By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
During mammalian spermiogenesis, spermatid head shaping takes place in the presence of the acrosome-acroplaxome complex and the transient microtubular manchette. Transgenic mouse models support the hypothesis that the acrosome-acroplaxome-manchette complex may impact on spermatid head shaping (reviewed in Kierszenbaum and Tres,2004; Kierszenbaum et al.,2007). Examples are the acrosome-deficient Hrb (Kang-Decker et al.,2001; Kierszenbaum et al.,2004), the GOPC mutant, displaying spermatids with spherical heads (Yao et al.,2002), and the azh mutant, characterized by head shaping abnormalities, ectopic manchette, head dislocation, and eventual decapitation (Mochida et al.,1999). These observations emphasize a need for a more extensive search of genes involved in sperm head shaping to understand the pathogenesis of round-headed sperm (globozoospermia) leading to infertility (Christensen et al.,2006).
Several aspects of acrosome biogenesis have been previously described (Moreno et al.,2000; Ramalho-Santos and Moreno,2001). Acrosome development occurs in close proximity to the acroplaxome, an F-actin-keratin 5/Sak57-containing cytoskeletal plate outlined by a marginal ring (Kierszenbaum et al.,2003a). The acroplaxome plate corresponds in part to the perinuclear theca (Clermont et al.,1993). One of the distinguishing features of the acroplaxome from the perinuclear theca is the acroplaxome marginal ring, a tripartite desmosome-like assembly that fastens the descending recess of the acrosome to the spermatid nuclear envelope-dense nuclear lamina unit (Kierszenbaum et al.,2003a,b,2007). The acroplaxome is the docking site of Golgi-derived proacrosomal vesicles transported during acrosome development (Kierszenbaum et al.,2003a,b,2004). Golgi-derived proacrosomal vesicles are transported to the acroplaxome site by F-actin- and microtubule-driven molecular motors (actin: myosin Va, Kierszenbaum et al.,2003b; microtubules: kinesin KIFC1, Yang and Sperry,2003). Myosin Va-driven proacrosomal vesicles involve a linker bridging myosin Va to Rab proteins Rab27a/b (Kierszenbaum and Tres,2004). Rab27a/b belong to the monomeric GTPases of the Ras superfamily (reviewed in Seabra and Coudrier,2004). It appears that two different molecular motor systems using F-actin and microtubule tracks participate in the transport of proacrosomal vesicles to the acroplaxome. F-actin is also found in Sertoli cell hoops encircling the elongating spermatid head.
We have addressed the question of whether F-actin undergoes a molecular rearrangement during spermatid head shaping by focusing on the non-receptor tyrosine kinase Fer. Fer is a member of the Src family of tyrosine kinases regulating numerous cellular processes, including cytoskeletal reorganization, cell adhesion, vesicular transport, and intracellular signaling (reviewed in Greer,2002). Fer tyrosine kinase is of interest for two reasons. First, Fer somatic (designated FerS) and testicular-type (designated FerT) transcripts are expressed in testis (Keshet et al.,1990; Fischman et al.,1990), thus suggesting a role in spermatogenesis. Second, cortactin, an F-actin-bound substrate for Src tyrosine kinases (including Fer, Fyn, and c-abl tyrosine kinases), can modulate the dynamics of the F-actin network by cortactin tyrosine phosphorylation (Kanner et al.,1991; Wu et al.,1991; Fan et al.,2004). The presence of Fer tyrosine kinase in Sertoli cell-spermatid cell adhesion has been reported and its functional roles during spermatogenesis discussed (Maekawa et al.,2002; Chen et al.,2003). However, the localization of Fer tyrosine kinase in spermatids, including the acroplaxome, has not been determined.
Preliminary observations on the distribution of FerT tyrosine kinase and phosphorylated cortactin in the acroplaxome have been published (Kierszenbaum,2006; Kierszenbaum et al.,2007). Here we report the expression of novel FerT transcript isoforms in pachytene spermatocytes and round spermatids. We show FerT antigenic sites in the Golgi and propose the involvement of this tyrosine kinase in acrosome development. Phospho(Tyr421)-cortactin coexist in the acroplaxome and Sertoli cells F-actin-containing hoops embracing the elongated spermatid head, supporting a role of phosphorylated cortactin in the modulation of F-actin dynamics during spermatid head morphogenesis.
Characterization of FerS and FerT Transcripts
A comparison of the available amino acid sequence of FerT and FerS proteins showed very high sequence conservation between rat, mouse, and human (Supp. Fig. 1, which is available online). For example, FerT and FerS proteins in rats and mice are 94 and 98% identical, respectively. BLAST analysis of published mRNAs confirmed that FerT and FerS are products of the same gene in all species (Supp. Fig. 2). The last ten exons at the carboxyl-end are shared in FerT and FerS and encode 411 amino acids. The difference between these two forms is at the amino-end. FerT has four specific exons, but translation starts only in the last one, resulting in a 43 FerT-specific amino acid domain. There are eight FerS-specific exons, encoding 412 amino acids. Interestingly, the coding exon 8 of FerS coincides with the non-coding exon 2 of FerT.
A comparative RT-PCR analysis of Fer mRNAs at the 5′-ends showed that both FerT and FerS forms are expressed in testis and pachytene spermatocytes (Fig. 1A, B). Screening the pachytene spermatocyte and round spermatid cDNA expression libraries detected FerS transcripts in pachytene spermatocytes at a very low level; transcripts were detected only by repeated RT-PCR (Fig. 1B). FerS transcripts were not detected in round spermatids even by repeated RT-PCR. Surprisingly, the FerT band pattern in adult testis differed from the pattern in pachytene spermatocytes and round spermatids (Fig. 1A). It should be noted that FerT isoforms seen in pachytene spermatocytes and round spermatocytes are not seen in the testis sample (Fig. 1A). This result can be attributed to the common observation that when several mRNA/cDNA templates are available in a single RT-PCR reaction, the shortest ones are amplified preferentially with the possible exclusion of the larger templates. The observed results are also dependent on the relative concentration of the templates that are relatively abundant in the screened pachytene spermatocyte and round spermatid cDNA libraries. Sequencing of all the FerT bands revealed developmental changes in splicing of the Fer primary transcripts (Supp. Fig. 3). Figure 1C shows an additional exon (106 bp) present between exons 3 and 4 in pachytene spermatocytes and round spermatids. In round spermatids, a longer variant of this exon (133 bp) is also present as well as transcripts without this extra exon. In addition, there is a small (4 bp) deletion at the beginning of the exon 4 (Fig. 1C, vertical red line) in some of the transcripts. All these changes occur upstream of the translation start site (in exon 4). Therefore, they should not affect the FerT protein sequence directly. The changes may affect the stability of resulting transcripts or be related to efficiency of translation.
Western blot analysis defined the specificity of our anti-FerS and anti-FerT antibodies (Fig. 2). These antibodies recognized Fer proteins in rat, mouse and human tissues. FerS (94kDa) was detected in somatic tissues and testis (Fig. 2A), whereas FerT (51kDa) was predominant in testis (Fig. 2B). A similar tissue specific distribution of FerS and FerT was reported by Fischman et al. (1990) and Hazan et al. (1993). Specificity of the antisera was confirmed by competition studies. Figure 2C shows a complete block of anti-FerT serum pre-incubated with FerT-specific peptide (Fig. 2, lane B) in comparison with non-preincubated affinity purified serum (Fig. 2, lane A) or after pre-incubation of affinity purified anti-FerT serum with a non-specific peptide (Fig. 2, lane C).
Localization of FerT in Rat Spermatids
Confronted with the undetectable levels of FerS transcripts in spermatids, our next approach was to focus on FerT to determine its temporal and spatial distribution during rat spermiogenesis. FerT was initially observed in the Golgi region adjacent to the acroplaxome-nuclear pole and at the periphery of the acroplaxome (S4-S6 spermatid; Fig. 3A,B). The localization of FerT in the Golgi was confirmed by immunogold electron microscopy (see below). FerT immunofluorescence was seen in the manchette region (S8-S9 spermatids; Fig. 3C,D). S12 elongating spermatids displayed two characteristic FerT immunoreactive sites: the acroplaxome and the manchette (Fig. 3E,F). The FerT immunoreactive acroplaxome was still visible in S19 spermatids (Fig. 3G,H), in addition to a broad cytoplasmic region adjacent to the concave face of the spermatid head. This cytoplasmic region was regarded as containing tubulobular complexes (see below) extending from the spermatid cytoplasm into narrow clefts in the adjacent Sertoli cell (Russell and Clermont,1976). The persistence of the Sertoli cell tubulobulbar complex region attached to spermatids highlights its stress-resistant nature in spite of the hypotonic effect of the fixative.
Additional data derived from immunogold electron microscopy. Figure 3I (S7 spermatid) shows FerT immunoreactivity in the media/trans-Golgi, Golgi-derived vesicles, and along the outer acrosome membrane (denoted by red circles) and inner acrosome membrane, the latter closely associated to the acroplaxome complex (denoted by dashed blue circles). Figure 3J (S10 spermatid) illustrates a Sertoli cell cytoplasmic process containing F-actin bundles interposed between two adjacent spermatids with their acrosomes facing each other. FerT immunoreactivity was associated with Sertoli cell F-actin bundles (indicated by arrowheads) and the acroplaxome (denoted by dashed blue circles). The distribution sites suggested that FerT may participate in acrosome development and also become cargo passengers along the acrosomal membrane before gaining access to the acroplaxome.
Distribution Sites of Phospho(Tyr421)-Cortactin and Unphosphorylated Cortactin in Spermatids
Cortactin is a prominent substrate of the Src family of cytoplasmic tyrosine kinases (including Fer, Fyn, and c-abl) and a cross-linking protein of F-actin (reviewed in Daly,2004), one of the components of the acroplaxome. There are three tyrosine phosphorylation sites in cortactin (Tyr421, Tyr466, and Tyr482). Our next step was to determine whether phosphocortactin was present in the acroplaxome and shared similar localization sites with FerT tyrosine kinase. The presence of phosphorylated cortactin could indicate sites where F-actin remodeling was taking place during the round-to-elongated transition of the spermatid head. We used a polyclonal antibody produced against a synthetic cortactin peptide phosphorylated at Tyr421. The antibody was affinity purified and adsorbed against the non-phosphorylated peptide to remove immunoglobulins that may crossreact with the non-phosphorylated form of the protein. Figure 4A–C shows a rat S8-S9 spermatid with its acrosome-acroplaxome unit and subjacent manchette. Anti-phospho(Tyr421)-cortactin stains the acrosome-acroplaxome region but not the manchette region (Fig. 4D–F, S12 spermatid). Figure 4G,H displays S19 rat spermatids displaying two characteristic phospho(Tyr421)-cortactin immunoreactive sites: the acroplaxome and the tubulobulbar complex region (TbCR). Figure 4I,J shows that anti-β-actin stains the acroplaxome and the TbCR of a S19 rat spermatid. Therefore, a correlation exists in the distribution sites of FerT, phospho(Tyr421)-cortactin, and actin in the acroplaxome and the TbCR. In contrast, FerT is detected in the manchette but phosphorylated cortactin is not apparent. This difference was clarified when a monoclonal antibody to unphosphorylated cortactin was used. Figure 4K–M demonstrates that unphosphorylated cortactin is present in the acroplaxome and the manchette. Phosphorylated cortactin was not apparent in the manchette, even though it contains mainly microtubules but also F-actin (Mochida et al.,1998,1999; Kierszenbaum et al.,2003b). An explanation for this discrepancy needs to be determined.
We report here the presence of the FerT non-receptor tyrosine kinase in the Golgi, acrosome membranes, and acroplaxome of spermatids. It was previously reported that FerT is restricted to spermatocyte nuclei (Keshet et al.,1990; Fischman et al.,1990; Hazan et al.,1993; Orlovsky et al.,2000). Our data extend their earlier observations to specific cytoplasmic sites in spermatids. We also show that FerT colocalizes with phosphorylated cortactin in the acroplaxome. Cortactin phosphorylation is tightly coupled to the kinetic properties of F-actin. F-actin is a component of Sertoli cell apical cortical cytoplasm adjacent to elongating spermatids and the acroplaxome, a cytoskeletal plate that stabilizes the acrosome during spermatid head shaping. Although cortactin is a well-known tyrosine kinase substrate, a specific role of FerT in cortactin phosphorylations has not been determined. FerS and FerT have significant autophosphorylation activity (Orlovsky et al.,2000). We also report for the first time the expression of FerT transcript variant in rat pachytene spermatocytes and round spermatids. An unexpected finding was a very low level of FerS transcripts in pachytene spermatocytes and the absence of FerS transcripts in round spermatids. In contrast, pachytene spermatocytes and round spermatids express FerT transcript isoforms at a level comparable with that in adult testis but with nucleotide changes at the 5′-ends that do not affect the amino acid sequence of the proteins. In agreement with the RT-PCR data, anti-FerS serum failed to detect immunoreactive sites in spermatids whereas spermatocytes were not significantly stained (whole cells and seminiferous epithelium preparations; data not shown).
Fer is the only other known member of the proto-oncogene fps/fes (fps, Fujinami poultry sarcoma; fes, feline sarcoma) encoding a structurally distinct member of the non-receptor protein-tyrosine-kinase characterized by a Src-homology-2 (SH2) modular protein domain. FerS contains an amino-terminal FCH (for Fps/Fes/Fer/CIP4 homology) domain implicated in the regulation of cytoskeletal rearrangements and vesicular transport. The FCH domain is not present in FerT. Instead, a unique 44–amino acid amino terminal sequence is observed. This difference suggests that FerS and FerT can bind to distinct substrates with diverse biological function, a possibility highlighted by the prevalence of FerT in pachytene spermatocytes and round spermatids. The coexistence of FerS, albeit at a low level, and FerT transcripts in pachytene spermatocyte suggests that the two forms of Fer are required during meiotic prophase. This does not seem to be the case in round spermatids where FerS is not detected.
The FCH domain in FerS is followed by three coiled-coil regions required for cis-oligomerization and intramolecular regulation of tyrosine kinase activity (reviewed in Greer,2002). The coiled-coil domain is continuous with a central SH2 domain with strong binding affinity for several protein substrates, including cortactin. The tyrosine kinase domain is located at the carboxyl terminus. Four structural features of FerS and FerT should be stressed. First, FerT lacks the FCH and coiled-coil domains but retains, like FerS, the SH2 and tyrosine kinase domains. Second, SH2 and tyrosine kinase domains enable FerT and FerS kinase activity and protein substrate binding. Third, FerT induces the phosphorylation of several nuclear proteins in COS1 cells, indicating that a lack of the FCH and coiled-coil regions does not repress the activity of the catalytic domain (Orlovsky et al.,2000). Yet, there is no direct demonstration that cortactin in Sertoli cells and spermatids is a substrate for FerT phosphorylation. Fyn tyrosine kinase, like FerT, is another member of the Src family of tyrosine kinases seen in Sertoli cells (Maekawa et al.,2002) and spermatids (Kierszenbaum,2006). Fyn kinase could also be a candidate for cortactin phosphorylation, a possibility that points to overlapping, downstream, or redundant roles of FerT and Fyn during spermiogenesis. Fourth, Drosophila and Caenorhabditis elegans express Fer isoforms involved in the regulation of F-actin dynamics (Putzke et al.,2005; Murray et al.,2006), suggesting that the molecular pathways involving Fer signaling are likely to be evolutionarily conserved.
The presence of Fer T in the medial/trans-Golgi network and adjacent to acrosome membranes suggests a possible involvement of this tyrosine kinase in acrosome biogenesis preceding FerT relocation to the acroplaxome. Figure 5 is a summary of the visualized FerT Golgi-acrosome-acroplaxome pathway and proposed role of FerT-cortactin phosphorylation in the acroplaxome. Although FerT kinase may phosphorylate a protein substrate at the acrosome membrane site, a non-enzymatic role of FerT cannot be disregarded. In fact, the SH2 domain of Fer-related kinase-1, an ortholog of mammalian FerT in C. elegans, can interact with a variety of cytoskeletal and cell signaling proteins in a kinase-independent manner during shaping of the spheroid embryo into the typical vermiform shape (Putzke et al.,2005). Unlike mammals, which produce full-length FerS and truncated FerT, C. elegans has only a FerT-equivalent that performs the function of FerS and FerT forms in mammals. A comparable situation is seen in rat spermatids. Whether truncated FerT isoforms expressed in spermatids can substitute for the absence of full-length FerS in the acrosome-acroplaxome complex needs to be determined. Possible approaches are the construction of genetic models and monitoring the expression of FerT truncated transcripts in animals with defective spermatid head shaping.
Phosphorylated cortactin in the acroplaxome and the F-actin-contaning Sertoli cell hoops suggests a role in spermatid head shaping. We have postulated that subtle modifications of the F-actin cytoskeletal scaffold may permit the acroplaxome to gradually adjust to Sertoli cell–derived biomechanical stress curbing the associated acrosome and nucleus from blunt deformation (reviewed in Kierszenbaum et al.,2007). Supporting a role of cortactin in spermatid head shaping is a mutation in the gene Cttn (for cortactin) in the mouse Dspd (for dominant spermiogenesis defect, Kai et al.,2004). Dspd mutants display elongated spermatids with abnormal head shape and premature detachment of round spermatids from Sertoli cell niches. As discussed above, FerT tyrosine kinase and Fyn tyrosine kinase coexist in the acroplaxome and may account for cortactin phosphorylation in the acroplaxome. The proto-oncogene c-abl is another tyrosine kinase of the Src family highly expressed in spermatids (Ponzetto and Wolgemuth,1985). In situ hybridization has shown that c-abl transcripts are most abundant in elongating spermatids (Meijer et al.,1987). We have not detected c-abl immunoreactivity in the acroplaxome (data not shown), thus suggesting that c-abl may not have a prevalent role in the phosphorylation of cortactin in the acroplaxome.
Unrelated to the possible role of FerT or any other tyrosine kinase in cortactin phosphorylation in spermatids is the question whether actin filaments are the first to appear in the acroplaxome when the initial Golgi-derived proacrosomal vesicles reach the acroplaxome. A striking number of actin-related and actin-like proteins are expressed in spermatids. For example, actin-related proteins (Arp1, Fouquet et al.,2000; Arp-T1/Arp-T2, Heid et al.,2002); testis actin-like proteins (mouse Tact1/Tact2, Hisano et al.,2003; mouse T-ACTIN 1 and T-ACTIN 2, Tanaka et al.,2003; human actin-like 7A [ACTL7B] and actin-like 7A [ACTL7A], Chadwick et al.,1999) and actin-capping proteins (rat, Hurst et al.,1998) have been reported in spermatids. ACTL7A in human and the equivalent Tact2 and T-ACTIN2 proteins in mouse are testis-specific actin-related proteins, which may play a role in actin dynamics by interacting with specific binding protein partners (Coutts et al.,2003).
In summary, our findings in spermatids of several FerT transcripts isoforms and the absence of FerS expression, lend support to a functional role of FerT in the acrosome-acroplaxome complex and indirectly in spermatid head shaping. However, it is likely that a deficiency in the function of FerT may be compensated for by an additional tyrosine kinase. In fact, it has been reported that mice devoid of Fer protein-tyrosine kinase activity are viable and fertile but display reduced cortactin phosphorylation (Craig et al.,2001). Our report of the distribution sites of FerT and phosphorylated cortactin represent an initial step in understanding the emerging molecular intricacy of the acrosome-acroplaxome-manchette complex. Further studies should determine the functional significance of specific subsets of actin-related proteins and the role of additional protein kinases, including Fyn tyrosine kinase, during spermatid head shaping.
Total RNA was prepared from rat (Sprague-Dawley, adult) and mouse (wild-type 129X1/SvJ, adult) tissues and human testis using TRI Reagent (MRC, Cincinnati, OH). Single-stranded cDNAs were synthesized using a Superscript III kit (Invitrogen, Carlsbad CA). The expression of the Fer gene was examined in rat testis and using pachytene spermatocyte and round spermatid cDNA libraries prepared and described previously (Rivkin et al.,1997). RT-PCR was performed on a GenAmp 2700 system (Applied Biosystems, Foster City, CA) using TaKaRa Taq DNA Polymerase (PanVera, Madison, WI). Standard conditions included incubation of RT-PCR reactions at 94°C/2.5 min followed by 40 cycles of 94°C/10 sec, 58-60°C/10 sec, 72°C/30 sec. The following primers were used:
Numbers in Fer primers indicate the corresponding exon; F and R, identify forward and reverse primers, respectively. Final concentrations of Fer and β-actin primers in RT-PCR reactions were 160 nM and 25 nM, respectively. Sequencing of the RT-PCR products was done by GENEWIZ (South Plainfield, NJ). Fer transcripts in rat pachytene spermatocyte and round spermatid cDNA expression libraries were detected by repeated RT-PCR cycles in which 2% (1 μl) of the first RT-PCR was used as a template for the second RT-PCR cycle with the same FerS primers.
Production, Affinity Purification, and Characterization of FerS and FerT Antisera by Immunoblotting Analysis
Immunogenic peptides from the amino end of FerS and FerT were selected using a protein analysis program (Protean, DNASTAR, Madison, WI). Peptides were synthesized and peptide antibodies raised in rabbits by Sigma-Genosys (Woodlands, TX). Antibodies were affinity purified using the corresponding peptide antigens. The specificity of the antisera was confirmed by immunoblot analysis using pre-immune sera and affinity-purified sera pre-incubated with the specific peptide LEPESDPQFSKKC (see below) and a non-specific peptide LRSRIEHQGDKLELA in a peptide competition assay. The sequence of the peptides and their locations are as follows: Immunogenic peptide FerT, the mouse exon 4T (14LEPESDPQFSKKC26). Immunogenic peptide FerS, the mouse exon 6S (162KETEKAKERYDKATMC176).
FerS and FerT proteins in rat, mouse, and human tissues were analyzed by SDS-PAGE followed by immunoblotting. Proteins were extracted from tissues using 8M urea-containing buffer as described previously (Kierszenbaum et al.,2002), resolved by 12% SDS-PAGE, and blotted using semi-dry Trans-Blot SD (BioRad, Hercules, CA). Immunoblotting was performed using affinity-purified antibodies as described (Kierszenbaum et al.,2003a). Human testis samples were from surgical specimens. Their use was approved by the Institutional Human Subjects Committee (assurance number M-1111-XM-4XM; protocol number H-0021).
Indirect Immunofluorescence and Immunogold Electron Microscopy
Spermatogenic cells from adult rats (Sprague-Dawley) were collected from seminiferous tubular fragments (identified with a dissecting stereomicroscope as corresponding to stage I–XIV of rat spermatogenesis according to their transillumination pattern). Spermatogenic cells were gently extruded from the open ends of seminiferous tubular fragments onto a drop of phosphate-buffered saline (PBS) as previously described (Kierszenbaum et al.,2003a). Dispersed spermatogenic cells were fixed for 15 min by adding a drop of 3.7% paraformaldehyde (electron microscopy grade) in 0.1 M sucrose in phosphate buffer, pH 7.4, on microscope slides coated with Vectabond (Vector Laboratories, Burlingame, CA). After 15-min fixation, a coverglass was placed on top of the preparation. The glass coverslip was removed and the preparations used for immunocytochemistry. This fixation procedure yields spermatid nuclei with associated Golgi, acrosome, acroplaxome, manchette, and chromatoid body structures (see Kierszenbaum et al.,2003a for the characterization of these structures using the same cell fixation protocol).
Cells were immunoreacted with affinity-purified FerS (data not shown) and FerT sera (produced and characterized as described above), anti-phospho (Tyr421)-cortactin (polyclonal antibody, catalog number KAP-TH001; Stressgen, Victoria, BC, Canada) and anti-cortactin (monoclonal antibody, catalog number 05-180; Upstate Biotechnology, Lake Placid, NY). The specificity of phosphorylated and unphosphorylated cortactin antibodies was determined by the manufacturer and verified in our laboratory by immunoblotting (data not shown). Second antibodies were Alexa Fluor 488-conjugated goat anti-rabbit, anti-mouse, and anti-goat IgG (working dilution 1:200; Invitrogen). Specimens were mounted with Vectashield (Vector Laboratories) without or with propidium iodide (to detect nucleic acids by a red emission color). Specimens were observed using a fluorescence microscope equipped with episcopic illumination (Zeiss Universal microscope). Images were recorded using a Magnafire digital CCD camera (Optronics, Goleta, CA).
For immunogold electron microscopy, testes from adult rats were fixed in a mixture of 1.5% glutaraldehyde and 3.4% paraformaldehyde (electron microscope grade) in 0.1 M phosphate buffer, pH 7.2, embedded in Lowicryl K4M (Polysciences, Warrington, PA), and processed for immunogold microscopy as described previously (Rivkin et al.,1997). Anti-FerT polyclonal affinity purified sera (see above) was used at working dilutions of 1:100 in phosphate-buffered saline containing 0.1% Tween 20, 1% bovine serum albumin, and 1% goat serum. Bound antibody was detected by incubating the sample overnight at 4°C with goat anti-rabbit IgG IgG conjugated with 10-nm gold particles (Amersham Biosciences, Piscataway, NJ), by using a 1:25 working dilution. Sections were stained for 5 min with 5% uranyl acetate in deionized water or methanol, and specimens were examined using a JEM-100CX transmission electron microscope operated at an accelerating voltage of 60 kV.
We appreciate the outstanding technical assistance of Cindy Zhang. We are grateful to Dr. Edward W. Gresik for his critical reading of the manuscript. Partial funding for this work was provided by the National Institutes of Health (HD36477 to L.L.T; HD37282 to A.L.K.).