| In Drosophila sex determination, the X-linked bHLH gene product SISB activates the SxlPe promoter in dipo-X (female) embryos by dimerizing with maternally supplied Daughterless (Da). The autosomal gene, dpn, in contrast, modulates the activation threshold of SxlPe to ensure Sxl's appropriate on (female) or off (male) response. Although bHLH proteins have defined DNA-binding sites, the distribution of such sites at Sxl Pe cannot explain how SISB/Da and Dpn regulate Sxl Pe. Using DNase I footprinting and gel-shift, I mapped SISB/Da and Dpn binding sites in SxlPe. The majority of sites lack the E-box CANNTG or D-box CACGCG motifs typical of bHLH proteins. Instead, the proteins bound non-canonical 6 and 7 bp sequences: CAC(C/G)CG and CAC(C/G)TTG for SISB/Da, and CACACT, CACGTT, and CAAGTA for Dpn. Cell culture experiments demonstrated that the noncanonical sites can mediate SISB/Da or Dpn-VP16 activated transcription. Using the single-embryo transient transcription assay (SETTA) I developed, I showed that the SISB/Da sites act in the context of SxlPe, and that multimerized Dpn sites repress an otherwise active 1.1 kb SxlPe. Loss of single SISB/Da sites had little effect but multiple site mutations eliminated promoter activity in transgenic Ries, suggesting the position, number, and affinity of SISB/Da sites are critical for SxlPe regulation.; I also addressed the mechanism of double-stranded RNA interference, or RNAi. I established SETTA as a quantitative in vivo assay that recapitulated all known features of RNAi. Although long dsRNAs were more active than short ones, I found that a 30 bp dsRNA could mediate RNAi. Consistent with recent in vitro findings that 21--23 base RNAs template mRNA for degradation, 21--23 base RNAs processed from dsRNA were present throughout the time when RNAi occured in vivo. RNAi is independent of new RNA synthesis and the antisense-strand of dsRNA determines target specificity, suggesting processed dsRNAs act directly on mRNA. Excess unrelated dsRNA and cognate ssRNAs can inhibit RNAi, suggesting that non-specific dsRNA processing, and sequence-specific target recognition are critical steps in RNAi. Since excess antisense RNAs can inhibit RNAi, I propose that processed RNAs are double-stranded. |
