Hydra meiosis reveals unexpected conservation of structural synaptonemal complex proteins across metazoans

Abstract

The synaptonemal complex (SC) is a key structure of meiosis, mediating the stable pairing (synapsis) of homologous chromosomes during prophase I. Its remarkable tripartite structure is evolutionarily well conserved and can be found in almost all sexually reproducing organisms. However, comparison of the different SC protein components in the common meiosis model organisms Saccharomyces cerevisiae, Arabidopsis thaliana, Caenorhabditis elegans, Drosophila melanogaster, and Mus musculus revealed no sequence homology. This discrepancy challenged the hypothesis that the SC arose only once in evolution. To pursue this matter we focused on the evolution of SYCP1 and SYCP3, the two major structural SC proteins of mammals. Remarkably, our comparative bioinformatic and expression studies revealed that SYCP1 and SYCP3 are also components of the SC in the basal metazoan Hydra. In contrast to previous assumptions, we therefore conclude that SYCP1 and SYCP3 form monophyletic groups of orthologous proteins across metazoans.

Most eukaryotic organisms reproduce sexually, a process that involves the fusion of gametes of the two sexes to form a new organism. A prerequisite is that the chromosome complement becomes reduced from diploid to haploid during germ cell differentiation. This is achieved during a special type of cell division, the meiosis. The original diploid state is then reconstituted during fertilization.

The reduction of the chromosome number is achieved by the succession of two rounds of chromosome segregation after a single round of DNA replication. During the first and most crucial round, homologous chromosomes need to pair and go through cross-over recombination to ensure their correct segregation into two new daughter cells. The association of sister chromatids is maintained and will not be abrogated until the onset of the second round of cell division, which resembles a mitotic one.

As the essential features of meiosis, i.e., pairing and cross-over recombination of homologs, can be found in a wide range of different taxonomic groups, it is assumed that meiosis evolved only once early in eukaryotic life (1). The meiotic recombination machinery, which is responsible for the synthesis of crossing over, shows strong evolutionary conservation and can be found in the meiotic model organisms from the yeast Saccharomyces cerevisiae across Caenorhabditis elegans and Drosophila melanogaster to mammals with only little variation (2, 3). The mechanism of homology search is still not well understood, but the assembly of a proteinaceous structure known as synaptonemal complex (SC), which mediates synapsis of homologs, seems likely to be an important facilitator or stabilizer of homologous pairing (4, 5). Synaptic defects consequently impair meiosis and lead to failure of cross-over recombination and chromosome segregation (chromosome nondisjunction) in diverse organisms [for reviews see (6, 7)] that can cause aneuploidy as well as infertility (5, 8). Therefore, the functions of the SC, which is implemented by its remarkable tripartite structure, seem to be evolutionarily conserved (5).

Electron microscopic data revealed that the fine structure of the SC is also well conserved in evolution. It is similar in shape to a ladder in which two parallel rodlike lateral elements (LEs) are joined by several crosswise arranged transverse filaments (TFs). Chromatin of homologous chromosomes is associated with one LE each. The central region of the SC where the TFs of opposing LEs overlap is referred to as central element (CE).

In mammals, protein SYCP1 is the main constituent of TFs (9) whereas LEs are largely composed of the proteins SYCP2 and SYCP3 (10, 11). In addition to the N-terminal head domain of SYCP1 molecules, the CE is composed of the CE-specific proteins SYCE1, SYCE2, SYCE3, and Tex12 [(12–14); for review see (15)]. SYCP1 and SYCP3 are particularly interesting as they are the only bona fide structural SC proteins so far characterized in mammals (16). SYCP3 can self-assemble in the absence of other SC proteins, thereby forming thin fibrils that can associate laterally to generate higher-order structures resembling LEs (16, 17). SYCP1 forms so-called polycomplexes on its own that represent stacks of normally appearing SC central regions (17).

As the fine structure of the SC is well conserved in evolution, it is reasonable to assume that this would be also the case for the structural protein components. However, according to the literature, this is apparently not the case. Comparison of LE or TF protein components in the most commonly used meiosis model organisms—S. cerevisiae, Arabidopsis thaliana, C. elegans, D. melanogaster, and Mus musculus—revealed very little sequence similarity between these proteins.

The emerging picture is rather puzzling. On the one hand, the SC shows a conserved ladderlike organization from yeast to human and plays a key role in meiotic recombination. On the other hand, SC's protein composition appears to be not conserved, which suggests that the SC might have evolved independently in different taxonomic groups (5, 6, 18). To address this apparent discrepancy we performed comparative bioinformatic and expression studies focused on SC structural proteins in the basal metazoan Hydra. We provide clear evidence that, in contrast to previous assumptions, SYCP1 and SYCP3 form monophyletic groups of orthologous SC proteins across metazoans down to organisms as basal as Hydra.

Results and Discussion

SYCP1 and SYCP3 Orthologs in a Basal Metazoan.

Mammalian SYCP1 is a long fibrillar molecule composed of a central α-helical domain, most likely involved in the formation of coiled coil structures, and the flanking globular N and C termini (9, 17). Recently, Iwai and colleagues (19) reported the presence of an SYCP1 ortholog also in a nonmammalian vertebrate, the medaka fish. Comparing the sequence of mammalian (rat) and fish (medaka) SYCP1 proteins we identified, as in the case of SYCP3 (20), conserved features within taxonomically distant vertebrate species (Fig. S1). Sequence analysis of medaka SYCP1 (GenBank accession no. JQ906936) revealed that the protein has the same general organization as the rat ortholog, with a sequence identity of about 20% (sequence similarity of about 45%) in the nonhelical terminal domains and the α-helix. It is notable that we identified two previously unnoticed domains that display a much higher homology [conserved motif 1 (CM1) and 2 (CM2)]. CM1 is 83 amino acids long and comprises part of the nonhelical N terminus and the head of the coiled coil (amino acids 106–188 in the rat). It is this domain that shows one of the highest conservation among vertebrates (65% identity and 80% similarity between rat and medaka; Fig. S1). CM2 consists of 31 amino acids and is located in the last third of the protein (amino acids 724–754 in the rat). It reveals a sequence identity of 74% (97% sequence similarity) between rat and medaka (Fig. S1).

We next asked whether SYCP1 is actually restricted to vertebrate species. To address this issue, we performed extensive TBlastN searches (i.e. searches of a translated nucleotide database using a protein query) with the sequence corresponding to the highly conserved domain of rat SYCP1 (i.e., CM1). By this means we could find a great number of apparently homologous sequences across a broad range of metazoans including the cnidarians as one of the earliest phylogenetic stages. The fact that TBlastN search returned significant hits even to cnidarian Hydra magnipapillata EST and genomic sequences was rather astounding because it was previously suggested that SC proteins are not well conserved over metazoans and might have evolved independently in distant taxa (5, 6, 18). Therefore, we intended to test whether these sequences actually relate to a real Hydra SYCP1 ortholog, which would share a common origin with the mammalian SYCP1 from the same ancestral gene. A good indicator of orthology of proteins is the similarity of sequence, structure, and function. Thus, we checked for sequence similarity and structural similarity of the SYCP1 proteins in the first instance. We cloned and sequenced the respective full-length cDNA of H. vulgaris (GenBank accession no. JQ906934) and performed a detailed comparison of the deduced Hydra protein, which is 1,016 amino acids long, and the rat SYCP1 protein. It revealed that the general organization of the putative SYCP1 protein was conserved between cnidarians and mammals (Fig. 1; see also below). The highly conserved SYCP1 domains we defined above were also contained in the Hydra ortholog (41% identity and 67% similarity of CM1 and 32% identity and 65% similarity of CM2 between rat and Hydra). Alignment of CM1 sequences of a representative range of organisms further confirmed its sequence conservation during metazoan evolution (Fig. S2; for further sequences see Table S1).

Fig. 1.

Schematic comparison of SYCP1 proteins of rat (GenBank accession no. NM_012810) and Hydra (GenBank accession no. JQ906934). Sequence identity and sequence similarity (in parentheses) at the amino acid level are given (%). The gray boxes represent putative coiled coil domains. Position of conserved motifs CM1 (amino acids 106–188 in the rat; amino acids 56–138 in Hydra) and CM2 (amino acids 724–754 in the rat; amino acids 674–704 in Hydra) is denoted by red boxes. NLS, nuclear localization signal.

We next extended our study to SYCP3, the other major structural SC protein in mammals. In previous studies it has been shown that SYCP3 is composed of a central coiled coil region and the flanking N- and C-terminal domains (10). We have reported that the coiled coil domain and two short highly conserved flanking sequence motifs of the N and C termini (CM1 and CM2) are not only essential but sufficient for SYCP3 polymerization (20, 21). It is interesting to note that the calculated length of the coiled coil domain according to the Lupas algorithm may differ among vertebrates. However, the distance between CM1 and CM2 was remarkably constant from fish to mammals, namely 146 amino acids (20). TBlastN searches using the described SYCP3 domains involved in polymerization also revealed homologous sequences across metazoans including the cnidarians. As in the case of SYCP1, we cloned the full-length cDNA of the putative H. vulgaris SYCP3 (GenBank accession no. JQ906932). Analysis of the deduced Hydra SYCP3 peptide again disclosed an evolutionarily conserved domain structure of the protein, including the constant distance of 146 amino acids between the first and second CM (Fig. 2; see also below). In addition, sequence alignment of these domains (CM1, CM2, and the 146 amino acids in between) from different species once more revealed sequence conservation across metazoans (Fig. S3; for further sequences see Table S2).

Fig. 2.

Schematic comparison of SYCP3 proteins of rat (GenBank accession no. NM_013041) and Hydra (GenBank accession no. JQ906932). Sequence identity and sequence similarity (in parentheses) at the amino acid level are given (%). The gray boxes represent putative coiled coil domains. Position of conserved motifs CM1 (amino acids 86–105 in the rat; amino acids 64–82 in Hydra) and CM2 (amino acids 252–257 in the rat; amino acids 230–235 in Hydra) is denoted by red boxes.

Hydra SYCP1 and SYCP3 Are Bona Fide SC Proteins.

The obtained results of a given similarity of the proteins in sequence and structure provided compelling evidence that the Hydra genome contains genes that appear to be the orthologs of mammalian SYCP1 and SYCP3. However, it remained to be proven whether they also exhibit a similar functionality by being expressed in the same cellular context and playing a similar role in Hydra meiosis. Therefore, we started a detailed expression analysis of the Hydra Sycp1 and Sycp3 genes.

To investigate the expression pattern of these genes in Hydra we performed RT-PCR on the H. vulgaris strain AEP. To this end, we isolated mRNA from head, midpiece, and foot. With the aid of sequence-specific primers, we were able to selectively amplify Hydra Sycp1 and Sycp3 cDNAs in the midpiece fraction where testes are associated with the body column (Fig. 3 A and B). Using the same assumption that SYCP1 and SYCP3 are meiotic proteins, we could amplify the message from isolated testis samples as well. Coanalysis of Hydra actin expression served as internal control for RT-PCR (Fig. 3 A and B).

Fig. 3.

Identification and characterization of Hydra SYCP1 and SYCP3. Expression pattern of SYCP1 (A) and SYCP3 (B) mRNAs as shown by RT-PCR. Expression pattern of SYCP1 (C) and SYCP3 (D) at the protein level as shown by immunoblotting. Testis-specific expression of SYCP1 (E and F, arrow) and SYCP3 (G and H, arrow) mRNA as shown by in situ hybridization on whole-mount Hydra.

In a next step, we applied Western blot analysis to monitor the expression of SYCP1 and SYCP3 proteins in the different tissue fractions. As expected, specific protein bands with the calculated molecular mass of Hydra SYCP1 and SYCP3 appeared in the midpiece and the testis tissue. We could not detect any protein signal either in the head or in the foot fraction. Again, visualization of Hydra actin was used as loading control (Fig. 3 C and D).

Once we were able to demonstrate the tissue-specific synthesis of SYCP1 and SYCP3, we now aimed to narrow down the spatial expression within the Hydra testis. Whole-mount in situ hybridization with digoxigenin (DIG)-labeled RNA probes against mRNA of Hydra Sycp1 and Sycp3 showed that both mRNAs are selectively synthesized in the basal located cells of the testis (Fig. 3 E–H). This finding fits well with the general organization of meiotic cells in Hydra testis where primary spermatocytes locate in peripheral chambers, forming a basal layer close to the mesoglea (22).

As expected, electron microscopic analysis of Hydra spermatocytes revealed the presence of SCs showing the typical fine structure (Fig. 4A). Hydra SYCP1 and SYCP3 proteins specifically localized to the spermatocytes as shown by immunofluorescence microscopy on cryosections of Hydra testes (Fig. 4B). No signals were detected in spermatogonia or spermatids (Fig. 4B). In pachytene spermatocytes, both antibodies selectively stained typical threadlike structures corresponding to the SCs (Figs. 4 and 5, Insets). As expected, the number of 15 threads in spread pachytene spermatocytes (Fig. 5) correlates with the number of synapsed homologous chromosomes in H. vulgaris (23).

Fig. 4.

Ultrastructure of a Hydra SC showing the typical components: LE, CE, and NE, nuclear envelope (A). Immunolocalization of SYCP1 and SYCP3 on cryosections of Hydra testis (B). Sg, Spermatogonia; Sp, Spermatocytes; St, Spermatids. Bar, 20 μm.

Fig. 5.

Immunolocalization of SYCP1 and SYCP3 on spread meiotic chromosomes as shown by confocal microscopy. During zygotene and pachytene, SYCP1 selectively localizes to the synapsed areas of homologous chromosomes (arrows) and SYCP3 labels the whole length of lateral elements. During diplotene, SYCP1 dissociates from the lateral elements (arrows). Image enlargements of diplotene SCs show the localization of SYCP3 in the two aligned LEs (arrows); whereas, SYCP1 is restricted to the synapsed regions in between. Bar, 10 μm.

From previous studies in mice it is known that the assembly of the SC begins with the formation of the axial elements (AEs) during leptotene by the recruitment of SYCP3 to the chromosome axis. During zygotene, the TFs consisting of SYCP1 start to assemble between the aligned AEs (now called lateral elements) thereby mediating synapsis of the homologous chromosomes. From several initiation sites synapsis finally spreads along the axis in a zipperlike manner until it is completed during pachytene. In diplotene stage, the SC disassembles and SYCP1 disappears gradually from the chromosome axis [for a review see (15)]. Consistent with these studies in mammals, in Hydra, zygotene meiocyte chromosome axes are formed by SYCP3; but synapsis is just initiated at several sites as denoted by SYCP1 labeling that apparently colocalizes with SYCP3. In pachytene, all chromosomes are fully synapsed and SYCP3 and SYCP1 show the same localization pattern. Finally during diplotene when the homologous chromosomes split, SYCP1 starts to dissociate from the axes at several sites (Fig. 5). As shown at higher magnification, SYCP3 localizes to the two parallel running axes of a bivalent and SYCP1 localizes in between (Fig. 5).

In summary, our results provide compelling evidence that Hydra SYCP1 and Hydra SYCP3 are bona fide components of the cnidarian SC.

Structural Synaptonemal Complex Proteins SYCP1 and SYCP3 Form Monophyletic Groups of Orthologous Proteins in Metazoans.

According to the results presented above, Hydra SYCP1 and SYCP3 are SC proteins representing orthologs to the mammalian SYCP1 and SYCP3. This notion received support from the calculation of evolutionary trees for metazoan SYCP1 and SYCP3 (Tables S1 and S2). As shown in Figs. 6 and 7, sequences from most multicellular clades (cnidaria, annelida, mollusca, echinodermata, and chordata) cluster together as would be expected for monophyletic groups of proteins. Beyond that, we also found homologous SYCP1 and SYCP3 sequences in the early branching (and much divergent) metazoan taxa of demosponges, placozoans, and ctenophores, which suggests that the origin of these SC proteins might even be as old as metazoan themselves [for metazoan phylogeny, see (24)].

Fig. 6.

Molecular phylogeny of SYCP1 proteins. The tree was calculated from an 83 amino acid alignment using Bayesian inference [MrBayes (36); posterior probability values are indicated]. CCDC39 sequences were used as an outgroup. Phylogenetic grouping of SYCP1 proteins was also supported using the neighbor-joining method (34).

Fig. 7.

Molecular phylogeny of SYCP3 proteins. The tree was calculated from a 171 amino acid alignment corresponding to the Cor1/Xlr/Xmr conserved region (CM1 to CM2) using the neighbor-joining method [(34); 1,000 pseudosamples; bootstrap values are indicated]. SYCP2 sequences were used as an outgroup. Phylogenetic grouping of SYCP3 proteins was also supported using Bayesian inference MrBayes (36).

Only in the clade of ecdysozoans (i.e., molting animals) hardly any sequence homologs could be identified. In the case of SYCP1 (Fig. 6) sequences of the crab appear in the cluster as sole representatives of this clade. This is an interesting point as the two current multicellular invertebrate meiosis model organisms C. elegans and D. melanogaster belong to the clade of ecdysozoans. An explanation for the strong differences in SC protein sequences between these two species and the other metazoans described here may be the evolutionary drift of ecdysozoans. This view is not without precedent in the literature as it has been already proposed for other structural proteins as is the case of lamins, the major components of the nuclear lamina (25, 26), as well as for ribosomal DNA (27). In support of this, our bioinformatic analysis with C(3)G [the TF protein of D. melanogaster (28)] and SYP1-4 [the TF proteins of C. elegans (29)] in each case revealed only protein sequences from the respective genus, underscoring the proposal of a great sequence diversity and a rapid evolution within this clade (27).

Evolutionary Conservation of the SC: More than Just the Ladderlike Organization.

So far it was assumed that only the organization of the SC, but not the protein components, would be evolutionarily conserved. However, according to our bioinformatic and expression analysis we concluded that the major structural SC protein components SYCP1 and SYCP3 form monophyletic groups of orthologous proteins across metazoans down to organisms as basal as Hydra. In other terms, evolutionary conservation of metazoan SCs would involve more than just the ladderlike organization. SYCP1 and SYCP3 sequences are consistent between quite distant metazoan clades and share a common origin more than 500 million years ago. Therefore, we can assume that there is selection pressure operating not only on the ladderlike structure of the SC but also on certain sequence motifs of single SC protein components. This casts another light not only on the topic of chromosome synapsis but also provides further support to the hypothesis of a single origin of the SC.

Finally, an interesting outcome of our work is the understanding that Hydra fulfils the major criteria to become an important model organism also in meiosis research: The genome of H. magnipapillata and the transcriptome of H. vulgaris AEP are fully sequenced, genetic manipulation methods are available, and clones of genetically manipulated animals can be easily expanded via asexual reproduction (30–32). Thus, studies comparing Hydra and mammals promise to provide deep insight into the essential features of meiosis over the last 500 million years of metazoan evolution.

Materials and Methods

Experiment details are described in SI Material and Methods.

Sequence Analysis.

Multiple amino acid sequence alignments were generated using ClustalX (33). Phylogenetic trees were constructed with the distance-based neighbor-joining method [(34); 1,000 bootstrap replicates] with Poisson correction for multiple substitutions, as implemented in Seaview (35). Bayesian inference of phylogeny was done with MrBayes v3.0b4, a program using Markov Chain Monte Carlo to approximate the posterior probability of trees (36). Four Markov chains were run from random starting trees for 10⁶ generations. The mixed model option was used to choose the appropriate fixed-rate model for amino acid substitutions. Trees were sampled every 100 generations and the first 25% of samples were discarded as burn-in. BLAST analysis was performed using sequence databases accessible from the National Center for Biotechnology Information (NCBI) server (www.ncbi.nlm.nih.gov).

Annotated sequence alignments were designed using CHROMA Version 1.0 (www.llew.org.uk/chroma/index.html) (37). Identity threshold for grouping of the residues was set to 80%. Depending on the different features of residues, the following seven groups were created: identical, charged, Ser/Thr, aliphatic, aromatic, polar, and hydrophobic.

Animal Strains and Culture Conditions.

Experiments were performed using the animal strain AEP from the H. vulgaris group (38; for sequence database see ref. 39). Animals were cultured following the standard procedures (32).

Isolation of RNA, Reverse Transcription, PCR, and Cloning of cDNA.

Total RNAs from whole animals or tissues were reversely transcribed into cDNAs. These were used for either cloning of full-length cDNA sequences or tissue-specific expression analysis.

Antibodies.

Antibodies recognizing Hydra SYCP1 were generated against a peptide (ENRTDANVLKKNSDYK), as well as the C-terminal region of Hydra SYCP1. Anti-Hydra SYCP3 antibody was generated against the peptide SDEAPALLSKSSLKRK. All antibodies were applied as affinity-purified antibodies (40).

SDS/PAGE and Immunoblot Analysis.

SDS/PAGE and immunoblot analysis (41) were carried out on protein fractions of different tissues of 50–100 animals. For protein detection the following affinity-purified antibodies were used: guinea pig anti-HySYCP1 (1:1,500), guinea-pig anti-HySYCP3, mouse anti-actin (1:10,000). Peroxidase-conjugated secondary antibodies (Dianova) were used as recommended by the manufacturer.

In Situ Hybridization.

Whole mount in situ hybridization on Hydra was performed as described in (42) with minor modifications. Digoxigenin-labeled RNA probes targeted on Hydra Sycp3 and Sycp1 were synthesized according to the manufacturer’s instructions (DIG RNA labeling Kit SP6/T7; Roche).

Immunocytochemistry.

The 6-μm cryosections of Hydra midpieces with associated testis were cut and transferred to Superfrost Plus slides (Thermo Scientific).

Preparation of Hydra meiotic chromosome spreads was performed as described by De Boer et al. (43) with minor modifications. Immunostaining of cryosections was performed following the protocol described in (44), whereas immunostaining of the chromosome spreads was carried out in accordance to De Boer et al. (43). The following affinity-purified primary antibodies were used: guinea pig anti-HySYCP1 (1:50), rabbit anti- HySYCP1 (1:400), and guinea-pig anti-HySYCP3 (1:200). Secondary antibodies (Dianova) were applied following instructions provided by the manufacturer.

Microscopy and Imaging.

Fluorescent images were taken on a confocal laser scanning microscope and processed using Adobe Photoshop (Adobe Systems). For electron microscopy, fixation and Epon embedding of Hydra pieces were done according to standard procedures (14).