PARS: a web server for the prediction of Protein Allosteric and Regulatory Sites

BIOINFORMATICS APPLICATIONS NOTE Structural bioinformatics Vol. 30 no. 9 2014, pages 1314–1315 doi:10.1093/bioinformatics/btu002 Advance Access publication January 9, 2014 PARS: a web server for the prediction of Protein Allosteric and Regulatory Sites Alejandro Panjkovich1 and Xavier Daura1,2,* 1 Institute of Biotechnology and Biomedicine (IBB), Universitat Autònoma de Barcelona (UAB), Bellaterra 08193, Spain and 2Catalan Institution for Research and Advanced Studies (ICREA), Barcelona 08010, Spain Associate Editor: Alfonso Valencia ABSTRACT Received on June 5, 2013; revised on November 30, 2013; accepted on December 31, 2013 1 INTRODUCTION Tight regulation of protein function is fundamental to life. Proteins involved in metabolic pathways, signalling cascades and genomic transcription, among other processes within the living cell, are commonly under allosteric regulation, i.e. their activity is modified through the binding of a ligand molecule to the protein in a site different from the active site. In fact, allostery is one of the most powerful protein-function regulation mechanisms, as it allows proteins to sense and immediately respond to changes in their environment (Fenton, 2008). Traditional drug-design approaches focusing on active or primary binding sites can be, therefore, extended by exploiting allosteric sites, as shown by current efforts on GPCRs (Melancon et al., 2012) or farnesyl pyrophosphate synthase (Jahnke et al., 2010). An advantage of targeting allosteric sites therapeutically is a reduced risk of secondary adverse effects. This is because allosteric sites appear to be significantly less conserved than active sites across homolog proteins (Waelbroeck, 2003), enabling the design of allosteric drugs with high specificity for a single protein within a family. This observation has motivated the development of allosteric drugs for the regulation of phosphodiesterase 4D, for which active site inhibitors cause emesis, a dose-limiting side effect (Burgin et al., 2010). Moreover, a drug-discovery approach *To whom correspondence should be addressed. 1314 based on allosteric sites may result in the development of not only novel drug-like inhibitors but activators as well (Peracchi and Mozzarelli, 2011). Other concepts such as the notion of serendipitous allosteric sites, which have no known ligand in nature but can become functional in the presence of an ‘opportunistic’ ligand (Hardy and Wells, 2004) or the idea that almost any protein may be regulated allosterically (Gunasekaran et al., 2004), contribute to the current interest in the pharmacological exploitation of allosteric sites. In this context, it is expected that a deeper understanding of the properties of allosteric sites and their identification would help streamlining the design and discovery of novel therapeutic drugs (Nussinov and Tsai, 2013). Despite growing efforts, the atomic-level details that explain the functional relationship between distant sites in the same protein molecule have not been elucidated for most of the known cases of allostery (Peracchi and Mozzarelli, 2011). This has motivated a growth in the number of computational studies on protein allosteric sites, including the recent publication of three web servers: SPACER (Goncearenco et al., 2013) and MCPath (Kaya et al., 2013), which study the allosteric communication across residues in a protein structure, and Allosite (Huang et al., 2013), which predicts allosteric sites on protein structures using a machine-learning approach. Here, we present PARS, which queries protein flexibility and structural conservation to predict the location of allosteric sites. The implementation emphasizes ease-of-use and speed, whereas maintaining high accuracy as benchmarked against a large set of known allosteric proteins. Even though our method (as the servers mentioned earlier in the text) requires a protein structure to work, we have added an initial homology modelling step that allows the user to run predictions starting from protein sequence in cases where a known structure is not available. 2 DESCRIPTION The current perspective on allosteric transitions and associated regulatory events relies on the ‘population shift’ concept (Cui and Karplus, 2008). Briefly, the protein or protein complex explores different conformations (both active and inactive) in solution, and the allosteric ligand ‘shifts’ the population or ensemble of conformations on binding, effectively modulating the protein’s activity rate (Kar et al., 2010). The conformational space explored by the protein can be sampled using computational methods such as molecular dynamics (Chiappori et al., 2012) or normal mode analysis (NMA) (Dykeman and Twarock, 2010). In our approach, protein dynamics are queried ß The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: Summary: The regulation of protein activity is a key aspect of life at the molecular level. Unveiling its details is thus crucial to understanding signalling and metabolic pathways. The most common and powerful mechanism of protein-function regulation is allostery, which has been increasingly calling the attention of medicinal chemists due to its potential for the discovery of novel therapeutics. In this context, PARS is a simple and fast method that queries protein dynamics and structural conservation to identify pockets on a protein structure that may exert a regulatory effect on the binding of a small-molecule ligand. Availability: PARS is freely available as a web server at http://bioinf. uab.cat/pars. Contact: Supplementary information: Supplementary data are available at Bioinformatics online. PARS and 18% on the second position of the PARS ranking, whereas a total of 73% are found in the top three positions. Full benchmark details and performance statistics are available as Supplementary Material. We expect that this easy-to-use and relatively fast web server will prove useful for medicinal chemists and other researchers studying the regulation of protein function for biochemical characterization and other applied tasks such as binding-site prioritization for virtual drug screening. Funding: Seventh Research Framework Programme of the European Union (ref. HEALTH-F3-2009-223101). FPU Scholarship from MICINN, Spanish Government (to A.P.). Conflict of Interest: none declared. REFERENCES Burgin,A.B. et al. (2010) Design of phosphodiesterase 4d (PDE4D) allosteric modulators for enhancing cognition with improved safety. Nat. Biotechnol., 28, 63–70. Chiappori,F. et al. (2012) Molecular mechanism of allosteric communication in hsp70 revealed by molecular dynamics simulations. PLoS Comput. Biol., 8, e1002844. Cui,Q. and Karplus,M. (2008) Allostery and cooperativity revisited. Protein Sci., 17, 1295–1307. Dykeman,E.C. and Twarock,R. (2010) All-atom normal-mode analysis reveals an RNA-induced allostery in a bacteriophage coat protein. Phys. Rev. E Sta (...truncated)