Topaz1, an essential gene for murine spermatogenesis, down-regulates the expression of numerous testis-specific long non-coding RNAs

Spermatogenesis involves coordinated processes, including meiosis, to produce functional gametes. We previously reported Topaz1 as a germ cell-specific gene highly conserved in vertebrates. Topaz1 knockout males are sterile with testes that lack haploid germ cells because of meiotic arrest after prophase I. To better characterize Topaz1−/− testes, we used RNA-sequencing analyses at two different developmental stages (P16 and P18). The absence of TOPAZ1 disturbed the expression of genes involved in microtubule and/or cilium mobility, which was consistent with testicular histology showing the disruption of microtubules and centrosomes. Moreover, a quarter of P18 dysregulated genes are long non-coding RNAs (lncRNAs), and three of them are testis-specific and located in spermatocytes, their expression starting between P11 and P15. The suppression of one of them, 4939463O16Rik, did not alter fertility although sperm parameters were disturbed and sperm concentration fell. The transcriptome of P18-4939463O16Rik−/− testes was altered and the molecular pathways affected included microtubule-based processes, the regulation of cilium movement and spermatogenesis. The absence of TOPAZ1 protein or 4930463O16Rik produced the same enrichment clusters in mutant testes despite a contrasted phenotype on male fertility. In conclusion, TOPAZ1 appeared to stabilize the expression of numerous lncRNAs. Its suppression is not essential for fertility but required during the terminal differentiation of male gametes. Author Summary The Topaz1 gene was initially characterized during the initiation of meiosis in the sheep fetal ovary. In order to determine its function, a KO of the murine gene was performed. In this species, only males were sterile and spermatogenesis was blocked before the first meiotic division. Here, we show that cytoskeletal elements are markedly disturbed in mutant testes, indicating that these elements play an important function in spermatogenesis. While the mitotic spindle of spermatogonia was normal, the meiotic spindle of spermatocytes was hemi-spindle-shaped and the homologous chromosome pairs could position themselves on the equatorial plate. In addition, lncRNAs account for 25% of genes whose expression in testes varies significantly in the absence of Topaz1. This suggests a key role for these factors in spermatogenesis. Largely testis-specific, they may be involved in spermatogenesis and play a more or less critical role in mouse fertility, which probably also depends on their redundancies.

2 Abstract 22 Spermatogenesis involves coordinated processes, including meiosis, to produce functional gametes. 23 We previously reported Topaz1 as a germ cell-specific gene highly conserved in vertebrates. Topaz1 In mammals, an organism derives from two parental haploid gametes, a maternal oocyte and paternal 53 sperm. Meiosis is a highly specialized event that leads to the production of these haploid germ cells 54 (Kleckner, 1996). In females, meiosis is initiated during fetal life while male germ cells are involved in 55 the meiosis process around puberty. In males, meiosis is essential during spermatogenesis that 56 involves mitotic division and the multiplication of spermatogonia, the segregation of homologous 57 chromosomes and the spermiogenesis of haploid germ cells. This complex process of spermatogenesis, 58 which progresses through precisely timed and highly organized cycles, is primordial for male fertility.

73
Since discovery of the maternal H19 lncRNA (Brannan et al., 1990) and the Xist (Brockdorff et al., 1992) 74 genes that regulate the structure of chromosomes and mediate gene repression during X chromosome 75 inactivation, interest in studying the role of non-coding RNAs (ncRNAs) has grown considerably. Non-76 coding RNAs are present in many organisms, from bacteria to humans, where only 1.2% of the human 77 genome codes for functional proteins (Carninci et    . Expression of the Testis-specific X-linked gene was specific to, and highly abundant 99 in, mouse pachytene-stage spermatocytes and could regulate germ cell progression in meiosis 100 (Anguera et al., 2011). Moreover, in male germ cells, it has been shown that the Dmrt1-related gene 101 negatively regulates Dmrt1 (doublesex and mab-3 related transcription factor 1) and that this 102 regulation might be involved in the switch between mitosis and meiosis in spermatogenesis (Zhang et

106
Following a previous study, which presented comparative microarray analyses of wild-type and 107 Topaz1 -/testis RNAs at P15 and P20 (Luangpraseuth-Prosper et al., 2015), we have now performed 108 deep sequencing by bulk RNA-sequencing (RNA-seq) of these testes collected at P16 and P18 in an aim 109 to refine the developmental stages that display transcriptional differences between the two mouse 110 lines. Since the proportion of deregulated lncRNAs represented about a quarter of the differentially 111 expressed genes (DEGs), we studied the testicular localization of three of them. In order to approach 112 the role of testicular lncRNAs, we created a mouse line in which one of them was deleted 113 (4930463O16Rik). These knockout mice displayed normal fertility in both sexes, but the male mutants 114 produced half as much sperm as wild-type controls.   Table 1). At P16, out of the 205 DEGs, 97 genes were significantly down-regulated (Log2FC<-1 or 135 FC<0.5) and 108 were up-regulated (Log2FC>1 or FC>2). However, at P18, down-regulated DEGs 136 accounted for 91% (2491 genes) and up-regulated genes for only 9% (257 genes). Among all these 137 DEGs, 120 were common to both P16 and P18 ( Figure 1A). According to the mouse gene atlas, the 138 2748 DEGs at developmental stage P18 were largely testis-enriched DEGs in mouse testis-specific 139 genes ( Figure 1B).    In order to evaluate a potential role in spermatogenesis for one of these lncRNAs, 4930463O16Rik (the 286 nuclear expressed gene), it was decided to suppress its expression in a mouse knockout model.

314
The absence of 4930463O16Rik does not affect mouse fertility 315 Fertility was then investigated in 4930463O16Rik-deficient mice. Eight-week-old male and female 316 4930463O16Rik -/mice were mated, and both sexes were fertile. Their litter sizes (7.5 ± 2.10 pups per 317 litter, n=28) were similar to those of their WT counterparts (6.9 ± 2.12 pups per litter, n=20). There 318 were no significant differences in terms of testicular size, testis morphology and histology and cauda 319 and caput epididymis between WT and 4930463O16Rik -/adult mice ( Figure 6A, B). In addition, the 320 different stages of seminiferous tubules divided into seven groups were quantified between 321 4930463O16Rik -/and WT adult mice. No significant differences were observed between the two 322 genotypes ( Figure 6C). These results therefore demonstrated that 4930463O16Rik is not required for 323 mouse fertility.     Table 6). An analysis that discriminated up-regulated from down-regulated genes only

480
Two days later, at P18, just before the prophase I-metaphase 1 transition, ten times more genes were 481 deregulated. Among the 120 DEGs common to P16 and P18, there was at least one gene that might be 482 involved in meiosis such as Aym1, an activator of yeast meiotic promoters 1. The absence of Topaz1 483 27 led to a lack of the testicular expression of Aym1. This gene is germ cell-specific (Malcov et al., 2004).

484
In male mice, Aym1 is expressed from 10 dpp in early meiotic spermatocytes. The small murine AYM1 485 protein (44 amino acids) is immunolocalized in the nucleus of primary spermatocytes, mainly late 486 pachytene and diplotene, suggesting a nuclear role for AYM1 in germ cells during the first meiotic 487 division (Malcov et al., 2004).

488
At P18, the testicular transcriptome of Topaz1 -/mice was largely disturbed when compared to WT 489 animals, and most DEGs were down-regulated ( Figure 1F), suggesting that TOPAZ1 promotes gene 490 expression in normal mice. As TOPAZ1 is predicted to be an RNA-binding protein, it is tempting to 491 speculate that its absence disorganized ribonucleic-protein complexes, including their instabilities and 492 degradation. This could partly explain why 90% of DEGs were down-regulated at P18 and included a 493 large proportion of lincRNAs. These down-regulated genes at P18 concerned microtubule-based 494 movement and microtubule-based processes, and cellular components relative to motile cilium, ciliary 495 part, sperm flagellum and axoneme. In addition, DAVID analysis revealed GO terms such as centriole, 496 microtubule and spermatogenesis. All these terms relate to elements of the cytoskeleton that are 497 indispensable for mitotic and/or meiotic divisions, motility and differentiation and are also widely 498 involved in spermiogenesis, as might be expected with this latter GO term because most DEGS are 499 testis-specific. The centriole is a widely conserved organelle in most organisms. A pair of centrioles is 500 located at the heart of the centrosome, and the whole is grouped together as the main microtubule- DEGs between Topaz1 -/and WT mouse testes also revealed a high proportion of deregulated lncRNAs.

553
We showed that three lincRNAs, whose expression was almost abolished as early as P16 in Topaz1-554 deficient mouse testes, were testis-and germ cell-specific. We showed that these genes are expressed

616
Nevertheless, in our 4930463O16Rik-knockout mouse model, several sperm parameters were altered, 617 including reduction in epididymal sperm concentrations (by more than half) and sperm motility. In 618 Tslrn1 knockout mice (testis-specific long non-coding RNA 1) the males were fertile and displayed 619 significantly lower sperm levels (-20%) but no reduction in litter size, or major defects in testis histology 620 or variations in sperm motility (Wichman et al., 2017). In Kif9-mutant male mice, no testes 621 abnormalities were found (Miyata et al., 2020b). They were sub-fertile due to impaired sperm motility: 622 the VSL and VAP velocity parameters were reduced, as in 4930463O16Rik knockout mice. The authors 623 concluded that Kif9 mutant mice were still fertile and this was probably due to variations in the motility 624 of individual spermatozoa; those with good motility could still fertilize oocytes. The same conclusion 625 may apply to 4930463O16Rik -/mice.

626
The suppression of a gene -in this case 4930463O16Rik lincRNA -whose expression is markedly down-627 regulated in the testes of sterile Topaz1 -/mice (FC = 40), has no effect on spermatogenesis. Our data 628 suggest that the expression of 4930463O16Rik is not essential for meiotic division but adds to the 629 terminal differentiation of male germ cells.

631
Various genes, either testis-specific or highly expressed in the testes, exert no effect on reproduction