Deficient spermiogenesis in mice lacking Rlim

The X-linked gene Rlim plays major roles in female mouse development and reproduction, where it is crucial for the maintenance of imprinted X chromosome inactivation in extraembryonic tissues of embryos. However, while females carrying a systemic Rlim knockout (KO) die around implantation, male Rlim KO mice appear healthy and are fertile. Here, we report an important role for Rlim in testis where it is highly expressed in post-meiotic round spermatids as well as in Sertoli cells. Systemic deletion of the Rlim gene results in lower numbers of mature sperm that contains excess cytoplasm, leading to decreased sperm motility and in vitro fertilization rates. Targeting the conditional Rlim cKO specifically to the spermatogenic cell lineage largely recapitulates this phenotype. These results reveal functions of Rlim in male reproduction specifically in round spermatids during spermiogenesis.


Introduction
In testes of adult animals, the differentiation of spermatogonial stem cells during spermatogenesis occurs within seminiferous tubules. Spermiogenesis represents a late stage during spermatogenesis, in which post-meiotic round spermatids differentiate into mature spermatozoa by condensation of the spermatid DNA and formation of the sperm head and tail (O'Donnell, 2014;O'Donnell et al., 2011). Spermatozoa are then released into the lumen of seminiferous tubules, in a process called spermiation, which involves the remodeling and reduction of cytoplasm (Franç a et al., 2016). Even though these processes are crucial for male reproduction, they are poorly understood due to lack of knowledge on spermiation molecular mechanisms. The ubiquitin proteasome system (UPS) plays important roles in male reproduction (Richburg et al., 2014) with a testis-specific version of the proteasome complex (Kniepert and Groettrup, 2014). Indeed, the ubiquitination of proteins in cells of the testis is required for functional spermatogenesis including spermiogenesis, and multiple steps during the progression of spermatogonial stem cells to mature spermatozoa critically depend on the UPS (Richburg et al., 2014). The UPS pathway is critically dependent upon E3 ubiquitin ligases, which provide substrate specificity by selecting target proteins for ubiquitination (Metzger et al., 2014;Pickart, 2001).
The X-linked gene Rlim (also known as Rnf12) encodes a RING H2 type E3 ligase (Metzger et al., 2014;Joazeiro and Weissman, 2000;Bach et al., 1999). While Rlim mRNA is widely expressed in many organs and cell types, RLIM protein is more selectively detected (Bach et al., 1999;Ostendorff et al., 2006). In cells RLIM shuttles between the cytoplasm and nucleus. Nuclear translocation is regulated by phosphorylation, and in many cell types RLIM is primarily detected in the nucleus (Jiao et al., 2013), where it controls not only levels and dynamics of various proteins and protein complexes involved in transcriptional regulation (Bach et al., 1999;Ostendorff et al., 2002;Krämer et al., 2003;Güngö r et al., 2007;Gontan et al., 2012;Johnsen et al., 2009;Her and Chung, 2009;Huang et al., 2011;Wang et al., 2019), but also its own expression via autoubiquitination (Ostendorff et al., 2002). In female mice, Rlim functions as a major epigenetic regulator of nurturing tissues. RLIM promotes the survival of milk-producing alveolar cells in mammary glands of pregnant and lactating females (Jiao et al., 2012). Moreover, RLIM is crucial for imprinted X chromosome inactivation (iXCI) (Shin et al., 2010;Wang et al., 2016;Gontan et al., 2018), the epigenetic silencing of one X chromosome in placental trophoblast cells early during female embryogenesis to achieve X dosage compensation (Payer, 2016). Indeed, due to inhibited placental trophoblast development, the deletion of a maternally inherited Rlim allele results in peri-implantation lethality specifically of females (Shin et al., 2010;Wang et al., 2016). In contrast, males systemically lacking Rlim appear to develop normally, are born at Mendelian ratios and are fertile as adults (Shin et al., 2010).
Here, we report high and dynamic Rlim mRNA and protein expression in the testis of male mice, where expression of RLIM protein is highly detected in round spermatids during spermiogenesis as well as in Sertoli cells. Indeed, we show that Rlim is required for the generation of normal sperm numbers with normal sperm cytoplasmic volume. Even though Sertoli cells are known to regulate cytoplasmic reduction in spermatozoa (O'Donnell, 2014;O'Donnell et al., 2011), our genetic analyses reveal that this activity is mediated by Rlim expressed in the spermatogenic cell lineage. These results assign important functions of Rlim during spermiogenesis.

Rlim expression in testis is highly regulated
To investigate potential functions of Rlim in mice in addition to XCI, we examined mRNA expression in various tissues isolated from adult mice via Northern blots. Rlim mRNA was detected in many tissues with highest levels in testis ( Figure 1A), consistent with published results (Bach et al., 1999;Ostendorff et al., 2000). However, while the Rlim-encoding mRNA in most tissues migrated around 7.5 kb, a variant band at 2.4 kb was detected in testis. Based on published RNA-seq data sets on mouse testes isolated at various post-partum stages (Margolin et al., 2014), mapping of reads to the Rlim locus revealed relatively homogenously distribution over all exons in 6-20 days post-partum (dpp) animals. However, most of the reads in sexually mature (38 dpp) animals mapped in exonic regions upstream of the TGA Stop codon, encompassing the 5' non-coding region and the entire open-reading frame (ORF), while most of the 3' noncoding region was underrepresented ( Figure 1B). Indeed, consistent with the length of the observed variant Rlim mRNA, closer examination the Rlim cDNA sequence revealed a consensus alternative polyadenylation site (Proudfoot, 2011) starting 69 bp downstream of the TGA Stop codon ( Figure 1C). Thus, in mature mouse testes a short, variant Rlim RNA is generated by alternative polyadenylation (Tian and Manley, 2017).
Because expression of Rlim mRNA and protein can be strikingly different (Ostendorff et al., 2006), we examined RLIM protein in testes using an established RLIM antibody (Ostendorff et al., 2006;Ostendorff et al., 2002). Consistent with our mRNA analyses (Figure 1), western blots on protein extracts of testis, brain, and spleen confirmed high expression of full-length RLIM protein in testis ( Figure 2-figure supplement 1A). Using immunohistochemistry (IHC) on testes sections, we detected strong immunoreactivity in specific regions of some seminiferous tubules, representing differentiating spermatogenic cells ( Figure 2A). Moreover, we detected single RLIM-positive cells at the periphery of all tubules. Consistent with published results (Wang et al., 2016), little to no RLIM staining was detected in males carrying a Sox2-Cre (Hayashi et al., 2002) -mediated conditional knockout of the Rlim gene (cKO/Y Sox2-Cre ), as these animals lack RLIM in somatic tissues as well as the germline (Wang et al., 2016;Shin et al., 2014). Lack of RLIM protein in testes of cKO/Y Sox2-Cre animals was confirmed by western blot (Figure 2-figure supplement 1B). Because high RLIM levels appeared to be expressed at specific stages during spermatogenesis (Figure 2A), we performed costaining using an antibody against peanut agglutinin (PNA) that stains acrosomal structures allowing staging of spermatogenic cells within seminiferous tubules (Kotaja et al., 2004;Oakberg, 1956a;Oakberg, 1956b). Around seminiferous stage II / III, RLIM levels are detectable but low in all spermatogenic cells, including post-meiotic step 2/3 spermatids ( Figure 2B). However, indicated by RLIM/PNA co-staining, RLIM levels are high in step 6-8 round spermatids (stages VI-VIII), a midtimepoint in spermiogenesis before spermatids begin to elongate (O'Donnell, 2014;Qian et al., 2014). Thus, RLIM protein levels are low in step 1-5 spermatids, dramatically upregulated in step 6-8 spermatids, and then downregulated in elongating spermatids at step 10 prior to their release (O'Donnell et al., 2011;Figure 2B). We noted that the nuclei of RLIM-positive cells located at the periphery of seminiferous tubules displayed a triangular shape characteristic for Sertoli cells (Figure 2-figure supplement 1C). To identify this RLIM-positive cell type (Figure 2A), we performed IHC, co-staining with antibodies against RLIM and the Sertoli cell marker GATA1 (Yomogida et Figure 1. A short Rlim mRNA variant highly expressed in testis is generated via alternative polyadenylation in mature male mice. (A) A Northern blot containing RNA extracts from various adult mouse tissues (WT for Rlim) was hybridized with an Rlim probe (upper panel) and a probe recognizing bactin as loading control (lower panel). (B) Modified from the UCSC Genome Browser: Cumulative raw reads from RNA-seq datasets of testes RNA isolated from post-natal mice at 6, 12, 16, 20, and 38 days post-partum (dpp) (Margolin et al., 2014) were mapped on the Rlim locus (variable scales). Structure of the Rlim gene is shown below in blue with boxed exon regions. Protein coding regions are indicated in thicker stroke. Blue arrow indicates direction of transcription. ATG start codon, TGA stop codon and site of alternative polyadenylation sequence (red dotted line) is indicated. Note low relative read density 3' of the alternative polyadenylation site specifically in 38 dpp animals. (C) Nucleotide sequence containing an alternative polyadenylation site downstream of the TGA stop codon. Conserved motifs including a T-rich sequence, A/TATAAA, and CA motifs are boxed. The cleavage position is indicated (green arrow). 1994). Positive co-staining identified these cells as Sertoli cells ( Figure 2C). This dynamic and regulated expression of Rlim suggests roles during male reproduction.

Diminished production and functionality of Rlim KO sperm
Next, we investigated functions of the Rlim gene in males using our conditional knockout (cKO) mouse model (Shin et al., 2010). Results on male mice systemically lacking Rlim induced by Cre recombinase transgenes (Shin et al., 2010;Wang et al., 2016;Shin et al., 2014)   embryogenesis and post-natal development. The germline KO/Y animals were generated by crossing heterozygous Rlim KO females (fl m /KO p ) with WT/Y males (Shin et al., 2010;Shin et al., 2014). To explore potential roles of the Rlim gene during male reproduction, we compared testes of 8week-old animals systemically lacking Rlim either by a germline Rlim KO or with a Sox2-Cre-mediated Rlim cKO (cKO/Y Sox2-Cre ) with fl/Y littermate controls. Indeed, in cKO/Y Sox2-Cre and KO/Y mice the size and weight of the testis was significantly decreased when compared to their respective fl/Y littermates, (Figure 3A . Even though these numbers are biased towards motile sperm, these data suggest diminished sperm production. To examine plausible causes for these phenotypes, we analyzed PASstained testis sections. However, we did not detect obvious abnormalities of specific testicular cell types during spermiogenesis in animals lacking Rlim, and sperm release into tubules appeared normal ( Figure 3-figure supplement 1D, not shown). Moreover, in IHC staining testes sections with antibodies against cleaved caspase 3, there were no signs of increased apoptosis in the spermatogenic cells lacking Rlim (not shown). Examining epididymal epithelia, which also express RLIM ( Visualizing tubules that form the inner epididymal layers via de-lipidation (Sylwestrak et al., 2016;Tomer et al., 2014), we did not detect major differences between genotypes (Figure 3-figure supplement 2C). Our results indicate testicular phenotypes in mice lacking Rlim.
As Rlim is a transcriptional regulator and the structural integrity of testes appeared to be generally intact in cKO/Y Sox2-Cre males (Figure 2A), we next examined effects of Rlim on genome wide gene expression. Thus, we performed RNA-seq experiments on RNA isolated from total testis of 8week-old animals, comparing global gene expression in cKO/Y Sox2-Cre and fl/Y male littermates. These experiments including library construction, sequencing, and data processing were performed as previously described (Wang et al., 2017). Consistent with findings that attribute both positive and negative functions of RLIM for gene transcription (Bach et al., 1999;Gontan et al., 2012), statistical analyses revealed 118 down-regulated and 83 up-regulated genes (threshold p<0.05) in cKO/ Y Sox2-Cre animals ( Figure 3C,D). Gene ontology analyses revealed that functions of these genes fell mostly in eight categories with around half of all differentially expressed genes involved in signaling (27.5%) and regulation of metabolism (22.5%), in particular regulatory functions on lipid metabolism ( Figure 3E). Other genes are involved in transcription/chromatin (14%), cell organization (6%), transport (6%) or the UPS (3%), while 9.5% of gene functions occupy multiple other cellular pathways. Around 10.5% of transcripts constituted non-coding (nc) RNAs. No major differences in functions between up-and down-regulated genes were detected in cKO/Y Sox2-Cre testes. Combined, these results suggest functions of Rlim during spermatogenesis.
Next, we investigated the characteristics and functionality of sperm isolated from the Cauda of 8week-old males via swim-out. Indeed, cKO/Y Sox2-Cre sperm displayed elevated rates of morphological abnormalities, including coiled midpieces and head malformations ( Figure 4A,B). However, the sperm capacitation-induced phosphorylation pathways, acrosomal status and induced acrosome reaction in Rlim KO sperm were similar to controls as judged by western blot and PNA staining, respectively (Figure 4-figure supplement 1A-E), indicating that the deletion of Rlim did not affect general signaling. To examine sperm motility, we used CEROS computer-assisted semen analysis (CASA) in swim-out experiments at T0 and after 60 min (T60) under conditions that allow capacitation comparing cKO/Y Sox2-Cre males with fl/Y littermates. The results revealed significant motility deficiencies of cKO/Y Sox2-Cre sperm. Indeed, cKO/Y Sox2-Cre sperm displayed decreased total motility, and out of the total motile sperm the percentages of sperm with progressive motility was lower, while the percentages of slow and weakly motile sperm populations were higher ( Figure 4C-E). Thus, cKO/Y Sox2-Cre sperm is less motile. Next, we tested for possible functional consequences of these deficiencies during in vitro fertilization (IVF) (Sharma et al., 2016), using oocytes originating from WT females and sperm isolated from either fl/Y or cKO/Y Sox2-Cre littermates. In these experiments, the numbers of oocytes and sperm cells were adjusted to 100-150 and 100,000, respectively, to achieve fertilization rates of around 80% for the control samples as judged by the number of embryos reaching cleavage stage 24 hr after adding sperm ( Figure 4F,G). The development of IVF embryos was monitored up to (96 hr) at which point blastocyst stage was reached by the majority of embryos generated by control sperm. Indeed, sperm isolated from fl/Y control animals yielded in metabolism transcription signaling other cell organization transport UPS ncRNA up-regulated down-regulated   84.3% cleavage stage and 77% blastocyst stage embryos, compared to 67.0% and 46.8% for cKO/ Y Sox2-Cre sperm, respectively ( Figure 4F,G). Moreover, the rate of blastocysts per cleavage staged embryos was 92% of for fl/Y sperm and only 70% for cKO/Y Sox2-Cre (Figure 4-figure supplement 1F), suggesting that lack of Rlim negatively affects embryonic development under in vitro conditions. Thus, sperm isolated from cKO/Y Sox2-Cre animals yielded in significantly less embryos reaching the appropriate developmental stage when compared to fl/Y littermate controls, both for cleavage and blastocyst stages ( Figure 4F,G). Combined, these data indicate that deletion of Rlim in male mice results in reduced sperm functionality.

Increased size of cytoplasmic droplets in Rlim cKO sperm
Because of the decreased motility of Rlim cKO sperm ( Figure 4C-E), we examined the energetic status including amino acids, glycolysis, TCA cycle, pentose phosphate pathway, and nucleotide biosynthesis via metabolomic profiling of polar metabolites. We isolated 40-45 Mio cauda swim-out sperm each of 3 fl/Y and 5 cKO/Y Sox2-Cre animals, and polar metabolites were measured using liquid chromatography coupled with mass spectrometry (LC-MS). Results showed that, unexpectedly, 66 metabolites were significantly more abundant in the cKO/Y Sox2-Cre sperm, in particular levels of many amino acids, in contrast to only five metabolites that displayed lower levels ( Figure 5A). Analyses of these diverse metabolites did not yield in the identification of a specific energy production pathway but rather revealed that metabolite intermediates of many cellular pathways are increased in the Rlim cKO sperm, suggesting a more general problem. Thus, we included electron microscopy (EM) in our analyses. Interrogating sperm via scanning EM (SEM) revealed significantly more Rlim cKO/Y Sox2-Cre sperm displaying cytoplasmic droplets in the midpiece, and these droplets were also increased in size as measured using ImageJ software ( Figure 5B-D). Because Cauda sperm has matured a considerable time in the epididymis that also expresses Rlim (Figure 3-figure supplement 2A), to distinguish testicular versus epididymal functions of Rlim in droplet formation we extended these studies to testicular sperm. Indeed, isolating testicular sperm from 30 fl/Y and cKO/ Y Sox2-Cre animals each, the numbers of isolated sperm per Rlim cKO testis were lower but this was no longer significant when compared to controls ( Figure 6A). We noted that the midpieces from cKO sperm were highly vulnerable to rupturing, while the prevalence of coiled midpieces appeared similar between cKO and control sperm ( Figure 6B). Moreover, we detected a low number of sperm that exhibited duplicated axonemes specifically in Rlim cKO/Y Sox2-Cre but not in fl/Y sperm ( Figure 6B; Figure 6-figure supplement 1A). Again, the sizes of cytoplasmic droplets in testicular sperm were significantly increased ( Figure 6C,D). Transmission EM (TEM) on testes sections confirmed the occurrence of duplicated axonemes in cKO/Y Sox2-Cre sperm as well as head malformations, while the overall structural integrity of the sperm tail appeared normal ( Figure 6-figure supplement 1A-D), and no signs of decreased chromatin packaging in sperm heads was detected as judged by sperm head density. Interestingly, the cytoplasmic pockets in the Rlim cKO sperm heads appeared more pronounced in sperm of the epididymal Caput region ( Figure 6E,F). These data reveal excessive cytoplasm in sperm heads and midpieces in males lacking Rlim and are consistent with testicular as opposed to epididymal functions of Rlim.

Functions of Rlim during spermiogenesis specifically in the spermatogenic cell lineage
Next, we addressed the question as to the cell type of Rlim action during spermiogenesis as RLIM protein is highly detectable both in the spermatogenic cell lineage specifically in round spermatids and in Sertoli cells (Figure 2). Because Sertoli cells play major roles in the regulation of spermiogenesis/spermiation (O'Donnell, 2014;O'Donnell et al., 2011;Franç a et al., 2016) and cell numbers are correlated with sperm production capacity (Griswold, 1995), we compared number of cells positive for Sertoli cell marker GATA4, as in adult mice GATA4 expression does not vary with the cycle of the seminiferous epithelium (Yomogida et al., 1994). Counting GATA4-positive cells within seminiferous tubules, IHC revealed similar numbers of Sertoli cells in testes with or without Rlim (Figure 7-figure supplement 1A,B), indicating that Rlim is not required for Sertoli cell development and differentiation. Moreover, analyzing specific Sertoli cell structures involved in spermiation in testis sections via TEM, we did not detect major structural deficiencies in cells lacking Rlim including the formation of apical ectoplasmic specialization (ES) or the apical tubulobulbar complex (TBC), and no obvious signs of defective cytoplasmic reduction (Figure 7-figure supplement 1C,D).
In order to genetically elucidate the cell identity of Rlim function, we targeted the Rlim cKO via Ngn3-Cre (also known as Neurog3-Cre) (Schonhoff et al., 2004) to the spermatogenic cell lineage (Yoshida et al., 2004) and via Sf1-Cre (Dhillon et al., 2006) to Sertoli cells (Kim et al., 2007). IHC on testis sections using RLIM antibodies confirmed the correct, specific and penetrant targeting of both Cre drivers ( Figure 7A). Because of a mixed background of these Cre-driver mouse lines, cKO males were compared to their respective littermate controls. Indeed, Rlim cKO/Y Ngn3-Cre mice displayed decreased testis weights and numbers of mature sperm isolated in caudal swim-out experiments ( Figure 7B,C, respectively). In contrast, no significant effects on testes weights and sperm numbers were measured in cKO/Y Sf1-Cre animals. Consistent with these findings, SEM analyses of Caudal swim-out sperm revealed significantly increased numbers of sperm containing cytoplasmic   Figure 7 continued on next page droplets in cKO/Y Ngn3-Cre animals when compared with cKO/Y Sf1-Cre animals and their respective littermate controls ( Figure 7D). Moreover, focusing on sperm with cytoplasmic droplets, the droplet sizes in cKO/Y Ngn3-Cre sperm were increased ( Figure 7E,F), similar to those of cKO/Y Sox2-Cre sperm ( Figure 5C,D). IVF experiments using swim-out sperm yielded in 76% cleavage stage and 58% blastocyst stage embryos for control, compared to 62% and 33% for cKO/Y Ngn3-Cre sperm, respectively ( Figure 7G), confirming the functional deficiency of sperm. Thus, the phenotypes in males systemically lacking Rlim are largely recapitulated by cKO/Y Ngn3-Cre but not cKO/Y Sf1-Cre mice, demonstrating functions of Rlim in the spermatogenic cell lineage. Our combined results provide strong evidence that the upregulated expression of Rlim in round spermatids plays important functions for the reduction of cytoplasmic volume in sperm.

Discussion
Our results reveal robust expression of Rlim in male reproductive organs, particularly in testis, where the expression pattern was highly dynamic both at the mRNA and protein levels (Figures 1 and  2). The appearance of a variant mRNA in males coincides with sexual maturation, suggesting that this form is predominantly expressed in differentiating round spermatids, which express high levels of RLIM protein (Figure 2A), and, because Rlim acts in the spermatogenic cell lineage (Figure 7), also coincides with the exertion of its in vivo function in males. Indeed, alternative mRNAs display different time of synthesis, mRNA stability and/or translational efficiency (Tian and Manley, 2017). At the protein level, we find RLIM expression in spermatogenic cells is post-meiotically upregulated from low levels in step three spermatids to high levels peaking in round spermatids at step 6-8 and then downregulated again in elongating step nine spermatids and thereafter ( Figure 2). Thus, Rlim joins many X-linked genes that are reactivated after meiotic sex chromosome inactivation (MSCI) (Ernst et al., 2019) and it is tempting to speculate that this reactivation is connected with the occurrence of the alternative Rlim mRNA. Moreover, the downregulation of RLIM in elongating spermatocytes coincides with a change in transcriptional and chromatin dynamics at this stage (Ernst et al., 2019). Even though Rlim is X-linked, the finding that the RLIM protein is detected in most/all round spermatids, including those that presumably harbor a Y chromosome, is explained by the fact that cytoplasmic bridges exist between spermatids and that RLIM efficiently shuttles between nuclei in heterokaryon cells (Jiao et al., 2013). Our RNA-seq analyses indicate that many genes are affected in testes lacking Rlim influencing various cellular functions including signaling and metabolism ( Figure 3). As Rlim regulates transcriptional factors (Bach et al., 1999;Ostendorff et al., 2002;Krämer et al., 2003;Güngö r et al., 2007;Gontan et al., 2012;Johnsen et al., 2009;Her and Chung, 2009;Huang et al., 2011;Wang et al., 2019), it is thus likely that many of the differentially expressed genes ( Figure 3D,E) might be affected indirectly. Because the vast majority of cells in the testis reflect spermatogenic cells, where Rlim exerts its functions (Figure 7), many differentially expressed genes reflect those expressed in this cell lineage and therefore may collectively contribute to the observed Rlim KO phenotypes in testes. However, we cannot exclude the possibility that subtle and hence undetected changes in testicular cell types in mice lacking Rlim may contribute to the observed differences in gene expression as well as testes weight (Figure 3).
Sperm produced by males lacking Rlim is dysfunctional with decreased motility and increased cytoplasm and head abnormalities. Because excess cytoplasm affects sperm motility, morphology including head morphology as well as fertilization potential (Cooper, 2011;Rengan et al., 2012), and the cytoplasmic volume is regulated during spermiogenesis just after RLIM protein is highly detected in spermatids, it is likely that the increased cytoplasmic volume is responsible for much of the defects detected in sperm lacking Rlim. Moreover, as the midpiece cytoplasm is particularly important for sperm osmoregulation (Cooper, 2011;Rengan et al., 2012), an increased size of cytoplasmic droplets is predicted to render sperm more vulnerable to osmotic challenges, which may ultimately lead to midpiece rupturing ( Figure 6). Thus, while the finding of increased metabolite content in cKO sperm ( Figure 5A) suggests defective cytoplasmic reduction during spermiation even though major defects were not observed (Figure 7-figure supplement 1), it is likely that the some of the increase in cytoplasmic size might have occurred in released sperm after spermiation. In this context, an increased sperm cytoplasm has been associated with higher activities of specific enzymes of the energy pathway including G6PDH (Aitken et al., 1994;Yuan et al., 2013), which diverts glucose metabolism away from glycolysis toward the pentose pathway. It is thus tempting to speculate that increased G6PDH activity might be partially responsible for the decreased Acetyl-CoA levels measured in sperm lacking Rlim ( Figure 5A). Our combined data suggests that inefficient cytoplasmic reduction during spermiogenesis/spermiation renders Rlim KO sperm vulnerable to adequately adjust to the changing environment that occurs during transit through the epididymis (Gervasi and Visconti, 2017;Sullivan and Saez, 2013) leading to functional defects.
Concerning the cell type of Rlim action, RLIM protein is detected in spermatogenic cells specifically in round spermatids, in Sertoli cells and in epididymal epithelial cells. Our combined data reveals that lack of RLIM specifically in round spermatids is responsible for much of the observed sperm phenotype. This is demonstrated by targeting the Rlim cKO via Ngn3-Cre in the spermatogenic cell lineage, resulting in defective spermiogenesis, and similar sperm phenotypes when compared to the systemic Rlim cKO (Figures 3 and 7). In contrast, targeting the Rlim cKO via Sf1-Cre to Sertoli cells failed to induce a testis/sperm phenotype and providing strong evidence that Rlim in Sertoli cells is not involved in the regulation of spermiogenesis provided by this cell type. Thus, while we cannot exclude minor functions of Rlim in Sertoli cells and possibly also epididymal epithelial cells, our results provide strong evidence that Rlim in round spermatids is required for normal spermiogenesis. Therefore, Rlim adds to a very limited number of genes in the spermatogenic cell lineage that regulate the cytoplasmic reduction/droplet size (Zheng et al., 2007;Mikolcevic et al., 2012).
Considering early lethality of female mouse embryos with a maternally transmitted Rlim mutation (Shin et al., 2010), mathematical modeling of this exclusively female phenotype indicates that a deleterious mutation in the Rlim gene would shift selective evolutionary pressure entirely on females leading to a gender bias toward males in a mouse population over time (Jiao et al., 2012). Because gender biases in mouse populations represent an unfavorable strategy for reproduction (Hamilton, 1967), it is likely that the observed functions of Rlim during male reproduction will contribute to counter-act gender biases induced by its female function.
In summary, our data provide first evidence that in addition to crucial epigenetic functions in female embryogenesis and reproduction, the E3 ubiquitin ligase RLIM also occupies important roles during the reproduction of male mice. These results have major implications for epigenetic regulation and emphasize the importance of the UPS in male reproduction.

De-lipidation of epididymis
De-lipidation of epididymis was performed based on previously described protocols (Sylwestrak et al., 2016;Tomer et al., 2014). Briefly, isolated organs were fixed with 4% paraformaldehyde (PFA) in PBS for 32 hr at room temperature (RT), then rinsed with PBS for three times of at least 2 hr. Tissues were kept at 4˚C in PBS with 0.02% sodium azide until the time of tissue processing. In order to visualize tubules that form the inner layers of the epididymis, a de-lipidation step was performed. De-lipidation was done passively by incubating the organ with 4% SDS/PBS at RT in an orbital shaker for 2 weeks. The 4% SDS/PBS solution was changed every other day. At the end of the second week, the organ was rinsed with PBS for three times of least 4 hr and finally placed on a refractive index matching solution (RIMS: 0.17M iodixanol; 0.4M diatrizoic acid, 1M n-methyl-d-glucamine, 0.01% sodium azide), 24 hr prior to imaging.

Collection of sperm
Epididymal Cauda sperm collected via swim-out and testicular sperm which were analyzed in this study was collected from 8 weeks-old fl/Y or cKO/Y mice. Briefly, Cauda epididymides were dissected and placed in 1 ml of modified Krebs-Ringer medium (m-TYH; 100 mM NaCl, 4.7 mM KCl, 1.2 mM KH 2 PO 4 , 1.2 mM MgSO 4 , 5.5 mM Glucose, 0.8 mM Pyruvic Acid, 1.7 mM Calcium Chloride, 20 mM HEPES). Sperm was allowed to swim-out for 10 min at 37˚C and then the epididymides were removed. Concentration of all sperm was calculated using a Neubauer hemocytometer. For mature testicular spermatozoa isolation, testes from one 8-week-old mouse were minced in a 35 mm Petri dish containing 1 ml 150 mM NaCl. Finely minced tissue slurry was then transferred to a 15 ml conical tube and set aside for 3-5 min to allow tissue pieces to settle down. Next, the cell suspension was loaded onto 10.5 ml of 52% isotonic percoll (Sigma). The tubes were then centrifuged at 15000 x g for 10 min at 10˚C. The pellet was resuspended in 10 ml of 150 mM NaCl and spun at 900 x g at 4˚C for 10 min followed by three washes with 150 mM NaCl at 4000 x g at 4˚C for 5 min.

Sperm analyses
Concerning the assessment of sperm morphology, after swim-out, sperm suspensions (50 ml) were fixed with paraformaldehyde 4% (w/v; EMS, Hatfield, MA) in phosphate buffered saline (PBS) for 10 min at room temperature. Then samples were centrifuged at 800 x g for 5 min, washed with PBS, and let air-dry on poly-L-lysine-coated glass slides. Samples were mounted using VectaShield mounting media (H-1000, Vector Laboratories, Burlingame, CA) and differential interference contrast (DIC) images were taken using a 60X objective (Nikon, PlanApo, NA 1.49) in an inverted microscope (Nikon Eclipse TE300). Images were analyzed using the freeware ImageJ v1.52e (http://imagej.nih. gov/ij/download.html), and sperm morphology was classified into the following categories: normal, coiled midpiece, head malformation, flipped head, and residual cytoplasmic droplet. Results are shown as percentages, at least 200 sperm per sample were counted in single-blinded experiments.
Concerning the analysis of sperm acrosomal status, sperm from the swim-out suspension were loaded into glass slides and let air-dry for 15 min. After that samples were fixed, and the acrosomes were stained as described below in the sperm acrosome reaction section. Fluorescence and phase contrast images were taken in a Nikon Eclipse TE300 fluorescence microscope using a 40x objective (Nikon, Phase 2 DL, NA 0.55). Images were analyzed using the freeware ImageJ v1.52e, and at least 200 sperm per sample were counted in single-blinded experiments. Results are shown as percentage of acrosome intact sperm. Concerning sperm capacitation and analysis of swim-out cauda sperm motility, sperm were incubated at 37˚C for 60 min in m-TYH (Non-Cap) or in m-TYH supplemented with 15 mM NaHCO 3 and 5 mg/ml BSA (Cap). Sperm motility was evaluated in the swim-out (T = 0) and after 60 min of incubation in capacitating conditions (T = 60). Briefly, sperm suspensions (30 ml) were loaded into prewarmed chamber slides (Leja slides, Spectrum Technologies, Healdsburg, CA) and placed on a warmed microscope stage at 37˚C. Sperm motility was examined using the CEROS computerassisted semen analysis (CASA) system (Hamilton Thorne Research, Beverly, MA). Acquisition parameters were set as follows: frames acquired: 90; frame rate: 60 Hz; minimum cell size: four pixels; static head size: 0.13-2.43; static head intensity: 0.10-1.52; and static head elongation: 5-100. At least five microscopy fields corresponding to a minimum of 200 sperm were analyzed in each experiment. Data were analyzed using the CASAnova software (Goodson et al., 2011).
Concerning SDS-PAGE and western blotting, swim-out cauda sperm samples were centrifuged at 12,000 x g for 2 min, washed in 1 ml of PBS, and then centrifuged at 12,100 x g for 3 min. Sperm proteins were extracted by resuspending the remaining pellets in Laemmli, 1970 sample buffer, boiled for 5 min and centrifuged once more at 12,100 x g for 5 min. Protein extracts (supernatant) were then supplemented with b-mercaptoethanol 5% (v/v), and boiled again for 4 min. Protein extracts equivalent to 2.5 Â 10 5 sperm/lane were subjected to SDS-PAGE, and electro-transferred to PVDF membranes (Bio-Rad, Waltham, MA). PVDF membranes were blocked with 5% (w/v) fatfree milk in tris buffered saline containing 0.1% (v/v) Tween 20 (T-TBS) and immunoblotted with anti-pPKAs antibody (clone 100G7E, 1:10,000) overnight at 4˚C to detect phosphorylated PKA substrates. Then, membranes were incubated with HRP-conjugated anti-rabbit secondary antibody diluted in T-TBS (1:10,000) for 60 min at room temperature. Detection was done with an enhanced chemiluminescence ECL plus kit (GE Healthcare) as per manufacturer instructions. After developing of pPKAs, membranes were stripped at 55˚C for 20 min in 2% (w/v) SDS, 0.74% (v/v) b-mercaptoethanol, 62.5 mM Tris (pH 6.5), blocked with fish gelatin 20% (v/v; Sigma cat # G7765, St. Louis, MO) in T-TBS for 60 min at room temperature and re-blotted with anti-PY antibody (clone 4G10, 1:10,000) to detect proteins phosphorylated in tyrosine residues. Membranes were then incubated with HRP-conjugated anti-mouse secondary antibody diluted in T-TBS (1:10,000) for 60 min at room temperature. Detection was done with an enhanced chemiluminescence ECL plus kit (GE Healthcare) as per manufacturer instructions.
Concerning swim-out cauda sperm acrosome reaction, after 60 min of capacitation in m-TYH Cap medium, sperm samples were incubated with progesterone (10 mM) or with DMSO (vehicle) in m-TYH Cap at 37˚C for 30 min. Then, sperm were loaded into glass slides and let air-dry for 15 min. Sperm were fixed by incubation with paraformaldehyde 4% (w/v) in PBS at room temperature for 15 min, washed three times (5 min each) with PBS and permeabilized with 0.1% (v/v) Triton X-100 for 3 min. After permeabilization, samples were washed three times with PBS, and then incubated with Alexa Fluor 488-conjugated lectin peanut agglutinin (PNA) in PBS at room temperature for 30 min. Before mounting with Vectashield (Vector Laboratories, Burlingame, CA), samples were washed three times with PBS for 5 min each time. Epifluorescence images were taken in a Nikon Eclipse TE300 fluorescence microscope using a 40x objective (Nikon). Phase contrast images were taken in parallel. Images were analyzed using the freeware ImageJ v1.52e, and at least 200 sperm per sample were counted in single-blinded experiments. Results are shown as percentage of acrosome reacted sperm.

In vitro fertilization (IVF)
IVF experiments were performed essentially as described (Sharma et al., 2016) with some modifications. Briefly, female C57BL/6J mice (age 6-8 weeks) were superovulated by intraperitoneal injection of pregnant mare's serum gonadotropin (PMSG; 5 U; Calbiochem) followed by human chorionic gonadotropin (hCG; 5 U; Sigma-Aldrich) 48 hr later (Yamashita et al., 2008). Metaphase II-arrested oocytes tightly packed with cumulus cells were collected from the oviductal ampulla 14 hr after hCG injection and placed in a 100 ml drop of human tubal fluid (HTF; Millipore) medium covered with mineral oil. Fresh cauda epididymal sperm of fl/Y and cKO/Y littermates (>6 animals per genotype both for Sox2-Cre and Ngn3-Cre, aged 2-4 mo) were swum-up in 1 ml HTF medium for 10 min and capacitated by incubation for another 30 min at 37˚C under 5% CO 2 . An aliquot (1.0 Â 10 5 cells) of the capacitated sperm suspension was added to 100 ml drop of HTF medium containing the oocytes. After incubation at 37˚C under 5% CO 2 for 4 hr, the presumed zygotes were washed with KSOM medium to remove cumulus cells, sperm and debris, and then incubated in a 50 ml drop of KSOM medium. In vitro fertilized embryos were analyzed at cleavage (24 hr) and blastocyst stages (96 hr).

Metabolomic profiling
Polar metabolite measurements were performed by the Metabolite Profiling Core Facility at the Whitehead Institute for Biomedical Research (Cambridge, MA). For LC-MS analyses, 45 Mio epididymal cauda swim-out sperm were collected each from 5 cKO/Y Sox2-Cre and three fl/y animals in extraction mix containing 80% methanol and isotopically labeled amino acids. Peak area ratios of each metabolite were determined and log 2 ratio relative to the mean of control (fl/Y) sperm determined.

Electron microscopy
SEM was performed on caudal epididymis sperm and sperm isolated from testes. Caudal sperm was let to dissociate in HTF medium and then fixed in 2.5% glutaraldehyde, 2% paraformaldehyde in 0.1M Na Cacodylate buffer. The fixed sperm was then placed on poly Lysine covered cover slips and left to adhere for 10 min. The samples were rinsed three times in the same fixation buffer, post fixed in aqueous 1% (w/v) OsO4 for 1 hr at RT, dehydrated through a graded ethanol series to ethanol 100% (x3), and then they were critically point dried. The cover slips were then mounted on double sided carbon tape onto aluminum SEM stubs and grounded with colloidal silver paint, sputter coated with 12 nm of gold-palladium and were imaged using secondary electron (SEI) mode with a FEI Quanta 200 MKII FEG SEM.
TEM was performed on ultrathin sections on Caput epididymides and testes. Tissues were dissected and immediately immersed in 2.5% glutaraldehyde in 0.1 M Na Cacodylate buffer, pH 7.2 for 60 min at RT. The samples were rinsed three times in the same fixation buffer and post-fixed with 1% osmium tetroxide for 1 hr at room temperature. Samples were then washed three times with ddH 2 O for 10 min, and in block stained with a 1% Uranyl Acetate aqueous solution (w/v) at 6˚C overnight. After three rinses in ddH2O the samples were dehydrated through a graded ethanol series of 20% increments, before two changes in 100% ethanol. Samples were then infiltrated first with two changes of 100% Propylene Oxide and then with a 50%/50% propylene oxide / SPI-Pon 812 resin mixture. The following day 5 changes of fresh 100% SPI-Pon 812 resin were done before the samples were polymerized at 68˚C in flat embedding molds. The samples were then reoriented, and thin sections (approximately 70 nm) were placed on copper support grids and contrasted with Lead citrate and Uranyl acetate. Sections were examined using the a CM10 TEM with 100Kv accelerating voltage, and images were captured using a Gatan TEM CCD camera.