Solving a Meiotic LEGO^® Puzzle: Transverse Filaments and the Assembly of the Synaptonemal Complex in Caenorhabditis elegans

Abstract

The structure of the meiosis-specific synaptonemal complex, which is perhaps the central visible characteristic of meiotic prophase, has been a matter of intense interest for decades. Although a general picture of the interactions between the transverse filament proteins that create this structure has emerged from studies in a variety of organisms, a recent analysis of synaptonemal complex structure in Caenorhabditis elegans by Schild-Prüfert et al. (2011) has provided the clearest picture of the structure of the architecture of a synaptonemal complex to date. Although the transverse filaments of the worm synaptonemal complex are assembled differently then those observed in yeast, mammalian, and Drosophila synaptonemal complexes, a comparison of the four assemblies shows that achieving the overall basic structure of the synaptonemal complex is far more crucial than conserving the structures of the individual transverse filaments.

In this commentary, R. Scott Hawley explains why the analysis of synaptonemal complex structure in Caenorhabditis elegans by Schild-Prüfert et al., published in this issue of GENETICS, provides one of the clearest pictures of the architecture of a synaptonemal complex to date.

WHEN I was in graduate school, one of my professors told me that we could really understand a biological structure only when we could describe its architecture to the degree of detail that matched the then-current structure for the bacteriophage T4 (for review see Rao and Black 2010). Those words have lingered with me these past few decades as I have watched our understanding of a meiosis-specific structure known as the synaptonemal complex become increasingly more detailed. In many ways, the synaptonemal complex is the centerpiece of early meiotic prophase—a railroad track-like structure that is almost ubiquitous among organisms. The synaptonemal complex runs between paired homologs and plays a variety of roles in the meiotic process (Page and Hawley 2004; de Boer and Heyting 2006; Kleckner 2006). While some of those roles remain controversial, it seems increasingly clear that synaptonemal complex assembly (a process referred to as synapsis) plays a critical role in mediating the intimate pairing of homologs (Hiraoka and Dernburg 2009) and in facilitating the maturation of double-strand breaks into mature reciprocal recombination events (for reviews see Hunter 2003 and Page and Hawley 2004). Other roles, such as mediating crossover interference, also have been proposed for the synaptonemal complex (for review see Youds and Boulton 2011).

But the purpose of this Commentary lies less in describing advances in understanding the synaptonemal complex’s function than it does in discussing recent progress in elucidating the roles of a class of proteins known as transverse filaments in assembling this intriguing nuclear structure in a number of organisms, but most especially in Caenorhabditis elegans. The basic techniques available to investigators to discern the blueprint of the synaptonemal complex are few, so ingenuity plays a huge role in success. Standard immunofluorescence or the use of fluorescently tagged proteins is usually sufficient only to localize proteins to the space between paired homologs and thus map them within the vicinity of the synaptonemal complex. Immunogold electron microscopy can provide a sublocalization within the synaptonemal complex, whose general structure is described below. However, characterizing the relationships (and most importantly the physical interactions between various synaptonemal complex proteins) requires techniques such as co-immunoprecipitation, co-expression in heterologous systems, and the use of the yeast two-hybrid system.

In most organisms, the synaptonemal complex is composed of two lateral elements, which are connected by transverse filament proteins (for review see Hunter 2003). The transverse filament proteins project outward from each of the lateral elements and then overlap with transverse filament proteins emanating from the other lateral elements, rather like the teeth on a zipper. The region between the lateral elements is referred to as the “central region” and the region in the middle of the central region where the transverse filaments overlap is known as the “central element” (see Figure 1A). The transverse filament proteins from most organisms have a globular N-terminal domain, which is located in the central element, a coiled-coil central domain which spans half the central region, and a globular C-terminal domain that abuts the lateral elements (Dobson et al. 1994; Liu et al. 1996; Schmekel et al. 1996; Dong and Roeder 2000; Anderson et al. 2005). The central element of the synaptonemal complex appears to be a highly ordered structure that contains pillar-like proteins that link multiple layers of central region proteins together (Schmekel and Daneholt 1995; Anderson et al. 2005) (Figure 1A).

Figure 1 .

Schematic diagrams of the organization of transverse filament proteins within the synaptonemal complex in Drosophila (A) and in C. elegans (B). (A) Adapted from Anderson et al. (2005). (B) Adapted from Schild-Prüfert et al. (2011).

As reviewed in Hunter (2003) and Page and Hawley (2004), the Zip1 protein corresponds to the yeast transverse filament (Sym et al. 1993). Like other transverse filament proteins, Zip1 contains an internal coiled-coil domain flanked by N- and C-terminal globular domains. Careful immunoelectron microscopy analysis by Dong and Roeder (2000) positioned the C terminus of Zip1 within the lateral element, while the N-terminal region localized to the center of the synaptonemal complex. They also showed that the two copies of the Zip1 protein were capable of forming a homodimer with both Zip1 components oriented in the same direction. Mutations that alter the length of the Zip1p coiled-coil domain correspondingly increase or decrease the width of the synaptonemal complex (Sym et al. 1993; Tung and Roeder 1998). These data support a structure in which the antiparallel association of Zip1p homodimers in the N-terminal region of the coiled-coil domain is essential to form transverse filaments within the synaptonemal complex. Tung and Roeder (1998) also demonstrated that removal of the C-terminal region of Zip1 prevents its binding to the chromosomes, but still allows Zip1 proteins to assemble into highly ordered aggregates known as polycomplexes.

The first insights into the molecular organization of the mammalian synaptonemal complex came from the seminal work on the mammalian transverse filament protein SCP1 (also known as SYCP1) by Heyting and her collaborators (Meuwissen et al. 1992). Structurally similar to Zip1, SCP1 contains a central coiled-coil domain flanked by globular domains at the N terminus. On the basis of immunogold localization these authors correctly inferred that the C terminus of SCP1 is associated with the lateral element and that the remainder of the protein projects out into the central region of the complex. These observations were greatly extended by Liu et al. (1996), who went on to show that the N terminus of SCP1 was indeed localized within the central element and confirmed the localization of the C terminus to the vicinity of the lateral element. As pointed out by the authors, these observations support “the notion that SCP1 is an extended filamentous protein and that the two molecules of the putative SCP1 dimer are likely to have the same polarity,” allowing the authors to suggest a model in which SCP1 homodimers, each projecting their N termini into the region of the central element, can effectively interlock the two lateral elements by physical contacts between those N-terminal regions (Liu et al. 1996). In support of this idea, the authors used the yeast two-hybrid system to demonstrate the ability of the N-terminal region of SCP1 to interact with itself, but not with other domains of the protein, leading Liu et al. to suggest “that a transversal filament consists of one or more pairs of SCP1 dimers, each pair being organized in a head-to-head arrangement with the C-termini anchored in the lateral elements and the two N-termini being joined in the central element” (Liu et al. 1996). Similar conclusions were reached for the rat SCP1 protein by Schmekel et al. (1996).

Ollinger et al. (2005) went on to show that SCP1 is capable of self-assembling into large arrays of synaptonemal complex-like structures in the cytoplasm, known as polycomplexes, when expressed in mitotic cell culture. Zickler and Kleckner (1999) noted that “the existence of polycomplexes is generally considered to be a manifestation of the intrinsic capacity of synaptonemal complex components to self-assemble.” The polycomplexes observed by Ollinger et al. formed in the absence of chromatin, and changes in the length of the coiled-coil domain altered the width of the individual synaptonemal complex-like elements of these polycomplexes. Ollinger et al. also explored the effects of deletions of the N- and C-terminal globular domains on the formation of these polycomplexes and observed that, although deletion of the N terminus reduces the efficiency of polycomplex formation, it does not prevent the appearance of synaptonemal complex polycomplex structures with apparently normal structures. On the other hand, deletion of the C terminus fully ablates the ability of SCP1 to form polycomplexes, suggesting a critical role of the C terminus in the self-assembly of transverse filaments. But perhaps most important, these data argue that, at least in the mouse, transverse filaments are fully sufficient to assemble into synaptonemal complex-like structures even in the absence of other meiosis-specific accessory proteins.

In addition to SCP1, the identification of a number of other components of the mouse has also helped to flesh-out the mechanisms by which the SCP1 transverse filament proteins are integrated into the structure of the synaptonemal complex. For example, Yuan et al. (2000) identified a component of the lateral element of the murine synaptonemal complex known as SCP3, showing that a mutaton in its gene prevents proper localization of SCP1. In addition, Costa et al. (2005) have identified two central element proteins whose recruitment to the synaptonemal complex is dependent on SCP1, both in meiocytes and in the context of the heterologous expression system described by Ollinger et al. (2005).

The analysis of transverse filament structure and function in Drosophila has been similar in trajectory and in the outcome of the story described above for the mouse SCP1 protein. The structure of the Drosophila synaptonemal complex was elucidated in four distinct phases. First, Carpenter described the basic (and canonical) structure of the synaptonemal complex by serial-section electron microscopy in a series of articles in the mid and late 1970s (Carpenter 1975a,b; Carpenter 1979a,b). Next, on the basis of the molecular characterization of the c(3)G gene, whose absence ablates synaptonemal complex formation, Page and Hawley (2001) described a protein [C(3)G] that is indeed the Drosophila transverse filament and whose absence eliminates synapsis and recombination. Third, studies by Bickel and McKim and their collaborators identified two genes, ord (Bickel et al. 1996) and c(2)M (Manheim and McKim 2003), whose protein products were likely candidates for lateral element components, and Page et al. (2008) identified a gene (cona) whose protein product would subsequently be shown to be a central element protein (see below). Anderson et al. (2005) used immunogold electron microscopy to position C(3)G and C(2)M proteins within the synaptonemal complex, demonstrating that both the N terminus C(2)M and the C terminus of C(3)G localize at or near the lateral element while the N terminus of C(3)G localized within the central element (Figure 1A). Finally, Lake and Hawley (2011) showed that Cona did indeed localize to the edges of the central element, suggesting that it is analogous to a ‘pillar protein” (Figure 1A).

The significance of these positional relationships was to a large degree elucidated by the studies of Jeffress et al. (2007) and Page et al. (2008). Using transgene constructs, Jeffress et al. (2007) demonstrated that the C-terminal globular domain of C(3)G is required to link the transverse filaments to the lateral elements (or chromosome axis), but not to form the synaptonemal complex structure itself. Oocytes expressing only such terminal deletions of C(3)G form elaborate polycomplexes whose structure (as assayed by electron microscopy) mimics that of the cylindrical columns of synaptonemal complexes. The formation of such polycomplexes is dependent on the Cona protein (Page et al. 2008). Constructs missing the N-terminal region of C(3)G failed to form polycomplexes entirely, but formed small foci at various places in the nucleus, suggesting that interactions between N-terminal regions are necessary for synaptonemal complex formation. These results suggest a different set of rules for the assembly into polycomplexes of the C(3)G transverse filament for flies than was observed for SCP1 in the mouse by Ollinger et al. (2005). First, the C terminus of C(3)G is not required for transverse filament assembly into polycomplexes in Drosophila (contrary to the results obtained for SCP1; see above). Second, and again unlike the data for SCP1, the N terminus for C(3)G is absolutely required for polycomplex formation. Finally, the assembly of C(3)G into polycomplexes requires a meiosis-specific accessory protein (Cona). Thus, despite their similarity in overall structure, the rules that govern the interaction of transverse filament proteins appear to be very different in mammals than in flies.

But perhaps the most unusual set of instructions for transverse filament assembly may be found in C. elegans. In comparison to the structural studies described above for yeast, mice, and Drosophila, the architectural analysis of the C. elegans synaptonemal complex has been more complex. A combination of genetic and yeast two-hybrid studies from the Villeneuve and Colaiácovo labs has identified four proteins—SYP-1 (MacQueen et al. 2002), SYP-2 (Colaiácovo et al. 2003), SYP-3 (Smolikov et al. 2007), and SYP-4 (Smolikov et al. 2009)—that localize to the space between paired chromosomes and thus are candidate synaptonemal complex components. Curiously, genetic studies showed that the proper localization of each of the four proteins is dependent on the presence of all three other proteins, suggesting that they interact within a common structure. Indeed, immunogold electron microscopy analysis by Schild-Prüfert et al. (2011) demonstrates that all four SYP proteins are indeed components of the C. elegans synaptonemal complex.

Prior to the Schild-Prüfert et al. (2011) study, the only known physical interaction between these proteins was between SYP-3 and SYP-4. Indeed, SYP-4 had been initially identified on the basis of its ability to interact with SYP-3 in a yeast two-hybrid study (Smolikov et al. 2009). In addition, the assembly of SYP-1 and SYP-2 into a structural module had been inferred on the basis of the predicted sizes of these proteins (Colaiácovo et al. 2003); both consist of coiled-coil domains flanked by N and C globular domains. As pointed out by Hunter (2003), such a proposed interaction might solve the dilemma that the transverse filament proteins in most other organisms were predicted to be ∼70–90 nm in length, long enough that head–head overlaps of the N-terminal regions of these proteins could create a structure spanning the ∼100-nm distance that separates the lateral elements in most organisms, including C. elegans. However, SYP-1 was predicted to have a length of only 35 nm (for discussion see Hunter 2003).

With only these data, one is left with a bit of a LEGO puzzle. With four differently colored blocks corresponding to the four SYP proteins, one must deduce the rules that allow their assembly into a model of a synaptonemal complex without the usual instruction pamphlet. In Schild-Prüfert et al. (2011), the authors primarily used three techniques to deduce which proteins interact with the other three proteins within the synaptonemal complex: co-immunoprecipitation, a fascinating variant of yeast two-hybrid analysis, and immunogold localization via electron microscopy. The yeast two-hybrid studies are based on an innovative strategy described in Boxem et al. (2008) in which full-length or domain-specific constructs are tested for interactions against what is essentially a protein “fragment library.” This method appears to greatly increase the sensitivity of the yeast two-hybrid assay.

The results of all of these studies are summarized in Figure 1B, a drawing based on the major conclusions of Boxem et al.’s work (as summarized in their figure 5). First, these authors showed that only SYP-1 can homodimerize (an interaction that strictly requires only the coiled-coil domains but is strengthened by the presence of the C-terminal globular domains). That said, the coiled-coil domain of SYP-1 can interact with SYP-2 as demonstrated by co-immunoprecipitation and the yeast two-hybrid assay. Surprisingly, further yeast two-hybrid studies demonstrated that it is the C terminus of SYP-2 that interacts with the coiled-coil domain of SYP-1.

But perhaps the most interesting of the findings regarding SYP-1 was the ability of its C terminus to interact with the N terminus of SYP-3. As the authors point out, this is a particularly important result because it links the SYP-1/SYP-2 and SYP-3/SYP-4 modules (SYP-2 and SYP-3 do not appear to interact directly). Finally, the authors verified that the SYP-3::SYP-4 interaction requires the N terminus of SYP-3 and apparently the full length of SYP-4.

All of this tells us only which colors of the LEGOs physically interact. It does not give us positional information regarding how those blocks will fit within our model of the synaptonemal complex. That requires the use of domain-specific antibodies to facilitate the localization within the synaptonemal complex by immunogold electron microscopy. These studies clearly demonstrated that the C terminus of SYP-3 abuts the chromosome axes, while N termini of SYP-1 and SYP-4 and the C terminus of SYP-2 are located in the middle of the synaptonemal complex. Combining all of these data, we arrive at the model diagrammed in Figure 1B. Interestingly, the ability to span the ∼100-nm distance between the lateral elements appears to lie not in the ability of SYP-1 to interact with SYP-2, but in the interaction SYP-1 with SYP-3.

The beauty in the structure of the C. elegans synaptonemal complex lies both in the manner of its elucidation by Schild-Prüfert et al. (2011) and in the quite unusual combination of at least four proteins (SYP-1, SYP-2, SYP-3, and SYP-4), which together appear to mimic the function of the single transverse filament proteins [Zip1, SCP1, and C(3)G)] described above. There may be a reason that the discovery of the synaptonemal complex did not come with an instruction manual to describe its assembly—there are apparently a great many ways to solve the puzzle of building this amazing structure.