Characterization of a novel meiosis-specific protein within the central element of the synaptonemal complex

The mammalian meiotic cell division, in which DNA replication is followed by two successive rounds of chromosome segregation (meiosis I and II), gives rise to genetically diverse haploid gametes. During meiosis I the homologous chromosomes, each consisting of one pair of sister chromatids, move to opposite poles, whereas during meiosis II the sister chromatids are separated into haploid cells. The prophase of the first meiotic division is highly regulated and can be cytologically subdivided into four stages called leptonema (chromatin condensation), zygonema (synapsis of homologous chromosomes), pachynema (full synapsis) and diplonema (visible chiasmata) (Zickler and Kleckner, 1999). Throughout the first meiotic prophase, the sister chromatids are held together by cohesin complex proteins whereas proper alignment and pairing of the homologous chromosomes are achieved by homologous recombination and formation of the synaptonemal complex (Petronczki et al., 2003; Page and Hawley, 2004).

The synaptonemal complex is a large zipper-like protein complex that connects one pair of sister chromatids to the homologous pair. In humans, failure in formation of the synaptonemal complex is known to cause male infertility (Miyamoto et al., 2003; Judis et al., 2004) and a high aneuploidy rate in oocytes (Hassold and Hunt, 2001; Hunt and Hassold, 2002). The details of how mammalian synaptonemal complex formation is regulated are currently unknown and more knowledge about this complex structure is essential for the understanding of many fertility problems and the striking differences between male and female meiosis in humans.

Synaptonemal complex formation starts during leptonema when the synaptonemal complex proteins 2 and 3 (SYCP2 and SYCP3) begin to form the axial elements. Knockout of the mouse Sycp3 gene, which also abolishes recruitment of SYCP2 to the axial elements, leads to male sterility. The Sycp3^-/- spermatocytes fail to form normal axial elements, display severely disrupted synapsis between the homologous chromosomes and fail to develop beyond the zygotene stage (Yuan et al., 2000; Pelttari et al., 2001; Liebe et al., 2004). By contrast, the Sycp3^-/- females are fertile, although absence of SYCP3 in oocytes does cause a severe disruption of synapsis and a reduction of meiotic recombination and chiasmata formation, which subsequently leads to a high aneuploidy rate (Yuan et al., 2002; Lightfoot et al., 2006).

Later, when the homologous chromosomes become synapsed during the zygotene stage, the axial elements (now referred to as lateral elements) are joined by the transverse filaments formed by the synaptonemal complex protein 1 (SYCP1). SYCP1 molecules are long coiled-coil proteins with two globular heads that form parallel homodimers, with their C-termini embedded in the lateral elements and their N-termini interacting in a more dense region in the middle called the central element (Liu et al., 1996; Schmekel et al., 1996; Ollinger et al., 2005).

Although they have different primary amino acid sequences, the predicted secondary protein structure and function of the transverse filament proteins in other organisms, such as C(3)G in Drosophila melanogaster, SYP-1 and SYP-2 in Caenorhabditis elegans and Zip1p in Saccharomyces cerevisiae, are remarkably conserved (Page and Hawley, 2004). Apart from abrogated synapsis, mutation of the corresponding genes causes a severe reduction of crossover formation in S. cerevisiae (Sym et al., 1993; Borner et al., 2004) and even a complete abolishment of crossovers in D. melanogaster (Page and Hawley, 2001) and C. elegans (MacQueen et al., 2002; Colaiacovo et al., 2003). Knockout of the mouse transverse filament gene Sycp1 leads to both male and female infertility, as most meiotic cells undergo apoptosis during the pachytene stage (de Vries et al., 2005). As described for most other organisms, axial element formation and alignment of the homologous chromosomes proceed normally in Sycp1^-/- spermatocytes, whereas their synapsis never occurs. Additionally, SYCP1 has been suggested to play a role in the development of early meiotic recombination intermediates into crossovers and in formation of the XY body (de Vries et al., 2005).

In comparison to the transverse filaments, composition, formation and function of the central element, the dense structure in the center of the synaptonemal complex, are currently under characterized. A functional study in yeast, however, did indicate separate functionality of the central element region by showing that the N-terminal region of Zip1p, although affecting crossover formation, is not required for chromosome synapsis or localization of Zip1p to chromosomes (Tung and Roeder, 1998). Moreover, two additional non-transverse filament proteins (Zip2p and Zip3p) were shown to be associated with Zip1p and to function in both synapsis and recombination (Chua and Roeder, 1998; Agarwal and Roeder, 2000).

The morphology of the central element has been studied in detail by electron microscopy using insect (D. melanogaster and Blaps cribosa) and rat spermatocytes (Schmekel and Daneholt, 1995). The central elements consist of three to four layers of transverse filament components that are longitudinally connected by pillar-shaped protein structures (Schmekel and Daneholt, 1995). Moreover, it has been noted that these structures contain material in addition to the transverse filament proteins (Solari and Moses, 1973). Although it has been estimated that the synaptonemal complex as a whole consists of more than ten meiosis-specific proteins (Heyting et al., 1989), only two proteins, called synaptonemal complex central element 1 and 2 (SYCE1 and SYCE2), have been identified and specifically localized to the central element region (Costa et al., 2005).

Here we present and characterize a novel meiosis specific protein called testis expressed sequence 12, or TEX12, which we show to be expressed during both male and female meiosis. We have studied the spatio-temporal behavior of this protein, using immunohistochemistry, immunocytology and immunogold electron microscopy. Additionally, we studied its interactions with the transverse filament protein SYCP1 and the central element proteins SYCE1 and SYCE2 using co-immunoprecipitations from total testis lysates. We identified TEX12 to be a novel central element protein that co-localizes and interacts with SYCE2, as opposed to SYCE1, which more strongly interacts with SYCP1 and thus the transverse filaments. Our studies uncover some of the complexity and molecular networks of the synaptonemal central element structure.

From a set of genes, initially identified using a male-germ-cell-specific RNA expression screen (Wang et al., 2001), we selected a subset of genes for more-detailed studies with the aim of identifying meiosis-specific genes. Using in situ hybridizations, we identified one gene, with the name Tex12 (testis expressed sequence 12), whose mRNA expression appeared similar to that of the Sycp3 gene in both testis (Fig. 1A) and ovary (Fig. 1B) sections.

For analysis of the TEX12 protein, we generated several antibodies against different regions of the predicted protein sequence. Two of these antibodies were found to label the expected 14 kDa band on western blots (Fig. 1C) and were both used for further experimentation.

By using these antibodies for immunohistochemistry on testis sections we were able to accurately determine spatio-temporal protein expression during spermatogenesis. Both antibodies stained the nuclei from leptotene spermatocytes until step 5 round spermatids, peaking in intensity during pachynema (Fig. 1D), consistent with a role for TEX12 in meiotic cell division.

In order to determine localization of the TEX12 protein on the meiotic chromosomes, we performed immunofluorescence microscopy on spread spermatocytes (Fig. 2A). For proper determination of the meiotic stages we co-labeled the cells for SYCP3 and the centromeric marker CREST. The TEX12-antibodies exclusively marked the synapsed axes of the meiotic chromosomes, giving a punctate signal at zygonema and a more continuous staining at the pachytene stage. Notably, not all pachytene cells stained at equal intensity. Whereas early and mid pachytene cells displayed a very bright and continuous staining, TEX12 appeared punctate and less bright in pachytene cells progressing towards diplonema (late pachytene in Fig. 2). These cells could also be referred to as early diplonema because they start to display thickened chromosome ends. TEX12 localization during zygonema and late pachynema is fragmented but, notably, TEX12 does not form clear countable foci like MSH4 or MLH1, nor does it co-localize with these proteins (data not shown). Zygotene cells can be distinguished from diplotene cells because the centromeric regions are not yet synapsed during this stage (two CREST dots), whereas the diplotene cells display de-synapsed regions elsewhere on the chromosomes, and later homologues that are held together by chiasmata until the two meiotic divisions. On the XY chromosomes only the synapsed pseudo-autosomal synapsed regions stained for TEX12. In cells at the diplotene stage, TEX12 appeared bright on synapsed axes but was completely absent from de-synapsed axes.

To further investigate whether TEX12 specifically localizes to synapsed chromosomes, we labeled meiotic cells for TEX12 and the central element protein SYCP1. Indeed, TEX12 follows the localization of SYCP1 at the synapsed chromosomes (Fig. 2B) but SYCP1 appeared earlier and disappeared later from the synapsed chromosomes. Again, TEX12 displayed a more punctate localization pattern in late pachytene cells, whereas SYCP1 staining appears continuous in all pachytene cells (Fig. 2B).

To show that TEX12 localization is not male specific, we also performed the localization experiments using oocytes. TEX12 displays the same localization pattern in oocytes as in spermatocytes (supplementary material Fig. S1), indicating that TEX12 function is not sex specific but a general part of the mammalian synaptonemal complex.

Finally, in order to determine TEX12 localization within the synaptonemal complex, we performed immunogold electron microscopy on mouse and rat spermatocytes (Fig. 3). Like the central element proteins SYCE1 and SYCE2 (Costa et al., 2005), the TEX12 protein also appeared to be specifically present at the dense central element region in which the SYCP1 homodimers interact through their N-termini (Fig. 3A). The gold particles staining TEX12 (Fig. 3A) were more centrally located than immunogold labeling of the coiled-coil domain of SYCP1 and thus the transverse filaments (Fig. 3B).

Hence, both timing and localization show that TEX12 plays a role in the central element of the synaptonemal complex at the synapsed homologous chromosomes.

To further characterize TEX12 localization in relation to the other two central-element-specific proteins, we performed additional co-localization studies on spread meiotic cells for TEX12 with both SYCE1 and SYCE2 (see Fig. 4 and supplementary material Fig. S1 for antibody characterization). Interestingly, we found TEX12 to have the same staining pattern as SYCE2, in contrast to SYCE1, which instead displayed a staining pattern very similar to SYCP1. Whereas SYCE1 (Fig. 4), like SYCP1 (Fig. 2B), displayed a more continuous staining from zygonema onwards, both TEX12 and SYCE2 often appeared in a punctate manner (Fig. 4). Moreover, in a double-labelling experiment, TEX12 and SYCE2 staining precisely overlap (Fig. 4), whereas axial areas that label for SYCE1 (or SYCP1) do not necessarily also label for TEX12 (Fig. 2B, Fig. 4). We obtained the same results when using oocytes instead of spermatocytes (data not shown). Based on timing and localization, TEX12 seems to be associated with SYCE2, whereas SYCE1 seems to interact more directly with SYCP1.

Using mouse knockout models, we investigated whether absence of the lateral element protein SYCP3 (Yuan et al., 2000), the cohesin subunit SMC1β (Revenkova et al., 2004) or the central element protein SYCP1 (de Vries et al., 2005) would influence TEX12 localization. Neither absence of SYCP3 (Fig. 5A) nor SMC1β (data not shown) influenced timing or localization of TEX12 on the synapsed meiotic chromosomes. However, aberration of SYCP1 completely abolished TEX12 staining on meiotic chromosomal axes (Fig. 5B), showing that the presence of SYCP1 is required for proper TEX12 localization. Additionally, we also found that localization of SYCE1 and SYCE2 (Fig. 5B) depended on SYCP1. Again, the same results were obtained when oocytes instead of spermatocytes were used (data not shown). These results demonstrate that only the transverse filaments, and not the lateral elements or cohesins, are essential for TEX12 localization.

To further characterize the co-localization between TEX12 and SYCE2, we performed co-immunoprecipitations from total testis lysates using antibodies against SYCE1, SYCE2 and SYCP1 (Fig. 6). Using western blot analysis, we found that TEX12 interacted solely with SYCE2 (Fig. 6), which is not unexpected because these proteins precisely co-localized throughout the first meiotic prophase (Fig. 4). Because, the rabbit anti-SYCE2 antibodies do not work on western blots we performed the reverse co-immunoprecipitation experiments using the Seize® X Protein A Immunoprecipitation kit (Pierce), which allowed both pull down and detection to be performed with our guinea pig antibodies. In these experiments we found that only SYCE2 was pulled down with TEX12 (Fig. 6).

We were also able to confirm the previously described interaction between SYCP1 and SYCE1 (Costa et al., 2005). However, like Costa et al., we did not detect this interaction in the reverse experiment, which is probably due to the difficulty of co-precipitating SYCP1 from the large synaptonemal complex (Heyting et al., 1985; Heyting et al., 1989).

When overexpressed in cultured somatic cells, SYCP1 forms fibres (Ollinger et al., 2005) that interact with the central element proteins SYCE1 and SYCE2 (Costa et al., 2005). Considering the spatio-temporal distribution of TEX12 on meiotic chromosomes, we studied possible interactions of these proteins with TEX12 by co-overexpressing them in cultured cells. However, it appeared that TEX12, when overexpressed ex vivo, generated non-specific interactions with co-overexpressed proteins (data not shown).

Our results suggest a strong direct interaction between TEX12 and SYCE2. SYCE1, on the other hand, seems to interact more directly with SYCP1, possibly establishing a link between the transverse filaments and the central element proteins.

We have characterized a novel meiotic protein, TEX12, which specifically localizes to the central element of the synaptonemal complex at the synapsed axes of homologous meiotic chromosomes. When the central elements are disrupted by the absence of SYCP1, TEX12 no longer localizes to the meiotic chromosomes, whereas disruption of the lateral elements or a meiosis specific cohesion subunit does not influence TEX12 localization. Furthermore, TEX12 co-localizes and co-precipitates with SYCE2 during the meiotic cell division. Taken together, our data show that TEX12, together with SYCE2, is an important part of the central element of the synaptonemal complex after synapsis of the homologous chromosomes has occurred.

The Tex12 gene is exclusively expressed during the meiotic cell division and its promoter, as for many other germ cell specific genes, is suppressed in somatic cells by the transcription factor E2F6 (Pohlers et al., 2005). The TEX12 protein is small (14 kDa) and does not contain any known protein domains. Reciprocal BLAST searches primarily reveal orthologues in mammals. However, albeit with low sequence homology, a predicted TEX12 orthologue also exists in zebrafish (Danio rerio), which is in line with SYCE1 and SYCE2, which are conserved in vertebrates (Costa et al., 2005). The absence of sequence homologues in lower eukaryotes can be explained in two ways. Either the transverse filaments are formed differently in these organisms, or there are analogous proteins that carry out similar functions without their gene sequences being conserved. The latter explanation applies to the known transverse element proteins in various organisms such as SYCP1 in mammals, C(3)G in Drosophila, SYP-1 and SYP-2 in C. elegans and Zip1p in S. cerevisiae. Although these proteins have different primary amino acid sequences their secondary protein structure and function are conserved (Page and Hawley, 2004).

The transverse filament proteins are not only involved in synapsis of the homologous chromosomes but, as described for various other eukaryotes, are also involved in the formation of meiotic crossovers (Hunter, 2003; Page and Hawley, 2004). By analyzing Sycp1^-/- spermatocytes, the mammalian SYCP1 has also been suggested to be involved in the development of early recombination intermediates into crossovers and the formation of XY bodies (de Vries et al., 2005). If this holds true, TEX12, SYCE2 and SYCE1 are almost sure to be involved because absence of SYCP1 abolishes the presence of these central element proteins on meiotic chromosomes. To be more concise, one should approach the Sycp1^-/- mouse as a quadruple knockout model for Sycp1, Syce1, Syce2 and Tex12.

Although most studies do not functionally distinguish the transverse filaments from the central elements, one study showed that the N-terminal region of the yeast Zip1p affects crossover formation without being essential for chromosome synapsis (Tung and Roeder, 1998), indicating separate functionality of the central element region. Additionally, two Zip1p-associated non-transverse filament proteins (Zip2p and Zip3p) were shown to function in both synapsis and recombination (Chua and Roeder, 1998; Agarwal and Roeder, 2000), demonstrating that the transverse filament proteins use other proteins to link synapsis with recombination. Using electron microscopy, the insect (D. melanogaster and B. cribosa) and rat synaptonemal complex have also been shown to contain central element specific proteins, which are visible as pillar-shaped protein structures that connect to the transverse filaments in the center of the synaptonemal complex (Solari and Moses, 1973; Schmekel and Daneholt, 1995). Although it has been estimated that the synaptonemal complex as a whole consists of at least ten meiosis-specific proteins (Heyting et al., 1989), only SYCE1 and SYCE2 have been identified and specifically localized to the central element region to date (Costa et al., 2005). We have now not only added TEX12 to the list of central-element-specific proteins, but also uncovered some of the molecular networks that link the central element proteins to each other and the transverse filaments.

How exactly the central element proteins affect meiotic events is currently unknown. One hypothesis postulates that the mechanical properties of the synaptonemal complex, like robustness versus flexibility, affect how meiotic double-strand breaks are resolved (Moens, 1978; Blat et al., 2002; Borner et al., 2004). In this case TEX12, and other proteins, would only influence the physical properties of the synaptonemal complex. Alternatively, the central element proteins also influence other proteins that more directly relate to crossover formation, like the mismatch repair proteins MHL1 and MLH3 that were both shown to depend on the transversal filament and/or central element proteins for their localization to meiotic crossovers (de Vries et al., 2005).

Since SYCE2 has been shown to interact with both SYCP1 and SYCE1 (Costa et al., 2005), we propose a model in which TEX12 forms a complex with SYCE2, which in turn interacts with SYCE1. SYCE1 is likely to function as a bridge between the central element proteins and the transversal filaments by binding to SYCP1 and thus anchoring TEX12 and SYCE2 to the transversal filaments.

In situ hybridization was performed using standard protocols as described (Schaeren-Wiemers and Gerfin-Moser, 1993) and in more detail in the DIG application manual (Roche), using DIG-labelled antisense RNA probes and the fluorescence antibody enhancer set for DIG detection (Roche).

Antibodies against TEX12, SYCE1 and SYCE2 were raised in guinea pigs using short peptides coupled to keyhole limpet hemocyanin as described (Kouznetsova et al., 2005). Anti-TEX12 antibodies were raised against amino acids 15-34 and 35-61. Anti-SYCE1 antibodies were raised against amino acids 302-329. Anti-SYCE2 antibodies were raised against amino acids 1-28. The individual antisera were affinity purified on columns coupled to the corresponding peptide. For western blot analysis of the antibodies, cell (C2C12) or testis lysates were separated on a 4-12% poly-acrylamide gel (Invitrogen) and blotted onto an Immobilon-P membrane (Millipore) (supplementary material Fig. S2). The antibodies were used at 1:250 (TEX12), 1:1000 (SYCE1) and 1:400 (SYCE2) dilutions and detected using horseradish-peroxidase-conjugated donkey anti-guinea pig secondary antibodies (Jackson Immunoresearch Laboratories) diluted 1:5000.

Wild-type mice and Sycp1^-/- (de Vries et al., 2005), Sycp3^-/- (Yuan et al., 2000) or Smc1β^-/- (Revenkova et al., 2004) mice were used and maintained according to regulations provided by the animal ethical committee of the Karolinska Institute who also approved the experiments.

Immunohistochemistry was performed as described (Hamer et al., 2001) using guinea pig anti-TEX12 (1:500). Immunocytochemistry was performed as described (Kouznetsova et al., 2005) using a `drying-down' technique (Peters et al., 1997) and the following antibodies: rabbit anti-SYCP1 (1:50) (Liu et al., 1996), rabbit anti-SYCP3 (1:200) (Liu et al., 1996), human anti-CREST (1:1500) (Hadlaczky et al., 1986), rabbit anti-STAG3 (Pezzi et al., 2000), guinea pig anti-SYCE1 (1:1500) and -SYCE2 (1:400), rabbit anti-SYCE1 (1:500) and anti-SYCE2 (1:100) (Costa et al., 2005) and guinea pig anti-TEX12 (1:200). Secondary antibodies were applied as described (Kouznetsova et al., 2005). Electron microscopy and immunogold labeling were performed on 5 nm cryostat sections of shock-frozen rat testes according to standard protocols (Smith and Benavente, 1992) using guinea pig anti-TEX12 and rabbit anti-SYCP1 (Ollinger et al., 2005).

Slides were viewed at room temperature using Leica DMRA2 and DMRXA microscopes and 100× objectives with epifluorescence. Images were captured with a Hamamatsu digital charge-coupled device camera C4742-95 and Openlab™ software version 3.1.4. Images were processed using Adobe Photoshop version 9.0.

For co-immunoprecipitation, whole testes were lysed in immunoprecipitation (IP) buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 5 mM EDTA) complemented with complete protease inhibitor (Roche). Volumes equivalent to 0.5 mg of protein were taken from whole testis lysates and IP buffer was added to a total volume of 200 μl. 15 μl protein-A-agarose beads (Repligen) per sample were washed five times with 1% BSA in PBS (by spinning down the beads and adding fresh buffer). 10 μl rabbit anti-SYCP1 (Liu et al., 1996), rabbit anti-SYCE1 (Costa et al., 2005), rabbit anti-SYCE2 (Costa et al., 2005) or rabbit serum (Vector Laboratories) was incubated with the beads in 300 μl IP buffer for 2 hours at 4°C. Simultaneously, the prepared lysates were pre-cleared with washed beads for 2 hours at 4°C. After spinning down the beads, the pre-cleared lysates were added to the incubated beads giving a total volume of 500 μl and incubated for another 2 hours at 4°C. The beads were then washed once in IP buffer, dried and analyzed by western blot analysis as described, using guinea pig anti-SYCP1 (Kouznetsova et al., 2005) (1:250) or the guinea pig antibodies against SYCE1, SYCE2 or TEX12.

The `reverse' IPs, using TEX12 as bait, were performed using the Seize® X Protein A Immunoprecipitation kit (Pierce) in which the first antibodies are covalently linked to the protein-A-agarose beads, thus allowing guinea pig antibodies for both precipitation and detection. However, after diluting the interacting proteins from the beads, we used centrifugation (30 seconds, 10,000 g), instead of the filters provided by this kit, to separate the beads from the interacting proteins as the testis samples were clogging these filters.

We thank Yael Costa and Howard J. Cooke for sharing their antibodies against SYCE1 and SYCE2; Christa Heyting and Albert Pastink for the Sycp1-deficient mice; Rolf Jessberger for the Smc1β-deficient mice. This work has been supported by the Swedish Cancer Society, the Swedish Research Council, the Axel Wenner-Gren Foundation and the Karolinska Institute.