Traditional screens aiming at identifying genes regulating development have relied on mutagenesis. A new gene has been identified involved in bristle development, identified through the use of natural variation and selection. Drosophila melanogaster bears a pattern of 11 macrochaetes per heminotum. From a population initially sampled in Marrakech, a strain was selected for an increased number of thoracic macrochaetes. Using recombination and single nucleotide polymorphisms, the factor responsible was mapped to a single locus on the third chromosome, poils au dos (French for 'hairy back'), that encodes a zinc-finger-ZAD protein. The original, as well as new, presumed null alleles of poils au dos are associated with ectopic achaete-scute expression that results in the additional bristles. This suggests a possible role for Poils au dos as a repressor of achaete and scute. Ectopic expression appears to be independent of the activity of known cis-regulatory enhancer sequences at the achaete–scute complex that mediate activation at specific sites on the notum. The target sequences for Poils au dos activity were mapped to a 14 kb region around scute. In addition, pad has been shown to interact synergistically with the repressor hairy and with Dpp signaling in posterior and anterior regions of the notum, respectively (Gibert, 2005).

The large bristles (macrochaetes) on the notum of fruit flies are arranged in a stereotyped array. Development of these bristles has been the focus of detailed genetic analysis. Bristle development is dependent on the activity of basic Helix-Loop-Helix transcription factors encoded by the achaete (ac) and scute (sc) genes. The positioning of bristles is achieved through precise spatial regulation of ac-sc expression in small clusters of cells, the proneural clusters, at the sites of the future bristles in third larval instar wing imaginal discs. Expression in proneural clusters is regulated by multiple, independently-acting cis-regulatory enhancer modules scattered throughout the 100 kb or so of the ac-sc complex (AS-C). These mediate activation by transcription factors such as Pannier and Araucan/Caupolican. Expression of ac-sc is then progressively refined within each cluster to the bristle precursors by autoregulation and Notch-mediated lateral inhibition. The bristle pattern also relies on the activity of repressors such as Hairy and Extramacrochaetae, which prevent accumulation of Ac-Sc at positions where proneural clusters do not form (Gibert, 2005 and references therein).

The pattern of bristles on the notum varies between Dipteran species and a few differences are even found within the genus Drosophila. Differences in the arrangements of bristles between species correlate with changes in the temporal-spatial expression patterns of ac-sc and at least one of their upstream regulators, pannier. So there must exist sufficient variation in the genes regulating bristle patterning within a species (Skaer, 2000) to serve as a substrate for the evolution of bristle patterning between species (Gibert, 2005).

Variation of bristle number in Drosophila has been a classical model for quantitative genetics. Quantitative trait locus (QTL) analysis of lines selected for an increased number of bristles, has shown that a small number of factors of large effect are usually involved. They often map close to genes with a known role in bristle development or more generally in nervous system development. Naturally occurring variation at some of these loci such as the AS-C, scabrous, the Delta-Hairless region, contribute to quantitative variation in bristle number. Most quantitative genetic studies have focussed on the sternopleural and abdominal bristles, which are highly variable compared to the bristles found on the notum. Indeed, a 'wild-type' pattern of 11 macrochaetes per heminotum is assumed to be fixed in D. melanogaster. However, natural variants of this pattern can be found and enough genetic variation exists in nature to select flies for an increased number of notal bristles. Selection experiments for ectopic dorso-central bristles have uncovered the influence of the genetic background and shown that the anterior and posterior dorso-central bristles can to some extent respond independently to selection (Gibert, 2005 and references therein).

In some cases, QTLs do not map close to genes known for their role in bristle development. The study of natural variants can therefore lead to the discovery of new genes and give insights into bristle patterning mechanisms. This study presents the analysis of a natural variant for the thoracic bristle pattern from a selected population of D. melanogaster initially sampled in Marrakech, Morocco, and uncovers a new gene involved in the regulation of ac-sc (Gibert, 2005).

An important question in evolutionary biology is to understand the relationship between intraspecific variability in morphological traits and their interspecific divergence. Selection experiments can help to answer this question, since they allow identification of the genetic variation relevant to a specific trait present in the selected population. In this study, the quick response to artificial selection has relied on the fixation of a single loss of function allele with a strong phenotypic effect. Interestingly, other examples from natural populations have shown that some important morphological variations, such as pelvic reduction in fresh-water sticklebacks can be caused by a major or even a single locus (Shapiro, 2004). However, previous studies on bristle QTL in selected lines of Drosophila have always identified several major QTLs. In the case of the DC bristles, the three major chromosomes contributed to the response to artificial selection in a previous experiment (Dominguez, 1993). It is therefore unusual that the phenotype in the selected line relies on a single bristle QTL (other loci may contribute quantitatively, but the pad¹ mutation is necessary and sufficient for the phenotype). The pad¹ mutation is probably present at very low frequency in Marrakech, since a second sample of flies collected in the same location did not give a strong response to selection, suggesting an absence of the mutation. Furthermore, the phenotype appears to be due to a single deleterious lesion in the pad gene; many single QTLs contain multiple mutations. It is not known whether deleterious variants of this kind contribute to long term evolution but such polymorphisms need not necessarily be devoid of evolutionary advantage. Indeed, some natural situations have been described where loss of function mutations presenting a particular advantage are rapidly fixed during an adaptative radiation. Recent examples in multicellular eukaryotes include loss of function mutations in a pigmentation gene inducing a shift of pollinators in Petunia (Gibert, 2005 and references therein).

The gene uncovered encodes a zinc-finger transcription factor with a ZAD domain. ZAD domains have only been described in insects where they are highly represented: there are about 80 of them in the Drosophila genome (Chung, 2002). Almost all are found in association with zinc finger domains. Few have been studied, but in several cases, point mutations in the ZAD domain have been shown to completely disrupt the function of the gene (Crozatier, 1992; Gaszner, 1999). The ZAD domain of the Grauzone transcription factor has recently been crystallized and shown to be a dimerization domain (Jauch, 2003). The global structure of the ZAD domains is remarkably conserved and so it is possible that different transcription factors are able to form heterodimers through their ZAD domains. The ZAD domain encoded in pad might therefore have a crucial role in establishing molecular interactions between different transcription factors (Gibert, 2005).

In addition to the bristle phenotype, the alleles pad², pad³, and pad⁴ die as late pupae with twisted legs, a phenotype caused by abnormal eversion of leg imaginal discs during pupal development. Unlike these alleles which are predicted not to encode any protein, pad¹ with a predicted truncated protein does not display this phenotype. A similar phenotype is seen in mutants of the gene crooked-legs which encodes a zinc-finger transcription factor showing significant similarity to Pad in the zinc fingers. It is the most closely related gene to pad in the D. melanogaster genome according to Blast analysis (Gibert, 2005).

pad is involved in the regulation of ac-sc. In pad mutants, expression of ac and sc is increased. The increase appears to be general in that the proneural clusters are enlarged, but in addition, other cells with high levels of Ac-Sc are seen in regions where these products are normally absent or barely detectable. This phenotype is reminiscent of the Hw mutations that are associated with a generalized increase in the levels of Ac and Sc (Balcells, 1988). Enlargement of the proneural clusters need not necessarily result from increased activity of the cis-regulatory sequences that normally drive them but can be seen as a consequence of a ubiquitous increase in gene expression. Indeed, these results suggest that ectopic bristles do not arise from proneural clusters. Use of the DC enhancer with a lacZ reporter shows that the ectopic bristles are outside the area of activity of this enhancer. Furthermore, DC bristles also form in the absence of this regulatory sequence [In(1)ac³ and Df(1)91B]. In wild-type flies, the precursors for the pDC and then later the aDC arise from the DC cluster. In pad mutants, a bristle immediately anterior to the pDC, the 'aDC', is outside the area of lacZ staining. This suggests that a bristle at this position, forms earlier in pad mutants from cells with a high level of ac-sc expression not driven by the endogenous enhancer. The presence of such a bristle would then prevent the formation of a precursor from the DC-driven cluster itself, by means of Notch-mediated inhibition. Scutellar bristles can also form in the absence of the scutellar enhancer [In(1)sc⁴] and a reporter construct for the L3-TSM enhancer is unchanged in pad¹. Therefore, pad is unlikely to act via each of the enhancer modules that mediate activation in proneural clusters. It is more probable that Pad acts as a repressor to prevent generalized accumulation of Ac-Sc over the notum and in particular outside the sites of the proneural clusters. The strong genetic interaction between pad and hairy, a known transcriptional repressor of ac-sc, as well as the synergism with Dpp signaling is in agreement with this hypothesis (Gibert, 2005).

This study identifed a 13 kb region around sc that is likely to contain sequences necessary for the formation of all ectopic bristles in pad mutants. It is hypothesized that these sequences direct a weak expression of ac-sc over the entire notum that is up-regulated in pad mutants. One possibility is that Pad acts on the sc promoter. Within the region delimited above, a 3.7 kb stretch upstream of the scute ATG, containing both the SOP enhancer and an enhancer for the wing L3-TSM region, is expressed in several proneural clusters and some other regions. In the absence of Ac and Sc, expression, albeit weak, of a 3.7 kb-lacZ reporter construct can still be detected in the regions of the DC and PSA bristles. Therefore, this fragment drives expression of sc prior to the onset of autoregulation. pad may act, directly or indirectly, via these sequences, allowing a level of Sc high enough in some cells for autoamplification and the adoption of a neural fate. This is visible in cells where the EE4-lacZ is activated, prior to expression of sens and activation of the SOP enhancer. Further studies are required to determine the mechanism of Pad function and whether these sequences are indeed a target for Pad (Gibert, 2005).

Variation of bristle number in Drosophila has been a classical model for quantitative genetics. Traditionally, the phenotype of interest is generated by selection followed by a search for the genetic factors responsible. Several QTLs causing variation in bristle number have been identified in this fashion (Gurganus, 1998; Long, 1995). However, their resolution at the single gene level has only been successful when candidate genes known to affect bristle development were found to map within the QTLs, as confirmed by the more detailed study of some of them. This is also true for studies on other models. In fact, there are very few examples where a new, previously unknown gene, has been shown to be responsible for a QTL. This can be attributed to the difficulty of obtaining enough informative recombinants within the QTL. Indeed, one of the rare successful examples corresponds to a QTL located in a recombination hot spot (Fridman, 2000). Furthermore, most quantitative traits are influenced by several QTL, which makes the mapping more difficult than in this study. The intensive focus on D. melanogaster as a genetic model has led to the development of a number of tools that allow the efficient mapping of mutations and the rapid, precise resolution of QTLs. The genome sequence now provides access to all genes and allows the sites of recombination to be mapped using SNPs. Furthermore, the generation of thousands of precisely located transposon insertions labeled with convenient markers such as white⁺, means that informative recombinants can be efficiently identified. Indeed, the use of Pw⁺ insertions located at proximity to the QTL significantly reduces the effort involved in SNP mapping, and also the cost, since the proportion of uninformative recombinants is much lower. In addition to quantitative variation of bristles, morphometric traits such as body weight, wing and thorax length, ovariole number or pigmentation, vary significantly between wild populations which have adapted locally (Gibert, 2004). These too could be amenable to studies allowing the identification of new regulatory genes. An advantage of natural variants is that the mutations responsible for phenotypic differences are likely to be less severe in general than the complete loss of function mutations frequently generated by traditional mutagenesis. Such hypomorphic mutations facilitate the study of adult phenotypes. Therefore, the combined use of natural variation observed in lines selected from wild populations of Drosophila and the powerful genetic tools provided by laboratory strains permits identification of new genes as illustrated here (Gibert, 2005).

GENE STRUCTURE

Exons - 5

PROTEIN STRUCTURE

Amino Acids - 925

Structural Domains

Using Blast, orthologues of CG10309 were identified in Drosophila pseudoobscura and Anopheles gambiae. The protein sequences are extremely conserved between the two Drosophila species. Two well conserved domains between D. melanogaster and A. gambiae are discernable in the N-terminal and C-terminal regions. The conserved domain in the C-terminal region corresponds to four C2H2 zinc fingers likely to be involved in DNA binding. The conserved domain in the N-terminal region has recently been identified as a zinc-finger-associated domain, ZAD (Chung, 2002). The ZAD has so far been found only in insects and is apparently a dimerization domain involved in protein interactions (Chung, 2002; Jauch, 2003). In the pad¹ mutant, the ZAD is present but the DNA binding domain is predicted to be missing (Gibert, 2005).

C2H2 zinc-finger proteins (ZFPs) constitute the largest family of nucleic acid binding factors in higher eukaryotes. In silico analysis identified a total of 326 putative ZFP genes in the Drosophila genome, corresponding to approximately 2.3% of the annotated genes. Approximately 29% of the Drosophila ZFPs are evolutionary conserved in humans and/or Caenorhabditis elegans. In addition, approximately 28% of the ZFPs contain an N-terminal zinc-finger-associated C4DM domain (ZAD) consisting of approximately 75 amino acid residues. The ZAD is restricted to ZFPs of dipteran and closely related insects. The evolutionary restriction, an expansion of ZAD-containing ZFP genes in the Drosophila genome and their clustering at few chromosomal sites are features reminiscent of vertebrate KRAB-ZFPs. ZADs are likely to represent protein-protein interaction domains. It is proposed that ZAD-containing ZFP genes participate in transcriptional regulation either directly or through site-specific modification and/or regulation of chromatin (Chung, 2002).

date revised: 5 March 2005

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.