Two sets of interacting collagens form functionally distinct substructures within a Caenorhabditis elegans extracellular matrix

Abstract

A ubiquitous feature of collagens is protein interaction, the trimerization of monomers to form a triple helix followed by higher order interactions during the formation of the mature extracellular matrix. The Caenorhabditis elegans cuticle is a complex extracellular matrix consisting predominantly of cuticle collagens, which are encoded by a family of approximately 154 genes. We identify two discrete interacting sets of collagens and show that they form functionally distinct matrix substructures. We show that mutation in or RNA-mediated interference of a gene encoding a collagen belonging to one interacting set affects the assembly of other members of that set, but not those belonging to the other set. During cuticle synthesis, the collagen genes are expressed in a distinct temporal series, which we hypothesize exists to facilitate partner finding and the formation of appropriate interactions between encoded collagens. Consistent with this hypothesis, we find for the two identified interacting sets that the individual members of each set are temporally coexpressed, whereas the two sets are expressed approximately 2 h apart during matrix synthesis.

Figure 1

Cuticular structures in C. elegans. (A) Diagram of a section of a C. elegans adult showing cuticular structures (annuli and alae) and a cross section of the underlying hypodermal cells. (B) Diagram of a longitudinal cross section through the cuticle and hypodermis, indicating the location of the annular furrows in the outer cortical layer of the cuticle. (C) Cross section as in B but during cuticle synthesis, with a new outer cortical layer underneath a displaced old cuticle, showing the position of the transient circumferential actin bundles in the hypodermal cells and the associated furrows above these bundles on the hypodermal membrane surface. Note the juxtaposition of these bundles, the furrows in the hypodermal apical membrane, and the annular furrows in the newly formed cortical layer.

Figure 2

Dpy phenotype. Images of adult hermaphrodites are shown at the same magnification. (A) Wild-type strain N2. (B) dpy-7(qm63) mutant. (C) dpy-13(e458) mutant. (D) dpy-13(e458);dpy-7(qm63) double mutant.

Figure 3

Immunolocalization of DPY-7 in C. elegans embryos. (A–F) DPY-7 localization in a comma stage embryo (A), a threefold embryo during cuticle secretion (B), and a late threefold embryo after cuticle secretion (C). (D–F) Nomarski images of the same embryos. In A, nuclei are seen as nonstained circles surrounded by a halo of DPY-7 fluorescence (white arrow). In B, some extracellular localized DPY-7 (red arrow) and some remaining perinuclear-localized DPY-7 (white arrow) are both seen. This animal seems to be in the process of secreting DPY-7. In C, secretion seems to be complete, with DPY-7 being detected in mature circumferential stripes. (G–I) Embryo double labeled with the DPY7-5a mAb (G and I) and anti-LIN-26 polyclonal antibody (H and I). LIN-26 is present in all hypodermal cell nuclei at this stage. White arrows, the three rows of cells that make up the main body hypodermis of one side of the animal are indicated: P cells that form the ventral hypodermis (bottom row); V cells that form the lateral seam (middle row); hyp-7 cells that form the main body hypodermal syncytium (top row). (J and K) Comma-stage embryo (J) and a late threefold embryo after cuticle secretion (K) double stained with DPY7-5a and MH27 mAbs. The hypodermal cell boundaries are visualized as thin lines by MH27 detection of adherens junctions. Arrow in J, the nucleus of one of the seam cells; the halo of DPY-7 localization is visible around it. By the late threefold stage, the hypodermal cells have elongated and the cuticle has polymerized on the hypodermal surface; arrows in K indicate the dorsal and ventral boundaries of an elongated seam cell.

Figure 4

Immunolocalization of DPY-7 in the cuticle of larval and adult stages. (A–D) Nomarski (A and C) and fluorescence (B and D) images of two adults. A and B are in the some focal plane, as are C and D; C and D are at higher magnification. The alae are visible in A, running left to right across the specimen. The annuli are at right angles to the alae and extend to the edges of the animal. White arrows, a region of the cuticle where the regular spacing of the annuli is disrupted; every second delineating furrow is absent. The DPY-7 bands are seen to locate to the furrows. (E and F) Midstage larvae (L2 or L3), which do not have alae, stained with the DPY7-5a mAb (E) or double stained with the DPY7-5a and MH27 mAbs (F). White arrow, the boundary of an underlying seam cell detected with MH27. (G) DPY7-5a localization in an adult fixed by the modified Ruvkun method involving partial reduction of the specimen. The thread-like nature of the DPY-7 bands after partial reduction is evident.

Figure 5

Scanning electron micrographs of the surfaces of adult wild-type and mutant worms. A–K are at the same magnification; L is at twice this magnification. (A) Wild-type strain N2. (B) dpy-7(qm63). (C) dpy-7(e88). (D) dpy-13(e458). (E) dpy-2(e1359). (F) dpy-2(e8). (G) dpy-3(e27). (H) dpy-8(e130). (I) dpy-8(sc44). (J) dpy-10(sc48). (K) dpy-10(e128). (L) sqt-1(e1350). Alae are visible in most images as three distinct parallel lines; annuli are visible in A and D, running perpendicular to the black lines that indicate the span of ten annuli. The white arrow in D indicates the region close to the alae on the dpy-13(e458) mutant where annuli are disorganized in comparison with wild type.

Figure 6

Immunolocalization of the DPY-7, DPY-10, and DPY-13 collagens in adults of the various genetic backgrounds. (A–J) DPY-7 visualized with the DPY7-5a mAb in wild-type (A), dpy-7(e88) (B), dpy-13(e458) (C), dpy-2(sc38) (D), dpy-3(e27) (E), dpy-8(e130) (F), dpy-10(e128) (G), dpy-10(sc48) (H), sqt-1(e1350) (I), and sqt-3(e2117) (J) animals. In C, the region adjacent to the alae where the DPY-7 stripes are disorganized in the dpy-13(e458) mutant, is indicated by a white arrow. In B, E, F, and G, extended exposures were required to visualize the relatively faint fluorescent signals. (K–M) Transgenic epitope-tagged DPY-10::Ty visualized with the anti-Ty antibody in a dpy-10(e128) mutant (K) (note phenotypic rescue by the transgene), a dpy-7(e88) mutant (L), and a dpy-7(qm63) null mutant (M). In L and M, extended exposures were used in attempting to detect any faint DPY-10::Ty signal; the long exposure led to the visualization of background fluorescence from the alae in M. (N–P) Visualization of DPY-13::Ty and DPY-7 (O) in dpy-13(e458) animals in which the mutant phenotype is rescued by transgenic expression of dpy-13::Ty. In O, both primary antibodies were detected with the same secondary antibody. In P, the animal had been treated with dpy-7 RNAi.

Figure 7

Timing of expression of the dpy-2, dpy-3, dpy-5, dpy-8, and dpy-10 genes. The results of semiquantitative RT-PCR for the cuticle collagen genes dpy-2, dpy-3, dpy-5, dpy-8, and dpy-10, performed on a time course of RNA samples taken at 2-h intervals from early Ll to young adult, displayed from left to right in the figures. Molts occurred during the time points represented by samples 6 (L1 to L2), 9 (L2 to L3), 12 (L3 to L4), and 15–16 (L4 to adult). (A) Negative image of an ethidium bromide-stained agarose gel of the RT-PCR samples for dpy-8. The bands labeled 1 and 2 are amplified dpy-8 cDNA and genomic dpy-8 sequences, respectively. The RNA samples used are contaminated to various extents with genomic DNA, but amplification across an intron allows amplified cDNA and genomic DNA to be distinguished by size. In samples where a cDNA species is abundant, reduced amplification of the relevant contaminating genomic band is generally seen, as a result of competition for oligonucleotides. Bands labeled 3 are amplified control ama-1 cDNA. (B and C) Graphs showing relative transcript abundances, generated from data like those of A (see MATERIALS AND METHODS). (B) dpy-2, dpy-8, and dpy-10. (C) dpy-8, dpy-3, and dpy-5. In each cycle, the dpy-5 peak is one sample later than those for the other genes.