem. The protein-protein docking problem is the prediction of a complex between two proteins given the three-dimensional structures of the individual proteins (Zielenkiewicz and Rabczenko, 1984; Santavy and Kypr, 1984). Connolly (1986d) developed a method based upon juxtaposing knobs and holes (critical points in the solid angle curvature function). It worked for hemoglobin, but not BPTI/trypsin. Another shape-matching algorithm, using a more sophistical shape measure, performed similarly (Connolly, 1992). The approach of mating knobs and holes has also used by Wang (1991). The critical-point approach has been made workable by a group at the National Cancer Institute (Norel, Fischer, Wolfson, Nussinov, 1994; Norel, Lin, Wolfson and Nussinov, 1994; Norel, Lin, Wolfson, Nussinov, 1995; Fischer, Lin, Wolfson and Nussinov, 1995).

Shoichet and Kuntz, 1991; adapted their protein-ligand DOCK program to protein-protein docking. Jiang and Kim (1991) developed a protein-protein docking method that performed a search over three degrees of rotational freedom, but avoided a search over the three translational degrees of freedom by a clever use of surface normal vectors. The method of Bacon and Moult (1992) docks together surface patches represented by spirals of points. There is a search over the rotational degree of freedom about the protein-protein axis for each juxtaposition of spiral patches, and a check for a correlation in the relative heights of points along the spiral. These methods work well for individual protein structures taken from the cocrystal, but not as well using structures from individual crystallizations. Protein docking methods using a simplified, one-sphere-per-residue model for the protein have been used to predict complexes involving antibodies (Cherfils, Duquerroy and Janin, 1991; Cherfils, Bizebard, Knossow and Janin, 1994). Cherfils and Janin (1993) have reviewed recent protein docking work. Grid-search protein docking methods involving shape complementarity have been developed by Badel, Mornon and Hazout (1992), and by Walls and Sternberg (1992). Searches over the three relative rotational degrees of freedom have been used to predict a ferritin-antibody complex (Helmer-Citterich and Tramontano, 1994; Helmer-Citterich, Rovida, Luzzago and Tramontano, 1995).

Katchalski-Katzir, Shariv, Eisenstein and Friesem (1992) developed a grid search method that worked quickly because the positionings were evaluated by a Fourier transform method of evaluating surface complementarity. Gerstein (1992) has also developed a Fourier method for comparing protein surface shape, and has used it to compare antigen-combining sites. Duncan and Olson (1994) have developed a docking algorithm that is based upon approximating the protein surface by parametric surfaces. Electrostatics are involved in the Docking of Cytochrome f with Plastocyanin.

Database searching methods have been used by Seidl and Kriegel (1995). Artificial intelligence methods have been applied by Ackermann, Herrmann, Kummert, Posch, Sagerer, Schomburg (1995). Cummings, Hart and Read (1995) have used the desolvation energies of amino acids to rank predicted protein-protein complexes. Hydrophobicity has been used in some protein-protein docking work (Vakser and Aflalo, 1994; Vakser, 1995). Jackson and Sternberg (1995) analyzed a set of predicted protein complexes (obtained from Shoichet and Kuntz) and have shown that the smooth molecular surface is more useful than the Lee and Richards accessible surface in evaluating the hydrophobic contribution to binding. An approach to peptide-protein docking combining virtual reality and genetic algorithms can be found at the STALK molecular docking systemof Argonne National Lab.

Below is a (wall-eyed) stereo pair of the trypsin / trypsin inhibitor interface. The inhibitor (BPTI) is red, and trypsin is green. The front surface has been clipped.

[ 1. Introduction ] [ 2. Physical Molecular Models ] [ 3. Electron Density Fitting ] [ 4. Molecular Graphics ] [ 5. Solvent-Accessible Surfaces ] [ 6. Molecular Surface Graphics ] [ 7. Molecular Volume and Protein Packing ] [ 8. Shapes of Small Molecules and Proteins ] [ 9. Structure-based Drug Design ] [ *** 10. Protein-Protein Interactions *** ] [ 11. Surface Biology, Chemistry and Physics ] [ 12. Bibliography ]

All material in ths article Copyright © 1996 by Michael L. Connolly

NetSci, ISSN 1092-7360, is published by Network Science Corporation. Except where expressly stated, content at this site is copyright (© 1995 - 2010) by Network Science Corporation and is for your personal use only. No redistribution is allowed without written permission from Network Science Corporation. This web site is managed by:

Network Science Corporation
4411 Connecticut Avenue NW, STE 514
Washington, DC 20008
Tel: (828) 817-9811
Website Hosted by Total Choice