The achaete–scute gene complex (AS-C) contains four genes encoding transcription factors of the bHLH family, achaete, scute, lethal of scute, and asense located in 40 kb of DNA containing multiple cis-regulatory position-specific enhancers. These genes play a key role in the commitment of epidermal cells toward a neural fate, promoting the formation of both sensory organs in the peripheral nervous system (bristles) of the adult and of neuroblasts in the central nervous system of the embryo. The analysis of the AS-C initially focused on the variations in positional specificity of effects of achaete (ac) and scute (sc) alleles on macrochaete bristle pattern in the Drosophila adult epidermis, and from there it evolved as a key entry point into understanding the molecular bases of pattern formation and cell commitment. In this perspective, we describe how the study of the AS-C has contributed to the understanding of eukaryotic gene organization and the dissection of the developmental mechanisms underlying pattern formation.
Anecdotal, Historical and Critical Commentaries on Genetics
PATTERN formation consists of the generation of constant distributions of cell types in a developing tissue or organism. The analysis of the causal mechanisms underlying pattern formation has had a major impact in developmental genetics, due in part to the identification of genetic variants affecting the formation of sensory organs at specific spatial positions in the thorax and head of the fruit fly. In particular, the study of the achaete–scute gene complex has provided the bulk of information and concepts about gene organization, the spatial regulation of gene expression, the genetic and cellular mechanisms of cell commitment, and, more recently, the developmental bases of the evolution of both the genes and the patterns they determine. In this Perspectives we summarize some of the key aspects of the achaete–scute complex that have made a significant contribution to the understanding of the developmental mechanisms regulating pattern formation. We summarize the particular characteristics of achaete and scute alleles that made them attractive from the genetic point of view, the information gained by the molecular analysis of the genes, and the different aspects of bristle pattern formation that made the study of the achaete–scute complex a paradigmatic case of the analysis of developmental genes and the process they regulate.
GENETIC COMPLEXITY OF scute AND achaete MUTATIONS
The story began with the variations in positional specificity of achaete (ac) and scute (sc) mutations in the Drosophila adult epidermis, and, as we shall see, it progressed to identify crucial roles for the wild-type genes in neural development. At the time of their discovery, genes were just hereditary factors whose allelic variants allowed their mapping to chromosomes. The functional nature of these genes could be inferred only from the phenotype of their mutant alleles. For William Bateson, at the beginning of the 20th century, mutant alleles corresponded to the loss of function, but this idea started to be reconsidered when noncomplementing multiple alleles in the same gene appeared. For enzyme coding genes, this notion was understood as partial failures of a basic enzymatic function, e.g., in eye pigment formation. Multiple alleles in the white gene, leading to varied tones of red, were more difficult to explain. It was found later that they were related to mutations in functional domains of a carrier protein displaying distinct affinities for different eye pigments. The allelic series of achaete–scute mutants defied a quantitative, lineal interpretation of the function of the genes in the terms suggested by H. J. Muller (amorphs, hypomorphs, and hypermorphs) to classify mutations on the basis of the results of genetic tests (Muller 1932). Thus ac alleles showed specificity for the removal of microchaetae (“hairs” at the time) and some macrochaetae (“bristles”) of the notum. The sc alleles eliminated only a subset of macrochaetae, those not affected by ac mutations (see Figure 1A). Some sc alleles behaved as noncomplementing in certain macrochaetae positions, but other sc alleles with different pattern specificities would complement for the positions not affected by these individual alleles. The positions of affected macrochaetae in individual alleles and allelic combinations followed a topological order (“seriation”) that was clearly nonlinear in the thorax, but discontinuous. The colleagues of Muller in Moscow (A. S. Serebrovsky, N. P. Dubinin, and A. A. Prokofieva, et al.) designated these sc alleles step-alleles (Treppen allelomorphism in the original) (Agol 1931; Sturtevant and Schultz 1931; Dubinin 1932; Muller and Prokofyeva 1935).
The key point was that different alleles showed specificity in the positions of chaetae they affected. ac–sc alleles are readily inducible by x-ray irradiation and easily detected by changes in the otherwise constant pattern of thoracic chaetae. These alleles could not be mapped meiotically, because meiotic recombination does not occur in the tip of the X chromosome. This prevented the study of possible cis-effects among different alleles, unless they were induced in mutant chromosomes, as, for example, the sc10.1 allele (see Figure 1A and below). With the increase in the number of alleles of sc, in particular, new combinations of affected positions continued to appear. This situation was a genetic challenge for many exceptional geneticists (e.g., Alfred Sturtevant, Curt Stern, and Hermann Müller). It was also a challenge for developmental geneticists: it presented an opportunity to confront the fundamental problem of how “position” is encoded in the genome.
Many of the induced sc mutations were associated with chromosome breakpoints with one break in the “ac–sc” region (distal tip of the X chromosome) and another in the centromeric heterochromatin or in the euchromatin of the X chromosome (chromosomal inversions), or in any other autosomic arm (chromosomal translocations). Raffel and Müller (1940) used X chromosome inversions to generate deficiencies (loss of a chromosomal segment) and duplications of chromosomal segments in the as–sc region (the so called left–right test). This test revealed the existence of the “ac” region, the “sc” region, and a “lethal of scute” (l'sc) region located between sc breakpoints whose deletion was lethal (Figure 1, A and B). The results of a similar approach extending this test to more breakpoints, as well as to internal and terminal deletions and duplications constructed using autosomal translocations, confirmed the existence of the l'sc function and uncovered certain symmetries in the phenotypes of deletions at both sides of l'sc (named sc-α and sc-β) (Garcia-Bellido 1979). Interestingly, the phenotype of the breakpoints was more extreme the closer they mapped genetically to l'sc. In addition, certain duplications showed a phenotype in which extra chaetae differentiated in novel positions of the fly thorax, called the Hairy-wing (Hw) phenotype. Later, the Hw alleles were shown to correspond to a gain of function of ac or sc functions, when it was found that revertants of Hw were ac or sc mutations, and that Hw alleles showed overexpression of the ac and sc genes in normal or ectopic positions (Campuzano et al. 1986; Garcia Alonso and Garcia-Bellido 1986; Balcells et al. 1988). These alleles were enlightening in proposing an instructive “bristle promoting” function for the ac and sc genes. The cytological breakpoints leading to ac, sc, and Hw phenotypes were shown by the work of the Juan Modolell group to extend over ∼100 kb of DNA (Campuzano et al. 1985). This implied that partial or noncomplementation between alleles of the same gene extended through huge distances of DNA!
DEVELOPMENTAL FUNCTION OF THE achaete–scute GENE COMPLEX
A genetic analysis of ac and sc mutations appeared in Genetics at about the same time as another article on the developmental genetics of the ac–sc system (Garcia-Bellido and Santamaria 1978; Garcia-Bellido 1979). The obvious question, at the time, was to know the phenotype of the total lack of function of the ac and sc genes including the l'sc region. This question was first addressed by Stern (1935), when he used a major deletion of the genes that also included the gene yellow (y) [the deficiency element of T(1:2)sc19]. Stern looked for yellow chaetae in somatic recombination spots but found none, therefore concluding that Df(1)sc19 was cell lethal. The analysis by Stern of achaete gynadromorphs, flies in which the male tissue was mutant for achaete, did show, however, that the mutation was acting in a cell-autonomous manner. Stern suggested that achaete was involved in “the response of cells to a predetermined invisible pattern” (Stern 1954, p. 240), the “prepattern,” a concept that had a major influence in the understanding of bristle pattern formation (see below). We repeated the experiment labeling the Df(1)sc19 cells with another cell marker, forked (f36a) in addition to y, and found large spots in the notum devoid of chaetae, but composed of forked trichomes (Garcia-Bellido and Santamaria 1978). This result implied that Df(1)sc19 cells have normal growth and viability, but failed to differentiate chaetae in the notum and in most of the adult cuticle. Cells homozygous for the Df(1)sc19 deletion did, however, differentiate some chaetae in the wing margin—the mechanosensory chaetae of the triple row—suggesting the existence of yet another bristle-promoting gene located outside the region deleted by Df(1)sc19. A subsequent analysis of a terminal deletion with a more proximal breakpoint [Df(1)260-1] concluded that this novel bristle promoting function, named asense, was also critical for the formation of a subset of sensory elements in the larval cuticle (Dambly-Chaudiere and Ghysen 1987). Clearly the ac–sc region included genes for several related functions and has since been called the achaete–scute complex (AS-C).
The best-known AS-C functions at the time were related to the formation of the adult peripheral nervous system (PNS), but the function of the l'sc region was still mysterious, because the smallest deletion including this gene [Df(1)sc4L9R] was embryonic lethal. In gynandromorph mosaics, the Df(1)sc4L9R spots that appeared in the adult cuticle could form large territories that lacked only the chaetae missing in the In(1)sc4 and In(1)sc9 alleles (Figure 1B). In this manner, no adult phenotype distinct from that of the original inversions used to generate this deficiency could be associated with the deletion of the l'sc region. Interestingly, mutant territories that crossed the ventral midline killed the embryo, suggesting that the l'sc function was related to the formation of an essential ventral structure in the embryonic fate map, possibly the central nervous system (CNS) (Garcia-Bellido and Santamaria 1978). Subsequent work by the Campos-Ortega group confirmed the existence in the l'sc region of a function required for the formation of the CNS (Jimenez and Campos-Ortega 1979, 1990). We can now generalize by saying that the AS-C encodes four functions related to both PNS and CNS development in Drosophila. These functions promote the formation of sensory organs in the embryonic and adult peripheral neural systems and of neuroblasts in the central neural system. Accordingly, AS-C genes were named “proneural genes” (Ghysen and Dambly-Chaudiere 1988; Romani et al. 1989). The proneural genes were truly morphogenetic, as indicated by the existence of gain-of-function alleles causing supernumerary sensory organs. The developmental analysis of the AS-C also suggested that the same mutants affected the developmental pathways leading to the formation of chaetae and the CNS. The detailed analysis of cells deficient for the AS-C induced by mitotic recombination in the last stages of larval development indicated that the differential divisions of the chaetae mother cell occurred between 48 hr before puparium formation and 12 hr after puparium formation (APF) and that the macrochaetae completed their development earlier than the microchaetae. This analysis also showed that chaetae were able to develop in the absence of the AS-C, but only when its removal took place at the time of the differential divisions of the chaetae mother cell (Garcia-Bellido and Merriam 1971).
MOLECULAR ORGANIZATION OF THE achaete–scute GENES
Many questions related to the genetic organization and function of the AS-C had to wait for molecular analysis, which was undertaken by Juan Modolell and his colleagues during the 1980s and 1990s (Carramolino et al. 1982; Campuzano et al. 1985; Romani et al. 1987; Ruiz-Gomez and Modolell 1987; Villares and Cabrera 1987; Alonso and Cabrera 1988; Gonzalez et al. 1989; Dominguez and Campuzano 1993; Martin-Bermudo et al. 1993). The first challenges were to clone the genes, identify the coding regions, and molecularly map the existing breakpoints. This molecular information was critical to understanding how the large DNA extent of the AS-C related to the observed interactions between the sc-α, sc-β, and asense regions. The mapping of mutant alleles established the molecular limits of the ac, sc, l'sc, and ase genes and identified the transcripts corresponding to these genetic regions, which were called T5 (ac), T4 (sc/sc-α), T3 (l'sc), and T8/T1a (ase) (Figure 1B). The sc-β region contained the T2 transcript, but this was not related to a bristle promoting function.
The sc alleles mapped in ∼40 kb of DNA 3′ of the T4 (sc) transcript, and their phenotypes were more severe the closer they were to this transcript. At the time, these mutants were thought to cause long-range structural perturbations leading to a reduction in sc transcription, which were stronger the closer they were to the transcript, which combined with a differential sensitivity of each macrochaeta to a given reduction in the amount of Ac/Sc (“threshold”). This scenario would result in the observed seriation of affected bristles (Campuzano et al. 1985). However, a subsequent analysis of ∼70 X chromosome terminal deficiencies ending in the 5′ region of the T4 transcript uncovered a completely different order of sensitive bristles, although still the phenotypes depended on the distance from the gene of each deficiency endpoint (Ruiz-Gomez and Modolell 1987). To explain the existence of these two different 3′ and 5′ seriations, it was proposed that sc alleles disconnected flanking cis-regulatory elements from the transcript: the closer any breakpoint was from the T4 transcript, the stronger the phenotype because more cis-regulatory elements would be disconnected from the gene (Ruiz-Gomez and Modolell 1987).
Other work, reviewed in Ghysen and Dambly-Chaudiere (1988), Campuzano and Modolell (1992), and Gomez-Skarmeta et al. (2003), showed that the four AS-C proteins contain basic-HLH motifs previously found in the Myc oncogene, and that they regulate gene expression. The large noncoding DNA regions within the complex should then contain cis-regulatory regions (“enhancers”) with sequences regulating the expression of the coding regions in specific epidermal positions and cell types. These enhancers are the key to understanding positional specificity and the peculiarities of step allelomorphism, because they act at a distance in cis to regulate the distantly located coding regions. Thus, breakpoints disconnect enhancer regions from the promoters, preventing trans-regulatory interactions, and lead to the absence of particular neuroblasts or sensory mother cells in the CNS and PNS, respectively. The discovery of enhancer sequences in the AS-C located at large distances from the coding regions was one of the first examples of the positional specificity of cis-regulatory DNA likely targeted by trans-regulatory proteins. This example was later shown to be a general aspect of eukaryotic gene organization and extends to many other gene complexes (such as the bithorax complex) where enhancer sequences were thought to correspond to specific genes (Lewis 1998). The question of positional specificity and temporal and spatial specification was thus transferred to the molecular recognition of cis-regulatory sequences by the products of trans-regulatory genes. These notions were fully developed in the emergent analysis of embryonic segmentation, when the hierarchy of maternal, gap, and pair-rule genes was dissected (Clyde et al. 2003). In this manner, the regulatory region of eukaryotic genes includes two types of sequences: one, inherited from prokaryotes that correspond to the “promoter” or “operator” region, and a second, one that can be very complex, exclusive of eukaryotes that can be called the “modulator” region containing the enhancer sequences. The first region is where the interaction with RNA polymerases and other multiprotein regulatory complexes occurs, whereas the modulator region defines when and where the gene will be transcribed through interactions with sequence-specific transcription factors.
EXPRESSION OF THE AS-C GENES
Trans-regulatory genes of the AS-C were searched for, using a “gene titration” approach. Like in bacteria, it was expected that an increase in the number of doses of the promoter could lead to a relative insufficiency of repressor trans-regulatory gene products. Following random mutagenesis in these potential genes, phenotypes of extrachaetae (similar to the Hw) were found in two loci, extramacrochaetae (emc) and hairy (h) (Botas et al. 1982). Both genes also encode HLH proteins that regulate the expression of ac and sc (Hairy) and that interact with the Ac and Sc proteins, antagonizing their function (Emc) (Ellis et al. 1990; Garrell and Modolell 1990; Van Doren et al. 1991, 1994; Ohsako et al. 1994; Campuzano 2001). Similar mutagenesis experiments using a mutant background heterozygous for AS-C deficiencies did not yield any candidate for an activator gene. However, these were later found using other approaches (see below).
The cloning of the AS-C genes allowed a number of experiments that deepened the understanding of the mechanisms leading to bristle patterning. First, the expression of the genes was described in detail in the wing imaginal disc—the epithelium that gives rise to the thorax and wing of the fly—in the embryonic peripheral and central nervous system and in several other tissues from which sensory organs developed (Cabrera et al. 1987; Romani et al. 1987, 1989; Cubas et al. 1991; Martin-Bermudo et al. 1991; Skeath and Carroll 1991, 1992; Dominguez and Campuzano 1993; Ruiz-Gomez and Ghysen 1993). The visualization of the expression patterns of the AS-C genes, by in situ hybridization first and then by using antibodies directed against the proteins, revealed a common scenario in which the genes were first expressed in groups of cells, the so-called “proneural clusters,” and then accumulated at higher levels in the cell that enters the neural fate, the sensory mother cell (SMC) or the neuroblast (Figure 1C). The expression of the AS-C genes, the modifications to this pattern observed in AS-C mutants, and the use of additional cell markers specifically expressed in the neural precursors, together with the finding that AS-C proteins are transcription factors, marked a high point in the developmental analysis of chaeta formation and its relationships with the AS-C genes. In this manner, the complex pattern of sensory elements could be largely reduced to the generation of a landscape of proneural clusters where the AS-C genes were expressed. Similarly, the complex and puzzling complementation patterns among sc mutations came to be understood as a consequence of the existence of cis-regulatory regions directing gene expression in individual proneural clusters (Ghysen and Dambly-Chaudiere 1988; Campuzano and Modolell 1992; Modolell and Campuzano 1998). Interestingly, the same enhancers control the expression of ac and sc, and therefore both genes are expressed in the same pattern of proneural clusters (Ruiz-Gomez and Ghysen 1993; Gomez-Skarmeta et al. 1995). If the specific effects of ac and sc mutations involved only the enhancers they each affect, why do the two sets remove complementary subsets of bristles? This question could be only partially solved when point alleles in the sc (Gomez-Skarmeta et al. 1995) and ac (Marcellini et al. 2005) genes were characterized. Surprisingly, hemizygous males for a sc null allele (scM6) lost only a few micro- and some macrochaetae, whereas flies null for ac (accami) were entirely normal (Figure 1, A and B). The ac and sc double mutant (sc10.1) lacks all macro- and microchaetae (Figure 1, A and B), reinforcing the notion that the corresponding proteins have some degree of functional redundancy. These observations suggested that there are no qualitative position-specific differences between the Ac and Sc proteins with regard to their proneural function, although the mutant phenotypes of individual alleles and the study of ac and sc overexpression phenotypes indicated that the proneural activity of Sc is more effective than that of Ac (Rodriguez et al. 1990; Gomez-Skarmeta et al. 1995; Marcellini et al. 2005).
The positional specificity of the sc null mutant must be related in part to differences in the expression of other genes that somehow determine the probability of SMC formation for each amount and activity of proneural protein. These positional differences are likely to explain the puzzling observation that transient and generalized expression of Sc is able to direct bristle formation in the correct positions in homozygous AS-C mutant backgrounds (sc10.1), i.e., in wing discs lacking patterned expression of AS-C (Rodriguez et al. 1990). This inferred underlying layer of positional information may be conferred by the heterogeneous distribution of Emc, which affects the activity of AS-C proteins (Cubas and Modolell 1992), and by the heterogeneous expression of other genes involved in “lateral inhibition” (Vassin et al. 1987; de Celis and Garcia-Bellido 1994; Parks et al. 1997; Joshi et al. 2006). In this manner, it appears that bristle positions are determined both by the restricted expression of AS-C genes, controlled by modular enhancers, and by the heterogeneous expression of other genes that modulate the proneural activity of the Ac/Sc proteins or locally modify the response of the tissue to the neuralizing effects of these proteins.
Several related questions, not yet satisfactorily solved, have emerged from the description of AS-C expression and the dynamics of SMC appearance and differentiation. The first question relates to the nature of the elusive positive regulators responsible for the activation of AS-C expression in each proneural cluster. The second issue concerns the mechanisms of SMC selection among the cells expressing AS-C in the proneural cluster. Finally, the actual role of the AS-C proteins in conferring neural potential remains mysterious, in part because the genes ac and sc are no longer expressed once the SMC starts its differential divisions (Cubas et al. 1991); only ase expression persists in the SMC (Dominguez and Campuzano 1993).
Cis-REGULATION OF THE AS-C
The understanding of the regulation of AS-C expression followed the complementary approaches of (1) dissecting each regulatory region (the “position-specific enhancers”) by making fusion constructs with a reporter gene and (2) searching for mutants affecting the formation of specific subsets of macrochaetae. The first mutation identified as a candidate to participate in the position-specific activation of the AS-C in the thorax was named iroquois (iro), because it caused a phenotype in which several of the lateral macrochaetae were missing (Dambly-Chaudière and Leyns 1992; Leyns et al. 1996). The genetic and molecular analysis of iro uncovered yet another gene complex, the iro-C, formed by three genes, caupolican, araucan, and mirror (Gomez-Skarmeta et al. 1996; McNeill et al. 1997; Kehl et al. 1998). These genes encoded related nuclear proteins characterized by the presence of a conserved homeodomain. They have different functions depending of the developmental context, and during the appearance of the proneural clusters they are expressed in a pattern that partially overlaps some of the clusters (Cavodeassi et al. 2001). The mechanism of action of the Iro proteins is still unknown, and although they are required for the correct expression of the AS-C genes in the most lateral proneural clusters, they seem to act as transcriptional repressors (Cavodeassi et al. 2001; Bilioni et al. 2005). Other candidate transcriptional activators of AS-C expression were also identified by virtue of their restricted expression pattern in the thorax and their effects on specific macrochaetae. For example, the GATA-containing protein Pannier (Pnr) is expressed in the region from which the dorsocentral macrochaetae form and directly regulates the expression of ac/sc by binding to a specific AS-C enhancer (Garcia-Garcia et al. 1999). Similarly, the Zn-finger proteins Spalt and Spalt-related are also expressed in specific domains of the thorax and are required for the formation of the anterior notopleural macrochaetae (de Celis et al. 1999). The identification of ac/sc activators in the thorax suggests that pattern formation in this tissue is the consequence of a progressive deployment of transcription factors whose expression is restricted to specific territories. Therefore, the epithelium contains a landscape of transcription factors acting in a combinatorial manner to confer a genetic identity on each region of the thorax. This landscape has been referred to as the prepattern, following the classic definition that Stern used to explain the competence to develop bristles in genetic mosaics bearing ac mutant clones (Stern 1954; Ghysen and Dambly-Chaudiere 1988; Campuzano and Modolell 1992). In this scenario, the AS-C cis-regulatory regions work as a decoding device that reads out different combinations of transcriptional regulators, the prepattern proteins, and converts them into ON and OFF states of transcription for both ac and sc, resulting in the formation of individual proneural clusters (Gomez-Skarmeta et al. 2003). Much work is still needed to understand the dynamics of proneural cluster formation and extinction and the manner in which they then relate to the singling out of individual cells in constant positions. This is the final issue addressed in this overview.
CELL INTERACTIONS WITHIN PRONEURAL CLUSTERS
The transition from proneural clusters to individual SMCs became a paradigmatic example of a patterning mechanism that refines cell commitment from groups of cells to individual cells. Two classic observations relate to this mechanism. First, in the pioneering description of the development of Notch mutant embryos by Poulson in the 1940s, it was reported that an excess of neural tissue in Notch mutant embryos developed at the expense of the epidermis (Poulson 1940). This observation suggested that ventral ectodermal cells have the potential to develop as neural elements and that Notch activity was somehow involved in the repression of this fate, therefore allowing the formation of epidermal cells. Thus in the Notch mutant embryo all ventral ectodermal cells follow what was understood to be the primary fate, i.e., neural development. Curt Stern, who analyzed the behavior of ac mosaics in the thorax, made the second key observation: when ac mutant tissue includes the position of the anterior or the posterior dorsocentral bristles, they fail to differentiate (see above), but instead, in a number of cases, a nonmutant macrochaeta appears close to, but not in, the wild-type position (Stern 1954). This indicated that several cells near the position of a normal macrochaeta are competent to form a bristle, and that in a normal situation the formation of one bristle in this field prevented other cells from accomplishing the same fate. These observations were followed by cell ablation experiments, carried out in grasshopper embryos, which indicated that the epidermal cells in the vicinity of a developing neuroblast entered the neural pathway when this neuroblast was killed (Taghert et al. 1984). Several authors realized that the Notch phenotype and the mechanism of cell fate inhibition by the SMC or neuroblast (lateral inhibition) were related phenomena (Knust and Campos-Ortega 1989). Furthermore, the mutagenesis screens carried out by Nusslein-Volhard and Wieschaus (1980) identified additional genes with hyperplasic CNS (the “neurogenic” phenotype) (Campos-Ortega and Knust 1990). This opened the possibility of connecting the function of a group of genes (the neurogenic genes as they were called) with the cellular mechanism of lateral inhibition (Knust and Campos-Ortega 1989; Campos-Ortega 1993). Today we know that the neurogenic genes encode members of a universally conserved signal transduction pathway, the Notch signaling pathway, that during SMC singling out prevents the accumulation of proneural protein in a cell by interfering with a loop of ac and sc self-stimulation mediated by SMC-specific enhancers present in the AS-C (Culí and Modolell 1998; Giagtzoglou et al. 2003). Thus, the Ac/Sc proteins in the proneural cluster cells promote activation of Delta, the ligand of the pathway, which in turn activates the Notch receptor in neighboring cells. This impairs the activity of the SMC-specific enhancers and maintains most of the cells of a proneural cluster in a non-SMC state (“mutual inhibition”). The cell with the highest levels of Ac and Sc escapes from the inhibition, activates the SMC enhancers, accumulates maximal levels of Ac and Sc, and becomes the SMC. The SMC then signals most strongly to the remaining cells of the cluster and prevents them from becoming additional SMCs (lateral inhibition). In summary, by linking Ac/Sc expression in proneural clusters to Delta and Notch signaling, and this to repression of the SMC-specific enhancers, differences in proneural gene activity lead to the selection of single SMCs. The detailed molecular analysis of the regulatory relationships between the Notch signaling pathway and the proneural genes is still a work in progress (Jennings et al. 1994; Giagtzoglou et al. 2003; Castro et al. 2005; Acar et al. 2006; Pi and Chien 2007).
BEYOND NEUROGENESIS: OTHER ROLES OF THE AS-C GENES
The AS-C was perhaps the first example of a gene or group of genes positively linked to a key developmental decision, that of forming a neural precursor, and consequently the AS-C has been mostly studied in the context of neurogenesis. Surprisingly, a more exhaustive analysis of its expression pattern and phenotype uncovered several functions not related to neural development. For example, the AS-C genes are expressed in clusters of mesodermal cells, from which muscle progenitors form through a mechanism of lateral inhibition mediated by Notch signaling (Bate et al. 1993; Carmena et al. 1995). In this system, the loss of AS-C function leads to the absence of individual muscle progenitors. Similarly, the AS-C genes are also required for cell fate assignment of specific cells in the gut, a tissue of endodermal origin (Tepass and Hartenstein 1995). In this manner, a key invariant aspect of the AS-C genes is their participation in the selection of committed cells from groups of competent cells, through processes of lateral inhibition. One can speculate that the connection between the AS-C and the Notch pathway is phylogenetically old and has been retained during evolution and adapted to a variety of developmental contexts as a device ensuring single-cell resolution in cell-fate allocation.
Since the identification of the AS-C proteins as transcription factors bearing a bHLH domain, many orthologs have been identified and characterized within the framework of the developmental mechanisms of sensory organ pattern formation (Bertrand et al. 2002; Sugimori et al. 2007). The proteins belonging to the AS-C family identified in other invertebrate and vertebrate genomes share functional features with the fly orthologs. Thus most vertebrate AS-C genes are expressed principally in the developing nervous system, where they participate in the selection of neural progenitor cells and mostly in the differentiation of specific neuronal lineages. As happens in the fly, the expression of vertebrate proneural genes is turned off before the progenitor cell begins to differentiate, suggesting that a key aspect of their function is to initiate a cascade of transcriptional regulation leading to sequential steps of cell determination and differentiation (Bertrand et al. 2002; Chang et al. 2008).
The analysis of the AS-C uncovered several trends common to many developmental processes and provided a framework to dissect the molecular bases of pattern formation, regional specification, and cell commitment. It is a very illustrative example of the difficulties of applying genetic analysis to complex genes, because many sound and internally consistent proposals could be contrasted and accounted for only after the cloning and molecular study of the genes. Apart from telling the story of the analysis of chaetae pattern formation, the AS-C has been instrumental in understanding the organization of eukaryotic genes, with their complex arrays of cis-regulatory modules influencing the expression of adjacent transcription units, and was also a key entry point for the analysis of the genetic subdivisions of developmental territories by partially overlapping domains of gene expression (prepattern). Finally, the identification of the AS-C genes as proneural also helped to identify their vertebrate counterparts and to begin to understand the molecular mechanisms of neural cell-type specification.
We thank J. Modolell, S. Campuzano, and M. Ruiz-Gomez for critically reading this manuscript and for suggestions that greatly improved it, and many colleagues and friends that immensely contributed to the unwinding of the AS-C tale. We also thank J. Modolell and S. Campuzano for the gift of original pictures shown in Figure 1 and Adam Wilkins for his constructive suggestions.
↵1 These authors contributed equally to this work.
- Copyright © 2009 by the Genetics Society of America