Epithelial invagination is an important developmental process, fundamental to gastrulation and organ formation. Invagination can involve cell movement, changes in cell shape, and changes in cell-cell or cell-extracellular matrix adhesion. We are interested in learning what molecules are involved in this morphogenetic process and are using C. elegans vulval formation as a model system. The hermaphrodite vulva is a tube that connects the uterus to the outer epithelium and is formed by the invagination of a specialized set of epithelial cells during the third and fourth larval stages of development. We screened for mutants with a defective vulval invagination but normal vulval cell lineages and designed the screen to allow isolation of mutations that cause sterility. From over 12,000 haploid genomes screened, we isolated 26 mutations, which we placed into eight complementation groups. Seven appear to define new genes,
sqv-1 to 7 (squashed vulva), and alleles of the eighth fail to complement
spe-2(
mn63). The strongest alleles of all eight genes result in identical vulval phenotypes: a wild-type division pattern and detachment from the cuticle, but indistinct separation between the anterior and posterior halves of the invagination from the late L3 stage onward. We found no mutations that caused any other abnormal vulval phenotype without also affecting vulval cell lineages. We have cloned
sqv-3 and
spe-2, both of which mapped to areas sequenced by the genome project; we have entered their positions into the Acedb database.
sqv-3 is predicted to encode a homolog of mammalian
beta-1,4 galactosyltransferase, an enzyme localized to the trans Golgi, where it adds galactose to proteins, many of which are then transported to the cell surface. We have so far been unable to demonstrate that SQV-3 itself has galactosyltransferase activity. The predicted SPE-2 protein has similarities to two classes of human expressed sequences from the WashU-Merck EST project but has no similarity to any protein of known function. Each of our eight genes involved in vulval formation has alleles that additionally cause hermaphrodites to be self-sterile. Homozygous mutants generate eggs that have egg-shells but fail to hatch (and probably even to divide). Genetic evidence suggests that
spe-2 hermaphrodites have inviable sperm but viable oocytes: they can have homozygous
spe-2 progeny if crossed with
spe-2/+ males[1].
sqv-1 to 7 hermaphrodites, on the other hand, cannot have progeny even if mated with wild-type males. We have so far shown, using the technique of artificial insemination[2], that
sqv-3 hermaphrodites do have viable sperm, implying that their oocytes or gonad must be defective. It is therefore possible that SPE-2 may be required in sperm and SQV-3 (or a SQV protein glycosylated by SQV-3) in another cell with which the sperm must interact (e.g. the oocyte) for fertilization to proceed beyond the cue for egg-shell formation. Alternatively, it is possible that SPE-2 must simply be provided to the zygote via sperm and that SQV-3 (or a SQV protein glycosylated by SQV-3) must be provided via the oocyte and that both are required for an early embryonic process. We are currently trying to define the infertility defect caused by these genes more carefully and are hopeful that information about this defect might also have implications for how the genes are affecting vulval invagination. We have raised antibodies against both SQV-3 and SPE-2 and, after affinity-purification, have used them to stain whole worms. In each case, we have so far been unable to see staining of wild-type worms but do see reproducible staining patterns in worms overexpressing either SQV-3 or SPE-2 by means of an integrated array containing the appropriate minimal rescuing fragment. Our preliminary results indicate that SQV-3 may be expressed in subcellular dots (Golgi perhaps) in the intestinal cells, possibly spermatheca, head mesodermal cell, and vulva. The
sqv-3 fertility defect may therefore result from a need for expression in the intestine, which produces yolk and possibly other proteins that are packaged into oocytes. Our preliminary results concerning SPE-2 indicate that it may be expressed in subcellular dots in many cells in the head and pharynx, in the tail, and, as expected, in adult sperm. We have not yet seen SPE-2 staining in the vulva. [1] Sigurdson et al. (1984). Genetics 108: 331. [2] LaMunyon and Ward (1994). Genetics 138: 689.