Caenorhabditis elegans exoskeleton collagen COL-19: An adult-specific marker for collagen modification and assembly, and the analysis of organismal morphology



The integral role that collagens play in the morphogenesis of the nematode exoskeleton or cuticle makes them a useful marker in the examination of the collagen synthesizing machinery. In this study, a green fluorescent protein–collagen fusion has been constructed by using the Caenorhabditis elegans adult-specific, hypodermally synthesized collagen COL-19. In wild-type nematodes, this collagen marker localized to the circumferential annular rings and the lateral trilaminar alae of the cuticle. Crosses carried out between a COL-19::GFP integrated strain and several morphologically mutant strains, including blister, dumpy, long, small, squat, and roller revealed significant COL-19 disruption that was predominantly strain-specific and provided a structural basis for the associated phenotypes. Disruption was most notable in the cuticle overlying the lateral seam cell syncytium, and confirmed the presence of two distinct forms of hypodermis, namely the circumferentially contracting lateral seam cells and the laterally contracting ventral–dorsal hypodermis. The effect of a single aberrant collagen being sufficient to mediate widespread collagen disruption was exemplified by the collagen mutant strain dpy-5 and its disrupted COL-19::GFP and DPY-7 collagen expression patterns. Through the disrupted pattern of COL-19 and DPY-7 in a thioredoxin mutant, dpy-11, and through RNA interference of a dual oxidase enzyme and a vesicular transport protein, we also show the efficacy of the COL-19::GFP strain as a marker for aberrant cuticle collagen synthesis and, thus, for the identification of factors involved in the construction of collagenous extracellular matrices. Developmental Dynamics 226:000–000, 2003. © 2003 Wiley-Liss, Inc.


Collagens represent the major structural extracellular matrix (ECM) proteins in metazoans and are formed from the trimerization of procollagen polypeptides that have a Gly-X-Y repeat motif, where X and Y are most commonly proline and hydroxyproline. The free-living nematode Caenorhabditis elegans forms two major collagenous ECMs, the cuticular exoskeleton and the tissue-surrounding basement membranes, and represents an excellent model system to study these proteins and the ECM (Kramer, 1997; Hutter et al., 2000; Johnstone, 2000; Page, 2001). There are 154 collagen genes expressed in the C. elegans genome that are involved in the formation of the cuticle (Johnstone, 2000), which predominately encode small Gly-X-Y–rich proteins with interruptions in their collagenous domains. The co- and posttranslational hydroxylation of the numerous Y-position prolines has been established as being an essential step in the assembly and maturation of these structural proteins (Friedman et al., 2000; Winter and Page, 2000; Myllyharju et al., 2002). The interruptions contain clusters of conserved cysteines, the number and spacing of which allows the grouping of these collagens into 5–6 distinct families (Johnstone, 2000). Mutations in some of the cuticle collagen genes (dpy-2, dpy-7, dpy-10, dpy-13, sqt-1, sqt-3, rol-6, bli-1, and bli-2) result in a range of morphologic defects, including short and fat (dumpy) and blister phenotypes (Kramer, 1997; Johnstone, 2000; Page, 2001). Likewise, mutation in collagen processing enzymes (BLI-4; Peters et al., 1991), modifying enzymes (DPY-18, PHY-2, and PDI-2, Winter and Page, 2000; PDI-3, Eschenlauer and Page, 2003; DPY-11, Ko and Chow, 2002), and the duox cross-linking enzymes (Edens et al., 2001) cause similar severe body form defects.

The nematode cuticle collagens are synthesized and secreted from the underlying hypodermis and assemble to produce larval and adult stage-specific cuticles, that are both structurally and chemically distinct (Page, 2001). The cuticle protects the nematode from the environment, prevents desiccation, and acts as a hydrostat, thereby permitting locomotion by means of opposed longitudinal muscles. Two major forms of hypodermis are present in C. elegans, the syncytial ventral and dorsal hypodermal cells that cover the head, tail, ventral and dorsal surfaces, and the distinct lateral seam cells that terminally differentiate to form two parallel syncytial bands or cords at the fourth larval to adult molt (Singh and Sulston, 1978; Sulston et al., 1983). At this final molt, the cuticle attains many adult-specific features, including the alae, a set of three raised cuticular ridges that form lateral stripes above the seam cell cords.

The majority of examined cuticle collagens have temporal expression patterns that peak in a cyclical manner in the preadult stages corresponding to the molting cycle (Cox and Hirsh, 1985; Kingston et al., 1989; Park and Kramer, 1994; Johnstone and Barry, 1996). The collagen encoded by col-19 is unusual in that it is temporally stage-specific (Liu et al., 1995). This gene is expressed in both hypodermal cell types, and has been used as a heterochronic marker due to its strict temporal regulation by the transcription factor LIN-29 (Liu et al., 1995; Abrahante et al., 1998).

In this study, we describe the detailed characterization of the adult-specific collagen, COL-19 in C. elegans. We demonstrate that a C-terminal green fluorescent protein (GFP) -tagged collagen is expressed exclusively in the terminally differentiated adult cuticle and provides an invaluable molecular tool to identify and study defects in the secretion, modification, processing, and cross-linking of the proteins that constitute this essential ECM. This collagen marker also elucidates a structural basis for the defective body morphology resulting from mutations in the collagen biosynthetic pathway components.


COL-19 Is Expressed in a Stage- and Structure-Specific Manner

The cuticle collagens COL-19, ROL-6, and C39E9.9 were translationally fused to a GFP marker protein, by means of their C-termini, and analyzed by transgene experiments to determine whether GFP-tagged collagens can be incorporated and expressed in the nematode cuticle. All three collagens belong to the same gene family, namely group I collagens, and all have short noncollagenous C-terminal domains (domain III) (Johnstone, 2000); being 14, 25, 15 residues, respectively. As a consequence, these collagens are expected not to be susceptible to C-terminal posttranslational proteolytic processing. COL-19 has the additional advantage of being under tight regulatory control, being expressed stage-specifically at the dauer molt and the final L4- adult molt (Liu et al., 1995). The two remaining collagens tested were expressed in multiple stages, with transcripts being detected in both larval and adult stage cuticles (Park and Kramer, 1994; Jiang et al., 2001). These expression experiments revealed that, of the three sets of transgenic lines, only COL-19::GFP was expressed in the cuticle of live worms, and extrachromosomal arrays carried by these lines were integrated into the genome of C. elegans to provide the strain TP12:kaIs(col-19::gfp). Two additional COL-19::GFP integrated lines were obtained and found to express identical GFP patterns to strain TP12:kaIs(col-19::gfp). The C39E9.9 and ROL-6 constructs did generate viable transgenic animals as determined by coexpression of the rol-6 dominant marker but were, however, GFP negative.

COL-19::GFP is expressed exclusively in the adult cuticle, with a distinctive spatial expression pattern (Fig. 1A–C). The tagged protein is expressed in both forms of hypodermally derived cuticle, specifically, the matrix of the circumferential transverse annuli (Fig. 1D,F) and the tri-laminate lateral alae (Fig. 1E,F). This dominant cuticle matrix expression was similar between adult hermaphrodites and males. In addition, the hermaphrodite-specific vulva fluoresced strongly (data not shown), as did the male-specific tail structures, including the cuticular ray and fan structures (Fig. 1B). Significantly, both of these structures are derived from the hypodermal epithelia (Sulston and Horvitz, 1977). The widespread adult cuticle expression pattern is remarkably similar to that previously described for rabbit polyclonal antibodies raised against total adult cuticles and probed against the surface of live worms (Cox et al., 1980; Kramer et al., 1988), supporting the prediction that the GFP-tagged COL-19 expression may reflect the endogenous COL-19 expression pattern. Additionally, a recent study analyzing the localization of a Ty-tagged DPY-13 transgene displayed a remarkably similar annuli expression pattern to that described here for COL-19::GFP (McMahon et al., 2003). The integration of the GFP-tagged construct did not cause any visible effect on the morphology of wild-type adult nematodes, as wild-type regularly space (1.2 μm wide) annuli and normal trilaminate alae were observed (data not shown). COL-19::GFP expression was also examined in dauer stage larvae, as this resistant arrested development stage previously was also determined to express this transcript (Liu et al., 1995). Of interest, apart from the adult and first larval stages (L1), the dauer larvae is the only other developmental stage to possess lateral alae. After allowing TP12:kaIs(col-19::gfp) strain to starve, a large population of dauer larvae were obtained, examined, and found not to express the COL-19::GFP protein in their cuticles (data not shown). Likewise, no annuli or ala-specific GFP expression was detected in temperature-induced dauers of a daf-2(e1370) mutant allele strain crossed with TP12:kaIs(col-19::gfp) males.

Figure 1.

COL-19::GFP expression in wild-type nematodes. A: COL-19::GFP expression in the TP12:kaIs(col-19::gfp) adult hermaphrodite cuticle. Annuli (an) and lateral ala (la) express a uniform staining pattern. Annular furrows are green fluorescence protein (GFP) negative (arrowed). The close apposition of the annular furrows to the lateral alae are depicted by double-headed arrows. B: COL-19::GFP expression in TP12:kaIs(col-19::gfp) adult male tail. Only the rays are in focus. C: Merged image of DPY-7 antibody staining of wild-type TP12:kaIs(col-19::gfp) hermaphrodite, displaying specific DPY-7 localization to annular furrows (red, arrowed) and COL-19::GFP in annular cortex (green). The characteristic wild-type close apposition of the annular furrows to the lateral alae (la), is depicted by double-headed arrows. D: Transmission electron photomicrograph of longitudinal transverse section of the adult cuticle, depicting a single annulus to schematically illustrate the relative positions of COL-19 and DPY-7. DPY-7 localises to the annular furrows (arrowed) and COL-19 to the annulus (between vertical bars). The cuticle layers are cortex (c), medial (m), basal (b), and hypodermis (hy). E: Scanning electron photomicrograph of a wild-type adult hermaphrodite cuticle. The regularly spaced annuli (an) are depicted and are separated by the annular furrows (black arrows). The annuli and the furrows closely appose the tri-laminate lateral alae (la) as depicted by double-headed arrows. F: Schematic representation of the adult cuticle and the predicted roles the adult hypodermis and cuticle play in maintaining normal nematode body shape. The distinct dorsal–ventral hypodermis (shaded) and the lateral seam cell cords and overlying alae are depicted. The horizontal arrow A denotes the direction of longitudinal contraction due to annuli and lateral seam cell–derived cuticle. The vertical arrow B denotes circumferential direction of contraction due to seam cell-derived cuticle. Scale bars = 10 μm in A–C, 1.2 μm in D, 10 μm in E.

By using the integrated strain TP12:kaIs(col-19::gfp) costained with a monoclonal antibody against the cuticle collagen DPY-7, we were able to substantiate the annuli-specific expression pattern of COL-19 (Fig. 1C). The DPY-7 protein is temporally expressed in the cuticle during all stages of development (Johnstone and Barry, 1996) and is spatially restricted to the transverse grooves or furrows of the cuticle cortex called the annulations or annular furrows (Fig. 1D; McMahon et al., 2003). This collagen is absent from the annuli and is not expressed in the seam cell-derived cuticle, including the adult lateral alae (McMahon et al., 2003). The colocalization studies reported here confirmed this annular furrow-specific DPY-7 pattern (Fig. 1C). In contrast, COL-19::GFP was expressed in a transverse banding pattern of approximately 1.2 μm in thickness, consistent with the annuli, and was absent from the annular furrows. The merged image of DPY-7 antibodies on the fixed COL-19::GFP strain clearly reveals the adjacent but nonoverlapping expression patterns of these two collagens (Fig. 1C). In the wild-type adult cuticle, both the annuli and annular furrows closely appose the lateral alae; a feature clearly depicted in the costained nematodes (Fig. 1C, double-headed arrows), in a scanning electron photomicrograph of a wild-type nematode (Fig. 1E, double-headed arrows) and in the schematic representation of the adult cuticle (Fig. 1F). A transmission electron photomicrograph of a single annulus and bordering annular furrows from an adult stage hermaphrodite is depicted to illustrate the relative localizations of DPY-7 (Fig. 1D, annular furrows, arrows) and COL-19 (Fig. 1D, annulus, bars).

COL-19 Is Nonessential But Its Disruption Leads to Adult-Specific Ala Defects

RNA interference (RNAi) was applied to determine the importance of COL-19 in the assembly and maintenance of the adult nematode cuticle. Three separate RNAi approaches were used in this study, feeding a bacterial vector that synthesizes the corresponding col-19 dsRNA in vivo, soaking L4 larvae in a solution containing col-19 dsRNA, and finally injecting a solution of col-19 dsRNA into the syncytial gonad of C. elegans. These experiments were carried out on wild-type nematodes, TP12:kaIs (col-19::gfp) nematodes, and a dauer constitutive strain CB1370(daf-2). In general, no dramatic effects on growth or gross morphology were observed following the RNAi procedures in the strains tested, consistent with COL-19 being an adult-specific collagen (Abrahante et al., 1998). The most conspicuous effects noted were on wild-type and TP12:kaIs(col-19::gfp) adult stage nematodes and took the form of subtle ala structural defects; these structures being variably mutant, multiple, or discontinuous (Fig. 2). Before col-19 RNAi, the cuticle of TP12:kaIs(col-19::gfp) adult stage nematodes displayed the normal wild-type triple-track alae structure (Fig. 2A). After col-19 RNAi by feeding, the adult progeny displayed disrupted cuticular ridges that were discontinuous; however, the cuticular annuli and the annular furrows were evenly spaced and remained wild-type in appearance (Fig. 2B). After col-19 RNAi by means of injection, the ala of the adult progeny were likewise found to be mutant, with a characteristic discontinuous broken appearance and/or multiple ala being present (Fig. 2C). No differences were detected between RNAi performed on wild-type N2 or TP12:kaIs(col-19::gfp) nematode strains (data not shown). The RNAi treatments in the TP12:kaIs(col-19::gfp) strain did lead, however, to the complete ablation or the disruption of the wild-type pattern of COL-19::GFP expression (data not shown), therefore, confirming the efficiency of the depletion procedures and that COL-19::GFP was functioning as a translational fusion. In general, RNAi injection then soaking and finally feeding procedures were most effective. The effect of col-19 RNAi on dauer larvae was likewise assessed in daf-2(e1370) nematodes following the three different RNAi approaches. The dauer larvae remained wild-type in appearance following RNAi, and presented none of the ala-specific defects noted in adults (data not shown); however, similar to wild-type and TP12:kaIs(col-19::gfp) nematodes, the alae of adult stage daf-2(e1370) nematodes were abnormal (data not shown).

Figure 2.

The effect of col-19-specific RNA interference (RNAi) in the TP12:kaIs(col-19::gfp) strain. A: Nomarski differential interference contrast microscopy image of the cuticle of untreated TP12:kaIs(col-19::gfp) adult hermaphrodite. Characteristic well-defined triple-track lateral ala ridges are denoted (la) and individually arrowed. Inset depicts a ×2 magnification of alae region. B: RNAi by means of col-19 L4440 feeding. F1 adult from feeding plate, denoting discontinuous broken lateral ala (arrowed). Annuli are in focus and remain morphologically wild-type in appearance. Inset depicts a 2× magnification of disrupted alae region. C: F1 adult from injection experiment (1 mg/ml dsRNA injected gonadally to parent), denoting discontinuous ala (large arrow) and multiple ala (four instead of three, arrowed individually). Inset depicts a 2× magnification of multiple ala. Scale bars = 10 μm for A–C.

COL-19 as a Marker for Defects in Adult Cuticle ECM Morphology

Experiments were performed to examine COL-19::GFP expression in the context of it being used as a specific collagen marker to characterize defective ECM formation in adult hermaphrodite and male nematodes. The COL-19::GFP–expressing strain TP12:kaIs(col-19::gfp) was crossed with several cuticle-related morphologic mutants of C. elegans, and two examples producing distinct phenotypes are given here, namely dpy-5(e61) and dpy-11(e224) (Fig. 3). Additionally, a range of mutant alleles that affected postembryonic body morphology were also examined, including roller (Rol), blister (Bli), dumpy (Dpy), small (Sma), squat (Sqt), and long (Lon) mutants, and these results are summarized in Table 1, Figures 4 and 5, and in the supplementary Figures I–III. The TP12:kaIs(col-19::gfp) strain was likewise used in a RNAi feeding-based study to examine potential collagen-modifying enzymes, collagen cross-linking enzymes, and secretory pathway transport proteins expressed in the C. elegans genome. An example of a class of dual oxidases cross-linking enzymes (duox; F56C11.1 and F53G12.3) and a potential ER to Golgi transport protein (ZK1098.5) are presented (Fig. 6).

Figure 3.

Aberrant COL-19 and DPY-7 expression in morphologic mutants. A: Scanning electron microscopy image of cuticle collagen gene mutant dpy-5(e61) adult hermaphrodite. Dpy phenotype in mid-body and tail due to failed contraction of the lateral seam cell cuticle (double-headed arrow). Lateral ala (la) are depicted by single-headed arrows; head (h) and vulva (v) also are depicted. B: TP14:dpy-5(e61)I;kaIs12(col-19::gfp) strain showing aberrant COL-19::GFP expression overlying the lateral seam cell hypodermis. The double-headed arrow depicts the noncontracted seam cell cuticle; the single-headed arrow depicts the position of lateral ala (la); tightly constricted annuli are indicated (an). C: TP14:dpy-5(e61)I;kaIs12(col-19::gfp) strain probed with DPY-7 antibody (red) showing aberrant COL-19::GFP (green) and aberrant DPY-7 expression overlying the lateral seam cells. The double-headed arrow depicts the noncontracted seam cell cuticle; the single-headed arrow depicts the position of the lateral ala (la). D: TP14:dpy-5(e61)I;kaIs12(col-19::gfp) strain highlighting the noncontracted seam cell region (double-headed arrow). E: Differential interference contrast microscopy (DIC) image of D, showing the smooth appearance of the surface cuticle overlying the noncontracted seam cell (double-headed arrow), the tight arrangement of the annuli (an), and the relatively wild-type appearance of the lateral ala (la, arrowed). F: Scanning electron microscopy image of a thioredoxin mutant dpy-11(e224) adult hermaphrodite. The Dpy phenotype in mid-body and especially tail, due to failed contraction of the lateral seam cell hypodermis (double-headed arrows). The branched, bifurcating lateral ala (la) is depicted between single-headed arrows; the head (h) also is depicted. G: TP16:dpy-11(e224)V;kaIs12(col-19::gfp) nematode showing aberrant COL-19::GFP expression overlying the lateral seam cell cuticle. The double-headed arrow depicts the partially contracted seam cell cords, the single-headed arrows depict the position of bifurcating lateral ala (la), and annuli are indicated (an). H: TP16:dpy-11(e224)V;kaIs12(col-19::gfp) strain probed with DPY-7 antibody (red) showing aberrant COL-19::GFP and DPY-7 expression overlying the lateral seam cell cord. The double-headed arrow depicts noncontracted seam cell cuticle; the single-headed arrow depicts the position of the lateral ala (la). I: TP16:dpy-11(e224)V;kaIs12(col-19::gfp) strain highlighting the noncontracted seam cell region (double-headed arrow). The position of lateral alae (la) is depicted with a single-headed arrow. J: DIC image of I, showing the smooth appearance of the surface cuticle overlying the noncontracted seam cell (double-headed arrow) and the broken irregular appearance of the lateral ala (la, single-headed arrow). Scale bars = 150 μm in A,F, 10 μm in B–E,G–J.

Table 1. Definitions of Classes of COL-19 Adult Cuticle Collagen Disruptiona
GenotypeIdentity (cosmid)LesionPhenotypeAlaeAnnuli
  • a

    Dpy, dumpy; Bli, blister; Mta, male tail abnormal; Rol, roller; ts, temperature sensitive. COL-19::GFP expression is indicated for two areas. Alae: +, wild-type; 1, multiple or discontinuous ala; 2, abnormal branched ala; 3, severe disruption with branching. Annuli: +, wild-type (N, narrow annuli; W, wide annuli); 1, abnormal branched annuli (lateral hypodermis); 2, missing or amorphous annuli; 3, severe disruption.

I: Mild disruption or wild-type     
 TP12:N2;kaIs12[col-19::gfp]NA Wild-type++
 TP29:bli-4(e937)I;kaIs12[col-19::gfp]Subtilisin protease (k04f10.4)[Exon deletion]Bli++
 TP11:rol-6(su1006)II;kaIs12[col-19::gfp]Cuticle collagen (t01b7.7)[N-protease R71C]Rol++
 TP27:sma-2(e502)III;kaIs12[col-19::gfp]Dwarfin (zk370.2)?Small (mta)++N
 TP31:lon-3(e2175)V;kaIs12[col-19::gfp]Cuticle collagen (zk836.1)?Long11W
 TP42:lon-1(e185)III;kaIs12[col-19::gfp]PR protein (f48c8.1)?Long11W
 TP44:lon-2(e678)X;kaIs12[col-19::gfp]Not yet known (c39e6.1)?Long11W
 TP52:lon-2(n1630)X;kaIs12[col-19::gfp]Not yet known (c39e6.1)?Long11W
II: Amorphous annuli     
 TP17:dpy-7(e88)X;kaIs12[col-19::gfp]Cuticle collagen (f46c8.6)[Gly156Arg]Medium Dpy+2
 TP36:dpy-7(qm63)X;kaIs12[col-19::gfp]Cuticle collagen (f46c8.6)[null]Medium Dpy+2
 TP22:dpy-3(e27)X;kaIs12[col-19::gfp]Cuticle collagen (egap7.1)?Medium Dpy12
 TP19:dpy-10(e128)II;kaIs12[col-19::gfp]Cuticle collagen (t14b4.7)[g766a splice]Small Dpy12
 TP24:dpy-8(e130)X;kaIs12[col-19::gfp]Cuticle collagen (c31h2.2)?Medium Dpy12
 TP35:dpy-9(e12)IV;kaIs12[col-19::gfp]Not yet known?Large Dpy12
 TP37:dpy-2(e489)II;kaIs12[col-19::gfp]Cuticle collagen (t14b4.6))[Gly247Arg]Dpy Rol12
 TP49:dpy-2(sc38ts)II;kaIs12[col-19::gfp]Cuticle collagen (t14b4.6)[Gly129Glu]Dpy Rol12
 TP47:dpy-2(e1359)II;kaIs12[col-19::gfp]Cuticle collagen (t14b4.6)?Dpy Rol12
III:Branched seam cell annuli     
 TP21:dpy-13(e458)IV;kaIs12[col-19::gfp]Cuticle collagen (f30b5.1)[deletion of Gly-X-Y]Medium Dpy+1
 TP25:dpy-4(e1166)IV;kaIs12[col-19::gfp]Not yet known?Large Dpy+1
 TP28:bli-1(e769)II;kaIs12[col-19::gfp]Cuticle collagen (c09g5.6)?Bli (ts) (mta)+1
 TP15:bli-3(e767)I;kaIs12[col-19::gfp]Not yet known?Dpy/ Bli (mta)+1
 TP18:bli-5(e518)III;kaIs12[col-19::gfp]Not yet known?Dpy/Bli (mta)+1
 TP14:dpy-5(e61)I;kaIs12[col-19::gfp]Cuticle collagen (f27c1.8)[Gly202opal]Strong Dpy11
IV:Severe global disruption     
 TP23:sqt-3(sc63)V;kaIs12[col-19::gfp]Cuticle collagen (f23h12.4)[Gly227Glu]Medium Dpy (ts)23
 TP20:sqt-3(e2117)V;kaIs12[col-19::gfp]Cuticle collagen (f23h12.4)[Gly114Arg]Extreme Dpy (ts)23
 TP26:sqt-1(e1350)II;kaIs12[col-19::gfp]Cuticle collagen (b0491.2)[N-protease R69C]Medium Dpy (mta)33
 TP40:sqt-1(sc1)II; kaIs12[col-19::gfp]Cuticle collagen (b0491.2)[N-protease R66C]Medium Dpy (mta)33
 TP33:sqt-1(sc13)II;kaIs12[col-19::gfp]Cuticle collagen (b0491.2)[C-domain C303Y]Medium Dpy Rol23
 TP45:sqt-2(sc3)II;kaIs12[col-19::gfp]Not yet known?Medium Dpy33
 TP48:sqt-2(sc108)II;kaIs12[col-19::gfp]Not yet known?Medium Dpy33
 TP32:dpy-17(e164)III;kaIs12[col-19::gfp]Cuticle collagen (f54d8.1)?Medium Dpy33
V:Branching alae and annuli     
 TP13:dpy-18(e364)III;kaIs[col-19::gfp]Prolyl 4-hydroxylase (T28D6.1)[Trp76amber]Medium Dpy (mta)21
 TP16:dpy-11(e224)V;kaIs12[col-19::gfp]Thioredoxin (F46E10.9)[G76E partial]Medium Dpy (mta)21
 TP30:dpy-11(e1180)V;kaIs12[col-19::gfp]Thioredoxin (F46E10.9)[null]Medium Dpy (mta)22
Figure 4.

Aberrant COL-19::GFP expression in class II, III, IV, and V morphologic mutants. A: An example of a class II mutant (amorphous annuli), TP17:dpy-7(e88)X;kaIs12(col-19::gfp) strain, showing amorphous annular COL-19::GFP expression. This image represents a glycine substitution mutant allele of dpy-7. Position of lateral alae is indicated (la) and is slightly out of focus compared with the annuli. B: An example of a class III mutant (branched seam cell–derived annuli), TP28:bli-1(e769)II;kaIs12(col-19::gfp) strain, showing aberrant branched seam cell annuli COL-19::GFP expression. The constricted annuli are denoted (an). The area of mutant seam cell annuli lies between the dorsal–ventral hypodermally derived annuli and the lateral alae (la) and is highlighted by a double-headed arrow. C: An example of the severe disruption in a class IV mutant (severe disruption), TP22:sqt-1(e1350)II;kaIs12(col-19::gfp) at 20°C. This image represents an N-terminal protease processing-site mutant allele of sqt-1 and displays the severely disrupted and irregular annuli (an) and branched, broken, and disrupted lateral ala (la). Ala bifurcation is highlighted by arrows. D: An example of a class V mutant (branching ala and annuli), TP30:dpy-11(e1180)V;kaIs12(col-19::gfp). This image represents a severe null allele of the thioredoxin enzyme DPY-11 and displays the corresponding branching ala (arrowed) and amorphous annuli. Scale bars = 10 μm in A–D.

Figure 5.

Comparison of allele-specific effects and RNA interference (RNAi) of selected cuticle collagen mutants. A: TP17:dpy-7(e88)X;kaIs12(col-19::gfp) represents a DPY-7 glycine-substitution mutant with characteristic class II amorphous annuli COL-19::GFP expression pattern. B: TP36:dpy-7(qm63)X;kaIs12(col-19::gfp) represents a DPY-7 null mutant that also displays the characteristic class II amorphous annuli COL-19::GFP expression pattern. C:dpy-7 RNAi in TP12:kaIs12(col-19::gfp) background displays an identical amorphous annuli expression pattern to A and B. D: TP23:sqt-3(sc63)V;kaIs12(col-19::gfp) represents a C-terminal glycine substitution mutant of SQT-3 that exhibits the characteristic severe class IV disruption pattern, which effects the annuli and the lateral alae (arrowed). E: TP20:sqt-3(e2117)V;kaIs12(col-19::gfp) represents an N-terminal glycine-substitution mutant of SQT-3 that also exhibits the characteristic severe class IV disruption pattern, effecting both the annuli and the lateral alae (arrowed). F:sqt-3 RNAi in TP12:kaIs12(col-19::gfp) strain results in mild disruption of the annuli, and a wild-type COL-19::GFP expression pattern in the lateral ala (class I). G: TP40:sqt-1(sc1)II;kaIs12(col-19::gfp) represents an N-terminal protease processing-site mutation in SQT-1 that results in severe disruption of COL-19::GFP in the lateral ala (bifurcating ala are arrowed) and the annuli and is characteristic of the class IV mutants. H: TP33:sqt-1(sc13)II;kaIs12(col-19::gfp) represents a C-terminal cross-linking residue mutation in SQT-1 and results in a milder version of the global disruption (class IV). Both the annuli and ala COL-19::GFP expression patterns are aberrant but display less of the disorganization denoted in G (class IV). I:sqt-1 RNAi in TP12;kaIs12(col-19::gfp) results in mild disruption of the annuli, and a relatively wild-type COL-19::GFP expression pattern in the lateral ala (class I). The relative position of the lateral alae is depicted in all images (la). Scale bars = 10 μm in A–I.

Figure 6.

Aberrant COL-19::GFP expression in RNA interference (RNAi) treated TP12:kaIs(col-19::gfp) nematodes. A: RNAi feeding effect of cross-linking enzyme homologue duox (F56C11.1) on adult F1 progeny of treated TP12:kaIs(col-19::gfp) nematodes. COL-19::GFP is severely disrupted, most notably in the region corresponding to the lateral ala (la). Severe disorganization of the adjacent annuli is also evident (arrowed). B: Effects of RNAi injection with ER to Golgi transport homologue Bet3 (ZK1098.5) on TP12:kaIs(col-19::gfp) strain. COL-19::GFP is retained in small vesicles (arrowed) and is not incorporated into the cuticle of F1 adults. The relative position of the lateral ala is indicated (la). Scale bars = 10 μm in A,B.

dpy-5, a Structural Cuticle Collagen Mutation Causes Aberrant Cuticle Collagen Localization

dpy-5(e61) is a medium Dpy mutant affecting the cuticle of L2 to adult stages (Ouazana et al., 1985). The phenotype is due to a lesion in a cuticle collagen gene (F27C1.8; Ann Rose, Vancouver, personal communication). Scanning electron micrograph (SEM) analysis of this strain revealed that dpy-5 mutant nematodes have a relatively normal head morphology; however, the mid-body and tail regions were markedly shorter and fatter than their wild-type counterparts (Fig. 3A). It is interesting to note that the annuli overlying the ventral and dorsal hypodermis are present but differ markedly from wild-type nematodes (Fig. 1E, double-headed arrows) in that they do not oppose the lateral alae (Fig. 3A,E, double-headed arrows). As a result, the cuticle above the lateral hypodermal seam cell cords, having failed to contract normally, is extended, and this results in the characteristic Dpy phenotype. When the strain TP14:dpy-5(e61) I;kaIs12(col-19::gfp), resulting from crossing TP12:kaIs(col-19::gfp) with the dpy-5(e61) mutant, was examined under epifluorescence, the ala, apparent by means of Nomarski imaging (Fig. 3E), were shown to express COL-19::GFP in a variable manner (Fig. 3B, arrowed). In addition, two distinct regions became apparent: a region overlying the dorsal/ventral hypodermal cells, which showed annuli fluorescing as distinct bands as observed in wild-type TP12:kaIs(col-19::gfp) worms (Fig. 3B–D, denoted an), and a second, extended region of disruption overlying the seam cell cords (Fig. 3B–D, double-headed arrows). The disrupted region did not have the characteristic annuli and instead appeared featureless when viewed by Nomarski microscopy (Fig. 3E, double-headed arrow). The regular annular fluorescence did appear more closely packed than in wild-type nematodes (Fig. 3B–D, denoted an), having a periodicity of approximately 0.75 μm compared with 1.2 μm, respectively, an observation consistent with the shorter length of these adult nematodes. The region of dominant COL-19::GFP mutant expression and disruption overlying the lateral seam cell hypodermis has a complex fibrous and broken appearance (Fig. 3B–D, double-headed arrows). In some areas corresponding to the seam cell cords, large vesicles are evident and the tagged protein is found concentrated at the periphery of these structures (Fig. 3D), perhaps indicating a failure in the complete secretion of COL-19::GFP. These data show that mutation of the DPY-5 collagen in these nematodes is sufficient to alter the expression patterns of the COL-19::GFP collagen. This ability of a mutation in one collagen to affect others was further confirmed for the non–stage-specific collagen DPY-7, after costaining of the TP14:dpy-5(e61)I;kaIs12(col-19::gfp) strain with the DPY-7 monoclonal antibodies (Fig. 3C, red). In the dorsal/ventral hypodermal cells, DPY-7 localized in a regular but constricted pattern that corresponded to the annular furrows (Fig. 3C, denoted an) and was similar to the COL-19::GFP patterns observed in wild-type and in dpy-5 mutants. Conversely, in the matrix overlying the lateral seam cell cords, DPY-7 expression was abnormal: the DPY-7 staining pattern appeared broken (Fig. 3C, denoted by double-headed arrow) and generally deviated from the wild-type staining pattern in which both annuli (COL-19) and annular furrows (DPY-7) normally extend to appose the lateral ala (Fig. 1C). Despite having this altered COL-19::GFP expression in the area below the seam cells, Nomarski imaging of the alae revealed a relatively wild-type appearance (Fig. 3E). By using a hypodermal junction-specific monoclonal antibody, MH27 (Francis and Waterston, 1985), we were able to confirm that the hypodermal seam cells were relatively normal in dpy-5(e61) mutant nematodes, whereas the overlying cuticle was abnormal (supplementary Fig. II, F, available online at index.html).

dpy-11, a Thioredoxin Enzyme Mutation That Causes Aberrant Cuticle Collagen Localization

The Dpy mutant dpy-11 has been identified recently as the thioredoxin-like enzyme encoded by F46E10.9 (Ko and Chow, 2002). The point-mutant dpy-11(e224) results in a partial-loss-of-function mutation, characterized by a medium Dpy phenotype and additional male tail-specific morphologic defects. More severe phenotypes are associated with loss-of-function mutations and RNAi, both of which result in severe dumpiness, slow growth, and severe male tail defects. All of these observations are consistent with this enzyme playing an important role in the proper assembly or cross-linking of the nematode cuticle components (Ko and Chow, 2002). When examined at the SEM level, dpy-11(e224) nematodes were particularly Dpy at the mid-body and tail regions and the alae are bifurcated and branched (Fig. 3F, denoted by double-headed arrows). TP12:kaIs(col-19::gfp) and CB224(dpy-11) nematodes were crossed to produce the strain TP16:dpy-11(e224)V;kaIs12(col-19::gfp) that exhibited irregular COL-19::GFP patterns similar but not identical to that of TP14:dpy-5(e61)I;kaIs12(col-19::gfp) in which annuli in the dorsal/ventral hypodermally derived cuticle appeared as a regular, although slightly constricted pattern of bands, again consistent with the Dpy appearance (Fig. 3G,H, denoted an). Similar to dpy-5 mutants, in these dpy-11 mutant animals COL-19::GFP expression in the cuticle overlying the noncontracted seam cell cords was abnormal with annuli appearing branched, prematurely terminated, and occasionally absent, a pattern reflected in the fragmented and branched DPY-7 collagen pattern after antibody staining (Fig. 3H, double-headed arrow region). Lateral ala are present, as depicted with COL-19::GFP and Nomarski, but unlike dpy-5 mutants, in many regions the alae are abnormal, commonly having an unusual bifurcated (Fig. 3G, arrowed) or broken appearance (Fig. 3J, arrowed).

Classification of Morphologic Mutants Based on Aberrant COL-19::GFP Expression Patterns

Several additional crosses were carried out between the TP12:kaIs12 (col-19::gfp) strain and a range of morphologic mutants, the results of which are presented in Table 1, Figures 4 and 5, and in the supplementary Figures I–III. Definable phenotypes became apparent and consequently, strains were assigned to five distinct classes (classes I–V) that ranged from wild-type/mild disruption to severe global COL-19::GFP disruption. Although some mutants (bli-4, rol-6, sma-2, lon-1, lon-2, and lon-3) displayed COL-19::GFP patterns similar to TP12:kaIs12(col-19::GFP) wild-type expression (class I: mild or wild-type, supplementary Fig. I, A–D), the majority of the crossed lines showed remarkably distinctive patterns of disruption. As with the dpy-5(e61) class III and dpy-11(e224) class V crosses, each strain exhibited differing degrees of annuli and ala disruption and was marked accordingly for both categories (Table 1).

Four categories of ala disruption were defined as wild-type (+), multiple or discontinuous ala (1), abnormal branched ala (2), and severe disruption with branching (3). Comparison of the mild disruption, virtually wild-type alae in dpy-7(e88) (Fig. 4A, denoted la) with the severe disruption and ala branching in sqt-1(e1350) (Fig. 4C, denoted la) reveals the range of ala disruption observed with this collagen marker. Furthermore, by comparison to the well-defined longitudinal ala ridges in the wild-type nematodes (Fig 1A,E), the alae of sqt-1(e1350) mutants appear to be composed of short, wispy fragments (Fig. 4C). Less severe examples of aberrant ala are exemplified in supplementary data: Multiple or discontinuous ala are observed in dpy-2(sc38), dpy-3(e27), dpy-10(e128), and dpy-8(e130) mutants (supplementary Fig. I, E–H), whereas branching alae are apparent in sqt-3(sc63), sqt-3(e2117), sqt-1(sc1), sqt-1(sc13) mutants (Fig. 5D,E,G,H, respectively); sqt-2(sc3) and dpy-17(e164) (supplementary Fig. III. , A, B); and in dpy-11 (supplementary Fig. III, C, G) and dpy-18(e364) mutants (supplementary Fig. III, E, D).

With regard to the annuli, four categories of disruption are described, ranging from wild-type/mild disruption (+) to severe disruption (3) (Table 1). Three types of aberrant COL-19::GFP annuli expression, deviating from the regular pattern observed in wild-types, are shown in Figure 4. The bli-1(e769) allele is an example of type I branched annuli (Fig. 4B and supplementary Fig. II, C), whereas dpy-7(e88) and dpy-11(e1180) illustrate what we have termed as missing or amorphous annuli (Fig. 4A,D and supplementary Fig. I, E–H). Severe annuli disruption is exemplified by the sqt-1(e1350 and sc1) and sqt-3(sc63 and e2117) mutant strains (Figs. 4C, 5G,D,E, respectively). As described for the dpy-5(e61) and dpy-11(e224) mutants, much of the observed COL-19::GFP disruption in the strains examined occurs in the region of the cuticle overlying the lateral seam cell cords, whereas a wild-type pattern of expression becomes established in the cuticle associated the ventral and dorsal hypodermis. This finding is especially evident for the class III mutants, including bli-1(e769), dpy-5(e61) (Figs. 4B, 3B, respectively) and dpy-13(e458), dpy-4(e1166), bli-3(e767), and bli-5(e518) (supplementary Fig. II, A,B,D,E, respectively). It is interesting to note that three of the four blister mutants examined in this study fall into the class III pattern of disruption. Many of the mutant nematodes examined also exhibit a mutant COL-19::GFP pattern in the circumferential annuli that overlay the dorsal/ventral hypodermis, including dpy-7(e88), sqt-1(e1350), dpy-11(e1180), dpy-2(sc38), dpy-3(e27), dpy-10(e128), dpy-8(e130), sqt-2(sc3), dpy-17(e164), sqt-3 alleles (sc63 and e2117), and sqt-1 alleles (sc1 and sc13) (Fig. 4A,C,D, supplementary Fig. I, E–H, supplementary Fig. III, A,B, Fig. 5D,E,G,H, respectively). It is significant to note that all the class V mutants, which exhibit both annuli and ala disruption, are mutants in collagen biosynthetic enzymes, whereas classes II, III, and IV are composed predominantly of cuticle collagen mutants (Table 1).

Comparison of Allele-Specific Effects and RNAi-Induced Depletion of Selected Cuticle Collagens

The types of mutations that are present in individual collagens and their consequential effects on body morphology can provide important information regarding the level of redundancy, and the interactions formed between the individual collagens, to shape the final ECM. In C. elegans, a range of mutations affect the cuticle collagens; and some examples are given in Table 1. In most cases, the incorporation of mutant collagens, for example with glycine substitutions, into the final matrix produces the observed phenotypes, whilst the removal of the collagen can also result in a strong phenotype. The removal of some collagens, however, will have no effect on body morphology. In general, a complex range of outcomes can result from the incorporation of individual mutant collagens and in the deletion of single collagens, and this finding reflects the complexity of the interactions between individual collagens in the context of the final exoskeleton. We have applied the TP12:kaIs12;(col-19::gfp) marker strain to examine different mutant alleles of three well-characterized cuticle collagens, and in combination with RNAi-depletion experiments in the TP12:kaIs12;(col-19::gfp) background, the effects of removing five individual collagens have also been assessed (Fig. 5 and supplementary Fig. II, G,H). The class II mutant dpy-7 produced a very similar amorphous annuli COL-19::GFP expression pattern, independent of whether a glycine substitution (e88; Fig. 5A), a deletion mutant (qm63; Fig. 5B), or an RNAi knockdown (Fig 5C) was being examined. In addition, all three groups of nematodes had a similar gross morphologic appearance, as all were medium Dpy with a smooth dorsal/ventral cuticle (data not shown). This result indicates that the DPY-7 collagen is an important, nonredundant cuticle component. Three separate alleles of another class II mutant dpy-2 were also examined, two representing glycine substitution mutants (e489 and sc38) and one that is uncharacterized (e1359), and all three displayed identical COL-19::GFP disruption patterns (Table 1). Two mutant alleles and RNAi disruption of the class IV collagen sqt-3 was also assessed. The mutant alleles examined were both glycine substitution mutants (sc63 and e2117) and both gave similar COL-19::GFP disruption patterns (Fig. 5D,E). The RNAi depletion of sqt-3, however, resulted in morphologically wild-type animals that expressed wild-type to mild disruption of the COL-19::GFP marker (Fig. 5F), confirming that the incorporation of mutant SQT-3 collagen into the ECM is the cause of the mutant phenotype. The analysis of Rol-non-Dpy heterozygotes from crossing TP12:kaIs12(col-19::gfp) males with sqt-3(sc63) hermaphrodites also revealed a relatively wild-type but helically twisted COL-19::GFP expression pattern (data not shown). Examination of another class IV mutant, sqt-1, produced similar findings to the sqt-3 results. In total, three separate alleles of sqt-1, two N-terminal protease recognition site mutants (e1350 and sc1), and a C-terminal putative cross-linking residue mutant (sc13) were examined in addition to RNAi disruption analysis (Figs. 4C, 5G–I, respectively). The N-terminal protease processing site mutants resulted in similar severe phenotypes (Figs. 4C, 5G), displaying both annuli and alae COL-19::GFP disruption patterns with complex branching patterns of the lateral alae. The C-terminal cross-linking domain mutant also displayed a severe type IV COL-19::GFP disruption pattern; however, the severity of this phenotype was less marked that that found for the two N-terminal processing site mutants (Fig. 5H). As described for sqt-3, the sqt-1–specific RNAi resulted in progeny that were morphologically wild-type by Nomarski microscopy (data not shown) and expressed a predominantly wild-type COL-19::GFP pattern (Fig. 5I). Analysis of TP26:sqt-1(e1350) II;kaIs(col-19::gfp) heterozygotes likewise revealed a relatively wild-type COL-19::GFP pattern in the roller–non-Dpy animals (data not shown). As found for sqt-3 mutants examined, the incorporation of mutant SQT-1 collagen into the ECM is the contributing factor to the body form defects. The mutations in the N-terminal processing site, however, do cause more severe disruption effects than mutations affecting the C-terminal cross-linking domain, indicating that perhaps a larger more defective collagen is being incorporated into the exoskeletons of the sqt-1(e1350 and sc1) mutant alleles compared with the sqt-1(sc13) mutants. RNAi depletion of class III collagen mutants was also performed on the TP12:kaIs12(col-19::gfp) strain for the collagens dpy-13 and dpy-5, both of which resulted in identical phenotypes to the premature termination mutants previously examined (compare supplementary Fig. II, A,E to G,H). This finding supports the contention that both DPY-13 and DPY-5 are important components of the normal cuticle.

RNAi-Analysis of Potential Collagen Secretion and Cross-linking Components

Dual oxidase cross-linking enzyme RNAi causes aberrant COL-19::GFP expression.

The open reading frames F56C11.1 and F53G12.3 correspond to a gene duplication at the end of linkage group I, and both encode dual oxidases, large multidomain proteins that have an N-terminal peroxidase domain involved in cuticle cross-linking through di- and tri-tyrosine formation (Edens et al., 2001). Disruption of either of these transcripts by means of RNAi results in morphologic phenotypes consistent with this cross-linking function, namely, Dpy and Bli phenotypes and with a corresponding lack of detectable tyrosine cross-linked residues in the cuticles of treated nematodes (Edens et al., 2001). The effects of RNAi against the dual oxidase enzymes on COL-19::GFP expression was examined by feeding the TP12:kaIs12 (col-19::gfp) nematodes with a bacterial strain expressing dsRNA corresponding to either F56C11.1 or F53G12.3. Our RNAi experiments replicated the Dpy and Bli phenotypes previously reported (Edens et al., 2001; data not shown), with COL-19::GFP providing structural clues as to the cause of this phenotype. Both duox constructs produced a range of RNAi effects on the COL-19 expression pattern, ranging from mild disruption to severe disruption and were ultimately dependent on the quantity and integrity of dsRNA ingested. Severe COL-19::GFP disruption is depicted as a mesh of matrix fibers, lacking the distinctive wild-type dorsal–ventral annular and lateral seam alae expression (Fig. 6A, arrowed).

Depletion of a Bet3-like secretory pathway transport protein (ZK1098.5) causes aberrant COL-19::GFP expression.

A similar RNAi-based examination of a protein with homology to the yeast and human ER to Golgi vesicular transport protein Bet3 was also undertaken. This protein is encoded by the gene ZK1098.5 on linkage group III and was described as inducing a Dpy phenotype in a previous RNAi-based study (Gonczy et al., 2000). In our experiments, the effects of RNAi on COL-19::GFP and DPY-7 collagen expression were re-examined by using feeding, soaking, and injection protocols. All three methods produced distinctive morphologic defects. A small (Sma) and not Dpy phenotype was observed, however, as nematodes viewed at the light microscope level were short but circumferentially relatively normal in size (data not shown). As well as being small, these nematodes contained numerous visible vesicles and lacked any discernible annuli and alae. RNAi injection was the most reliable method in producing this Sma phenotype and, therefore, was applied to introduce dsRNA into the TP12:kaIs12(col-19::gfp) strain. This procedure resulted in a severely mutant COL-19::GFP expression pattern. No fluorescent annuli or alae were detectable, and upon shifting the focal plane into the matrix of the cuticle, COL-19::GFP was localized in small hypodermal vesicles (Fig. 6B, arrowed), and additionally in a diffuse hypodermal nuclear excluded pattern (data not shown). This mutant COL-19::GFP pattern is consistent with a block in the transport of this cuticle collagen from the ER to the final matrix. The RNAi effect of ZK1098.5 on collagen expression was further examined by staining treated nematodes with the DPY-7 monoclonal antibody, revealing that the DPY-7 collagen is absent from these nematodes, a fact consistent with the notable lack of annuli and annular furrows in these animals (data not shown).


COL-19::GFP represents a specific marker to aid in the characterization of the synthesis and modification of the cuticular ECM components and the resulting adult exoskeleton structures in C. elegans. It also illustrates a structural basis for morphologic defects resulting from mutations in single collagens genes and their associated biosynthetic enzymes. This role is significantly demonstrated by the studies on the cuticle collagen mutant dpy-5 and the collagen modifying enzyme mutant dpy-11, and elucidates the probable cause of the morphology defects in these strains. The two closely related collagens (ROL-6 and C39E9.9) also examined in this study failed to express the GFP-tagged transgene, a fact that may relate to their earlier expression in the nematode developmental pathway having a detrimental effect on survival of the transgenic animals, or alternatively the processing and destruction of the GFP marker in these earlier stages. A similar explanation may account for the exclusive localization of the COL-19::GFP marker to the adult cuticle and its unexpected absence from the dauer cuticle, a stage that has been shown previously to express this gene (Liu et al., 1995).

The cuticle is an exoskeleton, composed predominately of numerous small collagen-like proteins that are secreted from the underlying hypodermis. Over 150 cuticle collagen genes are expressed in C. elegans (Johnstone, 2000), and the interactions formed between individual collagens to construct distinct structures, remains relatively uncharacterized. Nematode cuticle collagen genetics is, however, complex, with different mutations of the same collagen resulting in widely different phenotypes. Glycine substitution mutants are common and generally produce strong phenotypes in most collagens, excluding sqt-1. Likewise, mutations in the N-terminal procollagen processing site are found in several mutants, and together, indicate that the incorporation of mutant or improperly processed collagens into the final ECM is a common feature in nematode cuticle collagen-related phenotypes (Kramer, 1997). Null mutations in certain collagens such as sqt-1 are weak or wild-type, whereas nulls in others such as dpy-10 and dpy-13 have strong phenotypes (Kramer, 1997). These observations indicate that there is a high degree of redundancy and complexity between the cuticle collagens, their interacting partners, and the higher-order structure that they form.

The hypodermis of C. elegans is composed of four major longitudinal ridges (Fig. 1F) that are joined circumferentially by thin sheets of cytoplasm (Sulston and Horvitz, 1977). The main components of this tissue are the ventral and dorsal hypodermal syncytium (predominantly hyp7 derived) and to a lesser extent the smaller individual hypodermal cells in the head and tail (hyp3 to 6 in the head and hyp8 to 11 in the tail). The lateral hypodermis comprises the seam cells that fuse together at the last molt (L4-adult) to form a continuous band, and these in turn are overlaid by the adult-specific lateral alae (Sulston et al., 1983; Fig. 1E,F). Laser ablation studies of the seam cells in L4 larvae confirmed that they are responsible for the formation of the alae (Singh and Sulston, 1978). Studies of the dauer alae and seam cells also revealed that the lateral hypodermis is responsible for the diametric-shrinkage that occurs in this stage and, therefore, may actually be circumferentially contractile. Mutant nematodes that lack individual seam cells, and in turn have gaps in their ala, also have a greatly weakened cuticle, and the underlying hypodermis was found to instead resemble the dorsal and ventral hypodermis (Singh and Sulston, 1978). The lateral seam cells were also observed to be areas of intense biosynthetic activity and were packed with Golgi bodies before the cuticle molt (Singh and Sulston, 1978). Together with our observations of distinct regions corresponding to disrupted and regular COL-19::GFP expression patterns, these previous studies indicate that the lateral seam cell hypodermal syncytium is a highly specialized tissue that is morphologically and functionally distinct from the ventral–dorsal hypodermis. This finding also provides a convincing reason for the adult Dpy phenotypes that we detect with dpy-5(e61) and dpy-11(e224) mutants in that dumpiness is concurrent with the failure of the seam cell-derived cuticle to contract normally and may also be associated with branching or bifurcated alae. It is significant to note that both the dpy-5 and dpy-11 mutant alleles depicted have a relatively normal head morphology, an observation consistent with the fact that the lateral seam cell cords terminate posterior to the head, and the anterior region is composed of ventral–dorsal hypodermal cells (hyps 3–6). These mutants are also notably less resistant to physical stress, and dpy-11(e224) in particular regularly exploded during preparation for microscopic examination. dpy-11 was characterized recently (Ko and Chow, 2002) and was shown to encode a novel hypodermally expressed thioredoxin-like enzyme that is predicted to modify matrix proteins in this tissue. The loss-of-function dpy-11 mutant allele (e1180) and RNAi disruption leads to more severe morphologic phenotypes, further supporting the important role played by this enzyme (Ko and Chow, 2002).

A range of crosses was performed between the COL-19::GFP strain TP12:kaIs12(col-19::gfp) and selected morphologic mutants that together support the critical role played by individual collagens in the overall maintenance of the normal nematode body form. It may be hypothesized that the lateral seam cell hypodermis and its associated ECM are important in the circumferential contraction of the cuticle, whereas the ventral–dorsal hypodermal-derived ECM is responsible for the lateral contraction (Fig. 1F). This finding obviously represents a simplified model, and it should also be noted that the ventral–dorsal hypodermis is likewise responsible for circumferential contraction, as noted in the class II Dpy mutants. The dpy-7(e88) mutants for example, do not have normal circumferential annuli but do have a relatively normal COL-19::GFP expression in the seam cell-derived cuticle and in the overlying alae (Table 1; Fig. 4A). This finding again supports the existence of two discrete forms of hypodermis and overlying cuticle in C. elegans, with both forms playing distinct but overlapping roles in the maintenance of longitudinal and circumferential body morphology (Fig. 1F). A function for the annuli in permitting longitudinal body extension, while maintaining a constant diametric size under the high hydrostatic pressure, has also been hypothesized previously for the cuticular annuli in many parasitic nematodes (Lee, 1966).

In addition to highlighting the presence of the two distinct hypodermal components underlying the cuticle, the resulting phenotypes enabled us to define five distinct classes of COL-19::GFP disruption that ranged from wild-type/mild disruption to severe global disruption. We were able thus to group mutants together on the basis of the interruption patterns. It is significant to note that only a subset of all the morphologic mutants examined produced a disrupted pattern of COL-19::GFP expression (classes II–V) and that class I mutants such as lon-1, lon-2, lon-3, sma-2, rol-6, and bli-4 displayed relatively normal expression patterns. This fact supports the contention that the COL-19::GFP marker is not significantly contributing to the overall mutant phenotypes observed. It is, however, interesting to note that, where COL-19::GFP expression in the ala and annuli was mutant, a corresponding morphologic defect could be detected in these structures by Nomarski microscopy or scanning electron microscopy.

Our five categories of disruption grouped modifying enzyme mutants into a single group, namely, “class V; branching alae and annuli disruption.” This class was to some extent expected, because mutations in the enzymes involved in collagen biosynthesis would affect all collagens and, therefore, all regions of the cuticle. The dpy-11(e224) point mutant resulted in a less severe phenotype than dpy-11 null allele (e1180) for the cross-linking enzyme, thioredoxin, and the amorphous appearance of this null mutant suggests a lack of redundancy in this system. Visualization of differing degrees of severity of different alleles of a mutated gene encoding a biosynthetic enzyme will be a useful tool in characterizing first the alleles themselves and, furthermore, the mechanisms, roles/functions, and important functional domains of the enzymes they encode.

With the exception of the class V mutants, the other three categories of disruption are predominantly made up of cuticle collagen mutants and as yet uncharacterized mutants (Table 1). From this observation, it is probable that the uncharacterized mutants may additionally encode mutant collagens. The different classes of disruption possibly indicate distinct roles for the mutant collagens in the cuticle, and it may be hypothesized that collagens belonging to the same group will interact in the assembly of a common structure; for example, class II may represent collagens (DPY-7, DPY-2, DPY-3, DPY-8, and DPY-10) that interact together to form normal dorsal/ventral annuli, and class III mutants may likewise represent collagens (DPY-13, DPY-5, BLI-1) that interact to form the seam cell–derived cuticle. A direct correlation between the degree of disruption and the collagens that associate in the final cuticle remains to be firmly established; however, a recent study (McMahon et al., 2003) lends support to this contention. The class II collagens, dpy-2, dpy-8, dpy-7, and dpy-10, were all found to be expressed in the same early cyclical manner during the postembryonic life cycle, whereas the class III collagens, dpy-5 and dpy-13, are both found to have an intermediate larval transcript cycling pattern (Johnstone and Barry, 1996; McMahon et al., 2003), and finally, the class IV collagen, sqt-1, was found to peak in a temporal expression pattern that falls between the early (class II) and the intermediate (class III) collagens (Johnstone and Barry, 1996). The temporal control of collagen gene expression, therefore, may be the key factor that determines which collagens will interact to form specific structures within the cuticle. These interactions may exist between individual collagen chains but more probably occur between the preformed trimers and the higher-order structures that they ultimately form.

There is, however, no direct correlation between the classes of disruption and the predicted C. elegans procollagen structural groups (Johnstone, 2000); however, as stated above, the interactions may be occurring at the final assembly stage between preformed collagen trimers and their higher-order structures. Additionally, no obvious relationship between the nature of the mutation in the collagens and our defined classes could be detected. However, for some of the individual collagens tested (sqt-1 and sqt-3), the nature of the mutation did affect the resulting COL-19:GFP patterns, whereas for others it did not (dpy-7, dpy-13, and dpy-5). Collagens with glycine substitutions were present in class II dpy-7(e88) and dpy-2(e489 and sc38), and class IV sqt-3(sc63 and e2117). Whereas class I rol-6(su1006) and class IV sqt-1(e1350 and sc1) both contained examples of mutants in the N-terminal collagen-processing domain.

The group defined as having branched annuli overlying the seam cell cords (class III) also contains many cuticle collagens (Table 1) and clearly exemplifies the two distinct regions of hypodermally derived cuticle. The dorsal/ventral hypodermally derived annuli are present in this class of mutants, albeit in a highly constricted form, a fact that is significantly consistent with the Dpy phenotype of these animals. The wild-type nematodes express 7–8 annuli per 10 μM division (Fig. 1), giving each annulus an average periodicity of 1.3 μM, whereas in the class III mutants there are 10–13 annuli per 10 μM division (Fig. 3 and supplementary Fig. II), giving an average of periodicity of 0.75–1 μM per annulus. A similar observation was noted for the class I mutants sma-2 and lon-3, which generally expressed a normal COL-19::GFP profile but did have a correspondingly constricted (sma-2, 0.66 μM per annulus) or enlarged (lon-3, 1.6 μM per annulus) average annuli periodicity (supplementary data, Fig. I). The seam cell-derived cuticle in the class III mutants, however, is clearly disrupted. This observation confirms that this class of collagens are potentially more seam cell localized. An alternative hypothesis is that this seam cell area is more susceptible to mutations in the collagens that comprise it. Double staining with a hypodermal cell desmosome-specific antibody, MH27 (Francis and Waterston, 1985), on the TP14:dpy-5(e61)I:kaIs12(col-19::gfp) strain has revealed a relatively normal seam cell shape (supplementary Fig. II, F), thereby confirming that it is the overlying cuticle that is in fact mutant in this strain. The two molecularly defined class III collagen mutants examined, dpy-5(e61) and dpy-13(e458), represent a premature termination and a deletion mutant, respectively. It is significant to note, that RNAi of both these genes resulted in COL-19::GFP disruption patterns that were identical to the corresponding genetic mutants (supplementary Fig. II, G,H), thus confirming the important nonredundant function of these collagens in the maintenance of normal organismal morphology.

The mutants in the class IV category are predominantly members of the sqt collagen gene family, and it is interesting to note a very similar pattern between these mutants and the RNAi animals of the dual oxidase cross-linking enzymes. The sqt-1 alleles examined are mutants in the N-terminal procollagen processing site (e1350 and sc1) and also the C-terminal putative cross-linking domain (sc13); with the N-terminal protease mutants producing a more severe COL-19::GFP disruption pattern than the C-terminal cross-linking mutant; however, both patterns exhibit class IV characteristics. The RNAi disruption of sqt-1 together with the examination of heterozygotes TP26:sqt-1(e1350)II;kaIs12 (col-19::gfp) revealed little or no COL-19::GFP disruption and is consistent with the prediction that phenotype is caused by the normally redundant, larger mutant nonprocessed SQT-1 collagen being incorporated into the cuticle (Kramer et al., 1988; Yang and Kramer, 1994). A similar observation has been found for the sqt-3 mutant alleles examined (sc63 and e2117) both of which represent temperature-sensitive glycine substitutions; therefore, these mutant collagens will also be incorporated into the cuticle (Vanderkeyl et al., 1994). sqt-3 RNAi and heterozygotes in TP23:sqt-3(sc63)II;kaIs12(col-19::gfp) are practically wild-type with regard to COL-19::GFP expression, further supporting the conclusion that the incorporation of a mutant collagen into the ECM is the main reason for the severe Sqt phenotypes. In addition to this observation, larvae from sqt-3(e2117) embryos raised at 25°C have a structurally abnormal cuticle, lacking a medial striated layer (Priess and Hirsh, 1986). It is also feasible that the sqt-1 and sqt-3 mutant COL-19::GFP patterns may be due to abnormal collagen cross-linking caused by the accumulation of these structurally significant but mutant collagens.

The minor allele-specific differences noted between the sqt-1 mutants, however, was not a feature shared with the class II mutant, dpy-7. A glycine substitution allele dpy-7(e88), a null allele dpy-7(qm63), and the dpy-7-specific RNAi all resulted in identical amorphous COL-19::GFP disruption patterns in the annuli; with all nematodes displaying a smooth featureless surface, devoid of annulations and having only lateral alae, when viewed by Nomarski. These observations support an important role for this cuticle collagen in the maintenance of normal body morphology. The allele-specific effects of a small proportion of the nematode cuticle collagens have been addressed in this study, revealing that a range of mutations in individual collagens can result in a complex array of disruption patterns, and validates the COL-19::GFP marker to be effective tool to dissect the interactions of individual collagens in the construction of this complex cuticular ECM.

The C. elegans genome has been sequenced completely (C. elegans genome Consortium, 1998), and an RNAi-based gene disruption approach in combination with the COL-19::GFP strain additionally may prove to be a useful tool in the identification of components of the cuticle collagen biosynthesis, modification, and assembly pathways. Examples of both early events (Bet3) and late effecter components (duox) in collagen biosynthesis are presented in this study. The C. elegans ZK1098.5 gene is an orthologue (36.4% identical and 60% similar) of the yeast secretory protein Bet3, a 22-kDa hydrophilic protein that is involved in both ER to Golgi and post-Golgi transport (Rossi et al., 1995). Yeast mutants for this protein were shown to have elaborated ER membranes and large accumulations of both small and large vesicles when viewed at the electron micrograph (EM) level (Rossi et al., 1995). The accumulation of vesicles is consistent with our ZK1098.5 RNAi observations on the effect of COL-19::GFP secretion, which likewise was localized in vesicles and is not incorporated into the cuticular ECM (Fig. 6B). The overall effect of the resulting mutant collagen expression led to a small phenotype in these animals, perhaps consistent with a reduction in the amount of secreted cuticular components. The duox genes were found to be important for normal cuticle assembly, and their disruption led to visible blistering of the cuticle, as well as a general Dpy phenotype. These RNAi-induced mutants were also shown to lack the important tyrosine-based cross-links common in wild-type cuticles and, therefore, are implicated in the late-stage cross-linking and assembly of this ECM (Edens et al., 2001). Our RNAi experiments also confirmed these observations and provided a structural basis for the observed gross morphologic defects, as the COL-19::GFP expression was severely disrupted in these mutants, having a random disorganized appearance, illustrating the aberrant assembly of the protein. Hence, this marker can help in characterizing both early stage secretory mutants and later stage cross-linking mutants in the assembly of the ECM components.

Collagens are found in all metazoan phyla and constitute the major structural proteins in extracellular matrices. The COL-19::GFP collagen marker will provide an invaluable tool for the future dissection of the ECM collagen biosynthetic pathway in C. elegans and may help further elucidate the relationship between genes and their effects on adult body morphology. This approach is potentially valuable in the characterization of these essential processes in higher vertebrates, and several important human diseases, including osteogenesis imperfecta and Ehlers-Danlos type VI and VIIC, which are direct consequences of mutations in collagens and their processing enzymes, respectively (Myllyharju and Kivirikko, 2001). A detailed understanding of collagen biosynthesis in nematodes additionally may lead to the identification of components with significance in parasitic nematode vaccine and chemotherapeutic intervention.


Nematode Strains

Wild-type (Bristol N2), CB769(bli-1), CB767(bli-3), CB937(bli-4), CB518(bli-5), CB1370(daf-2), CB489(dpy-2), BE38(dpy-2), CB1359(dpy-2), CB27(dpy-3), CB1166(dpy-4), CB61(dpy-5), CB88(dpy-7), CB130(dpy-8), CB12(dpy-9), CB128(dpy-10), CB224(dpy-11), CB1180(dpy-11), CB458(dpy-13), CB364(dpy-18), CB164(dpy-17), CB185(lon-1), CB678(lon-2), MT3847(lon-2), CB4123(lon-3), HE1006(rol-6), CB502(sma-2), CB1350(sqt-1), BE1(sqt-1), BE13(sqt-1), BE3(sqt-2), BE108(sqt-2), BE63(sqt-3), and CB4121(sqt-3), strains were received from the C. elegans Genetics Center. The dpy-7 null strain MQ375 was received from Jonathan Ewbank. Nematodes were maintained on standard OP50 seeded agar plates at 20°C unless stated otherwise.

Construction and Integration of Collagen GFP Translational Fusions

The promoter and the full-length coding sequences of the group I (Johnstone, 2000) collagen genes ZK1193.1 (col-19), C39e9.9, and T01b7.7 (rol-6) were cloned “in frame” to the green fluorescent protein GFP-C3 (Crameri et al., 1996) in a pSP65 backbone (Promega). The following primers were used to generate products with compatible restriction sites to the GFP expression vector: col-19, col19-A 5′-cgcctgcagCATTTGAAAATTTGCACC-3′ (PstI underlined) and col19-B 5′-ggggtaccGCCTTGTAAGCTGCACG-3′ (KpnI underlined); C39e9.9, C39e9.9-A 5′-cgcgcatgcTTCAAATGCCTGGCTTCATCG-3′ (SphI underlined) and C39e9.9-B 5′-cgcggtaccAAGACAAGACGTGCCGAAC-3′ (KpnI underlined); rol-6, T01b7.7-A 5′-gcgctccagAACAGGTGAATCTGAACCTCC-3′ (PstI underlined) and T01b7.7-B 5′-cgcggtaccAATTGGAATTTGCGATGACG-3′ (KpnI underlined). Amplified genomic products were digested and ligated into similarly cut vectors and transformed into XL10 competent cells. Miniprep DNA was either microinjected at 20 μg/ml with pBluescript carrier DNA (100 μg/ml) or in combination with 100 μg/ml marker DNA pRF4 (rol-6 su1006). Transgenic lines were selected based on GFP expression for the col-19 construct or by the “rolling” phenotype of the coexpressed pRF4 marker for C39E9.9 and T01B7.7. Several lines expressing the COL-19::GFP extrachromosomal array were integrated into the genome after irradiation (38 Gy) from a 60Co source. A GFP-expressing strain TP12:kaIs12(col-19::gfp) was then selected and backcrossed 4× against wild-type nematodes to remove deleterious mutations. Males were induced from the TP12:kaIs12(col-19::gfp) strain after heat shock treatment of L4s (33°C for 8 hr) and selection of male progeny. Male TP12:kaIs12(col-19::gfp) nematodes were crossed into the strains listed above. Refer to Table 1 for strain designation, e.g., CB61(dpy-5e61) × TP12:kaIs12(col-19::gfp) males to produce TP14:dpy- 5(e61)I;kaIs12(col-19::gfp).

Microscopy and Imaging

Live nematodes were mounted on 2% agarose pads and immobilized in a solution of M9 salts supplemented with 0.01% sodium azide. Samples were then viewed under Nomarski optics or epifluorescence with a Zeiss Axioskop 2 microscope, and images were captured on a CCD Hamamatsu digital camera, pseudocolored, and processed by using Improvision Openlab and Adobe Photoshop software.

Transmission and Scanning Electron Microscopy.

Nematodes were washed in PBS before fixation in 2.5% glutaraldehyde, 1% paraformaldehyde in PBS. After extensive washes, samples were fixed in 1% osmium tetroxide, and dehydrated in ethanol (transmission EM, or TEM) or acetone (SEM). For TEM, samples were embedded in epoxy spurs resin, sectioned, and viewed on a LEO 902 TEM at 80kV. For SEM, samples were processed in a critical-point drier, mounted, and viewed in a Phillips 500 SEM.


TP12:kaIs12(col-19::gfp) nematodes were collected from agar plates and washed extensively in ice-cold PBS. Adults were then pipetted onto poly-L-Lysine–coated slides and permeabilized by freeze-cracking (Rogalski et al., 1993). Samples were blocked in PBST (PBS, 0.1% Tween-20) containing 1% dried skimmed milk, then probed with monoclonal anti-DPY-7 (1/50) or MH27 (1/100) then incubated in a mixture of 1/200 Alexa Fluor 594 goat anti-mouse (Molecular Probes). Samples were washed and viewed by epifluorescence on a Zeiss Axioskop 2 microscope, and images were captured as described.


The effects of col-19 disruption on wild-type, CB1370(daf-2), and TP12: kaIs12(col-19::gfp) strains was examined by using standard RNAi feeding, soaking, and injection protocols (Winter and Page, 2000; Timmons et al., 2001). The col-19 coding sequence was cloned from C. elegans mixed stage cDNA by polymerase chain reaction (PCR) using the primers col19F 5′-ATGGGCAAGCTCATTGTGGTTG-3′ and col19R 5′-TTAAGCCTTGTAAGCTGCACG-3′. The PCR product was cloned into pGEM, digested with NotI and NcoI, and ligated into similarly cut L4440 vector (Timmons et al., 2001). Double-stranded RNA for RNAi injection and soaking was produced in vitro as described previously (Winter and Page, 2000). Fifteen to 20 young adults from wild-type, TP12:kaIs12(col- 19::gfp), and CB1370 strains were microinjected (0.4 mg/ml dsRNA), allowed to recover overnight, transferred singly to fresh plates, and progeny were scored. For soaking experiments, L3–L4s from the above strains were incubated at 20 or 25°C for 1–3 days in a solution of dsRNA (0.4 mg/ml) before plating out onto fresh plates. To perform bacterially mediated RNAi, the construct L4440 was transformed into HT115 (DE3) cells and plated according to published methods (Timmons et al., 2001). Ten L4 animals from the strains N2, TP12:kaIs12 (col-19::gfp), and CB1370 were transferred to feeding plates, incubated for 2 days at 20 or 25°C, adults were then transferred to fresh plates and allowed to egg-lay for 24 hr when progeny were scored. The RNAi effects of selected cuticle collagen genes were observed in TP12:kaIs12(col-19::gfp) following the feeding protocol outlined above. The dpy-7 and dpy-13 genes were cloned from genomic DNA, into the NotI/HindIII site of L4440 vector, after an intermediate cloning stage in the PCR transfer vector PCRScript, by using the following primers: dpy-7, dpy7fwd 5′-cgcggtaccATGGAGGAGCCCAGTTCGGGGGCC-3′ and dpy7rev 5′-cgcggtaccTTCTTTCCATAACCACCACCAGAA-3′; dpy-13, dpy13fwd 5′-cgcggtaccATGGACATTGACACTAAAATCAAG-3′ and dpy13rev 5′-cgcggtaccCGGCGAGTTCCGTCCTCGAAGAAG-3′. The sqt-1 and sqt-3 genes were cloned from mixed stage cDNA into the NotI/XhoI sites of L4440, after an intermediate cloning stage in the PCR transfer vector PCRScript, by using the following primer combinations: sqt-1, sqt1fwd 5′-ccgcggccgcATGTCTGTAAAACTTGCGTG-3′ and sqt1rev 5′-gcctcgagTTAGATATTTCTGTATCCACG-3′; sqt-3, sqt3fwd 5′-cggcggccgcATGGGAAACTGACGGTAGGCTC-3′; and sqt3rev 5′-gcctcgagTTATCGTCTGGTTCCGTCCTCGAAG-3′. The dpy-5 RNAi feeding clone was constructed as described previously (McMahon et al., 2003). The RNAi effects of ZK1098.5 (Bet3) were observed in TP12:kaIs12 (col-19::gfp) and wild-type nematodes by means of the feeding, soaking, and injection protocols outlined above. RNAi-treated nematodes were co-stained with DPY-7 as described above. The ZK1098.5 gene was cloned into pGEMT (Promega) from C. elegans genomic DNA by PCR with the following primer pairs: ZK1098.5F 5′-CCTTAAAATCGGTCTCTTTCGTT-3′ and ZK1098.5R 5′-AATGAGAAAATTTGAACAAGCCA-3′. The 1,663-bp insert was then digested with NotI and NcoI and ligated into a similarly digested L4440 vector, before transforming into HT115 cells. RNAi of Duox encoding genes F53G12.3 and F56C11.1 were performed by the above feeding protocols using the chromosome I RNAi library clones produced by Dr. J. Ahringer and distributed through the UK HGMP Resource Centre, Cambridge.


The authors thank the C. elegans Genetics Center for providing the nematode strains used in this study and Dr. Julie Ahringer (Cambridge) for providing the duox feeding clones. Thanks also to Prof. Ann Rose (Vancouver) for communicating unpublished results regarding the identity of the DPY-5 collagen. A.P.P. received a Senior Fellowship in Basic Biomedical Research from the Medical Research Council of Great Britain.