Targeted Disruption of Glycogen Synthase Kinase 3a (Gsk3a) in Mice Affects Sperm Motility Resulting in Male Infertility

ABSTRACT

The signaling enzyme glycogen synthase kinase 3 (GSK3) exists as two isoforms—GSK3A and GSK3B. Protein phosphorylation by GSK3 has important signaling roles in several cells. In our past work, we found that both isoforms of GSK3 are present in mouse sperm and that catalytic GSK3 activity correlates with motility of sperm from several species. Here, we examined the role of Gsk3a in male fertility using a targeted gene knockout (KO) approach. The mutant mice are viable, but have a male infertility phenotype, while female fertility is unaffected. Testis weights of Gsk3a^−/− mice are normal and sperm are produced in normal numbers. Although spermatogenesis is apparently unimpaired, sperm motility parameters in vitro are impaired. In addition, the flagellar waveform appears abnormal, characterized by low amplitude of flagellar beat. Sperm ATP levels were lower in Gsk3a^−/− mice compared to wild-type animals. Protein phosphatase PP1 gamma2 protein levels were unaltered, but its catalytic activity was elevated in KO sperm. Remarkably, tyrosine phosphorylation of hexokinase and capacitation-associated changes in tyrosine phosphorylation of proteins are absent or significantly lower in Gsk3a^−/− sperm. The GSK3B isoform was present and unaltered in testis and sperm of Gsk3a^−/− mice, showing the inability of GSK3B to substitute for GSK3A in this context. Our studies show that sperm GSK3A is essential for male fertility. In addition, the GSK3A isoform, with its highly conserved glycine-rich N terminus in mammals, may have an isoform-specific role in its requirement for normal sperm motility and fertility.

INTRODUCTION

Glycogen synthase kinase 3 (GSK3), a serine/threonine protein kinase, is a key component of a large number of cellular processes, including glucose regulation, inflammation, and immune responses, proliferation, migration, and apoptosis (see reviews [1, 2]). GSK3 is expressed in virtually all mammalian tissues, and is encoded by two genes that generate two related proteins: GSK3A and GSK3B. GSK3 has been implicated in a variety of diseases, including mood disorders, Alzheimer disease, diabetes, and cancer.

Unlike most protein kinases, GSK3 is constitutively active under resting conditions. In this constitutively active state, GSK3 is phosphorylated on Tyr279 (GSK3A) or Tyr216 (GSK3B); this phosphorylation is most likely catalyzed by GSK3 itself (auto-phosphorylation) [3]. Since GSK3 is active under resting conditions, its regulation occurs through inhibition or the retargeting of its activity. This inhibition of activity occurs via phosphorylation on Ser21 (GSK3A) or Ser9 (GSK3B) [2]. Several protein kinases are able to phosphorylate these sites and thereby negatively regulate GSK3 activity, including PKB/Akt, a serine/threonine kinase located downstream of phosphatidyl-inositol 3-kinase, and cAMP-dependent protein kinase (PKA) [4, 5]. Dephosphorylation of GSK3 can be accomplished by protein phosphatase (PP) 1 or PP2A [6]. GSK3 is also unique among protein kinases in that most of its substrates must be prephosphorylated (primed) before being phosphorylated by GSK3 [7, 8]. Thus, two sets of protein kinases are integrated to regulate the actions of GSK3: those that directly phosphorylate GSK3 and those that prime its substrate. In some cases, a single protein kinase can serve both functions. For example, PKA is capable of both phosphorylating GSK3 and priming GSK3 substrates. Thus, the timing and location of these signaling enzymes may combine to contribute to the individualized control of the actions of GSK3 toward each substrate [9]. The constitutively active nature of GSK3 suggests that it may contribute to maintaining steady, resting conditions within a cellular compartment. In accordance with this idea, many GSK3 substrates, like transcription factors, are inhibited in their respective functions when phosphorylated by GSK3. Therefore, deactivation of GSK3 often leads to activation of its substrates.

Both α and β isoforms of GSK3 are present in mammalian spermatozoa [10, 11]. Immotile caput sperm contain six-fold higher GSK3 activity than motile caudal sperm [11]. Serine phosphorylation of GSK3 increases significantly in sperm during their passage through the epididymis [11]. Stimulation of bovine sperm motility by isobutyl-methyl-xanthine, 2-chloro-2′-deoxyadenosine, or calyculin A, is accompanied by a dramatic increase GSK3 serine phosphorylation [11]. In porcine sperm, a parallel increase in serine phosphorylation of GSK3 is observed after treatments that also induce a significant increase in porcine sperm velocity parameters. Therefore, a significant positive correlation among straight-line velocity, circular velocity, average velocity, rapid-speed spermatozoa, and GSK3 serine phosphorylation levels exists. Pharmacological inhibition of GSK3 activity leads to a significant increase in the percentage of rapid and medium-speed spermatozoa, as well as in all sperm velocity parameters. Moreover, pretreatment of porcine spermatozoa with a GSK3 inhibitor significantly increased the percentage of capacitated porcine spermatozoa [12]. These studies suggest a role for the enzyme during fertilization of the egg by sperm.

Tyrosine phosphorylation of GSK3 also correlates with sperm motility. The tyrosine phosphorylation of GSK3 is much higher in motile caudal sperm than in immotile caput sperm [10, 13]. Experimental inhibition of motility in motile caudal sperm by PKA anchoring inhibitors led to the virtual disappearance of tyrosine phosphorylation, while addition of motility activators, IBMX or 8-bromo-cAMP, to activate PKA resulted in an increase in tyrosine phosphorylation of GSK3. Thus, both tyrosine and serine phosphorylation of GSK3 are correlated with increased sperm motility.

An important role for GSK3 in regulation of sperm function is indicated by its greater activity in immotile caput sperm and lower activity in motile caudal sperm. Moreover, tyrosine phosphorylation (which stimulates catalytic activity) and serine phosphorylation of GSK3 (an inhibitory mechanism) both increase significantly in sperm during their passage through the epididymis. The activation of GSK3 by tyrosine phosphorylation, along with an inactivating phosphorylation during epididymal sperm maturation, indicates that the exact signaling role of GSK3 in the testis and sperm is complex and requires clarification. In this report, we examine the role of Gsk3a in sperm maturation and motility in mice using a targeted knockout (KO) approach. Our results show an indispensable role for Gsk3a in male fertility.

MATERIALS AND METHODS

Ethics Statement

All procedures with transgenic and wild-type (WT) mice used in the present study were performed at the Kent State University animal facility, and were approved by the National Institute of Environmental Health Sciences Animal Care and Use Committee and the Kent State Animal Ethics Committee under the Institutional Animal Care and Use Committees protocol number 362DK 13-11.

Generation and Genotyping of Gsk3a Conditional KO Mice

The targeting construct for Gsk3a was created using a recombineering strategy [14]. Briefly, a Roswell Park Cancer Institute (RPCI) bacterial artificial chromosome (BAC) clone containing the murine Gsk3a locus (RPCI-23 228E7), obtained from Children's Hospital of Oakland Research Institute, was transformed into EL350 cells. PCR was used to retrieve approximately 400-bp homology arms that flanked exons 2, 3, and 4 of Gsk3a by 5 kb on each side. Deletion of exon 2 has been shown to result in a nonfunctional kinase [15]. After cloning homology arms into a retrieval vector (pBluescript modified to contain a thymidine kinase cassette), this construct was transformed into the EL350 BAC-containing cells, and induced homologous recombination via gap repair, resulting in a 10-kb fragment of Gsk3a being retrieved into the pBluescript derivative. Next, mini-targeting vectors were created by introducing identical 34-bp loxP sites on both sides of a neomycin resistance cassette (neomycin phosphotransferase gene [Neo^R]). The loxP/Neo^R cassette was then introduced into gap repaired retrieval plasmids using a similar approach as described above for retrieving the Gsk3a fragment. Transformation of these constructs into bacteria with arabinose-inducible Cre allowed recombination, removing the Neo^R cassette and one loxP site. An additional round of this cloning approach was used to introduce a second loxP/Neo^R cassette, now flanked by FRT sites, on the 3′ side of exon 2. Plasmids were linearized with NotI, gel purified, and electroporated into mouse ES cells by the ES Cell and Transgenic Core Facility at the Research Institute at Nationwide Children's Hospital. Targeted ES cells were grown in the presence of G418 (350 μg/ml) and FIAU (5-iodo-2′-fluoro-2′-deoxy-arabinouridine; 200 nM) for the selection of cells that contained correct gene targeting. Southern blotting of G418/FIAU-resistant clones revealed that 22 out of 156 colonies (14%) contained a correctly targeted Gsk3a floxed allele. Two clones were selected for injection into blastocysts, from which we obtained several male chimeric mice. These mice were mated to either ROSA26-FLPe mice, a well-characterized transgenic mouse that ubiquitously expresses Flp at high levels [16], resulting in the highly efficient removal of the Neo^R cassette, or WT C57BL/6 females.

Isolation and Analysis of Sperm Morphology

The cauda epididymis and vas deferentia from adult mice, aged 7–10 wk, were isolated and placed in 1 ml of human tubal fluid buffer (HTF; Embryo Max HTF from Millipore). The HTF buffer was equilibrated in a 5% CO₂ chamber at 37°C for 2 h before use for sperm suspension. Each cauda epididymis was punctured several times with a 26-G (45-mm) needle and the sperm were allowed to swim out. Swim out was also aided by squeezing with surgical scissors. Sperm was also extruded from the vas deferens by squeezing it along its length. Sperm were then allowed to disperse in the media for ∼10 min at 37°C with occasional swirling and then transferred to micro centrifuge tubes with large-bore pipette tips; 10 μl of the sperm isolate was diluted 10 times (1:10) in water for determination of sperm concentration by counting in a Neubauer hemocytometer.

For examination of morphology, sperm were resuspended in freshly prepared 3.75% paraformaldehyde in 1× PBS and incubated at 4°C for 1 h. Fixed sperm were mounted on clean poly-l-lysine-coated slides and sealed with coverslips. Sperm morphology was observed under 20× and 60× objective lens with an Olympus 81 microscope using differential interference contrast.

Sperm Motility Analysis

Within 10 min of isolation in 1× PBS media, as described above, caudal sperm were diluted to a concentration of 2 × 10⁷ sperm/ml, and 25 μl of diluted sperm suspension was loaded using a large-bore pipette into a 100-μm Leja chamber slide, prewarmed to 37°C on a stage warmer. Sperm motility was analyzed with a computer-assisted sperm motility analyzer equipped with the CEROS sperm analysis system (software version 12.3; Hamilton Thorne Biosciences, Beverly, MA) [17]. For each chamber with the sperm sample, three to five random fields were recorded and analyzed using the following settings: 90 frames acquired at 60 frames/sec; minimum contrast of 30; minimum cell size at 4 pixels; default cell size at 13 pixels; static cell intensity of 60; low size gate of 0.17; high size gate of 2.26; low-intensity gate of 0.35; high-intensity gate of 1.84; minimum static elongation gate of 0; maximum static elongation gate of 90; minimum average path velocity (VAP) of 50 μ/sec; minimum path straightness (STR) of 50%; VAP cut off of 10 μ/sec; and straight line velocity cut off of 0 μ/sec. Motility was recorded independently for sperm collected from WT (n = 3) and Gsk3a^−/− (n = 3) mice.

Preparation of Mouse Testis and Sperm Extracts

Testes from WT, Gsk3a^+/−, and Gsk3a^−/− mice were homogenized in homogenization buffer with protease inhibitors (HB⁺; 10 mM Tris [pH 7.2] containing 1 mM EDTA, 1 mM EGTA, 10 mM benzamidine-HCl, 1 mM PMSF, 0.1 mM N-p-tosyl-l-phenylalanine chloromethyl ketone, and 0.1% [V/V] β-mercaptoethanol). The homogenates were centrifuged at 16 000 × g for 20 min at 4°C, and the supernatants were collected as testis extracts. Caudal epididymal sperm isolated in HTF media was centrifuged at 700 × g for 10 min at 4°C. The sperm pellet was resuspended in 1% SDS at a final concentration of 2 × 10⁸ sperm/ml. The sperm suspension in 1% SDS was then boiled in a water bath for 5 min and centrifuged at 12 000 × g for 15 min at room temperature. The supernatant was collected and used for Western blot analysis.

Western Blot Analysis

Testes and sperm extracts were denatured by boiling with Laemmli buffer for 3 min. The proteins were separated by electrophoresis on 12% SDS-PAGE and transferred onto Immobilon-P PVDF membranes (Millipore Corp.). Following transfer, nonspecific protein binding sites on the membrane were blocked by incubation with 5% nonfat dry milk diluted in TTBS (0.2 M Tris, pH 7.4, 1.5 M NaCl, 0.1% thimerosol and 0.5% Tween 20). The blots were then incubated with primary antibody diluted in 5% nonfat dry milk in TTBS. The following primary antibodies were used: anti-GSK3A/B mouse monoclonal antibody (1:1000 dilution; 44610; Invitrogen); anti-hexokinase1 rabbit monoclonal antibody (1:1000 dilution; 2024; Cell Signaling); anti-PP PP1 gamma2 (PPP1CC2) antibody (1:5000, commercially prepared using a synthetic peptide corresponding to the 22 amino acids at the carboxy terminus of PPP1CC2 as the antigen); anti-phosphotyrosine mouse monoclonal antibody (1:1000 dilution, clone 4G10^R; Millipore); anti-beta-tubulin rabbit polyclonal Ab (1:6000 dilution, ab52901; Abcam). Following a brief wash with TTBS, the blots were incubated with the appropriate horseradish peroxidase-conjugated secondary antibody (1:2000; GE Amersham) for 1 h at room temperature. Blots were washed with TTBS twice for 15 min and twice for 5 min. The blots were finally developed with enhanced chemiluminescence substrate (Thermo Scientific Super Signal West Pico ECL).

Protein Phosphatase Assay

Caudal epididymal sperm resuspended in HB⁺ at a final concentration of 2 × 10⁸ sperm/ml were sonicated on ice with an ultrasonic cell disruptor (Q-Sonica) for three times at 10 sec each at 30% amplitude. The sonicate was centrifuged at 16 000 × g for 20 min at 4°C, and the supernatant was collected as the soluble fraction. The pellet (insoluble fraction) was resuspended in an equal volume of HB⁺. The soluble and insoluble fractions of the sonicates were utilized for phosphatase enzyme activity measurement on the same day. Radiolabeled phosphorylase a was used as a substrate to measure the activity of PP1 by procedures previously reported [18]. Aliquots of the fractions of sperm extracts were preincubated at 30°C for 15 min in either the presence or absence of protein phosphatase inhibitors. Purified protein phosphatase inhibitor-2 (I-2) was used at a final concentration of 25 nM to inhibit PP1 or 2 nM okadaic acid final concentration to selectively inhibit PP2A. Phosphorylase a was then added and the samples were further incubated at 30°C for 10 min. The reaction was terminated by addition of 10% trichloroacetic acid and centrifuged for 10 min at 12 000 × g. Supernatants were quantitated for ³²P released from phosphorylase a.

Histology of Testis Sections

Testes were fixed in Bouin Fixation Fluid (Harleco) for 6 h and washed in 70% ethanol to remove excess Bouin solution. Fixed testes were then processed by washing in a graded, increasing ethanol concentration (70%, 80%, 95%, and 100%) for 45 min each, and then permeabilized in CitriSolv (Fisher Scientific) for 30 min [17, 19]. The processed testes were then embedded in paraffin, sectioned, and the 5-μm-thick sections were transferred to poly-l-lysine-coated slides. Sections were stained with periodic acid-Schiff staining kit (Leica Biosystems) using the manufacturer's protocol. The slides were counterstained with Gill II hematoxylin for 3–4 min, rinsed again in running tap water for 5 min, dehydrated through two changes of 95% and 100% alcohol (2 min each), and finally cleared in two changes of xylene before addition of mounting medium.

Immunocytochemistry of Spermatozoa

Caudal epididymal spermatozoa in PBS were spun down at 700 × g for 10 min at 4°C. The cells were fixed in 4% paraformaldehyde, EM grade (Electron Microscopy Sciences) at 4°C for 20 min, followed by permeabilization with 0.2% Triton-X (5 min). Fixed spermatozoa were attached to poly-l-lysine-coated coverslips. The coverslips were washed three times with TTBS to remove excess paraformaldehyde and incubated for 4 h at room temperature in a blocking solution containing 5% normal goat serum and 5% BSA in TTBS at room temperature. The coverslips were then incubated overnight at 4°C with anti-GSK3α antibody (1:150 dilution; SAB4300292; SIGMA) and sperm protein sp56 mouse monoclonal antibody (1:200 dilution; MA1-10866; Thermo Scientific), washed three times 5 min each with TTBS, followed by incubation with the appropriate secondary antibody conjugated to Cy3 for 1 h at room temperature. The coverslips were then washed three times, 10 min each with TTBS, mounting medium was applied, and the sperm cells were examined by brightfield and fluorescence microscopy.

Mitochondrial Structure Assessment

Sperm were collected from C57BL6 WT and Gsk3a^−/− mice and was resuspended in PBS with a concentration of 1 × 10⁷ sperm/ml. It was then incubated in 20 nm of MitoTracker green (Molecular Probes) for 30 min at 37°C to stain the live mitochondria. To counterstain the nucleus, samples were subsequently incubated in Hoechst dye (Molecular Probes) for 20 min at 37°C in 1:1000 dilution. The samples where mounted on poly-l-lysine-coated slides and observed with an Olympus 1X81 microscope.

ATP Assay

Cauda epididymal sperm were isolated in HTF medium as described above and sperm concentration determined. At the specified time points, triplicate 30-μl aliquots were diluted into 270 μl of boiling Tris-EDTA buffer (0.1 M Tris-HCl and 4 mM EDTA; pH 7.75) as described previously [20]. The diluted suspensions were boiled for 5 min and then frozen in dry ice. The frozen samples were thawed and centrifuged at 15 000 × g for 5 min at 4°C. The supernatant was then diluted at least 1:10 using the Tris-EDTA buffer and 100 μl of diluted sample was then utilized for quantifying ATP using the Bioluminescence Assay Kit CLS II (Roche Applied Science). Luminescence was measured in a Turner Biosystems 20/20 Luminometer.

RNA Isolation, cDNA Synthesis, and RT-PCR Analysis

Total RNA isolation from testis of WT, Gsk3a^+/−, and Gsk3a^−/− mice was performed using TriZol reagent (Sigma), phenol-chloroform extraction (Amresco), and isopropanol precipitation. The pellet was washed with 75% ethanol and dissolved in DE-PC-treated water. The concentration of the RNA was measured in a Nanodrop ND-1000 Spectrophotometer (Nanodrop Technologies); 900 ng of RNA was used to prepare cDNA by following the QuantiTect Reverse Transcription Kit (Qiagen). The PCR analysis was done using 1 μg of cDNA prepared from RNA of testis of WT, heterozygous, and KO mice. The following primer sets were used: L 5′-ACCCTTGGACAAAGGTGTTC-3′ (from exon 8/9 junction), and L 5′-TCAGTCCTGGTGAACTGTCC-3′ (from exon 10) expected to produce an amplicon 300 bp in size.

Quantitative PCR

Quantitative PCR analysis of Gsk3a mRNA expression was done using the QuantiTect SYBR Green RT-PCR Kit according to the manufacturer's protocol. Quantitative RT-PCR (qRT-PCR) experiments were performed on the Rotor-Gene Q series. For SYBER green (Quanti-Tect SYBR Green RT-PCR Kit) qRT-PCR, average threshold cycles were determined from triplicate reactions. PCR reactions were carried out in a 25-μl volume for 5 min at 95°C for initial denaturing, followed by 35 cycles at 95°C for 5 sec and at 59°C for 20 sec. A housekeeping control gene, Gapdh, was used as an internal control. Each primer set was first tested to determine optimal concentrations, and products were run on a 1% agarose gel to confirm the presence of the predicted amplification products. Data were analyzed by ΔCt method considering the average Ct values of each sample. Error bars indicate standard deviation or standard deviation normalized to a reference sample, as indicated in the legends accompanying the figures.

Statistics

Statistical analyses were performed using the nonparametric one-way ANOVA and two-tailed unpaired t-test using the GraphPad Prism 6.03 (GraphPad Software Inc.). In all cases, differences between samples were considered significant for P values ≤ 0.05.

RESULTS

Targeted Disruption of Gsk3a

In order to create conditional alleles of Gsk3a, homologous recombination was used to insert flanking loxP sites into the Gsk3a locus (Fig. 1). The gene targeting strategy was designed to remove exons 2, 3, and 4 from Gsk3a. A doublet of lysine residues necessary for catalytic activity (Lys148/149 in Gsk3a) lies within exon 2 of both isoforms; therefore, loss of exons 2, 3, and 4 from Gsk3a was predicted to introduce premature stop codons in the mRNA, resulting in a truncated message and severely truncated protein if translated. It has been previously shown that the loss of exon 2 results in loss of detectable GSK3A protein [15]. Chimeric mice were generated by injection of targeted ESC clones into C57BL/6 blastocysts. Chimeric males were mated to WT C57BL/6 females to generate heterozygous mice. Heterozygous Gsk3a (+/loxP) mice were crossed with mice possessing the Sox2-Cre transgene, which expresses Cre throughout the epiblast, resulting in the effective removal of the floxed sequence in all tissues of the adult mice. Heterozygous Gsk3a^+/− males and females were crossed to obtain homozygous Gsk3a^−/− mice.

FIG. 1.

Targeting strategy to generate Gsk3a conditional alleles. The targeting construct was generated via recombineering to include loxP sites flanking exons 2, 3, and 4. Additionally, the downstream loxP site includes an adjacent PGK-NeoR cassette flanked by FRT sites to facilitate removal of this cassette by Flp recombinase. The hatched boxes represent homology arms used to retrieve Gsk3a from the BAC clone, and also represent the edges of Gsk3a homology in the targeting construct. HSV-thymidine kinase was included outside of the Gsk3a homology region for negative selection against random integrations in the mouse genome. Two Southern blot probes (black boxes) detect a 20.5-kb SpeI fragment in the WT allele, and either a 3.9-kb or 14.8-kb fragment in the PGK-NeoR floxed allele. Genotyping PCR primers using a common reverse primer (R) were designed to detect the WT allele (F1-R, 482 bp), PGK-NeoR floxed allele (F2-R, 330 bp), Flp deletion product (F1-R, 579 bp), and Cre deletion product (F3-R, 252 bp).

Surprisingly, homozygous Gsk3a^−/− male mice failed to produce offspring. Previous results, which employed a different KO strategy, suggested that fertility of Gsk3a^−/− was normal [15]. More recently, a serendipitous observation made during studies on the role of the Gsk3a in brain noted that males apparently lacking Gsk3a were infertile [21]. The reasons for the apparent infertility in the null mice were not explored. We therefore undertook a more detailed analysis of the Gsk3a^−/− mice that we had generated using this different gene disruption strategy.

GSK3 Isoforms in Sperm and Testis of Gsk3a^+/+ and Gsk3a^−/− Mice

Western blotting of testis extracts probed with a monoclonal antibody that recognizes both GSK3 isoforms shows that GSK3A and B are present in WT and Gsk3a^+/− mice (Fig. 2A). As expected, GSK3A was not detectable in testis of Gsk3a^−/− mice, while GSK3B was present at apparently unaltered levels. Similarly, both GSK3 isoforms were present in WT and Gsk3a^+/− sperm, whereas Gsk3a^−/− sperm contained only GSK3B (Fig. 2B). It should also be noted that GSK3A protein levels were lower in Gsk3a^+/− compared to WT testis; however, its levels in sperm were comparable to that in WT sperm. The levels of the sperm- and testis-specific PPP1CC2 were unaltered. Immunoreactivity to beta-tubulin also showed equal protein loading.

FIG. 2.

Protein and mRNA expression in WT and Gsk3a^−/− mice. A and B) Western blot analysis of testis and sperm from WT (+/+), Gsk3a heterozygous (+/−), and Gsk3a^−/− showing absence of GSK3A in KO extracts compared to WT and Gsk3a^+/−. PPP1CC2 levels in all the samples remained almost equal for both testis (A) and sperm (B) extracts. The blot with testis extract was probed with beta-tubulin antibody to demonstrate equal protein loading. Extract from 2 million sperm was loaded in each lane for the sperm blot. C) Quantitative PCR analysis of RNA from testis of WT, Gsk3a^+/−, and ^−/− mice shows drastically reduced levels of Gsk3a mRNA. D) Quantitative PCR showing increase in Gsk3a mRNA levels in postnatal developing testis. The primers used in C and D were from exon 8–9 junction (L5′ACCCTTGGACAAAGGTGTTC3′) and exon 10 (R 5′TCAGTCCTGGTGAACTGTCC3′). The results are represented as fold change after normalizing the Gsk3a mRNA levels with Gapdh mRNA. These data are representative of three independent experiments in triplicates, and error bars represent SE.

Determination of mRNA levels for Gsk3a by quantitative PCR showed that the Gsk3a^+/− mRNA levels in testis was roughly one-half of that in WT testis. The mRNA for Gsk3a in Gsk3a^−/− testis was drastically reduced. In both cases, the reduction was likely due to reduced stability of the truncated transcript arising from the mutant alleles lacking exons 2, 3, and 4.

It is known that a 2.5-kb mRNA for Gsk3a is expressed in adult testis [22, 23]; however, it is not known if and how expression of Gsk3a mRNA changes in postnatal developing testis. Gsk3a mRNA in testis increases in postnatal developing testis, reaching a maximum in adult testis (Fig. 2D). A temporal expression pattern similar to genes expressed in developing spermatocytes and spermatids coinciding with the onset of spermatogenesis (Day 15 testis), when differentiating spermatocytes lead to the formation of spermatids.

Immunocytochemistry of WT sperm showed the presence of GSK3A along the entire length of the principal and midpieces. Staining was also predominant in the acrosomal region of the sperm head (Fig. 3, A–D). In contrast, staining was absent in sperm exposed only to the secondary antibody (data not shown) and in sperm from Gsk3a^−/− mice (Fig. 3, E and F). We used MitoTracker staining to check the status of mitochondria in KO sperm. MitoTracker green staining showed that Gsk3a^−/− sperm mitochondria appeared to be normal, and that the intensity of staining was comparable with Gsk3a^+/+ (Fig. 4, J–M). Staining with acrosomal matrix protein sp56 showed no apparent defects in acrosomal structure of Gsk3a^−/− sperm, as they appeared as normal crescent shapes seen in WT sperm (Fig. 4, G–I).

FIG. 3.

Immunocytochemistry of spermatozoa and histological analysis of testis sections from WT and Gsk3a^−/− mice. Brightfield and fluorescence immunocytochemistry of spermatozoa from WT mouse (A–D). Sperm were labeled with rabbit polyclonal antibody against GSK3A as described in Materials and Methods. GSK3A was present in the acrosomal region of the head and in the entire length of the flagellum, including the midpiece (B and D). There was not staining in sperm from Gsk3a^−/− (E and F). G–J) Testis sections stained with hematoxylin show normal morphology in Gsk3α^−/− (H and J) mice and apparently normal spermatogenesis as seen in WT mice (G and I). Stage IX spermatids were identified by the absence of round spermatids (K–N). For both WT (K, L) and Gsk3α^−/− (M and N), number of stage IX spermatids are similar. Bars = 100 μm in both 20× and 40×.

FIG. 4.

Morphological features of sperm from a Gsk3a^−/− mice. A–C) A proportion of KO sperm was normal. D–F) A hairpin bend (arrow) at the mid- and principle piece junction was observed in some cases. Bent heads were also observed in a fraction of sperm from Gsk3a^−/− mice. G–I) Immunostaining of acrosome protein sp56 shows normal crescent moon-shaped structure (boxed area) for Gsk3a^−/− sperm. Inset showing stained acrosomal structure. J–M) Staining with MitoTracker green shows normal mitochondrial structure for Gsk3a^−/− (K–M) when compared with WT sperm (J). The pictures were taken by 1×81 Olympus Microscope at 60× (oil emulsion) objective lens.

Sperm Motility and Fertility of Gsk3a Null Mice

Gsk3a^−/− mice appeared healthy, with normal body weight, and displayed no overt phenotype, except that Gsk3a^−/− males were infertile. Testis weights and sperm numbers appeared normal in Gsk3a^+/− compared to Gsk3a^+/+ mice (Table 1). Testis weights and sperm numbers in Gsk3a^+/− and Gsk3a^−/− were comparable, but reduced compared to Gsk3a^+/+ mice (Table 1). Morphologically normal sperm were 86% and 71% in Gsk3α^+/− and Gsk3a^−/− mice, respectively. A proportion of these sperm with normal-looking heads were bent at the mid- and principal-piece junction (Table 1, Fig. 4). A higher proportion of sperm from both Gsk3a^+/− and Gsk3a^−/− mice were bent at the head/connecting piece junctions (Table 1, Fig. 4). A proportion of sperm from WT mice also showed this bent head morphology. Testis sections show that testis morphology was unaltered, and that spermatogenesis in KO mice appeared to be normal compared to WT mice (Fig. 3, G–J). Cross-section of seminiferous tubules from WT and Gsk3a^−/− testis showed normal morphology. Moreover, stage IX tubule cross-sections of testis from WT and KO mice (Fig. 3, K–N) showed comparable numbers of spermatids (mean numbers of spermatids were 80 for WT and 76 for KO). It thus appears that absence of GSK3A in testis does not affect spermatogenesis. Morphological analysis of Gsk3a^−/− testicular, caput, and caudal sperm showed bent midpiece, and bent heads were present in epididymal, but not in testicular, sperm (Table 2).

TABLE 1.

Testis weights, sperm number, and morphology of WT, Gsk3a^+/−, and Gsk3a^−/− mice.

^*Values shown are means ± SEM from different determinations.

^†n denotes number of samples/groups.

^a,b,cSignificantly different groups.

TABLE 2.

Testicular, caput, and caudal sperm morphology of Gsk3a^−/−.

^*n denotes number of samples/groups.

^†Values shown are means ± SEM from different determinations.

^a,b,cSignificantly different groups.

Independent mating of 11 Gsk3a^−/− males with WT or Gsk3a^+/− females did not produce any offspring (Table 3). Gsk3a^+/− males and females were fertile. Homozygous Gsk3a^−/− females were also fertile, as they were used in breeding for maintaining the Gsk3a^−/− line.

TABLE 3.

Fertility of Gsk3a^−/− and Gsk3a^+/− mice.

^*n denotes number of samples/groups.

^†Values shown are means ± SEM from different determinations.

Sperm motility of Gsk3a^−/− mice was reduced compared to Gsk3a^+/− mice. All parameters of motility, including percent and progressive motility and velocity parameters, were significantly reduced in KO compared to WT sperm (Fig. 5). Velocity of KO sperm was about 50% that of WT sperm. A notable feature of motility of KO sperm was the visibly reduced flagellar beat amplitude (Supplemental Movies S1–S5; all Supplemental Data are available online at www.biolreprod.org).

FIG. 5.

Motility analysis of mature caudal sperm from Gsk3a^−/− and WT animals. Computer-assisted sperm analysis of freshly prepared cauda epididymal spermatozoa was performed from adult (8–10 wk old) WT and KO sperm. A) Both percent motility and progressive motility in Gsk3a^−/− sperm were significantly decreased compared to WT (B). There was alteration in the velocity parameters VAP, straight line velocity (VSL), and curvilinear velocity (VCL). Mean values of all these velocity parameters of KO mice were significantly lower than the WT control. The motility parameters were expressed as mean of n = 3 ± SEM. For each animal, eight or more nonoverlapping fields were recorded for analysis. A two-tailed unpaired t-test was used for analysis between the groups. The motility parameters were expressed as significant differences from WT control values observed. **Significant difference (**P < 0.01 and *P < 0.05). C) Still images of Supplemental Movies S1–S5 were taken from the first frame of the recordings. WT1 and KO1 stills are from recordings used for computer-assisted sperm motility analyzer; WT2, KO2, and KO3 are high-speed recordings of tethered sperm. The recordings were made with high-performance digital camera DVC 340M (Thor Labs Scientific Imaging) mounted onto a Zeiss microscope with a 10× phase contrast objective lens. The DVC camera was set to record 5 sec videos with readout of 20 MHz pixel rate at 12 bits, 2 msec exposure and 2 × 2 binning for a final output of 100 frames/sec.

Protein Phosphatase Activity and ATP Levels in Sperm from Gsk3a^−/− Mice

We first identified sperm GSK3 as an enzyme involved in activation of PPP1CC2 [18, 24]. This activation was thought to involve phosphorylation of the protein phosphatase inhibitor PPP1R2 (inhibitor I-2) by GSK3. To determine if there was a relationship between GSK3A and protein phosphatase, we assayed for protein phosphatase activity in sperm from Gsk3a^−/− mice. Protein phosphatase activity was measured using phosphorylase a as a substrate. Data in Figure 6B show that total protein phosphatase (PP1 + PP2A; Fig. 6A) and PP1 activities (Fig. 6B) in both the soluble and insoluble fractions of sperm sonicates from Gsk3a^−/− mice were significantly higher than in WT mice.

FIG. 6.

Protein phosphatase activity and ATP levels. Total protein phosphatase activity (A) and PP1 activity (B) was measured both in supernatant (sup) and pellet (insoluble fraction) for equal sperm numbers (2 × 10⁵). A significant increase in both total and PP1 activity was observed in sperm from Gsk3a^−/− compared to Gsk3a^+/− and WT mice. C) Sperm incubated for 40 min in HTF complete medium were analyzed for ATP levels. A significant decrease in ATP levels were observed in KO^−/− compared to heterozygous (+/−; Het) and WT mice. A–C) Values expressed as percent of control are means ± SEM from three different experiments (n = 3). A nonparametric one-way ANOVA was used for comparison of all groups. *P < 0.05, **P < 0.01, significant differences compared to WT control.

Based on gene KO approaches, it appears that sperm ATP levels and cAMP levels may be regulated by PKA [25]. ATP levels in sperm from sAC and PKA catalytic subunit (Cs) KO mice are also low compared to WT sperm [26]. Because it is possible that sperm GSK3A could be a mediator or target of PKA action, we measured ATP levels in sperm from Gsk3a^−/− mice. ATP levels in Gsk3a^−/− sperm were reduced to nearly 50% of those in WT sperm (Fig. 6C).

Protein Tyrosine Phosphorylation in Sperm from Gsk3a Null Mice

Changes in global protein tyrosine phosphorylation, accompanying sperm capacitation, occur in the presence of bicarbonate in the sperm suspension buffer [27, 28]. The blot in Figure 7A shows that tyrosine phosphorylation detected by a phosphotyrosine-specific antibody was diminished in Gsk3a^−/− sperm compared to WT sperm. The blots reprobed with beta-tubulin antibody show equal protein loading.

FIG. 7.

Tyrosine phosphorylation of sperm from Gsk3a^−/− mouse compared to WT mouse. A) Western blots developed with anti-phosphotyrosine antibodies. Protein extracts from 2 × 10⁶ sperm without incubation (0 h) or after 1 h of incubation (1 h) in HTF medium. In both cases, significant decreases in phosphotyrosine levels for KO (−/−) compared to WT were observed. B) The same protein extracts were used for comparing hexokinase-1 levels in WT and KO samples by probing the blot with anti-hexokinase-1 antibody. The blot shows equal levels of hexokinase protein in both KO and WT samples. All the blots were subsequently developed with anti-beta-tubulin antibody to show equal protein loading. Each experiment was repeated three times with preparations from different animals producing similar results.

Surprisingly, immunoreactivity of a tyrosine phosphorylated protein band around 120 kDa, thought to be hexokinase [29], is absent. Sperm hexokinase, which is constitutively tyrosine phosphorylated, is often used as an internal loading control to denote equal protein loading [30]. Figure 7B shows that immunoreactivity to hexokinase is the same in WT and KO sperm extracts, eliminating the possibility that the lack of hexokinase phosphorylation could be due to a lack of the protein. It is likely that immunoreactivity seen at ∼25 kDa was likely due to light chain of IgG from blood contamination in the sperm preparation.

DISCUSSION

Testicular sperm undergoing maturation in the epididymis are terminally differentiated cells with little or no transcription and protein synthesis. Regulation of sperm function during their acquisition of motility and fertilizing ability in the epididymis and during fertilization of the egg in the female reproductive tract must involve changes in phosphorylation of pre-existing proteins. Cyclic AMP-mediated protein kinase activation is essential for activated and hyperactivated motility. Consequently, disruption of Cs of PKA in testis and sperm, and of soluble adenylyl cyclase, the source of sperm cAMP, results in impaired sperm motility and male infertility. We now show that GSK3 is a key kinase essential for male gamete function and fertility.

There are two conflicting accounts of the phenotype resulting from disruption of Gsk3a [15, 21]. It was first reported that homozygous null mice for Gsk3a display normal fertility [15]. More recently, observations from Gsk3a^−/− mice, which were unexpectedly generated during attempts to effect a brain region-specific disruption, showed that male mice lacking Gsk3a are infertile [21]. The male infertility phenotype in this report was not fully explored. The reasons for the discrepancy in the phenotypes observed in these two reports are not known. Generation of null mice in both these reports used the same disruption strategy that should have resulted in the removal of exon 2 of Gsk3a. In the first report, deletion of exon 2 was accomplished with pCAGGS (chicken β-actin) CRE mice [15]. It is possible that, in this study, there could have been incomplete CRE-mediated removal of Gsk3a in testes, accounting for the normal fertility of male mice that were thought to globally lack GSK3A. In this current report, we conclusively show, using a targeted disruption strategy that removes exons 2, 3, and 4, that Gsk3a is essential for normal sperm function and male fertility.

The only obvious phenotype of the targeted KO of Gsk3a is male infertility. The most likely reason for this infertility is compromised sperm motility. In addition to reduced progressive motility and velocity parameters, sperm lacking Gsk3a have a markedly reduced flagellar beat. Activated and hyperactivated sperm motility characterized by high-amplitude flagellar beat are essential for penetration and fertilization of the egg [31]. A strikingly similar attenuated flagellar beat and impaired sperm hyperactivation was also seen in sperm lacking the sperm-specific Cs of PKA, Cαs [25]. Flagellar beat waveform features seen with Gsk3a null sperm were observed in membrane-permeabilized sperm activated in the presence of AMP [32] and also in sperm lacking the nucleoside transport protein, Slc29a (ENT1) [33]. Further studies with sperm from Gsk3a^−/− null mice are required where quantitation of flagellar beat amplitude and frequency can be coupled to biochemical changes in null sperm. Just as seen in Cαs KO mice, we anticipate that sperm lacking Gsk3a will also be unable to fertilize eggs in vitro, due to impaired capacitation and their inability to undergo hyperactivation. Because testis morphology, spermatogenesis, and sperm numbers appear normal, it is unlikely that Sertoli or Leydig cell functions are compromised in Gsk3a^−/− mice. The major dysfunction in mice lacking Gsk3a is in male gametes rather than in somatic cells in testis and other tissues. However, definitive verification of this conclusion must await conditional spermatocyte cell-specific knockdown of Gsk3a.

It is generally accepted that protein tyrosine phosphorylation accompanies sperm capacitation. Reduced tyrosine phosphorylation during incubation of sperm under capacitating conditions supports the possibility that biochemical events that normally accompany sperm capacitation are affected in sperm lacking Gsk3a. We have also determined that PKA-mediated serine/threonine phosphorylation, as measured by PKA phospho domain antibodies, is unaltered in Gsk3a^−/− compared to WT sperm (data not shown). Furthermore, cAMP levels were not altered in Gsk3a^−/− compared to WT sperm (data not shown). Reduced tyrosine phosphorylation (Fig. 7), in spite of the presumed presence of PKA and its activation by cAMP, shows that GSK3A is required for sperm capacitation-associated tyrosine phosphorylation. Our data support the novel suggestion that events accompanying sperm capacitation require both PKA and GSK3A.

Some of the key changes occurring in sperm during capacitation, attributed to PKA, are likely to be mediated by GSK3A. It is possible that GSK3A may regulate the activity of a tyrosine kinase or phosphatase responsible for changes in sperm protein tyrosine phosphorylation. It is notable that, in addition to lowered tyrosine phosphorylation in general, tyrosine phosphorylation of hexokinase in particular is reduced or is virtually absent. To our knowledge, there is but one documented instance where tyrosine phosphorylation of mouse sperm hexokinase is absent [29]. This is in infertile t-complex mice with defective sperm motility [29]. The biochemical basis for the loss of tyrosine phosphorylation in the t-complex mice, or the significance of how and why sperm hexokinase is tyrosine phosphorylated, is not understood [34–36]. Future studies should reveal how GSK3A is involved in tyrosine phosphorylation of sperm hexokinase. The knowledge that GSK3A may interact with or regulate protein tyrosine kinases or phosphatases should also help in the understanding the mechanistic basis for sperm protein tyrosine phosphorylation.

Morphology is normal in a majority of the sperm isolated from testis of Gsk3a^−/− mice; however, a significant proportion of sperm from null mice are bent at the midpiece/principal piece and head/connecting piece junctions (Table 1). This hairpin bend at the midpiece was also seen in sperm from sAC KO mice [37, 38] and mice lacking the PKA regulatory subunit 1A [39]. Interestingly, in both these KO mouse lines, cAMP-mediated sperm PKA action is missing or lacking. We propose, as discussed above, that downstream effects of PKA may be compromised in Gsk3a^−/−. Bends at the head are also seen in a majority of sperm from mice lacking SPEM1, a protein expressed in spermatids and localized in the sperm cytoplasmic droplet [40]. Unlike Gsk3a^−/−, sperm from the SPEM1 KO mice do not show bends at the sperm midpieces. Bends at the head connecting piece junctions are also seen in sperm from transgenic mice, where there are reduced levels of testis and sperm PPP1CC2 [17, 41]. It is possible that these bent head and midpiece morphological features could result from altered steady-state levels of sperm protein phosphorylation. Consistent with previous observations, it is interesting that these morphological characteristics are present only in developing epididymal sperm but not in testicular sperm (Table 2) [40, 41]. Moreover, although the proportions of bent heads are similar in both caput and caudal sperm, the proportion of sperm with bent midpieces is significantly higher in caudal compared to caput sperm (Table 2). It is possible that bends at the head and midpieces could occur due to improper removal of the cytoplasmic droplets and/or due to mechanical shear during the passage of sperm through the epididymis. Staining of acrosome with antibodies against sp56 show that morphologically normal sperm and sperm with bent heads and midpieces from KO mice have normal-shaped acrosomes. Thus, acrosome biogenesis is normal in Gsk3a^−/− testis. This is in contrast to PICK1 KO mice, where sperm acrosomes are malformed [42]. Moreover, mitochondria of sperm from Gsk3a^−/− mice also appear normal.

The reasons why ATP levels are low in Gsk3a^−/− sperm are not known. Reduced ATP levels may be due to the lower energy requirements of submotile sperm. Alternatively, it is possible that lower ATP levels may be due to reduced glycolysis and/or respiration. Because serine/threonine phosphatase activity is higher in sperm lacking GSK3A, it is possible that higher protein phosphatase activity may result in altered phosphorylation and catalytic activity of glycolytic enzymes. We have previously reported that sperm glycolytic enzymes are among proteins known to contain a PP1 binding motif and bind to sperm PPP1CC2 [43, 44]. It is, however, generally thought that increased phosphorylation is a negative feedback mechanism regulating glycolytic and mitochondrial enzymes. Why decreased or absent phosphorylation in the absence of GSK3A should result in lower glycolysis or respiration is puzzling.

We have previously shown that GSK3 could activate sperm PPP1CC2, presumably by phosphorylating the PP1 inhibitor protein I-2 [18, 24]. We therefore anticipated that PP1 activity should be low in Gsk3a^−/− sperm. However, protein phosphatase activity is higher in sperm lacking GSK3A compared to WT sperm. It is possible that PP1 regulators other than I-2, PPP1R11 (I-3), and PPP1R11 (sds22), known to be present in sperm [45, 46], may be activated/inactivated in the absence of GSK3A. The ability of I-3 and sds22 to bind and inhibit PPP1CC2 may depend on an inhibitory or activating phosphorylation effected by GSK3A along with yet another protein kinase, such as PKA. It is known that PPP1CC2 and other PP1 isoforms in several species, including yeast, have a highly conserved threonine phosphorylation site (TPPR). Phosphorylation at this threonine residue is known to reduce catalytic activity of PP1 [47–50]. It is possible that, in sperm, GSK3A may be responsible for threonine phosphorylation of PPP1CC2. We have previously shown that PPP1CC2 inhibition results in increased phosphorylation and decreased GSK3A catalytic activity. Whatever the mechanism by which GSK3A may regulate PP1, a mutual regulatory relationship between GSK3A and PPP1CC2 exists in spermatozoa. The existence of a feedback regulatory cycle between GSK3 and PP1 has been shown in somatic cells [51].

It is intriguing that some of the phenotypes of Gsk3a KO sperm are similar to those seen in sperm lacking the PKA Cs. These include altered sperm flagellar beat, low tyrosine phosphorylation and ATP levels. We suggest that GSK3A is a key mediator of PKA action in spermatozoa. Sperm PKA, in turn, regulates phosphorylation and activity of GSK3A. Similar to PPP1CC2 and GSK3A, sperm PKA and GSK3A also appear to be mechanistically interconnected. We propose that there exists a mechanistic interrelationship between GSK3A, PKA, and PPP1CC2 essential for normal sperm function.

In many cell types, the biochemical functions of GSK3A and GSK3B are interchangeable. However, GSK3B KO results in embryonic or perinatal lethality, suggesting that GSK3A may not substitute for GSK3B during embryo development [52–54]. This apparent lack of biochemical equivalence between GSK3B for GSK3A may also be due to a nonoverlapping expression of the two isoform in cells of the developing embryo. Both isoforms of GSK3 are present in testis. Microarray analysis also shows that, for selected probe sets, GSK3B expression is highest in postmeiotic germ cells compared to somatic cells in both mouse and rat testis (Mammalian Reproductive Genetics database). It is known that, in rat testis, GSK3B expression follows a similar temporal expression pattern [55]. We have shown here, for the first time, that GSK3A expression increases during the onset of spermatogenesis, reaching a maximum in adult testis (Fig. 2). Thus, there is temporal and spatial overlap in the expression of both GSK3 isoforms in testis. We have also shown that both GSK3 protein isoforms are also present in sperm (Fig. 2). While GSK3A is present in several species, its unique glycine-rich N terminus is highly conserved only in mammals. Therefore, it appears likely that the N terminus could be responsible for an isoform-specific function for GSK3A in mammals [56, 57]. Whether our observations suggest that GSK3A has an isoform-specific function during spermatogenesis and in mature sperm, or whether the lack of substitution of GSK3B for GSK3A in sperm is a gene dosage effect, remains to be verified. Testis-specific KO of GSK3 and other double-KO experiments are required to definitively conclude that GSK3A and GKS3B are not interchangeable in sperm. While the exact details of how GSK3A operates in sperm during epididymal sperm maturation and during fertilization should emerge from future studies, it is clear that GSK3A is essential for normal sperm function required for male fertility.