about web server help people prc

Server for Determination of Potential DNA Binding Sites of Proteins
The website is free and open to all users and there is no login requirement.

Prediction of the DNA binding sites in proteins is of interest for understanding the protein function and the molecular mechanism of the protein-DNA recognition. Several approaches have been developed to identify DNA binding sites in proteins using sequence and structural properties of proteins.
Here, a new approach is introduced for the prediction of DNA binding residues and the residues that the DNA binding residues possibly interact using the fluctuations of residues in high frequency modes by the Gaussian Network Model (GNM) (1, 2). High frequency fluctuations involve tightly packed and severely constrained residues; i.e., the centers of localization of the energy in the structure (3). The high frequency fluctuating residues appears to have a role in maintaining the structure or its stability. These residues are implicated to be associated with binding as well as folding core residues (4-6). Recently, high frequency modes reflecting the local events at residue level have been shown to describe a network of residues whose fluctuations are strongly coupled and responding strongly to energy fluctuations (7). With the present approach, a network of residues is identified by the high frequency modes and these residues are analyzed with respect to the DNA binding. These residues are then filtered with their evolutionary conservation profiles from ConSurf (8) and ranked according to their propensities for the DNA binding. The high frequency fluctuating residues that are not tagged as DNA binding residues in the network of residues suggested are those that display correlated fluctuations with the identified DNA binding sites and that have possibility to overlap other functional residues, such as small ligand binding, protein-protein interaction sites etc..

The algorithm implemented as a web server, DNABINDPROT, is tested on a data set of 54 unbound proteins, verified with the bound structures. If the analyses are done based on the exact outcome of the highest fluctuation mode, using a conservation threshold of 5, the results have a sensitivity, specifity, precision and accuracy of 9.3%, 90.5%, 18.1% and 78.6% respectively. These values increase up to 24.3%, 93.4%, 45.3% and 83.3% for the respective cases, when the neighboring two residues are considered. Taking the high frequency modes into account enhances the specificity and the accuracy of the results when compared with the conservation analysis. The low sensitivity of the results can be explained by the susceptible association of the high frequency modes with the binding hotspots rather than the all possible binding residues. The identified residues are thus more selective, suggesting an average of only 4.8% of the structure as a potential DNA binding site. For half of the structures, at least one of the first three suggested sites is a DNA binding site. The remaining ones, on the other hand, are possibly part of the identified network of residues that interact with the DNA.

DNABINDPROT is available at http://www.prc.boun.edu.tr/appserv/prc/dnabindprot/ and has been up and running for the past 3 months on a local server, which is now accessible to all users. Entering the PDB ID or an uploaded structure with the chain ID as an input, the user has the option to choose conservation threshold, fast threshold percentage, the number of sequence neighbors, and the number of fast modes. With an optional email address provided, the results can be sent to the user as well as being displayed on the web page. The output consists of the suggested DNA binding sites obtained using the high frequency modes only and the results obtained from the combination of the high frequency modes and conservation data, wherever applicable. The given results are ranked based on the DNA binding propensities of the residues, ascribing higher weights for the DNA binding residues among the predicted residues that are functionally predisposed and are plausible for the their interaction with the DNA binding sites.
Keywords: protein-DNA binding, GNM (Gaussian Network model), conservation, fast modes

Keywords: protein-DNA binding, GNM (Gaussian Network model), conservation, fast modes


1. Haliloglu, T. Bahar, I., and Erman, B. (1997) Gaussian dynamics of folded proteins, (1997) Phys. Rev. Lett., 79, 3090
2. Bahar, I., Atilgan, A.R., and Erman, B. (1997) Direct evaluation of thermal fluctuations in proteins using a single-parameter harmonic potential, Fold. Des., 2, 173
3. Bahar, I., Atilgan, A. R., Demirel, M.C. and Erman, B. (1998) Vibrational dynamics of folded proteins: Significance of slow and fast modes in relation to function and stability, Phys. Rev. Lett. 80, 2733-2736.
4. Haliloglu, T., Seyrek, E., and Erman, B. (2008) Prediction of binding sites in receptor-ligand complexes with the Gaussian network model, Phys. Rev. Lett, 100, 228102
5. Haliloglu, T., Keskin, O., Ma, B., Nussinov, R. (2005) How similar are protein folding and protein binding nuclei? Examination of fluctuations of energy hot spots and conserved residues, Biophysical Journal, 88, 3, 1552-1559
6. Ertekin, A., Nussinov, R., Haliloglu, T. (2006) Association of Putative Protein Binding Sites with the Fluctuation Behavior of Residues, Protein Science, 15, 2265-2277
7. Haliloglu, T, Erman, B, (2009) Analysis of Correlations between Energy and Residue Fluctuations in Native Proteins and Determination of Specific Sites for Binding, Phys. Rev. Lett, 27;102(8):088103
8. Landau M., Mayrose I., Rosenberg Y., Glaser F., Martz E., Pupko T., and Ben-Tal N. (2005) ConSurf: the projection of evolutionary conservation scores of residues on protein structures, Nucl. Acids Res. 33:W299-W302.


DNABINDPROT: fluctuation-based predictor of DNA-binding residues within a network of interacting residues
Pemra Ozbek; Seren Soner; Burak Erman; Turkan Haliloglu
Nucleic Acids Research 2010; Web Server issue, 38:W417-423, 2010

For questions and comments: "appserv AT prc.boun.edu.tr"