Using a microinjection assay, we identified several regions of the X chromosome that feminize triploid males (McCoubrey et al. Science 242:1146, 1988). We then used one of these regions for gel retardation assays using nuclear protein extracted from worms of mixed stages and sex. There is specific binding to a 64 base feminizing element from the first intron of
act-4 (WBG 11(1):69). Further analysis has indicated that most, if not all, of the mobility shift can be attributed to specific binding to each single strand. Single-stranded oligonucleotides representing the sense and antisense strands of the
act-4 feminizing region were end labelled, gel purified and used as probes for the retardation assay. A variety of single-stranded oligos (see figure) were used to get an idea of the region(s) involved in binding to each strand. Binding to the sense strand is completely competed by a 20-fold excess of cold oligo 64. It is not competed by a 5000-fold excess of non-feminizing sequences including another region of the
act-4 intron and a ribosomal gene enhancer from Xenopus. The sense strand of a 56 nt oligo, which is identical to the 64 but lacks the core octamer 5'TATTGAAA3' common to all feminizing clones, also competes but somewhat less effectively. Approximately four-fold more DNA is required to achieve the same level of competition. This sequence is feminizing in the microinjection assay although to a lesser extent than is the 64mer. The 44 nt oligo containing the core octamer shows virtually no competition on this strand even in 4000-fold molar excess. This suggests that the binding on the sense strand depends on sequences 5' of the TATTGAAA. Binding on the opposite strand, however, is more complicated. Both the 56mer and the 44mer compete for binding almost as well as the 64 itself. In addition, concatamers of the octamer show poor but detectable competition for binding to the antisense strand, requiring a 300-fold excess. Most significantly, the two strands of a 200 base feminizing element from the X-linked myosin light chain genes compete for these activities in amounts comparable to the oligos from the
act-4 intron. These results suggest the possibility that one binding activity depends on the core octamer (in the antisense strand) and another activity (on both strands) depends on sequences outside the octamer. They also leave open the possibility that the binding proteins may interact. They further suggest that the same proteins are binding to two different feminizing elements. We are currently attempting to determine whether the factors also bind to RNA from this region. We are also using methylation interference and hydroxyl-radical footprinting to map protein contacts on the single-stranded DNA. [See Figure 1]