24Gene regulation in the germline ensures the production of high-quality gametes, long-term 25 maintenance of the species, and speciation. Germline transcriptomes undergo dynamic 26 changes after the mitosis-to-meiosis transition in males and have been subject to evolutionary 27 divergence among mammals. However, the mechanism that underlies germline regulatory 28 divergence remains undetermined. Here, we show that endogenous retroviruses influence 29 species-specific germline transcriptomes in mammals. We show that the expression of 30 endogenous retroviruses, particularly the evolutionarily young K family (ERVK), is 31 associated with gene activation after the mitosis-to-meiosis transition in male mice. We 32 demonstrate that accessible chromatin and H3K27ac, a marker of active enhancers, are 33 tightly associated with ERVK loci as well as with the activation of neighboring evolutionarily 34 young germline genes. Thus, ERVKs serve as evolutionarily novel enhancers in mouse 35 spermatogenesis. These ERVK loci bear binding motifs for critical regulators of 36 spermatogenesis such as A-MYB. The genome-wide transposition of ERVKs might have 37 rewired germline gene expression in a species-specific manner. Notably, these features are 38 present in human spermatogenesis, but independently evolved ERVs are associated with 39 expression of germline genes, demonstrating the prevalence of ERV-driven mechanisms in 40 mammals. Together, we propose a model whereby species-specific transcriptomes are fine-41 tuned by endogenous retroviruses in the mammalian germline.42 43 84for germline genes 25,26 . These enhancers drive expression of evolutionarily novel germline genes 85 after the mitosis-to-meiosis transition, thereby defining the species-specificity of germline 86 transcriptomes in mammals. Notably, we demonstrate the prevalence of ERV-driven germline 87 genes in humans. We propose a model whereby ERVs fine-tune species-specific transcriptomes in 88 the mammalian germline.
90
Results
91A large group of endogenous retroviruses are differentially expressed at the mitosis-to-92 meiosis transition in spermatogenesis.To determine the expression dynamics of repetitive elements interspersed throughout the 94 genome in spermatogenesis, we analyzed previously published RNA-seq datasets for representative 95 stages of spermatogenesis 5,6,27 using RepBase, a database of consensus sequences for repetitive 96 DNA elements 28 . We included in the analysis the transcriptomes of THY1 + undifferentiated 97 spermatogonia from postnatal day 7 (P7) testes, which contain spermatogonial stem cells and 98 progenitor cells; KIT + differentiating spermatogonia from P7 testes; purified pachytene 99 spermatocytes (PS) in the midst of meiosis; postmeiotic round spermatids (RS) from adult testes; 100 and mature sperm ( Fig. 1a). Among the 1,195 consensus repetitive DNA elements of Mus musculus 101 in RepBase, we detected 307 differentially expressed types of repetitive elements between at least 102 two stages of spermatogenesis ( Fig. 1b, Supplemen...