Research ArticleFertility

Dephosphorylation of protamine 2 at serine 56 is crucial for murine sperm maturation in vivo

See allHide authors and affiliations

Science Signaling  26 Mar 2019:
Vol. 12, Issue 574, eaao7232
DOI: 10.1126/scisignal.aao7232

Protamine dephosphorylation for fertility

During the final stage of spermatogenesis, protamines tightly package DNA in the mature sperm. Itoh et al. generated mice deficient in the heat shock protein and chaperone Hspa4l, which is implicated in spermatogenesis. The mice were infertile with malformed sperm heads, a phenotype similar to that of mice deficient in the phosphatase Ppp1cc2. The authors showed that Hspa4l was required to release Ppp1cc2 from a complex with other chaperones, enabling its translocation to chromatin. In vitro studies showed that Ppp1cc2 dephosphorylated protamine 2 at Ser56. Expression of the unphosphorylatable protamine 2 S56A mutant reversed the infertility of Hspa4l-deficient mice, suggesting that the dephosphorylation of protamine 2 at Ser56 is important for its role in sperm maturation.


The posttranslational modification of histones is crucial in spermatogenesis, as in other tissues; however, during spermiogenesis, histones are replaced with protamines, which are critical for the tight packaging of the DNA in sperm cells. Protamines are also posttranslationally modified by phosphorylation and dephosphorylation, which prompted our investigation of the underlying mechanisms and biological consequences of their regulation. On the basis of a screen that implicated the heat shock protein Hspa4l in spermatogenesis, we generated mice deficient in Hspa4l (Hspa4l-null mice), which showed male infertility and the malformation of sperm heads. These phenotypes are similar to those of Ppp1cc-deficient mice, and we found that the amount of a testis- and sperm-specific isoform of the Ppp1cc phosphatase (Ppp1cc2) in the chromatin-binding fraction was substantially less in Hspa4l-null spermatozoa than that in those of wild-type mice. We further showed that Ppp1cc2 was a substrate of the chaperones Hsc70 and Hsp70 and that Hspa4l enhanced the release of Ppp1cc2 from these complexes, enabling the freed Ppp1cc2 to localize to chromatin. Pull-down and in vitro phosphatase assays suggested the dephosphorylation of protamine 2 at serine 56 (Prm2 Ser56) by Ppp1cc2. To confirm the biological importance of Prm2 Ser56 dephosphorylation, we mutated Ser56 to alanine in Prm2 (Prm2 S56A). Introduction of this mutation to Hspa4l-null mice (Hspa4l−/−; Prm2S56A/S56A) restored the malformation of sperm heads and the infertility of Hspa4l−/− mice. The dephosphorylation signal to eliminate phosphate was crucial, and these results unveiled the mechanism and biological relevance of the dephosphorylation of Prm2 for sperm maturation in vivo.


Spermatogenesis is a stepwise process that requires coordination of mitosis, meiosis, and spermiogenesis, the process by which haploid spermatids mature into sperm (1, 2). During the elongation phase of spermiogenesis, several critical processes occur to facilitate sperm motility, including nuclear compaction and chromatin condensation to generate the distinct head shape, as well as the redistribution of mitochondria within the flagella (3). Packaging haploid genomes is essential for nuclear condensation (4) and is achieved by replacing histones with transition proteins and, subsequently, with protamines, which are believed to be crucial in chromatin condensation, determination of sperm shape, and protection of the haploid genome (5).

A few species, including humans and mice, have two genes encoding protamines (Prms, P1/Prm1, and P2/Prm2), and the gene for Prm2, but not for Prm1, encodes a precursor protein, which undergoes proteolytic processing (6). It was reported more than 30 years ago that Prm2 is posttranslationally modified by phosphorylation and dephosphorylation in vivo (7); however, the underlying mechanism and biological consequences of this modification have not been clarified. Heat shock proteins (Hsps) are evolutionarily highly conserved, and they make up a group of structurally unrelated protein families (8). Hsps of 70 kDa (Hsp70s), such as Hsp70 (also termed Hspa1b) and Hsc70 (also termed Hspa8), are molecular chaperones involved in various protein-processing reactions, including protein transport across membranes (9). Hsp110 family members have a similar architecture to that of Hsp70s, except for the existence of a longer linker domain between the N-terminal adenosine triphosphatase domain and the C-terminal peptide-binding domain (10). These proteins act as nucleotide exchange factors (NEFs) (11, 12) and accelerate the dissociation of nucleotides from Hsp70/adenosine diphosphate complexes and the subsequent release of Hsp70-bound substrates in an adenosine 5′-triphosphate (ATP)–dependent manner. Studies of the crystal structures of Hsp70 or Hsc70 in a complex with the NEF of Hsp110 have revealed how these chaperones cooperate in protein folding (13, 14).

In a screen to identify proteins involved in spermatogenesis, we previously identified Hspa4l (also termed Apg-1) (15), which belongs to the Hsp110 family and is predominantly expressed in the testis (16). In the present study, we generated Hspa4l gene knockout mice to clarify the biological function of this Hsp110 member protein in vivo. Consistent with a previous report (17), we found that Hspa4l-deficient male mice were sterile and impaired in spermiogenesis. Furthermore, we found that the phosphatase Ppp1cc2 was a substrate of Hsp70s and that Hspa4l acted as a regulator of the binding of Hsp70s to Ppp1cc2, resulting in increased chromatin localization of Ppp1cc2 and dephosphorylation of Prm2 in vivo.


Male mice deficient in Hspa4l are infertile and impaired in spermiogenesis

We generated Hspa4l-deficient mice by homologous recombination in E14 embryonic stem (ES) cells derived from the 129/Ola strain (fig. S1, A to D). The heterozygous and homozygous mice were viable, and intercrossing of heterozygous mice generated progeny at near Mendelian ratios (Hspa4l+/+:Hspa4l+/−:Hspa4l−/− = 25:55:31, n = 111). Hspa4l-null mice showed no gross abnormalities, but male mice were sterile after eight generations of backcrossing with C57BL/6 mice (Fig. 1A and fig. S2).

Fig. 1 Male mice deficient in Hspa4l are infertile, and their sperm heads show a rounded morphology.

(A) Male mice of the indicated genotype (Hspa4l+/+ or Hspa4l−/−) were mated with the indicated female mice, and the resulting litter sizes was determined. Data are means ± SD. ***P < 0.001 by two-tailed Student’s t test. (B) Representative images of the heads of sperm in the cauda epididymis of wild-type and Hspa4l-deficient mice. Scale bar, 10 μm. (C and D) Fertilization ratios with Hspa4l−/− sperm in vitro. We analyzed in vitro fertility using two male mice of each genotype with ZP-intact oocytes (C) or ZP-free oocytes (D) in five or three independent experiments, respectively. Data are means ± SD. *P < 0.05, ***P < 0.001 by two-tailed Student’s t test.

Spermiogenesis was observed in the seminiferous tubules of Hspa4l−/− mice (fig. S3), and we found no abnormality in the number of spermatogonia or sperm of Hspa4l−/− mice (fig. S4). However, Hspa4l−/− spermatozoa in the cauda epididymidis had round heads (Fig. 1B). These spermatozoa could neither bind to zona pellucida (ZP)–intact oocytes (table S1) nor fertilize ZP-intact or ZP-free oocytes as efficiently as wild-type spermatozoa (Fig. 1C). Furthermore, these fertilized oocytes did not develop to the blastocyst stage (table S2).

To examine whether haploid cells in these mice contained a functional nucleus, we performed intracytoplasmic sperm injection (ICSI) and round spermatid injection (ROSI) assays. After ICSI, oocytes injected with Hspa4l−/− spermatozoa developed less efficiently than those injected with wild-type spermatozoa (table S3) and did not give rise to offspring (Table 1). After ROSI, in contrast, oocytes injected with Hspa4l−/− round spermatids developed comparably to those injected with wild-type spermatids (table S4) and efficiently gave rise to offspring (Table 2). Together, these results indicate that postmeiotic spermatids fail to differentiate into functional spermatozoa in Hspa4l−/− mice.

Table 1 Development of oocytes after intracytoplasmic injection of sperm.

Results are from three independent experiments using two male mice of each genotype.

View this table:
Table 2 Development of oocytes injected with round spermatids.

Results are from three independent experiments using two male mice of each genotype.

View this table:

Localization of Ppp1cc2 in a chromatin-binding fraction is impaired in Hspa4l-deficient mice

The phenotype of Hspa4l−/− mice is similar to that of Ppp1cc [protein phosphatase 1 gamma subunit (Pp1γ)]–deficient mice (18, 19), and spermatozoa express a testis- and sperm-specific isoform of Ppp1cc (Ppp1cc2) (20, 21). Because molecular chaperones can modulate protein localization (9), we examined whether the subcellular localization of Ppp1cc2 was altered in Hspa4l−/− spermatozoa. The distribution of Ppp1cc2 was comparable between Hspa4l−/− and wild-type sperm cells in cytoplasmic and nuclear fractions (fig. S5A). However, the amount of Ppp1cc2 localized in the chromatin-binding fraction was statistically significantly less in Hspa4l-null spermatozoa than that in wild types (Fig. 2, A and B, and fig. S5B). These results indicate that the subcellular localization of Ppp1cc2 is impaired in Hspa4l-deficient sperm.

Fig. 2 Ppp1cc2 forms a chaperone-substrate complex with Hsc70, which includes Hspa4l, and Ppp1cc2 dephosphorylates the pSer56of Prm2.

(A) Chromatin-binding fractions from sperm of mice of the indicated genotypes were analyzed by Western blotting (IB) with antibodies (Abs) against the indicated proteins. Blots are representative of three experiments. (B) Ppp1cc2 to Prm2 ratios in the chromatin-binding fractions. Results indicate the relative Ppp1cc2:Prm2 ratio in Hspa4l-deficient sperm compared to that in wild-type sperm. Data are means ± SD of three independent experiments. ***P = 0.0003 by Welch’s t test. (C) Interaction between Ppp1cc2 and Hsc70. Sperm cell lysates from the indicated mice were subjected to immunoprecipitation (IP) with control or anti-Ppp1cc antibodies and then analyzed by Western blotting with antibodies against the indicated proteins. Blots are representative of three experiments. IgG, immunoglobulin G. (D) Interaction between Hspa4l and Hsc70. Sperm cell lysates from the wild-type mice were subjected to immunoprecipitation with control or anti-Hspa4l antibodies and then analyzed by Western blotting with antibodies against the indicated proteins. Blots are representative of three experiments. (E) Interaction between Ppp1cc2 and Prm2. Sperm cell lysates from wild-type mice were subjected to GST pull-down experiments with the indicated fusion proteins. The samples were then analyzed by Western blotting with anti-GST antibody. Blots are representative of three experiments. (F) Ppp1cc2 dephosphorylates Ser56 of Prm2. Phosphatase assays were performed using the indicated fusion proteins as substrates. After the reactions were complete, phosphorylated or total substrates were detected by biotinylated Phos-tag or Coomassie blue staining (CBB), respectively. Data are representative of three experiments. (G and H) Results of three independent phosphatase assays using pGST-Prm2 (G) or pGST-Prm2 S56A (H). These substrates were dephosphorylated by samples immunoprecipitated by an anti-Ppp1cc antibody from the chromatin-binding fraction of epididymal sperm as described in Materials and Methods. Results indicate the relative ratios of phosphorylated proteins after the reaction using an anti-Ppp1cc antibody to those using a control antibody. Data are means ± SD. For (G), P = 0.0323; for (H), P = 0.2785 by Welch’s t test.

To unveil an underlying mechanism, we tested whether Ppp1cc2 was a substrate of Hsp70s. An anti-Ppp1cc antibody co-immunoprecipitated Hsc70 (Fig. 2C), consistent with the protein phosphatase-1 (Pp1) interactome analysis of the human testis (22). It also co-precipitated Hsp70 (fig. S6). These results suggest that Ppp1cc could be a putative substrate of Hsp70s. Next, we tested whether Hspa4l acted as a co-chaperone for Hsp70s in sperm. An anti-Hspa4l antibody co-immunoprecipitated both Hsc70 (Fig. 2D) and Hsp70 (fig. S6B) from cauda epididymal sperm lysates. Together, these results suggest the formation of a chaperone-substrate complex between Hsp70s and Ppp1cc2, which includes Hspa4l as a co-chaperone. We postulated that the NEF activity of Hspa4l might accelerate the release of Ppp1cc2 from the complex, enabling the freed Ppp1cc2 to localize to chromatin. Consistent with this hypothesis, we co-immunoprecipitated Hsc70 with an anti-Ppp1cc antibody from the cytoplasmic and nuclear fractions of sperm cell lysates from wild-type mice, but not from the chromatin-binding fraction (fig. S7).

Ser56 of Prm2 is a putative target of the phosphatase Ppp1cc

A deficiency in Prm1 or Prm2 causes male infertility due to impaired sperm maturation (23), and Prms are posttranslationally modified (5, 24). We examined whether Ppp1cc2 could bind to and modify Prms in chromatin. First, we performed a pull-down assay with recombinant Prm1 [glutathione S-transferase (GST)–Prm1] or the mature form of Prm2 (GST-Prm2). GST-Prm2, but not GST-Prm1, pulled down Ppp1cc2 from sperm cell lysates (Fig. 2E). Next, we substituted a putative phosphorylation site at Ser56 in GST-Prm2 with alanine (GST-Prm2 S56A) (24, 25) and prepared phosphorylated GST-Prm2 or GST-Prm2 S56A (pGST-Prm2 or pGST-Prm2 S56A, respectively) using testicular cell lysate as a kinase. In phosphatase assays, we used these proteins as substrates, whereas for a source of phosphatase, proteins immunoprecipitated from the chromatin-binding fraction with an anti-Ppp1cc antibody or a control antibody were used. The former immunoprecipitate contained Ppp1cc2 (Fig. 2F) and dephosphorylated pGST-Prm2 but not pGST-Prm2 S56A (Fig. 2, G and H). Together, these data suggest that Ppp1cc2 can dephosphorylate Prm2 Ser56 in sperm chromatin.

Substitution of Prm2 Ser56 with alanine rescues the phenotype of Hspa4l−/− mice

To investigate the biological relevance of the dephosphorylation of Prm2-pSer56, we generated Prm2S56A/wt and Prm2S56D/wt C57BL/6 mice, which harbored mutations of Ser56 to alanine or aspartate, respectively, by using the technology of CRISPR-Cas9 (2629) (fig. S8, A and B). No mutations were identified in either strain in potential off-target sites (fig. S9, A and B), nor in neighboring sites of mutated regions (fig. S10). The ensuing heterozygous and homozygous mice were viable and showed no gross abnormalities. Similarly, both Prm2S56A/S56A and Prm2S56D/S56D male mice were fertile (fig. S8C), and their spermatozoa fertilized oocytes as efficiently as wild types in vitro (table S5 and table S6). These results suggest that these mutations were unlikely to affect the structural integrity of the nucleochromatin.

We intercrossed male Prm2S56A/S56A and Prm2S56D/S56D mice with Hspa4l−/− females. Although both Hspa4l−/− and Hspa4l−/−; Prm2S56D/S56D male mice were infertile, the litter size from pairings with Hspa4l−/−; Prm2S56A/S56A male mice was comparable to that with wild types and was statistically significantly greater than that with Hspa4l−/− or Hspa4l−/−; Prm2S56D/S56D mice, indicating that the mutation of Prm2 Ser56 to an alanine (S56A), but not to a phosphomimetic of phospho-serine, aspartate (S56D), rescued the phenotype of infertility in Hspa4l-null mice (Fig. 3A). An in vitro fertilization experiment confirmed that Hspa4l−/−; Prm2S56A/S56A spermatozoa could fertilize oocytes more efficiently than did Hspa4l−/− spermatozoa (Fig. 3B), and oocytes fertilized with Hspa4l−/−; Prm2S56A/S56A spermatozoa developed to the blastocyst stage in culture similarly to those fertilized with wild-type spermatozoa (table S7). Furthermore, morphological examination showed that the percentage of spermatozoa with a round head in Hspa4l−/−; Prm2S56A/S56A mice was statistically significantly less than that in Hspa4l−/− mice (Fig. 3, C and D). Mitochondrial staining revealed that the mitochondria were localized to the round heads of spermatozoa from Hspa4l−/− mice. In contrast, the mitochondria in Hspa4l−/−; Prm2S56A/S56A sperm localized in the mid-piece as did those in the sperm of wild-type mice (fig. S11). Together, these data suggest that the substitution of Ser56 with alanine rescued several aspects of the infertile phenotype and abnormal sperm morphology of Hspa4l−/− male mice. Furthermore, these results demonstrate that posttranslational modifications of Prms, in addition to those of histones (5), are biologically relevant and suggest that the dephosphorylation signal is a crucial biological process for murine sperm maturation in vivo.

Fig. 3 Conversion of the codon encoding Ser56 to a codon encoding alanine in the Prm2 locus rescues the infertile phenotype of Hspa4l−/−mice.

(A) Hspa4l−/−; Prm2S56A/S56A male mice are fertile. Male mice of the indicated genotypes were mated with wild-type (wt) female mice, and the resulting litter sizes were determined. Data are means ± SD. ***P < 0.001 by two-tailed Student’s t test. “#” indicates Welch’s t test. (B) Analysis of in vitro fertility. Two male mice of each the indicated genotypes were used. Data are means ± SD of three independent experiments. **P < 0.01, ***P < 0.001 by two-tailed Student’s t test. (C and D) Substitution of Ser56 with alanine in Prm2 rescues the abnormal morphology of Hspa4l−/− sperm. (C) Results are from two male mice of each of the indicated genotypes. Data are means ± SD of three independent experiments. *P < 0.05, ***P < 0.001 by two-tailed Student’s t test. (D) Representative images showing the morphology of the sperm from mice of the indicated genotypes. Scale bar, 10 μm.


Humans and mice have two genes that encode protamines (Prm1 and Prm2) (30). Previous reports demonstrated that chimeric male mice generated by injecting Prm1+/− or Prm2+/− ES cells into C57BL/6 blastocysts transmit neither the mutant nor the wild-type allele from these cells to the next generation (23). Prm1+/− and Prm2−/− male mice were also generated using TT2-XO ES cells (31) or by applying CRISPR-Cas9–mediated gene editing (32), respectively, and both sets of mice are infertile. Wu et al. (33) reported that the amount of Prm2 is reduced in step-15 spermatids of Camk4 (Ca2+/calmodulin-dependent protein kinase IV)–deficient mice, which are infertile on a mixed genetic background and that Prm2 is a putative substrate of Camk4. These studies showed the importance of protamines and their posttranslational regulation in mouse models. In humans, biochemical studies of male infertility demonstrated that changes in the protamine content of sperm chromatin, incomplete protamine P2 precursor processing, or alterations in the P1:P2 ratio contribute to male infertility (34, 35).

Protamines have a series of small anchoring domains that contain positively charged amino acids (35), which bind to negatively charged DNA of the minor groove. In addition, protamines have an element containing serine and threonine residues, which are phosphorylated after synthesis. Phosphorylation may make protamines more negatively charged and decrease their affinity for negatively charged DNA. About 50 years ago, the serine residues of trout protamine were reported to be dephosphorylated after phosphorylation (36). More than 30 years ago, human Prm2 was reported to be posttranslationally modified by phosphorylation and dephosphorylation in vivo (7), and the Ser59 in human Prm2 can be phosphorylated (25). Furthermore, Ser56 in mouse Prm2, which is equivalent to Ser59 of human Prm2, is phosphorylated in vivo (24). However, it has not been clarified whether or how these sites are dephosphorylated.

Although the importance of pSer56 dephosphorylation was supported by experiments in which this residue was substituted with an alanine, the results of experiments in which Ser56 was substituted with an aspartate, a phosphomimetic of phospho-serine, raise the question of why Prm2S56D/S56D male mice are fertile (fig. S8C and table S6). Multisite phosphorylation of human Prm2 has been reported (37). Many regulatory proteins are regulated by multisite phosphorylation, and some of them have proved to be multisite-phosphorylated by sequential intramolecular phosphorylation cascade, suggesting a complex posttranslational network including kinases and phosphatases (38, 39). For example, the hepatitis C virus nonstructural protein 5A (NS5A) is first phosphorylated at Ser232, and this phosphorylation primes the sequential phosphorylation of other serine residues of NS5A, which are important for viral replication (40, 41). On the basis of these reports and our observations of the mutant mice in this study, we postulate that phosphorylation of Prm2 at Ser56 might prime a sequential intramolecular phosphorylation cascade. Ppp1cc2 might also dephosphorylate other serine residues of Prm2, and this step could be crucial for sperm maturation. Consistent with this postulation, Prm2S56A/S56A or Hspa4l−/−; Prm2S56A/S56A male mice were fertile (Fig. 3, A and B, and tables S5 and S7), because substitution of Ser56 with alanine would be expected to block the phosphorylation cascade. On the other hand, substitution of Ser56 with aspartate would not block this cascade. We suggest that because Ppp1cc2 could dephosphorylate the other phosphorylated serine residues primed by the aspartate substitution, Prm2S56D/S56D male mice were fertile (fig. S8C and table S7). However, Hspa4l−/−; Prm2S56D/S56D male mice were infertile because of the impaired recruitment of Ppp1cc (Fig. 3A and table S8). Our current study has just began to unravel the sperm maturation network by kinases and phosphatases, and detailed mechanisms remain to be clarified in experiments with Ppp1cc-null mice or with antibodies that recognize specific phosphorylated serine residues.

Pp1 is a major protein serine and threonine phosphatase and consists of a catalytic and a regulatory subunit (42). Although several isoforms exist that are involved in a wide range of cellular processes, the catalytic subunits are encoded by only three genes (Ppp1ca, Ppp1cb, and Ppp1cc) (43). Two groups have reported that Ppp1cc-deficient male mice are infertile without apparent abnormality in other tissues (18, 19). To unveil the underlying mechanisms, potential targets or interacting molecules of Ppp1cc2 were examined in the testis and sperm (21, 22, 4446). Here, we demonstrated that Hsc70 and Hsp70 regulated the localization of Ppp1cc2 to the chromatin-binding fraction, together with Hspa4l, and that Ppp1cc2 could dephosphorylate Prm2 at Ser56 in chromatin. The biological relevance of this dephosphorylation signal was confirmed by generating Hspa4l−/−; Prm2S56A/S56A mice.

Several limitations of this study need to be addressed. Regulatory subunits of Pp1 are thought to control its localization, activity, and substrate specificity (47). The identity of the Ppp1cc2-specific regulatory subunit in the chromatin-binding fraction of sperm remains to be clarified. Next, because of technical limitations, we could not show the phosphorylation state of Prm2 at Ser56 in Hspa4l-null mice directly. Such an experiment would provide additional and clear evidence that the dephosphorylation of Prm2 at Ser56 is crucial for sperm maturation.

Although genetic analysis of human Prm2 has been widely performed, pathogenic mutations are estimated to be rare in infertile patients (48). On the other hand, a report showed that the amount of Hspa4l in the sperm of patients with asthenozoospermia or teratospermia is less than that in the sperm of healthy controls (49). Because the posttranslational regulatory mechanism of Prm2 proposed in this study potentially associates with male infertility in humans, it would be worth examining the phosphorylation state of Ser59 of human Prm2 in the sperm of infertile patients, as well as investigating any alterations of Hspa4l or Ppp1cc.



Polyclonal antibodies against Hspa4l (Apg-1, N-96), Hsp70 (K-20), Hsc70 (K-19), protamine 1 (M-51), protamine 2 (N-20), Ppp1cc (Pp1γ C-19), acrosin (C-14), and GST (Z-5), as well as anti-mouse Hsc70 (B-6) and anti–α-tubulin (B-7) monoclonal antibodies were purchased from Santa Cruz Biotechnology. Phycoerythrin-conjugated anti–α6-integrin (G0H3) and anti-EpCAM (G8.8) monoclonal antibodies were from BD Pharmingen. Anti-mouse Hsp70 and Hsc70 (B-6) antibodies were biotinylated, whereas Alexa Fluor 647 or fluorescein isothiocyanate was conjugated to anti-mouse CD117 (c-Kit, ACK2) or anti-rat IgG κ (K4F5, Coulter), respectively. Horseradish peroxidase (HRP)–conjugated anti-rabbit, anti-mouse, and anti-goat IgG polyclonal antibodies, as well as HRP-conjugated streptavidin were from DakoCytomation.

Plasmids and preparation of recombinant proteins

To prepare pX330-Prm2, we designed a Prm2 gene–specific guide RNA sequence using the CRISPR design tool ( (50), and ligated annealed oligonucleotides of 5′-CACCGacaggcgctgctctcgtaag-3′ and 5′-AAACcttacgagagcagcgcctgtC-3′ into the Bbs I site of pX330-U6-Chimeric_BB-CBh-hSpCas9 (26). We amplified fragments containing the coding sequence of Prm1 and the mature form of Prm2 by reverse transcription polymerase chain reaction (PCR), which were ligated into pGEX-4 T-1 (GE Healthcare Japan) to construct pGEX/Prm1 and pGEX/Prm2, respectively. To generate pGEX/Prm2/S56A, we mutated the cloned sequence of the mature form of Prm2 to substitute Ser56 with alanine using the QuikChange site-directed mutagenesis kit (Stratagene) and ligated the fragment into pGEX-4 T-1. BL21 cells transformed with pGEX/Prm1, pGEX/Prm2, or pGEX/Prm2/S56A were lysed, and the recombinant proteins GST-Prm1, GST-Prm2, and GST-Prm2 S56A were prepared using a glutathione sepharose column (GE Healthcare Japan), respectively.

Generation of Hspa4l-deficient and Prm2 mutant mice

We constructed the Hspa4l-targeting vector, which was designed to replace a 0.4-kb genomic fragment containing its promoter and the first exon with a neomycin resistance gene cassette (fig. S1A). We generated Hspa4l-deficient mice by homologous recombination in E14 ES cells derived from the mouse strain 129/Ola. After targeting, we microinjected ES clones into C57BL/6 blastocysts and mated chimeric mice with C57BL/6 mice to obtain F1 Hspa4l+/− mice. The mutant mice were backcrossed to C57BL/6 mice for at least eight generations before being used in experiments, unless otherwise indicated in the figure legends. To genotype the mice, we performed PCR analysis using extracted genomic DNAs from tails and the following sets of primers (a and b in fig. S1A): for the targeted locus, 5′-ctgctaaagcgcatgctccaga-3′ and 5′-cttagaagagcgtccgcgagaca-3′; and for the wild-type locus, 5′-ctgccctagattttccgatagagagt-3′ and 5′-gctacagcgatgtagcagttgag-3′. We generated Prm2S56A/wt and Prm2S56D/wt mice using a CRISPR-Cas9–mediated genome-editing strategy, with minor modification as described by Mashiko et al. (2629) (fig. S8, A and B). We injected a circular plasmid, pX330-Prm2 (5 ng/μl), and single-stranded oligonucleotides (10 ng/μl; for Prm2S56A/wt, 5′-cacacaggggccaccaccaccacagacacaggcgctgcGc CAgGaagaggctacataggatccacaagag gcgtcggtcatgc-3′; and for Prm2S56D/wt: 5′-caggggccaccaccaccacagacacaggcgctgcGACcgtaagaggctacataggatccacaagaggcgtcgg-3′) into the pronuclei of fertilized oocytes of the C57BL/6 strain and transferred two-cell stage embryos into the oviducts of pseudopregnant ICR females. Genomic DNAs of offspring were extracted and analyzed by PCR using the following sets of primers: for the mutated locus of Prm2S56A/wt, 5′-agacacaggcgctgcGcCAgG-3′ and 5′-ggaaggcagcagggtttagatgg-3′; for the wild-type locus of Prm2S56A/wt, 5′-agacacaggcgctgctctcgt-3′ and 5′-ggaaggcagcagggtttagatgg-3′; for the mutated locus of Prm2S56D/wt, 5′-cacagacacaggcgctgcGAC-3′ and 5′-ggaaggcagcagggtttagatgg-3′; and for wild-type locus of Prm2 S56D/wt, 5′-cacagacacaggcgctgctct-3′ and 5′-ggaaggcagcagggtttagatgg -3′. To confirm the mutations, we amplified genomic fragments by PCR using a set of primers, 5′-gggcctggacaagaccatgaac-3′ and 5′-cttggctccaggcagaatccac-3′, and sequenced directly using a sequence primer, 5′-ggaaggcagcagggtttagatgg-3′. Potential off-target sites (OT-1 to OT-7) were found at the website CRISPR Design ( (50), and genomic fragments containing the off-target sequences were amplified and sequenced using the primers listed in table S9. All animal experiments were reviewed and approved by the Animal Committee of Kyoto University.

Preparation of subcellular fractions

Cauda epididymides were homogenized on ice with a glass Teflon homogenizer in buffer-150 [150 mM NaCl, 50 mM tris (pH 7.4), leupeptin (2 μg/ml), and 1 mM phenylmethylsulfonyl fluoride (PMSF)], passed through nylon mesh (40 μm), and incubated for 60 min at 4°C after adding Triton X-100 (to a final concentration of 1.0%). After centrifugation at 5000g for 5 min, the supernatant was analyzed as the nucleocytoplasmic fraction. The pellet was extracted with 2 M NaCl for 20 min at 4°C and centrifuged, and the supernatant was designated as the NaCl fraction. The remaining pellet was digested with deoxyribonuclease I for 20 min at 37°C and centrifuged and was used as the chromatin-binding fraction. In some experiments, homogenized cauda epididymides were passed through nylon mesh, solubilized in buffer-150 containing 0.5% Nonidet P40, and analyzed as the cytoplasmic fraction. The pellet was further solubilized in buffer containing 1.0% Triton X-100 and analyzed as the nuclear fraction.

Immunoprecipitation, GST pull-down assay, and Western blotting

For co-immunoprecipitation of Hspa4l or Ppp1cc, we prepared sperm lysates in buffer-150 containing 1.0% Triton X-100 or 0.5% Nonidet P40, immunoprecipitated with anti-Hspa4l or anti-Ppp1cc antibodies, respectively, and analyzed the samples by Western blotting using the antibodies indicated in the figure legends. For the GST pull-down assay, we incubated GST-Prm1 or GST-Prm2 in sperm cell lysates prepared in buffer-150 containing 0.5% Nonidet P40, pulled down protein complexes with glutathione sepharose beads, and subjected the immunoprecipitates to Western blotting analysis using an anti-Ppp1cc antibody. For detection of basic proteins, we used Bjerrum and Schafer-Nielsen buffer [48 mM tris (pH 9.2), 39 mM glycine, 20% methanol, and 0.01% SDS] as the transfer buffer. After Western blotting, in some experiments, band densities were quantitatively measured with Image Lab software (Bio-Rad).

In vitro phosphatase assay

We lysed dissociated testicular cells in kination buffer [150 mM NaCl, 10 mM tris (pH 7.5), 10 mM MgCl2, 0.5 mM dithiothreitol (DTT), leupeptin (2 μg/ml), 1 mM PMSF, 1 mM NaVO4, and 1 mM NaF] containing 1.0% Triton X-100, and phosphorylated GST-Prm2 or GST-Prm2 S56A using the lysates as a kinase in the presence of 2 mM ATP at 30°C for 45 min. After kination, phosphorylated GST-fusion proteins were prepared using a glutathione sepharose column, suspended in phosphatase buffer (10 mM Hepes, 1 mM MnCl2, and 1 mM DTT) and used as substrates. For the phosphatase, we used proteins precipitated by an anti-Ppp1cc antibody from the chromatin-binding fraction of epididymal sperm and incubated the substrates in phosphatase buffer at 30°C for 45 min. After the reaction, the products were subjected to SDS–polyacrylamide gel electrophoresis, and phosphorylated or total products were analyzed by chemiluminescence using biotinylated Phos-tag (Phos-tag biotin BTL-104, Manac Inc.) or Coomassie blue staining, respectively. Densities of products were quantitatively measured using Image Lab software.

Preparation of dissociated testicular cells and flow cytometric analysis of spermatogenic cells

We prepared dissociated testicular cells by a two-step enzymatic digestion using collagenase and trypsin and performed flow cytometric analysis of spermatogonia as described previously (51, 52). To analyze the DNA content of cells, dissociated testicular cells were fixed with methanol, stained with propidium iodide, and analyzed on an Epics XL flow cytometer (Coulter).

In vitro fertilization and sperm-egg binding assays

We performed an in vitro fertilization assay and a sperm-egg binding assay as described previously (53). In some experiments, we used ZP-free oocytes prepared as described previously (54).


We performed ICSI according to a previously described procedure (55). ROSI was performed according to a previously described procedure (56) with modification. Briefly, we prepared dissociated testicular cells and injected round spermatids into superovulated oocytes with a piezo-actuated micromanipulator. Oocytes that received round spermatids were placed in Ca2+-free CZB medium containing 10 mM SrCl2 for 20 min after the injection and then were cultured for 1 hour for their activation. We incubated oocytes that successfully received ICSI or ROSI in CZB medium and transferred them to the oviducts of pseudopregnant ICR females after reaching the two-cell stage unless indicated otherwise in the figure legends.

Sperm count

Sperm were collected from the vas deferens, and the number of sperm was counted using a hemocytometer.

Sperm morphology

Sperm from cauda epididymides were incubated in HTF medium for 1 hour, and their morphology was observed by microscopy. Sperm were also stained with Hoechst33342 (5 μg/ml) and 50 nM Mito Tracker Green FM (Invitrogen) in HTF medium at 37°C for 30 min and then were analyzed using a fluorescence microscope, BZ-9000 (Keyence, Japan). Signals of Hoechst33342 or Mito Tracker Green FM were visualized as blue or green, respectively, and images were merged using software provided with the BZ-9000 microscope.

Fertility tests

We placed male mice with females for about 2 months to observe whether pregnancy ensured, confirming copulation by checking for vaginal plugs every morning and recording the numbers of litters produced.

Histological and immunocytochemical analyses

Testes were fixed with paraformaldehyde and embedded in paraffin. Mounted sections were deparaffinized, rehydrated, stained with hematoxylin and labeled antibodies, and analyzed using fluorescence microscopy for immunohistochemical analysis. For toluidine blue staining, testes were fixed with osmium and embedded in Epon.

Data analysis

Statistical analyses were performed using the JMP statistical software package (SAS Institute Inc.). Differences between the means of two groups were analyzed for statistical significance using two-tailed Student’s t tests or Welch’s t test when Levene’s test rejected homoscedasticity (cutoff, 0.05). Two nominal variables were analyzed using Fisher’s exact probability test. A value of P < 0.05 was considered to be statistically significant for all analyses.


Fig. S1. Generation of Hspa4l−/− mice.

Fig. S2. Hspa4l−/− male mice of a mixed strain are not sterile.

Fig. S3. Histological analysis of testes.

Fig. S4. The numbers of spermatogonia and sperm are normal in Hspa4l−/− mice.

Fig. S5. Subcellular localization of Ppp1cc2.

Fig. S6. Ppp1cc2 forms a chaperone-substrate complex with Hsp70.

Fig. S7. Subcellular analysis of the interaction of Ppp1cc2 with Hsc70.

Fig. S8. Generation of Prm2S56A/wt and Prm2S56D/wt mice and demonstration that Prm2S56A/S56A and Prm2S56D/S56D male mice are fertile.

Fig. S9. No mutations in potential off-target sites.

Fig. S10. No mutations in neighboring sites of mutated regions.

Fig. S11. Morphological analysis of sperm.

Table S1. Sperm binding assays.

Table S2. Development of oocytes in culture after fertilization with sperm from Hspa4l−/− mice.

Table S3. Development of oocytes in culture after intracytoplasmic injection of sperm.

Table S4. Development of oocytes in culture after injection with round spermatids.

Table S5. Development of oocytes in culture after fertilization with sperm from Prm2S56A/S56A mice.

Table S6. Development of oocytes in culture after fertilization with sperm from Prm2S56D/S56D mice.

Table S7. Development of oocytes in culture after fertilization with sperm from Hspa4l−/−; Prm2S56A/S56A mice.

Table S8. Development of oocytes in culture after fertilization with sperm from Hspa4l−/−; Prm2S56D/S56D mice.

Table S9. Primers used for sequencing potential off-target sites.


Acknowledgments: We thank C. Stocking-Harbers and A. Tanaka for critical reading of the manuscript and preparation of figures, respectively. The plasmid pX330-U6-Chmeric_BB-CBh-hSpCas9 was obtained from Addgene (Addgene plasmid 42230). Funding: This work was supported, in part, by the Smoking Research Foundation of Japan and Grants-in-Aid for Scientific Research (KAKENHI, 16H01387) from the Japan Society for the Promotion of Science. Author contributions: K.I. and J.F. designed research; K.I., G.K., H.W., R.M., A.I., and Y.L. performed cellular or biochemical research; C.I. and K.T. performed histochemical analysis; H.M., Y.K., S.K., and S.I. generated mutant mice; K.I., G.K., H.M., M.S., and J.F. analyzed and interpreted the data; K.I. performed the statistical analysis; G.K., M.S., S.I., C.I., and K.T. made comments; K.I. wrote the manuscript; and G.K., M.S., C.I., and J.F. edited the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials. The gene-manipulated mouse lines Prm2S56A/S56A, Prm2S56D/S56D, and Hspa4l–/– were deposited into the Riken BioResource Center or Center for Animal Resources and Development, Kumamoto University, with stock numbers RBRC09931, RBRC09932, and 2379, respectively.

Stay Connected to Science Signaling

Navigate This Article