sli-1 is a negative regulator of
let-23-mediated vulval induction and encodes a protein similar to the mammalian proto-oncoprotein c-Cbl. The
sli-1 locus was isolated in screens that identified suppressors of the vulvaless (Vul) phenotype associated with reduction-of-function (hypomorphic) mutations in
let-23, a C. elegans homolog of the mammalian EGF receptor. Reduction-of-function mutations in
sli-1 alone are silent in the vulva. LET-60 Ras acts downstream in the signal transduction process initiated by LET-23. We have determined that
sli-1 does not measurably suppress the Vul phenotype associated with severe, non-null
let-60(rf) mutations. In contrast,
sli-1(rf) does strongly suppress the Vul phenotype associated with a severe, non-null
sem-5(rf) mutation. SEM-5 is a Grb-2-like adaptor that acts downstream of LET-23 RTK and upstream of LET-60 Ras. Furthermore, we have shown that SLI-1 can interact with SEM-5 in yeast two-hybrid assays; specifically, the C-terminal proline-rich domain of SLI-1, but not the N-terminal domain, interacts with SEM-5. Minigene constructs of the full-length
sli-1 cDNA driven by heatshock promoters have wild-type
sli-1 function in vulval signalling in germline rescue experiments. Using the heatshock minigene constructs, we have also identified domains of SLI-1 that can confer partial wildtype SLI-1 activity. In fact, N-terminal domain of SLI-1, which lacks the C-terminal proline-rich domain, has wild-type SLI-1 activity in the vulva.
sli-1(rf) mutation restores vulval signalling to near wild-type levels in
let-23(
sy97). The molcular lesion associated with
sy97 deletes from LET-23 the tyrosine phosphoylation sites that are the canonical SEM-5 binding sites. Although
sli-1(rf) also restores vulval signalling to wild-type levels in the severe hypomorph
sem-5(
n1619), there are significant differences in the suppression of
sem-5 in comparison to that of
let-23. We suggest the following model of SLI-1 action: SLI-1's N-terminal domain normally acts by preventing SEM-5's SH2 domain from binding non-canonical phospho-tyrosines in LET-23. In the absence of SLI-1, such sites are again accessible to SEM-5 and signalling can resume even in the absence of the canonical SEM-5 binding sites such as in LET-23sy97. These "cryptic" phospho-tyrosines may serve to bind other adaptor proteins which can substitute for SEM-5's vulval signalling when both SLI-1 and SEM-5 are absent; however, not all functions of SEM-5 can be replaced by these other adaptor proteins since the suppression of
let-23 and
sem-5 by
sli-1 are different. Finally, SLI-1's proline-rich C-terminal domain may bind SEM-5's SH3 domains in the wild-type RTK signlling complex and help to localize SEM-5's binding to the canonical phospho-tyrosines in LET-23. This data may have implications for the function of mammalian c-Cbl and the oncogenic v-Cbl. We are currently testing
sli-1/c-cbl chimeric constructs for their effects on vulval induction.