Dynamic evolution of precise regulatory encodings creates the clustered site signature of enhancers

Abstract

Concentration gradients of morphogenic proteins pattern the embryonic axes of Drosophila by activating different genes at different concentrations. The neurogenic ectoderm enhancers (NEEs) activate different genes at different threshold levels of the Dorsal (Dl) morphogen, which patterns the dorsal/ventral axis. NEEs share a unique arrangement of highly constrained DNA-binding sites for Dl, Twist (Twi), Snail (Sna) and Suppressor of Hairless (Su(H)), and encode the threshold variable in the precise length of DNA that separates one well-defined Dl element from a Twi element. However, NEEs also possess dense clusters of variant Dl sites. Here, we show that these increasingly variant sites are eclipsed relic elements, which were superseded by more recently evolved threshold encodings. Given the divergence in egg size during Drosophila lineage evolution, the observed characteristic clusters of divergent sites indicate a history of frequent selection for changes in threshold responses to the Dl morphogen gradient and confirm the NEE structure/function model.

Discovery and information-theoretic characterization of transcription factor binding sites that act cooperatively

“…It is also consistent with the mathematical definition of "specialized" sites from Erives et al [44] and the D β sites of Crocker et al [28] who defined these sites based on how they were detected (similar to MEME's One Occurrence Per Sequence setting (OOPS) [63], the specialized sites were one site per NEE CRM sequence, where each discovered site shared the highest sequence similarity between the selected sites between the CRMs), which in a sense, is the Dorsal site that had the slowest mutation rate (i.e., under the strongest purifying selection).…”

“…These sites are the D β sites of Table S2 of Crocker et.al. [28], the Dorsal sites from figure 2 of Crocker et al [43], as well as the "specialized" NEE (Neurogenic Ectoderm Enhancers) and NEE-like Dorsal binding sites of Erives et al and Crocker et al [43,44]). Those sites are specialized in the sense that they have been shown to evolve slower than flanking Dorsal binding sites in homotypic clusters of Dorsal binding sites in the NEE [28], and possibly specialized to the cooperative interaction with Twist (which we aim to characterize through information techniques).…”

“…However, this is complicated for a number of reasons: first, binding site turnover within CRMs leaves remnants of functional sites such as "half sites" that have partial matches to motifs [28], second the multiple half sites (that are easier to evolve) may be able to compensate for a strong full site. Therefore, even with a confirmed functional CRM, functional binding site discovery is a daunting task, due to vestigial sites that have fuzzy or poor matches to bioinformatic motifs.…”

“…Our known binding sites, to a degree, come with the class labels already attached. The D DCmel data is the known Dorsal binding site data set based on the definition of D β or 'specialized' sites, or NEE-like Dorsal binding sites (neuroectoderm Dorsal sites that were linked to Twist sites, but were not linked to the canonical 5'-CACATGT Twist sites) [28,44]. However, our DC detector is different than a detector built strictly from the D DC data set (the set of all 12 orthologs for each melanogaster locus), since we included additional ortholog CRMs of the NEEs.…”

Regulation of broad by the Notch pathway affects timing of follicle cell development

“…The common motif of the Su(H) binding site is YGTGDgAA (R=[AG], Y=[CT], M=[AC], D=[AGT], adjusted based on YGTGRGAAM (Crocker et al, 2010); YGTGDGAA (Rebeiz et al, 2002), and the third “g” in the sequence is critical for Su(H) binding (Bailey and Posakony, 1995; Barolo et al, 2000). The bioinformatic program “Patser” (Hertz and Stormo, 1999) and a position-specific scoring matrix (Krejci et al, 2009; Bernard et al, 2010) were applied to predict putative Su(H) binding sites.…”