Nuclear envelope remodelling during rat spermiogenesis: distribution and expression pattern of LAP2/thymopoietins

Spermatogenesis is a complex differentiation process that can be divided into three phases: (i) mitotic proliferation of the stem cells (spermatogonia), (ii) meiotic divisions of spermatocytes, and (iii) spermiogenesis (see Erickson, 1990). During spermiogenesis the postmeiotic cells (spermatids) undergo a profound remodelling that affects the whole cell and gives rise to a mature sperm (see Clermont et al., 1993). In the nuclear compartment these changes are characterized by the removal of histones and, consequently, the nucleosomal organization of chromatin. At the end of the process chromatin is largely composed of DNA and protamines, and is highly condensed and transcriptionally inactive (e.g. Clermont et al., 1993; Eddy and O’Brien, 1994; Hecht, 1995).

Changes in the nuclear interior are accompanied by changes in the nuclear envelope, which, however, have been characterized to a much lesser extent. It has been shown that, as a reflection of the progressive decrease in nuclear activity, the number of pore complexes per spermatid nucleus diminishes dramatically (for review see Clermont et al., 1993). When compared with somatic cells and meiotic stages of spermatogenesis, the nuclear envelope of spermatids also shows remarkable differences in the expression of structural protein components (for an overview see Alsheimer and Benavente, 1996). Of the known mammalian somatic lamins, only lamin B1 has been demonstrated to be expressed during rat spermatogenesis. Somatic lamins A, C and B2 are not detectable in any phase (Vester et al., 1993). Instead, the expression of two short germ line-specific lamin isoforms, the lamins C2 and B3, has been demonstrated, but only during meiotic stages (Smith and Benavente, 1992a; Furukawa and Hotta, 1993; Furukawa et al., 1994; Alsheimer and Benavente, 1996). In early round spermatids lamin B1 is observed along the whole nuclear periphery (Vester et al., 1993). However, in the mature sperms it is undetectable by immunochemical (Kaufmann, 1989) and immunocytochemical methods (Vester et al., 1993). In mammals, the expression of additional lamins during spermiogenesis, such as the spermiogenesis-specific lamin LIV of amphibians (Benavente and Krohne, 1985), has not yet been demonstrated (see Vester et al., 1993). Since chromatin is intimately associated with the nuclear envelope of spermatids, a detailed knowledge of the composition and organization of the nuclear periphery of these cells is important for a better understanding of the changes in the topological organization of chromatin. Moreover, this knowledge would also be relevant to our understanding of how the male pronucleus is assembled and organized from a sperm head after fertilization.

Thymopoietins (TPs) are a family of proteins, first described in humans, which are derived by alternative splicing from a single gene. Molecular characterization revealed that both TPβ and γ (but not TPoi) contain one potential transmembrane domain (Harris et al., 1994, 1995). The lamina-associated polypeptide 2 (LAP2) was first characterized in rat somatic cells by biochemical methods as a transmembrane protein of the inner nuclear membrane that binds chromatin and lamin B1 (Foisner and Gerace, 1993; for review see Gerace and Foisner, 1994; Georgatos et al., 1994; Cowin and Burke, 1996). Sequencing of LAP2 (Furukawa et al., 1995) surprisingly revealed that it is the rat homologue of TPβ (see also Harris et al., 1995; Cowin and Burke, 1996). In the present study we have investigated the expression pattern and localization of the TPs during rat spermatogenesis with the aid of the novel monoclonal antibody (mAb) 13d4. For comparison, the distribution and expression of lamin B1 during this differentiation process was also investigated.

The following cell lines were used for immunocytochemistry and protein analysis: RV-SMC (rat vascular smooth muscle cells; Franke et al., 1980), HeLa (human epitheloid carcinoma cells), PtK2 (kidney epithelia cells of the marsupial Protorous tridactylis) and A6 (kidney epithelial cells ofXenopus laevis). Cell lines were cultured according to standard procedures. Rat liver was dissected from adult animals and cut into small pieces. The pieces were immediately shock frozen in methylbutane cooled with liquid nitrogen to − 140°C. Alternatively, small pieces of liver tissue were homogenized and boiled in lysis buffer for analysis by SDS-PAGE. Enriched populations of spermatogenic cells were obtained by elutriation of a cell suspension prepared from Wistar rat testes (Meistrich, 1977; Heyting and Dietrich, 1991). Sperm heads were prepared after sonication of a sperm suspension obtained from rat epididymus (see Longo et al., 1987).

The novel hybridoma cell line 13d4 was isolated after the fusion of mouse myeloma cells with spleen cells of a Balb/c mouse immunized with a karyoskeletal-enriched protein fraction of rat pachytene spermatocytes (for a detailed description of the protocol see Smith and Benavente, 1992b). Immunoglobulins secreted by the cell line 13d4 are of the IgG2a class (Smith, 1992). Monoclonal antibody L3f4 (IgM) against A-and B-type somatic lamins (Vester et al., 1993) has been characterized previously. Appropriate secondary antibodies coupled to FITC, Texas Red or colloidal gold (6 nm) were purchased from Dianova (Hamburg, Germany). Immunolocalizations on frozen sections of rat testes at the light and electron microscopical levels were conducted as previously described (Smith and Benavente, 1992b). Alternatively, immunofluorescence microscopy was performed on smears of testicular cells. For this, spermatogenic cells obtained by elutriation were gently dropped on glass slides. After drying, the cells were fixed in methanol (10 minutes, −20°C) and acetone (1 minute, −20°C). They were then air dried and incubated with the antibodies as described (Smith and Benavente, 1992b).

SDS-PAGE was carried out on 10% polyacrylamide gels according to the method of Thomas and Kornberg (1975). The proteins were transferred to nitrocellulose membranes by using the semi-dry western blotting system described by Matsudaira (1987). The membranes were blocked for 2 hours at room temperature with TBST buffer (10 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% Tween-20) containing 10% milk powder. After washing with TBST, membranes were incubated for 2 hours at room temperature with hybridoma supernatant containing mAb 13d4 or L3f4. Bound antibodies were detected with either alkaline phosphatase (Dianova), or the enhanced chemiluminescence system (ECL, Amersham, Braunschweig, Germany).

A XZAP II cDNA expression library prepared from rat pachytene spermatocytes (Alsheimer and Benavente, 1996) was screened with mAb 13d4. One positive plaque was isolated through three successive rounds of screening. The cDNA insert (rp1.8 clone) was sequenced using the Taq DyeDeoxy Terminator cycle sequencing kit (Applied Biosystems, Weiterstadt, Germany) and the Applied Biosystems automated sequencer (model 373A). The cDNA sequence and the deduced protein sequence were analyzed using the GCG software package (Genetics Computer Group, Madison WI).

cDNA clone rp1.8 was cut out of the original plasmid vector pB1 (obtained by in vivo excision of the insert containing X phage) with BamHI and KpnI, and ligated in frame into vector pQE31 (Qiagen, Hilden, Germany). The plasmids were then transformed into E. coli strain SG13009 (Qiagen) according to standard protocols. Transformed bacteria were selected and amplified by overnight culture at 37°C in 10 ml Luria-Bertani medium (LB) containing 50 μg/ml ampicilin. A new 25 ml culture was inoculated with 100 μl of the overnight culture and incubated for 4 hours at 37°C. Protein synthesis was activated by the addition of isopropyl-β -D-thiogalactopyranoside (IPTG) to a final concentration of 1 mM. After incubation for 4 hours at 37°C, the induced bacteria were harvested and processed for SDS-PAGE and immunoblotting.

To obtain a cDNA clone coding for the first 187 amino acids only, rp1.8 clone was used as a template. The desired sequence was amplified by PCR using bluescript internal primer T3 (ATTAAC-CCTCACTAAAG) and primer L7 (TTCGTCATTGTCACTGTATC), which is complementary to nucleotides 564–583 of rp1.8 clone (corresponding to nucleotides 791-810 of the clone 4d described by Furukawa et al., 1995). The amplifying reaction was carried out according to standard protocols except that 5% DMSO was added to the reaction mixture. The amplified sequence (rp1.8/A 1-187) was cloned into the PCR direct cloning vector pTAg (R&D Systems, Wiesbaden, Germany). The insert was then cut out of the cloning vector with restriction enzymes BamHI and HindIII and ligated into vector pQE31. Correct subcloning of the inserted fragment was checked by sequencing of the construct. Plasmids were transformed into E. coli strain SG13009 and protein synthesis was induced with IPTG as described for rp1.8 clone.

In immunofluorescence analysis of cultured cells and on frozen sections of rat liver, mAb 13d4 selectively labelled the nuclear periphery of interphase cells (Fig. 1A,B). This was confirmed by confocal microscopy of cultured cells (Fig. 1C). Furthermore, at the ultrastructural level binding of the antibody was observed to be restricted to the nucleoplasmic side of the inner nuclear membrane (Fig. 1D). During mitosis, the antigens recognized by mAb 13d4 were found dispersed in the cytoplasm (Fig. 1B). In immunoblot experiments the mAb 13d4 recognized three protein bands of Mr 75,000, 53,000 and 39,000, respectively, of which the Mr 53,000 protein band was labelled the strongest. The Mr 39,000 band appeared as a double band, and in some cell types, particularly in certain cultured cell lines, was barely detectable (Fig. 2). These proteins are conserved among mammals, as they were detectable in immunoblots from marsupials to humans. However, the epitope recognized by mAb 13d4 is apparently not conserved throughout vertebrates, as negative results were obtained in the amphibian cell line A6 by both immunofluorescence microscopy (not shown) and immunoblotting (Fig. 2).

In order to further characterize the proteins recognized by mAb 13d4, we screened a rat cDNA expression library using this antibody. We isolated one 1.8 kb long cDNA (rp1.8 clone), which was subsequently sequenced. A comparison of the deduced amino acid sequence revealed that our clone coded for the full-length sequence of rat LAP2 (100% identity) previously reported by Furukawa et al. (1995). Further evidence for the identity of the proteins recognized by mAb 13d4 is provided in Fig. 3. SG13009 bacteria were transformed with the expression vector pQE31 containing the rp1.8 clone (pQE31/rp1.8). As expected, induced bacteria expressed a Mr 53,000 protein that was recognized in immunoblots with mAb 13d4 (Fig. 3). No protein band was recognized in transformed, but non-induced bacteria (Fig. 3).

As summarized in the introductory paragraph, the available evidence indicates that LAP2 is the rat homologue of human protein TPβ (Harris et al., 1994, 1995; Furukawa et al., 1995). Based on their molecular masses (Mr 75,000 and 39,000) and their immunological relationship to LAP2, it appeared reasonable to postulate that the two additional major protein bands recognized by mAb 13d4 (see Fig. 2) are the rat homologues of human TPa and y. If this is the case, the epitope recognized by mAb 13d4 is likely contained within the first 187 amino acids of the LAP2 sequence, i.e. the only domain shared by all TPs (Harris et al., 1994; Berger et al., 1996). This point was addressed in the experiment described in Fig. 3. A deletion mutant coding for the leading sequence and the first 187 amino acids of rp1.8 clone was inserted in vector pQE31 (construct pQE31/Δ 1–187) and the protein expressed in bacteria. As expected, only one low molecular mass band was recognized by mAb 13d4 on blots of lysates from induced bacteria.

Taken together, we conclude that mAb 13d4 recognizes three nuclear envelope proteins of Mr 75,000, 53,000 and 39,000, which correspond to TPα, LAP2/TPβ and TPγ, respectively.

Hybridoma cell line 13d4 was initially selected in a large scale immunocytochemical screen, seeking mAbs that differentially label spermatogenic nuclei. In this report we describe the expression and localization patterns of the mAb 13d4 immunoreactive TPs during spermatogenesis.

When tested by immunoblotting, the mAb 13d4 detected in lysates of whole pachytene spermatocytes and early round spermatids the same three protein bands as in somatic cells (Fig. 4), indicating that the same complement of TPs is expressed in these spermatogenic cells. Fig. 5A shows the localization of TPs on cryosections of rat testes using mAb 13d4. Although weaker than in somatic cells present in the same section, the signal was restricted to the whole periphery of spermatogonia and spermatocyte (meiotic stages) nuclei. However, this labelling pattern was not maintained during the later postmeiotic stages, in which dramatic changes were observed (Figs 5 and 6). In round spermatids the staining with mAb 13d4 was at the nuclear periphery, although the first indications of a polarization of the signal to one half of the nuclear periphery were apparent (Fig. 6B). With the progression of spermiogenesis this polarization became more evident (Figs 5A, 6C). Finally, in elongating spermatids the labelling was restricted to one spot at the nuclear periphery. This spot was located at the centriolar pole of the nucleus (Fig. 6D,E). Interestingly, the changes in the distribution of TPs during spermiogenesis are accompanied by a redistribution of lamin B1, the only lamin known to be expressed in postmeiotic stages of mammals (see Vester et al., 1993). As shown in Figs 5B and 6F, lamin B1-staining was more intense in half of the nuclear periphery of late round spermatids, which later became concentrated to a spot in the centriolar pole of elongated nuclei. However, in contrast to TPs, a weak staining of the whole nuclear periphery remained till late spermiogenesis (Fig. 6F). This weak nuclear periphery labelling with mAb L3f4 indicates that the more limited reactivity obtained with mAb 13d4 in elongated spermatids is not due to a general inaccessibility of the immunoglobulins to their epitopes. Later on, neither TPs nor lamins were detectable by immunofluorescence microscopy of mature sperms present in the epididymus (not shown; see below).

The expression pattern of TPs during spermiogenesis was also followed by immunoblotting. For this, lysates of spermatids (enriched in three different maturation stages) and mature sperms were prepared and electrophoresed in parallel. As shown in Fig. 7A (lanes 1 and 2), during earlier stages of spermiogenesis the expression pattern of TPs remained virtually unchanged. The first differences were observed in the elongating spermatids (Fig. 7A). Here TPy was barely detectable. However, the intensity of the bands corresponding to the other two TPs remained essentially the same. At the end of spermiogenesis LAP2/TP0 also became undetectable, so that the mature sperm appears to contain only TPa (Fig. 7A). As judged from the band intensity, the amount of TPa per cell appears to remain constant throughout spermiogenesis. According to these data, the negative results obtained by immunofluorescence microscopy with mAb 13d4 in mature sperms appear to be due to the limited accessibility of nuclear structures to the immunoglobulins. This is not surprising since similar limitations have been reported repeatedly in the past using different antibodies against nuclear components of mammalian mature sperms.

For comparison, we also investigated the expression of lamin B1 by immunoblotting. As shown in Fig. 7B, the protein band corresponding to lamin B1 becomes weaker as spermiogenesis progresses. This is particularly evident in the elongating spermatids. Finally, in mature sperms isolated from the epididymus no lamin B1 was detectable. This is in agreement with the results of Kaufmann (1989), who reported that in mature rat sperms somatic lamins are undetectable (below 2% of the level of somatic cells).

In the present study we have introduced a new mAb (13d4) that is specific for the TPs. The epitope recognized by this antibody is located in the amino-terminal domain of LAP2/TPβ, i.e. the only domain shared by all members of this protein family. The epitope is conserved in mammals, as indicated by our results in marsupial and human cells, but it is apparently not conserved throughout vertebrates, as shown by the negative results in one amphibian species (Figs 1–3). In addition to the three major bands corresponding to the TPs α, β (LAP2) and y, our mAb 13d4 also recognizes some minor bands in certain mammalian tissues, which, according to their apparent molecular masses, most likely correspond to the recently described additional splicing variants of the TP gene (Fig. 2, lane 1; see Berger et al., 1996). However, our results (Figs 1 and 2) contrast with those of Ishijima et al. (1996), who reported that any form of TPs was barely detected in the liver.

Previous immunolocalizations using an antibody specific for LAP2/TP0 have shown that this protein is a component of the inner membrane of the nuclear envelope (Foisner and Gerace, 1993). Furthermore, cell fractionation experiments and sequence data strongly support the notion that LAP2/TPβ is a transmembrane protein associated with the nuclear lamina (Foisner and Gerace, 1993; Furukawa et al., 1995). Our immunolocalization experiments with mAb 13d4 showed a specific labelling of the nuclear envelope. No other cell components were stained with the antibody, as shown at the light and electron microscopical levels (Fig. 1). These results indicate that all members of the TP family are located in the inner nuclear membrane of interphase cells. An interesting aspect to be elucidated is how TPα becomes targeted to the nuclear envelope since it lacks a potential transmembrane domain as well as the lamin binding domain described in LAP2/TPβ.

In this study we present evidence that TPs and lamin B1 undergo dramatic changes in their expression and distribution patterns during rat spermiogenesis. As shown by immunoblotting, the amount per cell of TPs β (LAP2) and γ and of lamin B1 decreases during late spermiogenesis. Concomitantly, at the immunofluorescence level we observed that labelling with both anti-TP and anti-lamin antibodies becomes progressively concentrated to the centriolar pole of the nucleus of elongating spermatids. In the case of lamin B1, a weak signal along the whole nuclear periphery of elongating spermatids was also observed. Of the nuclear envelope proteins investigated here, only TPα was consistently detectable during the whole process of spermiogenesis and in the mature sperm. Interestingly, the behaviour of the nuclear envelope proteins described here is comparable to that of pore complexes. In late round and elongating spermatids, pore complexes are either lost or shift to the nuclear centriolar pole (for review see Clermont et al., 1993).

The mechanisms that lead to a redistribution of TPs and lamin B1 during spermiogenesis are presently unknown. Also unknown is the mechanism responsible for the progressive disappearance of these proteins. It would be particularly interesting to establish how such a mechanism distinguishes between TPα and the other TPs. In this context, it is of interest to mention that the expression of enzymes involved in the degradation of proteins via ubiquitination has been recently reported in spermiogenic cells (Koken et al., 1996; Wing et al., 1996; see also Rivkin et al., 1997). At present, it is not clear what the reasons for the selective disappearance of only two of the three TP isoforms are. Does TPα play a role after fertilization? It may be of interest to keep in mind that TPα is the only member of the family that lacks a potential transmembrane domain (Harris et al., 1994).

The nuclear envelope at the centriolar pole (preimplantation fossa) of rat sperms has several peculiarities that are worth summarizing. (i) It is a specialized region showing interaction with chromatin. During spermiogenesis chromatin condensation is initiated at the level of preimplantation fossa, and following fertilization it shows late chromatin decondensation (see Collas and Poccia, 1996, and references therein). (ii) There is growing evidence that in the sperm nucleus chromatin is non-randomly distributed (e.g. Zalenski et al., 1995, 1997) and organized into topologically constrained DNA loop domains which are different from those of somatic cells (Ward et al., 1989). For example, in the rat, as in some other species, telomere sequences are clustered in the posterior part of sperm nuclei. To our knowledge, however, the existence of chromosomal proteins involved in the association of telomeres with the nuclear envelope of sperms has not yet been demonstrated (see Zalenski et al., 1997). (iii) Finally, in the preimplantation fossa region of mammalian sperm (and in other species also in the acrosomal region), Collas and Poccia (1995, 1996) have described the existence of the so-called lipophilic structures. These nuclear envelope structures are thought to be indispensable for the formation of the male pronuclear envelope around sperm chromatin following fertilization (for a recent review see Poccia and Collas, 1997).

What is the significance of redistribution and disappearance of TPs and lamin B1 during rat spermiogenesis? Two possibilities will be discussed here which, however, are not mutually exclusive.

(i) As previously described, LAP2/TP0 has affinity in interphase somatic cells for B-type lamins and chromatin (Foisner and Gerace, 1993; Furukawa et al., 1998). The binding partners of TPa and y are not known. Lamin B1 binds, in addition to LAP2/TPβ, to chromatin, as well as to A-type lamins and the lamin B receptor (for review see Gerace and Foisner, 1994; Georgatos et al., 1994; Cowin and Burke, 1996; Ye and Worman, 1996). According to these binding properties, it is tempting to speculate that the TPs and lamin B1 are involved in the changes in chromatin organization that occur during spermiogenesis. By moving to the posterior pole, the chromatin binding sites at the nuclear envelope would contribute to the achievement of the non-random, spermspecific chromatin distribution. Furthermore, evidence has been presented earlier indicating that nuclear envelope proteins, in particular the lamins, play an important role in chromatin dynamics and function. In cultured somatic cells, in which postmitotic lamin reassembly was severely perturbed by the microinjection of metaphase cells with lamin antibodies, daughter nuclei become arrested in a telophase-like configuration. In these cells, the chromatin was surrounded by a double membrane but remained highly condensed and apparently inactive (Benavente and Krohne, 1986; see also Yang et al., 1997). Comparable results were obtained in Xenopus egg extracts in which the major egg lamin (lamin LIII) was immunodepleted or its assembly inhibited by the addition of specific antibodies. In these cases, reformed nuclei remained small and the chromatin in a compact configuration (Newport et al., 1990; Meier et al., 1991; Goldberg et al., 1995; Spann et al., 1997). According to this evidence, we propose here that the progressive disappearance of lamin B1 and most of the TPs observed in elongating spermatids may be of significance in the process of chromatin condensation and inactivation that normally takes place during spermiogenesis.

(ii) During spermiogenesis the nuclear envelope undergoes a profound reorganization that leads to the formation of a domain in the posterior pole, which differs in morphology from other regions of the spermatid nuclear periphery. As described above, we have localized the TPs and lamin B1 to the posterior pole of elongating rat spermatid nuclei. These findings are reminiscent of the situation in the sea urchin in which B lamins and a lamin B receptor-like molecule were located to the posterior pole of mature sperm nuclei. In addition, these two proteins were also detected in the acrosomal pole of the same cells (Collas et al., 1996). Interestingly, in the sea urchin the lipophilic structures mentioned above were found associated with the posterior as well as with the acrosomal poles. However, in the rat, lipophilic structures were only found in the posterior pole (Collas and Poccia, 1996). At least in these two species, there is a good correlation between the localization of nuclear envelope proteins at the end of spermiogenesis and the localization of lipophilic structures (Collas et al., 1996, and the present study). Although the composition of these structures is not known and its origin has not yet been investigated, available data lead us to envisage the interesting possibility that the movement of nuclear envelope proteins during spermiogenesis described in this study may be somehow related to the assembly and positioning of lipophilic structures.

We thank Ulrich Scheer for continuous support, Avril Smith for preparing mAb 13d4, David Lourim and Georg Krohne for helpful discussions, Robert Hock for help with the confocal microscope, Corinna Zünkler and Claudia Gehrig for excellent technical assistance, and Lynn Leverenz, David Lourim, Rosie Rudd and Micheline Paulin-Levasseur for correction of the manuscript. Supported by a grant of the Deutsche Forschungsgemeinschaft to R.B. (Be 1168/4-2).