Fig. 1. Silencing activities of sequences from even skipped. (Top panel) Range of eye colors exhibited by heterozygous flies of different lines carrying each construct (indicated below both panels) is graphed in colors that approximate the eye color. The extent of each color bar represents the percentage of lines exhibiting that color. (Bottom panel) The percentage of lines showing pairing-sensitive silencing (PSS) is graphed. In both panels, the number of lines analyzed is indicated above each bar. The first two bars in each panel represent transgenes carrying mini-white without a PSS element. These showed no PSS activity, either with or without Glass-binding sites, and the presence of Glass activator binding sites strongly shifted the intensity of eye colors, indicative of increased expression of mini-white. PSS elements and derivatives were each assayed in the context of transgenes with Glass-binding sites (see Results). In addition, PRE300 caused PSS activity in 73% of lines and decreased the average eye color intensity; mutation of the Pho binding site (PRE300Pho) drastically decreased PSS activity (to 11%) and abolished the effect on eye color; mutation of either one, two or all three GAF-binding sites (PRE300gaga123) had a relatively weak effect on both PSS activity and eye color. The 3' region of PRE300 (PRE300-3') harbors most of its PSS activity (69% of lines for PRE300-3' compared to 5% for PRE300-5'), even though binding sites for several known PRE-binding proteins are removed (see Fig. 3). The eve promoter region (PSEpro) also confers PSS activity (53% of lines); combining it with PRE300 (PRE300 + PSEpro) did not significantly change this activity (50%), although it did increase the repression of heterozygous mini-white activity. Mutations in Pho- and GAF-binding sites, and deletions of PRE300, are as indicated in Fig. 3.