back to Table of Contents

16. Protein-protein complexes

16.1 Overview

EGAD has a few tools that may be useful for analyzing protein-protein complexes. The EGAD energy function is able to predict the effects of mutations on protein complex formation to within 1kcal/mol (Pokala and Handel 2005). Figure 16.1.1 shows the thermodynamic cycle used for predicting these effects. These data are shown in Figure 16.1.2 and Figure 16.1.3.

As discussed in the template pdb file section, if a subunit has multiple chains that are not disulfide linked, the chain IDs for these should be defined as a SUPER_CHAIN.

As discussed below, EGAD can also be used to design protein-protein complexes; this is an area of active research in the lab. See the Handel lab website. If you are seriously interested in a protein-complex design project, you may wish to collaborate with the lab.

16.2 JOBTYPE separate_complex or JOBTYPE split_complex

These jobs take a protein complex TEMPLATE structure and return a pdb file that has the two subunits separated along the centroid-centroid axis such that their solvent-accessible surfaces are ≥ 15Å apart (the larger subunit is stationary). If two chains are part of the same SUPER_CHAIN, they are moved together.

It also creates the OUTPUT_PREFIX.complex_energy.out file that contains energies of the free and bound complexes, as well as a list of interface residues (interface residues are defined as those that undergo a change in SASA upon complex formation), and their SASA change. The "level" entries describe how a residue is defined as being at the complex interface. Residues that undergo a change in SASA upon binding are "level_1." Residues whose energies, but not SASA, change are "level_2"

Interface residues are also listed in a comma-delimited line (line begins with COMMA_INTERFACE_RESIDUES); this line may be pasted into Rasmol for selecting the interface residues.

16.3 JOBTYPE complex_formation_energy

This job is similar to JOBTYPE separate_complex, except that the energies of residues are also calculated. The energies of interface residues are listed in OUTPUT_PREFIX.complex_energy.out. Interface residues are defined as described for JOBTYPE separate_complex, except that residues that undergo an energy change, even without a SASA change, are also defined as being at the interface.

Due to the fact that the energy is calculated for each residue individually, this job takes much longer than JOBTYPE separate_complex. The only advantage this job offers is the energy for each residue. It should be noted that there is very little difference between interface residues defined with energies compared to those defined by SASA.

16.4 JOBTYPE interface_alanine_scan and interface_saturation_mutagenesis

These two jobtypes automatically identify the interface residues, and then proceed to perform the defined scanning mutagenesis. The data are returned in a format similar to that described for scanning mutagenesis, except that the differences between the complex and free with respect to the wild-type sequence are returned in the .interface_scan file.

If the structures of the free forms are known, and in a pdb file which contains the structure of each subunit solved individually, with the coordinates transformed such that the subunits are 15Å apart, this structure may be listed in the inputfile as:
FREE structure_of_free_subunits.pdb
Otherwise, the statically separated structure will be generated and used.

The caveats and issues discussed for normal scanning mutagenesis also apply here. For example, unless OUTPUT_COORD_FLAG is set to 1, coordinates of the individual mutants will not be returned. Unless CLEAN_UP_FLAG is set to 1, the scanning mutagenesis files of the complex and free structures will not be saved.

Examples of interface alanine scanning and saturation mutagenesis are in examples/protein_complexes/ala_scan/ and examples/protein_complexes/sat_mut/ respectively. These files have a format similar to monomer scanning mutagenesis output files. As with mutant protein stabilities, predicted binding ∆∆Gs are unreliable if the difference if |∆clashes| > 2.

16.5 Implementation

The functions for these jobs are all in complex_formation_energy.cpp, and are called by main.
The interface scan jobs (interface_scan) creates the free structure and identifies interface residues by calling complex_formation_energy_master. It then generates inputfiles for scanning mutagenesis of the bound and free forms. These are placed in a batch inputfile. A batch master is launched as described above. The batch master launches the actual scanning mutagenesis jobs. Parallelization, if requested, occurs at the level of lookup table generation and rotamer optimization for the scanning mutagenesis of the bound and free forms. After these jobs are completed, the master parses the .scanning_mutagenesis files for the free and bound states, and calculates the energy differences.

back to Table of Contents