A conserved LIM protein that affects muscular adherens junction integrity and mechanosensory function in Caenorhabditis elegans

Abstract

We describe here the molecular and functional characterization of the Caenorhabditis elegans unc-97 gene, whose gene product constitutes a novel component of muscular adherens junctions. UNC-97 and homologues from several other species define the PINCH family, a family of LIM proteins whose modular composition of five LIM domains implicates them as potential adapter molecules. unc-97 expression is restricted to tissue types that attach to the hypodermis, specifically body wall muscles, vulval muscles, and mechanosensory neurons. In body wall muscles, the UNC-97 protein colocalizes with the beta-integrin PAT-3 to the focal adhesion-like attachment sites of muscles. Partial and complete loss-of-function studies demonstrate that UNC-97 affects the structural integrity of the integrin containing muscle adherens junctions and contributes to the mechanosensory functions of touch neurons. The expression of a Drosophila homologue of unc-97 in two integrin containing cell types, muscles, and muscle-attached epidermal cells, suggests that unc-97 function in adherens junction assembly and stability has been conserved across phylogeny. In addition to its localization to adherens junctions UNC-97 can also be detected in the nucleus, suggesting multiple functions for this LIM domain protein.

Figure 1

Schematic view of nematode muscle organization. (A) Schematic view of C. elegans body wall muscles. (B) Cross-section showing muscle quadrants. (C) Lateral view onto muscle quadrants. White dots, focal-adhesion like muscle attachment sites (dense bodies); these represent a top view of the black lines in B. (D) Muscle cell. (E) Structural composition of sarcomeres (actin-myosin based contractile units) and their anchorage to the hypodermis. See text for references.

Figure 4

UNC-97 is expressed in muscles and neurons. (A) Schematic representation of reporter gene constructs used. Both constructs contain regulatory regions up to the mec-2 gene (last exon is shown), which precedes unc-97 and reveal a similar expression pattern. The unc-97-prom::GFP reporter gene constructs (shown below in B–F) uniformly labels the whole cell, whereas the unc-97::GFP fusion gene product reflects the subcellular localization of the UNC-97 protein shown in Figs. 6–8. (B) Expression of the unc-97-prom::GFP reporter gene in embryos at the comma stage (∼400 min). (C) Expression of unc-97-prom:: GFP in L1 larvae muscle cells. (D) Expression in vulval muscles (arrow). (E and F) Expression in the touch neurons PLM (E) and ALM (F). Bottom arrows, axonal processes; top arrows, cell bodies. Expression in the other touch neurons PVM and AVM is as strong but not shown here. The expression pattern described can be observed with an chromosomally integrated unc-97-prom:: GFP reporter gene.

Figure 6

Subcellular localization of UNC-97. (A) Top, UNC-97::GFP localization to dense bodies (DB) and M lines; middle, PAT-3 localization visualized with the anti-MH25 antibody; bottom, overlay of the first and second panel showing the overlap between green (UNC-97::GFP) and red (PAT-3) in yellow. (B) Top two panels show lateral Nomarski images of wild-type and unc-97(su110) animals, the latter animal displaying a protruding vulva phenotype (p-vul). The p-vul phenotype displays a 56% penetrance (n = 32). Black arrows in top panel, approximate attachment sites of the vulval muscles. The shape of the vulval muscles can also be seen in Fig. 4 D. Bottom panel, ventral view of UNC-97::GFP localization in wild-type adult animals. UNC-97::GFP localizes to the attachment sites (white arrows) of the vulval muscles to the hypodermis (for a schematic drawing of the vulval muscles see White, 1988). (C) Top, UNC-97:: GFP localization in the muscle sarcomere. Thin stripes, M line structures; dots, dense body structures (for a schematic description of these structures see Fig. 1). In the center of the panel, fuzzy nuclear staining can be observed which is out of focus. Second panel, the same cell as in the first panel, but in a different plane of focus showing nuclear localization of UNC-97 in dots. Third panel, DAPI staining of the same nucleus. Bottom, overlay between the second and third panel. The animal shown in this series of micrographs has been fixed with formaldehyde to stabilize sarcomeres and nuclear localization of UNC-97.

Figure 7

unc-97 functions in muscle development. (A–C): UNC-97::GFP localization in live worms at embryonic stage ∼300 min (A), ∼400 min (comma stage) (B), and ∼500 min (threefold “pretzel” stage) (C). (A′–C′) Corresponding DIC micrographs. White arrows in C, a muscle nucleus and dense bodies, respectively. Note UNC-97::GFP localization to the periphery of nuclei in A–C. (D–F): Loss-of-function phenotype of unc-97 as determined by RNAi. D and E show the progeny of animals subjected to unc-97 RNAi. Embryos either arrest and die at the twofold stage or die shortly after hatching (both categories were scored as arrested). Note that the pharynx develops relatively normally but that elongation of the rest of the animal is arrested. All these phenotypes are hallmarks of the pat phenotype (Williams and Waterston, 1994). Bottom, progeny of animals subject to RNAi with a control ds RNA derived from a new homeobox gene (Hobert, O., and G. Ruvkun, unpublished data), which displays no lethality.

Figure 8

unc-97 genetically interacts with the mec-3 gene. (A) UNC-97::GFP expression in the ALM touch neurons. Left arrow, axonal process; right arrow, cell body. (B) Genetic interaction of unc-97 and mec-3. Mechanosensory assays were performed on the individual genotypes as described in Materials and Methods. The responsiveness of an individual animal is defined by how often an animal can respond to a total of ten touches on the anterior and posterior half of the animal (e.g., two responses from ten touches is recorded as a 20% responsiveness). Animals tested: unc-97(su110) n = 102; mec-3(u298) n = 102; unc-97(su110), mec-3(u298) n = 100; mec-3(e1338) n = 30; N2 wild-type n = 30. The fraction of the tested animals that show any of the ten different touch responses is plotted against the touch responsiveness.

Figure 8

unc-97 genetically interacts with the mec-3 gene. (A) UNC-97::GFP expression in the ALM touch neurons. Left arrow, axonal process; right arrow, cell body. (B) Genetic interaction of unc-97 and mec-3. Mechanosensory assays were performed on the individual genotypes as described in Materials and Methods. The responsiveness of an individual animal is defined by how often an animal can respond to a total of ten touches on the anterior and posterior half of the animal (e.g., two responses from ten touches is recorded as a 20% responsiveness). Animals tested: unc-97(su110) n = 102; mec-3(u298) n = 102; unc-97(su110), mec-3(u298) n = 100; mec-3(e1338) n = 30; N2 wild-type n = 30. The fraction of the tested animals that show any of the ten different touch responses is plotted against the touch responsiveness.

Figure 2

Analysis of muscle structure in unc-97 mutant animals. (A) High magnification views of individual body wall muscle cells in N2 wild-type and unc-97(su110) living adult animals using polarized light microscopy. In N2 wild-type dense bodies and M lines are easily detected. In the first unc-97(su110) panel, although the overall structure of the muscle cell is similar to that observed in wild type, individual dense bodies are not as easily detected and the edges of the A bands are often ragged. Arrowheads, A bands; arrows, individual dense bodies (adhesion plaques). Bottom four unc-97(su110) panels, varying levels of disorganization that can be observed in su110 muscle cells. Relatively mild disorganization with several sarcomeres are just beginning to detach can be observed (arrowhead). Note how the birefringent material accumulates at the ends of cells in the bottom two panels. (B) Immunohistochemical analysis of dense body structure in wild-type and unc-97(su110) with the monoclonal β-integrin/PAT-3 antibody MH25, the monoclonal vinculin antibody MH24, and the polyclonal anti–UNC-52 antibody. White arrow, dense bodies; white triangle; M line. Note the occasional fusion of diffuse stripes staining with anti-vinculin antibodies in unc-97(su110) animals. Bar, 10 μm.

Figure 3

UNC-97 and the PINCH family of LIM proteins. (A) Chromosomal localization of unc-97. The unc-97(su110) mutation is a G→ A transition in a highly conserved splice acceptor site. Due to the invariance of G at splice sites, these sites represent common targets for the EMS mutagen in C. elegans. (B) Schematic domain structure of UNC-97, D-PINCH, and other PINCH family members. (C) Alignment of the PINCH family. PIN-2 corresponds to F07C6.1. We propose the names h-PINCH-1 for the first original PINCH gene described by Rearden (1994), h-PINCH-2 for the incomplete human EST clone c-0qd01 (GenBank/EMBL/ DDBJ accession number Z42656), m-PINCH-1 for the mouse EST clone vc92b05.r1 (GenBank/EMBL/DDBJ accession number AA289280), and m-PINCH-2 for the mouse EST clones vf78h08.r1 (GenBank/EMBL/DDBJ accession number AA450826) and vg99h07.r1 (GenBank/ EMBL/DDBJ accession number AA471768). All of the ESTs are incomplete and were corrected for several obvious errors causing reading frame shifts. Black arrowhead, location of the splice site mutation that is predicted to disrupt the structural integrity of the last LIM domain in UNC-97. X, ambiguities in the EST sequences. (D) The relationship of the UNC-97 family to one another and to other LIM domain subfamilies that also consist of multiple LIM domains is shown in this dendrogram. Note that this dendrogram created by the pileup program represents a clustering based on similarity and is not representative of evolutionary distance. SLIM family members are characterized by the presence of four LIM domains preceded by an incomplete fifth LIM domain. Paxillin family members contain four LIM domains. Vertebrate paxillins, but not the C. elegans homologue C28H8.6 contain an NH₂-terminal extension. Note that all subfamilies contain at least one C. elegans homologue. Database accession numbers (incl. references) are: h-PAXILLIN: Swiss prot P49023; m-HIC-5: GenBank/EMBL/DDBJ L22482; C28H8.6: Swiss prot Q09476; Dm-PAXILLIN (EST GM04891.5): GenBank/EMBL/DDBJ AA696001; h-SLIM-3 (identical to h-DRAL): Swiss prot Q14192; h-SLIM-2: Swiss prot Q13643; h-SLIM-1 (the human orthologue of the mouse KyoT gene): Swiss prot Q13642; F25H5.1: GenBank/EMBL/DDBJ Z81068.

Figure 5

Distribution of d-pinch RNA during Drosophila development. (A) Lateral view of a stage 9 embryo, hybridized with antisense d-pinch probe; oriented with anterior left, dorsal side up. The transcript is concentrated in the developing mesoderm, whereas there is no specific staining seen with the negative control (data not shown). (B) Dorsal view of an early stage 13 embryo, hybridized with antisense d-pinch probe. D-pinch transcript is detected strongly in the visceral mesoderm flanking the developing gut (arrowheads), as well as in the somatic musculature. (C) Dorsal view of a late stage 16 embryo, hybridized with antisense d-pinch probe. The d-pinch transcript is still strongly expressed in the visceral mesoderm (white arrowheads) that now completely envelopes the gut and the somatic body wall muscles (black arrowheads). Pharyngeal muscle and epidermal tendon cell staining is also present at this stage; however, these tissues are not discernible in this plane of focus. (D) Isolated epidermis from a stage 16 embryo, hybridized with d-pinch probe. Note the strong hybridization at the segment boundaries where the somatic muscle cells attach to the epidermis. Arrowheads, two attachment sites.

Figure 9

Expression of the UNC-97–related LIM protein PIN-2. (A) The GFP reporter gene construct is depicted schematically. The last exons of the predicted gene F07C6.2 that precedes PIN-2/F07C6.1 are shown to demonstrate that virtually the whole 5′ upstream regulatory region of PIN-2 has been incorporated into the reporter gene construct. (B–D) Expression of the PIN-2:: GFP reporter construct. PIN-2::GFP expression in neurons is highly penetrant; however, the exact number of PIN-2::GFP– expressing cells varies slightly from animal to animal. We have not observed more than 10 head neurons and three tail neurons labeled in any one individual. The strongest expressing tail neuron (D) has a size and location consistent with the PVT interneuron, and sends a process into the nerve ring (C). At least one of the head neurons is an amphid sensory neuron (B). Arrows in C, variscosities in the ventral nerve cord. Those can be also seen in unanaesthetized animals, ruling out artefactual induction of variscosities as often observed, for example, with sodium azide.