The 5′ and 3′ untranslated regions of eukaryotic mRNAs play crucial roles in the post-transcriptional regulation of gene expression through the modulation of nucleo-cytoplasmic mRNA transport, translation efficiency, subcellular localization and message stability. UTRdb is a curated database of 5′ and 3′ untranslated sequences of eukaryotic mRNAs, derived from several sources of primary data. Experimentally validated functional motifs are annotated (and also collated as the UTRsite database) and cross-links to genomic and protein data are provided. The integration of UTRdb with genomic and protein data has allowed the implementation of a powerful retrieval resource for the selection and extraction of UTR subsets based on their genomic coordinates and/or features of the protein encoded by the relevant mRNA (e.g. GO term, PFAM domain, etc.). All internet resources implemented for retrieval and functional analysis of 5′ and 3′ untranslated regions of eukaryotic mRNAs are accessible at http://www.ba.itb.cnr.it/UTR/ .
Received September 15, 2004; Accepted September 18, 2004
One of the main challenges of the post-genomic era is the understanding of the mechanisms that control the spatio-temporal regulation of gene expression. The fate of newly synthesized mRNA with respect to its nucleo-cytoplasmic transport, stability, translation efficiency and subcellular localization is determined at the post-transcriptional level. Such regulation is mostly mediated by cis-acting elements located in the 5′ and 3′ untranslated regions of mRNAs (5′ UTR and 3′ UTR) ( 1 ).
In several cases, specific functional sequence elements have been identified and characterized. These usually correspond to short oligonucleotide tracts whose biological activity relies on a combination of their primary sequence and specific secondary structure. These motifs act either as target sites for RNA-binding factors or interact directly with the translation machinery.
The availability of a large collection of functionally related sequences—such as UTRs—is invaluable for the inference of structural and compositional features and for the identification of conserved candidate regulatory motifs. For this reason, we have developed UTRdb, a collection of 5′ and 3′ UTR sequences derived from eukaryotic mRNAs. Sequences collated in UTRdb were generated by custom software. UTRdb is a non-redundant database and annotation includes information not available in the primary databases such as genome localization and structure and presence of known regulatory elements.
We have also created UTRsite, a collection of regulatory elements located in 5′ and 3′ UTRs whose function and structure have been experimentally determined and published. The UTRsite collection may prove useful in automatic annotation projects of unknown sequences as well as for finding previously undetected signals in known sequences.
For the most recent release of the UTRdb and UTRsite databases, we have focused on the improvement of data quality, increasing the degree of integration with other resources and the incorporation of genome-related facilities. Besides a new graphical interface, we have introduced new specific UTR collections: (i) UTRef from RefSeq database ( 2 ); (ii) UTRait from TRAIT database of muscle-specific transcripts ( 3 ); and (iii) UTRexp, a collection of UTR sequences whose functional activity has been experimentally investigated (see below for details). The UTRsite collection of functional motifs has also been significantly expanded. Moreover, we have mapped human UTRs on genome assemblies, facilitating the direct comparison and integration of several annotated genomic features available through batch queries of Ensembl databases.
The integration of UTRs and protein/genomic resources is potent in that it allows the retrieval of specific UTR subsets based on their genomic coordinates and/or features associated with the encoded proteins (e.g. GO terms, PFAM domains, etc.).
GENERATION OF UTRdb AND ITS INTEGRATION WITH OTHER DATABASES
UTRdb entries are automatically generated through the accurate parsing of the Feature Table of entries in primary databases (e.g. EMBL). Entry curation includes the detection of contaminating vector sequences, the removal of sequence redundancy and the annotation of repetitive elements and known regulatory motifs collected in the UTRsite database. Details of this process can be found in ( 4 ).
The current release of UTRdb contains three further specialized divisions: UTRef, UTRait and UTRexp. Sequences collected in UTRef and UTRait have been generated from the RefSeq ( 2 ) and TRAIT ( 3 ) databases, respectively. UTRexp contains UTRs that have been investigated experimentally and shown to contain functional motifs. Some of these sequences are not present in primary sequence databases and have been manually extracted from literature resources.
In the current release, we have also determined the genomic coordinates of human UTR sequences using the program BLAT ( 5 ) with the human genome assembly (Release NCBI 34). Only those UTRs that unambiguously mapped to a single genomic location were considered. Exonic structure of mapped UTRs was then refined by applying the program Spidey ( 6 ) to compare the UTR and its corresponding genomic location.
We have tried to associate each mapped UTR to the specific protein encoded by the corresponding mRNA using the relevant Ensembl coordinates. A protein was defined a ‘neighbor’ of a 5′ UTR if its start site corresponds to the end of the 5′ UTR sequence (and the converse for 3′ UTRs) Once the neighbor protein of a given UTR entry had been defined, we were also able to identify the Ensembl transcripts cross-referenced to the neighbor protein. If, for a given UTR entry, no annotated protein matched our criteria, we associated any Ensembl gene overlapping the same genomic region with the UTR.
The cross-referencing of UTRs and Ensembl features (protein, transcript, gene) provides a valuable resource as UTRs automatically inherit the large body of functional features annotated with the Ensembl project ( 7 ).
We have also endeavored to cross-link the UTRdb human division with IPI (International Protein Index) ( 8 ), which contains a complete non-redundant data set representing the human proteome, derived from different curated protein databases.
In future releases of UTRdb, we plan to extend the cross-referencing between UTRs and protein/genomic resources to all other organisms included in Ensembl.
UTRdb entries (see Figure 1 for an example) are annotated for the occurrence of regulatory motifs whose activity has been assessed by experimental investigation, located in the 5′ or 3′ UTR of eukaryotic mRNAs. All these motifs are collected in the UTRsite database. Each UTRsite entry ( Figure 2 ) is prepared/reviewed/updated by expert scientists (in many cases, those who performed the experimental analysis). We have now developed a Submission Tool for the generation/management/update of UTRsite entries ( Figure 3 ). This tool allows selected annotators, to annotate/update all the information in the entry in a user-friendly manner via a personal login.
The databases UTRdb, UTRsite and the new specific UTR collections (UTRef, UTRait, UTRexp) have been organized into MySQL relational database management system.
The main section of UTRdb (Release 19) contains nine sequence collections, one for each of the eukaryotic divisions of the EMBL nucleotide database (Release 78), namely (i) human; (ii) mouse; (ii) rodent; (iv) other mammal; (v) other vertebrate; (vi) invertebrate; (vii) plant; (viii) fungi; and (ix) virus.
UTRef was generated from Reference Sequence collections (RefSeq Rel. 3). Table 1 reports a summary description of UTRdb which contains 298 036 entries and 128 286 081 nucleotides. UTRsite collects a total of 52 regulatory motifs, including upstream Open Reading Frames (uORFs) with known regulatory activity, whose occurrences have been annotated in 30 370 entries of UTRef collection.
| 5′UTR || 3′UTR || Total |
| 5′UTR || 3′UTR || Total |
AVAILABILITY OF UTRdb
UTRdb and UTRsite are accessible through an SRS retrieval system, which has been updated to include the new fields added. In particular, all the information derived from genome mapping of UTRs, the IPI cross-link, and cross-referencing to ‘neighbor proteins’ and Ensembl genes/transcripts is available through a new field named ‘Genomic Features’ as reported in Figure 1 . It is also possible to browse these fields by querying the SRS ‘Extended query form’ where relevant query fields have been added.
In addition, to access all of the information indirectly linked to UTRs, we have developed a custom browsing system—the ‘UTR genome browser’ ( Figure 4 ). Through this retrieval system, it is possible to select and extract specific UTR subsets defined by accession numbers derived from a variety of databases including IPI ( 8 ), Interpro ( 11 ), GO ( 12 ), GENEW ( 13 ), MIM ( 14 ), etc. as well as by genomic coordinates.
The user can choose to download selected entries in Fasta format or to display a summary of their relevant genomic features, including genomic coordinates and cross-references to a variety of genomic resources. A graphical representation displays selected UTRs in terms of genomic coordinates and shows other human UTRs, cDNAs, proteins and ESTs in the same genome location.
With this new tool, it is now possible to obtain specific UTR subsets from mRNAs coding for proteins of a selected protein family, containing a specific domain or belonging to a specific GO class. Further investigations of such homogeneous sets of UTRs may allow the identification of common features or conserved regions whose potential functional activity may then be experimentally characterized.
Further on-line utilities are UTRscan and UTRblast. The UTRscan feature allows the enquirer to search user submitted sequences for any of the motifs collected in UTRsite. The UTRblast utility allows database searches against any of the UTRdb sections.
UTRdb, UTRsite and other related resources are publicly available at http://www.ba.itb.cnr.it/UTR/ .
We thank David Horner for helpful comments on the manuscript. This work was supported by Ministero dell'Istruzione e Ricerca, Italy (projects: MIUR Cluster C03/2000-CEGB, PON 2000-2006 Progetto BIG, FIRB project ‘Bioinformatica per la Genomica e la Proteomica’) and Telethon. F.M. was the recipient of a EU Marie Curie fellowship at the European Bioinformatic Institute, Hinxton, UK.
1Dipartimento di Scienze Biomolecolari e Biotecnologie, Università di Milano, via Celoria 26, 20133 Milano, Italy, 2Sezione di Bioinformatica e Genomica, Istituto Tecnologie Biomediche del Consiglio Nazionale delle Ricerche (CNR), via Amendola 165/A, 70126 Bari, Italy, 3Dipartimento di Biochimica e Biologia Molecolare, Università di Bari, via Orabona 4, 70126 Bari, Italy and 4EMBL Outstation, The European Bioinformatics Institute, Wellcome Trust Genome Campus, Hinxton, Cambridgeshire CB10 1SD, UK