scute expression in Calliphora vicina reveals an ancestral pattern of longitudinal stripes on the thorax of higher Diptera

Insects bear sensory organs, such as bristles, over the cuticle of the body. In some dipteran flies the bristles are organised into stereotyped spatial arrays in which each bristle occupies a defined position. A single species, Drosophila melanogaster, has been the focus of investigation into the genetic control of the arrangement of sensory bristles. In Drosophila melanogaster there are eleven large bristles, or macrochaetes, on each heminotum and these occupy stereotyped positions. Bristle precursor development depends upon expression of the achaete-scute (ac-sc) genes. The ac-sc complex of Drosophila contains four genes that encode related basic helix-loop-helix (bHLH) proteins, transcriptional regulators that work as heterodimers together with the product of the gene daughterless (Alonso and Cabrera, 1988; Ghysen and Dambly-Chaudiere, 1988; Gonzalez et al., 1989; Villares and Cabrera, 1987). Expression of these genes provides cells with neural potential, allowing them to develop into nerve cells. The large bristles on the notum arise from small clusters of cells expressing ac-sc, called proneural clusters, that prefigure the sites of each of the future bristles (Cubas et al., 1991; Skeath and Carroll, 1991). Within domains of ac-sc expression, single, spaced cells are chosen to become sensory organ precursors. achaete and sc share cis-regulatory enhancer sequences scattered over nearly 100 kb (Gomez-Skarmeta et al., 1995; Ruiz-Gomez and Modolell, 1987). These enhancer sequences respond to local positional cues, conveyed by transcriptional activators. One trans-acting factor is the product of the pannier gene (Ramain et al., 1993). Pannier is a transcription factor of the GATA family and acts in a selector gene-like fashion to regulate pattern in the medial, dorsal half of the notum (Calleja et al., 2000). It has been shown to directly activate transcription of ac-sc through binding to target sequences in the dorsocentral enhancer element, that drives expression in a cluster from which the dorsocentral bristles arise (Garcia-Garcia et al., 1999). The genes of the Iroquois complex are required for the bristles of the lateral half of the notum (Gomez-Skarmeta et al., 1996; Leyns et al., 1996).

Other species of Diptera, particularly those of the derived cyclorraphous Schizophora (that includes Drosophila), have different, but equally stereotyped bristle patterns. Many of these patterns are phylogenetically old, suggesting that they are stable over long periods of evolutionary time (McAlpine, 1981; Grimaldi, 1987). Closely related species have similar arrangements of bristles, whereas evolutionarily more distant ones display more diverged patterns. Ceratitis capitata is separated from Drosophila by about 80 million years (McAlpine, 1981). It displays a stereotyped bristle pattern with some bristles occupying similar positions to those of Drosophila (Wülbeck and Simpson, 2000). The scute gene of Ceratitis is expressed in proneural clusters at the sites of each future bristle, suggesting a similar genetic organisation of the locus in this species (Wülbeck and Simpson, 2000).

Throughout the cyclorraphous Schizophora there is a basic arrangement of bristles on the dorsal notum (McAlpine, 1981). There are four rows of bristles on the scutum: the acrostichal (AC), dorsocentral (DC), intra-alar (IA) and supra-alar (SA) rows. The pattern of most species of Schizophora can be superimposed upon this basic ‘ground plan’, even though some species display all rows and others have only a subset (McAlpine, 1981). It has thus been postulated that the stereotyped bristle patterns of species such as Drosophila and Ceratitis are derived from an ancestral pattern similar to this ‘ground plan’ (McAlpine, 1981; Simpson et al., 1999). The cyclorraphous Schizophora are subdivided into two subordinate groups, the Calyptrata and the Acalyptrata. Both Drosophila and Ceratitis are acalyptrates. When compared with the Acalyptrata, Calyptrata generally bear more macrochaetes and many species display all four rows extending the full length of the scutum. Calliphora vicina is one such species with a pattern resembling the hypothetical ancestral one. It is separated from Drosophila by at least 100 million years.

To investigate evolutionary changes in ac-sc expression, we have isolated homologues of these genes, and also pannier, from Calliphora vicina and examined their expression patterns. We find that sc is expressed in two longitudinal stripes that prefigure the development of the AC and DC rows of bristles. This result suggests that a stripe-like expression pattern of sc may be an ancestral feature and may have preceded the evolution of the small discrete proneural clusters characteristic of Ceratitis and Drosophila. In contrast, bristles of the IA and SA rows of Calliphora arise from domains of sc-expressing cells some of which resemble proneural clusters. These observations reinforce the hypothesis that the stereotyped patterns are derived from an ancestral pattern of four rows of bristles on the scutum and suggest that this pattern may have been the result of a regulated expression of sc in four stripes. We have also examined the expression pattern of the pannier homologue in Calliphora. We find that it is expressed in a conserved domain in the medial dorsal notum, consistent with a possibly conserved selector gene function. The implication of these results for the evolution of the cis-regulatory elements and the function of Pannier in the regulation of sc expression in the proneural clusters of Drosophila is discussed.

Construction of embryonic cDNA libraries was performed using the cDNA Synthesis Kit, ZAP-cDNA Synthesis Kit and ZAP-cDNA Gigapack III Gold Cloning Kit (Stratagene) according to the instructions of the manufacturer. Total RNA was extracted with TRIZOL (Gibco BRL) according to instructions of the manufacturer from a 0- to 24-hour collection of Calliphora embryos. mRNA was purified using the Oligo(dT) Beads Kit (Dynall).

Fragments of Calliphora scute (sc) (729 bp); pannier (pnr) (1194 bp), and Delta (Dl) (555 bp) were isolated by RT-PCR using the following degenerated primers (5′ to 3′, forward then reverse):

sc: AAYGCIMGIGARMGIAAYCG, CRTCRTCIGGIGTRCARTCYTC;

pnr: GAYTTYCARTTYGGIGARGG, GCIGYYTGIATIACRTTRTGYTG;

Dl: CCIGGIACITTYWSIYTIATIRTIGARGC, RCAIGTICCICCRTTIVCRCAIGG.

cDNA was generated from mRNA extracted from a 0- to 24-hour embryo collection using Superscript II reverse transcriptase (Gibco BRL). This was then used as a template. PCR was performed according the following general scheme: 94°C 1 minute; annealing temperature 1 minute 30 seconds; 72°C 2 minutes; 35 cycles; 10 minutes 72°C. PCR products were cloned into pGem T easy vector (Promega).

The 1194 bp fragment of pnr recovered by RT-PCR was extended by 5′ RACE PCR using the 5′/3′ RACE kit from Roche. A composite sequence of 1533 bp was generated.

Homologues for lethal of scute (l’sc; also known as l(1)sc) and asense (ase) were isolated by low stringency screening performed at 42°C in buffer containing 20% formamide, 5× SSPE, 0.5% SDS, 5× Denhart’s solution. Washes were carried out at 50°C with 2× SSC, 0.5% SDS. A genomic Calliphora library (from M. Bownes) was plated and nylon replica filters (Amersham, Hybond-NX filters) were screened with a fragment containing the bHLH domain of Drosophila virilis achaete (ac) (from J. Modolell). Several phages containing either l’sc or ase were isolated. Complete coding sequences were subcloned into pBluescript vector (Stratagene).

To recover the full sequence of sc, the Calliphora genomic library was screened at high stringency with the 680 bp fragment recovered by RT-PCR using Amersham Hybond-NX filters and conditions according to the manufacturer.

All sequences were submitted to GenBank. Accession numbers: asense, AY061875; lethal of scute, AY061876; scute, AY061877; pannier, AY061878; Delta, AY061879.

Sequences were compared using the ClustalX software. Alignments were performed using default ClustalX parameters, and percentage identities calculated from the resulting alignments (Thompson et al., 1997).

Flies were kept at room temperature and fed with sucrose. Eggs were laid in fresh meat and kept at room temperature. Larvae were fed on fresh meat and kept at room temperature. White pupae were collected and staged at 25°C.

Digoxigenin-labelled RNA probes (DIG-UTP, Roche) were generated using the standard protocol of Roche. The resulting RNA was resuspended in 100:l preHyb solution (50% formamide, 5× SSC, 0.1% Tween-20, pH 6.0). RNA was transcribed from linearised DNA templates.

Wing discs and pupal thoraces were dissected in phosphate-buffered saline (PBS) and fixed using a modified version of the protocol of Pattatucci and Kaufmann (Pattatucci and Kaufmann, 1992) in a solution of 4% formaldehyde, 5% DMSO in PBS. In situ hybridisations were performed using a protocol adapted from Wülbeck and Campos-Ortega (Wülbeck and Campos-Ortega, 1997).

Wing discs and pupal thoraces were dissected in PBS, fixed in 4% formaldehyde/PBS for 20 minutes and stained. Mouse anti-22C10 and anti-HRP (horseradish peroxidase) primary antibodies were used at 1:200 dilution. Biotinylated anti-mouse secondary antibody was used and visualised using a standard ABC kit (Vector Chemicals). All preparations were mounted in 80% glycerol, 1× PBS.

Adult flies were collected 30-90 minutes after eclosion before the cuticle had tanned and darkened and stored in 70% ethanol, thereby allowing clearer visualisation of bristle patterns. Thoraces were dissected in 70% ethanol, transferred to 100% ethanol for 10 minutes and mounted under raised coverslips in Euparal (Fisher Chemicals).

We have searched for achaete-scute homologues from Calliphora vicina. We were able to recover sequences specific to scute (sc), lethal of scute (l’sc) and asense (ase) (Fig. 1). No sequences specific to achaete (ac) were recovered. Percentage identity with the Drosophila orthologues is as follows (overall/bHLH only): sc, 74.3/96.7; l’sc, 74.9/95.8; ase 71.1/90.2. Comparisons with Drosophila ac were also made (overall/bHLH only): Calliphora sc shared 68.7/87.5% identity, l’sc 66.2/87.2%, and ase 63/81.2%.

Examination of expression of these three proneural genes in Calliphora embryos by in situ hybridisation revealed expression of all three genes in the presumptive central nervous system similar to that seen in Drosophila. l’sc and sc are initially expressed in a dynamic pattern in clusters of cells, which are then progressively restricted to individual neural precursors; ase is expressed in single cells (not shown).

The bristle pattern of the dorsal notum of Calliphora is depicted in Fig. 2D. The four rows of large bristles on the scutum are labelled AC, DC, IA and SA for the acrostichal, dorsocentral, intra-alar and supra-alar rows, respectively. The transverse suture divides the scutum into pre-sutural and post-sutural domains. The scutellar suture also separates the scutum from the scutellum. The scutellum bears a single line of scutellar (SC) bristles round the lateral edge. The expression of sc in the developing notum was examined by in situ hybridisation. Expression starts at pupariation before the wing discs have started to evert and fuse along the midline. The general shape and morphology of the prospective notum is reflected in the shape of the discs, which bear strong similarity to the well-studied Drosophila discs (Usui and Simpson, 2000). Through examination of many different stages and comparison with Drosophila, we were able to determine the positions at which the various bristles arise. By 3 hours after puparium formation (h APF), two distinct longitudinal stripes of expression, aligned with the dorsal midline, are visible in the medial half of the future scutum at the positions of the future AC and DC bristles (Fig. 2A,D). Both stripes are interrupted by a band in which expression is absent; this corresponds to the future transverse suture (see Fig. 2D). A single stripe of expression along the posterior medial edge prefigures the row of SC bristles. In the lateral half of the notum expression appears in several broad domains and smaller clusters at the sites of the future IA and SA bristles (Fig. 2A,D). A stripe-like domain can be discerned but appears to include bristles of both IA and SA rows. In the lateral region, too, expression is lacking along the prospective transverse suture. The correspondence between the domains of expression and the future bristle rows is shown in Fig. 2D. Several clusters of expressing cells are also present in the region of the wing hinge and calypter, as well as on the wing blade, but we did not attempt to correlate these with specific sensory organs (Fig. 2A,B).

Expression is first observed in the DC stripe at, or just prior to, pupariation. It is rapidly followed by expression in the posterior domain of the prospective AC stripe, the posterior most IA and SA domains, and the more medial domain of the SC stripe (not shown). Expression is then observed more anteriorly, in the case of the AC, IA and SA domains, and laterally for the SC domain, and is finally visible at all sites of the prospective bristle organs by 3 hours APF (Fig. 2A).

By 8-10 hours APF, the stripes of sc expression are less coherent and later strong spots of high levels of expression are seen, which probably reflect the emergence of bristle precursors (Fig. 2B). At the same time ase starts to be expressed in single cells, the precursors of the bristles. By 10 hours, ase expression is widespread amongst the precursors (Fig. 2C,E). The rows of precursors arise from within each stripe or cluster of expression. However, ase expression is transient and the complete pattern of bristle precursors cannot be visualised at any one time. The order in which bristle precursors arise, as revealed by ase and high levels of sc, mirrors the progression of sc expression, and there is a general trend for precursors to arise in a posterior to anterior fashion. This holds true for the AC and DC domains, each of which have three presutural and three postsutural precursors. Exceptionally though, the final precursors to form in the post-sutural domain are the central ones of each triplet. These intercalate between the existing two, perhaps after growth of the epithelium by cell division generates more available space. This may also be the case for the three pre-sutural bristles in each row. By 16 hours APF sc expression has faded from the precursors.

22C10 is a marker of late precursors and the entire neural lineage, and is expressed later than sc and ase (Zipursky et al., 1984). The 22C10 antibody thus reveals neural precursors and staining for this marker reveals a similar pattern and time progression of precursor segregation (Fig. 2F). We also isolated a 555 bp fragment from the Delta gene of Calliphora. Delta has been shown to be downstream of Ac-Sc in Drosophila (Kunisch et al., 1994; Parks et al., 1997). In situ hybridisation with this probe, as well as staining with the cross-reacting antibody against horseradish peroxidase, a neural marker, confirmed the above sequence of events (not shown).

In addition to the large bristles (macrochaetes), the notum is also covered with numerous small bristles (microchaetes). At about 30 hours APF a second wave of sc expression takes place and is correlated with segregation of the precursors of the small bristles (Fig. 3A,B). Staining is fairly ubiquitous over the notum but is again excluded from the sutures. It is then refined to expression in single cells. sc expression at this stage reappears in the macrochaete precursors (Fig. 3A,B). By 33 hours APF ase staining is visible in single microchaete precursors (Fig. 3C), and by 50 hours 22C10 staining indicates that axonogenesis of the bristle neurons is taking place (Fig. 3D). We were unable to detect l’sc transcripts during development of the imaginal notum.

We isolated a 1533 bp fragment of the pannier (pnr) gene from Calliphora. It shows 76.3% of overall conservation with Drosophila pnr, and 99% in the two zinc finger motifs. This is significantly greater than with any of the other four Drosophila GATA factors that display between 54% and 63% overall identity to Calliphora pnr. In situ hybridisation revealed expression in the wing discs at pupariation and throughout the period of sc expression and segregation of bristle precursors. It is expressed in a broad domain, similar to that of Drosophila pnr, which covers the medial half of the notum (Fig. 4A). The lateral boundary of pnr expression appears to be aligned with the DC row of precursors (Fig. 4B,C).

We looked for variation in the bristle pattern between individuals of different size in our colony of Calliphora. Pupae of the same age were collected, weighed, and divided into four groups: 30 mg and smaller, 30-50 mg, 50-70 mg, and 70 mg and larger. They were left to hatch and the number of bristles per row on each individual was scored (Fig. 5F). The results indicate that there is some variation amongst the DC, IA and SA rows, and animals with fewer or additional bristles are found. Fig. 2D depicts the pattern seen in the majority of flies; we refer to this as the ‘wild type’. In total, 10.5% of individuals showed differences from the ‘wild type’ in the DC row, 6.5% in the IA row and 19% in the SA row. However, this variation is small by comparison to that observed in the AC (58%) and scutellar (54%) rows. Moreover, the variation in bristle number in these two rows correlated strongly with size of the individuals. The AC row, for example, generally includes six bristles, three pre-sutural, and three post-sutural ones. Smaller flies, however, may have five bristles, and larger flies seven bristles, in this row (Fig. 5A,E). Bristles are most frequently lost from the anteriormost position (Fig. 5E) but may also be lost from other sites (Fig. 5D). Supernumerary AC bristles may appear throughout the row (Fig. 5A,B). Absence of bristles in the AC, SC, and DC rows is often associated with displacement of the other bristles (Fig. 5B,E). This indicates that the precise position of each bristle in the row is variable, although the lower level of variation in the DC row does suggest that a more robust patterning mechanism may operate in this domain. However, in other individuals, both supernumerary and missing bristles can be superimposed on top of the ‘wild-type’ pattern such that the ‘wild-type’ bristles are not displaced (Fig. 5A,D). Interestingly, bristle displacement was not observed in the IA and SA rows, where precursors arise from a more cluster-like expression of sc.

We isolated three orthologues of the Drosophila AS-C genes, which show clear homology to sc, l’sc and ase. Despite extensive screening by both PCR and of cDNA and genomic Calliphora libraries we were unable to find ac, so this gene may not be present in this species. It was not found in the acalyptrate, Ceratitis capitata (Tephritidae), and has so far only been described in the genus Drosophila (Alonso and Cabrera, 1988; Gonzalez et al., 1989; Villares and Cabrera, 1987) [see references in Takano (Takano, 1998)]. achaete displays closest similarity to sc of all three species. It shares a largely conserved expression pattern and has been shown to be functionally redundant with sc in Drosophila (Balcells et al., 1988; Gomez-Skarmeta et al., 1995; Martinez and Modolell, 1991). Thus, it may have arisen from duplication of the ancestral sc orthologue, some time after the separation of the Tephritidae and Drosophilidae.

Our results indicate strong conservation of the roles of the different ac-sc genes. In Drosophila, l’sc is essential for development of the central nervous system, and its loss results in lethality. We find that, as in Drosophila, l’sc is expressed in the central nervous system of Calliphora during embryogenesis, but is not expressed in the developing notum. Similarly, expression of sc in proneural domains in the presumptive notum is conserved and expression of ase is restricted to sensory precursors. This suggests that specialisation of the functions of these three genes predates the separation of acalyptrate and calyptrate Schizophora.

An arrangement of four longitudinal rows of large bristles is characteristic of the scutum of a number of calyptrate flies and is thought to resemble an ancestral pattern or ‘ground plan’, from which the many different patterns seen in calyptrate and acalyptrate species are derived (McAlpine, 1981; Simpson et al., 1999). Thus an alignment of bristles into four rows may have been the first patterning event in a series of steps that culminated in the stereotyped bristle arrangements characteristic of Drosophila. The single AC and DC bristles of Ceratitis capitata, an acalyptrate species, come from two separate proneural clusters, suggesting a different developmental origin for AC and DC bristles, consistent with the hypothesis that each of them may be derived from an independent row (Wülbeck and Simpson, 2000). Drosophila does not bear any AC bristles, but does carry two DC bristles that interestingly arise from a single proneural cluster in a posterior to anterior sequence (Cubas et al., 1991; Skeath and Carroll, 1991). This cluster is controlled by a discrete cis-regulatory element, called the DC enhancer (Garcia-Garcia et al., 1999; Gomez-Skarmeta et al., 1995). The origin of this and the other positional enhancers is unknown. We have demonstrated that the row of DC bristles in Calliphora arises from a stripe of sc-expressing cells. It is thus tempting to speculate that a stripe-like domain may have preceded the cluster-shaped domain during the course of the evolutionary history of the lineage leading to Drosophila. If so, discrete regulatory elements may have been acquired to drive expression of sc in stripes on the scutum of a common ancestor of Calliphora and Drosophila. Identification of regulatory elements in Calliphora may help to resolve this hypothesis.

It is noteworthy that the two DC bristles of Drosophila are situated close to one another. If they are indeed derived from a complete longitudinal DC row present in an ancestor, through secondary loss of some of the bristles in the row, it seems likely that bristle loss would occur from the anterior downwards or from the posterior upwards, or both together. The DC enhancer may be derived from a single, discrete regulatory element that was responsible for a stripe of expression in the ancestor. If so, it is unlikely that bristles would be lost from the centre of the row, since this would entail a division of the stripe domain into two separate clusters of expression. We examined the distribution of DC bristles in 63 species of acalyptrate flies from 17 different families. 33% were found to have bristles missing from the anterior end of the row (see also Sturtevant, 1970), 8% from both anterior and posterior ends, and only one species (1.6%) had bristles missing from the middle of the row. The entire DC row was lacking in 1.6% of this sample. Similarly, examination of 52 species of calyptrate flies from 6 families, showed missing pre-sutural DC bristles in 9.6% of cases. The entire DC row was lacking in 7.7% of the animals.

The IA and SA bristles of Calliphora do not arise from stripes of sc expression but from apparent clusters. These resemble the proneural clusters of Drosophila and Ceratitis and are associated with a greater degree of determinacy of the positioning of these bristles (see below).

The pattern of four bristle rows appears to be an ancient, widespread one that has been retained regardless of considerable size differences between different species (McAlpine, 1981; Simpson et al., 1999). This suggests that bristle positioning along the mediolateral coordinate of the scutum was fixed a very long time ago. In contrast, anteroposterior patterning, that is the stereotyped positioning of bristles within rows, seems to have been acquired more recently in derived species. It is a characteristic of many acalyptrate flies such as Drosophila and Ceratitis, but is not a consistent feature of more basal species that frequently display a variable number of bristles within the rows. Such variability is thought to be an ancestral feature.

Calliphora appears to be intermediate with respect to this morphological feature. The number of bristles in the AC row of Calliphora varies between individuals: large flies may have more and small flies fewer AC bristles. It is clear that the precise position of each bristle is, to some extent, variable, since the bristles are often displaced when compared with those on the contra-lateral side (Fig. 5B,E). The AC bristles arise from a stripe of sc expression, so spacing of the bristle precursors could be simply achieved through Notch-mediated lateral inhibition (Wigglesworth, 1940; Kimble and Simpson, 1997; Simpson, 1990). The distance between bristles is a function of the range of Notch signalling, so in larger animals there would be room for more precursors. This view is supported by the order in which the precursors arise within the post-sutural AC and DC rows. Each row has three post-sutural bristles, with precursors for the posterior and anterior-most forming first, followed by the central precursor. Formation of the central precursor may only be possible after growth of the epithelium has provided sufficient space between the other two precursors. It is also noticeable that ‘missing’ post-sutural bristles are invariably those located between the two early forming bristles, which are never lost, and that ‘supernumerary’ bristles are also added in the middle.

However, in other individuals the positions of the ‘wild-type’ bristles do not change, and ‘additional’ ones are superimposed on top of the ‘wild-type’ pattern. This observation leads us to postulate that there may be another mechanism(s), in addition to the regulation of sc transcription, which helps to position the precursors. In Drosophila, one or two precursors are selected from each proneural cluster by means of Notch-mediated lateral inhibition (Wigglesworth, 1940; Hartenstien and Posakony, 1990; Heitzler et al., 1996a; Heitzler and Simpson, 1991). However the choice is often biased to a cell at a specific position within the cluster (Cubas et al., 1991; Simpson, 1997; Skeath and Carroll, 1991); it is not known how this is achieved.

The pnr gene of Calliphora was found to be expressed in a conserved domain, similar to that of Drosophila and Ceratitis, that covers the medial half of the notum. This suggests that pnr has retained its selector gene function (Calleja et al., 2000) in all three species. The bristle patterns and the domains of sc expression within the pnr expression domain differ, however, between the three species (Fig. 4). So, if the function of pnr has been conserved, other factors must have changed in order to account for these differences. It is not entirely understood how the broad domain of Pnr in Drosophila is translated into three small clusters of ac-sc expression, but this requires the activity of three discrete cis-regulatory elements, as well as modulation of Pnr function by at least one cofactor, the product of the u-shaped (ush) gene (Heitzler et al., 1996b; Ramain et al., 1993; Garcia-Garcia et al., 1999; Gomez-Skarmeta et al., 1995; Haenlin et al., 1997; Cubadda et al., 1997). Homologues of ush have not been isolated in Ceratitis and Calliphora, but the AC bristles of these species are situated within the domain where ush is expressed in Drosophila. Changes in the regulation of genes encoding cofactors for Pnr, such as ush, is thus a possible mechanism for evolutionary changes in bristle patterns.

Our observations suggest a model for the changes in gene regulation that may have occurred during evolution of the stereotyped bristle patterns of higher flies (Fig. 6). An ancestor of the Schizophora would have had a pattern of four longitudinal rows of large bristles on the scutum (McAlpine, 1981; Simpson et al., 1999). The bristle precursors would be spaced apart by lateral inhibition and the number of bristles in each row, and hence their position, would be variable. Four stripes of sc expression may have allowed the development of these rows, and they may have been a result of the activity of four discrete cis-regulatory elements. The two medial stripes would have been in the domain of pnr expression, and Pnr would have regulated activity of the two corresponding enhancers. During the course of evolution of the lineage leading to acalyptrate flies there has been a tendency to reduce the number of bristles through secondary loss (Grimaldi, 1987; Simpson et al., 1999; Sturtevant, 1970). At the same time the anterior-posterior positioning of individual bristles has become stereotyped. The cis-regulatory elements responsible for the stripes of sc expression may have been retained during this process and perhaps have been modified to drive expression in small proneural clusters. This development may have been accompanied by modulation of the activity of Pnr brought about by changes in the expression of regulatory cofactors such as Ush.

We warmly thank Mary Bownes and Klaus Scheller for the gift of the Calliphora DNA libraries, Sandra Grant and Klaus Scheller for Calliphora pupae, Juan Modolell for the ac probe from D. virilis and Jean-Michel Gibert for help isolating the full length sc. We are grateful to Claudine Ackerman, Ann Mackay and Claire Chapman for technical assistance, the members of the sequencing facility of the IGBMC, Corinna Wülbeck, Helen Skaer, Henry Disney and other members of the group for discussion and comments on the manuscript. We thank the following organisations for financial support: l’Institut National de la Santé et de la Recherche Médicale, le Centre National de la Recherche Scientifique, l’Hôpital Universitaire de Strasbourg, the Human Frontiers Scientific Programme Organization (RG0374/1997-M), the Human Capital and Mobility Programme of the European Community (FMRX CT 96 0065), l’Association pour la Recherche contre le Cancer and the Wellcome Trust (29156). D. P. and N. S. were the recipients of Marie Curie training fellowships.