Three MSYq-deficient mouse models were used in this study (Figure 1): XY^RIIIqdel, with deletion of about two-thirds of MSYq; XY^Tdym1qdelSry, with deletion of about nine tenths of MSYq; and XSxr^aY* ^X, in which the only Y specific sequences are provided by the Y short arm derived sex-reversal factor Sxr^a, and thus lack all of MSYq. These models are hereafter abbreviated as 2/3MSYq^-, 9/10MSYq^- and MSYq^-.

To analyze the testis transcriptomes of these MSYq deficient mice we used microarrayed testis cDNA clones enriched for cDNAs deriving from spermatogenic cells (see Materials and methods, below). Total testis RNA from each of the MSYq-deficient mice (labeled with the fluorochrome Cy3) and from matched controls (labeled with Cy5) was hybridized to the array; four technical replicates were obtained for each model. The initial raw data were filtered to select only clones for which fluorescence intensity data were available from at least two of the technical replicates for each model. After this filtering, 14,681 clones remained from the complete set. As expected, scatter plots of these filtered data show that expression of the vast majority of clones is unchanged between the MSYq deficient and control mice (Figure 2a).

For transcripts encoded by single or multiple copy Y genes mapping to MSYq, substantially reduced levels should occur in one or more of the deletion models. In order to focus on potential MSYq encoded transcripts we restricted our analysis to clones exhibiting a twofold or greater reduction in fluorescence intensity relative to control and a t test probability of under 1% for the comparison between the replicates for the Cy3 and Cy5 fluorescence intensities. Twenty-three clones were identified as substantially reduced by these criteria in one or more of the MSYq deletion models. BLAST (basic local alignment search tool) comparisons were used to identify matching cDNAs (if previously identified) and/or the chromosomal locations of the encoding sequences. Sixteen of the 23 proved to be Ssty cDNA clones, thus demonstrating the efficacy of the strategy. Of the remaining seven clones, five proved to be Y encoded Sly cDNAs (see below), one was X encoded and one was autosomally encoded (Figure 2a, b).

All five of the additional Y encoded clones were found to have homology to regions of the cDNA clone BC049626 previously identified in a large-scale cDNA sequencing project 26 (Additional data file 1). This cDNA clone initially had no chromosomal assignment, but it was subsequently ascribed to a gene, Sly, that maps to MSYq (MGI:2687328; Mouse Chromosome Y Mapping Project [Jessica E Alfoldi, Helen Skaletsky, Steve Rozen and David C Page at the Whitehead Institute for Biomedical Research, Cambridge, MA, USA, and the Washington University Genome Sequencing Center, St. Louis, MO, USA]). Sly is a member of a multicopy family, and in December 2004 a total of 65 Sly family members were predicted based on the Y sequence data. There is a related multicopy X gene family that includes Xmr and Xlr 2728.

To confirm the downregulation of the Sly transcript in the MSYq deletion mice, we probed a northern blot of total testis RNA from the three MSYq deficient models and their controls with microarray clone MTnH-K10 (Figure 3a) and subsequently with the full BC049626 cDNA (not shown). The hybridization with both probes revealed high transcript levels in the control testes, a clear reduction in 2/3MSYq^- testes, and further marked reductions in 9/10MSYq^- and MSYq^- testes. The reductions as estimated by phosphorimager analysis with clone MTnH-K10 (with the microarray estimates in brackets) were as follows: 2/3MSYq^-, 47% (37%); 9/10MSYq^-, 81% (84%); and MSYq^-, 83% (91%). Because Sly has substantial homology to Xmr, which is also strongly transcribed in the testis, we considered that the remaining hybridization in the two models with severe MSYq deficiency could be due to cross-hybridization with Xmr transcripts. We therefore designed primers for reverse transcriptase polymerase chain reaction (RT-PCR) that distinguished between the two transcripts and used these to amplify template from testis cDNAs derived from the three MSYq deficient models and controls (Figure 3b). A 266 base-pair (bp) product was amplified from all samples, and sequencing confirmed that this derived from Xmr. A 227 bp product was also amplified from the control, 2/3MSYq^- and 9/10MSYq^- testes (although clearly reduced), but there was no amplified product from MSYq^- testes. Sequencing of this 227 bp product from 2/3MSYq^- and 9/10MSYq^- testes confirmed that it derived from Sly. Thus, only MSYq^- completely lacks Sly transcripts.

In December 2004, a BLAST analysis of the array clone 8832_f_22 against the mouse genome registered 41 hits on the mouse X chromosome and 710 hits on the mouse Y chromosome. The arrayed cDNA clone was apparently X encoded, there being eight putative loci (for example, gi:4640881, 3118-6344) with four exons that would encode matching cDNAs; the remaining nine hits were from incomplete loci or short sequence fragments. To investigate the coding potential of the Y sequences identified in the initial analysis, a further BLAST analysis was carried out with a complete X locus, and this identified 123 putative Y chromosomal loci (for example, gi:33667254, 73667-76894) that retain the same intron/exon structure as the X loci, and with 92-94% sequence identity across exons and introns.

The arrayed X encoded cDNA clone shared homology with three previously identified testis cDNA clones - two apparently X encoded (CF198098, AK076884) and one Y encoded (AK016790) - and a BLAST of mouse expressed sequence tags (ESTs) identified further related testis transcripts together with others derived from small intestine, aorta and eight cell embryos. The testis cDNAs and ESTs are summarized in Figure 4a. The arrayed X encoded cDNA, together with the two exactly matching transcripts BF019211 and CF198098, derive from purified spermatocyte cDNA libraries 29. Of the Y encoded testis transcripts, EST AV265093 matches only a short stretch of exon 4 and the two remaining transcripts, namely AK016790 and BY716467, show only segments of homology to the arrayed clone. Nevertheless, amplification from testis cDNA with an Asty exon 1/exon 4 primer pair generated products of 437 bp and 622 bp, and sequencing confirmed that these derived from Y chromosomal loci, the smaller product lacking exon 3. We refer to the similar X and Y loci encoding these X and Y transcripts as Astx and Asty (Amplified spermatogenic transcripts X encoded/Y encoded), respectively (for Astx and Asty transcript sequences, see Additional data file 2).

Intriguingly, the Y encoded transcripts AK016790 and BY716467, in addition to the sequence matching Astx/Asty exons 2 and 3 (and part of the intervening intron), proved to have exonic sequence matching Ssty1 exon 1 and part of exon 3, together with further sequence matching another testis cDNA AK015935 (Figure 4a). BLAST searches identified 'recombinant' Y genomic loci (comprising partial Ssty1 and Asty loci followed by sequence that includes exons matching AK015935) that could encode these transcripts (Figure 4b, Additional data file 3). We refer to these loci as Asty(rec).

The close homology between Astx and Asty suggested that the arrayed Astx cDNA clone would cross-hybridize with Asty and Asty(rec) RNAs. We designed primers from Astx/Asty exon 4 for RT-PCR that we thought should specifically amplify either Astx or Asty, but further analysis (see below) identified Asty(rec) transcripts that also include exon 4. RT-PCR analysis using these primers showed that Asty and/or Asty(rec), rather than Astx transcripts, are reduced in MSYq deficient mice (Figure 4c). Probing the northern blot of total testis RNA from the three MSYq deficient models and their controls with an Asty exon 4 probe revealed a transcript of about 1 kilobase (kb), together with other larger transcripts with sizes ranging up to more than 9.5 kb. All bands exhibited progressive reduction in intensity with increasing MSYq deficiency (Figure 4d), and thus we are confident that it is not due to cross-hybridization to Astx. The size for the two transcripts identified by RT-PCR is of course unknown. However, given that the microarrayed Astx cDNA is 1.5 kb, it is reasonable to assume that the two Asty transcripts should be approximately 1.3 kb (lacking exon 3) and 1.5 kb; faint bands approximating these sizes are present. In further attempts to determine the origins of these multiple sized transcripts we probed the blot with a probe matching part of exon 6 of the Asty(rec) transcript AK016790 (Figure 4b) and found that the bands of about 7.5 kb and above hybridized to this probe (not shown). Thus, there are transcripts derived from the 'recombinant' loci that also include Asty exon 4. Because the recombinant loci lack Asty exon 1 we then probed the blot with an Asty probe from exon 1, but no convincing hybridization was obtained. We conclude that the transcripts detected by the exon 4 probe that dose with the extent of the MSYq deletions derive predominantly from the Asty(rec) loci.

We next used the microarrayed Sly clone MTnH-K10 and the Asty exon 4 probe to hybridize to Southern blots of DNA samples from the MSYq deficient and control mice. The Sly probe revealed multiple hybridizing bands in control males that were reduced in intensity in 2/3MSYq^- males and exhibited no detectable hybridization in 9/10MSYq^- or MSYq^- males, or in females (Figure 5a, c). However, after long exposure of a further blot that included DNA from XX and XO females, multiple bands were detected in 9/10MSYq^- and MSYq^- males that were also present in XX and XO females (Figure 5e). In the females the band intensities dosed with the number of X chromosomes and thus presumably were due to cross-hybridization with Xlr/Xmr. However, at least three bands that are absent from females were retained in 9/10MSYq^- males, but were absent from MSYq^- males. Furthermore, a genomic PCR for the first Sly intron amplified from 9/10MSYq^- but not from MSYq^- DNA. This is consistent with the finding of Sly transcripts by RT-PCR in 9/10MSYq^- males but not in MSYq^- males. The Asty exon 4 probe also detected multiple hybridizing bands in the control males that reduced in intensity in the 2/3MSYq^- males (Figure 5b). There was a faintly hybridizing band in females, presumably due to cross-hybridization with Astx. The blot with all three MSYq deficient DNAs exhibited no hybridizing bands in 9/10MSYq^- and MSYq^- that were of the sizes of the four predominant Y-specific bands seen in the controls, but there was some unexplained intense hybridization at the level of the presumed Astx band and above.

Probing a multiple mouse tissue polyA northern blot (Figure 6a) with Sly clone MtnH-K10 indicated that Sly transcription is restricted to the testis, but the Asty exon 4 probe also detected transcripts in heart, consistent with there being a Y encoded EST from aorta (CA584558). In northern blots of RNAs from 12.5 days postpartum (dpp) to 30.5 dpp testes (Figure 6b), Sly and Asty/Asty(rec) transcripts were first detectable at 20.5 dpp, suggesting they are restricted to spermatid stages. To confirm spermatid expression we used the same probes for RNA in situ on testis sections, and for Sly (Figure 6c, d) high level spermatid specific expression was confirmed. For Asty/Asty(rec) hybridization was at a level not markedly above background, but it did appear to locate to round spermatids, particularly with respect to hybridization within the nucleus; however, there was a similar localization to round spermatids in the sense control (Figure 6e). We then found that the previously described transcript BU936708 derives from Y 'loci' that transcribe through Asty loci in the antisense orientation and include antisense sequence from Asty exon 4 (Additional data file 3); this transcript will hybridize to the sense control probe. Thus, we suspect that the round spermatid nuclear localization with the antisense probe does reflect the presence of Asty/Asty(rec) transcripts. However, it is apparent from the northern analysis and in situ analysis that Asty/Asty(rec) transcripts are much less abundant than those of Sly, which is consistent with there being five Sly clones but no Asty/Asty(rec) clones on the testis cDNA microarray.

Because we believe the transcripts detected by the Asty exon 4 probe derive predominantly from the Asty(rec) loci that are almost certainly driven by the spermatid specific Ssty1 promotor, we made attempts to determine whether true Asty transcripts are also spermatid specific. For this we used an Asty exon 1-4 primer pair (previously used to provide evidence for Asty transcripts) to amplify testis cDNA samples from 1.5 dpp to adult. Transcripts were weakly detected by these primers at 14.5 and 18.5 dpp, but the predominant expression was at 22.5 dpp onward when there were transcripts of two sizes (Figure 6f).

We know these primers can also amplify Astx transcripts, and so we sequenced cloned RT-PCR products to confirm their identity. This confirmed the presence of the two previously identified Asty transcripts (one of which lacks exon 3). The longer transcript was detected from 14.5 dpp onward, and the shorter transcript from 22.5 dpp onward. Because spermatids first appear at about 20 dpp, we conclude that the shorter Asty isoform appears to be spermatid specific, but that the longer one is not spermatid specific.

Of the 65 loci of the Sly gene family, 34 have retained coding potential for a protein that is related to three previously described chromatin associated nuclear proteins: XLR and XMR, which are encoded by members of a complex multicopy X gene family 28303132, and the autosomally encoded SYCP3 protein, via a conserved COR1 region 3334 (Figure. 7a, b). The putative SLY protein is most similar to XMR (amino acid identities: XMR 48%, XLR 46% and SCP3 30%), but within the COR1 region it is most similar to XLR (amino acid identities: XLR 79%, XMR 50% and SCP3 37%).

All eight Astx loci have a modest open reading frame (ORF) in exon 4, six of which would encode a protein of 106 amino acids and two would encode a protein with a carboxyl-terminal extension to 122 amino acids (Figure 8a). Surprisingly, given the greater than 92% sequence identity of the Asty and Astx loci, this exon 4 ORF is not conserved in any of the putative Asty loci so far identified. There are six Asty loci with an ORF in exon 1 that could encode proteins of 116 or 120 amino acids (Figure 8b). However, the Asty sequence we obtained by RT-PCR from testis cDNA (Additional data file 2) did not match the sequence of this subset of loci with an exon 1 ORF; thus, it appears likely that the Y-linked members of this family no longer produce a functional protein. Assessing whether any of the novel Asty(rec) transcripts have significant protein encoding potential will require detailed characterization of this complex family of transcripts.

All mice were produced on a random bred albino MF1 strain (NIMR colony) background. The mice used to provide RNA for microarray analysis were the three MSYq deficient genotypes (Figure 1) that we have previously analyzed with respect to Ssty expression and sperm abnormalities, together with appropriate age- and strain-matched controls.

Testes were obtained from two of each of the MSYq deficient males and the two control genotypes, at 2 months of age. Total RNA was isolated using TRI reagent (Sigma-Aldrich, Poole, Dorset, UK) and cleaned using RNEasy columns (Qiagen, Crawley, West Sussex, UK), in accordance with to the manufacturers' protocols. Microarray hybridizations and analysis were carried out as described by Ellis and coworkers 43, except that microarray data normalization was based on the global median signal for Cy3 and Cy5 channels rather than on a panel of control genes. This form of normalization is valid because the majority of genes do not vary between the mutant and control samples for any of the models (Figure 2a). Briefly, 10 μg total RNA was fluorescently labeled using an indirect protocol, and the test and control samples were allowed to co-hybridize to the array. Four technical replicate slides were obtained for each test/control comparison. Cy3 was used to label the test (mutant) sample and Cy5 the control (normal) sample in all cases. Fluorescence intensities provide measures of the relative abundance for each hybridizing testis transcript for each genotype, whereas Cy3/Cy5 ratios for individual clones provide a measure of the levels in each of the MSYq deficient models relative to their matched controls.

The arrayed clones consisted of two subtracted adult mouse testis libraries 43, six testis cell type separated libraries (IMAGE clone plate numbers 8825-8830, 8831-8836, 8846-8850, 9339-9342, 13869-13871, 13872-13874 29), and appropriate controls 43. From sequence analysis performed to date, the combined gene set included cDNAs derived from more than 4,000 genes, often represented by multiple clones, enriched for cDNAs deriving from spermatogenic cells. Slides were scanned using an ArraywoRx CCD-based scanner and the resulting images quantitated using GenePix. Raw data were processed in Excel to remove data from spots flagged as bad or not found, and from features with a background subtracted intensity of under 100 in both channels. Global median normalization, t tests, and fold change filtering were performed using GeneSpring. Full details of the array experiment were submitted to the ArrayExpress database (accession number E-MEXP-251).

The probes used for Southern analysis were Sly cDNA clone BC049626 26 and an Asty 314 bp exon 4 probe amplified with primers CAGCAAGGAGAGTGGGGAGTA and CAGTGGGATGTTGGTTTCTAATG. Genomic DNA was extracted from tail biopsies and 15 μg (or 4 μg for the control XY on the long exposure Southern blot) was digested with EcoRI, electrophoresed through a 0.8% agarose gel and transferred to a Hybond-N membrane (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK). After UV crosslinking (StrataLinker™. Stratagene, La Jolla, CA, USA), the membrane was hybridized overnight with ³²P-labelled probes either at low stringency (55°C; hybridization buffer: 6 × salt sodium citrate (SSC), 5 × Denhart's, 0.1% sodium dodecyl sulfate (SDS), 100 μg/ml salmon sperm DNA; two 30 minute washes (one with 0.5 × SSC and 0.1% SDS, and one with 0.1 × SSC and 0.1% SDS)) or at high stringency (60°C; hybridization buffer: 6 × SSC, 5 × Denhart's, 0.5% SDS, 100 μg/ml salmon sperm DNA; two 30 minute washes (one with 0.5 × SSC and 0.1% SDS, and one with 0.1 × SSC and 0.1%SDS)). The membrane was exposed to X-ray film or Phosphorimager screen overnight.

The probes used for northern analysis were as follows: Sly cDNA clones BC049626 26 and MtnH_K10 from the microarray (Additional data file 1), and the Asty 314 bp exon 4 probe used for northern analysis. An actin probe that recognizes α- and β-actin transcripts 44 served as a control for RNA integrity. Total RNA (20 μg) was electrophoresed in a 1.4% formaldehyde/agarose gel and transferred to Hybond-N membrane (Amersham) using 20 × SSC buffer. The RNA was cross-linked to the membrane with UV (StrataLinker™), the membrane was fixed for 1 hour at 80°C, and hybridized overnight at 60°C with ³²P-labelled probes in hybridization buffer (6 × SSC, 5 × Denharts, 0.1% SDS, 50 mmol/l sodium phosphate, 100 μg/ml salmon sperm DNA). After two 60°C washes (30 min with 0.5 × SSC and 0.1% SDS, and 30 min with 0.1 × SSC and 0.1% SDS) the membrane was exposed to X-ray film for 5 hours and subsequently to a Phosphorimager screen to allow quantitation of hybridization using ImageQuant software.

RNA samples were treated for DNA contamination using DNAse I amplification grade kit (Invitrogen, Paisley, UK). For the Sly/Xmr RT-PCR, 2 μg total RNA was reverse transcribed in a 40 μl reaction using standard procedures. A 2.5 μl aliquot was then added to a 25 μl PCR reaction. RT-PCR for Astx/Asty was performed using the Qiagen OneStep RT-PCR kit, following the manufacturers' instructions. In both cases amplification was for 30 cycles at an annealing temperature of 60°C.

The primers used were as follows: Sly and Xmr, forward primer GTGCGGTTTGGAAGTGT and reverse primer CTCAAGCAGAAGCAGATG; Asty and Astx exon 4, forward Asty primer GRGGAGTAGAACTCATCATC and forward Astx primer GGGGAGTAGAACTCATCTTTA, with common reverse primer CAGGAGATGACTAACATAGCA; Asty exon 1 to exon 4, forward GGCCTTGCTCTTATGTCATC and reverse CGATGATGAGTGACCTAAAGAT; and Astx exon 1/2 (spanning intron 1) to 3/4 (spanning intron 3), forward GCTCCAGAAGACAGAGATAC and reverse AGACTTCAAACCTCATGCAGT.

RT-PCR product was purified using the QiaQuick kit (Qiagen) and cloned using the pGEM-T easy vector system I kit (Invitrogen). Sequencing reactions were performed from 5' and 3' ends using standard protocols (BigDye; Amersham). Completed reactions were analyzed by the Cambridge Department of Genetics sequencing service using a 3130xl capillary system (Applied Biosystems, Warrrington, Chesire, UK). Cycling conditions for the sequencing reactions were 96°C, 55°C and 60°C for 10 s, 5 s and 4 minutes, respectively.

In situ hybridization using digoxigenin-labeled cRNA probes from Sly clone MtnH_K10 and Asty and Astx exon 4 were used to localize each mRNA on Bouin-fixed paraffin-embedded mouse testis sections using procedures previously described 45, with hybridization and washing temperatures up to 55°C. Both antisense and sense (negative control) cRNAs were used on each sample, in every experiment, and for each set of conditions tested.

RNA fluorescence in situ hybridization (FISH) was performed basically as described previously for Cot1 RNA FISH 46 using an X chromosomal BAC clone RP23-83P7 that contains the Astx locus encoding the cDNA AK076884 (BACPAC Resources, Oakland, CA, USA). Hybridization reactions consisted of 100 ng biotin-labeled BAC probe, 3 μg mouse Cot1 DNA and 10 μg salmon sperm DNA, and were carried out overnight at 37°C. Staging of spermatogenic cells was based on DAPI fluorescence morphology, together with immunolabelling for the synaptonemal complex protein SYCP3 (rabbit anti-SYCP3; Abcam, Cambridge, UK) and the phosphorylated form of histone H2AX (mouse anti-gamma H2AX; Upstate, Dundee, UK), as previously described 4647. The chromosomal source of the RNA FISH signals was first checked by DNA FISH with a digoxigenin labeled RP23-83P7 BAC probe prepared using the Digoxigenin Nick Translation Kit (Roche Diagnostics, Lewes, East Sussex, UK). Hybridizations were carried out as for RNA FISH with stringency washes as described previously 46, and the DNA FISH signals were developed using anti-DIG-FITC (Chemicon, Chandlers Ford, Hampshire, UK), diluted 1:10, for 1 hour at 37°C. Further confirmation of the X chromosomal source of the Astx transcripts was provided by X chromosome painting, as previously described 46.