PrionScan: an online database of predicted prion domains in complete proteomes

Abstract

Background: Prions are a particular type of amyloids related to a large variety of important processes in cells, but also responsible for serious diseases in mammals and humans. The number of experimentally characterized prions is still low and corresponds to a handful of examples in microorganisms and mammals. Prion aggregation is mediated by specific protein domains with a remarkable compositional bias towards glutamine/asparagine and against charged residues and prolines. These compositional features have been used to predict new prion proteins in the genomes of different organisms. Despite these efforts, there are only a few available data sources containing prion predictions at a genomic scale.

Description: Here we present PrionScan, a new database of predicted prion-like domains in complete proteomes. We have previously developed a predictive methodology to identify and score prionogenic stretches in protein sequences. In the present work, we exploit this approach to scan all the protein sequences in public databases and compile a repository containing relevant information of proteins bearing prion-like domains. The database is updated regularly alongside UniprotKB and in its present version contains approximately 28000 predictions in proteins from different functional categories in more than 3200 organisms from all the taxonomic subdivisions. PrionScan can be used in two different ways: database query and analysis of protein sequences submitted by the users. In the first mode, simple queries allow to retrieve a detailed description of the properties of a defined protein. Queries can also be combined to generate more complex and specific searching patterns. In the second mode, users can submit and analyze their own sequences.

Conclusions: It is expected that this database would provide relevant insights on prion functions and regulation from a genome-wide perspective, allowing researches performing cross-species prion biology studies. Our database might also be useful for guiding experimentalists in the identification of new candidates for further experimental characterization.

Figure 1

Distribution in all taxa of the predictions compiled in PrionScan. The pie chart depicts the distribution of prion predictions in all different taxa, from archaea to humans. The number of predictions in each taxon is shown in parenthesis. The predictions in each taxon are organized following the structure of UniprotKB [28] taxonomic subdivisions, in which the proteins in the taxa rodents, mammals and human are distributed separately.

Figure 2

The Simple Search option for direct access to protein prion prediction information. A) The Simple Search option is used for querying the database using the UniprotKB identifier of a given protein. B) The Detailed Output Page retrieved by the query for a protein with putative prion domains. At the top there is a button for complete download of the results.

Figure 3

The Simple Search option for searching with text keywords. A) The Simple Search option is used for querying the database with a text keyword –i.e. the complete name of an organism in this case. B) The General Output Page retrieved by the query with rows corresponding to multiple entries in the database –i.e. putative prion proteins in the genome of this organism– each one redirecting to a Detailed Output Page. At the top there is a button for complete download of the results and at the bottom there is a summary of the number of predictions and a functionality for browsing throughout multiple pages containing all the results returned.

Figure 4

Complex Search option for searching with multiple text keywords. A) The Complex Search option is used for querying the database combining the information of two columns of the database –e.g. in this case we search the putative prions in the genome of an organism related to a specific molecular function as described in Gene Ontology. B) The General Output Page retrieved by the query with rows corresponding to multiple entries in the database –i.e. putative prion proteins in the genome of this organism– each one redirecting to a Detailed Output Page. At the top there is a button for complete download of the results and at the bottom there is a summary of the number of predictions and a functionality for browsing throughout multiple pages containing all the results returned.

Figure 5

Sequence Analysis from file and text for processing user’s own sequences. A) The Sequence Analysis from text option in which the user can modify the cutoff used in our methodology to scan the sequences that can be pasted in the text box below. B) The Sequence Analysis from file option useful for processing a high number of sequences by submitting the file to be processed in our server. In this case there is an obligatory text box for providing the e-mail address for sending the results upon completion and the cutoff could also be adjusted at user’s discretion.