|Home | About | Journals | Submit | Contact Us | Français|
Males homozygous for the repro32 ENU-induced mutation produced by the Reproductive Genomics program at The Jackson Laboratory are infertile, have low epididymal sperm concentrations, and produce sperm with abnormally shaped heads and poor motility. The purpose of the present study was to identify the mutated gene in repro32 mice and to define the structural and functional changes causing infertility and the aberrant sperm phenotype. In repro32/repro32 mice, we discovered a failure to shed excess cytoplasm and disorganization of the middle piece of the flagellum at spermiation, resulting in the outer dense fibers being wrapped around the sperm head within a bag of cytoplasm. Using a candidate-gene approach, a mutation was identified in the spermatid-specific “capping protein (actin filament) muscle Z-line, alpha 3” gene (Capza3). CAPZA3 protein localization was altered in spermatids concurrent with altered localization of a unique CAPZB variant isoform and disruption of the filamentous actin (F-actin) network. These observations strongly suggest the missense mutation in Capza3 is responsible for the mutant phenotype of repro32/repro32 sperm and regulation of F-actin dynamics by a spermatogenic cell-specific CAPZ heterodimer is essential for removal of the cytoplasm and maintenance of midpiece integrity during spermiation in the mouse.
Spermiogenesis is the process during which spermatids undergo a series of complex morphological changes that result in formation of the highly specialized spermatozoon. In the mouse, as spermatids are remodeled from round (steps 1–8), to elongating (steps 9–12), and then to condensing (steps 13–16), they form an acrosome, assemble a flagellum, and reshape and condense their nucleus (reviewed by Kerr et al., 2006). During this process, spermatids become attached to Sertoli cells by an apical ectoplasmic specialization (ES) consisting of a heterotypic adherens junction. The ES are underlain in the Sertoli cytoplasm by hexagonal arrays of F-actin bundles sandwiched between a cistern of endoplasmic reticulum and the plasma membrane. Proteins associated with the ES include F-actin, members of the laminin and integrin adhesion molecule families, and transmembrane actin-complex binding molecules, including nectins and CEACAM6 (Yan et al., 2006; Ozaki-Kuroda et al., 2002; Kurio et al., 2008).
Near the end of spermiogenesis much of the spermatid cytoplasm in shunted into a lobule that is detached during spermiation and remains behind to be phagocytosed by Sertoli cells (Russell, 1979b). A small amount of cytoplasm remains in the head and neck region and forms the cytoplasmic droplet, which moves to the midpiece-principal piece junction and usually is shed in the epididymis. Aberrant retention of excess cytoplasm by human sperm is associated with impaired sperm motility, increased reactive oxygen species (ROS) production and decreased in vitro fertilization (IVF) success, resulting in sub-fertility or infertility (reviewed by Cooper, 2005).
The mechanisms responsible for removal of spermatid cytoplasm during spermiation in mammals are unclear. However, filamentous actin (F-actin) has been shown to be involved in this process in insects (Noguchi et al., 2003; Sahara and Kawamura, 2004). The potential role(s) of F-actin during spermiogenesis in mammals have been inferred from morphological studies, and include involvement in shaping the acrosome (Welch and O’Rand, 1985), attachment of the acrosome to the spermatid nucleus (Russell et al., 1986), and/or removal of cytoplasm during spermiation (Russell, 1979a; Russell, 1979b). In support of the latter proposed role, intratesticular injection of cytochalasin D to depolymerize actin resulted in retention of excess cytoplasm by testicular sperm along with disruption of tubulobulbar complexes (TBCs). Intratesticular injection of taxol to disrupt microtubules did not alter removal of cytoplasm at spermiation (Russell et al., 1989).
Unbiased approaches for identification of mutant phenotypes can be helpful in the dissection of complex processes such as spermatid cytoplasmic remodeling. The present study was undertaken to identify the mutation responsible for the aberrant sperm phenotype and male infertility in repro32 mice. The repro32 mutation was produced by the Reproductive Genomics program at The Jackson Laboratory (http://reproductivegenomics.jax.org/), using random N-ethyl-N-nitrosourea (ENU) mutagenesis and a breeding strategy to identify recessive mutations affecting male and/or female fertility (Handel et al., 2006; Lessard et al., 2007). Spermiogenesis in repro32/repro32 mice was abnormal and sperm were present in low concentrations in the epididymis, had abnormally shaped heads, exhibited poor motility, and showed very low IVF success. We identified a missense mutation in a spermatogenic cell-specific gene that is expressed solely in spermatids and encodes a protein involved in regulating F-actin dynamics. This study provides strong evidence that F-actin has an important role in removal of cytoplasm and maintenance of middle piece organization during spermiogenesis and spermiation, and that it is the disruption of F-actin organization during these processes that results in the structurally aberrant sperm and infertility of repro32/repro32 male mice.
The chromosomal localization of the repro32 mutation was determined by genome scanning with a panel of polymorphic microsatellite markers that discriminate between C57BL/6 and C3H chromosomal regions. Genotyping was performed by PCR on DNA extracted from tail tips of G3 progeny using standard conditions (Lessard et al., 2007). PCR products were analyzed on a 3% low EEO agarose (Fisher) gels. The map position was refined by breeding additional pairs of repro32/+ mice to generate offspring with recombinant chromosomes. The offspring were phenotyped by fertility testing and genotyped with additional polymorphic microsatellite markers to narrow the candidate region.
All animal procedures were performed in accordance with the NRC Guide for the Care and Use of Laboratory Animals and approved by the Animal Care and Use Committees of the National Institute of Environmental Health Sciences (NIEHS) and The Jackson Laboratory. In vitro fertilization (IVF) assays were performed as described previously (Eppig and O’Brien, 1996). Animals were euthanized and weighed and then one testis was removed, weighed, and immediately snap-frozen in liquid nitrogen and stored at −80°C for later extraction of protein or RNA. The other testis was processed for histology or transmission electron microscopy as described previously (Miki et al., 2002). Sperm were collected by carefully dissecting cauda epididymides to remove blood vessels and fat, making several small cuts with iridectomy scissors and allowing the sperm to swim out into 1X PBS (Ca2+/Mg2+-free) at room temperature. Sperm were prepared for scanning electron microscopy (SEM) and demembranated in 1% Triton X-100 as described previously (Miki et al., 2002; Miki et al., 2004). Sperm counts were done using a hemocytometer.
Genomic DNA was isolated from tail tip biopsies of male repro32/+, repro32/repro32, and founder strain (C57/Bl6 and C3H) mice using the DNeasy kit (Qiagen) according to manufacturer’s instructions. Primer pairs were designed to flank the exons of each of the candidate genes and PCR was used to amplify each exon and adjacent intronic sequence (sequences are available upon request). RNA was isolated from repro32/repro32 testes using the RNeasy kit (Qiagen), and 1 μg total RNA was reverse-transcribed into cDNA using oligo dT primers and MuLV reverse transcriptase (Applied Biosystems). PCR and RT-PCR products were purified using the QIAquick PCR purification kit (Qiagen) and sequenced using Big Dye Version 3.1 (ABI) on an Avant 3100 sequencer (Applied Biosystems), both according to the manufacturers’ instructions. The sequences obtained were identical to those in the public database.
RNA was isolated using Trizol reagent (Invitrogen) according to the manufacturer’s instructions, and cDNA was generated from 1 μg total RNA using oligo dT primers and MuLV reverse transcriptase (Applied Biosystems). 50 ng of cDNA was amplified for 35 cycles in a reaction containing: 10 mM Gene Amp PCR buffer II, 2 mM MgCl2, 1.5 mM dNTPs, 1.25 μM of each primer, and 1.25 U Taq Gold (Perkin-Elmer). Primers used for amplification of the unique Capzb_v3 5′ end (upper F: 5′-ACACGATGCATCCTAGCAGGC, upper R: 5′-CAGATGACAGGAGATCTTCACAC), the 3′ end common to all three known Capzb isoforms (lower F: 5′-TGAGTGACTGTTCCCCACAC, lower R: 5′-TGCTGCTTTCTCTTCAAGGC), and for cytoplasmic Actb (F: 5′-TCCGATGCCCTGAGGCTCTTTTC, R: 5′-CTTGCTGATCCACTATCTGCTGGAA) were designed to span introns to control for possible amplification of genomic DNA.
Protein was extracted by homogenization of wild type and repro32/repro32 testes as described previously (Goto and Eddy, 2004). Twenty μg of the insoluble fraction from whole testis extracts were boiled in SDS buffer for 10 minutes, separated on a 10% Tris-glycine gel (Biorad), and immunoblotted using standard procedures. Antibodies to ACTB (A-5441, Sigma), CAPZA3 (GP-SH4, Progen), and CAPZB3 (GP-SH5, Progen) were used at 1:10,000, 1:1,500, and 1:500, respectively. Two additional polyclonal antisera to CAPZA3 were the kind gifts of Dr. Roy Jones, Babraham Institute and Dr. Hiromitsu Tanaka, Osaka University. HRP-conjugated goat-anti-mouse (for ACTB, from Sigma) and donkey-anti-guinea pig (for CAPZA3 and CAPZB, from Jackson Immuno) secondary antibodies were each used at 1:20,000. Detection was done using the ECL Plus Detection System (Amersham).
Spermatids were prepared for immunostaining using a procedure modified from one described previously (Vogl et al., 1986). Briefly, repro32/+ and repro32/repro32 mice were euthanized and their testes removed, decapsulated, and placed into cold 1X PBS. The tubules were gently teased apart and then incubated in the presence or absence of 1 mg/ml (0.1%) trypsin (Sigma) for 15 minutes at 33°C. The tubules were washed 3 times with 1X PBS, fixed in 4% paraformaldehyde (in 1X PBS, pH 7.2) for 10 minutes at room temperature, and washed 3 more times with 1X PBS. The remaining incubations were performed at room temperature in a humid chamber. Short (~1 mm) tubule segments were placed onto a Superfrost Plus positively charged slide (Fisher) and permeabilized with Automation buffer (Biomeda) for 10 minutes. Blocking was done for 20 minutes in Automation buffer containing 4.5% horse serum, and tubule segments were then incubated with primary antibody for 1 hour. Following three washes in 1X Automation buffer, tubules were incubated in secondary antibody with or without fluorescent dye-conjugated phalloidin-568 (Invitrogen) diluted 1:25 (yielding approximately 265 nM final concentration, or 1 unit/section) for 1 hour in the dark. Three additional washes in Automation buffer were performed in the dark, and coverslips were mounted using Vectashield containing DAPI (Vector Laboratories) and sealed with clear nail polish. For phalloidin-only labeling, tubule segments were incubated following permeabilization with phalloidin for 1 hour in the dark, washed, and mounted as above. Images were obtained using a Zeiss LSM 5 Pascal UV inverted laser scanning microscope with an Omnichrome argon-krypton laser and a Zeiss Plan-Apo 100 (1.4 na) oil immersion objective. Images were scanned and analyzed using LSM PASCAL version 3.2 software.
Statistical analyses of body, testis, and seminal vesicle weights and of cauda epididymal sperm counts of repro32/+ and repro32/repro32 mice were conducted using Student’s t-test. Differences were considered significant when P<0.01.
Male C57BL/6J (herein designated B6) mice treated with ENU to produce point mutations were mated with C3HeB/FeJ (herein designated C3H) female mice to generate G1 progeny. The G1 male mice were mated with C3H female mice to produce G2 offspring, females of which were backcrossed to the G1 male to generate G3 progeny. Approximately 25% of the G3 mice were expected to be homozygous for any recessive mutation transmitted by the G1 male. The G3 mice were mated to wild type mice to identify infertility phenotypes (Handel et al., 2006; Lessard et al., 2007), which led to identification of the repro32 mutation. Further mating of G2 females to the G1 male determined that approximately 25% of males derived from these backcrosses were infertile. This process confirmed that the recessive mutation was transmissible and generated mice for initial low-resolution gene mapping (see below).
Initial analysis by the Reproductive Genomics Program found no apparent gross phenotypic differences between wild type, heterozygous (repro32/+), and homozygous (repro32/repro32) mutant mice (http://reproductivegenomics.jax.org/mutants/G1-510-4.html). In addition, the sperm from repro32/repro32 males had very low in vitro fertilization (IVF) capability in 7 separate experiments. Using sperm from 10 repro32/repro32 males and scoring an average of 65 oocytes per male (range 26 to 118), a mean of 4.2% of eggs were found to proceed to the 2-cell stage. Using sperm from 7 C57BL/6J males and scoring an average of 73 oocytes per male (range 49–88), a mean of 75.5% of eggs proceeded to the 2-cell stage. We confirmed that the repro32/repro32 males were infertile, had significantly decreased cauda epididymal sperm counts as compared to repro32/+ males (P<0.01), poor sperm motility (Table 1; Supplementary Movies), and abnormal sperm morphology (oligoasthenoteratozoospermia). The cauda epididymis contains a considerable amount of cell debris and sperm that are difficult to identify (Fig. S1). However, the general structure and organization of the seminiferous epithelium appeared normal until spermiogenesis, when abnormalities were apparent.
Condensed spermatids (step 16) become testicular sperm upon release during stage VIII of the seminiferous epithelium cycle. By stage IX, the seminiferous epithelium should contain only early elongating spermatids (step 9). However, both condensed spermatids and elongating spermatids were present during Stage IX in repro32/repro32 mice (Fig. S2). By stage X in repro32/repro32 mice, condensed spermatids were no longer present and the seminiferous epithelium appeared comparable to that in wild type mice (data not shown). Although condensed spermatids were retained, there were no overt differences in the appearance or orientation of early elongating spermatids (step 9) in repro32/repro32 during stage IX tubules (Fig. S2).
During spermiogenesis, the outer dense fibers (ODFs) develop from proximal to distal around the microtubule-containing axoneme, the mitochondria migrate into the flagellum and surround the ODFs to form the midpiece, and the fibrous sheath develops from distal to proximal to delineate the principal piece (Irons and Clermont, 1982a; Irons and Clermont, 1982b). In the latter steps of these processes, most of the cytoplasm is redistributed from around the head and flagellum into a narrow-necked lobule that is shed during spermiation (Clermont et al., 1993). The first overt change in spermatogenic cells of repro32/repro32 mice seen by transmission electron microscopy (TEM) was in late condensing spermatids (step 15–16), with a dramatic disruption of structural organization becoming apparent. The middle piece region of the flagellum was disorganized and its structural components surrounded the sperm head within an abnormal accumulation of cytoplasm (Fig. 1A and S3A), in contrast to what was observed in spermatids of wild type mice at the same stage (Fig. 1B). These features also were seen in repro32/repro32 sperm in the cauda epididymis (Figs. 1C and S3B), but not in wild type sperm (Fig. 1D). Cytoplasmic droplets (CDs) were present at the junction of the midpiece and principal piece of the flagellum of a majority of sperm from the cauda epididymis of repro32/+ mice, but were absent from repro32/repro32 sperm (Fig. 2 and Supplemental movies).
By scanning electron microscopy (SEM), the plasma membrane was seen to be closely applied to the head of wild type sperm (Fig. 3A), while a bag of cytoplasm loosely encased the head of repro32/repro32 sperm (Fig. 3B). It was seen in repro32/repro32 sperm demembranated by detergent treatment that ODF and other flagellar components were disorganized and wrapped around the sperm nucleus (Fig. 3D).
Genome scans using 2–3 polymorphic microsatellite markers per autosomal chromosome were performed on DNA from G3 progeny males. A region on Chr 6 between D6Mit254 and D6Mit14 homozygous for B6 genomic DNA in all infertile repro32 males was identified as the candidate region for the ENU mutation. This was confirmed by further genotyping for polymorphic microsatellite markers flanking the candidate region. The D6Mit254 and D6Mit14 markers were used for subsequent progeny genotyping and colony maintenance. Matings were carried out between heterozygous progeny for several generations to produce additional Chr 6 recombinants. A total of 353 mice were genotyped for the flanking markers and 65 recombinant individuals were identified. Typing of the recombinant individuals for microsatellite markers within this region allowed narrowing of the candidate region for the repro32 mutation to an 8-Mb genomic interval between D6Mit291 and D6Mit14, defined by 11 independent recombination events.
The 8-Mb genomic interval on Chr 6 contains 43 known and 15 hypothetical genes. Searches of the literature, EST databases, and GEO profiles were used to identify candidate genes likely to be expressed exclusively or principally in spermatogenic cells. The number of genes potentially responsible for the repro32 phenotype was reduced by excluding genes that were: 1) knocked out by gene targeting [14 genes], 2) knocked down by transgenic RNAi [1 gene] (Knott et al., 2005), 3) not expressed in the testis [13 genes], or 4) expressed in numerous tissues [11 genes]. The knocked out or knocked down genes were excluded because the phenotype produced was unlike the repro32 phenotype. Four genes (Igbp1b, Tuba3b, Ifltd1 and Capza3) remained as candidates based on these exclusion criteria. Exons and intron-exon junctions of Igbp1b, Tuba3b, and Capza3 were sequenced using genomic DNA isolated from tail tissue of repro32/repro32 male mice, while Ifltd1 was amplified and then sequenced from both genomic DNA from tail tissue and repro32/repro32 testis cDNA.
Mutations were not found in Igbp1b, Tuba3b, or Ifltd1, but an A/T nucleotide transversion resulting in a methionine-to-lysine missense mutation (M44K) was found in the gene for “capping protein (actin filament) muscle Z-line, alpha3” (Capza3; UniGene Mm.12817) in repro32/repro32 mice. CAPZA functions in somatic cells as a heterodimer with a CAPZB subunit to form a functional CAPZ complex capable of binding to the barbed ends of actin filaments with high affinity and preventing the addition or loss of actin subunits (Wear and Cooper, 2004). The mutation was confirmed by sequencing DNA from the B6 and C3H parental strains (Fig. 4A). The Capza3 gene lacks introns and encodes a 298 amino acid protein shown by northern analysis (Tanaka et al., 1994) and UniGene EST profiles (http://www.ncbi.nlm.nih.gov/UniGene/ESTProfileViewer.cgi?uglist=Mm.12817) to be expressed specifically in testis and by microarray analysis data to be expressed specifically in spermatids (Shima et al., 2004; http://mrg.gs.washington.edu/index.cgi?mrg_id=85248). The CAPZA3 protein is highly conserved, with at least 90% identity and 96% similarity at the amino acid level occurring in 7 eutherian mammals for which the CAPZA3 sequence has been determined. The M44K substitution in CAPZA3 of repro32 mutant mice occurred in a 39 amino acid region with 100% identity in these species (Fig. 4B).
There were no apparent differences in the amount of CAPZA3 in the insoluble fraction of lysates from testes of repro32/repro32 and wild type mice (Fig. 5C). In repro32/+ spermatids, CAPZA3 was detectable first in a narrow arc at the periphery of the developing acrosome in step 8 round spermatids (Fig. 5A.i). It surrounded the acrosome in elongating (Fig. 5A.ii) and early condensing spermatids (Fig. 5A.iii) and localized mainly in the post-acrosomal region of the sperm head in late condensing spermatids (Fig. 5A.iv). This pattern persisted in testicular and epididymal spermatozoa, which is consistent with the distribution of CAPZA3 seen in rat spermatids and testicular spermatozoa (Hurst et al., 1998) and in human spermatozoa (Miyagawa et al., 2002).
The localization of CAPZA3 in round spermatids from repro32/repro32 mice was similar to that in repro32/+ mice (Fig. 5B.i). However, in elongating spermatids it remained associated mainly with the convex margin of the acrosome (Fig. 5B.ii) and in condensing spermatids it failed to achieve the consistent pattern of localization seen in repro32/+ mice (Fig. 5B.iii–vi). The altered localization of CAPZA3 in isolated elongating and condensing spermatids from repro32/repro32 mice preceded the appearance of the overt mutant phenotype seen by TEM (Fig. 1A). The results shown were obtained with a commercial CAPZA3 antiserum (GP-SH4), but similar results were observed with antisera to CAPZA3 kindly provided by Drs. Jones and Tanaka (data not shown).
Because CAPZA forms a complex with CAPZB in somatic cells, CAPZB was investigated in wild type and repro32/repro32 mice. Two well-characterized mouse CAPZB isoforms are expressed in muscle or ubiquitously and are products of alternative transcripts from the Capzb gene on Chr 4 (Schafer et al., 1994). A third CAPZB isoform isolated from bovine sperm is the product of another alternative transcript abundant in testis and detectable in brain (von Bülow et al. 1997). An RT-PCR fragment generated from mouse brain cDNA had high identity with a region of this bovine sequence (von Bülow et al. 1997).
To determine which CAPZB isoforms are present in mouse testis, RT-PCR was performed with primers designed from the sequences of the two known alternative transcripts (Capzb_v1 and Capzb_v2) (Fig. 6A). The Capzb_v2 transcript differs from Capzb_v1 due to an alternative splicing event that removes 113 nucleotides within the C-terminal exon (Schafer et al., 1994). Capzb_v1 was identified in cDNA from skeletal muscle and at a very low level in Sertoli cells, while the Capzb_v2 sequence was distributed more widely (Fig. 6B). The possible presence of Capzb_v3 in the mouse testis was investigated at selected ages from 15 to >60 days post partum (dpp) using a primer set with the 5′ primer located within the mouse sequence corresponding to the bovine gene alternative first exon (von Bülow et al., 1997) (labeled 1b, Fig. 6A). A Capzb_v3 sequence was first detected on day 22, coinciding with the first wave of round spermatid development. Using these primers and RNA isolated from highly enriched populations of germ cells (a gift of Dr. John McCarrey, University of Texas at San Antonio), Capzb_v3 transcripts were detected in round spermatids, but not in pachytene spermatocytes or in the other mouse tissues examined (Fig. 6B). These amplification products were sequence-verified (data not shown), and confirmed the predicted usage of an alternate exon 1 (von Bülow et al. 1997) downstream of somatic exon 1 as well as exon 9 from Capzb_v2 (Fig. 6A). Therefore, the isoform encoded by mouse Capzb_v3 (referred to hereafter as CAPZB3) differs from the other CAPZB isoforms by the presence of an N-terminal extension resulting from utilization of an alternative first exon and is detectable only in round spermatids (GenBank accession no. FJ692320).
The CAPZB3 isoform was characterized using antisera specific to its unique N-terminal extension. Immunoblot analysis determined that a protein of approximately 36 kDa (predicted = 33.8 kDa) was present in the insoluble portion of lysates of whole testis from wild type mice (Fig. 5C). Spermatids were isolated mechanically and treated with trypsin to remove Sertoli cell remnants and examined by indirect immunofluorescence microscopy to determine the location of CAPZB3. The localization pattern of CAPZB3 was similar to CAPZA3 in elongating and condensing spermatids (data not shown) and in testicular sperm from repro32/+ mice (Fig. 6C). We found that CAPZB3 protein distribution was altered in spermatids concurrently with the alteration of CAPZA3 (Fig. 5A) in elongating and condensing spermatids (data not shown) and in testicular sperm from repro32/repro32 mice (Fig. S4A.i–ii). Because CAPZA and CAPZB function as heterodimers in somatic cells, we postulate that CAPZA3 and CAPZB3 form a novel spermatid-specific CAPZ heterodimer.
The role of the CAPZ complex in regulating actin filament growth (reviewed by dos Remedios et al., 2003) and the spermatid-specific expression of Capza3 in the mouse (Tanaka et al., 1994; Yoshimura et al., 1999) and rat (Hurst et al., 1998) suggested that a mutation in CAPZA3 would lead to perturbation of F-actin structures in repro32/repro32 spermatids. Bundles of F-actin are seen at two sites in the luminal region of the mouse seminiferous epithelium. One is within the Sertoli cell ES, which is composed of closely packed F-actin filaments and a cistern of endoplasmic reticulum. The ES is present in the cytoplasm adjacent to junctional complexes that anchor developing spermatids to Sertoli cells (Russell, 1977). This complex also has been implicated in spermatid translocation within the seminiferous epithelium prior to spermiation (reviewed by Vogl et al., 2000). The ES F-actin organization was examined using phalloidin on mechanically isolated spermatids to which Sertoli cell remnants remained attached. No differences in ES F-actin fiber quantity, orientation, or appearance were apparent between spermatids from repro32/+ and repro32/repro32 mice (Fig. 7A–B), and the patterns were similar to those reported previously (Vogl et al., 1985; Young et al., 2008).
The other location where F-actin is seen is within developing spermatids. To view the F-actin in spermatids unobscured by Sertoli cell ES F-actin, seminiferous tubules were mechanically disrupted and incubated with trypsin to separate Sertoli cell remnants from developing spermatids (Vogl et al., 1986; Masri et al., 1987) prior to fixation and labeling with phalloidin. The ES removed from condensing spermatids can be seen after trypsin treatment (Fig. S5A–C). The F-actin fibers first became detectable within round spermatids (~step 8) as a band between the cell surface and the developing acrosome (Fig. 8A.i). In elongating spermatids (steps 9–12), F-actin is present in a crescent overlying the convex aspect of the acrosomal region (Fig. 8A.ii), and in condensing spermatids (~step 13) the crescent begins to resolve into a distinct band along the convex margin with a diffuse accumulation over the lateral aspect of the acrosomal region (Fig. 8A.iii). The F-actin in late condensing spermatids (steps 14–16) was seen in a distinct band along the convex margin and dispersed unevenly over the lateral aspect of the acrosomal region (Fig. 8A.iv). However, F-actin was no longer detectable with phalloidin in testicular (Fig. 8A.v) or epididymal sperm (data not shown), consistent with observations in other studies (e.g., Halenda et al., 1987; Flaherty et al., 1986).
The distribution of F-actin fibers in round spermatids from repro32/repro32 mice was similar to those from repro32/+ mice, although they appeared thinner and shorter (Fig. 8B.i). The differences become more apparent in elongating spermatids, with the F-actin fibers being fewer, shorter, and more tangled in repro32/repro32 mice and located more towards the anterior-dorsal portion of the head (Fig. 8B.ii) than in elongating spermatids from repro32/+ mice (Fig. 8A.ii). Despite the differences in F-actin distribution in elongating spermatids from repro32/repro32 and repro32/+ mice, their overall morphological appearances were comparable. Differences became apparent at the TEM level during spermiation, after differences were seen in F-actin organization between repro32/repro32 (Figs. 8B.iii–vi) and repro32/+ spermatids (Figs. 8A.iii–v). However, the shape of nuclei in spermatids and sperm of repro32/repro32 mice did not differ significantly from those of wild type mice, suggesting that F-actin does not have a major role in shaping the mouse sperm nucleus (Fig. 7A & 7B).
This study identified a missense mutation in the Capza3 gene in repro32 mice produced by ENU mutagenesis and a screening protocol to discover mutations causing infertility. The mutant phenotype was inherited by repro32 offspring in the expected Mendelian ratios and the mutation was found using a candidate gene sequencing approach. CAPZA3 mislocalization occurred in condensing spermatids coincident with mislocalization of CAPZB3 and disruption of the F-actin network. The repro32/rero32 spermatids failed to shed excess cytoplasm during spermiation and the axoneme, mitochondria, and ODFs of the middle piece were present in a tangled array within a bag of cytoplasm surrounding the sperm head.
ENU mutagenesis provides a powerful and nonbiased forward genetics approach to identify novel genes involved in male and female infertility (Harris et al., 2007; Ward et al., 2007; Philipps et al., 2008). The missense mutation in Capza3 was identified by sequencing exons and exon-intron junctions of candidate genes. While there remains a possibility that an additional mutation might contribute to the phenotype, the combined results from 4 studies suggest there is less than a 5% likelihood of another mutation occurring in the 8-Mb critical region (Keays et al., 2006). However, we expect the likelihood of another mutation causing the repro32 phenotype to be considerably less than 5% because mutations in exons may be silent and much of the 8-Mb region consists of intragenic and intronic sequence.
A recent study identified 4 QTLs involved in controlling sperm head shape (L’Hôte et al., 2007). One of these (Sh1) mapped to the region on chromosome 6 containing Capza3 and three other testis-specific genes (Igpb1b, Plcz1, and Tuba3b) identified as potential candidate genes affecting the trait. However, the abnormal phenotype associated with the head-shape QTLs was a hammerhead shaped sperm nucleus, while the nuclear shape in repro32/repro32 sperm is relatively normal, suggesting that Capza3 alone cannot determine the head shape trait defined by the QTLs. However, it is possible that different Capza3 alleles and/or interactions with genes in the other QTLs are associated with the trait.
The CAPZA and CAPZB heterodimer (CAPZ) regulates actin dynamics by preventing addition or loss of G-actin subunits to the fast-growing or barbed end of F-actin fibers (reviewed by dos Remedios et al., 2003; Wear and Cooper, 2008). Previous studies determined the intronless Capza3 gene encodes a testis-specific isoform of CAPZA that is expressed in spermatids in mouse (Tanaka et al., 1994; Yoshimura et al., 1999) and rat (Hurst et al., 1998), and a testis-specific isoform of CAPZB is present in bull sperm (von Bülow et al., 1997). The current studies: 1) confirmed Capza3 gene expression is confined to spermatids, 2) demonstrated the Capzb_v3 variant transcript and CAPZB3 protein are expressed in spermatids, 3) determined the CAPZA3 and CAPZB3 proteins are present in spermatids and sperm, 4) demonstrated F-actin filaments are present in spermatids but not in sperm of the mouse, as is seen in other species (Welch and O’Rand, 1985; Vogl et al., 1986; Halenda et al., 1987), and 5) observed that F-actin, CAPZA3 and CAPZB3 are found in the same cytoplasmic regions in condensing and elongating spermatids. Taken together, these results strongly suggest that a spermatid-specific CAPZA3/CAPZB3 heterodimer plays a significant role in regulating F-actin dynamics in condensing and elongating spermatids.
We have shown the organization and distribution of CAPZA3, CAPZB3, and F-actin is disrupted in condensing and elongating spermatids of repro32/repro32 mice. The M44K missense mutation occurs in the N-terminal portion of CAPZA3 with nearly 100% identity among the eutherian mammals for which the sequence is known. This implies the novel N-terminal region imparts a conserved spermatid-specific function, interaction, or localization for CAPZA3 that would be subject to disruption by the substitution of a positively charged lysine for a hydrophobic methionine.
There are several possibilities for how the missense mutation in CAPZA3 would result in the mutant sperm phenotype: 1) Substitution of a positively charged lysine for a hydrophobic methionine (M44K) might disrupt the tertiary structure of CAPZA3. Because misfolded proteins often are targeted to the proteasome for elimination (reviewed by Goldberg and Dice, 1974), we examined CAPZA3 protein levels by immunoblot analysis. However, there was no apparent difference in the amount of CAPZA3 in lysates of testes from wild type and repro32/repro32 mice; 2) CAPZA3 and CAPZB3 have novel N-terminal regions and the M44K mutation might disrupt assembly of the CAPZ heterodimer or its interaction with F-actin. Arguing against this is the finding that the N-terminus of chicken CAPZA1 is not required to dimerize with CAPZB or to bind F-actin (Casella and Torres, 1994); 3) The novel CAPZA3 and CAPZB3 N-terminal regions might target or anchor CAPZ in spermatids. In addition, novel N-terminal regions are a common feature of spermatid- and sperm-specific proteins and are thought to be involved in localizing or anchoring these proteins (e.g., Vemuganti et al., 2007). Although we favor the latter possibility for how the missense mutation in CAPZA3 produces the repro32/repro32 sperm phenotype, the results are correlative and further studies will be needed to define the effects of the substitution of a lysine for a methionine on the structure and/or function of CAPZA3.
It previously was hypothesized that up to 70% of the spermatid cytoplasmic fluid volume is taken up by Sertoli cells through the TBC (Russell, 1979b; Russell and Malone, 1980). Most of the remaining cytoplasm is pinched off during spermiation to form the residual body that is phagocytosed by Sertoli cells (reviewed by Kerr et al., 2006. A small amount of cytoplasm remaining at the junction of the sperm head and neck becomes the cytoplasmic droplet that moves down the sperm flagellum as sperm traverse the epididymis (reviewed by Cooper, 2005).
An abnormal retention of cytoplasm can be seen by TEM in step 15–16 repro32/repro32 condensing spermatids in the region that will become the head and neck of the sperm. In addition, a greater than normal amount of cytoplasm remains associated with repro32/repro32 sperm following spermiation. Typical residual bodies are not apparent in the seminiferous epithelium, but vesicles are abundant in the epididymis (Supplementary Movie) and may represent residual cytoplasm shed from sperm after spermiation, as well as cytoplasm-encased sperm heads. Cytoplasmic droplets (CDs) present on most wild type and repro32/+ cauda epididymal sperm are absent from repro32/repro32 sperm. If the hypothesis for cytoplasmic fluid removal is correct, disruption of F-actin distribution or function in repro32/repro32 condensing spermatids might result in defective TBC function as well as disruption of cytoplasmic removal at spermiation.
The results of this study suggest: 1) the repro32 phenotype is caused by a missense mutation in the Capza3 gene, 2) CAPZA3 is required for maintenance of male fertility in the mouse and probably other mammalian species, 3) a novel CAPZA3/CAPZB3 heterodimer regulates F-actin dynamics in spermatids, 4) F-actin has an essential role in formation and stabilization of the nascent spermatozoon, and 5) F-actin does not have a major role in shaping the sperm head. They also raise the question of what roles are served by the novel N-terminal regions of CAPZA3 and CAPZB3. Finally, it should be noted that testis-specific genes expressed in spermatids have been identified for the actin-like proteins ACTL7A and ACTL7B (Tanaka et al., 2003) and actin-related proteins ACTRT1 and ACTRT2 (Heid et al., 2002) and it remains to be determined if they also serve roles in spermiogenesis and spermiation.
Figure S1: Thin sections of the cauda epididymis from wild type (A) and repro32/repro32 (B) mice. Nearly all of the sperm in B exhibit the abnormal wrapping of flagella around their heads (as seen in Figs. 1 and and3);3); several examples are identified with white arrows.
Figure S2: Thin sections of the seminiferous epithelium from wild type (A) and repro32/repro32 (B) testes. Both stage IX tubules contain typical step 9 elongating spermatids with their acrosomes correctly oriented towards the basal aspect of the tubule (indicated by white arrows). Delayed spermiation is seen in the repro32/repro32 testis, with condensed step 16 spermatids aberrantly retained (circled by dashed white line) within the apical portion of the epithelium adjacent to the lumen. Residual bodies derived from spermatid cytoplasm engulfed by Sertoli cells and located near the tubule lumen are intensely stained by the basic dye toluidine blue used on plastic sections.
Figure S4: CAPZB3 protein localization in condensing spermatids. (A) IIF using anti-CAPZB3 (green) on repro32/+ early (A.i) and late (A.ii) condensing spermatids. (B) The same as in (A), except for repro32/repro32 spermatids. Top panels = merged fluorescent and DIC images, bottom panels = fluorescent images for CAPZB3 only. DAPI-labeled nuclei are in blue. Scale bars = 2μ.
Figure S5: The ES structures are detached from developing spermatids by trypsinizing mechanically disrupted seminiferous epithelia. (A) IF using dye-conjugated phalloidin (568, red) on a merged fluorescent and DIC image, (B, C) IF using dye-conjugated phalloidin (488, green) on a merged fluorescent and DIC (B) and a fluorescent-only image, respectively. DAPI-labeled nuclei are in blue. Scale bars = 2μ.
We would like to extend our sincere appreciation to Drs. Fanny Odet and Lisa Chadwick for helpful discussions, Gina Goulding and Linwood Koonce for technical assistance, and to Drs. John Schimenti and Jason Mercer for critical reading of the manuscript. This research was supported in part by the Intramural Research Program of the NIH, National Institute of Environmental Health Sciences (ZO1-ES070076, EME) and a grant from NIH, NICHD (PO1 HD-42137, MAH).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.