Ablation of Ggnbp2 impairs meiotic DNA double‐strand break repair during spermatogenesis in mice

Abstract Gametogenetin (GGN) binding protein 2 (GGNBP2) is a zinc finger protein expressed abundantly in spermatocytes and spermatids. We previously discovered that Ggnbp2 resection caused metamorphotic defects during spermatid differentiation and resulted in an absence of mature spermatozoa in mice. However, whether GGNBP2 affects meiotic progression of spermatocytes remains to be established. In this study, flow cytometric analyses showed a decrease in haploid, while an increase in tetraploid spermatogenic cells in both 30‐ and 60‐day‐old Ggnbp2 knockout testes. In spread spermatocyte nuclei, Ggnbp2 loss increased DNA double‐strand breaks (DSB), compromised DSB repair and reduced crossovers. Further investigations demonstrated that GGNBP2 co‐immunoprecipitated with a testis‐enriched protein GGN1. Immunofluorescent staining revealed that both GGNBP2 and GGN1 had the same subcellular localizations in spermatocyte, spermatid and spermatozoa. Ggnbp2 loss suppressed Ggn expression and nuclear accumulation. Furthermore, deletion of either Ggnbp2 or Ggn in GC‐2spd cells inhibited their differentiation into haploid cells in vitro. Overexpression of Ggnbp2 in Ggnbp2 null but not in Ggn null GC‐2spd cells partially rescued the defect coinciding with a restoration of Ggn expression. Together, these data suggest that GGNBP2, likely mediated by its interaction with GGN1, plays a role in DSB repair during meiotic progression of spermatocytes.

GGNBP2 affects meiotic progression of spermatocytes remains to be established. In this study, flow cytometric analyses showed a decrease in haploid, while an increase in tetraploid spermatogenic cells in both 30-and 60-day-old Ggnbp2 knockout testes. In spread spermatocyte nuclei, Ggnbp2 loss increased DNA double-strand breaks (DSB), compromised DSB repair and reduced crossovers. Further investigations demonstrated that GGNBP2 co-immunoprecipitated with a testis-enriched protein GGN1. Immunofluorescent staining revealed that both GGNBP2 and GGN1 had the same subcellular localizations in spermatocyte, spermatid and spermatozoa.
Ggnbp2 loss suppressed Ggn expression and nuclear accumulation. Furthermore, deletion of either Ggnbp2 or Ggn in GC-2spd cells inhibited their differentiation into haploid cells in vitro. Overexpression of Ggnbp2 in Ggnbp2 null but not in Ggn null GC-2spd cells partially rescued the defect coinciding with a restoration of Ggn expression. Together, these data suggest that GGNBP2, likely mediated by its interaction with GGN1, plays a role in DSB repair during meiotic progression of spermatocytes.

K E Y W O R D S
DSB repair, GGN1, GGNBP2, knockout, meiosis, spermatocyte, testis 1 | INTRODUCTION Spermatogenesis consists of three major stages which include proliferation and differentiation of spermatogonia, meiotic divisions of spermatocytes and metamorphosis of haploid spermatids, which are tightly controlled by a highly orchestrated expression of several thousand different genes. 1,2 Failure to faithfully and timely complete one or more of these stages leads to male infertility.
Gametogenetin binding protein 2 (GGNBP2) is an evolutionarily conserved zinc figure protein across humans, mice and rats. 3,4 The biological functions of GGNBP2 are diverse. It has been reported that GGNBP2 participates in the regulation of the G2/M transition of somatic cell cycle 5,6 and in the modulation of the differentiation of placental trophoblastic stem cells. 3 9 ovarian cancer 10,11 and brain glioma. 12 GGNBP2 deficiency promotes cell growth, drug resistance and tumorigenic progression. 10,11 GGNBP2 is highly expressed in the adult testis and its expression is tightly associated with the start of spermatogenesis. 4,13 Recently, we and others demonstrated that testicular GGNBP2 is indispensable for spermatogenesis. 4,14 Ggnbp2 knockout (Ggnbp2KO) males are infertile and show extensive defects in spermatogenesis. We previously discovered that Ggnbp2 resection in mice caused dramatic structural defects during spermatid metamorphosis and exhibited a complete absence of mature spermatozoa phenotype. 4 Given the evidence that GGNBP2 immunoreactivity was strongest in pachytene spermatocytes 4,13 and the number of spermatids started to decline at the initial stage despite Ggnbp2 null mutants having a complete continuous process of spermatogenesis, 4 which raises the possibility that Ggnbp2 loss may compromise spermatocyte meiosis during spermatogenesis.
Meiotic homologous recombination of spermatocytes is important for proper chromosomal segregation during spermatogenesis to generate haploid spermatids and is highly regulated to ensure at least one crossover per homolog pair. Meiotic division is induced by programmed DNA double-strand breaks (DSB) that form at preferred sites or hotspots. [15][16][17] A small number of DSBs form crossover recombinations, and the majority of DSBs are repaired as non-crossover recombinations in spermatocytes. 17,18 These processes involve a number of meiotic-specific and ubiquitously expressed nuclear proteins, such as cH2AX, RAD51, RAD18, DMC1, BRCA1 and MLH1. [19][20][21][22] These proteins are co-ordinately recruited to DSB sites to facilitate DNA DSB repair via homologous recombination. Defects in DNA DSB repair can cause chromosome mis-segregation underling aneuploidy-related pathologies or meiotic arrest of spermatocytes leading to male infertility. 17,20,23,24 Gametogenetin (GGN) is a testis-enriched germ cell-specific protein and has three splice variants. 25 The largest variant GGN1 is demonstrated to be the predominant form in the testis and localizes in spermatocytes, spermatids and spermatozoa but not in spermatogonia in mouse and human testes. [25][26][27] GGN1 is identified to interact with various proteins, which include FANCL, FANCD2, BRCC36, OAZ3, CRISP2, GGNBP1 and GGNBP2. 26,[28][29][30] As such, GGN1 is proposed to play a role in meiosis involving in DNA DSB repair 25,27,30 as well as sperm tail development and/or motility in the testis. 29 Complete deletion of Ggn leads to embryonic lethality before implantation. Haploinsufficiency of Ggn in mouse pachytene spermatocytes results in an increased incidence of unrepaired DSBs, which supports that GGN1 plays a role in spermatogenesis, especially in DNA DSB repair during spermatocyte meiosis. 30 This study aims to determine whether meiotic progression of spermatocytes is compromised in Ggnbp2KO mice and provides further biological insight into the function of GGNBP2 in spermatogenesis. Here, we report that in addition to its indispensability in spermatid transformation, GGNBP2 plays a role in DNA DSB repair during spermatocyte meiosis, which is likely mediated by interaction with GGN1 to modulate its nuclear translocation and consequently affect the DNA DSB repair activity.

| Mice and cell line
The Ggnbp2 Knockout (Ggnbp2KO) mice with a mixed C57BL/6/ 129/Sv genetic background were generated as previously described. 4 Mouse tail genomic DNA was extracted using ZR genomic DNA-tissue mini prep kits according to the procedure recommended by the manufacturer (Zymo Research Corp, Irvine, CA, USA) and genotyped by PCR using primer sets as listed in Table 1  USA). The day before Ggnbp2 or Ggn plasmid transfection, GC-2spd cells were seeded into a 6-well plate at a density of 2 9 10 5 per well. Prior to transfection, the growth medium was replaced with fresh antibiotic-free growth media (Sigma). 300 lL plasmid DNA/ UltraCruz transfection reagent complexes (Santa Cruz Biotech) were added to each well, followed by incubation overnight. The medium was replaced by the growth medium 24 hours post-transfection.

CRISPR knockout plasmids
Three days after transfection, the medium was replaced with fresh medium containing 4 lg puromycin/mL (Sigma). After 2 weeks of puromycin selection, green fluorescent protein (GFP) positive and puromycin-resistant clones were dispersed by trypsin digestion (Sigma) and transferred to 96-well plates at a density of 1-2 cells per well. The procedure of single cell cloning was repeated three times to obtain 100% clonal purity. The stable Ggnbp2 and Ggn deleted cell lines (Ggnbp2 À/À and Ggn À/À ) were verified by RT-PCR and Western blotting. GC-2spd cells transfected with a control double nickase plasmid that contains a non-targeting scramble guide RNA was used for a control.

| Propidium iodide staining and flow cytometric assay
For flow cytometric analysis, testicular germ cell suspension was prepared. Briefly, the decapsulated testes were incubated in phosphate- and Ggn À/À GC-2spd cells colonies.

| RT-PCR
Total RNA was extracted from the testes using TRIzol reagent (Invitrogen) according to the manufacturer's instructions. Two microgram total RNA was reverse transcribed to cDNA with random primers (Invitrogen) and avian myeloblastosis virus reverse transcriptase (MP Biochemicals, Santa Ana, CA, USA). The cDNA was amplified by PCR with the primer sets of the target gene and housekeeping genes. PCR cycle consisted of denaturation for 45 seconds at 94°C, annealing for 1 minutes at 57°C and extension for 1 minutes 30 seconds at 72°C.
The amplified products were separated by electrophoresis in 1% agarose gels and stained by ethidium bromide. The intensity of specific bands was scanned and semiquantified using the image analysis software Total Lab (Nonlinear USA Inc., Durham, NC, USA). The results were presented as the ratio of the target gene over a housekeeping gene, b-actin or Rpl19. All RT-PCR primers, as listed in Table 1, were designed according to the sequences obtained from GenBank and synthesized by Operon Technologies (Alameda, CA, USA).

| Preparation of testicular germ cells and nuclear spread slides and immunofluorescence staining
Dispersed testicular germ cells were prepared by incubation of the seminiferous tubules in PBS supplemented with 0.5% bovine serum albumin, 2 mg/mL collagenase, 0.1 lg/mL DNase I and 0.5% trypsin (Sigma) for 10 minutes at 32°C. The dissociated cells were then resuspended in 4% paraformaldehyde, washed with PBS, spread on slides and air-dried.
Nuclear spreads were prepared by the "drying down technique" as previously described. 31 Briefly, the seminiferous tubules were isolated from 28-day-old testes. The germ cells were dissociated from the seminiferous tubules by repeatedly up-down pipetting in PBS.

| Transient transfection of Ggnbp2 expression plasmids
The Ggnbp2 cDNA fragment was obtained by PCR using testis cDNA as the template and validated by DNA sequencing. The correct fulllength Ggnbp2 cDNA was cloned into pcDNA3-6HisC vector (Invitrogen) to construct the mammalian expression plasmid pcDNA3-HisC tagged-Ggnbp2 (pcDNA3-Ggnbp2). Ggnbp2 À/À or Ggn À/À GC-2spd cells were cultured in DMEM medium to approximately 70% confluence in a 6-well plate and were transfected with 2 lg of either pcDNA3-Ggnbp2 or pcDNA3 plasmid (as a control) with a plasmid transfection reagent (Santa Cruz Biotech). The cells were cultured in the same medium for 72 hours, then harvested for Western blot and flow cytometric assays.

| Statistics
All statistical analyses were performed with a version 3.06 Instat program (GraphPad Software, San Diego, CA, USA). Comparisons of means between more than two groups were performed by one-way analysis of variance (ANOVA). Student's t test was used when comparing means of two groups. A P-value <.05 was considered statistically significant.

| Ggnbp2KO leads to aberrant distribution of testicular germ cells
We previously reported that Ggnbp2KO resulted in an absence of mature spermatozoa and a significant reduction in spermatids in the adult testes using morphometric analysis methods. 4 In this study, we examined the DNA contents of testicular germ cells isolated from 30-and 60-day-old WT and Ggnbp2KO mice using flow cytometry.
Ggnbp2KO mice showed markedly decreased haploid (1C) cell fraction and increased tetraploid (4C) cell fraction compared to agematched either 30-or 60-day-old WT littermates ( Figure 1A,B). In agreement with this data, RT-PCR analysis of 60-day-old Ggnbp2KO testes showed a significant decrease in both round and elongated spermatid markers Prm2 and Acrv1 ( Figure 1C,D).

| Ggnbp2KO does not impact chromosome synapsis, but results in compromised meiotic DSB repair
To determine whether deletion of Ggnbp2 causes meiotic defects, we examined the dynamics of spermatocyte chromosome synapsis and recombination using antibodies directed against the synaptonemal complex proteins, axial/lateral element SYCP3 and central element SYCP1, which are main components of the meiotic nodules and chromatin markers. 32 The labelling pattern of SYCP1 and SYCP3 was colocalized and the mutant spermatocytes were indistinguishable from that of WT littermates (Figure 2A), suggesting that Ggnbp2 We next examined whether Ggnbp2KO affected meiotic DSB repair. cH2AX is a meiotic marker that rapidly appears in the form of nuclear foci concomitantly with DNA DSB formation and repair. 33,34 In WT spermatocytes, globally distributed cH2AX foci were observed during the leptotene, correlating with DNA DSB formation, then diminished as breaks were repaired during the zygotene, eventually concentrated in the bivalent XY body by the pachytene and diplotene stages ( Figure 2B). Although staining pattern of cH2AX in Ggnbp2KO spermatocytes in the leptotene stage was similar to that of WT, its staining persisted not only in the XY body but also in several patches along autosomes in the pachytene and diplotene stages ( Figure 2B). Pachytene cells that contain more than one cH2AX foci in Ggnbp2KO spermatocytes were significantly higher than that of WT ( Figure 2C). In addition, we checked the distribution of several DNA DSB repair and meiotic recombination marker molecules RAD51, RAD18, BRCC36 and MLH1. The immunostaining pattern showed that RAD51 foci per pachytene cell on the XY body were not significantly different between the two groups. In contrast, a marked increase in autosomal RAD51 foci was observed in the Ggnbp2KO spermatocytes ( Figure 2D,E). In WT spermatocytes, RAD18 was accumulated as RAD51-like foci on the synaptonemal complexes only seen in the zygotene stage and was restricted to the XY body in the pachytene stage ( Figure 2D). In Ggnbp2KO pachytene spermatocyte, RAD18 staining was sustained in synapsed F I G U R E 2 Ggnbp2 deletion does not affect synapsis but compromises the repair of meiotic DNA double-strand breaks (DSB). A, Nuclear spreads of spermatocytes stained with antibodies against SYCP1 (red) and SYCP3 (green) show that the assembly and disassembly of the synaptonemal complexes are not affected. B, Chromosome spreads of spermatocytes are stained with antibodies against SYCP3 (red) and cH2AX (green). Four meiotic stages of spermatocyte, leptotene (lepto), zygotene (zygo), pachytene (pachy) and diplotene (diplo) are presented. D, Chromosome spreads of pachytene spermatocytes are stained with SYCP3 antibody (red), DNA DSB repair protein antibodies RAD51, RAD18 and BRCC36 and a crossover marker MLH1 (green) as indicated. Arrows indicate the XY body stained by cH2AX, RAD51 and RAD18. C,F, The percentage of pachytene cells with >1 cH2AX (C) and >1 RAD18 foci (F) are significantly higher in Ggnbp2KO than WT mice, which is calculated from 100 randomly selected pachytene spermatocytes per testis and data are from three 60-d-old testes. E,G,H, The numbers of RAD51 (E), BRCC36 (G) and MLH1 (H) foci counts per pachytene cell are significantly higher in Ggnbp2KO than WT mice, which is counted from 50 randomly selected pachytene spermatocytes per testis and results are from three 60-d-old testes. Data are presented as the mean AE SEM (**P < .01, scale bar = 5 lm) autosomal homologs in addition to the XY body ( Figure 2D,F).
BRCC36 is a major component of the BRCA1-A complex, which participates in promoting BRCA1 function in response to DNA damage. 35 Immunofluorescence staining revealed that BRCC36 foci dramatically increased ( Figure 2D,G) and MLH1 foci, a marker of chiasma formation, significantly decreased in Ggnbp2 null pachytene cells ( Figure 2D,H).

| GGNBP2 interacts with GGN1 and both have the same subcellular localization during spermatogenesis
It has been reported that GGNBP2 interacts with testicular proteins, GGN1, GGNBP1, CRISP2 and OAZ3. 26,[28][29][30] Reciprocal immunoprecipitation (IP) with GGN ( Figure 3A) and GGNBP2 ( Figure 3B) antibodies demonstrated that GGN1 and GGNBP2 formed a protein complex in the testis. Our previous use of immunohistochemical staining showed that GGNBP2 predominately localized in spermatocytes and round spermatids, 4 and GGN1 was reported to detect in spermatocytes, spermatids and sperm. 28 To determine subcellular localizations of GGNBP2 and GGN1 in spermatocytes, spermatids and spermatozoa, dispersed testicular cells were stained by immunofluorescence. Both GGNBP2 and GGN1 were observed in the nucleus and cytoplasm of spermatocyte and they colocalized very well with each other (Figure 3C). In round and elongating spermatids, both GGNBP2 and GGN1 were detected in the nucleus and cytoplasm as well as acrosome. Immunofluorescence signal of these two proteins was found in the tail of spermatozoa in addition to the nucleus and acrosome (Figure 3D). These results suggest GGNBP2 shares similar trafficking dynamics with GGN1 during spermatogenesis. testes. Both mRNA ( Figure 4A,C) and protein ( Figure 4B,C) levels in the Ggnbp2KO testes were noticeably reduced compared to WT testes. Immunofluorescence staining of GGN1 in nuclear spread testicular germ cells revealed that GGN1 immunostaining localized in the cytoplasm surrounding the nucleus and GGN1 foci were present in the nucleus associated with chromosomes in WT spermatocytes ( Figure 4E). Overall immunofluorescence intensity of GGN1 in

| Ggnbp2KO results in a decrease in
Ggnbp2KO spermatocytes was remarkably reduced ( Figure 4E).
GGN1 foci that associated with chromosomes in the nucleus were significantly lower than that of WT spermatocytes ( Figure 4E-G).

| Ggnbp2 and Ggn knockout share the similar phenotype in GC-2spd cells
Immortalized GC-2spd cell line is originally derived from mouse spermatocytes. 36 It has been shown that these cells are capable of undergoing meiosis to produce a haploid population, expressing some post-meiotic marker genes and possessing morphological features in common with early spermatids following low-temperature culture (32°C). 37 To evaluate the role of GGNBP2 and GGN1 in spermatocyte meiotic differentiation, we generated stable Ggnbp2 and Ggn knockout (Ggnbp2 À/À and Ggn À/À ) GC-2spd cells using CRISPR/Cas9 technique, respectively. Complete abolishment of Ggnbp2 and Ggn expression was verified by RT-PCR and Western blot ( Figure 5A,B). Flow cytometric analysis of the DNA contents showed that either Ggnbp2 or Ggn deletion resulted in a significant reduction in haploid cells (1C) compared to Ggnbp2 +/+ and Ggn +/+ counterparts, while the tetraploid (4C) cells increased and diploid cells (2C) were not significantly changed ( Figure 5C,D). Both of the haploid spermatid marker genes, Prm2 and Acrv1, were noticeably down-regulated, while the spermatocyte marker gene Pms2 was increased in both Ggnbp2 À/À and Ggn À/À GC-2spd cells. The same was observed in the testis, Ggnbp2 loss in GC-2spd cells also impeded Ggn expression ( Figure 6A-D).
We next asked whether re-expression of Ggnbp2 would rescue the phenotype of Ggnbp2 À/À and Ggn À/À GC-2spd cells. The results showed that GGNBP2 protein was readily detected in Ggnbp2 À/À GC-2spd cells following transfection of Ggnbp2 expression plasmids ( Figure 6E). Interestingly, an increased GGN1 protein level was observed in overexpressed Ggnbp2 À/À GC-2spd cells ( Figure 6E). In our Ggnbp2KO model, induction of DNA DSBs appears to be generated on time, which is manifested by accumulation of cH2AX at leptotene and zygotene stages. 19 Synapsis and meiotic sex chromosome inactivation seem to occur normally. However, inappropriately accentuated and persistent autosomal cH2AX staining in pachynema and even diplonema point to DNA DSB repair defect as one primary cause of meiotic impairment in the absence of GGNBP2. This is further buttressed by the evidence that commensurately increases in cytological markers for DSB repair proteins RAD51, RAD18 and BRCC36 in meiocytes. BRCC36 is a major component of the BRCA1-A complex, which participates in promoting BRCA1 function in response to DNA damage. 35 The RAD18 is expressed highest levels in the testis. 41 Its recruitment to meiotic DSBs is thought to occur if these breaks persist. 21 Hence, a decrease in MLH1 focus formation may possibly be caused by the inefficient DSB repair that likely promotes an aberrant activity of alternative DNA repair via the non-homologous end joining pathway 17 and results in a reduction in crossovers in the absence of GGNBP2.
F I G U R E 6 Enforced expression of Ggnbp2 in Ggnbp2 À/À GC-2spd cells restores Ggn expression and increases haploid differentiation. A-D, Ggnbp2 deletion decreases Ggn mRNA (A,B) and protein expression (C,D) in three stable Ggnbp2 À/À GC-2spd cell clones. E, Ggnbp2 À/À GC-2spd cells are transfected with pCDNA3-HisC empty vectors (EV) or pCDNA3-His-Ggnbp2 expression plasmids (OE). Ggnbp2 overexpression increases Ggn protein determined by Western blot. F,G, Representative histograms (F) and quantitation (G) of flow cytometric assays show Ggnbp2 overexpression increases the number of haploid cells in Ggnbp2 À/À but not in Ggn À/À GC-2spd cells. The experiments are repeated three times, and data are presented as mean AE SEM. (**P < .01, *P < .05) GGN1 is known to contain nuclear targeting signalling and GGNBP2 has a zinc finger motif in the N-terminus and a nuclear protein-binding domain in the C-terminus. 3,9,25 Current and previous studies demonstrate that GGNBP2 and GGN1 have a similar testicular expression profile and interact specifically in the testis. The highest levels of these two proteins are found in meiotic spermatocytes and persisted through spermatids and displayed a developmentally translocation between the nucleus and cytoplasm during meiosis and spermiogenesis. Ggnbp2 loss profoundly reduces GGN1 nuclear localization suggesting that binding of GGNBP2 to GGN1 may be necessary for nuclear accumulation. There is the possibility that GGNBP2 regulates the distribution, thus the activity of GGN1.
Intriguingly, Ggn expression is down-regulated in Ggnbp2 null testes implying that GGN1-GGNBP2 complex may be also required for Ggn More studies will help us to understand the involvement of GGNBP2 and GGN1 in the process of DNA DSB repair and to elucidate the nature of the interaction among GGNBP2, GGN1, FA and BRCA proteins to control meiotic prophase progression.
Given that GGNBP2 is a conserved zinc finger protein in human and mouse, 3,4,13 it presumptively has similar functions in regulation of spermatogenesis in humans. A recent comprehensive analysis revealed that sperm counts continually and significantly declined over a 38-year period in men from Western countries. 42 Numerous reports have provided evidence that a decline in sperm count links to chronic exposure to environmental endocrine disrupting chemicals (EDCs). 43,44 Dioxins are a group of highly persistent and omnipresent EDCs. 45 Decrease in sperm count is one of the most sensitive deteriorations of dioxin toxicity in both human and experimental studies. 45,46 However, the mechanism by which dioxins impair spermatogenesis remains elusive. The expression of GGNBP2 has been demonstrated to be regulated by dioxin and GGNBP2 plays critical roles in spermatocyte meiosis and spermatid differentiation. 4,13 Thereby, it would be worthwhile in future studies to investigate whether GGNBP2 is a factor that associates with dioxininduced reproductive toxicity by interfering with spermatogenesis.
In summary, the present data indicate that GGNBP2, together with GGN1, plays a regulatory role in germ cell meiosis. The results demonstrate that loss of Ggnbp2 increases DNA DSBs and compromises DSB repair during male meiosis, suggesting that male infertility in Ggnbp2KO mice is partially attributed to the role of GGNBP2 in spermatocyte meiosis. Our findings provide a framework for further molecular studies of GGNBP2 in spermatogenesis.

ACKNOWLEDG EMENTS
The authors are thankful for Ms. Robin J. Lawrence's helpful advice in the preparation and editing of this manuscript.

CONFLI CT OF INTEREST
The authors declared no potential conflict of interests with respect to the research, authorship and/or publication of this article.