Huwe1 was highly expressed in Sertoli cells and inactivation of Huwe1 caused the defects in spermatogenesis

In order to investigate the functions of Huwe1 in testis development, the expression of HUWE1 in testis was examined by immunofluorescence (IF) staining. As shown in Fig. 1, in addition to germ cells, HUWE1 was also highly expressed in SOX9-positive Sertoli cells (Fig. 1A-C, white arrows). This finding suggested that HUWE1 may play a role in Sertoli cell development. To further examine the function of HUWE1 in Sertoli cell development, we specifically deleted Huwe1 in Sertoli cells by crossing Huwe1+/flox mice with Amh-Cre mice. Quantitative PCR and Western blot analyses revealed a dramatic reduction in both the mRNA and protein levels of Huwe1 in testes of Huwe1 CKO mice (Huwe1flox/Y; Amh-Cre) (Supplementary Fig. 1A-C).

Fig. 1

Inactivation of Huwe1 in Sertoli cells caused germ cell loss and male infertility. HUWE1 was highly expressed in SOX9-positive Sertoli cells (A-C, white arrows). DAPI (blue) was used to stain the nuclei. No developmental anomalies were observed in Huwe1 CKO mice (Huwe1flox/Y; Amh-Cre) (D). The size of testes from Huwe1 CKO mice was significant reduced (n = 3) (E). The results of H&E staining showed well organized seminiferous tubules in control males (G), the seminiferous tubules in Huwe1 CKO mice was atrophied and the germ cells were disorganized (H). The epididymis in control mice were filled with mature sperm (I), whereas no mature sperm were observed in epididymis of Huwe1 CKO mice (J). SOX9-positive Sertoli cells were located at periphery region of seminiferous tubules in control males (K, indicated by black arrows), whereas the SOX9-positive Sertoli cells were scattered in the seminiferous tubules of Huwe1 CKO mice (L, black arrowheads). MVH-positive germ cells were well organized in the seminiferous tubules in control males (M, white arrows), whereas germ cells were disorganized in the seminiferous tubules of Huwe1 CKO mice (N, white arrowheads) and completely absent in some tubules (N, white asterisk). The data are presented as the mean ± SEM. ***, P < 0.001

Adult Huwe1 CKO male mice were grossly normal with no apparent developmental abnormalities. However, the size of testes from Huwe1 CKO mice was significantly reduced compared to that in control males (Fig. 1D-F). Histological analysis revealed well-organized seminiferous tubules in control mice (Fig. 1G). In contrast, seminiferous tubules in Huwe1 CKO mice were atrophic, and the number of germ cells was dramatically reduced (Fig. 1H). A large number of mature sperm were detected in the epididymis of control mice (Fig. 1I). By contrast, no mature sperm were observed in epididymis of Huwe1 CKO mice (Fig. 1J). The results of immunohistochemical (IHC) staining showed that SOX9-positive Sertoli cells were located at the peripheral region of seminiferous tubules (Fig. 1K) in control mice. Conversely, Sertoli cells in Huwe1 CKO mice were disorganized, and numerous SOX9-positive signal were detected at the center of seminiferous tubules (Fig. 1L). MVH-positive germ cells in the control mice were orderly arranged (Fig. 1M), whereas the number of germ cells in Huwe1 CKO mice was significantly reduced, and the location of germ cells was also disorganized (Fig. 1N).

Inactivation of Huwe1 caused polarity loss in Sertoli cells

We further examined the development of testes in Huwe1 CKO mice at early developmental stages. At 1 week of age, SOX9-positive Sertoli cells were localized at the center of seminiferous tubules in both control and Huwe1 CKO mice (Fig. 2A, indicated by black arrows). In control testes, Sertoli cells were well organized, whereas in Huwe1 CKO mice Sertoli cells appeared disordered (Fig. 2D, black arrowheads). At 2 weeks, most Sertoli cells in control mice had migrated to the periphery region of seminiferous tubules (Fig. 2B, black arrows), while in Huwe1 CKO mice, the nuclei of Sertoli cells remained centrally located (Fig. 2E, black arrowheads). By 3 weeks, the nuclei of Sertoli cells in control mice were positioned close to the basement membrane of seminiferous tubules (Fig. 2C, indicated by black arrows). By contrast, the nuclei of Sertoli cells in Huwe1 CKO mice were not in contact with the basement membrane, and most of them appeared disorganized (Fig. 2F, black arrowheads). Western blot analysis showed that the protein level of ZO1 was significantly decreased in Huwe1-deficient Sertoli cells. In contrast, the protein levels of CDH1, CTNNB1, VIMENTIN, SNAI1&2 were significantly increased in Huwe1-deficient Sertoli cells (Fig. 2G, H). Immunofluorescence staining further revealed aberrant localization of adhesion junction proteins and polarity-related protein in Huwe1-deficient Sertoli cells (Supplementary Fig. 2A-H). These results indicated that the cell polarity was disrupted in Huwe1-deficient Sertoli cells.

Fig. 2

Inactivation of Huwe1 caused polarity loss in Sertoli cells. The Sertoli cells in control and Huwe1 CKO mice were labeled with SOX9. SOX9-positive Sertoli cells were localized at the center of seminiferous tubules in both control and Huwe1 CKO mice at 1 week. The Sertoli cells in control testes were well organized (A, indicated by black arrow), whereas the Sertoli cells in Huwe1 CKO mice were disordered (D, black arrowhead). Most of Sertoli cells were moved to the periphery region of seminiferous tubules in control mice (B, black arrows) at 2 weeks. The Sertoli cells were still localized at the center of seminiferous tubules in Huwe1 CKO mice (E, black arrowhead). The nuclei of Sertoli cells were closed to the basement membrane of seminiferous tubules in control mice (C, indicated by black arrow) at 3 weeks. The nuclei of Sertoli cells in Huwe1 CKO mice were not contacted with basement membrane and most of them were disorganized (F, black arrowheads). The results of Western blot analysis showed the differential expression of adhesion junction molecules (e.g., ZO1, CTNNB1), cytoskeletal proteins (e.g., VIM), and epithelial-mesenchymal transition (EMT) trans-factors (e.g., TWIST, SNAI1&2) in Huwe1-deficient Sertoli cells at 2 weeks (G, H). In H, the data are presented as the mean ± SEM. *, P < 0.05. **, P < 0.01. ***, P < 0.001

The protein level of WT1 was significantly elevated in Huwe1-deficient Sertoli cells

Given that HUWE1 functions as an E3 ligase, quantitative mass spectrometry was conducted using whole-cell lysates of primary Sertoli cells to screen the potential downstream targeting proteins. A total of 5,001 proteins were detected by mass spectrometry analysis, of which 386 proteins upregulated and 219 proteins downregulated in Huwe1-deficient Sertoli cells. The differentially expressed proteins included adhesion molecules and cytoskeletal proteins. Interestingly, we found that the protein level of WT1 was increased in Huwe1-deficient Sertoli cells (Fig. 3A). To further confirm this finding, the expression of Wt1 and other Sertoli cell-specific genes were examined by Western blot analysis. As shown in Fig. 3B, the protein level of WT1 was significantly increased in Huwe1-deficient Sertoli cells, whereas the expression of DMRT1 and SOX9 remained unchanged (Fig. 3C). Surprisingly, the mRNA level of Wt1 gene was significantly reduced in Huwe1-deficient Sertoli cells (Supplementary Fig. 5B).

Fig. 3

The expression of Wt1 was increased in Huwe1-deficient Sertoli cells. Quantitative mass spectrometry of whole-cell lysates from control and Huwe1-deficient Sertoli cells. List of the most differentially expressed proteins in Huwe1-deficient Sertoli cells (A). (B-C), Western blot analysis of WT1 and other Sertoli cell specific genes. The protein level of WT1 was significantly increased in Huwe1-deficient Sertoli cells. The data are presented as the mean ± SEM. ***, P < 0.001

Overexpression of Wt1 caused disorganization of Sertoli cells and the defect of spermatogenesis

Previous studies have demonstrated that Wt1 gene is specifically expressed in Sertoli cells and plays critical roles in lineage specification and functional maintenance of Sertoli cells [23, 24]. Inactivation of Wt1 causes aberrant testis development and male infertility [25, 26]. To investigate whether the defects in Sertoli cell development observed in Huwe1 CKO mice are caused by increased WT1 protein level, we overexpressed Wt1 -KTS isoform in Sertoli cells by crossing Wt1+/ctg mice [27] with Amh-Cre mice. Quantitative PCR and Western blot analyses confirmed that both the mRNA and protein levels of Wt1 were significantly increased in Sertoli cells of Wt1-overexpressed (OE) mice (Wt1+/ctg; Amh-Cre) (Supplementary Fig. 4A-C).

No obvious developmental abnormalities were observed in adult Wt1 OE mice. However, the size of testes in Wt1 OE mice was smaller compared to those of control males (Fig. 4A-C). The epididymis in control mice were filled with mature sperm (Fig. 4D), whereas no mature sperm were observed in the epididymis of Wt1 OE mice (Fig. 4E). SOX9-positive Sertoli cells were located at the periphery region of seminiferous tubules in control males (Fig. 4H, black arrows), while a small number of SOX9-positive Sertoli cells were disorganized in the seminiferous tubules of Wt1 OE mice (Fig. 4K, black arrowheads). PNA-positive acrosomes (red) and AQP3-positive sperm tails (green) were detected in control testes (Supplementary Fig. 6 A, white arrows), while in testes of Wt1 OE mice, PNA-positive acrosomes (red) were noted but AQP3 signal was absent (Supplementary Fig. 6B, white arrowheads). Similarly, AQP3-positive sperm tail (green) was detected in the epididymis of control mice (Supplementary Fig. 6 C), but no AQP3 signal was detected in epididymis of Wt1 OE mice (Supplementary Fig. 6D). We further analyzed the process of spermiogenesis using periodic-acid-Schiff (PAS) and hematoxylin staining. Although stages 1–12 of the seminiferous epithelial cycle were observed in Wt1 OE mice (Supplementary Fig. 7A-L), the number of elongated spermatids was virtually absent at stage V (Supplementary Fig. 7E’) and VI (Supplementary Fig. 7F’), indicating the defect of spermiogenesis in Wt1 OE mice.

Fig. 4

Overexpression of Wt1 in Sertoli cells caused defects of spermatogenesis and disorganization of Sertoli cells in Wt1 OE mice. No developmental anomalies were observed in Wt1 OE mice (Wt1+/Ctg; Amh-Cre) (A). The size of testes from Wt1 OE mice was slightly reduced compared to control mice (n = 3) (B-C). The epididymis in control mice were filled with mature sperm (D), whereas no mature sperm were observed in epididymis of Wt1 OE mice (E). The Sertoli cells in control and Wt1 OE mice were labeled with SOX9. SOX9-positive Sertoli cells were localized at the center of seminiferous tubules in both control (F, black arrows) and Wt1 OE mice (I, black arrowheads) at 1 week. The Sertoli cells in control testes were well organized, whereas the Sertoli cells in Wt1 OE mice were disordered (I, black arrowheads). Most of Sertoli cells moved to the periphery region of seminiferous tubules in control mice (G, indicated by black arrow) at 2 weeks. The Sertoli cells were still localized at the center of Wt1 OE mice (J, black arrowheads). At 8 weeks, SOX9-positive Sertoli cells were located at periphery region of seminiferous tubules in control males (H, black arrows), whereas a small number of SOX9-positive Sertoli cells were disorganized in the seminiferous tubules of Wt1 OE mice (I, black arrowheads). The protein levels of these genes in control and in Wt1 OE Sertoli cells were examined by Western Blot analysis (L, M). The data are presented as the mean ± SEM. *, P < 0.05. **, P < 0.01. ***, P < 0.001

We also examined the testis development in Wt1 OE mice during early developmental stages. As shown in Fig. 4, SOX9-positive Sertoli cells were localized at the center of seminiferous tubules in both control and Wt1 OE mice at 1 week. The Sertoli cells in control testes were well organized (Fig. 4F, black arrows), whereas the Sertoli cells in Wt1 OE mice were disordered (Fig. 4I, black arrowheads). By 2 weeks, most of Sertoli cells moved to the periphery region of seminiferous tubules in control mice (Fig. 4G, indicated by black arrow), but in Wt1 OE mice, the Sertoli cells were still localized at the center of seminiferous tubules (Fig. 4J, black arrowheads), which was consistent with Huwe1 CKO mice. Quantitative PCR analysis (Supplementary Fig. 5 A) and Western Blot assay (Fig. 4L, M) showed the expression of Cdh1, Snai1&2 and Twist was significantly increased in Sertoli cells of Wt1 OE mice, consistent with the changes observed in Huwe1 CKO mice. These results suggested that the defects in Sertoli cell development in Huwe1 CKO mice are most likely caused by upregulation of WT1 protein.

The defects in spermatogenesis in Huwe1 CKO mice were partially rescued by deletion of one allele of Wt1 gene

To further test whether the defect in Sertoli cell development in Huwe1 CKO mice was caused by the abnormal accumulation of WT1, we generated Wt1+/−; Huwe1flox/Y; Amh-Cre mice, in which one allele of Wt1 gene was deleted. Western blot analysis revealed a marked reduction in WT1 protein levels in Wt1 rescue (Wt1+/−; Huwe1flox/Y; Amh-Cre) Sertoli cells compared to Huwe1 CKO Sertoli cells. (Supplementary Fig. 8 A, B). The results of histological studies showed that relatively normal seminiferous tubules with sperm heads (Fig. 5B, arrowheads) were observed in testes of Wt1 rescue mice, but not in Huwe1 CKO testes (Fig. 5C). A large number of mature sperm were detected in the epididymis of control mice (Fig. 5D), and no mature sperm were noted in the epididymis of Wt1 rescue (Fig. 5E) or Huwe1 CKO (Fig. 5F) mice. PNA-positive acrosomes (red, white arrows) and AQP3-positive sperm tails (green, black arrows) were detected in control (Fig. 5G) and Wt1 rescue mice (Fig. 5H). In contrast, while PNA-positive acrosomes (red, white arrowheads) were noted in the seminiferous tubule of Huwe1 CKO mice, AQP3 signals were absent (Fig. 5I). These results indicated that the defects in spermatogenesis in Huwe1 CKO mice were partially rescued by deletion of one allele of Wt1 gene.

Fig. 5

The defect of spermatogenesis in Huwe1 CKO mouse was partially rescued by deleting one allele of Wt1 gene. The histology of testes from control (A), rescue (B, Wt1+/−;Huwe1flox/Y; Amh-Cre), and Huwe1 CKO (C) mice. A small number of sperm heads (B, arrowheads) were observed in rescue testes, but not in Huwe1 CKO testes (C). Large number of mature sperm were detected in the epididymis of control mice (D), and no mature sperm were noted in the epididymis of rescue (E) and Huwe1 CKO (F) mice. PNA positive acrosome (red, white arrows) and AQP3 positive sperm tail (green, black arrows) were detected in control testes (G) and rescue mice (H). Only PNA positive acrosome (red, white arrowheads) was noted and AQP3 signal was absent in testes from Huwe1 CKO mice at 8 weeks (I)

HUWE1 mediated ubiquitination of WT1

To investigate whether WT1 is a direct target of HUWE1, protein immunoprecipitation (IP) assays were conducted. Whole-cell lysates from testes were immunoprecipitated using either anti-HUWE1 or anti-WT1 antibody. Then transferred membranes were immunoblotted with the corresponding antibodies. As shown in Fig. 6A, WT1 and HUWE1 proteins were mutually pulled down by each other, indicating that HUWE1 was interacted with WT1 protein in Sertoli cells.

Fig. 6

HUWE1 mediates ubiquitination of WT1. Immunoprecipitation assays confirmed a direct protein-protein interaction between HUWE1 and WT1 (A). Ubiquitination assays showed that HUWE1 was able to ubiquitinate WT1. Overexpression of HUWE1 markedly intensified ubiquitinated WT1 bands in IP samples (B). In vitro cultured primary Sertoli cells were treated by MG132 (10 µM, 6 h) or DMSO. Western blot assays showed the relative protein levels of WT1 was significant accumulated after MG132 treatment (C, D). Mutagenesis of lysine residues in WT1 aimed to identify sites of ubiquitination modification. Mutation of 320th and 444th lysine residues (K320S&K444S) completely abolished WT1 ubiquitination (E). The data are presented as the mean ± SEM. *, P < 0.05

To test whether WT1 could be ubiquitinated by HUWE1, ubiquitination assays were conducted. HEK293T cells were transfected with UB-V-GFP, WT1-FLAG, and HUWE1-HA expression plasmids. Whole-cell lysates were then immunoprecipitated using anti-FLAG antibody, and ubiquitinated WT1-FLAG was detected by immunoblotting with anti-GFP antibody. As shown in Fig. 6B, the level of ubiquitinated WT1 was increased in HUWE1-HA overexpressing HEK293T cells compared to the control group, indicating that WT1 could be ubiquitinated by HUWE1.

To examine whether WT1 protein is degraded by ubiquitination, in vitro cultured primary Sertoli cells were treated with the proteasome inhibitor MG132. We found that the protein level of WT1 was significantly increased after MG132 treatment, suggesting that WT1 protein is degraded via ubiquitin-proteasome system in Sertoli cells (Fig. 6C, D).

To identify the ubiquitination site(s) on WT1, computational predictions were performed, which identified three lysine residues (K320, K358, and K444) as potential ubiquitination sites. To assess the function of these lysine residues, three mutant plasmids were constructed in which the lysine residues were replaced with serine, and these plasmids were co-transfected into HEK293T cells along with HUWE1-HA and UB-V-GFP plasmids. Ubiquitination assays showed that the level of ubiquitinated WT1-K358S was comparable to that of wild-type WT1. By contrast, the levels of ubiquitinated WT1 were decreased in WT1-K320S and WT1-K444S transfected cells. To further confirm these findings, we generated a double mutant plasmid with mutations at both K320 and K444. Immunoblotting of double mutant revealed nearly complete absence of ubiquitinated WT1, indicating that both 320th and 444th lysine residues were essential for the ubiquitination of WT1 (Fig. 6E). Together, these results suggest that WT1 is a direct ubiquitination substrate of HUWE1.