piRBase: a web resource assisting piRNA functional study

Introduction

piRNAs are a recently discovered class of small RNAs that bind to PIWI proteins. piRNAs are mainly expressed in the germline, although expression is also observed in somatic cells. In most species, the piRNAs range in size between 24 and 33 nt, whereas in Caenorhabditis elegans, the small RNAs corresponding to piRNAs are 21 nt in length and are commonly called 21U RNAs. piRNAs share a strong preference for a 5′-uridine residue. Genomic mapping have shown that piRNAs mostly originate from a limited number of clustered loci, each cluster being several kilobases in extension, and in which piRNAs may be encoded by one or both strands (1, 2). The amount of publicly available piRNA data is presently increasing rapidly.

piRNAs were first shown to function in post-transcriptional regulation of transposons. Reuter et al. (3) discovered that extensive complementarity between piRNAs and targeted transposon transcript was required for cleaving of targets in male germ cells by the protein MIWI, the mouse homologue of PIWI, and that the cleavage position was located 10 nt downstream of the 5′-end of the guide piRNA. Enrichment for L1- and IAP-derived piRNAs in mouse testes similarly showed a 10-nt distance between the 5′-ends of sense and antisense partners (4). Kiuchi et al. (5) found a 10-nt overlap between piRNAs derived from the Fem and Masc mRNAs in silkworm embryos, suggesting that piRNAs might participate in post-transcriptional silencing of coding genes by cleaving the corresponding mRNAs. In addition, piRNAs appear to induce mRNA deadenylation and decay in mouse elongating spermatids (6) and in the Drosophila embryo (7).

Epigenetic roles for piRNAs have also been discovered. In the fruit fly, PIWI binds to heterochromatin protein 1a (HP1a), which, upon methylation of histone H3K9, maintains the heterochromatin state of specific chromosomal regions (8, 9). It has also been reported that the PIWI protein can reactivate the euchromatin state of some chromosomal regions (10). Upon mutation of the PIWI proteins in mouse testes, the DNA methylation of retrotransposon genes is lost and the elements show increased expression (4, 11–13). Besides this, the levels of histone modification H3K9me3 on sequences flanking full-length L1-A copies were reduced in Miwi2 knockout spermatogonia (14). These results indicate that the piRNAs function in the establishment of DNA methylation and H3K9me3 marks on retrotransposons. Another report indicated that the Piwi/piRNA complex from the Aplysia central nervous system facilitates methylation of a conserved CpG island in the promoter of the breast cancer-related CERB2 gene (15).

The varied roles and rapidly increasing numbers of piRNAs underscore the need for a web analysis platform for piRNAs. In RNAcentral (16), the main database for RNA sequences, piRNABank (17) is the only piRNA database. Outside the RNAcentral, the piRNAQuest database (18) also focuses on piRNAs. As both piRNABank and piRNAQuest only contain limited amounts of piRNA data (Table 1) and annotations, and barely touches on the functions of the piRNAs, we have developed a new database named piRBase. piRBase has assembled a larger amount of piRNA data than the presently existing databases, and is the only database that includes epigenetic data and experimentally or computationally generated piRNA target data.

Currently, piRBase contains 77 million piRNA sequences from nine organisms (Figure 1A), including data from worm (C. elegans), chicken, frog (Xenopus tropicalis) and silkworm (Bombyx mori) piRNAs which had previously not been collected by other piRNA databases. The amount of piRNA sequences derived from mouse, fruit fly and zebrafish is also much larger than in the other two databases (Table 1). More details on distinct piRNAs are provided, such as experimental method by which the piRNA was obtained, the tissues expressing the piRNAs, and annotations of the piRNA loci.

Construction and content of the piRBase database

More than 77 million piRNA sequences and their corresponding annotations have been collected by piRBase. The data were collected from the literature and external databases. Processed piRNA sequences (txt or fasta files) have been preferred to raw sequencing data (sra or fastq files). We have put much effort into harvesting piRNA datasets from the literature and in verifying that these sequences were regarded as piRNAs by the authors of the respective papers. The piRNAs presently assembled in piRBase were mainly obtained by four experimental methods: (i) small RNA sequencing, (ii) immunoprecipitation of Piwi or Piwi-associated proteins, (iii) Piwi protein crosslinking-immunoprecipitation and (iv) chromatography. The amounts of piRNA sequences obtained by each method are displayed in Figure 1B. Figure 1C shows the length distribution of unique piRNA sequences in piRBase.

After mapping the piRNAs to the genome, we took particular care to identify piRNAs that are derived from repeat elements and from coding genes, as these piRNAs might participate in the post-transcriptional regulation of the corresponding elements. In addition, piRBase also collected information on predicted and experimentally verified piRNA targets, DNA methylation data of tissues expressing piRNAs, and H3K9me3 data that may be related to piRNA function. The data collection and processing steps are illustrated in Figure 2 and in the Supplementary computational procedures.

Web interface

Future directions

The number of piRNAs that are being reported is increasing rapidly. We will therefore update piRBase and integrate more information supporting piRNA functional analysis at intervals depending on the rate with which new data appear, expecting to issue new versions of the database about once every half year. In the future, we will also integrate piRNA datasets that provide only raw sequencing data. We will continue to develop the piRNA target prediction software, and special attention will be paid to the possibility of constructing piRNA-gene regulatory networks and elucidate piRNA action in distinct environments.

Supplementary data

Supplementary data are available at Database Online.

Funding

The Ministry of Science and Technology of China (No. 2014AA021103, 2011CB504605); the National Natural Science Foundation of China (No. 31090253, 31210103912); grant (No. O529YX5105) from the Key Laboratory of the Zoological Systematics and Evolution of the Chinese Academy of Sciences and National Science Fund for Fostering Talents in Basic Research (Special subjects in animal taxonomy, NSFC-J1210002). Funding for open access charge: 2014AA021103.

Conflict of interest. None declared.