daf-4 ,
sma-2 ,
sma-3 ,and
sma-4 mutants show two phenotypes in common: small body size (Sma) and specific male tail ray and spicule defects (Mab).
daf-4 has recently been shown to encode a serine/threonine kinase receptor that binds human BMP-2 ,a TGF- -like ligand (Estevez et al., Nature 365: 644). The striking similarity in mutant phenotypes therefore implicates
sma-2 ,
sma-3 ,and
sma-4 in this TGF- -ike signal transduction pathway. We wondered whether other known genes could also be involved in TGF- -like signalling and whether they could be used to place
daf-4 ,
sma-2 ,
sma-3 ,and
sma-4 into an ordered genetic pathway. We began by examining the phenotypes of double mutants between these four genes and
lon-1 ,
lon-2 ,and
lon-3 .The Lon phenotype of
lon-1 is epistatic to all of the Sma phenotypes, while the Sma phenotypes of
daf-4 ,
sma-2 ,
sma-3 ,and
sma-4 are epistatic to the Lon phenotypes of
lon-2 and
lon-3 : In most instances, the lon genes did not affect the other phenotypes of these mutants (Mab and, for
daf-4 ,dauer constitutive, Daf-c). We did, however, notice one interesting exception:
lon-2 enhances maternal rescue of the
daf-4 Daf-c phenotypes. In particular, the
daf-4 ;
lon-2 progeny of heterozygous (
daf-4 /+;
lon-2 /+)mothers were more likely to develop into adults (10% of F2 progeny became Sma adults) than were the
daf-4 progeny of
daf-4 /+mothers (2% of F2 progeny were Sma adults). This suppression requires maternal
daf-4 product, since all of the progeny of
daf-4 ;10n-2Sma adults enter dauer at 25 C. This result strengthens the link between
lon-2 and this signalling pathway, leading us to propose that
lon-2 is a negative regulator of
daf-4 activity. For example,
lon-2 could encode an antagonistic ligand that forms heterodimers with an activating ligand for
daf-4 ,or it could encode a component of the extracellular matrix that binds the ligand for
daf-4 and restricts its access to the receptor. Alternatively, these genetic interactions could be the result of less specific effects of the mutant proteins. One additional observation, however, demonstrates that a Lon phenotype can be caused by over-activation of this pathway. Clones that rescue
sma-4 in transgenic animals often confer a Lon phenotype. Taken together, these results suggest that at least a subset of the lon genes, most likely including
lon-2 ,may be involved in this TGF- -like signalling pathway. Further testing of these hypotheses will require the identification of null alleles for these genes and the molecular cloning of their gene products.