Evolution of an insect-specific GROUCHO-interaction motif in the ENGRAILED selector protein

Abstract

Animal morphology evolves through alterations in the genetic regulatory networks that control development. Regulatory connections are commonly added, subtracted, or modified via mutations in cis-regulatory elements, but several cases are also known where transcription factors have gained or lost activity-modulating peptide motifs. In order to better assess the role of novel transcription factor peptide motifs in evolution, we searched for synapomorphic motifs in the homeotic selectors of Drosophila melanogaster and related insects. Here, we describe an evolutionarily novel GROUCHO (GRO)-interaction motif in the ENGRAILED (EN) selector protein. This "ehIFRPF" motif is not homologous to the previously characterized "engrailed homology 1" (eh1) GRO-interaction motif of EN. This second motif is an insect-specific "WRPW"-type motif that has been maintained by purifying selection in at least the dipteran/lepidopteran lineage. We demonstrate that this motif contributes to in vivo repression of the wingless (wg) target gene and to interaction with GRO in vitro. The acquisition and conservation of this auxiliary peptide motif shows how the number and activity of short peptide motifs can evolve in transcription factors while existing regulatory functions are maintained.

Fig. 1

The ehIFRPF motif is an evolutionarily novel, conserved insect-specific GRO-interaction motif. Location of the ehIFRPF motif relative to the previously characterized eh1 and eh2 sequence motifs, which physically interact with GRO and EXD, respectively. The D. melanogaster sequence of partially conserved peptide motifs between the eh1 and eh2 motifs are shown. ..., non-conserved (dipteran/lepidopteran clade) bases not shown. No function is known for the motif whose sequence in D. melanogaster is “LGSLCKAVSQIG”. Possible ehIFRPF motifs with one mismatch (lowercase) are also shown for T. castaneum, P. americana, and S. gregaria. No putative ehIFRPF motifs that match the consensus ψΩ(K/R)PΩ (or contain a single mismatch) were found in any other EN/INV homologs (see MATERIALS AND METHODS for organisms and sequences searched).

Fig. 2

The ehIFRPF motif contributes to in vivo repression of WG by ectopic expression of EN. A, WG expression in the presence of ectopic EN⁺ (A’, ectopic EN⁺ expression; A”, merge with EN⁺ expression in red, WG expression in green); B, ectopic expression of EN^IFRPF-; C, ectopic expression of EN^eh1-; D, ectopic expression of EN^{eh1- ehIFRPF-}. Note that ectopic EN⁺ strongly repressed the even numbered WG stripes (A). However, WG levels in the even stripes were only about half as repressed when EN^IFRPF- was ectopically expressed, instead of EN⁺ [B; 0.225 ± 0.088 WG expression index for EN^ehIFRPF- (N = 30) versus 0.119 ± 0.092 for EN⁺ (N = 23); P < 10^-4]. Note that the mutant EN proteins lacking the eh1 motif (C and D) had a more limited ability to repress WG, even at higher expression levels.

Fig. 3

The ehIFRPF motif physically interacts with GRO in vitro. Lanes 1-4 show relative amounts of labeled proteins pulled down by GST-GRO. Lane 1, EN⁺; lane 2, EN^eh1-; lane 3, EN^IFRPF-; lane 4, EN^{eh1- ehIFRPF-}. M2-GRO produced from an insect cell line gave similar results, while interactions with excess quantities of GST were much weaker (data not shown). Note that removing either the eh1 motif or the ehIFRPF motif reduced the interaction to 49 ± 10 % (N = 11) or 48 ± 9 % (N = 11), respectively, of the full-strength EN⁺-GRO interaction (both comparisons, P < 10^-4). Removing both motifs resulted in a slight additional reduction to 40 ± 7 % (N = 8; P < 10^-3 compared with EN⁺; P = 0.014 compared with either EN^IFRPF- and EN^eh1-).