Prediction of catalytic residues from X-ray protein structure refinement parameters
Abstract
Catalytic residue investigation is important for biologists to study protein functions. In previous studies, many researchers have successfully applied features, which were from sequence- and structure-based, to predict the position of catalytic residues in proteins. A highly correlation between atomic fluctuations and the catalytic positions have ever been observed. In this study, we were trying to investigate if this information was hidden in the X-ray diffraction data. The results from our test indicated the possibility to catch-up the catalytic residues in the protein structure refinement process.
Keywords
Download Options
Introduction
To realize the function of a protein is the major goal for structural biologists to solve the structure. However, many protein structures from structural genomic project [1] are usually functional unknown. In order to decrease spent time and money, researchers usually need to pursue theoretical methods for identifying the potential functional sites. For this demand, many in silicon methods have been designed to achieve this requirement. These methods including sequence based such as, L1pred [2] and PINGU [3], and structural based such as, Jorge Fajardo approach [4] and WCN approach [5].
Recently, Huang et al., 2011 [5] provided a prediction method based on the dynamics nature of catalytic residues.They filtered out the candidate residues by ranking the crowdedness value, which they called weighted contact number (WCN), for each residue in a given protein. The occurrences for each amino acid type of catalytic residues have also been investigated. Through their approach, potential catalytic residues can be thus simply predicted from a single protein structure. The concept of Huang’s method is from Lin et al. 2008 [6], in which article they have mentioned the WCN model can be simplified to a centroid model (CM) [7]. The CM has more than once been used as the alternative for the translation/libration/screw (TLS) model [8-13], which is frequently used in X-ray structural refinement process. Therefore, the application of the TLS model could be reasonably considered as a way to predict catalytic residues from a single structure.
The major advantage by using TLS model is its university in structural biology community. If the prediction can be finished coupled with protein structural refinement, it would be very useful for structural biologists. Here, we showed that the prediction of catalytic residues in enzymes could be achieved directly from X-ray structural refinement process by using TLS-derived B-factors.
We test our approach for a previous reported enzyme dataset [5], the results showed that the viability of our approach to locate the residues in proteins
Conclusion
In this study, we used TLS model to simulate the atomic fluctuations for investigating if the signature of catalytic residues is hidden in the X-ray diffraction data. Through several designed experiments, we found TLS-derived B-factors had ability to predict the positions of catalytic residues in proteins under a cut-off value -0.6. It’s not only reveal the viability for using TLS model for catalytic residue identification, but also a hint to show that the signatures of the catalytic residues might be hidden in X-ray diffraction data. Therefore, further studies are needed to clarify this point.