Many genes have been described that can be dominantly mutated to cause muscle activation defects (Mac-d) or hyper-activated muscle (Mac-h). These genes were proposed to encode ion channels or other regulators of cell excitability. Mutations in a third class of genes confer Mac-d phenotypes in some tissues and Mac-h phenotypes in other tissues (Mac-m; muscle activation mixed). Perhaps the Mac-m genes encode regulatory components whose functions are mediated by various gene products in different tissues to produce physiologically opposite phenotypes (1). If this were true, loss-of-function mutations in downstream genes should suppress the dominant Mac-m phenotypes. This notion was tested for the Mac-m gene
unc-43 by building double mutants between the
unc-43 gain-of-function mutation and loss-of-function mutations in other Mac genes. We found that loss-of-function mutations in the Mac-d gene
unc-103 completely suppressed the
unc-43(gf) enteric muscle contraction (EMC) defect in defecation, but did not suppress any other
unc-43(gf) phenotypes. Other double mutants indicate that the degree of this suppression is specific to
unc-103 and
unc-43. We propose that
unc-103 encodes a tissue-specific effector of
unc-43 in the enteric muscles, but that other proteins function downstream of
unc-43 in other tissues. Loss-of-function mutations in both
unc-103 and
unc-43 confer weak suppression of the
egl-36(gf) and
egl-19(lf) EMC defects. To explain these results, we hypothesize that both
unc-103(lf) and
unc-43(lf) cause the enteric muscles to be slightly hyper-activated. Based on sequence reported by the genome sequencing consortium (2), we identified a putative K+ channel homolog (C30D11.1) that was a likely candidate for the
unc-103 gene. Using C30D11 as a probe of a genomic southern of
unc-103 mutant DNAs, we found 1.6 kb insertions in fragments corresponding to C30D11.1 from strains containing the mutator alleles
e1597k103,
e1597k104,
e1597k105, and
e1597sa468. These mutations were generated by reverting the
unc-103(
e1597) dominant phenotypes in a
mut-6 background (3). Furthermore, the 1.6 kb insertion was absent in a revertant of
e1597k104. Finally, the putative
unc-103 gene was completely deleted in the EMS allele
e1597n1213. We conclude that
unc-103 corresponds to C30D11.1. RT-PCR was used to clone
unc-103 cDNAs, and the coding region was sequenced in several such clones. Overall, this predicted protein is 54% identical and 70% similar to the human ether-a-gogo related gene (HERG), an inward rectifying potassium channel shown to be mutant in a form of the hereditary cardiac disorder Long QT Syndrome (4-7). Furthermore, the
unc-103 and HERG sequence comparison is more striking in the putative transmembrane and pore domains, with 82% identity and 94% similarity. Given this extremely high similarity, we speculate that
unc-103 is a functional homolog of HERG, and that analysis of
unc-103, and perhaps
unc-43, is relevant to the understanding Long QT Syndrome. (1) Reiner, Weinshenker and Thomas (1995) Genetics 141: 961-976. (2) Wilson et al. (1994) Nature 368:32-38. (3) Uekusa, Nishiwaki, Uchida and Miwa (1994) WBG 13(3): 58. (4) Curran, Splawski, Timothy, Vincent, Green and Keating (1995) Cell 80: 795-803. (5) Sanguinetti, Jiang, Curran and Keating (1995) Cell 81: 299-307. (6) Trudeau, Warmke, Ganetzky and Robertson (1995) Science 269: 92-95. (7) Smith, Baukrowitz and Yellen (1996) Nature 379: 833-836.