BAC screening and STS content mapping

Eight different primer pairs were designed in conserved ovine RPSA regions identified by aligning previously described mRNA and expressed sequence tag (EST) sequences, representing each exon at least once. Using these primers, 34 bacterial artificial chromosome (BAC) clones, containing members of the RPSA family, were isolated by PCR screening of the INRA sheep BAC library [19], with an annealing temperature (Ta) that was at least 8°C lower than the melting temperature (Tm) of the primers to allow primer mismatches (Additional file 1). By sequence tagged site (STS) content mapping, performed with 54 unique STS primer pairs that were designed from the 68 BAC end sequences (BES) [GenBank:GS375851-GS375918], 6 mini-contigs could be constructed and another 6 single BAC clones could be identified, each containing a different family member of the ovine RPSA family (Figure 1 and 2; Additional file 2).

Characterization of the 12 RPSA gene family members

Each member of the RPSA gene family was sequenced by direct sequencing on BAC DNA starting with the PCR primers as sequencing primers and finishing by primer walking. The sequences were assembled with the CAP3 program [20] and annotated with BLAST [21].

The full length functional gene, that was first described by Marcos-Carcavilla et al. [18], was present in one of the contigs composed of 6 BAC clones [Genbank:GQ202529]. We have sequenced for the first time, the complete intron 3, comprising 8846 bp, which like the other introns, has consensus acceptor and donor splice sites. The full length ovine RPSA gene consist thus of 13287 bp.

Besides the full length functional gene, 11 other RPSA gene family members were sequenced [GenBank:GQ202530-GQ202540]. A schematic representation of all the family members, based on sequence alignments with the full length functional gene (Additional file 3), is included in Figure 3. They all are considered as processed pseudogenes and in accordance with RPSA pseudogenes described in other species, they have been assigned the names RPSAP1-RPSAP11. Pseudogenes arise in 2 different manners: either by retrotransposition of the mRNA of the ancestral gene into the genome or by duplication of genomic DNA [22]. The first class is known as processed pseudogenes, the second one as non processed pseudogenes. The majority of the pseudogenes are processed and originate from housekeeping genes, with ribosomal protein genes as largest subgroup [22, 23]. As processed pseudogenes are inserted without internal promoter, they are released from selection pressure and accumulate mutations during evolution leading to frameshift mutations and/or premature stopcodons which prevents them of encoding a functional protein [24]. In some cases nevertheless, they have obtained a (regulatory) function [17].

Analysis of the primer binding sites in the pseudogenes showed that in our experimental design the screening primers could anneal to targets down to 83% sequence identity, even in the case of RPSAP2.

All BAC clones and thus all RPSA family members were isolated with at least 2 primer pairs and there was no concordance between the number of BACs in a mini-contig and the level of sequence identity with RPSA. We conclude that it is most likely that we have isolated all the ovine RPSA pseudogenes sharing a high level of sequence identity and that therefore can interfere with the functional RPSA gene in genetic studies.

To obtain a first indication of possible functionality, in silico ORF and promoter prediction analysis were carried out.

The pseudogene RPSAP7 is the only member sharing almost an identical ORF with the full length RPSA gene. The only one amino acid difference (amino acid 31: D → G) is located in the intracellular part of the receptor that does not belong to any binding site. All the other pseudogenes either lack the start codon or contain a premature stop codon due to nonsense or frameshift mutations. The size of the potential ORF of the other pseudogenes varies and the largest reaches 171 amino acids sharing 100% identity with RPSA (Table 2). Most ORFs lie in the intracellular region of RPSA (amino acid 1-101). In case of RPSAP6, RPSAP8, RPSAP10 and RPSAP11, the ORF contains a part of the binding sites of RPSA with PrP (direct binding aa 161-180; indirect binding aa 180-285 [25]), but most of them have a low level of amino acid identity.

In silico promoter analysis predicted a possible promoter for RPSAP1, RPSAP2, RPSAP4, RPSAP8, RPSAP9 and RPSAP10 (Additional file 4). A consensus polyadenylation signal is present in 7 of the 11 pseudogenes (including RPSAP7).

Repeated sequences were identified with Repeatmasker [26] and showed that 4 pseudogenes are disrupted by interspersed repeats belonging to the class/family SINE/RTE-BovB, SINE/BovA, tRNA and SINE/tRNA-Glu, and that 7 pseudogenes were flanked by repeats belonging to the SINE, LINE, tRNA, LTR and simple repeat classes. According to Zhang et al. [27], processed pseudogenes are mostly found in genomic regions with a relatively low GC content, as do LINE repeats. Thus, it is not surprising that such repeats are present in the regions flanking many of the RPSA pseudogenes.

A remarkable observation is that part of the RPSA intron 4, containing the small nucleolar RNA (snoRNA) SNORA62, is present in the RPSAP8 and RPSAP9 pseudogenes. Therefore, these pseudogenes can be considered as semi-processed pseudogenes, which are very rarely reported and defined by Zhang et al. as "pseudogenes that contain remnant introns, which suggests that they were derived from semi-processed RNA transcripts" [28].

SnoRNAs are encoded in introns of ribosomal protein genes and other housekeeping genes [29, 30], and are responsible for both sequence-specific methylation and pseudouridilation of RNA [31]. SNORA62 isan H/ACA box snoRNA that guides the isomerization of uridine into pseudouridine [30] by binding with 2 uridines of 28S rRNA (U3830 and U3832). Sequence comparison shows that these important regions display mutations in RPSAP8 but are conserved in RPSAP9 (Figure 4). As a result, the paralog of SNORA62 is probably not functional in RPSAP8. In RPSAP9 on the other hand, the paralog of SNORA62 could, in addition to SNORA62, exhibit the function of pseudouridilation in case of transcription [29]. Marcos-Carcavilla et al. [18] had already postulated the existence of a non-processed pseudogene that differed from the active RPSA gene by the absence of a G at position 29 of intron 4. Thus, we hypothesize that this previously mentioned non-processed pseudogene is in fact the semi-processed pseudogene RPSAP9, because it lacks the G at position 29 of intron 4 and it can co-amplify with the active RPSA gene because of its high sequence identity (98%).

Annotation of the mini-contigs by comparative mapping and FISH localization

The genomic regions containing the 12 members of the RPSA gene family were further investigated by sequence comparison of both BES and internal BAC sequences using NCBI BLAST [21] (Figure 1 and 2). Sixty-two of the 68 BES were annotated while the remaining 6 contained either too many repeat sequences or no specific orthologous sequence to allow annotation. The different characteristics (length, repeat sequences and genes) are listed in the Additional file 5. Based on sequence annotation results, 40 ovine genes, of which 37 have not been described in sheep yet, could be mapped on the mini-contigs by comparative mapping with the bovine genome (Figure 1 and 2). The primers used to perform the PCR for annotating the genes, together with another 18 optimized primer pairs, amplifying genes not present in the mini-contig but flanking the genomic region of the different RPSA family members, are listed in Additional file 6.

The 11 pseudogenes were localized by FISH on different sheep chromosomes (see Table 4 and pictures of the FISH experiments in Additional file 7). All the localizations confirmed the positions predicted from the genes present in the mini-contigs by using the online tool Virtual Sheep Genome Assembly v2.0.

The genomic region around the ovine RPSA family members show conserved synteny (same genes, same orientation and same order) with the bovine genome. LOC784055, probably a processed pseudogene of GOLPH3L located in intron 2 of GON4L on Bos taurus chromosome (BTA) 3 and expected in the ovine mini-contig containing RPSAP2, was the only bovine ortholog not present in sheep and therefore is most probably a bovine specific pseudogene.

A BLAST analysis of the bovine genome (reference assembly, based on Btau_4.0) with both ovine and bovine RPSA and RPSA pseudogene sequences identified 60 potential RPSA family members (Additional file 8). These included the only bovine pseudogene described so far, designated as RPSAP1 and located on BTA4 [32]. No ortholog of this pseudogene was found in sheep. Twenty-five sequences were annotated as 'similar to Ribosomal protein SA pseudogene' but only one corresponded to an ovine ortholog i.e. RPSAP11. To date, the 35 remaining sequences have not been annotated, but we have identified an ovine ortholog in 4 cases (Table 4 and 5; Additional file 8). Apart from RPSAP3, the ORF of the ovine and bovine orthologs differ substantially, suggesting that there is no selective pressure to conserve the ORF of these pseudogenes.

No bovine ortholog was found for the 6 sheep RPSA pseudogenes sharing 86 to 99% nucleotide identity with RPSA whereas the 5, for which a bovine ortholog was identified, only displayed 74-83% sequence identity with RPSA. As the amount of mutations accumulated by the pseudogenes during evolution can be used to infer their age [27], it's not surprising that the first group, consisting of recently arisen pseudogenes which have not yet accumulated many mutations, is lineage specific and that the pseudogenes of the latter group, comprising the oldest pseudogenes, all have a bovine ortholog. In addition, none of the 11 ovine pseudogenes were orthologous with any of the 63 annotated human RPSA pseudogenes. As a result, we can conclude that all 11 ovine RPSA pseudogenes detected originated after the divergence between primates and ungulates and 6 of these after the divergence between cattle and sheep.

Transcription profiling by RT-PCR

To investigate whether some of the ovine RPSA pseudogenes were potentially functional, transcription profiling was performed by RT-PCR for all sheep RPSA family members in 7 tissues (Figure 5) i.e. cerebrum, cerebellum, spleen, muscle, lymph node, duodenum and blood. To be sure that no genomic DNA was present in the RNA samples, they were treated with DNase and checked by minus RT control PCR (Additional file 9). For 8 members of the RPSA family, gene-specific primers could be designed and their specificity was proven by checking that the primers did not amplify any other RPSA family member using the respective unique BAC clones as template (Additional file 10). Because RPSA, RPSAP1, RPSAP7 and RPSAP9 share a high level of sequence identity, no specific primers could be designed for these RPSA family members, they were tested with aspecific primers. All generated amplicons were sequenced. RPSA was expressed in all tested tissues. This agrees with the results of Marcos-Carcavilla et al. and Qiao et al. [18, 33]. None of the pseudogenes was transcribed in blood. RPSAP2 and RPSAP11 were transcribed in all other tested tissues, while RPSAP3, RPSAP5 and RPSAP8 were only transcribed in one or more brain regions and RPSAP4 was transcribed in brain regions and spleen. RPSAP6 and RPSAP10 were not expressed in any of the tested tissues. In the case of RPSAP1, RPSAP7 and RPSAP9, tested with aspecific primers which all could also amplify RPSA, we generated amplicons which, after sequencing, turned out to be all RPSA transcripts. Thus we can conclude that RPSAP1, RPSAP7 and RPSAP9 are not expressed or at a very low level compared to the active RPSA gene. Therefore it would be interesting to do RT-qPCR with specific probes in order to be sure if the pseudogenes are expressed at very low levels or not at all. No clear relationship between the transcription profile of the various pseudogenes and the in silico prediction of possible promoters was observed. For instance, RPSAP10 is not expressed in any tissue tested although we did predict a promoter in the upstream sequence. Thus it may be possible that RPSAP10 is expressed in other tissues not examined in this study or that it has a low level of transcription. In addition, the in silico predicted promoter might not act as a cis-regulatory element in vivo. In contrast, RPSAP3 is transcribed in certain brain regions although we did not predict any promoter, probably because the promoter is located more upstream than the region analyzed here.

Construction BAC mini-contigs

The ovine INRA BAC library, consisting of 90.000 clones with an average insert length of 123 kb and a genome equivalent of 3.4, was screened by PCR [19]. The primers were designed using Primer3, based on conserved regions in the sheep RPSA gene [34]. The conserved regions were detected by comparison of all ovine ESTs available in GenBank that shared similarity with the published ovine mRNA sequence of RPSA [GenBank:EF649775] with BLAST and ClustalW [21, 35]. PCR was conducted with Faststart Taq DNA Polymerase (Roche). PCR conditions were 5 min at 95°C, 40 cycles of 30 s at 95°C, 30 s at 50°C and 1 min at 72°C, and a final 10-min elongation step at 72°C. Thirty-nine superpools, each consisting of 44 pools (24 plates, 8 rows and 12 columns), were screened. Each positive combination was verified by colony PCR.

All isolated BACs were grown in a 200 ml culture from which DNA was purified with the Qiagen Plasmid Midi kit (Qiagen) according to the manufacturer's instructions. The BAC ends were sequenced with the universal primer (UP) (5'-CGACGTTGTAAAACGACGGCCAG-3') and reverse primer (RP) (5'-CACAGGAAACAGCTATGACCATGATTACG-3') primers with 1 μg of purified BAC DNA as template. Unique STS primer pairs (Additional file 2), based on the BESs, were used to screen all isolated BACs and to construct mini-contigs.

All sequencing was performed with the Big Dye Terminator mix (Applied Biosystems) and analyzed on an ABI-3730xl Analyser (Applied Biosystems).

Characterization of RPSA gene family members

The primers used to screen the INRA BAC library were used as initial sequence primers to sequence the RPSA family member in one BAC of each mini-contig by direct sequencing. The obtained sequence was then used to develop new sequencing primers until the whole gene and an additional ± 500 bp upstream and ± 500 bp downstream of the sequence showing similarity with RPSA, was sequenced. If the screening primer did not work as sequencing primer, the amplicon generated with the screening primer was cloned into a pCR 2.1 vector with the TA Cloning Kit (Invitrogen) and the vector was transformed in DH5α Competent Cells (Invitrogen). The insert was then sequenced with UP and RP primers. All sequences were assembled into continuous sequences with CAP3 and analyzed with FGENESH and NCBI ORF Finder [20, 36, 37]. Promoter sequences were searched with CISTER, Neural Network Promoter Prediction, FPROM and TFsearch [36, 38–40].

Annotation of the mini-contigs by comparative mapping

All mini-contigs were annotated by comparing the BESs against bovine and human genomic sequences with NCBI BLAST [21]. Internal sequences were also annotated by PCR with primers based on bovine sequences of genes that were expected to be present in the mini-contig. All amplicons were verified and repeats were detected with Repeatmasker [26].

FISH

Fluorescent in situ hybridization was performed at INRA in Jouy-en-Josas (France). To prepare the probes, BAC DNA was extracted according to standard protocols and purified with the S.N.A.P. K1900-01 Miniprep kit (Invitrogen). DNA was then nick-translated with biotin-14-dATP (BioNick 18247-015 labeling system, Invitrogen) and mixed with 100× total sonicated herring sperm DNA and 100× total sonicated sheep DNA. Subsequently, it was precipitated with ethanol, slightly dried and resuspended in hybridization buffer.

For R-banded sheep chromosomes, embryo fibroblast cell cultures were synchronized with an excess of thymidine and treated with 5-bromo-2'-deoxyuridine during the second half of S phase [41].

FISH, signal detection and R-banding were performed as previously described [42]. Briefly, labeled probes were denatured at 100°C for 10 min and pre-hybridized at 37°C for 30 to 60 min before hybridization to the chromosomes. Chromosome identification and band numbering followed the standard sheep ideogram reported in ISCNDB2000 [43].

Transcription profiling

Fresh tissue samples were obtained from a commercial sheep slaughterhouse, frozen in liquid nitrogen immediately after slaughtering, crushed into powder and frozen at -80°C. Total RNA was isolated with the Aurum Total RNA Fatty and Fibrous Tissue kit (Bio-Rad) as described in the instruction manual. Subsequently, a minus RT-PCR was performed with actin, beta (ACTB) primers on 1 μl RNA to confirm the absence of any DNA contamination (Additional file 9) as previously described [44]. If DNA was still present in the sample, an additional DNase treatment with RQ1 RNase-free DNase (Promega) and a spin column purification with Microcon YM-100 (Millipore) were carried out.

The RNA concentration and purity of the samples were measured with the Nanodrop ND-1000 Spectrophotometer (Isogen) and the RNA quality was determined by evaluation of the 28S and 18S ribosomal bands on a 0.8% agarose gel.

Then, 0.2-1 μg RNA was converted into cDNA with iScript cDNA synthesis kit (Bio-Rad) using random and oligo dT primers. A confirmation PCR on 10× diluted cDNA with ACTB primers (giving a different amplicon length on gDNA and cDNA) was performed.

Specific primers, based on the aligned sequences of the different RPSA family members (Additional file 3) were designed for 8 members of the RPSA family and specificity was proven (see above).

Due to the high level of nucleotide sequence identity among RPSA, RPSAP1, RPSAP7 and RPSAP9, it was not possible to develop specific primers for these, but we were able to develop several primer pairs which amplified different combinations of 2 to 6 RPSA family members. One primer for instance amplifies RPSA, RPSAP1, RPSAP7 and RPSAP11; another RPSA and RPSAP7 and a third one RPSA, RPSAP1, RPSAP5 RPSAP7 and RPSAP11.

The obtained amplicons were sequenced to determine/confirm which (pseudo)gene was transcribed.