Patterning and differentiation of developing musculatures require elaborate networks of transcriptional regulation. In Drosophila, significant progress has been made into identifying the regulators of muscle development and defining their interactive networks. One major family of transcription factors involved in these processes consists of homeodomain proteins. In flies, several members of this family serve as muscle identity genes to specify the fates of individual muscles, or groups thereof, during embryonic and/or adult muscle development. This study reports on the expression and function of a new Drosophila homeobox gene during both embryonic and adult muscle development. The newly described homeobox gene, termed lateral muscles scarcer (lms), which has yet uncharacterized orthologs in other invertebrates and primitive chordates but not in vertebrates, is expressed exclusively in subsets of developing muscle tissues. In embryos, lms is expressed specifically in the four lateral transverse (LT) muscles and their founder cells in each hemisegment, whereas in larval wing imaginal discs, it is expressed in myoblasts that develop into direct flight muscles (DFMs), which are important for proper wing positioning. The regulatory inputs of various other muscle identity genes with overlapping or complementary expression patterns towards the cell type specific regulation of lms expression were analyzed. Further it was demonstrate that lms null mutants exhibit reduced numbers of embryonic LT muscles, and null mutant adults feature held-out-wing phenotypes. A detailed description is provided of the pattern and morphology of the direct flight muscles in the wild type and lms mutant flies by using the recently-developed ultramicroscopy, and it was shown that, in the mutants, all DFMs are present and present normal morphologies. This study has identified the homeobox gene lms as a new muscle identity gene and shows that it interacts with various previously-characterized muscle identity genes to regulate normal formation of embryonic lateral transverse muscles. In addition, the direct flight muscles in the adults require lms for reliably exerting their functions in controlling wing postures (Müller, 2010).

The musculatures in both vertebrate and invertebrate animals are composed of a large variety of different muscles that are distinguished according to their specific size, morphology, and physiological properties. Whereas much progress has been made in defining the regulatory processes of myogenesis as such, the understanding of the developmental mechanisms underpinning muscle diversity is much less complete. To date, the fruit fly Drosophila has been one of the most profitable models for dissecting the mechanisms regulating muscle diversity. In this system, a number of mechanisms that provide specific identities to individual muscles have been described, particularly for the development of larval muscles during embryogenesis. By contrast, knowledge about the diversification of adult muscles, which takes place in a second round of muscle development during metamorphosis, is still much more limited (Müller, 2010).

Larval muscle development in Drosophila leads to the formation of ~30 distinct muscle fibers arranged in a stereotyped pattern within each embryonic trunk hemisegment. A large body of evidence has revealed that each of these muscles is seeded by a single myoblast, termed muscle founder cell, which already retains a defined identity that predetermines the characteristics of the particular muscle it will form. Upon myoblast fusion between founder myoblasts and fusion-competent myoblasts, which largely lack distinct identities, the identity of the founder cell is then imposed on the growing muscle syncytium and shapes its development. In the current view, the identity of individual founder cells is conferred through the expression and function of particular combinations of muscle identity genes. The muscle identity genes that have been characterized functionally to date all encode various types of transcription factors. The best-characterized muscle identity factors belong to the families of the homeodomain proteins (including Apterous/Ap, Slouch/S59, Ladybird/Lb, Even-skipped/Eve), Zinc-finger factors (Krüdppel/Kr), basic helix-loop-helix factors (Nautilus/Nau), and the COE (Col/Olf1/EBF) transcription factors (Collier/Col). In addition, the activities of these identity factors are modulated by transcription factors that are expressed within distinct broad domains in the somatic mesoderm, such as the homeodomain proteins Tinman/Tin, Muscle segment homeodomain/Msh, Six4, and Pox meso. Another notable example of this latter class of regulators are the Hox factors, which are expressed in broad domains along the anterior-posterior axis within the somatic mesoderm and are known to modulate the activities of muscle identity factors in a region-specific manner along this axis. As has been shown in some of these cases, different muscle identity genes and regional regulators are part of hierarchical and cross-regulatory networks during the development of a particular muscle. As a result, some of the identity factors and regional factors are expressed only transiently whereas the expression of others continues until a mature fiber is formed. Ultimately, the functions of these transcription factors in muscle fate determination must be mediated by their transcriptional target genes, but information about these targets and their roles in making muscles distinct is currently rather limited (Müller, 2010).

During metamorphosis, the majority of the adult muscles are generated anew from the descendants of undifferentiated myoblasts, termed adult muscle precursors, that are set aside during embryogenesis and start proliferating during larval stage. Currently, there is no clear evidence for muscle identity genes acting at the level of individual muscle founder cells during adult muscle development. However, like in embryos, Hox genes are known to play an important role during the regional diversification of muscle patterns along the anterior-posterior axis. Another example for genes involved in adult muscle diversification is ladybird, which is widely expressed in leg disc-associated myoblasts and required for normal leg muscle development. Hence, this embryonic muscle identity gene is redeployed during metamorphosis to participate in the control of the development of large subset of myoblasts, namely those forming the leg muscles. An instructive example of myoblast diversification during metamorphosis has also been described in the wing disc. The wing disc-associated myoblasts generate two fundamentally different types of muscles, which on the one hand include the indirect flight muscles that power the flight, and on the other hand the direct flight muscles that control wing positioning during steering and flight stabilization. It has been demonstrated that the myoblasts giving rise to the indirect flight muscles (IFMs), which form the majority of the wing disc-associated myoblasts and are located in proximal areas of the wing disc of the presumptive notum, are marked by the expression of the homeobox gene vestigial/vg. Conversely, the myoblasts forming the direct flight muscles (DFMs), which are located in adjacent areas near the future wing hinge, are marked by high-level expression of the homeobox gene cut. In this latter population of myoblasts, high levels of Cut repress vestigial, whereas in the IFM-forming population of myoblasts Vg down-regulates cut to low expression levels. In addition, Vg represses apterous (ap), which can therefore only be activated in the high-cut myoblasts. ap then helps specifying these myoblasts as DFM myoblasts. Altogether, these regulatory interactions and the functions ascribed to vg and cut/ap in IFM versus DFM development point to some mechanistic analogies of muscle diversification during larval and adult muscle development (Müller, 2010).

As the currently-known collection of muscle identity genes is still not sufficient to explain the entire muscle pattern during embryogenesis, and even less so during formation of adult muscle diversity, these studies have been aiming to identify additional genes of this type. This report describe a new homeobox gene, lateral muscles scarcer (lms), that fulfills the criteria for a muscle identity gene. During embryogenesis, lms is expressed specifically in the founders and syncytial fibers of the lateral muscles LT1-LT4 as part of a regulatory network that includes ap, which exhibits a closely related expression pattern, as well as lb, Kr, and msh. Null mutations for lms, which are homozygous viable, cause defects in LT muscle development that consist of a reduction in the number of muscles and morphological aberrations. These defects occur with a relatively low expressivity, similar to those reported for ap, and double mutants for lms and ap show additive effects. During adult muscle development, lms is expressed in wing disc-associated myoblasts within a small area that overlaps with the presumptive DFM myoblasts marked by high-cut expression. The held-out wing phenotype of lms null mutant flies is compatible with a requirement of lms for normal DFM differentiation. Because detailed analysis of the DFMs in lms mutant flies showed that the DFMs are present and lack any overt morphological alterations, it appears that lms is needed for the acquisition of the requisite functional properties of these muscles rather than their formation and morphogenesis (Müller, 2010).

Homeobox gene lms is the first representative among its orthologs in insects and primitive chordates (Ciona and amphioxus) that has now been characterized in terms of its expression and function. Although some expression data are available for Nkx-C, its ortholog from Ciona intestinalis, the exact tissues of expression of this gene remain to be characterized (Satou, 2008). Of note, in Drosophila the expression of lms is highly restricted and only found in specific domains of cells within the somatic mesoderm and the muscles derived from them. The lms gene is active within the somatic mesoderm during both larval and adult myogenesis, which suggested that it functions during both of these phases of muscle development (Müller, 2010).

In the embryo, lms is expressed like a typical muscle identity gene. Its expression in progenitors, founders and syncytia of the lateral muscles LT1 - LT4 is very similar to the mesodermal expression of the LIM homeobox gene apterous (ap), except that ap is activated slightly earlier in the corresponding myogenic preclusters. ap exerts regulatory inputs towards lms, which become most apparent upon ectopic expression of ap. However in ap mutants, lms is still expressed largely normally and the same is true for the expression of ap in lms mutants. Thus, although regulatory interactions between the two genes do exist, their expression seems to be established largely independently from one another by related upstream activators. Two candidates for these include msh and lb. msh expression is significantly broader but overlaps with lms, and loss of msh function causes delayed and less robust lms expression. Conversely, lb is expressed in adjacent cells and appears to play a role in the spatial restriction of lms expression. Because single mutations for any of the tested candidate genes do not lead to a total disruption of lms expression, either these regulators act redundantly or there are additional yet unidentified regulators of lms expression that play more indispensable roles (Müller, 2010).

Loss of lms can cause the absence of individual LT muscles or in some cases morphological changes, particularly insertions into inappropriate attachment sites. The absence of an LT muscle could be due to a transformation of its identity into another, although no clear examples of that have been observed. Alternatively, loss of lms function could lead to a failure of a muscle founder to acquire any specific identity or to progress only partially towards acquiring a normal LT muscle identity. This second interpretation, which is compatible with the occasional presence of a small, amorphous syncytium at the position of a missing fiber and the observation of mis-attached and mis-shapen LT muscle fibers, is favored. Interestingly, similar phenotypes with comparable low expressivity were also described for ap mutants. To explain the low expressivity, it was proposed that additional factors can partially compensate for the loss of ap function. Because the expression of lms in LT muscles and their progenitors is very similar to that of ap, Lms was a very good candidate for such a factor. However, in ap, lms double mutants the majority of LT muscles are still present as well, thus ruling out that the two genes are required for LT muscle specification in a mutually-redundant fashion. Rather, the roughly additive effects on the expressivity in the double mutants indicate that the two genes act in parallel with each other and in combination with yet other, perhaps more critical genes during the specification of LT muscle identities. These likely include Kr and msh, functional loss of which leads to a loss of over 30% and 50% of the LT muscles, respectively, as well as yet unknown genes (Müller, 2010).

Altogether, it appears that lms (and ap) is needed for securing the robustness of the program determining LT muscle identities. The findings reinforce the view that there is a significant degree of redundancy built into the muscle specification program in Drosophila. It is increasingly clear that the expressivities of phenotypes upon loss of function of different muscle identity genes occupy a wide range. Whereas lms and ap fall into the low end of this range, the expressivity of msh and Kr phenotypes is low for some muscle lineages and intermediate for others. At the high end of this spectrum are mutations for slou, col, and eve, which affect essentially all muscle lineages in which these identity genes are expressed. In addition, it must be considered that identity genes act within a hierarchically structured network of interactions and at different steps of muscle development. Some of them (e.g., slou, eve, lb) appear to have major roles during the initial diversification of founder cells, whereas others (e.g. ap, lms) may act mainly or purely in the execution of identity programs of specified muscle precursors and the acquisition of individual muscle features such as shape, attachment, and distinct functional properties (Müller, 2010).

The presence of a wing posture phenotype in almost all lms mutant flies, albeit with variable severity, argues for a rather strict requirement for lms during adult muscle differentiation. The major domain of expression during this phase occupies the area of wing disc-associated myoblasts marked by high-cut expression that give rise to the direct flight muscles (DFMs). Interestingly, ap is also activated in these cells, but unlike in embryos, in this case significantly later than lms, namely in pupal stages. Reduction of ap activity has been shown to severely disrupt the formation of DFMs. By contrast, all DFMs are formed and are arranged normally in lms null mutants, which implies that lms is not required for DFM muscle specification and morphogenesis. Instead, it is presumed that lms is needed in these muscles to fulfill their proper functions, which include the adjustment of wing positions and steering during flight. It is conceivable that ap acts together with lms in this pathway, even though ap has additional roles in regulating the formation or survival of DFMs. Only the lms mutant flies with mild or absent held-out wings phenotypes are able to fly, but their steering behaviour, which still may be disrupted, has not been examined. As shown in this study, loss of lms leads to ectopic expression of vg in the presumptive DFM myoblasts. This effect could in part explain the functional defects of the resulting DFMs since GAL4/UAS-driven vg is known to interfere with normal DFM development. In this study, 1151-GAL4-driven vg caused the reported held-out wing phenotype but, as with lms mutants, analysis of these flies via ultramicroscopy did not reveal any differences in the DFM muscle pattern. Whereas most genes with held-out or held-up wing phenotypes encode various muscle proteins, a few others such as Dichaete and mirror are expressed in proximal areas of the disc epithelium and, when mutated, cause disruptions of wing structures in the hinge region. By contrast, lms mutants with held-out wings show normal morphologies of proximal wing elements, which together with the myoblast-specific expression pattern of lms reinforces the notion of a DFM-specific role of lms (Müller, 2010).

A similar held-out wing phenotype as for lms and ectopic vg has been observed for Wnt2 mutant flies, which show a loss, mis-attachment, or ectopic location of usually several of the DFMs in each fly. Presumably as a result of these DFM patterning or attachment phenotypes, Wnt2 mutant flies hold out their wings more strongly as compared to lms mutants and are also unable to fly. The late expression at the epidermal wing hinge of Wnt2 and its temporal requirement, which occurs only during pupariation, rules out a role of Wnt2 in inducing lms, which is already expressed during third instar. However, it remains to be examined whether Wnt2 is needed for the maintenance of lms expression (and/or induction of ap) in the developing DFM myoblasts, which would be analogous to the known role of Wingless in the maintenance of vg in the presumptive indirect flight muscle (IFM) myoblasts (Müller, 2010).