Fig. 7

RT-PCR confirms the expression of different examples of selected putatively protein-coding splice variants. (A) Schematic representations of the splice variants (annotated and newly identified), and agarose gels showing their RT-PCR amplification. Ensembl annotations are depicted on the left as “ENSMUST” followed by the corresponding Ensembl numberings. Unannotated transcript isoforms are depicted with the label “MSTRG”. The designed primer sets for the amplification of either the annotated or the unannotated isoforms are shown for each case (arrows), together with the expected PCR product sizes (bp). The gray arrow above each diagram indicates transcription direction. Genomic location, as well as chromosome number, are indicated in each case. Whenever necessary, magnified insets are shown below each representation for better visualization of the amplified regions. A table with the coverage of the annotated and unannotated transcripts in the four cell populations is included in each case. The asterisks indicate the cell population in which the unannotated isoform was most highly expressed. In the agarose gels, A stands for the annotated splice variants, and U for the unannotated ones. Gels have been cropped for the sake of clarity (original agarose gels are presented as Supplementary Figure S7). (a) Msh5 splice variants that encode the canonical 833 amino acids protein (yellow), and an unannotated splice variant encoding a putative 362 residues isoform (red). (b) BC051142 most highly expressed annotated variant (red), and two unannotated putatively coding variants with a differential expression pattern along spermatogenesis (one is mostly differential of spermiogenesis, while the other progressively increases from early meiotic prophase to spermiogenesis; yellow and orange, respectively). In the lane corresponding to the annotated variant, two additional faint bands can be observed, most probably corresponding to the amplification of a couple of weakly expressed isoforms (due to the extremely high number of isoforms detected for this gene, it was not possible to design primer combinations to exclusively recognize only one variant). (c) Agbl5 canonical transcript (orange), which encodes an 846 residues protein, and a selected unannotated variant (red) encoding a putative much shorter isoform of 412 amino acids. (d) The chosen Larp1 unannotated splice variant (orange) encodes a putative not reported protein isoform of 760 amino acids, unlike the canonical one (yellow), whose encoded protein is 1,072 residues long. As shown, in PS the expression levels of the new variant are fifteen-fold higher than those of the canonical one. (e) The unannotated Stk31 isoform we chose for confirmation (red) encodes a shorter variant that is much more highly expressed than the canonical one (yellow). The comparatively poorer amplification of the unannotated variant is due to the fact that the region did not allow the design of a good pair of primers. (f) Representation of an annotated Baz1a isoform (light yellow), and the unannotated splice variant (red), which is much more highly expressed all along spermatogenesis, upregulates in PS, and directs the synthesis of a shorter protein. In this case, amplification was performed with a primer set that simultaneously amplifies a region of both the annotated (312 bp) and unannotated (216 bp) variants. The annotated isoform is poorly amplified, presumptively due to its competition with the newly identified one, which is expressed at much higher levels (see table). Besides, a band corresponding to the amplification product with a primer set that only recognizes the unannotated isoform is shown to the right. (g) Rsph1 was chosen as an example of a novel coding isoform generated through exon-skipping (yellow, while the canonical isoform is represented in red). Although the primer set was intended to amplify the annotated variant as well, yielding a larger, 265 bp band, the latter was not detected most probably because of its competition with the unannotated isoform. B) Representative schematic diagrams of two of the canonical and unannotated putative protein isoforms, to exemplify the differences between them. MSH5: The orange line in the “novel” isoform represents the first 133 amino acids, which are completely different from those of the canonical protein. STK31: While both isoforms present a Tudor domain, the predicted variant would lack the protein-kinase domain, which is essential for its function as a serin-threonine kinase in the canonical isoform