NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Riddle DL, Blumenthal T, Meyer BJ, et al., editors. C. elegans II. 2nd edition. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1997.

Cover of C. elegans II

C. elegans II. 2nd edition.

Show details

Section IICuticle

The cuticle is the animal's exoskeleton and is important for maintenance of morphology, protection from and/or interaction with the external environment, and motility. The cuticle is connected to the hypodermis via hemidesmosomes. Filaments extend from body wall muscles through the basement membrane and connect to the hypodermis (Francis and Waterston 1985; Francis and Waterston 1991). It is through these connections that the force of muscle contraction is transmitted to the cuticle. Nematodes do not have opposing muscles. The elasticity of the cuticle and the animal's high internal hydrostatic pressure provide the restorative force that allows the animal to straighten after contracting muscles on one side of the body (like a water-filled balloon). The cuticle is a highly complex extracellular structure, presumably due to the many functions it must perform.

A. Cuticle Structure

The cuticle has both surface specializations and internal layers that can differ at different developmental stages. Protruding ridges, termed alae, form over the left and right lateral rows of hypodermal seam cells. The L1-, dauer-, and adult-stage alae have distinct structures; however, alae are not present on L2, L3, or L4 cuticles (Cox et al. 1981b). Ablation of seam cells with a laser microbeam causes gaps in the overlying alae, indicating that seam cells are responsible for formation of the alae (Singh and Sulston 1978). The surfaces of all cuticles have narrow circumferential indentations, uniformly spaced about 1 μm apart, that define rings called annuli (Figs. 1 and 2). Annuli run continuously around the animal but are absent over the lateral hypodermal cords. Annuli may function like pleats, allowing the cuticle to fold on the inner radius of a bend and extend over the outer radius.

Figure 1. Scanning electron micrograph of adult C.

Figure 1

Scanning electron micrograph of adult C. elegans cuticle. On the left of the figure is the epicuticle/cortical layer, and an annulus is indicated. On the right, the cortical layer has been cracked off, revealing (more...)

Figure 2. Freeze-etch electron micrograph of adult cuticle.

Figure 2

Freeze-etch electron micrograph of adult cuticle. The cortical layer (C), strut (S), fiber layer (F), basal layer (B), and hypodermis (H) are indicated. An annular indentation can be seen in the (more...)

The ultrastructure of the adult cuticle has been most carefully characterized. Six layers (epicuticle, external cortical, internal cortical, medial, fiber, and basal) have been identified in the adult cuticle, although not all layers are well defined using any one fixation/staining technique (Figs. 1 and 2) (Cox et al. 1981a,b; Peixoto and de Souza 1992). In addition, a loosely associated, carbohydrate-rich surface coat external to the epicuticle can be detected using specific fixation and staining methods (Zuckerman et al. 1979; Blaxter et al. 1992; Blaxter 1993a). The structure of the epicuticle is not well understood. It is trilaminar in appearance (two electron-dense layers separated by an electron-lucent layer), and evidence from studies on other nematodes indicates that it contains lipid. The epicuticle of C. elegans can be fractured into two faces, suggesting a bilayer structure, but its properties are distinct from those of plasma membranes (Blaxter 1993a; Maizels et al. 1993; Peixoto and de Souza 1994).

The medial layer is composed of filamentous columns of material, termed struts, that connect the cortical and fiber layers. Struts spaced about 0.4 μm apart form two rows located on either side of the annular indentations (Fig. 1). A smaller number of struts are found scattered between the annular indentations. The space between the struts is presumably filled with fluid. The fiber layer contains two highly organized layers of fibers that spiral around the animal in opposite directions. Each layer is oriented at about 65o from the long axis, with the outer layer running counterclockwise and the inner layer running clockwise relative to the animal's tail-to-head axis.

L1-, dauer-, and adult-stage cuticles are structurally distinct and differ from the L2, L3, and L4 cuticles, which are very similar in structure (Cox et al. 1981b). The cuticles from all stages have epicuticle, external, and internal cortical layers, although the structure of the layers can differ between stages. A distinct medial layer with struts is only apparent in the adult cuticle. The L2–L4 and adult cuticles have similar-appearing fiber layers. In place of a fiber layer, the L1 and dauer larva cuticles have a striated layer characterized by darkly staining bands of about 18 nm periodicity. The striations in the dauer larva cuticle are broader and more distinct than those in the L1. Glancing sections through the dauer larva cuticle show that the striated layer is composed of interwoven orthogonal fibers or laminae (Popham and Webster 1978; Cox et al. 1981b). The dauer larva cuticle is especially thick, accounting for 10.2% of the animals' cross-sectional area versus 4.4% for the cuticle of other stages.

B. Molting

At the end of each larval developmental stage, nematodes undergo a molt in which a new cuticle is formed and the old cuticle is shed (Singh and Sulston 1978). At molts, animals enter a period of lethargus lasting approximately 2 hours, during which pharyngeal pumping and movement are suppressed. Two to four hours preceding lethargus, the cytoplasm of the lateral seam cells accumulates densely packed Golgi bodies. At the beginning of lethargus, connections between the hypodermis and cuticle are broken and a new cuticle begins forming. During the second half of lethargus, animals frequently spin or flip around their long axis. About 30 minutes before ecdysis (shedding of the old cuticle), the posterior bulb of the pharynx begins twitching and granules accumulate in the g1 pharyngeal gland cell bodies (Hall and Hedgecock 1991). Just preceding ecdysis, the granules are secreted and the pharynx begins spasmodic contractions. The cuticle lining the pharynx breaks, the old cuticle distends around the head, and the animal pulls back repeatedly to dislodge the cuticle remaining in the pharynx. The animal pushes with its head until the old cuticle breaks and then crawls out of the remainder of the old cuticle.

Circumferential microfilament bundles form in the hypodermis at larval molts and disperse between molts (B. Draper and J. Priess, pers. comm.; J. Kramer, unpubl.). These filament bundles appear to assume structural functions of the cuticle during molting and may have an important role in defining cuticle structure. During the second half of embryogenesis, similar bundles of circumferential microfilaments and microtubules form immediately under the apical plasma membrane of the hypodermis (Priess and Hirsh 1986). These filament bundles are involved in elongation of the embryo, and they define the location of the circumferential indentations of the cuticle that delimit annuli. Just prior to hatching, the filaments disperse, and maintenance of morphology shifts to the newly formed cuticle.

C. Cuticle Composition

In accord with its role in protection from the external environment, the cuticle is highly resistant to solubilization. The standard method for preparation of cuticles is to boil sonicated animals in 1% SDS. This treatment solubilizes essentially all other structures, leaving cuticles as the insoluble residue. Cuticles prepared in this manner retain their original ultrastructure remarkably well (Cox et al. 1981a). Treatment of the SDS-insoluble material with reducing agent solubilizes 70–75% of adult and L4-, 44% of L1-, and 26% of dauer-stage cuticle proteins (Cox et al. 1981b). Cuticle proteins are extensively cross-linked with disulfide bonds and also with nonreducible tyrosine-derived cross-links (see below). Solubilized cuticle proteins can be separated into about eight major and numerous minor molecular weight species on SDS-PAGE. The estimated molecular weights of the soluble cuticle proteins range from 53,000 to more than 200,000, with the majority being 90,000 and greater. Overlapping, but distinct, molecular weight forms are found in extracts from cuticles of different stages.

Most of the soluble cuticle proteins are digested by treatment with bacterial collagenase, indicating that they are collagenous in nature. The high glycine and imino acid content of adult soluble cuticle proteins (26% glycine, 11% proline, 12% hydroxyproline) is also consistent with a largely collagenous nature (Cox et al. 1981a,b; Ouazana and Herbage 1981; Ouazana et al. 1984). The insoluble cuticle proteins have lower, although still substantial, glycine and hydroxyproline content (22% glycine, 13% proline, 7.5% hydroxyproline), indicating the presence of less collagenous material. A small amount of hydroxylysine, which is a common constituent of vertebrate collagens, is found in dauer larva cuticles but not at other stages. The soluble cuticle proteins exhibit two unusual properties that have also been noted for vertebrate collagens; they run at different apparent molecular weights on gels of different polyacrylamide concentration, and they stain pink with Coomassie brilliant blue R-250. Both the soluble and insoluble cuticle fractions contain a small amount of carbohydrate, about 1% by weight (Cox et al. 1981c).

D. Cuticle Collagen Structure and Regulation

1. Gene Family

The collagens that constitute the major elements of the cuticle are encoded by an unusually large gene family estimated to contain between 50 and 150 genes (Cox et al. 1984). This estimate is based on the ratio of the number of collagen to actin-hybridizing clones in C. elegans genomic phage libraries, given that there are four actin genes in the genome. The genes are distributed throughout the genome, with multiple members mapping to each chromosome (Cox et al. 1985). In general, the genes are dispersed, but there are examples of two or more collagen genes in close proximity (Park and Kramer 1990; Bird 1992; Levy et al. 1993). Collagen genes in close proximity usually show strong sequence similarity, as would be expected if they arose by gene duplication. However, not all gene pairs that have high sequence similarity are in close proximity.

2. Protein Structure

Complete genomic DNA sequences for more than 30 cuticle collagen genes have been determined (Kramer et al. 1982, 1988, 1990; von Mende et al. 1988; Cox et al. 1989; Park and Kramer 1990; Bird 1992; Johnstone et al. 1992; Levy and Kramer 1993; Levy et al. 1993), and the genome sequencing project is adding to this number rapidly (Waterston and Sulston 1995). In general, the genes are small (<2 kb) and have just one to three short introns. The collagen chains they are predicted to encode range from 26 kD to 35 kD, with one exceptional gene that is predicted to encode a 107-kD product. All of the cuticle collagen chains have similar domain structures and several conserved motifs (Fig. 3). There is a long amino non-Gly-X-Y domain that is of variable length, a central Gly-X-Y repeat domain, and a variable length carboxyl non-Gly-X-Y domain.

Figure 3. General structure of C.

Figure 3

General structure of C. elegans cuticle collagens. The domain structure is derived from the sequences of 38 collagen genes. (Hatched boxes) Gly-X-Y domain; (more...)

Starting near the amino terminus are four short sequence motifs that are conserved in most or all cuticle collagens, named homology blocks D-A (HBD-A) in the amino to carboxyl direction (Fig. 4). HBD is very hydrophobic and is located within the predicted signal peptide. HBD is highly conserved in the col-1 and col-6 cysteine subfamilies (see below), but only weakly conserved in others. HBC is located six amino acids to the carboxyl side of HBD, and in most cases, it spans the predicted site for signal peptidase cleavage. HBB follows six amino acids after HBC, and it contains a conserved tryptophan, the only tryptophan in many of the cuticle collagens. About one half of the 38 cuticle collagen genes considered here (Fig. 3) have an intron between positions 6 and 7 of HBB. This is the only obviously conserved intron in the entire cuticle collagen gene family. HBA is located 19–44 amino acids after HBB, and it contains highly conserved arginine residues that constitute an endoproteolytic processing site (see below). The sequences between homology blocks do not have strongly conserved residues However, since the spacing between HBD, HBC, and HBB is conserved, these blocks may constitute a single functional unit in the cuticle collagens. Mutations in HBC and HBA have been shown to affect collagen function (Kramer and Johnson 1993; Levy et al. 1993; Yang and Kramer 1994).

Figure 4. Cuticle collagen homology block (HB) consensus sequences.

Figure 4

Cuticle collagen homology block (HB) consensus sequences. Consensus sequences were derived from 30 collagen gene sequences. The percentage of collagens that have a particular amino acid or (more...)

The region from HBA to the start of the Gly-X-Y repeat domain is highly variable in length (15–318 amino acids) and sequence. Closely preceding the Gly-X-Y domain, there are three cysteines in all but one collagen which has two cysteines. The Gly-X-Y repeat domain is broken into two major sections. The first section generally contains 10 repeats and is followed by an interruption of 10–21 amino acids that contains two or three cysteines. The second Gly-X-Y section contains about 40 repeats and can have one to three interruptions of from two to eight amino acids each. Closely following the end of the Gly-X-Y repeat domain are two cysteine residues in all cuticle collagens. The remainder of the carboxyl non-Gly-X-Y domain is short (9–19 amino acids) and conserved in the col-1 and col-6 subfamilies, but quite variable in length (13–393 amino acids) and sequence in other collagens.

The cuticle collagens can be divided into subfamilies based on the spacing of the cysteine residues that flank the Gly-X-Y domain (Fig. 5). Currently, there are nine cysteine subfamilies, but the number is likely to increase as more sequences are generated. Collagens in the same subfamilies generally have more sequence similarity to each other than to collagens in other subfamilies. The cysteine spacings are likely to be important for directing the formation of disulfide bonds between appropriate molecules in the cuticle. Whether cuticle collagen molecules form from three identical chains (homotrimeric) or a mix of nonidentical chains (heterotrimeric) has not been determined. Both homotrimeric and heterotrimeric collagens have been described in vertebrates, and it is possible that the cuticle contains both homotrimeric and heterotrimeric collagens.

Figure 5. Cuticle collagen cysteine subfamilies.

Figure 5

Cuticle collagen cysteine subfamilies. The amino acid sequences immediately amino and carboxyl to the Gly-X-Y repeat domain (written vertically in the figure) are shown. Also shown are the (more...)

Cuticle collagen gene sequences have been examined in several other nematode species. These collagens have all of the same conserved domain structures and sequence motifs that are found in C. elegans. One complete and two partial cuticle collagen sequences from the sheep parasite Haemonchus contortus have all of the conserved aspects of the col-1 subfamily (Shamansky et al. 1989; Cox 1990; Cox et al. 1990). Partial sequences of two cuticle collagen genes from the pig parasite Ascaris suum (Kingston et al. 1989; Kingston and Pettitt 1990) and the complete sequence of one gene from the human parasite Brugia malayi (Scott et al. 1995) fit in the col-6 subfamily. The complete sequence of a cuticle collagen from the root knot nematode, Meloidogyne incognita, places it in the col-8 subfamily (Vandereycken et al. 1994). On the basis of the numbers of Gly-X-Y hybridizing bands on genomic Southern blots, the size of the cuticle collagen gene families may be smaller in these parasitic nematodes than in C. elegans. However, the strong conservation of cuticle collagen structure among these diverse nematodes suggests that the roles of collagens in cuticle function have been conserved throughout the phylum.

3. Gene Expression

Synthesis of cuticle proteins occurs at high rates during molts and at lower rates between molts (Cox et al. 1981c). The levels of some cuticle collagen transcripts follow this same pattern (Park and Kramer 1994). Different members of the cuticle collagen gene family are expressed at different developmental stages. Two-dimensional gel analyses of 3H-proline-labeled in vitro translation products of RNA from the L4-adult- and L2d-dauer-stage molts identified at least 60 distinct collagenase-sensitive products in the 37,000 to 52,000 range (Politz and Edgar 1984). These apparent molecular weights are probably higher than expected (26,000–35,000) due to the fact that collagens migrate abnormally on SDS-PAGE (Furthmayr and Timpl 1971; Freytag et al. 1979). The number of spots on the gels is clearly an underestimate of the number of collagens expressed at these two molts, since different gene products may comigrate and a large class of collagens with more basic pIs would not have been detected (J. Kramer, unpubl.). Of the 32 collagenous products from L4 and 29 from L2d-dauer-stage molt RNA, only 3 were found in both molts. RNAs isolated early versus late in the L4 molt generate different sets of collagenase-sensitive products (Politz and Edgar 1984). Thus, there are qualitative and quantitative differences in collagen expression both between molts and at different times during a molt.

Expression of 20 cuticle collagen genes has been examined using several different techniques: hybridization of genomic collagen clones with labeled cDNA generated from poly(A) RNAs isolated from molting animals (Cox and Hirsh 1985), Northern hybridization with gene-specific probes (Kramer et al. 1985; Park and Kramer 1990, 1994), analysis of lacZ reporter constructs in transgenic animals (Liu and Ambros 1991; Liu et al. 1995), and reverse transcriptase–polymerase chain reaction (RT-PCR) (Levy and Kramer 1993). The results of these studies are summarized in a nonquantitative manner in Table 1. Nine different expression patterns are indicated in the table, but this is certainly an underestimate since most studies have not looked at RNA from all stages or from different times within a molt or between molts. Additionally, there are large quantitative differences in expression between different collagen genes and for the same gene at different stages. Further complexity arises from studies of collagen promoter-lacZ reporters that show expression in different subsets of hypodermal nuclei (Liu and Ambros 1991; Liu et al. 1995; I. Johnstone, pers. comm.; J. Kramer, unpubl.). The combination of temporal, quantitative, and spatial expression controls with the large number of cuticle collagen genes makes for a potentially bewildering level of complexity.

Table 1. Patterns of C. elegans cuticle collagen gene expression.

Table 1

Patterns of C. elegans cuticle collagen gene expression.

4. Nonreducible Cross-links

The primary translation products of the cuticle collagens have, with one exception, predicted molecular masses of 26–35 kD, but the collagens extracted from cuticles have apparent molecular weights of 53,000 and larger. Some of this discrepancy is due to the abnormal migration of collagens on SDS-PAGE. However, the major cause is the presence of nonreducible covalent cross-links between collagen chains. Vertebrate collagens contain nonreducible cross-links that form between modified lysine residues. In contrast, the cross-links that have been identified in Ascaris (Fujimoto 1975; Fujimoto et al. 1981) and C. elegans (D. Eyre and J. Kramer, unpubl.) cuticles are di-, tri-, and/or isotrityrosine residues. In Ascaris cuticles, isotrityrosine is primarily found in cuticle collagens, and di- and trityrosine in cuticlin. Cuticlin is defined as the insoluble cuticle residue that remains after extraction with reducing agents. After incubation of Ascaris in 3H-labeled tyrosine-containing medium, 20% of tyrosine incorporated into the cuticle was in the form of dityrosine and 6% was isotrityrosine (Fetterer et al. 1993). The rate of formation of dityrosine was greater in the β-mercaptoethanol-insoluble (cuticlin) fraction of the cuticle and that of isotrityrosine was greater in the soluble fraction, consistent with the distribution of the cross-links noted above. Formation of the cross-links was inhibited by several peroxidase inhibitors, suggesting the involvement of a peroxidase in their formation. Indeed, formation of dityrosine cross-links in the sea urchin fertilization envelope has been shown to be dependent on oxidation of tyrosine residues by ovoperoxidase (Deits et al. 1984).

During molting in C. elegans, low-molecular-weight (38,000–52,000) collagenase-sensitive proteins can be detected on Western blots using antisera produced against high-molecular-weight (53,000 to greater than 200 kD) adult cuticle proteins (Politz et al. 1986), suggesting conversion of low- to high-molecular-mass products. Western blots of cuticle extracts probed with antiserum specific for the SQT-1 cuticle collagen show that low-molecular-weight forms of SQT-1 are most abundant at molts and are replaced by higher-molecular-weight forms following the molt (J. Yang and J. Kramer, unpubl.). A cross-link containing tryptic peptide was isolated from C. elegans cuticles, and its sequence was found to be identical to the carboxyl end of COL-2 ( col-1 cysteine subfamily in Fig. 5), with the tyrosine located between the conserved cysteines involved in the cross-link (D. Eyre and J. Kramer, unpubl.). Generally, a single conserved tyrosine residue is found in the carboxyl non-Gly-X-Y domain of most cuticle collagens, and these are likely to be involved in cross-linking. Tyrosine residues in the amino non-Gly-X-Y domain must also be involved in cross-linking to account for the very high-molecular-weight multimers that form in the cuticle.

Rotary shadowing of collagens extracted from Ascaris cuticle under reducing conditions primarily shows individual 47-nm-long molecules (Betschart and Wyss 1990). The dimensions of these molecules match those expected for triple-helical molecules formed from products of the known cuticle collagen genes. Extraction of Ascaris cuticle under nonreducing, nondenaturing conditions results in long fibers that appear to be chains of triple-helical molecules. These results suggest that isotrityrosine cross-links primarily form between the three chains within a collagen molecule and that multiple molecules are linked end-to-end via disulfide bonds. The formation of higher-molecular-weight nonreducible material would result from further tyrosine cross-link formation occurring after the intramolecular cross-links had formed.

E. Cuticle Genetics

A large number of mutations have been identified that cause gross changes in the animal's overall morphology and may result from defects in cuticle function (Brenner 1974; Higgins and Hirsh 1977; Cox et al. 1980; Hosono 1980; Hosono et al. 1982; Kusch and Edgar 1986). These mutations can cause a variety of phenotypes, including blister (Bli), blisters on cuticle; dumpy (Dpy), short and fat; long (Lon), long and thin; left roller (LRol), twisted in a left-handed helix; and right roller (RRol), twisted in a right-handed helix (Figs. 6 and 7). Additionally, mutations that cause dominant Rol and recessive Dpy phenotypes have been termed squat (Sqt). Combinations of these phenotypes can occur, e.g., dumpy and left roller (DLR), and the severities of the phenotypes can vary widely. There are currently 6 bli genes, 27 dpy genes, 3 lon genes, 6 rol genes, and 3 sqt genes. Five of the dpy genes are involved in X chromosome dosage compensation (see Meyer, this volume) and only affect cuticle function indirectly. The dpy gene designation can also be misleading since many dpy genes have alleles that result in Rol and/or DpyRol, or Dpy, phenotypes. There is no evidence for lineage changes in these mutants, and their phenotypes appear to result from altered cell shape and position. In roller animals (Fig. 7), the cuticle and all of the internal organs are helically twisted (Higgins and Hirsh 1977). Roller animals rotate around their long axes and tend to move in circles. The fact that these severely abnormal animals can reproduce is testimony to the value of using a self-fertilizing hermaphroditic species for genetic studies, since most of these phenotypes render males incapable of mating.

Figure 6. Nomarski micrographs of adult wild-type and mutant animals illustrating morphological phenotypes.

Figure 6

Nomarski micrographs of adult wild-type and mutant animals illustrating morphological phenotypes. (a) Wild-type N2 strain animal; (b) dpy-10(e128) animal, Dpy phenotype; (more...)

Figure 7. Rhodamine phalloidin-stained RRol mutant animal, rol-6(e187).

Figure 7

Rhodamine phalloidin-stained RRol mutant animal, rol-6(e187). Helical twisting of the four muscle quandrants in this L2 animal is evident. Magnification, 315×. (more...)

Mutations in many of these genes only show phenotypes at particular developmental stages, and in some cases, the phenotype of a single allele can differ at different stages. For example, sqt-1(e1350) heterozygotes are RRol at all stages from L2 to adult, whereas homozygotes are wild type at L2, weak RRol at L3, weak Dpy at L4, and Dpy as adults (Cox et al. 1980; Park and Kramer 1994). Passage through the dauer larva stage can also affect the phenotype of some sqt-1 and sqt-2 alleles; e1350 homozygotes are variable RRol at L2d and dauer stages, as are adults that develop from dauer larvae. The phenotype is considered variable because individual animals range from strong Rol to apparently wild type. This stage specificity of phenotypes is likely to result from stage-specific expression of the particular gene as well as the influence of other gene products expressed at the same stages.

F. Mutations in Cuticle Collagens

Several of the genes that are involved in determination of overall morphology have been shown to encode cuticle collagens (Table 2): sqt-1 (Kramer et al. 1988; Kramer and Johnson 1993), dpy-13 (von Mende et al. 1988), rol-6 (Kramer et al. 1990; Kramer and Johnson 1993), dpy-7 (Johnstone et al. 1992), dpy-2 (Levy et al. 1993), dpy-10 (Levy et al. 1993), and sqt-3 (Vanderkeyl et al. 1994). Mutations in these cuticle collagens can generate all of the morphological phenotypes noted above and can additionally cause abnormal hermaphrodite tail morphology (Tal). The range of phenotypes indicates that these collagens can have different functions in the cuticle and that different mutations in a single collagen can alter its function in different ways.

Table 2. Adult phenotypes of molecularly characterized cuticle collagen mutations.

Table 2

Adult phenotypes of molecularly characterized cuticle collagen mutations.

Collagen chains with amino acid substitutions can aberrantly participate in the assembly pathway and interfere with normal collagen processing and assembly. For this reason, null mutations are critical for determining the function of a collagen. Null mutations in sqt-1 and rol-6 cause very weak phenotypes, Tal or very weak Dpy. Thus, the absence of these collagens has only a minor effect on morphology, even though the presence of abnormal SQT-1 or ROL-6 can produce severe morphological abnormalities. In contrast, null mutants of dpy-10 or dpy-13 have strong phenotypes, DLR or Dpy, demonstrating that these collagens are required for normal morphology.

In all organisms and collagens examined, the vast majority of collagen missense mutations are substitutions of glycine residues in the Gly-X-Y repeat domains. Commonly, glycine substitutions inhibit triple-helix formation, resulting in abnormal modification and degradation of much of the mutant collagen chain, as well as other chains associated with it (Prockop and Kivirikko 1995). As a result, a reduced amount of abnormal collagen is secreted into the matrix. Glycine substitutions cause weak phenotypes in sqt-1 but strong phenotypes in other genes. In the three genes that have both null and glycine substitution mutations, the phenotypes of both types of mutations are generally similar, possibly reflecting severe reduction in collagen level in both cases. However, the glycine substitution phenotypes are slightly more severe than the null phenotypes (weak Lon and weak LRol phenotypes for sqt-1 , and a more severe DLR in dpy-10 ). Since glycine substitutions result in more severe phenotypes than null mutations, the abnormal collagens must interfere with the function of other molecules involved in cuticle synthesis, assembly, or structure.

1. Mutations Affecting Homology Blocks A through C

Substitutions for the highly conserved arginine residues at homology block A positions 2 and 5 (see Fig. 4) have been identified in sqt-1 , rol-6 , and dpy-10 (Kramer and Johnson 1993; Levy et al. 1993). Replacement of arginine with cysteine in sqt-1 causes a dominant RRol, recessive Dpy phenotype, whereas the equivalent dpy-10 mutant is dominant LRol, recessive DLR. The arginine to cysteine substitution in rol-6 is semidominant RRol, whereas the arginine to histidine substitution is recessive RRol. The fact that the same mutation that causes RRol in sqt-1 and rol-6 causes the opposite, LRol, phenotype in dpy-10 suggests that these collagens function in a mirror image manner. The two fiber layers in the cuticle are mirror image structures that spiral around the animal in opposite directions. Possibly, SQT-1 and ROL-6 are localized in one of the fiber layers and DPY-10 in the other.

Transgenic analyses of in-vitro-generated sqt-1 and rol-6 mutations indicate that arginine or lysine is required at positions 2 and 5 of HBA for normal collagen function (Yang and Kramer 1994). The spacing of the conserved arginine residues in HBA suggests that they could form the cleavage site for a subtilisin-like endoproteinase (see bli-4 below). Western blot analyses using SQT-1-specific antisera show that HBA mutant forms of SQT-1 are larger than wild type by the amount expected if cleavage normally occurs at HBA and that sequences amino-terminal to HBA are retained in mutant but not wild-type SQT-1 (J. Yang and J. Kramer, unpubl.). These results indicate that cuticle collagens are synthesized as procollagens that are endoproteolytically processed at HBA to remove the amino-terminal pro-domain during their maturation. HBA mutant collagens retain the pro-domain, and this could interfere with their further processing and assembly into higher-order structures. Many, although not all, vertebrate collagens are also proteolytically processed during their maturation; however, the use of a subtilisin-like protease for this purpose is unique to nematodes. The inability to remove the amino pro-domain of vertebrate type I collagen causes dermatosparaxis, a fragile skin disease in which incorporation of the abnormal chains inhibits formation of collagen fibers (Smith et al. 1992).

Three ethylmethanesulfonate (EMS)-induced recessive LRol mutations of sqt-1 are substitutions of the first of the two conserved carboxyl domain cysteines (Figs. 3 and 5) with tyrosine or serine. Analyses of in-vitro-generated mutations show that replacement of either of these two cysteine residues with serine in sqt-1 or rol-6 causes an LRol phenotype, although the phenotype is less severe for rol-6 (Yang and Kramer 1994). Assembly of some vertebrate collagens requires disulfide bonding between the carboxyl domains of the three chains. Replacement of both carboxyl domain cysteine residues with serine in sqt-1 also results in an LRol phenotype. Since this is a non-null phenotype, carboxyl domain disulfide bonding is not essential for assembly of SQT-1, although it is necessary for normal SQT-1 function. Western blot analyses of cuticle extracts from sqt-1 LRol mutants show that nonreducible cross-link formation is severely inhibited (J. Yang and J. Kramer, unpubl.). A tyrosine immediately precedes the first carboxyl cysteine in both SQT-1 and ROL-6 (Fig. 5). Apparently, loss of the ability to form the adjacent disulfide bond inhibits cross-link formation at this tyrosine residue.

An unusual mutation has been identified in homology block C (Fig. 4) of dpy-10 (Levy et al. 1993). The dpy-10(m481m482) mutation was generated by a Tc1 excision (see Plasterk and van Leunen, this volume) that resulted in the deletion of an isoleucine codon at position 5 of HBC (Levy et al. 1993). This allele causes no apparent abnormal phenotype, but it fails to complement other dpy-10 alleles. Loss of the isoleucine residue may interfere with signal peptide cleavage, which is predicted to occur on the amino side of the deleted isoleucine residue. This mutant must provide enough dpy-10 function such that homozygotes are normal, but not enough function to complement more severe dpy-10 alleles.

2. Discordance of Expression and Phenotype in sqt-1 and rol-6

As noted above, sqt-1 and rol-6 mutants can exhibit abnormal phenotypes at L2-adult and at L2d and dauer stages. Both sqt-1 and rol-6 transcripts are detected at each of the molts preceding these stages, except for the dauer stage (Park and Kramer 1994). Dauer larvae show strong Rol phenotypes, but no expression of the genes is detected after completion of the L1-L2d molt. Why then do dauer larvae exhibit the mutant phenotype? A likely explanation is that the hypodermal filament bundles that form at each molt (see above) “lock” the hypodermis into whatever shape it has at the beginning of the molt. Since L2d animals are Rol, the hypodermis would be locked into a helical configuration at the L2d-dauer molt. The twisted hypodermis would synthesize a wild-type dauer cuticle that has the form of the underlying twisted hypodermis. Thus, the phenotype of the dauer larva is derived from the pattern created in the preceding stage. Support for this notion comes from rhodamine-phalloidin staining of sqt-1 and rol-6 mutants, showing that the hypodermis remains twisted throughout the entire L2d-dauer molt period.

G. Interactions between Cuticle Collagen Genes

Collagens generally do not function independently, but form complexes of increasing size and complexity during their maturation. The presence of an abnormal collagen chain can disturb the structure of the complex in subtle and unpredictable ways. As a result of their biochemical properties, collagens have somewhat unusual genetic properties (Cox et al. 1980; Kusch and Edgar 1986; Kramer and Johnson 1993). Collagen mutations are frequently dominant due to disruption of the complex in which they are a component. Genetic interactions between collagen genes, such as intergenic suppression or enhancement of phenotypes, occur frequently because multiple collagens are components of the same complex structure.

LRol or glycine substitution mutations of sqt-1 are dominant suppressors of rol-6 RRol alleles (Kramer and Johnson 1993), and rol-6 LRol mutations can suppress sqt-1 RRol phenotypes (Yang and Kramer 1994). Double mutants for RRol alleles of both genes are RRol, and thus no suppression occurs between these alleles. ROL-6 requires SQT-1 to function, since rol-6 phenotypes are suppressed in the sqt-1 null background. However, SQT-1 does not require ROL-6, since sqt-1 phenotypes are apparent in the rol-6 null background. These interactions suggest that SQT-1 and ROL-6 interact, possibly forming a single heterotrimeric collagen molecule. Recalling that the null phenotypes for both genes is nearly wild type, it is possible that suppression occurs by removal of the abnormal collagen chain(s). When certain mutations in the two collagen chains are combined, the end result may be complete loss of both collagens (mutual suicide), resulting in the null phenotype.

Mutant dpy-10(e128) animals have a Dpy phenotype that can be converted to DpyLRol (the more severe Dpy-10 phenotype) by the addition of a single copy of any non-null sqt-1 mutation (Kusch and Edgar 1986; Kramer and Johnson 1993). This enhancement of the Dpy-10 phenotype occurs with recessive LRol, dominant RRol, and recessive Lon alleles of sqt-1 . Enhancement does not occur with a sqt-1 null allele, showing that the effect is due to the presence of abnormal SQT-1 and not to the absence of normal SQT-1. Alleles of the dpy-2 and dpy-7 collagens also show enhancement by sqt-1 , but dpy-13 does not.

H. Surface Molecules

The cuticle surface of nematodes is a major focus of studies on the interactions of parasites with their hosts (Maizels et al. 1993). The C. elegans cuticle surface is being studied as a model for parasites, as well as for its roles in cuticle function (Politz and Philipp 1992). Cuticle surface molecules are discussed in detail by Blaxter and Bird (this volume).

I. Other Cuticle-related Genes

1. bli-4

The bli-4 gene was originally identified by a single allele, e937, that results in blistering (Bli) of the adult cuticle (Brenner 1974). Eleven more alleles were subsequently identified that cause early larval lethality (Peters et al. 1991). Two of the lethal (Let) alleles fail to complement the other Let alleles, but they do complement the Bli allele. The noncomplementing alleles arrest in late embryogenesis, and the complementing alleles arrest as L1 larvae. The complementing alleles may retain some bli-4 function, resulting in later arrest and the ability to complement the Bli allele. The bli-4 gene encodes a protein with strong similarity to the Kex2/subtilisin family of serine endoproteases that are involved in processing prohormones and other precursor proteins (Thacker et al. 1995). Alternative splicing generates three BLI-4 variants that share common amino-terminal sequences, but differ in their carboxyl termini. Two of the noncomplementing lethal alleles affect common exons present in all transcripts, whereas the Bli allele, e937, deletes an exon found in only one BLI-4 variant. This BLI-4 variant may only be required for proper function of the adult cuticle. As noted above, cuticle collagen homology block A appears to be a cleavage site for a subtilisin-like proteinase. It is possible that one or more of the bli-4 gene products are involved in cleavage of procollagens at HBA.

2. rol-3

The rol-3 gene was originally identified by a single allele, e754, that caused an adult LRol phenotype. Subsequently, 12 further rol-3 alleles were shown to be recessive larval lethals (Barbazuk et al. 1994). One temperature-sensitive mutant is lethal at high temperature but is a viable, weak LRol at low temperature. The temperature-sensitive period for rol-3 was shown to be from mid-L1 to mid-L3, well before formation of the adult cuticle. The LRol phenotype appears only in adult-stage animals, not in L4 heterochronic mutants that have adult cuticles, making it unlikely that rol-3 encodes an adult-specific cuticle component. Suppressors of the rol-3 temperature-sensitive lethality were localized to two genes, srl-1 and srl-2 (suppressor of roller lethal). Some rol-3;srl double mutants show abnormal development in the posterior region of the animal. The srl hermaphrodites show no obvious abnormalities, but males have abnormal tail morphology. It is possible that these genes are involved in morphogenesis of the posterior region of the animal (Barbazuk et al. 1994).

3. dpy-5

Mutations in dpy-5 result in a Dpy phenotype in animals from L2 to adult stages, and mutant animals have altered cuticle structure (Ouazana et al. 1985). The dpy-5 gene encodes a novel protein of approximately 25 kD. DPY-5 has a good predicted signal peptide and the carboxyl half of the protein is cysteine-rich. Whether DPY-5 is a cuticle component or affects cuticle structure indirectly is unknown.

4. dpy-20

Alleles of dpy-20 result in Dpy phenotypes of differing severities and can also cause a rounded-head phenotype (Hosono et al. 1982; Clark et al. 1995). The dpy-20 gene encodes a novel protein with no signal peptide, making it unlikely that it is a component of the cuticle. dpy-20 mRNA is most abundant in L2–L4 animals, corresponding to its temperature-sensitive period around the L2 stage (Clark et al. 1995). Whether dpy-20 has any role in cuticle function is not clear.

5. Cuticlin Genes

The material that remains insoluble after extraction of Ascaris cuticles with 0.5 M NaCl, followed by 1% β-mercaptoethanol at 37°C, was originally termed cuticlin (Fujimoto and Kanaya 1973). The C. elegans cuticle material that remains insoluble after boiling in 1–2% SDS, 5% β-mercaptoethanol has also been called cuticlin, although there may be differences between this material and Ascaris cuticlin. Two C. elegans genes, cut-1 and cut-2 , that encode proteins in the cuticlin fraction have been identified (Sebastiano et al. 1991; Lassandro et al. 1994). The cut-1 gene encodes a novel protein of about 40 kD that is cysteine- and tyrosine-rich. Transcripts are detected in animals undergoing dauer larva formation. Antiserum directed against CUT-1 detects a 40-kD protein on Western blots of animals that are forming dauer larvae, but CUT-1 becomes insoluble in the mature dauer cuticle. Immunofluorescence localized CUT-1 to a 2-μm-wide band underlying the dauer alae.

The cut-2 gene encodes a novel 231-amino-acid protein that is also tyrosine-rich. CUT-2 contains 13 variable-length repeats that begin with AAP(A/V/I) and have a tyrosine present in most of the repeats. Similar repeats have been seen in vitelline membrane, chorion, and larval cuticle proteins of insects. Transcripts are detectable in RNA from animals of all stages, but they may only be produced around the time of molts. Anti-CUT-2 antibodies react with the insoluble cuticle residue of all stages. Immunogold localization shows that CUT-1 and CUT-2 are in the cortical layer of cuticles in all stages and are also localized under the dauer alae (Ristoratore et al. 1994; Favre et al. 1995). Upon incubation with horseradish peroxidase and H2O2, recombinant CUT-2, but not CUT-1, is efficiently cross-linked via dityrosine residues. The structure of CUT-2 may promote the formation of tyrosine cross-links, resulting in its insolubility.

Copyright © 1997, Cold Spring Harbor Laboratory Press.
Bookshelf ID: NBK20029

Views

  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...