• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

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 IIIBasement Membranes

Basement membranes can be recognized as thin sheets of extracellular matrix material closely associated with cell membranes. Studies in vertebrates indicate that basement membranes have roles in cell adhesion, migration, and differentiation, as well as acting as molecular filters, barriers to cell migration, and mechanical supports. In C. elegans, basement membranes cover the pseudocoelomic faces of the hypodermis, pharynx, intestine, gonad, and some body wall muscles (White et al. 1976). The basement membrane covering the pharynx is thicker, 45 nm, than that found on other tissues, 20 nm (Albertson and Thomson 1976). There are generally few distinguishing features of basement membranes, but 30-nm striations have been noted in glancing sections (White et al. 1976), and thickening of the hypodermal basement membrane can be seen at sites of muscle cell attachment (Francis and Waterston 1985; Francis and Waterston 1991). The major constituents of vertebrate basement membranes are type IV collagen, laminin, nidogen, and proteoglycans (Yurchenco and Schittny 1990; Paulsson 1992). With the exception of nidogen, genes encoding all of these molecules have been identified in C. elegans.

A. Type IV Collagen

Type IV collagen is generally the most abundant constituent of basement membranes and is found only in basement membranes. The structure of type IV collagen has been conserved from sponges to humans. It has a short, non-triple-helical amino-terminal domain that contains several cysteines, a long Gly-X-Y repeat domain with numerous small interruptions, and a carboxy-terminal globular domain (NC1) with 12 highly conserved cysteines. Type IV collagen forms a polygonal network that is stabilized by disulfide-bonded NC1 domain dimers and amino-terminal domain tetramers, and poorly understood lateral interactions. There are at least six type IV collagen genes in mammals (Hudson et al. 1993). The ubiquitous form of type IV collagen, found in essentially all basement membranes, is a heterotrimer of two α1 and one α2(IV) chains. The α3, α4, and α5(IV) collagen chains are primarily localized to the kidney glomerular basement membrane, and mutations in these chains can cause the degenerative kidney disease Alport syndrome.

Two genes that encode type IV collagen chains, emb-9 and let-2 (previously called clb-2 and clb-1 , respectively) have been characterized in C. elegans (Guo and Kramer 1989; Guo et al. 1991; Sibley et al. 1993, 1994). They have strong similarity to the human type IV chains, being most conserved in the NC1 domain, where emb-9 and let-2 are 63% and 72% identical to the human α1 and α2 chains, respectively. Sequences of α2(IV) collagen genes have also been determined from the parasitic nematodes Ascaris (Pettitt and Kingston 1991) and Brugia (Caulagi and Rajan 1995), and they have 74% amino acid sequence identity with C. elegans over the entire protein, 87% in the NC1 domain. In mammals, the α1/α2, α3/α4, and α5/α6 type IV gene pairs are in head-to-head orientation, separated by short promoter regions, allowing transcriptional control of gene pairs encoding chains that form a single type IV molecule. The C. elegans genes, emb-9 and let-2 , are located on different chromosomes, III and X, respectively, ruling out such a control mechanism.

1. Type IV Collagen Genetics

Alleles of emb-9 and let-2 were originally identified at high frequency in genetic screens for embryonic lethal mutations (Meneely and Herman Herman, 1981; Miwa et al. 1980; Wood et al. 1980; Cassada et al. 1981; Isnenghi et al. 1983). Multiple alleles of both genes cause embryonic lethality, demonstrating that normal type IV collagen is required for embryogenesis. Most alleles are temperature-sensitive, such that animals raised at 15°C develop normally, but at 25°C (20°C in some cases), they die during embryogenesis. At intermediate temperatures, animals often arrest during larval development or become sterile adults. The strongest alleles convey nonconditional lethal phenotypes. Arrest occurs at about the twofold stage of embryogenesis and is accompanied by extensive herniation of the hypodermis and disorganization of body wall muscle. When mutant animals are shifted to 25°C as larvae, they arrest development or become subfertile adults, indicating that type IV collagen function is required throughout the life cycle.

Five alleles of emb-9 (Guo et al. 1991; M. Gupta and J. Kramer, unpubl.) and 15 of let-2 (Sibley et al. 1994) are substitutions of other amino acids for glycines in the Gly-X-Y repeat domain. Each type IV collagen chain normally contains about 20 interruptions in its Gly-X-Y repeat domain, but these “novel interruptions” are not compatible with normal function. Substitutions for different glycine residues can result in phenotypes of varying severity, ranging from embryonic lethality at 25°C, but not at 20°C, to unconditional lethality. The let-2(mn114) mutation is a glycine to glutamic acid substitution in the third Gly-X-Y repeat of the α2(IV) chain. It has a strong affect on adult fertility but only a mild affect on embryonic and larval development, suggesting that it affects gonad function preferentially. The let-2(mn126) mutation is an alanine to threonine substitution in the X position of a Gly-X-Y repeat immediately following a four-amino-acid interruption, and it causes a relatively severe phenotype. Substitutions for X- or Y-position amino acids are extremely rare in collagens. Such an alteration is not likely to have a major effect on triple-helix folding or stability, but it may define a region involved in interaction between type IV molecules or between type IV collagen and other basement membrane components.

Alleles of emb-9 and let-2 display temperature-sensitive dominant effects (P. Graham et al., unpubl.) that generally result in reduced viability and/or fertility in heterozygous animals. Some alleles, such as emb-9(g23) and let-2(mn103), are in fact temperature-sensitive dominant lethals. Two putative null alleles of emb-9 , a 500-bp deletion and a nonsense mutation, have been generated by reversion of the emb-9(g23) temperature-sensitive dominant lethal phenotype (M. Gupta and J. Kramer, unpubl.). The putative null alleles result in recessive nonconditional embryonic lethality, but they allow animals to develop further in embryogenesis, to about the threefold stage. Since missense mutations cause more severe phenotypes than the null mutations, the presence of abnormal type IV collagen must interfere with the function of other basement membrane components.

The distribution of mutations in the type IV collagen genes is nonrandom. There are 940 potential target Gly-X-Y glycines in the two genes. Yet, among the 20 independently isolated type IV mutations, there are four cases in which two mutations affect the same glycine. This clustering could result if substitutions for most glycines do not cause obvious phenotypes, and would therefore not have been identified in genetic screens, or if they cause dominant lethality or sterility, making them impossible to maintain. Since the existing alleles show strong dominant effects, the latter explanation seems most likely.

An unusual feature of the let-2 α2(IV) collagen gene is its complex pattern of interallelic complementation (Meneely and Herman 1981). Every let-2 allele complements at least one other allele to some extent, such that trans heterozygote animals have milder phenotypes than either homozygote. If each C. elegans type IV molecule contains one α2(IV) chain, as in vertebrate type IV, then complementation between different α2 chain alleles can only result from some type of intermolecular interaction. This interaction may be between type IV collagen molecules or between type IV collagen and other basement membrane components. A correlation exists between the degree of interallelic complementation for two alleles and the physical distance between the mutations (Sibley et al. 1994). Alleles must be at least 47 amino acids apart to show any complementation and at least 134 amino acids apart to show full complementation. The relationship between interallelic complementation and distance may reflect lateral interactions between C. elegans type IV collagen molecules.

2. Alternative Splicing of let-2

Transcripts of let-2 are alternatively spliced into two forms that contain either exon 9 or exon 10, but not both (Sibley et al. 1993). Exons 9 and 10 are separated by just 30 bp and have unusual 5′-splice donor sequences. Exons 9 and 10 appear to be duplicates, having the general structure (Gly-X-Y)5__9- or 10-amino-acid interruption __ (Gly-X-Y)4. The interruption is nine amino acids long in exon 9, and ten amino acids long in exon 10. The ratio of exon-9-containing to exon-10-containing transcripts changes dramatically during development. Exon-9-containing transcripts are 90% of the total in embryos, decline to about 30% within 1 hour after hatching, and slowly decline during larval development to a level of about 10% in adults. Transcripts of the Ascaris suum α2(IV) collagen gene are also alternatively spliced at exactly the same sites (Pettitt and Kingston 1994). Exon 9 is more highly conserved between the two nematodes (81% amino acid identity) than is exon 10 (65% identity). The interruption of exon 9 is 100% conserved, whereas the interruption of exon 10 is only 70% conserved. A similar temporal difference in expression of exon-9- versus exon-10-containing transcripts is seen in Ascaris and C. elegans.

The fact that alternative splicing of α2(IV) collagen has been maintained between these distantly related nematodes suggests that it may have an important role in basement membrane function. The shift from primarily the exon-9- to primarily the exon-10-containing variant of α2(IV) coincides with a dramatic shift in C. elegans development, from a phase of rapid and extensive morphological change (embryogenesis) to a phase that primarily entails symmetrical growth (early larval through adult stages). The morphological changes of embryogenesis may require basement membranes with properties different from those present in larvae or adults.

B. Proteoglycan

Proteoglycans consist of a protein core to which at least one glycosaminoglycan chain is attached. In vertebrates, there is a large family of proteoglycans that can have different core proteins and/or different attached glycosaminoglycan chains. The C. elegans unc-52 gene was shown to encode a homolog of the mammalian basement membrane heparan sulfate proteoglycan, perlecan (Rogalski et al. 1993). Perlecan is a common component of basement membranes and has been shown to interact with other basement membrane molecules as well as cell surface receptors. Three 65-kD heparan sulfate side chains are attached to domain I of mammalian perlecan. The C. elegans protein has no similarity to the mammalian protein in this region, but it does have two potential carbohydrate attachment sites. The C. elegans and mammalian proteins are very similar in domains II–IV. Domain II has three LDL receptor-like repeats, domain III has similarity to laminin, and domain IV contains 14 immunoglobulin C2-like repeats. Several variant forms of UNC-52 result from alternatively spliced transcripts that are missing one or more exons from domains II–IV. Antibody staining shows that UNC-52 is localized to basement membranes underlying body wall and anal muscles and surrounding the pharynx and gonad (G. Mullen and D. Moerman, pers. comm.).

Three classes of unc-52 alleles have been identified: class-1 mutants are viable and develop progressive paralysis, the class-2 mutant arrests at the twofold stage of embryogenesis and is paralyzed (Pat phenotype) (Williams and Waterston 1994), and the class-3 mutant arrests at the twofold stage but is not paralyzed. Paralysis of class-1 alleles is due to fracture of muscle-dense bodies and separation of the myofilament lattice from the plasma membrane (Waterston et al. 1980). Class-1 mutations also cause structural defects of the somatic gonad, possibly affecting the myoepithelial sheath cells (Gilchrist and Moerman 1992). The somatic gonadal defect of class-1 alleles is complemented by the class-3 allele, but it is more severe in animals heteroallelic for class-1 and class-2 alleles.

The unc-52 class-1 alleles are all localized to alternatively spliced exons in domain IV, including four nonsense mutations, one splice donor mutation, and one Tc1 insertion (Rogalski et al. 1995). Some of the alternatively spliced unc-52 transcripts would be unaffected by these mutations. Thirteen intragenic revertants of class-1 mutations (Gilchrist and Moerman 1992) were found to alter the splice acceptor sites of these same alternatively spliced exons. Thus, removal of the mutated exon by altered splicing can result in intragenic suppression. The class-3 allele has a transposon insertion in domain II. The class-2 allele is a nonsense mutation in a domain III exon present in all unc-52 transcripts. Thus, the Pat phenotype of the class-2 allele represents the complete loss of function for unc-52 .

Dominant intergenic suppressors of the Unc phenotype of unc-52 class-1 alleles have been identified (Gilchrist and Moerman 1992). Five of the suppressor mutations were mapped to a single gene, sup-38 IV. They are strong suppressors of the muscle defects but not of the gonadal defects. By themselves, the suppressor mutations cause mild muscle defects and variable gonadal defects. Putative sup-38 null mutations were generated by reverting the dominant suppressor activity and were found to cause maternal-effect lethality. The progeny of sup-38 null mutant hermaphrodites generally die during late larval development but have normal muscle structure. The function of sup-38 is therefore not required for normal muscle formation.

C. SPARC

SPARC (also known as BM-40 or osteonectin) is an anti-adhesive extracelluar matrix-associated glycoprotein that can modulate the interaction of cells with the matrix (Lane and Sage 1994). A SPARC homolog has been identified in C. elegans that has 38% amino acid sequence identity to mammalian SPARC (Schwarzbauer and Spencer 1993). Bacterial fusion proteins containing the amino- and carboxy-terminal domains of C. elegans SPARC bind calcium, as do the equivalent domains of mammalian SPARC. The C. elegans SPARC also contains intrachain disulfide bonds with properties similar to those in mammalian SPARC (Schwarzbauer et al. 1994). Staining of transgenic strains carrying a SPARC-lacZ fusion indicates that SPARC is expressed from late embryogenesis to adulthood and is restricted to body wall and sex muscles. Transgenic overexpression of wild-type C. elegans SPARC causes several abnormal phenotypes (Schwarzbauer and Spencer 1993). Animals develop an Unc phenotype as adults and in some cases become completely paralyzed in the posterior. The morphology of the vulva is abnormal, and sometimes the intestine or gonad protrudes through the vulva. Coinjection of the SPARC gene with the rol-6(su1006) RRol gene produced Unc, but not RRol, transgenic animals, suggesting that overexpression of SPARC can suppress the Rol phenotype. Injections with high concentrations of SPARC DNA resulted in no transgenic offspring, suggesting that strong overexpression of the gene may be lethal. Overexpression of SPARC may interfere with the normal interactions of muscle cells with their associated basement membranes.

D. Laminin

In mammals, the laminins are a family of basement membrane molecules composed of three disulfide-bonded subunits, αβγ (Tryggvason 1993; Timpl and Brown 1994). Association of different variant subunits generates at least seven distinct laminin molecules. Laminins can polymerize noncovalently and can bind to cell surface receptors. C. elegans homologs of the mammalian α and β laminin chains have been identified (K. Joh et al., pers. comm.). The α chain is encoded by the genetic locus epi-1 and the β chain is encoded by lam-1 . Mutations in these genes can cause rupturing of basement membranes, migration defects, and body wall muscle disorganization.

The unc-6 gene encodes a small laminin-related protein that is required for proper dorsal and ventral migrations of axons and mesodermal cells (Ishii et al. 1992). Closely related molecules named netrins have been identified in vertebrates where they have also been shown to promote axonal outgrowth (Serafini et al. 1994). For a complete description of the unc-6 gene, see Hedgecock et al. (this volume).

Copyright © 1997, Cold Spring Harbor Laboratory Press.
Bookshelf ID: NBK20211
PubReader format: click here to try

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...