Phenotypic robustness conferred by apparently redundant transcriptional enhancers

Abstract

Genes include cis-regulatory regions that contain transcriptional enhancers. Recent reports have shown that developmental genes often possess multiple discrete enhancer modules that drive transcription in similar spatio-temporal patterns: primary enhancers located near the basal promoter and secondary, or 'shadow', enhancers located at more remote positions. It has been proposed that the seemingly redundant activity of primary and secondary enhancers contributes to phenotypic robustness. We tested this hypothesis by generating a deficiency that removes two newly discovered enhancers of shavenbaby (svb, a transcript of the ovo locus), a gene encoding a transcription factor that directs development of Drosophila larval trichomes. At optimal temperatures for embryonic development, this deficiency causes minor defects in trichome patterning. In embryos that develop at both low and high extreme temperatures, however, absence of these secondary enhancers leads to extensive loss of trichomes. These temperature-dependent defects can be rescued by a transgene carrying a secondary enhancer driving transcription of the svb cDNA. Finally, removal of one copy of wingless, a gene required for normal trichome patterning, causes a similar loss of trichomes only in flies lacking the secondary enhancers. These results support the hypothesis that secondary enhancers contribute to phenotypic robustness in the face of environmental and genetic variability.

Figure 1. The svb cis-regulatory region in D. melanogaster

a, Drawing from the lateral perspective of a D. melanogaster first instar larva. The domain producing quaternary trichomes on the fifth abdominal segment is enclosed in a red outline. b, Diagram of the region upstream of the svb first exon, showing the positions of the five enhancers for this locus: DG2, Z, A, E, and 7. The expression driven by these enhancers in quaternary cells is shown in purple in the diagrams below each enhancer. The piggyBac elements used to generate Df(X)svb¹⁰⁸ are shown as blue triangles. c,f, Expression pattern driven by D. melanogaster Z::lacZ (c) and DG2::lacZ (f) in the 5th and 6th abdominal segments of a stage-15 embryo (purple). An anti-Dusky-like antibody was used to stain developing trichomes (green). d,g, Expression pattern driven by D. simulans Z::lacZ (d) and DG2::lacZ (g). e,h, Expression pattern driven by D. sechellia Z::lacZ (e) and DG2::lacZ (h). β-galactosidase protein produced by D. melanogaster Z::lacZ is expressed in the cytoplasm; β-galactosidase from all other constructs is localized to the nucleus.

Figure 2. Effect of Df(X)svb¹⁰⁸ on the number of quaternary trichomes

a, The lateral patch (green) and dorsal region (blue) in which trichomes were counted. The green and blue boxes correspond to the regions where the Z and DG2 enhancers are expressed strongly. The primary, secondary, tertiary, and quaternary cell types are indicated with horizontal lines above the photograph. The arrow marks the spiracle that was used to set the lower boundary for the green box. The blue box was positioned directly above the green box. The red box identifies the stout tertiary trichomes, which were excluded from the counts. b, Number of trichomes in the lateral plus dorsal region (blue and green boxes) of the fifth abdominal segment of the larva. Open circles give trichome numbers for each individual (n=10); the black symbols and lines show the mean ±1SD. Embryos from each of the two genotypes (C108 and Df(X)svb¹⁰⁸) were reared at three different temperatures: 17°C, 25°C, and 32°C. c, Cuticle images showing the quaternary trichomes in the lateral patch (below) and dorsal region (above) of Df(X)svb¹⁰⁸ first-instar larvae that developed at the three different temperatures. The genotype by temperature interaction term of a two-way ANOVA was highly significant (F = 27.57, P<0.0001).

Figure 3. Rescue of the temperature-dependent trichome loss in the lateral patch by a Z::svb transgene

a,b, Trichome number in the lateral patch (a) and dorsal region (b) of the 5th abdominal segment of larvae with the genotypes C108, Df(X)svb¹⁰⁸, and Df(X)svb¹⁰⁸; Z::svb. Open circles represent trichome numbers for each individual (n=10); the black symbols and lines show the mean ±1SD.

Figure 4. Effect of Df(X)svb¹⁰⁸; wg^-/+ on the number of quaternary trichomes

C108 and Df(X)svb¹⁰⁸ embryos that were heterozygous for a null allele of wingless were reared at 25°C. Quaternary trichomes were counted as described in the legend to Fig. 2. A two-way ANOVA reveals a highly significant genotype by temperature interaction (F=7.79, p=0.0084), which is caused by a large reduction in the number of trichomes on Df(X)svb¹⁰⁸; wg^-/+ larvae relative to all other genotypes.