We have previously reported that the divergent GATA-type regulators, MED-1,2, while essential for MS fate, contribute significantly to E specification, as ~50% of
med-1,2(RNAi) embryos lack endoderm (Maduro et al., 2001). The McGhee lab published a paper in 2005 that challenged this finding, reporting that embryos chromosomally lacking
med-1 and
med-2 made endoderm between 80-97% of the time. We subsequently found that germline med mRNAs in wild-type animals are not eliminated in strains that produce the
med-1,2(-) embryos studied by the McGhee lab, suggesting maternal med rescue. We used a strain in which maternal transcripts are largely eliminated by cosuppression with a
med-1(+) array integrated on the balancer sDp3, to form the new balancer irDp1. This array includes copies of
unc-32, for following transgenic animals in the strain used to generate irDp1. Captan et al. (2007) reported that
unc-32(RNAi) can strongly abrogate the production of gut granules in
med-1,2(-) embryos, and concluded that the increased penetrance of gutlessness that we see in
med-1,2(-) embryos produced by irDp1(+) mothers results from cosuppression of
unc-32, and not
med-1,2. This hypothesis predicts that we should be able to detect cosuppression of
unc-32 in our
med-1,2(-); irDp1 strain. However, using in situ hybridization and quantitative RT-PCR, we find no evidence for abrogation of
unc-32 transcripts (maternal or zygotic), even though we showed that gonadal
med-1,2 transcripts are substantially reduced (Maduro et al., 2007). A failure of
unc-32(+) cosuppression from irDp1 is consistent with the way we originally constructed the cosuppressing transgene, using a molar excess of
med-1(+) versus
unc-32(+) DNA. We are now building strains in which a partial
med-1 cDNA is driven by the
dpy-30 promoter -- an arrangement that is predicted to produce germline med cosuppression (e.g. Dernburg et al., 2000; Robert et al., 2005) -- in order to look for enhancement of the gut defect seen in
med-1,2(-) embryos produced by mothers carrying non-irDp1 balancers. Finally, our previously published work has shown weak input of
med-1,2 into activation of the endoderm specifying gene
end-1 (one of a pair of redundant genes that specifies gut; Maduro et al., 2005). We have now found that full-length reporter transgenes for the
end-1 paralogue
end-3 are dependent on the presence of MED sites for expression, providing a mechanism for contribution of
med-1,2 to endoderm specification through
end-3. Hence, we find no compelling reason to modify the model that we have most recently proposed (Maduro, 2006), in which both maternal and zygotic
med-1,2 provide a significant input into endoderm specification, along with SKN-1, PAL-1 and POP-1.