PDZscape: a comprehensive PDZ-protein database

The concept of domains has been one of the core themes in protein studies; one such being PDZ (Post synaptic density protein (PSD95), Drosophila disc large tumor suppressor (Dlg1), and Zonula occludens-1 protein (zo-1)) - a ubiquitous 70–90 residue protein–protein interaction module. PDZ helps localization and clustering of membrane receptors by binding primarily to their C-terminus in different pathways that drive life [1–4]. Structurally, PDZ domains have a conserved fold, though the secondary structure varies. Over 400 solved structures of PDZ domains suggest that the domain consists of 6 β-strands and two α-helices (one long and one short) (Fig. 1) [3, 5]. The specificity of PDZ-domain-based interactions is determined primarily by the sequence of the C-terminus of the proteins they bind, although few exceptions do exist [6]. Many PDZ containing proteins have more than one PDZ domains, which facilitate multi-protein interactions and thus have been referred to as a ‘Molecular Switch’ [3]. The only existing database ‘PDZBase’ on this domain comprises only about 339 proteins that was last updated way back in 2004 [6, 7] thus creating a need for a database on this domain family that would include the important developments of the last decade. With cellular PDZ proteins as common targets of pathogenic viruses emerging as an important new theme in virology, high throughput and wide spectrum studies on this family of proteins is of significant value [8]. Information about known PDZ proteins, collected together, should be of great importance to research community working with PDZ-proteins and their binding partners. All this information together can help identify role of various PDZ-proteins in different cellular processes.

The database ‘PDZscape’ provides information on 58,648 PDZ-containing proteins as well as their known and putative interacting partners on one platform. The user-friendly interface can be easily queried with external identifiers for the protein. Here, unique integration of all prominent databases such as NCBI, UniProtKB, Swiss-Prot, PubMed, Protein Data Bank (PDB), Protein Mutant Database (PMDB), DisGENET, STRING, IntAct, Pfam and KEGG lead to detailed information about PDZ proteins and pathways they are involved in, along with the specific nature of the interaction and the methodology involved. BLAST has been customized for PDZ protein homology search with an additional option of batch downloading the PDZscape. The organization of the database along with its different features is compiled in a self-explanatory flowchart as shown in Fig. 2. It enlists all the well-established and putative interacting partners thus providing a comprehensive picture of this domain family. It also includes information on mutations in the PDZ containing proteins, obtained not only from the PMDB but also through manual curation from literature. Our manually curated data from PMDB and PubMed on the association of PDZ-containing proteins with different pathological conditions significantly enhances the value of the provided information. Additionally, it also provides some tools – PDZ-Blast and in-house developed PBPFinder (PDZ-Binding Protein finder). PDZ-Blast can be used to find similarity between protein of interest and PDZ-proteins from PDZscape. PBPFinder helps to find whether a protein of interest is a known PDZ-binding partner. It also provides an option to scan small peptides for presence of PDZ-binding motif. Lastly, with the amount of data generated in the last decade (Fig. 3), and its involvement in a myriad of cellular pathways, it is essential to have updated information exclusively on this very important class of protein-protein interaction domains. Most importantly, this data will help in delineating various cellular functions and pathways involving PDZ-proteins. PDZScape can also be used as a tool to understand pathophysiology of diseases associated with aberrations in PDZ-proteins.

PDZscape Database is available online through ACTREC Home Page (http://www.actrec.gov.in/) under ‘Database & Tools’ (http://www.actrec.gov.in:8080/pdzscape) and has an easy-to-browse web interface. End user can retrieve complete information of a particular protein entry using any of the following: Keywords, Uniprot IDs, String IDs, PDB IDs, organism names or protein sequence. Search tool provides easy access to the database using various criteria as shown in Fig. 4. Keywords can be used to search the database using generic terms such as protein names, gene names or other related information for more specific search. Keywords can also be used in combination with ‘AND/OR’ for filtered results. Other search methods provide access to database using one or more external identifiers like Uniprot ID, PDB ID etc. Multiple IDs can also be entered separated by commas. Quicker access to the data is also possible where user can reach to a particular protein entry in PDZscape using URL: http://www.actrec.gov.in:8080/pdzscape/protein.jsp?accID=UniProtID, where term UniProtID should be replaced by protein identifier of required PDZ-protein. Figure 4 shows sample query string and the output of the search in PDZScape. Our manually curated data from PMDB and PubMed on the association of these proteins with different pathological conditions significantly enhances the value of the provided information. PDZ-BLAST facility allows user to search PDZscape using sequence of the protein of interest, which is accessible to the end-user through the hyperlink provided in the main menu. The query sequence can thereby be entered either manually by pasting FASTA sequence in the respective search box or by uploading a FASTA sequence file. Results can be downloaded in the pdf format from the BLAST result page. ‘Download’ option can be used to retrieve the complete database in CSV format; customized subsets can also be downloaded from this link. The links provided in the database highly expand the annotations of the PDZscape, and will enable end-user to have access to every bit of required information on a single arena. PDZscape is designed to include periodic updates and inclusion of new features to make it a useful portal for biologists. PBPFinder currently works on regular expressions for motifs described in previous literature. Development of robust machine learning model for prediction of PDZ-binding proteins is in process and will be a valuable addition to this database. While PDZbase is a well-known portal for PDZ information, it contains a limited number of PDZ-proteins with focus on the details of protein-protein interactions that perhaps needs updating. On the other hand, PDZscape encompasses all PDZ-proteins with known and putative binding partners.

A large number of genes have been annotated over the last decade with the exponential increase in the number of depositions on PDZ proteins in NCBI and other sources. This provides an exciting as well as challenging prospect of studying these newly identified and annotated proteins under a single platform. The plethora of research endeavors that have undertaken in the last decade on this protein domain underscores its importance in critical cellular processes and association with several disease conditions. Categorizing proteins, though important is a relatively simple job compared to their structural and functional characterization. With large number of PDZ-proteins being identified and structurally characterized (Fig. 5), there was a dire need of a compiled resource on these proteins that include information on their involvement in vital cellular processes, and their disease association.