Apterous mediates development of direct flight muscles autonomously and indirect flight muscles through epidermal cues

Myogenesis can arbitrarily be classified into events common to most muscles and others that confer muscle- or fibre-specific identities. Myoblast proliferation, migration, cell cycle exit and fusion are common to most muscles. In addition, specific properties of muscles determine the biochemical and physiological characteristics of each fibre (Bate, 1990, 1993; Bernstein, 1993).

During differentiation, these general and specific events must finally be integrated – myoblast fusion is a directed process that results in the formation of fibres with predictable, location specific properties. How is this specificity regulated? In the Drosophila embryo, the general and specific aspects of myogenesis have been integrated in the properties of the founder cell that directs the fusion process, while also conferring identity to the developing fibre (Rushton et al., 1995; Ruiz-Gomez et al., 1997). Other myoblasts, the feeders, fuse to founders but are incapable of fusing to each other. Founder identity is specified by the autonomous function of specific transcription factors (Dohrmann et al., 1990; Ruiz-Gomez et al., 1997; Ruiz-Gomez and Bate, 1997; see Baylies et al., 1998 and Frasch, 1999 for reviews). It would therefore appear that, in the Drosophila embryo, feeder myoblasts do not require their identity to be specified and ‘merely’ serve to increase fibre size. This is exemplified by examining the pattern of innervation, location and function of differentiating founders, in a situation where fusion is inhibited (Rushton et al., 1995).

Flight muscle development has features that suggest that other mechanisms may operate during myogenesis (Roy and VijayRaghavan, 1999). Adult myoblasts are set aside during embryogenesis and are siblings of embryonic founder cells (Bate et al., 1991). However, unlike their siblings, they proliferate during larval stages and are closely apposed, in the case of the dorsal mesothorax, to the region of the wing disc epithelium that forms the adult notum (Bate et al., 1991). These myoblasts are the progenitors of the indirect flight muscles (IFMs) and the direct flight muscles (DFMs). The development of the IFMs has been well studied (Fernandes et al., 1991; Fernandes and VijayRaghavan, 1993; Roy et al., 1998; Roy and VijayRaghavan, 1998) but development of the DFMs has been poorly examined (see Fig. 1 for a schematic of muscle development in the dorsal mesothorax).

What are the mechanisms that determine cell fate in adult flight muscles? Specifically, how are identities of these very different muscle fibre groups, the DFMs and the IFMs determined? We show that apterous (ap), which encodes a LIM-homeodomain protein and is a muscle ‘founder cell identity’ marker in the embryo (Bourgouin et al., 1992), is expressed in the progenitors of DFMs during pupal development. Using loss- and gain-of-function experiments we show that ap is also necessary for DFM development.

In addition to its early expression and role in DFM development, ap also has another, nonautonomous function, in patterning the IFMs. It regulates the early expression of stripe (sr), a gene that is a crucial determinant of epidermal muscle attachment sites (Volk and VijayRaghavan, 1994; Lee et al., 1995; Fernandes et al, 1996). We show that sr expression is absent or reduced in ap mutant wing discs, and ap and sr genetically interact with each other to affect IFM development. Further, ap mutant animals have adult thoracic muscle defects. We demonstrate tissue-specific requirements of ap by showing independent rescue of DFM and IFM phenotypes by mesodermal and epidermal ap expression, respectively.

Our report analyses the mechanisms underlying the dual roles for ap in adult myogenesis: autonomously for DFM development and an extrinsic epidermal function in patterning early events in IFM development. We discuss our results in the context of cell-fate specification in adult myogenesis where embryonic founder markers are expressed in groups of myoblasts rather than in single cells and relate these to possible similarities with vertebrate myogenesis.

Canton-S (CS) was used as wild type. ap lacZ (Cohen et al., 1992) is a hypomorphic enhancer trap insertion just 5′ of the ap coding region. ap Muscle Specific lacZ (ap MS lacZ) has been used by Becker et al. (1997) and was from Maria Capovilla and Juan Botas. P1618 is a P-element lacZ insertion in the sr locus (Callahan et al., 1996). sr^G11 is an embryonic null mutation (Frommer et al., 1996). The multiply marked ru, cu, ca/ru, cu, ca stock (Lindsley and Zimm, 1992), carrying the sr¹ allele that, in a homozygous condition, gives rise to a dark midnotal stripe and md237pnrGAL4 were obtained from Bloomington Stock Centre (IN). An ap⁴/CyO MHC lacZ; P1618/TM6 Tb stock was generated for this study. UAS-ap strains were gifts of John Thomas (Salk Institute, CA). Reporter gene activity in Ac(88F)lacZ (Hiromi et al, 1986) has been developmentally profiled (Fernandes at el., 1991). The srGAL4 strain is from Gines Morata (Madrid) and was recombined with UAS-GFP (srGAL4UAS, GFP/TM6 Tb) by Daniela Pistillo and Pat Simpson (Strasbourg).

For the epidermal rescue experiment, ap⁴/CyO MHC lacZ; UAS-ap/TM6 Tb flies were crossed to ap⁴/CyO MHC lacZ; pnrGAL4/TM6 Tb flies. The experiment was conducted at 18°C. Non Tb pupae were separated and eclosing adults scored for absence of wings. Thoracic muscle morphology of these flies was checked and compared with the wingless ap⁴ homozygotes. For the mesodermal rescue experiment, 1151/Y; ap⁴/CyO MHC lacZ males were crossed to ap⁴/CyO MHC lacZ; UAS-ap/TM6 Tb virgins. The cross was grown at 18°C and eclosing wingless females scored for muscle integrity.

Twi antibody (Seigfried Roth, Cologne) was used at 1:5000 dilution. Anti-β-galactosidase and anti-MHC antibodies were used at dilutions of 1:1000 and 1:500, respectively. Anti-Ap antibody (John Thomas, Salk Institute, CA), was preadsorbed overnight and used at a dilution of 1:500. For horseradish peroxidase (HRP) immunohistochemistry, biotinylated secondary antibodies were used and visualised using an standard A+B kit (Vector Chemicals). For fluorescent reactions FITC (green), Alexa 568 (red) or Alexa 488 (green) secondary antibodies were used. Confocal analysis was carried out on a Bio-Rad Model 1024 confocal microscope.

Larvae and pupae were dissected in phosphate-buffered saline (PBS), fixed in 4% paraformaldehyde in PBS and stained with X-gal (Fernandes et al,. 1991) or relevant antibodies. All preparations, except fluorescent samples, were mounted in 70% glycerol. Fluorescent preparations were mounted in Vectashield mounting medium (Vector Chemicals). For adult hemithoraces, nota were cut sagitally, dehydrated through 70%, 90% and absolute ethanol grades, cleared in methyl salicylate, mounted in Canada Balsam and observed under polarised light.

In examining the adult expression pattern of embryonic muscle founder markers, we observed expression of an ap reporter (Cohen et al., 1992) in a pattern that suggested specific roles in myogenesis. The DFMs, located dorsolaterally in each hemisegment of the adult mesothorax, showed reporter gene activity. No staining was seen in the IFMs, which constitute the bulk of the muscles of the dorsal mesothorax. Amongst DFMs, the most conveniently identifiable ones are muscles 49-58 (Miller, 1950; Bate, 1993). Muscles 49 and 51-55, showed staining for the ap lacZ reporter gene in different planes of foci (Fig. 2A-D), and are schematically represented in Figs 1E, 2E.

Epidermal attachment sites of muscles in the adult fly are identifiable by anatomical examination and by expression of stripe (sr). sr marks all muscle attachment cells, both in the embryo and the adult (Volk and VijayRaghavan, 1994; Fernandes et al., 1996). The pattern of sr expression during pupal development has been studied (Fernandes et al., 1996) and the attachment sites for the IFMs identified (Fig. 1D). ap lacZ adult expression is also seen in the thoracic epidermis in the regions where muscles attach (Fig. 2F).

On the third instar wing imaginal disc the presumptive notum showed low ap lacZ (Fig. 3A) and Ap protein expression (data not shown) distinct from the high levels seen in the presumptive dorsal wing. Although the presumptive notum is a dorsal structure, ap does not seem to have a selector function to define ‘dorsalness’ in the mesothoracic trunk as it does in the dorsal wing blade (Cohen et al., 1992; Blair et al., 1994).

ap expression in the pupal epidermis changed temporally beginning with an early expression in broad regions of the dorsal notal epidermis and a subsequent localisation to restricted domains including the attachment sites of the DLMs. At 18 hours after puparium formation (APF), ap lacZ expression was seen in regions that included the developing anterior attachment sites of the DLMs (Fig. 3D). This expression eventually narrowed down to the anterior attachment sites and very closely abutted the posterior attachment sites of the DLMs (Fig. 3E). This can be observed by simultaneous labelling of developing pupae for ap and sr (Fig. 3F-H). By 36 hours APF, when a complete set of DLM fibres is in place, ap co-localisation with individual muscle-attached sr-expressing tendon cells was clearly visualised (Fig. 3I-K). This attachment site expression continued in the adult as shown in Fig. 2F. The same pattern was seen on labelling with Ap-specific antibodies (data not shown).

We examined the developmental origins of the adult DFM-specific expression of ap. Three simple possibilities suggest themselves:

1. ap is expressed in all myoblasts on the third larval instar wing disc and later, perhaps during metamorphosis, becomes ‘switched off’ in IFMs or their progenitors.

2. ap is expressed in a subset of third larval instar wing disc associated myoblasts destined to form DFMs.

3. ap expression is absent from all the myoblasts on the third instar wing disc and begins later, during pupal development, in the developing DFMs.

We find no evidence of ap expression in the wing disc myoblasts. Several lines of evidence substantiate this. ap lacZ wing imaginal discs were double labelled with antibodies against the β-galactosidase protein and against Twist, which marks all myoblasts (Bate, 1990; Currie and Bate, 1991; Fernandes et al., 1991). Although epidermal expression of ap was clear, no co-localisation of ap lacZ expression with Twist was observed. This was confirmed by both confocal microscopy (Fig. 3A-C) as well as HRP immunohistochemistry and examination with Nomarski optics (data not shown). Wild-type wing discs labelled with anti-Ap antibody did not show any myoblast staining, nor did wing discs of a muscle-specific lacZ strain labelled with anti-β-galactosidase (data not shown). The dorsal cells of the disc epithelium were labelled with anti-Ap antibody as expected. These data suggest that ap expression in the DFMs or their progenitors begins during pupal development.

We examined the expression of an embryonic muscle (lateral transverse muscles LT1-4) specific ap reporter strain (see Materials and Methods) during adult muscle development and compared it with the expression of the ap lacZ strain described above and with Ap antibody labelling.

The embryonic muscle-specific ap lacZ (apMS lacZ) showed an adult DFM-restricted pattern of expression in a manner similar to reporter insertions in the ap locus. However, there was no detectable expression in the presumptive dorsal wing blade or in the presumptive notum (data not shown). The pupal attachment site expression observed with ap lacZ and Ap antibody staining was not seen (however, in the adult, staining was seen in the dorsocentral bristles). As in the embryo, the ‘muscle-specific enhancer’ showed ap expression in the developing mesoderm but, unlike the embryo, in clusters of myoblasts and not in single ‘founder’ cells. This is an important difference, the implications of which are discussed later.

We used this mesodermal expression of apMS lacZ to follow ap expression during adult development and this allowed a close observation, uncluttered by epidermal staining, of the dynamic expression pattern of ap in the developing DFMs. It also provided us with a broad developmental analysis of DFMs.

Myoblast expression of apMS lacZ was seen at 12-14 hours APF in clusters of cells all over the dorsal notum (Fig. 4A-C). The three larval muscles that escaped histolysis and served as templates for DLM formation did not express the reporter gene or Ap protein at detectable levels in our assays (Fig. 4A,B). The extent of ap myoblast expression as seen by reporter gene activity increased from 19-21 hours APF onwards, until about 24-26 hours APF and a number of clusters of myoblasts were seen. Between 26-28 hours APF to 34-36 hours APF, these clusters began to arrange themselves into distinct fibres (Fig. 4D,E) and by 36 hours APF completely formed DFMs were in place (Fig. 4F). Fibres increased in size considerably after 36 hours APF until about 48 hours APF at which time the complete complement of DFMs could be identified. Adults continued to express apMS lacZ in the DFMs (Fig. 4I). Immunohistochemistry with anti-Ap antibodies confirmed this (data not shown).

Many of the clusters of myoblasts that express ap can be identified as progenitors of specific DFMs. These inferences are based on the positions of these clusters and their correlation with Miller’s numbering scheme (Miller, 1950). One cluster is destined to form muscle 55 (Fig. 4G,H), based on position and orientation, as the developing fibre in this region is always noticed at the lateral edge of the last DLM fibre and corresponds to the position occupied by DFM 55 in the adult. Another cluster that is consistently noticeable is present below the DLMs and prefigures muscle 52 (Fig. 4F).

To decipher the functional significance of the ap expression pattern, we examined flight muscles of several viable ap alleles. Homozygous ap⁴ animals showed the strongest defects in the thoracic musculature. In particular, the DFMs were severely affected. For our study, we have concentrated on four of the DFMs: 51, 52, 53 and 54 (Fig. 5A). Most ap⁴ mutants showed a sliver of a fibre instead of four distinct muscles (Fig. 5C). Defects in ap lacZ homozygotes were less severe (Fig. 5B). IFM defects included a characteristic reduction in the width of the posterior attachment sites of the DLMs giving them a thin and ‘tapering’ appearance (Fig. 5E). This is consistent with prominent ap expression closely abutting the posterior attachment sites of the DLMs. Another IFM defect was the incomplete splitting of templates for DLM formation, resulting in three instead of six fibres (Fig. 5E). A third phenotype was formation of a single fibre (Fig. 5F). The DVMs also showed defects and were either absent (Fig. 5I) or reduced in size (Fig. 5H). The severity of phenotypes could be ordered as DFMs>DLMs>DVMs. A common feature of all mutants was a striking decrease in overall muscle size and volume.

We attempted rescue of DFM mutant phenotypes by expressing ap selectively in the mesoderm in a viable ap mutant background. ap was expressed in the developing adult muscles of ap⁴ homozygous animals using the GAL4/UAS system (Brand and Perrimon, 1993). We used the 1151 GAL4 driver, which shows reporter gene expression in all disc associated myoblasts until adult muscle differentiation (Roy et al, 1997; Roy and VijayRaghavan, 1997, 1998; Anant et al., 1998). A significant rescue of the DFM phenotypes in ap mutants was seen (compare Fig. 6A-C). While 68% (n=63) of mutant hemithoraces showed DFMs with different degrees of disorganisation, only 6% (n=18) of the rescued progeny showed disorganised hemithoraces. There was no corresponding rescue of the IFM phenotype (data not shown).

A large number of rescued progeny showed a single large muscle mass instead of four discrete fibres (Fig. 6C). This was noticed in 16/18 hemithoraces of the rescue progeny and 11/63 mutant hemitoraces analysed. We observed a similar phenomenon in the epidermal rescue experiment (described later). Our results strongly suggest that ap functions autonomously in DFM development.

We provided ap epidermally in the ap⁴ mutant background using the pannier (pnr) GAL4 driver. pnr is expressed in dorsal cells of the presumptive notum on the disc (Calleja et al., 1996) and the adult expression continues in the same region of the dorsal notum. No mesodermal expression is observed (A. G. and K. V. R., unpublished observations). pnr is also required for normal fusion of the two heminota – high levels of ap produced at 25°C, in the midline, caused a mid-notal cleft and IFM abnormalities (data not shown). To circumvent this and still produce ap in sufficient amounts to mediate a rescue, animals were shifted to 18°C during second instar stages. This resulted in eclosion of flies with a normal notum that could be screened for rescue. We screened the DLMs for morphology and muscle size.

There is a significant rescue of the DLM defects in ap mutants upon epidermal expression of ap (Fig. 6F), while DFM fibres continued to remain disorganised (data not shown). There is also a very striking restoration of muscle size (Fig. 6F). A sizeable population of rescued progeny shows DLMs arranged in one large mass instead of clearly separated fibres.

While 17% (n=60) mutants showed normal DLMs, 80% (n=45) of the rescued progeny had normal DLMs. 89% (n=45) of rescued progeny showed normal muscle size while this category was restricted to 27% (n=60) in mutants. These results substantiated the epidermal requirement of ap for patterning IFMs.

We examined the effects of misexpression of ap in IFM development using the 1151 GAL4 driver. IFM differentiation was assayed using an IFM specific Actin (88F) lacZ marker (Hiromi et al., 1986; Fernandes et al., 1991). When wild-type ap was expressed under control of 1151, the DLM templates showed either an absence of splitting or a delay. Fig. 7B shows a 15 hours APF pupa with obvious signs of degeneration (muscle clusters into a clump: compare with three normal long fibres in wild type control of same stage in Fig. 7A). Where splitting did occur, there was a marked lowering of the levels of IFM Actin staining as seen in Fig. 7D (compare with Fig. 7C). These phenotypes were seen most strongly when flies were transferred to 29°C.

Expression of ap in epidermal attachment sites of muscles and the phenotypes noticed in ap mutants suggested a regulatory role for the gene in the development of muscle attachment sites. We chose stripe (sr), the earliest known marker for epidermal attachment sites, as a potential target for regulation by ap in mediating its epidermal function and examined its expression in wing discs of ap⁴ homozygotes. sr encodes a zinc finger-containing DNA-binding protein, a member of the vertebrate early growth response (Egr) family of transcription factors and is crucial for differentiation of epidermal cells into muscle-attached tendon cells (Lee et al., 1995; Frommer et al., 1997; Vorbruggen and Jackle, 1997; Becker et al., 1997). sr is expressed in discrete domains in the wing disc (Fig. 8A), which will form the attachment sites of adult thoracic muscles (Fernandes et al., 1996). sr expression in the disc commences very late in third larval instar (A. G. and K. V. R., unpublished observations) and is consistently seen as pupation is initiated. Hence the 0 hours APF white prepupal stage was chosen for examination of sr expression in ap mutants. sr expression was either completely lost or drastically reduced in ap⁴ wing discs (Fig. 8B). Animals that reached pupal stages showed severe reduction in sr levels (compare Fig. 8D with wild-type control in 8C).

Our results were further strengthened by the observation that ap and sr interact with each other genetically to affect IFM development. The ap⁴ mutation is completely recessive, as is sr^G11, a recessive lethal at the sr locus (Frommer et al., 1997). In a transheterozygous combination, the two alleles showed defective IFMs in a significant population of animals (Fig. 8G). Further, P1618, an enhancer trap insertion at the sr locus that shows a very mild recessive phenotype, enhanced the IFM defects of ap⁴ (Fig. 8H) and such animals also showed a dark midnotal stripe that was characteristic of strong, viable sr alleles (Fig. 8I-K). This suggests that ap functions in IFM patterning by influencing attachment site development by the regulation of sr.

In this study, we have defined the functions of ap in adult myogenesis, in the development of two distinct thoracic muscle types, the DFMs and the IFMs. The IFMs have been well studied (Fernandes et al., 1991; Fernandes and VijayRaghavan, 1993; Roy et al., 1997; Roy and VijayRaghavan, 1998) but the development of the tubular DFMs has not been significantly analysed. ap is very likely the first transcriptional regulator protein that, by its presence in the DFM progenitors, discriminates between the DFMs and IFMs during development. Ewg, another transcription factor, has been shown previously to be expressed in IFM development during similar stages as ap (DeSimone et al., 1995)

There are important features of ap expression in DFMs that merit analysis: its adult spatiotemporal expression, differences in expression domains between the embryo and the adult, and the mechanism of mesodermal regulation.

Mesodermal ap expression starts early during pupal development and is absent on disc associated myoblasts. Expression is seen in close proximity to the developing epidermis that will overly the DFMs. ap expression is not seen in the developing IFMs.

Using the ap-lacZ reporter strains we have been able to track DFM development. The process shows similarities with IFM development. Both muscle groups are completely established by 36 hours APF, after which there is only a noticeable increase in size. There is a broad division into four temporal domains. Between 12-14 hours APF and 17-19 hours APF myoblast expression of ap becomes noticeable. Next, from 19-21 hours APF to about 24-26 hours APF an increase in ap expression takes place and myoblasts begin to arrange themselves in clusters. Subsequently, from 26-28 hours APF to 34-36 hours APF, distinct fibres appear. By 36 hours APF all DFMs are completely formed and subsequently there is only an increase in fibre size.

The expression pattern of ap in clusters of myoblasts in the developing DFMs is in striking contrast to the mesodermal expression of ap in the embryo where it is expressed in single muscle founder cells (Bourgouin et al, 1992). We would like to suggest that muscle identity in the adult could be determined by expression of genes such as ap in groups of ‘feeder’ myoblasts, as opposed to the situation in the embryo where identity is specified by expression of such genes in a single founder myoblast. This hypothesis thus suggests a role for feeder myoblasts that is not present in the embryo but could be similar to the role that vertebrate myoblasts may have – in vertebrate myogenesis, groups of myoblasts express myogenic genes in response to epidermal cues (see Tajbakhsh and Cossu, 1997, and Yun and Wold, 1996 for reviews).

The phenotypes observed by ectopic expression of ap in the developing IFMs are revealing. The reduction in IFM Actin lacZ staining suggests that ap may be imposing a DFM identity on the IFMs and this may impede their differentiation and cause the subsequent degeneration or failure to attach correctly. However, an absence of well characterised markers of DFM differentiation prevents immediate confirmation of this hypothesis.

What are the signals that regulate the specific onset of mesodermal ap expression temporally and spatially, and where do these signals come from? We have no clear answer to this question at present. However, a significant amount of circumstantial evidence suggests that most aspects of DFM development are configured by the developing epidermis (Roy et al., 1997; Anant et al., 1998). Potential mediators of epidermal regulatory cues are members of the Wnt signalling pathway. In vertebrates, there is strong evidence for Lim-Hd proteins being controlled by Wnts. Lmx1b which is so closely related to ap, specifies dorsal cell fate in the chick limb under control of Wnt7a (Vogel et al., 1995; Riddle et al., 1995). We do find DFM development to be affected in gain- and loss-of-function experiments with downstream components of the Wg pathway. Misexpression of a dominant negative TCF transgene (V. Sudarshan, unpublished observations) and a constitutively activated armadillo transgene (A. G. and K. V. R., unpublished observations) results in DFM defects.

A feature of epidermal patterning crucial for adult muscle development is the specification of attachment sites. This, interestingly, happens in the late third larval instar, although attachment sites themselves are used later during pupation (Fernandes et al., 1996). Experiments that follow the development of both the muscle and epidermal side of developing IFMs have shown the importance of the early association of developing muscle fibres with their attachment sites (Fernandes et al., 1996). This is particularly easily seen in DLMs, which use larval templates for their development. Their close association with attachment sites from the very early stages of their adult specific activity is striking. The expression of sr on the third instar disc is, therefore, the earliest marker of epidermal attachment sites and is an important step in fibre patterning. We have shown that sr expression requires ap for its onset and maintenance.

ap is expressed and functions in the dorsal wing blade and acts as a ‘selector’ gene for dorsal wing cells (Cohen et al., 1992; Diaz-Benjumea and Cohen, 1993; Blair et al., 1994). In the presumptive notum on the third instar wing disc, ap is expressed at low levels. Indeed, absence of wing disc sr expression in ap mutants is the best indicator of its role in the notum. Viable alleles of ap have no effect on notal patterning and clones of cells mutant for null alleles also have no externally visible phenotype (Blair et al., 1994).

The expression pattern of ap during pupal development changes dynamically to co-localise with the anterior dorsal sr-expressing cells, the presumptive anterior attachment sites for the DLMs. It closely juxtaposes, on either side, the posterior attachment sites too. In addition, expression is seen in the developing dorsocentral bristles. Consistent with this expression pattern, ap mutants have muscle defects and defects in bristle development (data not shown). ap encodes a Lim homeodomain protein that requires a co-factor (Chip) for its functioning (Morcillo et al., 1997; Fernandez-Funez et al., 1998) and is negatively regulated by a Lim-only protein dLMO (Bx – FlyBase; Milan et al., 1998; Shoresh et al., 1998; Milan and Cohen, 1999; van Meyel et al., 1999). Despite widespread, albeit low level, expression of ap on the presumptive notum the specificity of its interaction with Chip and dLmo could restrict the domain of ap action to activate sr. Indeed, both wingless (wg) loss- and gain-of-function mutants affect the domains, but not the onset of sr expression, on the wing imaginal disc (A. G. and K. V. R., unpublished observations), which could happen by altering the domain of ap activation.

In its adult thoracic expression, ap shows similarities with members of the Lim protein family that lack the homeodomain and have only Lim domains or Lim domains accompanied with other structural motifs. The most remarkable correlation is with the Caenorhabditis elegans protein UNC-97 and its Drosophila homologue dPINCH (Stck – FlyBase). UNC-97 co-localises with the β-integrin PAT-3 to the focal adhesion-like attachment sites and is required for the structural integrity of the dense bodies – Drosophila equivalents of muscle-attachment sites (Hobert et al., 1999). dPINCH resembles ap in its muscle and tendon-cell expression in the embryo (Hobert et al., 1999); so do Mlp60A and Mlp84B, the muscle Lim proteins from Drosophila (Stronach et al., 1996). ap, by virtue of its homeodomain, functions as a transcription regulator but resembles in expression these non-homeodomain-containing Lim proteins that probably behave as cytoskeletal organisers. While these comparisons do open up possibilities of alternate functions for ap, presumably as an ‘adaptor’ protein itself, it probably functions as a transcriptional regulator of components involved in specification or differentiation.

A combinatorial code of Lim-Hd protein expression appears to exist in neural systems of vertebrates and invertebrates (Thor et al., 1999), where sets of neurons and interneurons express partially overlapping or complementary sets of Lim-Hd proteins to determine properties of individual neurones. A similar mechanism, the combinatorial expression of founder markers, operates to generate muscle diversity in the Drosophila embryo. ap is one such embryonic founder marker but is expressed in clusters of myoblasts in the developing adult. It is possible that ap is part of a code that mediates the diversification of the different thoracic muscles and their individual subsets. Our recent finding that vestigial is expressed in a subset of myoblasts that contribute to the IFMs (V. Sudarsan, unpublished observations) further strengthens the view that groups of myoblasts express specific transcription factors and contribute to different adult muscle types.

We are grateful to John Thomas for UAS-ap stocks and the anti-Ap antibody, Seth Blair for the ap lacZ strain, Gines Morata and Pat Simpson for graciously providing the srGAL4, UAS-GFP stock before publication, Talila Volk for sr alleles and Juan Botas for ap MS lacZ. Our thanks to Veronica Rodrigues, Michael Bate and Stephen Cohen for their comments and stimulating discussions. The suggestions of anonymous referees greatly contributed to improving the manuscript and are gratefully acknowledged. This work was supported by the Departments of Biotechnology and Science and Technology, Government of India and funds from NCBS.