Mutation of S461, in the GOLGA3 phosphorylation site, does not affect mouse spermatogenesis

Animals

Mice were procured through the Experimental Animal Center of Nanjing Medical University, and subsequently randomly assigned to ventilated cages each housing four to five mice and were maintained in a standard facility under regulated conditions (20–22 °C; 50–70% humidity; 12-h/12-h dark/light exposure), with ad libitum water and food. The animals used in the experiments were over eight weeks old with mature male reproductive systems and were in good health. The mice were euthanized with CO₂ upon study completion. This investigation received approval through the Institutional Animal Care and Use Committees of Nanjing Medical University (Approval No. IACUC-2004020-3), and carried out in line with relevant professional recommendations and the Guide for the Care and Use of Laboratory Animals.

Mutation plasmid construction

Point mutant plasmids corresponding to the predicted GOLGA3 phosphorylation sites (Table 1) from the gnomAD database were constructed through designing forward and reverse primers (Table S1) to introduce the mutated base into the complete cDNA sequence leading to missense mutation of serine. PCR amplified complete cDNA encoding recombinant GOLGA3 which was then cloned into a pcDNA3.1 (+) vector (Thermo Fisher, Waltham, MA, USA) carrying a FLAG label. Next, the recombinant plasmid DNA was transformed into E.coli followed by extraction using the EndoFree Mini Plasmid Kit^® (DP118, Tiangen, Beijing, China) in accordance with the provided directions.

Cell culture

The HEK293T and HeLa cellular cultures were procured through ATCC and expanded within DMEM together with high-glucose, 10% FBS (Gibco, USA), and penicillin / streptomycin (100 U/ml, Thermo Fisher, USA). Transfection was performed with Lipofectamine^® 2000 (11668019, Thermo Fisher, USA), following the provided protocol.

Golga3 mutation mouse generation

Cytosine base editors were used to generate Golga3^S461L/S461L mice using BE3 sgRNA for the introduction of the missense mutation. The sequences of the BE3 sgRNA and donor were 5′-AGACTCCCTTAGTTCTGAAGTGG-3′ and 5′GGTGGAACACAGCCACAGCAGCCAGCAGAAGCAAGACTTGCTTAGTTCTGAAGTGGACACTTTGAAGCAGTCTTGCTGGGATCTAGAGCGGGCCATGACT-3′respectively. After separate annealing and ligation of the cDNA sgRNA oligos onto BsaI-digested pUC57-T7-sgRNA vector, with sgRNA templates amplified by PCR using the Trans-PCR Forward (GAAATTAATACGACTCACTATAGG) and Reverse (AAAAGCACCGACTCGGTGCCA) primers, with the aid of MinElute PCR^® Purification Kit (28004, Qiagen, Germany). The sgRNA was prepared through MEGAshortscript^® (AM1354, Ambion, USA). Purification was performed with a MEGAclear Kit (AM1908, Ambion). After linearization of the BE3 plasmids (Addgene No. 73021) with AgeI and PmeI and purification as described above, BE3 mRNA was generated using a mMACHINE^® T7 Ultra Kit (AM1345, Ambion) with subsequent purification through RNeasy Mini^® (74104, Qiagen). This BE3 mRNA (50 ng/µL) was then used for inoculating mouse zygotes together with DonorDNA (50 ng/µL) and, BE3 sgRNA (20 ng/µL), before implantation in pseudo-pregnant mice. Tissue samples were collected from the toes of the seven-day-old offspring, followed by extraction of DNA from the tissue using Mouse Direct PCR^® (B40013, Biotool, USA). and amplified through primers (Golga3^S461L/S461L mice: F5′-ACCAGCCAGGATGTCTAT-3′, R5′-AGTAGGAAGGAGAGCAGAA-3′) and PrimeSTAR^® HS DNA Polymerase (DR010A, Takara, Japan), adopting the following PCR protocol, consisting of 95 °C (five minutes), followed by 35 cycles (95 °C for thirty seconds; 62 °C (−0.2°/cycle) for thirty seconds; 72 °C for thirty seconds) with a final step of 72 °C for five minutes. The products were sequenced by Sanger sequencing.

Dephosphorylation assays

The transfected cells containing the constructed mutant plasmids and plasmids with the full-length cDNA sequence were grown for 36 h and lysed using SDS Medicon RIPA lysis buffer without SDS containing 1x complete^® EDTA-free Protease Inhibitor Cocktail (Roche, Switzerland). After sonication, proteomic mixtures were placed on a rotary shaker for 45 min at a 20–30 rpm/min followed by high-speed centrifugation at 4 °C and measurement of protein concentrations in the supernatants. Mutant and normal proteins were placed into incubation, while exposed to calf intestinal alkaline phosphatase (CIP) (180 min; 37 °C water-bath); controls were not treated with CIP. 6 × loading buffer was introduced with dephosphorylated proteins and the samples underwent denaturing within a metal bath (95 °C; ten minutes).

Western blotting

Cell lysates were prepared as above followed by electrophoresis of the proteins on SDS-PAGE, with consequent transport onto PVDF membranes. Blots underwent blocking (120 min; ambient temperature) using 5% skimmed milk within TBS-T (20 mM Tris, 150 mM NaCl, 0.1% Tween 20) before night-time incubation (4 °C) with anti-GOLGA3 (21193-1-AP, Proteintech, Waltham, MA, USA; 1: 5000) together with anti-p-CDK1(ab201008, Abcam, UK; 1: 5000). Following three TBS-T washing cycles, proteomic ontent samples were exposed (2 h, room temperature) to 2° antibodies (1:5000), evaluated visually through High-sig ECL^® Western Blotting Substrate (Tanon, China).

Genotyping

Genomic DNA deriving through murine tail-tissue samples for the detection of mutations using PCR amplification and standard Sanger sequencing. The Golga3-edited mouse founders (F0) were identified using F-Primer 5′-ACCAGCCAGGATGTCTAT-3′ and R-Primer 5′-AGTAGGAAGGAGAGCAGAA-3′ together with PCR amplification (Fast-Taq Master Mix^®, Vazyme, China). Amplicon was sub-cloned into the pMD19-T plasmid (TaKaRa, Wuhan, China), sequenced by Sanger sequencing. Founders showing base mutations at amino acid 461 of Golga3 were crossed through WT C57BL/6 murines (3–4 generations) for circumventing off-targeted editing and resulting in the production of pure heterozygous animals. Sanger DNA sequencing of Golga3^S461L/S461L mice was carried out, with SnapGene (version 1.1.3) was used to plot the results. Each mouse represented a unique label, including genotype, sex, and birth date.

Fertility test

The fertility of male mice was assessed through crossing male Golga3^S461L/S461L mice with female WT C57BL/6 mice; male Golga3^+/+ mice were used as controls. The formation of vaginal plugs in the female mice was checked each morning. Information about each litter was recorded including birth date and the number of pups. The experimental unit requires an adult wild type mouse and mutant male mouse each in eight weeks. All experiments were performed in triplicate.

Analysis of epididymal sperm

Sperm from Golga3^S461L/S461L and WT mice were assessed through computer-aided sperm analysis (CASA). Parameters for caudal epididymal sperm were determined, especially sperm quantity/ quality. After extrusion, samples were placed within human tubal-fluid culturing medium (In Vitro Care, Frederick, USA) and maintained at fixed temperature (37 °C) through a themostatic water bath. Sperm motility, progressive motility together with quantitative analysis for 10-µl aliquots of the sperm suspensions were assessed by CASA (Hamilton-Thorne Research, Beverly, MA, USA). An experimental unit requires an adult wild type mouse and mutant male mouse each in eight weeks. Experiments were performed in triplicate for ensuring statistical analysis robustness.

Histologicy

Testis/epididymal tissue samples were collected through at least three Golga3^S461L/S461L murines, together with three Golga3^+/+ male murines. Tissue samples were fixed overnight within Davidson’s fluid (MDF) and placed in 70% ethanol, followed by dehydration in an 80–90–100% ethanol gradient, paraffin-embedding, and sectioning (5-µm). Epididymal tissue underwent hematoxylin and eosin (HE) staining, while testis tissue underwent periodic acid Schiff (PAS) staining (395B, Sigma-Aldrich, USA). Smears of cauda epididymal sperm were also made, undergoing fixing with 4% paraformaldehyde within PBS (thirty minutes), rinsing, and HE staining.

Immunofluorescence analysis

Paraffin sections of testicular tissue from wild-type and Golga3^S461L/S461L mice were deparaffinized and rehydrated. Antigen retrieval of sperm smears and slides of rehydrated testis tissue was undertaken by boiling samples within sodium citrate buffer (10 minutes). The slides were cooled, treated with 3-fold PBS wash-cycles (5 minutes/cycle), followed by blocking through 1% BSA within PBS (120 min at ambient temperature) prior to probing through 1° antibodies (night-time; 4 °C) (Table S2). Following this step, segments were subjected to 3 × PBST wash-cycles (0.05% Tween 20 within 1 × PBS) followed by assessment through 2° antibodies (Table S3) (120 min; ambient temperature). Nuclei were counterstained through Hoechst 33342 (Invitrogen, Waltham, MA, USA) (five minutes). After mounting with glycerin, segments were imaged through LSM800^® confocal fluorescence microscopy (Carl Zeiss AG, Jena, Germany).

TUNEL assay

Quantitative analyses for apoptotic cells required TUNEL BrightRed^® Apoptosis Detection Kit (Vazyme, China), following the provided directions. After incubation with the BrightRed Labeling Buffer mixture, whereby wild-type and Golga3^S461L/S461L murine samples were treated to 3 × PBS wash-cycles (5 min each) in the dark,, stained with Hoechst 33342 (5 min, room temperature), mounted with glycerin, and examined and imaged under fluorescence microscopy as above.

Statistical analysis

At least three replicates per experiment were performed. Datasets reflected mean ± the standard error. Cohorts were compared through one-way ANOVAs and unpaired two-tailed t-tests. P < 0.05 was deemed to confer statistical significance. The gray values of Western blot bands and mean fluorescence intensities were analyzed by Image J software. Microsoft Excel^® and GraphPad Prism^® 6.0 were employed for all such evaluations.

Golga3 is conserved in various species and mutation of its phosphorylation site leads to abnormal localization of GOLGA3 in cells

The GOLGA3 protein consists of a coiled-coil domain in the C-terminal region and two reported functional domains (domains for interaction with GOPC and Golgi-targeting) in the N-terminal region (http://www.uniprot.org/uniprot/Q08378) (Fig. 1A). Protein sequence alignment showed that Golga3 was conserved in human, mouse, rhesus monkey, Xenopus laevis and zebrafish (Fig. 1B). Seven phosphorylation sites have been identified in the GOLGA3 protein and a further phosphorylation site inferred based on sequence similarity and mass spectrometry data (Bian et al., 2014; Olsen et al., 2010; Zhou et al., 2013). Mutations in the human GOGLA3 gene were searched using the Ensembl database (https://asia.ensembl.org/Homo_sapiens/Gene/Variation_Gene), specifying missense mutations as a filtering criterion to obtain the frequency of such mutations at phosphorylation sites in the human gene corresponding with those in the mouse Golga3 gene. By combining this information with that of conserved phosphorylation sites, five potential phosphorylation sites (Table 1; Figs. 1A and 1B) were identified. Constructed mutant plasmids with a flag tag were then transfected into HeLa cells prior to immunofluorescence co-localization staining with GM130 protein. It was found that GOLGA3^S465L showed reduced localization to the Golgi apparatus, with part of the protein dispersed within the cytoplasm. In contrast, GOLGA3^S272F, GOLGA3^S385R, GOLGA3^S389G and GOLGA3^S983P showed similar signals to the wild-type GOLGA3 protein (Fig. 1C). The fluorescence intensity of the GOLGA3 protein was subsequently quantified by analyzing gray values and statistical analysis of the mean fluorescence intensity in single channels, confirming the reduced localization of the protein to the Golgi apparatus (Fig. 1D). The S465 phosphorylation site in the human sequence corresponded to the S461 phosphorylation site in mouse. This site is located in the coiled coil domain and is strongly conserved in different species. To verify whether S461 was indeed phosphorylated, we transfected the constructed mutant plasmids and control plasmids into HEK293T cells. The mutant and normal proteins were then treated with CIP or were left untreated (Control) before western blotting. Quantification of the gray values of the protein bands showed a reduced band shift in the dephosphorylated GOLGA3^S465L, confirming that S461 was phosphorylated (Fig. 1E).

Golga3^S461L/S461L mice have normal fertility and normal spermatozoa

To investigate the effects of mutation of the GOLGA3 phosphorylation site in the testes of C57BL/6 mice, a missense mutation was generated at S461 in exon 7 of the Golga3 gene, in super-ovulated fertilized eggs using cytosine base editors (Fig. 2A). Compared with wild-type mice, the two cytosine bases responsible for encoding the serine at position 461 were simultaneously converted to thymine and adenine, respectively, in Golga3^S461L/S461L mice, and was confirmed by both Sanger sequencing and PCR genotyping (Fig. 2B).

It was found that male Golga3^S461L/S461L mice nevertheless exhibited normal development, including being viable and fertile. The testis morphology was also normal in adult Golga3^S461L/S461L mice, with no significant changes in the testis weight of adult Golga3^S461L/S461L mice (N = 3). The p-values for the average number of pups per litter as well as the average testis weight/body weight were 0.8298 and 0.6005, respectively, determined by t-tests performed with GraphPad Prism 8 software (Figs. 2C, 2D and 2E). Computer-assisted sperm analysis (CASA) was then used to evaluate semen quality of adult Golga3^S461L/S461L mice, including parameters such as the sperm motility ratio and progressive sperm ratio. In this case, the p-values for the latter two parameters were 0.7456 and 0.7182 respectively (Figs. 2F and 2G). In particular, the ratio of normal sperm in male Golga3^S461L/S461L mice was found to be similar to that of the control group, with no significant changes also noted in progressive motility (Fig. 2H).

Golga3^S461L/S461L mice exhibited normal spermatogenesis

Normal spermatogenic cells were observed in the PAS-stained seminiferous tubules of male adult Golga3^S461L/S461L. Furthermore, histological analysis of each phase of spermatogenic tubule revealed the presence of post-meiotic round spermatids in the seminiferous tubules of adult Golga3^S461L/S461L mice in contrast to the Golga3^+/+ mice (Fig. 3A). Normal numbers of spermatids and sperm were also observed in H&E-stained samples from these edited mice (Figs. 3B, 3C, 3D and 3E). Furthermore, no significant differences in sperm morphology were observed between the adult Golga3^S461L/S461L mice and wildtype mice based on the immunofluorescence staining of sperm smears (Fig. 3F). PLZF, γ-H2AX, PNA and Sox9 fluorescence signals were used to specifically identify proliferating spermatogonia, spermatocytes, spermatids, and Sertoli cells, respectively. The fluorescence images showed normal spermatogenesis in Golga3^S461L/S461L mice (Figs. 4A, 4B, 4C and 4D). Furthermore, to investigate the apoptosis of germ cells in spermatogenic tubules, TUNEL analysis of testicular sections was also performed. While sporadic dark orange fluorescence signals, indicative of apoptotic cells, were visible in the seminiferous tubules, these signals did not differ significantly between the Golga3^S461L/S461L and wild-type mice(Figs. 4E and 4F). Further t-test analysis showed that the p-values for the number of apoptotic cells/total tubules as well as the number of tubules/total tubules were 0.5070 and 0.0132, respectively (P > 0.05) indicating that there were no significant differences between the mutated and wild-type mice in terms of apoptosis in the seminiferous tubule (Figs. 4G and 4H, Table S4).

Immunofluorescence co-localization and GOLGA3 protein levels are normal in Golga3^S461L/S461L mice

During the early stages of spermatogenesis, the Golgi vesicles fuse to form sperm proacrosomal vesicles. The GOLGA3^S461L protein showed reduced Golgi localization, with some protein dispersed within the cytoplasm. To better understand the location of the GOLGA3 protein in Golga3^S461L/S461L mice, immunofluorescence co-localization was performed on mouse testicular sections. It was found that during stages 1–4 of spermatogenesis, GOLGA3 co-localized with Wheat Germ Agglutinin (WGA) (Fig. 5A), although there were no significant differences in fluorescence localization between mutant mice and wild type mice. Finally, to explore whether the S461L mutation affected the level of protein expression, western blotting and quantification of the protein bands were undertaken. This showed that the size of the GOLGA3^S465L band was similar to that of the wild-type, with no significant differences in the gray values between the two (Figs. 5B–5C).