Short sequences contribute to endogenous developmental zebrafish enhancers. (a,b) Alignments of 15-bp multiplexed sequences and 9-bp sequences identified by functional dissection were intersected with predicted enhancers (defined by H3K4me1-positive, H3K4me3-negative ChIP-seq regions ). We then used GREAT  to find enriched ontology terms for nearby genes. In each case, a term consistent with the expression pattern was identified, suggesting that the sequences identified by our screen are performing specific roles in developing vertebrates. Plotted are binomial fold enrichment values for the specific 15-bp or 9-bp alignments. Functional 9-bp sequences are depicted by the letter F. To control for common developmental gene ontology terms, we include the binomial fold enrichment values for the H3K4me1+/me3- alignments of all 184 constructs (a) or the non-functional 9-bp sequences identified by functional dissection (b). We tested enhancer activity of 20 1,000-bp regions from the zebrafish genome that overlap with H3K4me1/me3- signal and contain an alignment for 2GH12/notochord or the 2GH08(L9)/brain sequence. (c-e) Three of ten 2GH12/notochord-containing regions (Nc1-3) showed strong notochord (Nc) enhancer expression at 24 hpf. (f-j) Five of ten 2GH08(L9)/brain-containing sequences (Fb1-5) were strong forebrain (Fb) enhancers at 24 or 48 hpf. The chromosomal coordinates of each enhancer are indicated (Zv9 zebrafish genome assembly) as well as the time point with the strongest expression. (k,l) Deletion of the 2GH12/notochord or 2GH08(L9)/brain sequence from the 1,000-bp enhancers leads to a reduction and, in several cases, abolition of the observed enhancer signal. Bars represent the percentages of fish exhibiting expression in the notochord (k) and forebrain (l) at the indicated time point provided in the pictures above. The red dotted line represents the 40% consistency cut-off used for this study.
Smith et al. Genome Biology 2013 14:R72 doi:10.1186/gb-2013-14-7-r72