Any student of cell biology should be impressed by the tremendous complexity of even the simplest living cell, and how it manages to maintain, regulate, and use that complexity to achieve its functions. The idea of any cell being a mixed-up bag of enzymes is long gone.
Every known living cell is a highly organized entity with a densely crowded interior. Proteins within a cell do not act in isolation; they act together to carry out their functions.
Serious problems may arise when a protein-protein interaction is disturbed (e.g. a hormone may not be delivered to its intended target, causing a degenerative disease). Conversely, disturbing a malfunctioning interaction may reverse a disease in progress.
Great medical potential, yet great biochemical challenges.
Manipulating protein-protein interactions thus has great potential for drug discovery. The equally great challenge is that two typical proteins can interact over an area of roughly 15 to 20 square nanometers, involving on average 10 to 30 amino acids (a protein subunit).
Believe it or not, this is a huge surface area, with a wide range of chemistry which is challenging to unify in a manner useful for drug development. Although (usually) not all of the surface area is critical for two proteins to interact normally, there's no easy way to tell what sections are critical.
Imagine trying to score a free throw in a basketball court, blindfolded, without knowing the location of the basket, which may or may not be too small for your basketball. This is analogous to the situation scientists commonly face when designing a drug to selectively target a protein-protein interaction (no disrespect to the science or scientists is intended; this is meant solely to illustrate the difficulty of their task).
As you can imagine, experimental and theoretical investigations aimed at manipulating protein-protein interactions are costly, challenging, time-consuming, and labor-intensive. At times it's surely far more frustrating than spending a few hours obtaining a perfect image via atomic force microscopy, only to have a coworker unintentionally muck it up in the last few seconds (that was frequently my fate when I was a postdoc).
Consequently, the recent development reported by Narcis Fernandez-Fuentes (University of Leeds, United Kingdom) and coworkers will be highly appreciated by scientists working at manipulating protein-protein interactions. They have developed a user-friendly web server for identifying protein-protein "hot spots," i.e. the regions most critical to the interaction, which for many proteins requires only a few minutes of computation.
Operation of the web server.
The scientists' web server is called Presaging Critical Residues in Protein Interfaces Web Server (PCRPi-W). The first thing you need is the structure of the interacting proteins, which must be present in the Protein Data Bank, a vast public repository.
The web application then performs a number of quality checks, e.g. missing atoms are reconstructed. Any resulting computational changes to the data are recorded.
Other computational approaches commonly predict the contribution of specific amino acids to protein hot spots by some combination of amino acid sequence similarity, energy, and structural analysis. PCRPi-W utilizes seven different measures to account for this information, based on mathematical probabilities.
The web servers' performance has already been successfully tested against two independent data sets. One data set was comprised of 25 proteins which collectively interface at 636 amino acid units (300 of these amino acids are known to be critical or not based on actual experiments), and the other is a protein-protein interaction important in regulating cell division (disruption of which causes Costello syndrome and certain cancers).
Features of the web server.
The scientists' web server is extremely handy and user-friendly. The files provided to the user include:
An internal application enables you to highlight specific amino acid units and visualize their atomic-level interactions with neighboring amino acids. This is very useful for interpreting the results, and predicting future research directions for drug development.
The web application even emails you when the computations are complete, and assigns you an access code for easy data retrieval. A typical protein-protein interaction can be analyzed within a few minutes, but large protein complexes may require an hour or more.
The developers of this web server even offer their technical support if needed. They're clearly going out of their way to be useful to the scientific community.
Final comments.
If you don't have accurate, well-defined protein structures, the PCRPi-W web server unsurprisingly won't help you. Otherwise, I expect that it will be extremely useful for greatly cutting down on the cost and time inherent in probing protein-protein interactions, and therefore greatly accelerate drug discovery.
NOTE: The scientists' research was funded by the Research Councils United Kingdom, the Wellcome Trust, and the Leeds Institute of Molecular Medicine. Access to their web server is free at http://www.bioinsilico.org/PCRPi.
for more information:
Segura Mora, J. S., Assi, S. A., & Fernandez-Fuentes, N. (2010). Presaging Critical Residues in Protein interfaces-Web Server (PCRPi-W): A Web Server to Chart Hot Spots in Protein Interfaces PLoS ONE, 5 (8) DOI: 10.1371/journal.pone.0012352