Genome-wide analysis of clustered Dorsal binding sites identifies putative target genes in the Drosophila embryo

Abstract

Metazoan genomes contain vast tracts of cis-regulatory DNA that have been identified typically through tedious functional assays. As a result, it has not been possible to uncover a cis-regulatory code that links primary DNA sequences to gene expression patterns. In an initial effort to determine whether coordinately regulated genes share a common "grammar," we have examined the distribution of Dorsal recognition sequences in the Drosophila genome. Dorsal is one of the best-characterized sequence-specific transcription factors in Drosophila. The homeobox gene zerknullt (zen) is repressed directly by Dorsal, and this repression is mediated by a 600-bp silencer, the ventral repression element (VRE), which contains four optimal Dorsal binding sites. The arrangement and sequence of the Dorsal recognition sequences in the VRE were used to develop a computational algorithm to search the Drosophila genome for clusters of optimal Dorsal binding sites. There are 15 regions in the genome that contain three or more optimal sites within a span of 400 bp or less. Three of these regions are associated with known Dorsal target genes: sog, zen, and Brinker. The Dorsal binding cluster in sog is shown to mediate lateral stripes of gene expression in response to low levels of the Dorsal gradient. Two of the remaining 12 clusters are shown to be associated with genes that exhibit asymmetric patterns of expression across the dorsoventral axis. These results suggest that bioinformatics can be used to identify novel target genes and associated regulatory DNAs in a gene network.

Figure 1

zen and sog expression patterns. Precellular embryos are oriented with anterior to the left and dorsal up. A and C were hybridized with a digoxigenin-labeled zen antisense RNA probe, and B and D were hybridized with a sog probe. The staining patterns were visualized with anti-digoxigenin antibodies and histochemical staining. (A and C) Parasagittal and surface views of the same embryo. (B and D) Different planes of focus through a single embryo. Note that sog RNAs are detected in nuclei (D). (E) Diagram of the zen 5′ regulatory region showing distribution of the four Dl binding sites in the VRE.

Figure 2

Distribution of Dl clusters. (A) Frequency of clusters in genome containing a minimum of two, three, or four Dl binding sites in intervals of 1,000 or 400 bp. The Dl sequences searched are represented by the degenerate sequences GGGWWWWCCM and GGGWDWWWCCM, which encode a total of 208 unique sequences. Of the three clusters found to contain four sites in 400 bp, one is associated with zen and another with sog. (B) Statistical analysis of the expected (exp) vs. observed (obs) numbers of clusters with two, three, and four Dl sites found in windows of 1,000 and 400 bp. The number of observed clusters of three and four sites are many standard deviations (σ) from their expected frequencies, suggesting that their occurrence at the observed frequencies is not a random event. See Materials and Methods for details. (C) Distribution of Dl binding sites associated with sog, Ady, and Phm. Illustrated below the sog cluster are the three DNA fragments (sog A, B, and C) that were tested for regulatory activities in transgenic embryos.

Figure 3

The sog lateral stripe enhancer. Wild-type and transgenic embryos are oriented with anterior to the left and dorsal up. A–C were hybridized with a sog antisense RNA probe, and D–I were hybridized with a lacZ probe to monitor the activities of different sog-lacZ transgenes. (A–C) Endogenous sog expression pattern in precellular (A), gastrulating (B), and elongating (C) embryos. Staining is detected initially in broad lateral stripes (A and B) but is restricted to the mesectoderm during germ band elongation (C). (D–F) sog-lacZ transgene that contains a 6-kb region of sog intron 1. Staining is detected in broad lateral stripes before (D) and after (E) cellularization but is restricted to the mesectoderm in elongating embryos (F). The staining pattern is similar to the normal sog expression pattern except that there is progressive loss of staining in the mesectoderm (compare C with F; data not shown). (G–I) sog-lacZ transgene that contains a 393-bp fragment from sog intron 1, which encompasses all four high-affinity Dl binding sites. The lacZ expression pattern is similar to that obtained with the 6-kb sog DNA fragment except that staining may be somewhat weaker and mottled.

Figure 4

Comparison of the rhomboid NEE and sog lateral stripe enhancer. Transgenic, cellularizing embryos were hybridized with a lacZ RNA probe to visualize the activities of the 300-bp rhomboid NEE (A and B) and 393-bp sog lateral stripe enhancer (C and D). The lateral stripes generated by the NEE encompass 6–8 cells in the ventral half of the presumptive neurogenic ectoderm (A). In contrast, the stripes produced by the sog enhancer encompass 12–14 cells and include the entire neurogenic ectoderm (C). Both enhancers are inactive or attenuated in the ventral mesoderm (B and D). The NEE has been shown to be repressed by snail. The sog enhancer contains two potential snail repressor sites.

Figure 5

Expression patterns of putative target genes. Normal embryos were hybridized with either an Ady (A–C) or Phm (D–F) RNA probe. Both genes are expressed in ventral regions of precellular embryos (A and D), although Ady may be activated slightly earlier than Phm. Both genes are expressed in cellularized embryos (C and E), and expression persists during gastrulation (F; data not shown). The Ady gene exhibits broad expression in ventral and ventrolateral regions (B). This pattern appears to include the entire presumptive mesoderm and extends into the ventral-most regions of the neurogenic ectoderm.