programs on the protein docking benchmark 4.0. They demon-
strated that the docking performance depends on the target and no
program is able to perform well for all the cases of the benchmark,
though some algorithms are more robust than others [5]. They
concluded that ZDOCK3.0.2, SDOCK, and PIPER resulted in the
relatively higher success rates with 30.7%, 22.7%, and 21.0% for top
ten predictions. Moreover, ZDOCK3.0.2, ATTRACT, and FRO-
DOCK are all good choices if more predictions can be evaluated
which is often the case in a postdocking approach [5].
Not only the suitability of the programs but also other factors
might affect the docking results. Vajda et al. [38] demonstrated
that the success in protein–protein docking experiment depends on
three factors: (a) the quantity of conformational changes upon
binding, (b) the area of interface surface, and (c) the hydrophobi-
city of the interface which allows to classify the docking tasks into
five groups based on the difficulty of docking [16]. Briefly, Vajda
et al. [38] postulated that docking is facilitated by low conforma-
tional change upon complex formation (Ca RMSD<2A ̊)(see
Note 1), a large surface interface (700–1000 A ̊), and high hydro-
phobicity (DGdes<4.0 kcal/mol).
Some of the most recent examples of application of protein–-
protein docking (PPD) software to water-soluble and membrane-
anchoring proteins described below provide a general overview on
the potential and importance of the method.
In the field of cancer research Ramatenki et al. focused on the
ubiquitin-conjugating enzyme E2D4 and its binding partner, ubi-
quitin ligase CHIP. They used a range of tools to identify putative
binding sites, and dock the ligase with PatchDock server [19]. The
resultant complex structure was used for further virtual screening.
Unfortunately, the study lacks any experimental validation. How-
ever, it is an example of application of protein–protein docking for
design of small-molecule ligands.
Selent et al. [39] used protein–protein docking with Patch-
Dock to model the survivin/CDK4 complex. Survivin is the smal-
lest known inhibitor of apoptosis protein (IAP) and a valid target
for cancer research [39]. Selent et al. also assessed electrostatic
complementarity using APBS software and shape complementarity
based on fractal approach [40–43]. They performed molecular
dynamics simulation to obtain a refined survivin/CDK4 model
that can be used for structure-based design of inhibitors modifying
its interface recognition (seeNote 2).
Research on viral infections can also take advantage from PPD.
A very interesting and promising application was recently reported
by Yoshiyuki Suzuki [42]. It focuses on the problem of interspecies
transmission of viruses, which poses a potential threat for present
medicine. He investigated interactions of a viral protein from mea-
sles virus with a transmembrane protein—signaling lymphocyte
activation molecule (SLAM). The protein is known to be one of
Protein-Protein Docking 291