Control of ciliary transcriptional programs during 4 spermatogenesis by antagonistic transcription factors

22 Existence of cilia in the last eukaryotic common ancestor (LECA) raises a fundamental 23 question in biology: how the transcriptional regulation of ciliogenesis has evolved? One 24 conceptual answer to this question is by an ancient transcription factor regulating ciliary 25 gene expression in both unicellular and multicellular organisms, but examples of such 26 transcription factors in eukaryotes are lacking. Previously, we showed that an ancient 27 transcription factor XAP5 is required for flagellar assembly in Chlamydomonas . Here, we 28 show that XAP5 and XAP5L are two conserved pairs of antagonistic transcription 29 regulators that control ciliary transcriptional programs during spermatogenesis. Male mice 30 lacking either XAP5 or XAP5L display infertility, as a result of meiotic prophase arrest and 31 sperm flagella malformation, respectively. Mechanistically, XAP5 positively regulates the 32 ciliary gene expression by activating the key regulators including FOXJ1 and RFX families 33 during the early stage of spermatogenesis. In contrast, XAP5L negatively regulates the 34 expression of ciliary genes via repressing these ciliary transcription factors during the 35 spermiogenesis stage. Our results provide new insights into the mechanisms by which 36 temporal and spatial transcription regulators are coordinated to control ciliary 37 transcriptional programs during spermatogenesis. 38

The dysfunction of cilia in humans leads to an emerging class of genetic diseases called ciliopathies classified into first-order and second-order ciliopathies (Ishikawa & Marshall, 2011;Lovera & Lüders, 2021;Reiter & Leroux, 2017).First-order ciliopathies are caused by aberrations in ciliary proteins, and second-order ciliopathies occur due to defects in nonciliary proteins, such as transcription factors, that are required for cilia formation and function (Reiter & Leroux, 2017).
Cilia are complex organelles and are dynamically regulated during development and the cell cycle (Eggenschwiler & Anderson, 2007;Kasahara & Inagaki, 2021;Santos & Reiter, 2008).The transcriptional regulation of ciliary genes must be precisely coordinated during ciliogenesis (Choksi, Lauter, Swoboda, & Roy, 2014;Collins, Ventrella, & Mitchell, 2021;Lewis & Stracker, 2021;Thomas et al., 2010).In metazoans, several transcription factors, including FOXJ1 and RFXs, have been shown to be involved in directing the expression of ciliary genes (Stubbs, Oishi, Izpisua Belmonte, & Kintner, 2008;Swoboda, Adler, & Thomas, 2000;Yu, Ng, Habacher, & Roy, 2008).These key ciliary transcription factors show dynamic expression patterns during development and differentiation (Stubbs et al., 2008;Swoboda et al., 2000;Yu et al., 2008).However, how these key transcriptional programs are spatially and temporally controlled by cell type-specific transcription factors and signaling pathways in order to activate specific target ciliary genes and generate diverse cilia is largely unknown.Despite transcriptional regulation of ciliary genes is required for ciliogenesis in both unicellular and multicellular organisms, the underlying mechanism mediating ciliary gene expression may be fundamentally different, as the two major transcription factors, FOXJ1 and RFXs, are absent from many unicellular organisms (Chu, Baillie, & Chen, 2010;Li et al., 2018;Piasecki, Burghoorn, & Swoboda, 2010;Vij et al., 2012).Cilia and ciliary genes have been highly conserved throughout evolution, suggesting that the regulation of ciliary genes could be programmed by yet undiscovered transcriptional mechanisms, which possibly coevolved with multicellularity.Recently, an ancient transcription factor, XAP5, is found to regulate ciliary gene expression in a unicellular organism, Chlamydomonas reinhardtii (Li et al., 2018).XAP5 is highly conserved among different species and is widely expressed across tissues (Martin-Tryon & Harmer, 2008;Mazzarella, Pengue, Yoon, Jones, & Schlessinger, 1997).In vertebrates, XAP5-like (XAP5L) with an intronless open reading frame originated in Therians via retrotransposition of the ancestral XAP5, and it is highly expressed during spermatogenesis (Sedlacek et al., 1999;A. Zhang et al., 2011).However, whether XAP5 and XAP5L play essential roles in cilia development in multicellular organisms remains unknown.
Here, we show that XAP5/XAP5L are antagonistic transcription factors required for ciliary transcriptional programs during spermatogenesis.XAP5 is widely expressed in different tissues, while XAP5L is exclusively present in the testes.To explore the roles of XAP5/XAP5L in ciliogenesis, we generated a XAP5L knockout (KO) mouse line and found that XAP5L KO male mice are infertile due to the malformation of sperm flagella.
Interestingly, male mice with the loss of XAP5 protein in germ cells also exhibited infertile, due to the arrest of spermatogenesis in the meiotic prophase stage.These data suggest that XAP5/XAP5L play crucial roles in sperm flagellar assembly.Mechanistically, XAP5 positively modulates transcriptional rewiring of ciliary genes via activating the key regulators including FOXJ1 and RFXs during the early stage of spermatogenesis.
Conversely, XAP5L negatively regulates the transcriptional program controlling ciliogenesis by repressing these ciliary transcription factors during the spermiogenesis stage.Thus, ciliary transcriptional programs are spatially and temporally controlled by XAP5/XAP5L antagonistic transcription factors during spermatogenesis.

XAP5 and XAP5L are indispensable for male fertility
To explore the potential functions of XAP5/XAP5L during ciliogenesis in vivo, we performed Western blot analysis to investigate the spatiotemporal expression of XAP5/XAP5L in various mouse tissues.We observed that XAP5 protein was widely expressed in diverse tissues and present at all stages of postnatal testes maturation, whereas XAP5L protein was mainly present in the testes and its level increased dramatically from postnatal day 21 (P21) and continued into adulthood (Figures 1A, B).Next, we identified the specific testicular cell types expressing XAP5/XAP5L using published single-cell RNA-seq data (Jung et al., 2019).XAP5 mRNA was found predominantly expressed in spermatogonia, while XAP5L mRNA was observed in pachytene spermatocytes, and remained until elongating spermatids (Figure 1C).Consistent with the mRNA localization patterns, immunofluorescence staining showed that XAP5 protein was mainly detected in the nuclei of spermatogonia, whereas XAP5L protein was present in the nuclei of pachytene spermatocytes and spermatids within the seminiferous tubules, and neither of XAP5 or XAP5L was detected in mature sperm (Figure 1D and Figure S1).
To investigate the physiological function of the uncharacterized XAP5/XAP5L protein, we attempted to generate global KO mice using the CRISPR-Cas9 system.We successfully established a XAP5L KO mouse line (Figures S2A-C) and found that while all XAP5L KO mice survived to adulthood and appeared indistinguishable from the wildtype (WT) mice, XAP5L KO males were sterile (Figures 1E, F).Notably, generating mosaic F0 founders with XAP5 mutant allele proved challenging, and no F1 generation inherited the mutant alleles from the few F0 founders in over a year of breeding experiments, underscoring the critical roles of XAP5 in mouse development.Subsequently, we generated floxed-XAP5 mice and crossed them with Stra8-GFPCre knockin mice (Lin et al., 2017) to produce XAP5 germline-specific KO mice (XAP5 cKO) (Figures S2D-F).Similar in appearance to controls (XAP5 fl/Y ), all XAP5 cKO males were also found to be sterile (Figures 1G, H).
Collectively, these results suggest that XAP5/XAP5L play essential roles in spermatogenesis.

XAP5 and XAP5L function at different stages of spermatogenesis
To uncover the cause of male infertility, we conducted gross and histological analyses of the testes from WT and XAP5L KO mice.There were no differences in the gross appearance of testis or in the weight of mouse body and testis between WT and XAP5L KO mice (Figure 2A).Histological analysis did not reveal any severe abnormity in the seminiferous epithelial cycle of XAP5L KO testes (Figure 2B).We then investigated the motility of sperm collected from cauda epididymis, and found that total motility of sperm was significantly reduced in XAP5L KO mice (Figure 2C and Videos S1, 2).Further examination of the morphology and structure of cauda epididymal sperm in XAP5L KO mice revealed significantly higher rates of abnormal flagella (Figures 2D-F), indicating that XAP5L is crucial for flagellar assembly during spermiogenesis.
We also investigated the spermatogenic defects in the seminiferous tubules and epididymis of XAP5 cKO male mice.Notably, the testes of XAP5 cKO males were significantly smaller compared to WT controls at P16 and older (Figure 3A).Many seminiferous tubules in P16 and older XAP5 cKO males were vacuolated due to the absence of spermatocytes and spermatids, as indicated by the lack of XY body formation and the absence of PNA signal, a marker of spermatids and sperm, in XAP5 cKO germ cells (Figures 3B-D).As a consequence, no sperm were observed in the adult cauda epididymides of XAP5 cKO mice (Figure 3E).Taken together, these data indicate that germ cell development in XAP5 cKO males was arrested in meiotic prophase, and XAP5 is indispensable for meiotic progression during spermatogenesis.

XAP5 and XAP5L function as antagonistic regulators of ciliary transcriptional regulatory networks via regulating transcription factors
To gain insight into the regulatory mechanisms of XAP5L during sperm flagellar assembly, we employed RNA-seq analysis to examine mature sperm collected from the cauda epididymis of WT and XAP5L KO adult mice.Differential expression analysis identified 2093 upregulated and 267 downregulated genes in XAP5L KO sperm (Figure S3A).Gene ontology analysis revealed that the upregulated genes were associated with cilia assembly or function (Figure 4A and Figures S3B, C).When crossed with the previously compiled ciliogenesis-related gene list (Nemajerova et al., 2016), the upregulated genes were observed encompassing diverse structural and functional components of the ciliogenic program (Table S1 and Figure S3D).Notably, several key transcription factor (TF) regulators of ciliogenesis were among the upregulated genes, including FOXJ1 and RFX families (Figures 4B, C and Figure S3E).Moreover, several key regulators of spermatogenesis, particularly spermiogenesis, were identified, including RFX2 (Kistler et al., 2015;Wu et al., 2016), SOX families (Schartl et al., 2018;D. Zhang et al., 2018), TAF7L (Zhou et al., 2013) and TBPL1 (Martianov et al., 2001) (Figure 4B and Figure S3E).
Many core ciliary genes were also upregulated, such as CFAP206 (Beckers et al., 2020) and IFT81 (Qu et al., 2020) which are critical for male fertility (Figure S3E), consistent with the tight coregulation between spermiogenesis and cilia-related genes during spermatogenesis.These findings suggest that the genes identified in our study provide an excellent resource for candidates with novel ciliary or spermatogenesis-related functions, and the newly identified testes-specific protein TULP2 (Oyama et al., 2022;Zheng et al., 2021) was selected for validation.We observed that TULP2 KO male mice were sterile due to malformation of sperm flagella (Figure S4).Overall, our results clearly indicate that XAP5L acts as a central transcriptional repressor of ciliogenesis during mouse spermatogenesis.
Furthermore, we performed RNA-seq analysis using P16 testes samples from WT and XAP5 cKO mice to explore the underlying mechanism of XAP5-mediated germ cell loss.
Differential expression analysis identified 554 upregulated and 1587 downregulated genes in XAP5 cKO sperm (Figure S5A).We observed that the downregulated genes in XAP5 cKO males were involved in cilium assembly or functions (Figure 4D and Figures S5B-D).
Considering the evolutionary conservation of XAP5/XAP5L protein across different species and the expression patterns of XAP5, XAP5L, FOXJ1 and RFX factors during spermatogenesis (Figures 1C and S6), these results suggest that XAP5 and XAP5L have antagonistic effects, and they may function upstream of the transcription factor families, including FOXJ1 and RFX factors, to coordinate the ciliogenesis during spermatogenesis (Figure 4G).

Discussion
Cilia and flagella are ancient organelles present in the last eukaryotic common ancestor (LECA) (Mitchell, 2017).Cilia assembly and maintenance are under strict transcriptional regulation in both unicellular and multicellular organisms (Choksi et al., 2014;Collins et al., 2021;Lewis & Stracker, 2021;Thomas et al., 2010).Despite cilia and flagella having an ancient origin, the evolutionary history of ciliary gene regulation has remained an unsolved problem.The main reason is that the master ciliary transcription factors found in multicellular organisms, including FOXJ1 and RFXs, are absent from unicellular organisms (Choksi et al., 2014;Collins et al., 2021;Lewis & Stracker, 2021;Thomas et al., 2010).Previously, we showed that an ancient transcription factor, XAP5, was required for flagella assembly in the unicellular green algae Chlamydomonas (Li et al., 2018).XAP5 proteins are evolutionarily conserved across diverse organisms (Li et al., 2018;Martin-Tryon & Harmer, 2008), which offers the possibility to investigate the conservation of the role of XAP5 in multicellular organisms.In the present study, we report that XAP5 positively regulates transcriptional network of ciliary genes by activating the key regulators including FOXJ1 and RFXs during spermatogenesis.
The flagellar formation of FOXJ1-null sperm is severely impaired, and the FOXJ1 target genes are required for sperm motility (Beckers et al., 2020;Chen et al., 1998;Weidemann et al., 2016).RFX2 KO mice are completely sterile and RFX2 regulates the expression of ciliary genes during spermiogenesis (Kistler et al., 2015;Wu et al., 2016).Ciliary transcription factors (FOXJ1 and RFX2) are dynamically regulated during spermatogenesis (Jung et al., 2019).Intriguingly, our results show that XAP5 and XAP5L are two conserved pairs of antagonistic transcription factors that regulate these key transcriptional regulators required for spermatogenesis.
By using Western blot analysis, we found that XAP5 was ubiquitously expressed in all tissues examined, whereas XAP5L expression was restricted to the testes (Fig. 1a).
Interestingly, previous research found that XAP5L was widely expressed in normal tissue via using TCGA expression data (Thompson et al., 2021).Moreover, loss of XAP5/XAP5L perturbs transcriptional programs in cancer cells.At present, as shown by the recent data, we cannot rule out the possibility that XAP5L is dynamically expressed in other tissues during development.Although XAP5 and XAP5L are highly expressed in testes and critical for spermatogenesis, they are not present in mature sperm (Extended Data Fig. 1).Thus, XAP5/XAP5L are not needed for sperm maintenance.Given the dynamic expression of XAP5/XAP5L during spermatogenesis, it will be interesting to identify the key signaling factors that regulate the timely and spatial expression of XAP5 and XAP5L.
A role for XAP5 in human brain development is suggested by the association of X-linked intellectual disability (XLID) with rare XAP5 missense variants (Lee et al., 2020).However, how defects in XAP5 lead to XLID syndrome is unknown.Although there is as yet no evidence that XAP5 regulates ciliogenesis in any system besides Chlamydomonas and mice, the function in cilia could help clarify the unexplained role of XAP5 mutations in causing XLID.The possibility that XAP5 could be a ciliary transcription factor in humans would add XLID to the growing list of the second-order ciliopathies.

Mice
C57BL/6J strains (from Vital River Laboratories, Beijing, China) and Stra8-GFPCre knockin mice (from Cyagen Biosciences) were used.Floxed-XAP5 mice, XAP5L and (Applied Biosystems).Expression values were normalized using the ΔΔCt method with Gapdh as the internal control.The primer sequences are indicated in Table S2.

Histological analysis
Testes and epididymides were fixed in Bouin's solution (sigma, HT10132) overnight at 4 °C, embedded in paraffin and sectioned into 5 µm slices.The sections were then subjected to H&E staining using standard procedures and observed under a Zeiss Axio Vert A1 microscope.

Assessment of sperm motility and morphology
Cauda epididymides were collected from male mice and dissected in 37°C pre-warmed human tubal fluid (HTF) culture medium (EasyCheck, M1130), followed by incubation at 37°C for 15 min to release the sperm.To assess sperm motility, the sperm suspension was observed and recorded under a Leica total internal reflection fluorescence microscope.
Motility was defined as any movement of the sperm flagellum during a 10-second observation cycle.After immobilization with 1% paraformaldehyde, sperm morphology was observed by H&E staining using standard procedures.Images were captured using a Zeiss Axio Vert A1 microscope equipped with a ×100 oil immersion objective.Sperm without staining were also used to assess sperm morphology.For ultrastructural observation, sperm samples were fixed in 3% glutaraldehyde in 0.1M sodium phosphate buffer (pH7.4), and post-fixed with 1% (wt/vol) OsO4.After dehydration, the samples were placed in propylene oxide and embedded in a mixture of Epon 812 and Araldite.Ultrathin sections obtained by a Leica EM UC7 ultramicrotome were stained with uranyl acetate and lead citrate, then analyzed using a HT7700 TEM (Hitachi).

Fertility test
Each of the adult male mice was mated with two females for 2 months.All the mice used were aged 6-8 weeks.Vaginal plugs were checked every morning, and the number of newborn pups was counted.
Statistical significance between two groups was calculated using an unpaired, parametric, two-sided student's t test.

Figure 1 .
Figure 1.Loss of XAP5 and XAP5L proteins results in male mice infertility.

Figure 2 .
Figure 2. XAP5L is required for normal sperm formation.

Figure 3 .
Figure 3. Ablation of XAP5 in germ cells results in arrested spermatogenesis and

Figure 4 .
Figure 4. Aberrant ciliogenic and spermatogenic gene expression in XAP5L KO and

Figure S4 .
Figure S4.TULP2 KO male mice were infertile due to malformation of sperm.

Figure
Figure S5.XAP5-mediated regulation of ciliary gene expression during mouse