Transposon‐driven transcription is a conserved feature of vertebrate spermatogenesis and transcript evolution

Abstract

Spermatogenesis is associated with major and unique changes to chromosomes and chromatin. Here, we sought to understand the impact of these changes on spermatogenic transcriptomes. We show that long terminal repeats ( LTR s) of specific mouse endogenous retroviruses ( ERV s) drive the expression of many long non‐coding transcripts (lnc RNA ). This process occurs post‐mitotically predominantly in spermatocytes and round spermatids. We demonstrate that this transposon‐driven lnc RNA expression is a conserved feature of vertebrate spermatogenesis. We propose that transposon promoters are a mechanism by which the genome can explore novel transcriptional substrates, increasing evolutionary plasticity and allowing for the genesis of novel coding and non‐coding genes. Accordingly, we show that a small fraction of these novel ERV ‐driven transcripts encode short open reading frames that produce detectable peptides. Finally, we find that distinct ERV elements from the same subfamilies act as differentially activated promoters in a tissue‐specific context. In summary, we demonstrate that LTR s can act as tissue‐specific promoters and contribute to post‐mitotic spermatogenic transcriptome diversity. image Large‐scale transcriptome analyses of the mouse germline shows that a significant fraction of noncoding transcripts are expressed in late spermatogenesis and are associated with endogenous retroviral elements that drive their transcription. Transcriptome analyses of both noncoding and coding transcripts reveal a significant increase in gene expression during late stages of spermatogenesis. Noncoding transcription in particular is increased and is associated with endogenous retroviral elements ( ERV s) that overlay the transcriptional start sites of associated genes and drive their expression. This germline repeat‐driven transcription is observed in other species. Some of these ERV ‐driven loci appear to produce peptides even though they have computationally been determined to be non‐coding. The testis is a hotspot for this phenomenon, however this effect is also observed in other tissues based on publicly available data.