Evolutionary analysis of the well characterized endo16 promoter reveals substantial variation within functional sites

Abstract

The evolutionary mechanisms that operate on genetic variation within transcriptional regulatory sequences are not well understood. We present here an evolutionary analysis of an exceptionally well characterized cis-regulatory region, the endo16 promoter of the purple sea urchin. Segregating variation reveals striking differences in the intensity of negative selection among regulatory modules, reflecting their distinct functional roles. Surprisingly, transcription-factor-binding sites are as polymorphic and as likely to contain fixed differences as flanking nucleotides. Whereas nucleotides in protein-binding sites in the most proximal regulatory module exhibit reduced variation, those in other modules tend to be more polymorphic than putatively nonfunctional nucleotides. Two unrelated large insertions at the same position within the promoter are segregating at low frequencies; one is a strong ectodermal repressor that contains 16 verified transcription-factor-binding sites. These results demonstrate that a simple relationship between conservation and function does not exist within this cis-regulatory region and highlight significant population heterogeneity in the fine structure of a well understood promoter.

Fig. 1.

Structure of the endo16 promoter and coding sequence. (A) Variation within sequenced regions. Light green boxes represent activator modules, light pink boxes represent repressor modules, and light blue boxes represent protein-coding sequence. Hot pink squares represent locations of protein-binding sites. Red bars identify SNPs found within our sample of 20 individuals (12 individuals for intron 5). Within coding sequence, a bar with a ball on the end denotes an SNP that changes an amino acid. Above the SNPs are representations of length variation (blue) and simple tandem repeats (green) among the 20 individuals. Downward-pointing triangles represent insertions relative to a reference bacterial artificial chromosome sequence, whereas upward-pointing triangles indicate deletions. The length of the indel polymorphism is reflected in the width of the triangle. Filled arrowheads reflect a single-base indel. (B) Overview of endo16 locus up to exon 6. Purple bars denote regions used in this study. Intron–exon structure beyond exon 6 has not been determined.

Fig. 2.

Heterozygosity and divergence differ among modules, represented in scale for the reference allele (127I21) at the bottom of the figure. Within our sample, the promoter ranges from 1,850 to 2,297 bases in length. (A) Average pairwise differences (π) within each module. Blue (first bar) represents the entire module, dark green (second bar) includes only nucleotides within binding sites, and light green (third bar) includes all nucleotides excluding binding sites. (B) Observed nucleotide substitutions within each module as compared with S. droebachiensis. Colors are as described for A. (C) Sliding 300-base window of indel π (blue) and nucleotide π (red) across the endo16 promoter, in steps of 30 bases.

Fig. 3.

Modules E and F are the result of a novel insertion, polymorphic within S. purpuratus. Four promoter variants are “right-justified” within module D to detail the large length differences among them. Sequences corresponding to modules E and F (blue) are found only in the published sequence (GenBank accession no. S75835), embedded within a dinucleotide repeat (yellow). This insertion is the location of 16 verified binding sites. Most alleles (e.g., SpuLA23) possess only the dinucleotide repeat and a small section of module E containing two GCF1 sites, as does the sequence from S. droebachiensis. Interestingly, one allele we sequenced (Spu1107) contains an unrelated insertion (purple) within the same repeat sequence. Potential cAMP response element-binding protein (CREB)-binding sites are indicated by black ovals (see text).