| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Testis |
Departments of Molecular Genetics and Microbiology,3
Pharmacological Sciences,4
Obstetrics and Gynecology,5 State University of New York at Stony Brook, Stony Brook, New York 11794
Department of Molecular Biology and Biochemistry,6 Osaka University Graduate School of Medicine/Faculty of Medicine, Suite 565-0871, Osaka, Japan
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
fertilization, Sertoli cells, sperm, spermatid, spermatogenesis
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
Typical epithelial cells form junctional complexes between their apical lateral surfaces that establish cell polarity, ensure epithelial integrity, and limit paracellular diffusion through the combination of three types of junctions: tight junctions (zonula occludens), adherens junctions (zonula adherens), and desmosomes (macula adherens). Cadherins, a large family of calcium-dependent, homophilic adhesion molecules, have traditionally been regarded as the major adhesive unit at adherens junctions and link the junctional complex to underlying actin belts via catenins [1, 2].
More recently, nectins have been recognized as a novel, calcium-independent adhesion system at cadherin-based adherens junctions [3–5]. Nectins were originally isolated based on their homology to CD155, the human poliovirus receptor [6, 7]. To date, this emerging family of cell adhesion molecules consists of nectin-1, -2, -3, and -4 as well as CD155 and Tage4 [4–11]. These molecules are glycosylated, type Ia, single-pass transmembrane proteins belonging to the immunoglobulin superfamily with V-C2-C2 ectodomains and relatively short cytoplasmic domains. All nectins form cis-homodimers that interact in a trans-homophilic fashion [3–5, 12, 13]. In addition, nectin-3 cis-homodimers can trans-interact strongly with nectin-1 or nectin-2 cis-homodimers [4], and nectin-4 can interact in trans with nectin-1 [5]. Moreover, we have recently shown that the ectodomains of CD155 and nectin-3 express strong affinity to each other (see note added in proof). Nectin and E-cadherin interact through proteins associated with their cytoplasmic domains and are thought to act cooperatively in the formation of adherens junctions [14]. This novel nectin adhesion unit is linked to the actin cytoskeleton by interaction of its cytoplasmic domain with the F actin-binding protein afadin [3–5].
In mammalian testis, Sertoli cells create a unique seminiferous epithelium that cultivates differentiating germ cells during the process of spermatogenesis. Formation and maturation of spermatids depends on an intricate interplay between the Sertoli cell and the developing gamete. Sertoli cells maintain two types of specialized cell-cell junctions through structures termed ectoplasmic specializations. They form between Sertoli cells near the base of the epithelium (basal ectoplasmic specializations) and between Sertoli cells and heads of elongated spermatids (apical ectoplasmic specializations) [15–17]. Ectoplasmic specializations are characterized by hexagonal arrays of actin filaments intimately linked to both the cell membrane and endoplasmic reticulum (ER) cisternae of the Sertoli cell [18]. The apical ectoplasmic specializations is not only responsible for linking the spermatid to the Sertoli cell but also for the translocation of spermatids through the seminiferous epithelium and for the timely release of spermatozoa into the lumen on completion of spermiogenesis. The basal Sertoli-Sertoli junctional complex establishes a blood-testis barrier (BTB), which separates the basal, spermatogonia-containing compartment of the seminiferous epithelium from the immune-privileged, adluminal compartment [19]. The BTB functions primarily to limit paracellular diffusion and to shield haploid germ cells from autoimmune recognition. These barrier functions are thought to be mediated by occludins, claudins, and junctional adhesion molecules present at tight junctions [19–25]. It has been suggested that the formation and maintenance of tight junctions of both classical epithelia and endothelia of the blood-brain barrier are dependent on cadherin-based zonulae adherens [26–28]. In addition, Fukuhara et al. [29, 30] have now shown that in cultured epithelial cells, nectin-mediated adhesion is also required for the formation of tight junctions.
We have very recently described the presence of a novel, heterotypic adhesion system at Sertoli-spermatid junctions, which is maintained by nectin-2 expressed on Sertoli cells and nectin-3 expressed on elongated spermatid heads [31]. In the course of these studies, we have made use of a nectin-2-deficient mouse line (nectin-2LacZ/LacZ). We observed that the loss of nectin-2 leads to severely deformed spermatozoa, which likely is the result of improperly formed Sertoli-spermatid junctions. Moreover, a direct colocalization between nectin-2/nectin-3 and actin bundles was observed that was lost in nectin-2LacZ/LacZ testis [31].
Here, we describe in detail the generation and characterization of nectin-2LacZ/LacZ mice and analyze the resulting male-specific infertility phenotype. We have confirmed that nectin-2 and nectin-3 are strongly expressed at Sertoli-spermatid junctions and that nectin-3 localization at this junction is disturbed in the absence of nectin-2, which presumably leads to the aberrant head and midpiece morphology of nectin-2LacZ/LacZ spermatozoa. Furthermore, we report here that espin, an actin-bundling protein specifically expressed at Sertoli cell ectoplasmic specializations [32, 33], is virtually absent at Sertoli-spermatid junctions in nectin-2LacZ/LacZ testis. Through expression analysis of a LacZ knockin gene within the defunct mouse nectin-2 gene in nectin-2LacZ/LacZ mice, we provide additional evidence that nectin-2 present at Sertoli-spermatid junctions is contributed by the Sertoli cell.
While this work was in progress, Bouchard et al. [34] described the generation of a nectin-2 knockout mouse using a strategy different from the one described here. The similarity and differences between the two systems will be discussed.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
A gene targeting vector, pMPH-ko, was constructed in which a selectable pgk-neo cassette is flanked by two genomic fragments of the murine nectin-2 gene (Fig. 1). In detail, a HindIII/EcoRI fragment containing the pgk-neo cassette of pKJ1 [35] was subcloned into the HindIII/EcoRI sites of pBS-SK(+) (Stratagene, La Jolla, CA) to generate pBS-pgkneo. A polymerase chain reaction (PCR) fragment (oligonucleotides LacZ(+)/pUCRI(-)) containing the open reading frame (ORF) of Escherichia coli ß-galactosidase (LacZ) and a SV40 poly(A) signal was amplified from vector pSDK.LacZpA128 (J. Rossant, University of Toronto, Toronto, Canada) and inserted into XbaI/XmaI sites of pBS-pgkneo. The LacZ-pgkneo cassette was then liberated using XhoI/XbaI and inserted into the XhoI/XbaI linearized vector pBltk4 [36] containing the thymidine kinase gene of herpes simplex virus-1 (HSV-1) under its endogenous viral gene promoter. This allowed for negative selection against random insertion events of the final targeting vector. Into the resulting vector, pBS-ntz, two nectin-2 gene fragments were inserted. These fragments were generated by PCR from 129Sv/J genomic DNA using the LA-PCR kit (Takara, Shiga, Japan) in combination with oligonucleotide pairs Ex3(+)/Ex4(-) and Ex1(+)/Ex2(-), respectively. Fragment 1 (2.2 kilobases [kb]), spanning the C-terminal portion of exon 3, intron 3, and the N-terminal portion of exon 4, was inserted into the SalI/XhoI sites of pBS-ntz, resulting in vector pBS-ntz-2k. To generate the final targeting vector, pMPH-ko, fragment 2 (10 kb), spanning exon 1, intron 1, and the N-terminal part of exon 2, was inserted upstream and in frame with the LacZ ORF, into the NotI/AscI sites of pBS-ntz-2k. Thus, on successful homologous recombination, the mutated nectin-2 locus will express a chimeric reporter protein of the first 61 codons of nectin-2 fused to LacZ. In addition, the targeted locus has lost most of exons 2 and 3 (Fig. 1). The procedure, leading to the knockout of a functional nectin-2 gene, resulted in the knockin of the LacZ gene. We will refer to genetically altered, homozygous animals as nectin-2LacZ/LacZ mice.
|
The targeting vector was electroporated into embryonic stem (ES) cells derived from male 129Sv/J blastomeres (Genome Systems, Inc., St. Louis, MO). The ES clones were selected through a positive/negative selection procedure with G418 and gancyclovir. Targeted ES clones were identified by genomic PCR with oligonucleotides pgkpA(+)/Ex5(-), yielding a 3.5-kb reaction product, whereas random insertions do not produce a PCR product (Fig. 1B). Positive ES cell clones were confirmed by Southern blot analysis of NdeI (New England Biolabs)-digested genomic DNA (Fig. 1C). For Southern hybridization, a 5' external 32P-labeled probe was generated by random priming (RadPrime; Gibco BRL, Gaithersburg, MD) of a PCR fragment (oligonucleotides Ex5(+)/Ex6(-); 538 base pairs [bp] in length) of nectin-2 cDNA [8]. Eight ES clones that were positive by both screening procedures were injected into C57BL/6 blastocysts to generate chimera. One clone, EW63, resulted in highly chimeric male offspring, one of which transmitted the targeted nectin-2 locus through the germline when mated to DBA2xC57BL/6 hybrid females. Male and female offspring carrying the targeted allele were intermated to produce homozygous animals. After the correct targeting of the founder line was established by the procedure outlined above, routine genotyping of nectin-2 mice was done by a simpler genomic PCR analysis using a mix of three oligonucleotides, MPH126(+)/LacZ135(-)/Ex2-3'(-). Cycling conditions were 30 sec at 96°C, 45 sec at 60°C, and 45 sec at 72°C for 28 cycles. Whereas the 210-bp PCR reaction product of MPH126(+)/LacZ135 identifies the targeted allele, the 350-bp PCR product MPH126(+)/Ex2-3'(-) is indicative of the wild-type (wt) allele (Fig. 1D).
Oligonucleotide List
Immunofluorescence Analysis
Nectin-2wt/wt and nectin-2LacZ/LacZ mouse testes were fixed for 4 h in 4% paraformaldehyde before freezing in OCT cryoprotectant (Tissue-Tek, Sakura Finetek, Torrance, CA). Cryosections (thickness, 10 µm) were collected on poly-L-lysine-coated slides, air-dried, and postfixed in cold 1:1 methanol:acetone for 90 min at -20°C. Sections were blocked with PBS containing 5% normal horse serum and 2% normal goat serum (PBS/HG), followed by overnight incubation at 4°C with primary antibodies in PBS/HG. The following primary antibodies were used: anti-nectin-2 rat monoclonal antibodies (mAbs) 17B10 and 6B3 [12], anti-nectin-3 rat mAb 103-A1 (1:5 diluted hybridoma supernatants for each) [4], anti-espin mouse mAb 31 (2.5 µg/ml; BD Transduction Laboratories, San Jose, CA), and a rabbit polyclonal antibody against ß-galactosidase (1:500; Molecular Probes, Eugene, OR). After 1 h of washing with PBS, the sections were incubated with combinations of either Cy3-conjugated anti-rat (1:1000; Jackson Immuno Research, West Grove, PA) and Alexa488-conjugated anti-rabbit antibodies (1:500; Molecular Probes) or Cy2-conjugated anti-rat (1:200; Jackson Immuno Research) and Cy3-conjugated anti-mouse antibodies (1:1000; Jackson Immuno Research). Nuclei were counterstained by addition of 100 ng/ml of Hoechst 33258 (Molecular Probes). The sections were washed extensively with PBS and mounted with Immu-Mount (Shandon, Pittsburgh, PA). Images were acquired on a Zeiss Axioplan II fluorescence microscope equipped with a model SP401 camera (Diagnostic Instruments, Inc., Sterling Heights, MI) and processed with Adobe Photoshop software.
Scanning Electron Microscopy
Epididymal spermatozoa were fixed for 1 h in 2.5% glutaraldehyde and 2% paraformaldehyde in 100 mM cacodylate buffer (pH 7.4) containing 2.5 mM CaCl2. A drop of sperm suspension was then adhered to poly-L-lysine-coated coverslips, and fixed spermatozoa were rinsed three times in cacodylate buffer. The coverslips were then postfixed in a freshly prepared solution of 1% OsO4 in cacodylate buffer for 30 min at room temperature. After three more rinses in cacodylate buffer, the samples were dehydrated through a series of increasing concentrations of acetone in water and twice in 100% acetone. Finally, the samples were subjected to critical point drying. Images were acquired on a Jeol 5300 scanning electron microscope.
In Vivo Fertilization and Assessment of Oviductal Spermatozoa
CD1 female mice (age, 6 wk) were injected with 6 IU of eCG, followed by 6 IU of hCG 50 h thereafter to induce ovulation. Two females were caged with one male for one night following the hCG administration. Copulation plugs were observed 16 h post-hCG, at which time plugged females were killed and the sperm-exposed oocytes were collected from the ampullae in M2 medium (Specialty Media, Phillipsburg, NJ). Egg clusters were briefly treated with hyaluronidase (Sigma, St. Louis, MO) to disperse oocytes. The oocytes were then transferred to M16 medium (Specialty Media) and cultured for 6 h at 37°C under 5% CO2. At this point, the presence of a male pronuclei was determined by microscopic observation under Nomarski optics. After 24 h in culture, embryos were transferred to KSOM medium (Specialty Media) and allowed to develop for an additional 48 h to assess the formation of blastomeres. To assess the ability of spermatozoa to enter the oviducts following overnight mating, the spermatozoa were flushed out of the oviducts, collected in 1 ml of M2 medium, and briefly subjected to hyaluronidase treatment. Spermatozoa were then concentrated by centrifugation at 500 g for 5 min, resuspended in 20 µl of M2 medium, fixed with 2% paraformaldehyde, and counted in a hemocytometer. In the present study, P values were determined by the two-tailed, unpaired t-test.
Zona-Free Hamster Oocyte Sperm Penetration Assay
Epididymal spermatozoa of nectin-2wt/LacZ and nectin-2LacZ/LacZ males (age, 24 wk) were recovered by slashing the cauda epididymis several times with a hypodermic needle and allowing the sperm to swim out for 30 min at 37°C under 5% CO2 in a dish containing 2 ml of preequilibrated human tubal fluid (Specialty Media). Sperm penetration assay (SPA) was performed as described previously [37] with modifications [38]. Briefly, mouse sperm were recovered as described above, except that the medium was Biggers Whitten Whittingham (BWW) [39] containing 30 mg/ml of BSA (Sigma-Aldrich) and preequilibrated at 37°C under 5% CO2. Spermatozoa at concentrations between 25 x 106 and 30 x 106 per milliliter were capacitated by incubation at 37°C under 5% CO2 for 2 h. Oocytes of superovulated golden hamster females were collected in BWW containing 5 mg/ml of human serum albumin (HSA) and released from cumulus masses by treatment with 1 mg/ml of hyaluronidase. Zonae pellucidae were removed by incubation in 1 mg/ml of trypsin (Sigma). Oocytes were inseminated with spermatozoa at a final concentration of 105 per milliliter in 1 ml BWW with 5 mg/ml of HSA. After 45 and 90 min of incubation at 37°C under 5% CO2, oocytes were removed and stained with acridine orange to visualize penetrated, decondensing sperm nuclei. The number of bound spermatozoa was determined under phase-contrast microscopy by counting sperm tails protruding from the oocyte at a median focal plane. Penetration events were assessed under epifluorescence. All samples were counted by the same observer.
In Vitro Sperm Binding Assay to Zona-Intact Mouse Oocytes
Epididymal spermatozoa of three age-matched nectin-2wt/LacZ controls and three nectin-2LacZ/LacZ males were recovered and capacitated as described above. Oocytes of 12 ovulated CD1 females, recovered 15 h post-hCG injection, were pooled and subdivided into fertilization dishes containing 800 µl of BWW. Then, 5 x 105 spermatozoa were added to oocyte dishes in a final volume of 1 ml of BWW. After 15 min at 37°C under 5% CO2, sperm-exposed oocytes and bound sperm were washed through subsequent drops of M2 medium to remove loosely adhered sperm. Bound sperm were fixed with 2% paraformaldehyde and stained with 100 ng/ml of Hoechst 33258 to visualize sperm nuclei. The number of spermatozoa bound per sperm-exposed oocyte was determined by counting spermatozoan nuclei associated with each oocyte under a fluorescence microscope. Bound sperm nuclei were counted by slowly focusing through the whole oocyte. No attempts were made to assess the acrosome status of the spermatozoa.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
We have inactivated the nectin-2 gene locus by homologous recombination in mouse ES cells of the 129Sv/J genetic background. A targeting vector was constructed that on successful recombination replaces most of exons 2 and 3 by an in-frame fusion with the LacZ ORF and a selectable pgk-neo cassette [35] (Fig. 1). Thus, a fusion protein between the N-terminal 61 amino acids of nectin-2 and LacZ is expressed under control of the endogenous nectin-2 gene promoter. Of 355 ES cell clones screened, 8 were found to contain the targeted nectin-2 gene locus. Two independent screening procedures were used to confirm successful targeting. First, a primary PCR screen with oligonucleotide pair pgkpA(+)/Ex5(-) was employed that identified positive clones by the presence of 3.5-kb reaction product. Because the annealing site for Ex5(-) is located outside the replacement cassette, only homologous recombination events bring the two annealing sites into proximity to result in an amplicon of the predicted size. Consequently, random insertions should not produce a PCR product (Fig. 1B). Clones identified by this method were then subjected to Southern blot analysis of NdeI-digested genomic DNA. An external probe was chosen that comprises exons 5 and 6. This probe was amplified from nectin-2 cDNA and subsequently 32P-labeled by random priming. Whereas a wt allele produces a band of approximately 13 kb, the targeted allele is characterized by a 16-kb band (Fig. 1C). All clones identified in the primary PCR screen yielded the characteristic doublet of bands (Fig. 1C).
Of the eight targeted ES clones that were then injected into C57Bl/6 blastocysts and transferred to foster mothers, one produced chimeric offspring that transmitted the targeted allele through the germline. Their heterozygous offspring were then intermated to produce homozygous nectin-2-null mice (nectin-2LacZ/LacZ).
Nectin-2-Null Males Are Infertile
We found nectin-2LacZ/LacZ homozygous mice to be viable and without any gross morphological or behavioral phenotype. Homozygous offspring were present in a mendelian ratio, suggesting that no embryonal lethality is associated with the loss of nectin-2. On further analysis, we determined that homozygous males, but not females, were infertile. Ten nectin-2LacZ/LacZ males (age, 12–20 wk) were each housed with two CD1 females of proven fertility (age, 12 wk) for 3 mo. Although homozygous males showed normal mating behavior and each male produced copulation plugs on numerous occasions, no offspring were ever produced from these matings (data not shown). This confirms the earlier results of Bouchard et al. [34]. We next set out to systematically dissect the male infertility phenotype.
Nectin-2-Null Males Produce Severely Malformed Spermatozoa
No anatomical differences were detected in the reproductive system of nectin-2LacZ/LacZ males when compared to heterozygous and wt littermates. Testis size (data not shown) and sperm titers were comparable between the three genetic groups. Specifically, 27.4 ± 4.3 x 106, 29.1 ± 3.1 x 106, and 23.3 ± 4.9 x 106 spermatozoa (all values ± SD) were recovered from the cauda epididymis of nectin-2wt/wt, nectin-2wt/LacZ, and nectin-2LacZ/LacZ males, respectively; each value corresponds to the average of three age-matched males. However, the spermatozoa of nectin-2LacZ/LacZ males showed severe malformation of the head and the midpiece (Fig. 2). Although all spermatozoa were affected, the degree of malformation varied. No single common type of abnormality was observed; rather, a random disorganization of the spermatozoan head and midpiece was seen. The more prevalent shapes that were observed included those depicted in Figure 2, I and J, with the head being curled back onto the midpiece, or those depicted in Figure 2, F and H, with a forward-oriented but irregularly shaped head. We also frequently noticed a rough appearance of the midpiece (Fig. 2L) or a midpiece that appeared to be much thinner than normal (Fig. 2D). This result suggests a disorganization of the mitochondrial sheath (Fig. 2L) or the total lack thereof (Fig. 2D), and it confirms the findings of Bouchard et al. [34] with their nectin-2 knockout line derived by a different strategy. However, in all spermatozoa, the classic falciform shape of the normal spermatozoan head (Fig. 2B) was all but lost. In contrast to the phenotype described by Bouchard et al. [34], we did not observe an undulation of, or any other changes affecting, the principal piece of the tail.
|
Despite their severe defects, nectin-2LacZ/LacZ spermatozoa were viable and motile. However, on longer incubation (12–24 h), in vitro nectin-2LacZ/LacZ spermatozoan motility and viability declined more quickly than those of the controls (data not shown). It is conceivable that the disorganized mitochondrial sheath impairs the spermatozoan energy metabolism. Alternatively, fewer mitochondria might have been incorporated into nectin-2LacZ/LacZ spermatozoa during spermatogenesis.
Spermatozoa of Nectin-2-Null Males Fail to Fertilize Oocytes In Vivo
Because nectin-2-null spermatozoa were viable and motile despite their gross morphological defects, the question arose at which step during fertilization these spermatozoa fail to function. In a first set of experiments, we recovered potential zygotes from the oviducts of females after overnight mating for subsequent in vitro culture. Five nectin-2LacZ/LacZ males and three nectin-2wt/LacZ males were each housed with two CD1 females (see Material and Methods). Except for one nectin-2wt/LacZ male, all males mated with at least one female, as assessed by the presence of a copulation plug. Only females with clearly formed copulation plugs were used to recover sperm-exposed oocytes (potential zygotes) to assure that mating had, indeed, occurred. All plugged females had ovulated, and between 3 and 40 potential zygotes were recovered per animal (Table 1). Sperm-exposed oocytes were then cultured in vitro and periodically observed microscopically until the blastomere stage. We found that all but one oocyte of females mated to nectin-2wt/LacZ males were fertilized and proceeded to develop normally to the blastomere stage. In contrast, none of 73 sperm-exposed oocytes isolated from five females that had mated to nectin-2LacZ/LacZ males initiated the developmental program. Neither could we observe the formation of male pronuclei in this set of sperm-exposed oocytes, an observation suggesting that fertilization had not occurred (Table 1).
|
Spermatozoa of Nectin-2-Null Males Are Deficient in Migration to the Oviduct
To address the question of whether the observed absence of fertilized eggs in females mated to nectin-2LacZ/LacZ males results solely from defects in gamete interaction, we assessed the ability of spermatozoa to enter the oviducts. For this purpose, oviducts were dissected from females after natural overnight matings with either nectin-2LacZ/LacZ or nectin-2wt/LacZ control males. The contents of the oviducts were flushed out, and spermatozoa were processed as described in Materials and Methods. We found that nectin-2LacZ/LacZ spermatozoa were approximately 4-fold less efficient in reaching the oviducts compared to the heterozygous controls. On average, 880 spermatozoa were isolated from the oviducts of females mated to control males, but only 204 spermatozoa were found in the case of nectin-2LacZ/LacZ males (Fig. 3). This difference was determined to be statistically highly significant (P = 0.0002). Thus, it seems likely that the reduced ability of nectin-2LacZ/LacZ spermatozoa to reach the oviducts contributes to the observed infertility phenotype in nectin-2-deficient males.
|
Nectin-2-Null Spermatozoa Display Impaired Binding to Zona-Intact Mouse Oocytes
Zona-intact mouse oocytes were recovered from ovulated CD1 females 15 h post-hCG administration and incubated with 5 x 105 nectin-2wt/LacZ or nectin-2LacZ/LacZ sperm per milliliter (Fig. 4). We observed a 6-fold reduction in the binding capacity of nectin-2LacZ/LacZ spermatozoa to oocytes. An average of 40.8 nectin-2LacZ/LacZ spermatozoa bound per oocyte, but only 6.7 nectin-2wt/LacZ spermatozoa were found to bind per oocyte (Fig. 4).
|
Nectin-2LacZ/LacZ Sperm Bind to Oolemma Normally But Fail to Penetrate
The SPA using zona-free hamster oocytes is a well-established test that has been used extensively to assess the penetration capacity of a wide variety of mammalian sperm [37]. Removing the zona pellucida sidesteps the requirement for the species-specific sperm-zona binding reaction, and it allows one to evaluate sperm/oolemma binding and fusion. Unexpectedly, nectin-2LacZ/LacZ spermatozoa were able to bind to the oolemma at levels comparable those of nectin-2wt/LacZ controls (Table 2). After 90 min of incubation, an average of 42.1 nectin-2LacZ/LacZ versus 47.9 nectin-2wt/LacZ spermatozoa had bound per oocyte (counted as the number of protruding sperm tails at a median focal plane) (Fig. 5). In contrast, penetration of oocytes by nectin-2LacZ/LacZ spermatozoa, as assessed by the presence of decondensing sperm nuclei in the ooplasm, was 75-fold and 32-fold lower after 45 and 90 min of incubation, respectively (Table 2 and Fig. 5). After 90 min of incubation in the control group, every oocyte was penetrated by an average of 13.1 spermatozoa, but only 2 of 12 oocytes were penetrated when nectin-2LacZ/LacZ sperm was used, with an average of 0.42 penetrations per oocyte (counted as the total number of decondensing sperm nuclei per oocyte). We did observe one oocyte that was penetrated by two nectin-2LacZ/LacZ spermatozoa (Fig. 5b2), which indicates that fusion with nectin-2LacZ/LacZ spermatozoa can occur, albeit at a greatly reduced rate.
|
|
Expression of Nectin-2 in the Male Reproductive System
To confirm the absence of nectin-2 protein in nectin-2LacZ/LacZ testis and to relate nectin function to the observed phenotype, we analyzed the expression of nectin-2 and that of its heterophilic binding partner, nectin-3, by immunofluorescence in wt (Fig. 6, A and B) and in nectin-2LacZ/LacZ testis (Fig. 6, C and D). Both nectin-2 and nectin-3 were strongly expressed along the convex curvature of the heads of elongated spermatids (Fig. 6, A and B). In addition, nectin-2 protein, but not nectin-3, was found in an almost ring-like staining pattern encircling the seminiferous tubule throughout the basal compartment (Fig. 6A, arrowheads). The latter staining appears to be associated with inter-Sertoli junctions, and its location near the base of the tubule suggests the presence of nectin-2 at the BTB.
|
Because in nectin-2LacZ/LacZ mice the nectin-2 gene is interrupted by the in-frame insertion of a LacZ reporter gene, we were interested whether the LacZ protein is expressed and whether analysis of LacZ expression can provide further insight regarding the function of nectin-2 in testis. As expected, no intact nectin-2 protein could be detected in testes of nectin-2LacZ/LacZ males, confirming the successful inactivation of the nectin-2 gene (Fig. 6C). Instead, the targeted nectin-2 locus indeed expressed the LacZ protein as a fusion to the N-terminal 61 amino acids of nectin-2 (Fig. 6, C and D). Because these N-terminal 61 amino acids of nectin-2 only contain a signal sequence and a small portion of the first immunoglobulin-like domain, it is not expected that any residual function of nectin-2 was preserved. The nectin-2 N-terminal signal peptide will likely target the resulting nectin-2-LacZ fusion protein to the cell surface of cells in which the nectin-2 gene promoter is active. Staining for LacZ protein expression thus provided us with an alternative means of assessing the activity of the nectin-2 gene in the absence of secondary posttranslational events of the nectin-2 protein (e.g., protein-protein interactions). We found LacZ staining to outline the Sertoli cells, whereas no LacZ protein expression was associated with spermatid heads (Fig. 6, C, arrowheads, and D inset, asterisk). This indicates to us that LacZ was expressed by Sertoli cells. It can thus be inferred that the nectin-2 gene is active only in Sertoli cells and not in spermatids. This result supports those of our previous ultrastructural studies and germ cell transplantation experiments [31]. Expression of nectin-3, the only known heterophilic binding partner of nectin-2, remained associated with spermatids in nectin-2LacZ/LacZ testes (Fig. 6, D and D inset). However, nectin-3 protein localization was severely disturbed. Rather than evenly decorating the convex aspect of elongated spermatid heads, nectin-3 protein was found to be concentrated mainly in a single, dense speck of immunoreactivity located to the anterior on the spermatid head (Fig. 6D inset, asterisk). The absence of Sertoli cell-expressed nectin-2 appears to disrupt proper targeting of nectin-3 to Sertoli-spermatid junctions, or it may even indicate a defect in the formation of such junctions altogether. The present data support our previous finding of a heterophilic adhesion system maintained by nectin-2 on Sertoli cells and its binding partner, nectin-3, on heads of elongated spermatids [31].
Espin Fails to Localize at Sertoli-Spermatid Junctions in Nectin-2LacZ/LacZ Testis
The data presented above and our earlier results [31] suggest that Sertoli-spermatid junctions form improperly in nectin-2LacZ/LacZ testis. Furthermore, we have shown that actin assembly in Sertoli cells of nectin-2LacZ/LacZ mice is disturbed at their junctions with elongated spermatids [31]. To test whether other components of the junctional complex are affected, we analyzed the expression of the actin-bundling protein espin (ectoplasmic specialization + "in"), a known component of Sertoli cell ectoplasmic specializations [32, 33] (Fig. 7). In wt testis, espin colocalized with nectin-3 at Sertoli-spermatid junctions (Fig. 7C, yellow color). In contrast, the overall expression of espin was lower in nectin-2LacZ/LacZ testis, and remarkably, virtually no espin was found to be associated with Sertoli-spermatid junctions (Fig. 7F). This result suggests that in the absence of nectin-2, ectoplasmic specializations cannot be assembled.
Expression of Nectin-2 in Epididymis
Interestingly, as soon as spermatozoa were released from the seminiferous epithelium to be transported to the epididymis, no nectin-2 or nectin-3 immunoreactivity was associated with spermatozoan heads (Fig. 8, A and a1 and B and b1, respectively). This was expected for nectin-2, because the protein seems to be exclusively expressed by Sertoli cells (see above). In contrast, spermatid expression of nectin-3 appears to be downregulated, possibly to facilitate spermiation by detachment from nectin-2 containing ectoplasmic specialization on Sertoli cells.
|
Furthermore, we detected nectin-2 expression at apical junctions of epithelial cells lining the epididymal duct (Fig. 8, A, arrowhead, and a2, arrowheads). The localization of nectin-2 staining at the apical region of the epithelium is consistent with a role of nectin-2 as a homophilic adhesion molecule at cadherin-based adherens junction, as has been previously described in other epithelia [3, 4].
Unlike Bouchard et al. [34], we did not observe nectin-2 staining on the midpiece of spermatozoa at any stage of spermatogenesis (see, e.g., Figs. 6A and 8A), including mature spermatozoa of the cauda epididymis (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
Most morphological changes of the differentiating spermatid occur in the later stages of spermiogenesis (steps 8–16) as round spermatids elongate their nuclei, condense their chromatin, shed most of their cytoplasm, and extend the sperm tail. During this time, spermatogenesis in nectin-2-deficient males appears to deviate from the normal developmental program [31, 34]. Nectin-2 is only expressed on Sertoli cells and not on spermatids (see above). Consequently, the observed aberrant morphology of spermatozoa in nectin-2-deficient mice likely results from the absence of exterior forces exerted by the Sertoli cell onto the developing spermatid. In wt testes, these forces seem to be transmitted from the Sertoli cell to the spermatid by the nectin-2/nectin-3 adhesion unit. Although Sertoli cell-expressed nectin-2 is linked to actin filaments of the apical ectoplasmic specializations [31], this cortical actin is noncontractile, and no myosin is associated with it [40]. It therefore appears to have a merely structural function and is unlikely to act as a force-generating entity [40]. Although a possible link between microtubule-based transport in Sertoli cells and translocation of the Sertoli-spermatid junctional plaque together with the attached spermatid has been suggested previously [41–46], the precise mechanisms of shaping and translocating the differentiating spermatid within the seminiferous epithelium are still elusive. In this regard, it is noteworthy that we have recently identified an interaction between Tctex-1 and CD155, the closest known human relative of nectin-2 [47]. Interestingly, Tctex-1, a subunit of the dynein motor complex, is highly expressed in testis [48]. The dynein complexes, in turn, have been found in association with apical junctional plaques [45, 46]. In the future, it will be interesting to determine whether a direct link between nectins and dynein exists.
The differentiating germ cells evade autoimmune recognition by translocating to the immune-privileged adluminal compartment, which is separated from the basal compartment by the Sertoli cell-maintained junctional complexes of the BTB. This translocation requires the concerted action of breaking the tight and adherens junctions apically and rejoining them basally of the migrating germ cell, without compromising the BTB. We report here that nectin-2 protein localizes between Sertoli cells near the base of the tubule (Fig. 5). This is strikingly similar to expression patterns of occludin [22, 49] and claudin [23, 50], two molecules implicated in the formation of tight junctions at the BTB. Given the interdependence of adherens and tight junctions, our findings of nectin-2 expression at basal ectoplasmic specializations suggest a possible role of nectin-2 at adherens junctions of the BTB. In support of this theory, we also found expression of nectin-2 at apical junctions of the epididymal epithelium (Fig. 8), a site coinciding with the localization of the blood-epididymis barrier [51]. How the loss of nectin-2 affects these junctional complexes and, thus, the BTB or the blood-epididymis barrier is the subject of further investigation.
While this work was ongoing, Bouchard et al. [34] generated a similar nectin-2 knockout mouse line. In theirs as well as in our experiments, only one founder line was obtained. Therefore, our mouse nectin-2LacZ/LacZ line serves as an important validation of the infertility phenotype described earlier [34]. A significant difference in our line is the presence of an in-frame fusion of a LacZ reporter into the nectin-2 gene that provided us with an additional tool for assessing the activity of the nectin-2 promoter by LacZ detection in nectin-2LacZ/LacZ mouse tissue.
Although male infertility in nectin-2-deficient mice results from aberrant progression of spermiogenesis leading to aberrantly shaped sperm, at the functional level this infertility phenotype is only manifested later, at the time of sperm-oocyte interaction and penetration (Figs. 4 and 5 and Table 2). Nectin-2 itself is absent on mature mouse spermatozoa; therefore, it cannot play a direct role in gamete interaction. Thus, the question arises of how the distortion of molecular structure at the sperm surface of nectin-2LacZ/LacZ spermatozoa causes the failure in fertilization. Gamete interactions leading to successful fertilization are dependent on a number of discrete steps involving specific receptors and their ligands, both at the level of sperm interaction with the zona pellucida and, distinctly, with the oolemma. Acrosome-intact mouse sperm bind to the zona pellucida through specific receptor-ligand interactions and subsequently undergo an acrosome reaction. This step appears to have been impaired with sperm of nectin-2-deficient mice. Sperm binding to the oolemma is also mediated by a family of gamete receptors and their ligands. In mouse, both acrosome-intact and acrosome-reacted sperm are capable of binding to the oolemma of zona-free hamster eggs. Interestingly, this binding to the oolemma was not impaired in the nectin-2-deficient mice, although subsequent sperm penetration of the oocytes was abnormal. This observation suggests that the ability of knockout sperm to undergo an acrosome reaction may be altered.
Are mutations in the nectin-2 gene a reason for male infertility in men presenting with abnormal spermatozoan head/midpiece morphology? Numerous genetic causes are known for male infertility, deciphered mainly from experiments in animal models [52]. For example, targeted gene disruption of HsP70-2 in mice leads to failed meiosis, germ cell apoptosis, and male infertility [53], a phenotype that may relate to infertility in men [54 and references therein]. Cause and effect in this genetic trait, of course, is totally different from the nectin-2 knockout described here. As yet, no link is known between a defective nectin-2 gene and human male infertility, but a correlation between these phenotypes may be discovered in the future. If so, fertilization of eggs could possibly be achieved by intracytoplasmic sperm injection with nectin-2-null spermatozoa.
|
NOTE ADDED IN PROOF |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: Eckard Wimmer, Department of Molecular Genetics and Microbiology, State University of New York at Stony Brook, Stony Brook, NY 11794. FAX: 631 632 8891; ewimmer{at}ms.cc.sunysb.edu
Received: 14 December 2002.
First decision: 14 January 2003.
Accepted: 28 May 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONNOTE ADDED IN PROOFREFERENCES |
|---|
