The C. elegans dense body: anchoring and signaling structure of the muscle
- First Online:
- Cite this article as:
- Lecroisey, C., Ségalat, L. & Gieseler, K. J Muscle Res Cell Motil (2007) 28: 79. doi:10.1007/s10974-007-9104-y
- 238 Views
During evolution, both the architecture and the cellular physiology of muscles have been remarkably maintained. Striated muscles of invertebrates, although less complex, strongly resemble vertebrate skeletal muscles. In particular, the basic contractile unit called the sarcomere is almost identical between vertebrates and invertebrates. In vertebrate muscles, sarcomeric actin filaments are anchored to attachment points called Z-disks, which are linked to the extra-cellular matrix (ECM) by a muscle specific focal adhesion site called the costamere. In this review, we focus on the dense body of the animal model Caenorhabditis elegans. The C. elegans dense body is a structure that performs two in one roles at the same time, that of the Z-disk and of the costamere. The dense body is anchored in the muscle membrane and provides rigidity to the muscle by mechanically linking actin filaments to the ECM. In the last few years, it has become increasingly evident that, in addition to its structural role, the dense body also performs a signaling function in muscle cells. In this paper, we review recent advances in the understanding of the C. elegans dense body composition and function.
KeywordsStriated muscleCaenorhabditis elegansDense bodyZ-diskCostamereFocal adhesion sitesIntegrin signalingTitinMuscle LIM proteins
Caenorhabditis elegans is an invertebrate belonging to the nematode phylum (Eukaryota; Metazoa; Eumetazoa; Bilateria; Pseudocoelomata; Nematoda; Chromadorea; Rhabditida; Rhabditoidea; Rhabditidae; Peloderinae; Caenorhabditis). In spite of the phylogenetic distance between nematodes and vertebrates, genome analysis has shown that more than 50% of human genes have C. elegans counterparts (C. elegans Sequencing Consortium 1998). In addition to this high conservation of genes some of the C. elegans organs, most notably muscles, have a cellular physiology similar to that of vertebrates.
The overall structure and composition of the muscle basic functional unit, named sarcomere is very similar between vertebrates and invertebrates. The sarcomere is highly conserved between nematodes and vertebrates. As in vertebrates, each sarcomere is composed of myosin containing thick filaments associated at the M line, and actin containing thin filaments anchored to the Z-disk analogue, called the dense body. Body-wall muscles are striated due to the succession of sarcomeres. The striation observed by microscopy corresponds to the repetitive alignment of myosin-enriched A bands and actin-enriched I bands. In vertebrate skeletal muscles adjacent sarcomeres are aligned so that the striation appears perpendicularly orientated with respect to the length of the muscle fibre. In nematode body-wall muscles, adjacent sarcomeres are staggered and thus, the striation appears nearly parallel to the long axis of the muscle cell with which it forms an angle of 5.9° (Mackenzie and Epstein 1980) (Francis and Waterston 1985) (Fig. 1B and D).
Sarcomeres are formed at mid-embryogenesis during the two-fold embryonic stage. At hatching, the body-wall muscles are composed of 85 cells (Krause 1995). These cells are 2 sarcomeres wide with 5 μm long filaments. As development proceeds, new sarcomeres are added by a still unknown mechanism, resulting in cells which are 10 sarcomeres wide with 10 μm long filaments (Moerman and Williams 2006). Similarly, the A band grows from 100 filaments at the L1 stage to 600 filaments in the adult (Mackenzie et al. 1978). M lines and dense bodies also increase in size during this process.
The layer of contractile apparatus, which is approximately 1.5 μm thick, is parallel to the hypodermis and the cuticle. Dense bodies and M-lines, which are perpendicularly orientated go through the myofilament layer and anchor respectively actin and myosin filaments to the muscle cell membrane. Cellular organelles (nucleus, mitochondria...) are localized in the apical part of the muscle cells just above the contractile apparatus (Fig. 1).
The primary function of the dense body: anchoring sarcomeric actin filaments and transducing force
The essential role of muscles is to generate force, and thereby movement. From a purely mechanical point of view, muscles must carry out two functions: force generation and force transduction outside the muscle. Although other functions have been added through evolution, these remain the core ones. This strong conservation of function accounts for a strong conservation of the components and mechanisms. Force is generated by the sarcomere, by the ATP-dependent sliding of myosin heads on actin filaments.
Force transduction requires rigid structures, otherwise the movement created by the contractile machinery would be annihilated. In vertebrates, this problem is resolved at two levels: (1) by the anchoring of sarcomeres to the sarcolemma by specific muscle adhesion complexes, called costameres, which link Z-disks to the extra-cellular matrix (ECM) (Ervasti 2003; Samarel 2005), and (2) by the anchoring of the muscles ends to the bones by tendons.
In C. elegans, attachment of the sarcomeres to the muscle membrane and the underlying basement membrane is performed by the dense body, a protein complex which shares functional similarity with both the vertebrate Z-disk and the costamere. In addition, the muscles are linked via fibrous organelles to the underlying cuticle, the exoskeleton of the nematode (Cox and Hardin 2004).
Molecular composition of the dense body
Among the structural proteins, the vinculin homolog DEB-1 was the first dense body protein to be identified in C. elegans (Barstead and Waterston 1989). DEB-1 plays a crucial role in dense body assembly and requires integrin and probably talin (see below) for its localization (Hresko et al. 1994). Deb-1 null mutants are paralyzed at early embryonic stages and show disorganized muscles (Pat phenotype) (Barstead et al. 1991).
The C. elegans talin homolog (Moulder et al. 1996) is probably also implicated in the early steps of dense body and sarcomere assembly. Talin is a FERM (for protein 4.1, ezrin/radixin/moesin) domain containing protein. FERM domains contain binding sites for cytoplasmic tails of transmembrane proteins (Chishti et al. 1998). Like its vertebrate counterpart at focal adhesions the C. elegans talin is thought to bind to the PAT-3/ß-integrin cytoplasmic tail and might also interact with DEB-1/vinculin as well as with actin filaments (Horwitz et al. 1986; Calderwood et al. 1999; Cram et al. 2003). It has been suggested (Moulder et al. 1996), that talin acts upstream of vinculin and that vinculin is recruited to the integrin/talin complex at the forming dense body. Post-embryonic RNAi-mediated talin inactivation leads to a phenotype of uncoordination and paralysis because the muscles become disrupted. Talin null mutations are supposed to lead to an early arrest of embryonic development, but talin loss-of function mutants are not yet available (Cram et al. 2003).
Another component of the dense body is ATN-1, the C. elegans homolog of α-actinin (Francis and Waterston 1985; Barstead et al. 1991). ATN-1/α-actinin belongs to the spectrin super-family of proteins and contains three functional domains: a N-terminal actin binding domain, a central rod domain containing four spectrin-like domains and a C-terminal domain containing EF-hand calcium-binding motifs. However, unlike vertebrate α -actinin, the C-terminal end of ATN-1 is probably unable to bind calcium since it lacks the full calcium coordinating amino acids (Barstead et al. 1991). Surprisingly, atn-1 null mutants (genotype atn-1(ok84)) have nearly normal motility and muscle structure (by polarized light microscopy). However, extra accumulation of actin near the ends of muscle cells is observed and by electron microscopy, dense bodies appear shorter and larger than in wild type (Flaherty et al. 2002). This weak mutant phenotype suggests that ATN-1 has a limited function during embryonic sarcomere and dense body assembly in C. elegans or that it shares functional redundancy with other genes. However ATN-1 might play a post-embryonic physiological role by contributing to long-term sarcomeric stability.
UNC-112/MIG-2 (Rogalski et al. 2000), PAT-4/ILK (integrin-linked kinase) (Mackinnon et al. 2002), PAT-6/actopaxin (Lin et al. 2003) and UNC-97/PINCH (Hobert et al. 1999) are required downstream of integrin for proper assembly of dense bodies and sarcomeres during early embryogenesis. Their null mutations lead to the Pat phenotype.
In C. elegans, UNC-112, PAT-4, PAT-6 and UNC-97 form a complex in which UNC-112 and PAT-6 bind to the N-terminal part of PAT-4, whereas UNC-97 binds to the carboxy portion of PAT-4 (Fig. 2). The recruitment of this complex to the dense body is independent of DEB-1, suggesting that two independent sub-complexes are formed, one comprising UNC-112, PAT-4, PAT-6 and UNC-97, and the other one containing DEB-1, talin and ATN-1 (Labouesse and Georges-Labouesse 2003; Moerman and Williams 2006) (Fig. 2).
Numerous other cytoplasmic proteins are known to be associated or to co-localize with the dense body (see below). Interestingly, expression of all known dense body-associated proteins is maintained in muscles throughout the post-embryonic and adult life, suggesting a continuous requirement of these proteins for the structural integrity and function of muscles during growth and adulthood.
The dense body: a crossroad of signaling pathways
In addition to their role in attachment and mechano-transduction, vertebrate Z disks and costameres (focal adhesions) are platforms serving as anchoring sites for signaling/adapter proteins (Pyle and Solaro 2004; Samarel 2005). As the veil is uncovered on the C. elegans dense body, the same picture emerges. Like their mammalian counterparts, dense bodies probably receive extra-cellular signals, which are transduced into biochemical signals leading to sarcomeric assembly, modified gene expression and dense body stability.
1° integrin signaling
In the C. elegans body-wall muscle, the communication between the ECM and the contractile apparatus across the sarcolemma is mediated by PAT-2/α-integrin–PAT-3/ß-integrin heterodimers. In the whole animal kingdom, integrins are receptors for cell adhesion but are also implicated in many signal transduction pathways leading to various cellular responses such as proliferation, survival, gene expression control and cytoskeleton organization (Hynes 2002). However, integrins possess no catalytic activity, meaning that signal transmission necessarily implicates integrin-associated cytoplasmic proteins.
The C. elegans PAT-2/α-integrin–PAT-3/ß-integrin heterodimers form the foundation of dense bodies. The extra-cellular parts of PAT-2 and PAT-3 are thought to interact directly or indirectly with EMC proteins such as perlecan (UNC-52), while the intracellular domain of PAT-3 interacts with many cytoplasmic proteins, several of them are homologs of vertebrate proteins involved in integrin signaling (Fig. 2). Analysis of PAT-3 function in C. elegans shows that integrin heterodimers are required not only to anchor the nascent actin fibres to the adjacent basement membranes during embryogenesis, but also to maintain the sarcomere integrity in larvae and adults (Lee et al. 2001).
Among the proteins implicated in integrin function in vertebrates, talin has been shown to bind to ß-integrin (Horwitz et al. 1986). The interaction between talin and ß-integrin triggers integrin activation and ligand binding, perhaps by relieving an inhibitory interaction between α- and ß-integrin subunits (Calderwood et al. 2002, 1999). In addition, talin itself possesses binding sites for signaling molecules like focal adhesion kinase or phosphatidylinositol phosphate kinase type 1gamma indicating a possible role in signal transduction (Chen et al. 1995; Barsukov et al. 2003).
The sequence of the C. elegans talin is very similar to the mouse talin homolog (Moulder et al. 1996). C. elegans talin is thought to bind to PAT-3/ß-integrin cytoplasmic tail, similarly to its vertebrate homolog (Moerman and Williams 2006). Post-embryonic RNAi-mediated reduction of talin in C. elegans results in perturbation of muscle filament structure, animals becoming progressively paralysed. This phenotype is thought to be mediated by an alteration of integrin activity in the absence of talin (Cram et al. 2003).
In vertebrates, the focal adhesion protein integrin-linked kinase (ILK) also plays an important role in integrin signaling. ILK is an ankyrin repeat containing serine/threonin protein kinase, which interacts with the cytoplasmic domains of integrin ß1 and ß3 subunits (Hannigan et al. 1996). ILK mediates the phosphorylation of a variety of intracellular substrates, such as protein kinase B (PKB/AKT) and glycogen synthase kinase-3 (GSK-3) (Delcommenne et al. 1998).
The C. elegans homolog of ILK, PAT-4, has been proposed to function as an adaptor protein rather than as a signaling protein during integrin-dependent dense body assembly, because its kinase activity is dispensable for this process (Mackinnon et al. 2002). However, the functionality of the PAT-4 kinase domain has not been studied so far, nor has its eventual post-embryonic requirement. In accordance with a role as an adaptor, PAT-4 interacts with at least three dense body-associated proteins: UNC-112/MIG-2, PAT-6/actopaxin and UNC-97/PINCH. However, unlike its vertebrate counterpart, PAT-4 does not interact directly with the cytoplasmic tail of PAT-3/ß-integrin (Mackinnon et al. 2002). The PAT-4/UNC-112/PAT-6/UNC-97 complex might indeed be linked to PAT-3 through UNC-112 (Moerman and Williams 2006).
UNC-112 (Rogalski et al. 2000) is the homolog of the human MIG-2 product (Mitogen-Induced Gene-2, Wick et al. 1994). It possesses a region showing homology to the FERM domain of talin and other members of the FERM protein family, suggesting that the corresponding region of UNC-112 might be implicated in its interaction with PAT-3/ß-integrin. It has been proposed that UNC-112 acts as a scaffold protein linking the PAT-3/ß-integrin pathway to CDC-42 signaling in C. elegans (Fig. 2). Actually, UNC-112 interacts with UIG-1 (for UNC-112 Interacting GEF-1), which shows CDC-42-specific GEF (guanine nucleotide exchange factor) activity in vitro (Hikita et al. 2005). Interestingly, vertebrate integrin signaling has been shown to regulate the activity of Rho-family GTPases like Rac and CDC-42, which are implicated in cytoskeletal rearrangement and cell adhesion in response to extra-cellular signals (Price et al. 1998; Ren et al. 1999).
2° titin mediated signaling
Another protein that might be implicated in signaling transduction at the dense body is the C. elegans titin homolog. Titin belongs to the family of immunoglobulin (Ig) domain-containing giant elastic muscle proteins. In mammals, titin has a molecular mass of 3MDa, is 1.2 μm long and spans half a sarcomere. The C. elegans titin has two isoforms, one of 1.2 MDa, which is 0.5 μm long, and another of 2.2 MDa which is 0.9 μm long. The largest 2.2 MDa isoform contains 56 Ig domains and a single protein kinase domain (Flaherty et al. 2002). C. elegans titin might span an entire half I-band and extend into the edge of the A-band. The large titin isoforms are anchored to the dense body by an ATN-1/α-actinin-dependent mechanism (Fig. 2). In the absence of ATN-1, titin is not properly anchored (Flaherty et al. 2002). In vertebrates, the titin kinase domain has been implicated in control of gene expression and protein turnover (Lange et al. 2005). The C. elegans titin protein has kinase activity in vitro, but its in vivo substrates remain unknown (Flaherty et al. 2002).
3° the mysterious role of muscle LIM proteins
LIM proteins (named from the Lin-11, Isl-1 and Mec-3 proteins) are a large class of proteins carrying LIM domains, a motif composed of two Cys/His zinc finger motifs, (consensus CX2CX16-23HX2CX2CX2CX16-23CX2-3(C,H,D)) that mediate specific protein-protein interactions (Schmeichel and Beckerle 1994), but may also have the capacity to bind to nucleic acids (Schmeichel and Beckerle 1997). Some proteins of this family possess exclusively LIM domains (LIM-only proteins) (ex. zyxin and paxillin), while others carry LIM domains in combination with other functional domains (LIM-plus proteins) (Hobert et al. 1999). LIM proteins are involved in various pathways controlling cell growth, differentiation and cell attachment (Kadrmas and Beckerle 2004; Labouesse and Georges-Labouesse 2003). To date, four LIM proteins are known to be expressed in the C. elegans body-wall muscle and to localize to the dense body: ALP-1/ALP-Enigma, UNC-97/PINCH, UNC-95/paxillin and ZYX-1/zyxin (Fig. 2). Their respective vertebrate counterparts are localized at adhesion complexes.
ALP-1 is the single C. elegans homolog of vertebrate ALP/Enigma family (McKeown et al. 2006). In mammals, the ALP/Enigma family of LIM-plus proteins consist of at least 7 members, four α-actinin-associated LIM proteins (ALP) and three enigma proteins. These proteins carry an amino-terminal PDZ domain in addition to one (ALP-subfamily) or three (Enigma-subfamily) LIM domains (Kadrmas and Beckerle 2004). ALP/Enigma proteins localize and function at sites of actin anchorage in muscle and non-muscle cells where many of them interact with α-actinin. Lack of function of ALP/Enigma proteins has been shown to lead to defects in striated muscles and to cardiomyopathies, in mice as well as human patients (Pashmforoush et al. 2001; Zhou et al. 2001; Vatta et al. 2003). The alp-1 gene of C. elegans encodes four isoforms which are all expressed in body-wall muscles, where they localize to actin attachment sites, notably dense bodies. The PDZ domain is sufficient to localize the protein to the dense body, as shown in transgenic animals (McKeown et al. 2006).
The UNC-97 protein belongs to the PINCH family, which are LIM-only proteins possessing five tandem LIM domains. Mammal PINCH proteins interact with ILK and the adapter protein NCK-2 (Wu and Dedhar 2001) and localize to focal adhesions (Zhang et al. 2002). The C. elegans UNC-97/PINCH protein co-localizes with ß-integrin to dense bodies. UNC-97 has been shown to bind to the C-terminal part of PAT-4/ILK (Moerman and Williams 2006). UNC-97 null mutants exhibit a severe embryonic Pat phenotype suggesting that like UNC-112, PAT-4 and PAT-6, it is involved in early sarcomeric assembly (Hobert et al. 1999).
The UNC-95 is a LIM-only protein with a single LIM domain at the C-terminus. Its LIM domain shows strong similarity to the LIM domain of the vertebrate focal adhesion protein paxillin (Broday et al. 2004). UNC-95 localizes to dense bodies and is required for dense body assembly and integrity. Its null mutation affects myofilament organization and leads to paralysis (Zengel and Epstein 1980). Interestingly, UNC-95 interacts and co-localizes at dense bodies with the ring finger protein RNF-5, a protein with ubiquitin protein ligase (E3) activity (Fig. 2). In human, over-expression of RNF-5 causes exclusion of paxillin from focal adhesions, indicating that RNF-5 regulates the cellular localization of paxillin via its E3 ligase activity (Didier et al. 2003). In C. elegans, RNF-5 could regulate the localization of UNC-95 within the body-wall muscle cells (Broday et al. 2004). Further, via its interaction with RNF-5, UNC-95 might participate in the control of protein turnover at the dense body.
A fourth C. elegans LIM protein to be expressed in C. elegans body-wall muscle, is ZYX-1 (our unpublished results), the homolog to the vertebrate zyxin protein (Smith et al. 2002). ZYX-1 as well as zyxin are LIM-only proteins containing three LIM domains. In vertebrates, the LIM-only protein zyxin is present at sites of cell adhesion (Crawford and Beckerle 1991). It interacts with α-actinin and other cytoskeletal proteins (Crawford et al. 1992), as well as with signaling proteins such as the product of the proto-oncogene Vav (Hobert et al. 1996). Therefore, zyxin has been proposed to function in cytoskeletal dynamics and signaling. The zyxin homolog of C. elegans ZYX-1 localizes to the dense body, where it is probably anchored by ATN-1/α-actinin since in yeast two hybrid assays ZYX-1 and ATN-1 strongly interact (KG unpublished results) (Fig. 2).
What is the function of muscle LIM proteins in C. elegans? The four known muscle LIM proteins of C. elegans ALP, UNC-97, UNC-95, ZYX-1 are localized to the dense body, where they may function as adaptor proteins allowing the recruitment of various structural and signaling molecules. Interestingly, all four LIM proteins have also been detected in the nucleus of body-wall muscle cells (Fig. 2). This dual localization suggests that these proteins shuttle between the dense body and the nucleus and might therefore link structural and signaling processes. Shuttling between nucleus and sites of cell adhesion has been demonstrated for vertebrate LIM domain proteins, such as zyxin and paxillin (Nix and Beckerle 1997; Woods et al. 2002). These proteins are supposed to regulate gene transcription by interaction with transcription factors (Wang and Gilmore 2003). However, shuttling between dense bodies and nucleus has not yet been demonstrated in C. elegans.
An additional protein, UNC-98 has several properties in common with the four muscle LIM proteins (Mercer et al. 2003): (1) it is expressed in body-wall muscle cells, (2) localized at dense bodies and in the nucleus, (3) possesses zinc finger motifs (the sequence of UNC-98 reveals the presence of four C2H2 motifs). In addition, UNC-98 has been shown to interact with UNC-97/PINCH (Fig. 2). With respect to the possible function of UNC-98, it has been proposed, that this protein might be involved in maintaining muscle structure and mediate, like muscle LIM proteins, communication between dense bodies and the nucleus (Mercer et al. 2003).
The dense body is a cellular structure, which may be looked at from various angles. Its primary role is to provide some rigidity to the muscle in order to transduce force outside of the muscle. Without this rigid element, the force generated by the sliding of myosin on actin filaments would be lost. But beyond this basic mechanical function of force transduction (and probably also of force sensing), the dense body increasingly appears as a crossroad where numerous proteins interact. Among those proteins are the enigmatic LIM proteins, found both at the dense body and in the nucleus. In mammals, striated muscles respond to prolonged exercise by recruiting more cells and addition of new sarcomeres. It is therefore not surprising to find messenger proteins travelling from a force-sensing body (the Z-disk) to the nucleus to instruct it about the status of the contractile machinery. In C. elegans, the situation is somewhat different since the muscles are post-mitotic, and up to now, there is no evidence for changes in the number of sarcomeres with respect to the animal activity. Then, what may the LIM proteins and other shuttling proteins inform the nucleus about?
This question is still open. Thanks to the simplicity and the genetic amenability of the model, the C. elegans dense body makes a remarkable entry point for understanding muscle physiology, a physiology that appears to be more and more sophisticated. Given the high degree of conservation with mammals, it would be surprising if progress made on this topic would not, in the end, benefit to muscle research in general.
The authors thank John C. Sparrow, Renaud Legouis and Joëlle Thomas for critical reading this manuscript and Yannick Schwab for technical assistance concerning electron microscopy.