BIOLOGICAL OVERVIEW

How are boundaries formed between various tissues? One solution to this problem is illustrated by the action of the selector gene apterous. Selector genes establish the autonomy and direct the development of compartments with respect to other adjoining compartments. In the selector affinity model (Garcia-Bellido, 1975) it is proposed that the presence or absence of selector gene expression controls the affinity of cells for each other. Simply put, selector-expressing cells prefer to associate with other selector expressing cells, and to not associate with non-selector-expressing cells. A boundary is formed between cell groups to minimize contact between them (Garcia-Bellido, 1975 and Blair, 1995).

Two groups of cells separated by their adhesive properties are called compartments. The cells in different compartments are lineage restricted, that is, cells of one lineage do not cross into another compartment made up of cells from another lineage, because of their higher self affinity.

Such a boundary process occurs in wing morphogenesis. apterous expressing cells in the dorsal domain of the wing produce one type of integrin (alpha PS1 beta PS), and cells in the ventral half produce another type of integrin (alpha PS2 beta PS). myospheroid codes for beta PS integrin; multiple edematous wings codes for alpha PS1 integrin, and inflated codes for alpha PS2 integrin.

Integrins are cell surface adhesion molecules that act like specific glue, attaching the cells bearing them to specific protein molecules on the surface of other cells or on extracellular matrix molecules in their vicinity. Integrins can communicate information about the extracellular environment to the inside of the cell, and thus alter cell fate. apterous mutation or ectopic expression results in mis-expression of the adhesive integrin molecules, resulting in a breakdown in dorsal/ventral separation of wing cells (Blair, 1994).

The way Apterous works appears to be complex, having at least two distinct outputs. First, Apterous is responsible for making the dorsal cell distinct from ventral, a property that may be due to its activating the gene Dorsal wing (Tiong, 1995). Second, Apterous regulates the expression of boundary determining proteins such as Fringe and Serrate. Serrate is thought to act as a ligand of Notch (Kim, 1995), an adhesive protein that can communicate to the inside of the cell, information about the extracellular milieu. Serrate appears to signal from dorsal to ventral cells to elicit the production of a long range morphogen, perhaps Wingless (Lawrence, 1996). Fringe is remarkable because a boundary forms wherever fringe-expressing and nonexpressing cells meet, a boundary that can organize long-range pattern (Irvine, 1994). Fringe is actually responsible for the induction of Serrate, by an apterous independent mechanism (Kim, 1995). These kinds of proteins insure the separation of cellular compartments and the ongoing developmental pathways of the separated cell lineages.

Serrate, known to be regulated by Apterous, is not required for the initiation of wing development but rather for the expansion and early patterning of the wing primordium. apterous is expressed in dorsal cells of the wing disc; in the absence of ap, the wing blade does not develop, an effect thought to be due to the loss of Ser expression. Consistent with this, ectopic expression of Ser, or of fringe (fng), which leads to the expression of Ser, rescues the loss of the wing in ap mutants. However, the Ser and ap mutant phenotypes are not identical: while in the absence of ap there is no trace of the wing blade, Ser mutants do bear a small marginless wing blade. To explain this difference, the expression of wingless (wg) and vestigial (vg) were compared during wing development in these mutants. The expression of vg decays at the beginning of the third instar in Ser mutants, but is never activated in the wing region of ap mutants. In Ser mutants, the expression of wg never spreads along the wing disc DV interface and resolves into two rings, which fate map the small wing characteristic of Ser mutants. However, in ap mutants, wg expression comes to outline a single circle of expression, which defines proximal hinge structures and the absence of a wing blade, a deficit that is characteristic of these mutants. This is a phenotype very similar to that of vg null alleles. These results suggest that the phenotype of ap mutants cannot be accounted for simply by the absence of Ser. In ap mutants, the development of the wing blade is never initiated; in the absence of Ser, this process is initiated normally, but is aborted early on. Furthermore, since the expansion of wg expression in the AP direction is required for the establishment of the proper size of the primordium, its failure to occur in Ser mutants indicates that, in addition to its role in the establishment of the wing margin, Ser is required to define the proper size of the wing primordium (Klein, 1998).

apterous is also also required for the expression of a second Notch ligand, Delta. The initial stages of the development of the wing blade require Notch signaling and lead to the activation of the vestigial boundary enhancer (vgBE) at the interface between dorsal and ventral cells. Ser is not involved in this event and therefore there ought to be another Notch ligand, under the control of ap, that is responsible for the activation of the vgBE. The product of the Delta gene is a good candidate for this function. During the second instar, Dl is expressed throughout the wing disc, but it is slightly upregulated over the ventral region and shortly afterwards, its pattern of expression is identical to that of vgBE, i.e. a 2- to 3- cell-wide stripe that straddles the DV interface. Furthermore, the expression of Delta is similar to that of the vgBE in Ser and ap mutant discs: in ap mutant wing discs, expression of Dl is lost at the time when the wing primordium is induced, whereas in Ser mutants expression is detected until early third instar. This suggests that Delta might be the activating ligand for the Notch-dependent expression of the vgBE, which operates in the absence of Ser. Consistent with this possibility, ectopic expression of Dl can rescue the loss of wing blade tissue and of wing margin characteristic of ap and Ser mutants (Klein, 1998).

Serrate, in turn, along with Delta, refines Notch function at the DV interface. After the establishment and expansion of the wing primordium, there is a new requirement for Notch signalling in the growth and patterning of the wing blade. In this process, both Serrate and Delta act as ligands for Notch and, as in earlier stages, have different patterns of gene expression, which suggests that they might have different functions. However, ectopic expression of either Dl or Ser will rescue the loss of wing tissue and of wing margin characteristic of ap mutants, and this raises the question of why are there two different ligands to achieve the activation of Notch in these early stages of wing development. It is believed that coexpression of both ligands might result in a different degree of Notch activation than would be achieved individually -- this might allow for a finer degree of regulation for Notch activity. In the presence of Serrate, Delta is able to signal although with a reduced activity level, which perhaps reflects a competition between Serrate and Delta for Notch (Klein, 1998).

A feedback mechanism exists in which Fringe functions to inhibit Serrate by targeting Notch. In contrast to Delta, the effects of ectopic expression of Ser on wild-type discs are restricted to ventral cells. This has led to the suggestion that there is an inhibitor of Serrate activity in dorsal cells and that this inhibitor is under the control of ap. Consistent with this proposal, ectopic expression of Ser in ap mutants is found to be able to induce the expression of downstream targets of Notch in 'dorsal' cells. A variety of arguments have led to the proposal that the dorsal inhibitor of Serrate function is encoded by the fng gene. For example, ectopic expression of Ser with ptcGAL4 results in the activation of Notch targets in two parallel stripes in ventral cells of the developing wing blade, and this can be observed as early as the beginning of the third instar. When Ser is coexpressed with fng, the anterior stripe, but not the posterior one, is lost completely in late third instar discs. Correspondingly the ectopically induced margin structures are reduced to a posterior stripe with characteristics of the posterior compartment. While Fringe suppresses the function of Serrate cell autonomously, it enhances its signaling ability in a nonautonomous manner. Fringe is thought to dampen Serrate signaling by affecting its interaction with Notch, but no evidence has been presented to support this suggestion. Increasing the concentration of Notch appears to be able to titrate the effects of fng. Furthermore, the effects of ectopic expression of fng are partially suppressed by the expression of Notch with fng and are exaggerated by expressing dominant negative Notch molecules with fng. Altogether, these results strongly suggest that a target of Fringe activity is the Notch molecule itself (Klein, 1998).

Fringe functions to inhibit Serrate signaling via Notch. The activity of Fringe can inhibit Serrate signaling by enhancing the intrinsic dominant negative activity of Serrate over Notch. Expression of Ser throughout the late wing disc leads to a strong broadening of the wing veins and a moderate increase in the number of bristles in the notum. Both of these neurogenic phenotypes can be suppressed by coexpressing Notch with Ser, indicating that they are due to a dominant negative effect of Serrate. Ectopic expression of fng alone in the same pattern results in nicked wings with normal veins and a reduction of bristles in the notum, which is associated with the loss of sensory organ precursors. Coexpression of fng with Ser suppresses the extra vein phenotype caused by misexpression of Ser and, therefore, supports the notion that Fringe reduces the ability of Serrate to bind Notch. Fringe is shown to impinge on Notch signaling by the observation that the action of Fringe requires the activity of Su(H). Fringe is not able to rescue the defects caused by Su(H) mutants (Klein, 1998).

Establishment of a developmental compartment requires interactions between three synergistic cis-regulatory modules

The subdivision of cell populations in compartments is a key event during animal development. In Drosophila, the gene apterous (ap) divides the wing imaginal disc in dorsal vs ventral cell lineages and is required for wing formation. ap function as a dorsal selector gene has been extensively studied. However, the regulation of its expression during wing development is poorly understood. This study analyzed ap transcriptional regulation at the endogenous locus and identified three cis-regulatory modules (CRMs) essential for wing development. Only when the three CRMs are combined, robust ap expression is obtained. In addition, the trans-factors that regulate these CRMs were genetically and molecularly analyzed. The results propose a three-step mechanism for the cell lineage compartment expression of ap that includes initial activation, positive autoregulation and Trithorax-mediated maintenance through separable CRMs (Bieli, 2015).

Genetic and cis-regulatory analysis has provided information about the logic of ap expression during wing development. It is proposed that ap expression is controlled by at least three CRMs that act in combination. The first element, apE is the earliest to be activated in proximal wing disc cells via the EGFR pathway; its expression subsequently weakens in the wing pouch. Deletion of this early enhancer (e.g., apDG12 or apC1345) completely abolishes wing formation. The asymmetry of ap expression to the proximal domain of the wing disc is probably due to the localized activation of the EGFR pathway by its ligand Vn and a distal repression by Wg signaling. The initial activation of the apE by the EGFR pathway was genetically and molecularly confirmed; however, other inputs are required for the continuous activation of this CRM in later wing discs (Bieli, 2015).

A few hours after apE activation, a second CRM, apDV, is activated in a subset of apE positive cells. In contrast to apE, apDV is restricted to the dorsal-distal domain of the wing pouch by direct positive inputs from Ap and Vg/Sd. The direct Ap autoregulatory input defines the time window when the apDV element is activated; apDV can only be active after the induction of Ap by the early enhancer (apE). It has been shown that Ap induces vg expression by triggering Notch signaling at the D/V boundary. Thus, the (direct) input of Vg/Sd on apDV can be regarded as an indirect positive autoregulation, which delimits the spatial domain where apDV can be actived. Consequently, the interface of Ap and Vg expression defines the region of apDV activity via positive autoregulation (Bieli, 2015).

The third ap CRM is the ap PRE/TRE region (apP), that, when deleted, leads to a strong hypomorphic wing phenotype (apc1.2b). The apP requires Trx input and maintains ap expression when placed in cis with the apDV and apE CRMs. Only the combination of the three CRMs faithfully reproduces ap expression in the wing disc. Moreover, the regulatory in locus deletion and in situ rescue analysis provide strong functional relevance for these CRMs (Bieli, 2015).

Ultimately, this cascade of ap CRMs provides a mechanism to initiate, refine and maintain ap expression during wing imaginal disc development, in which the later CRMs depend on the activity of the early ones. A similar mechanism has been described for Distal-less (Dll) regulation in the leg primordia where separate CRMs trigger and maintain Dll expression in part by an autoregulatory mechanism (Bieli, 2015).

It has been proposed that positive autoregulation may help to maintain the epigenetic memory of differentiation. In the case of ap, this study demonstrates that autoregulation works in conjunction with a PRE/TRE system; this might make the system very robust and refractory to perturbations (Bieli, 2015).

ChIP experiments have shown that many developmentally important genes are associated with a promoter proximal PRE as found at ap. The role of such a PRE has been studied at the engrailed (en) locus. It has been demonstrated that in imaginal discs, the promoter as well as the promoter proximal PRE are important for the long-range action of en enhancers. It has been proposed that this PRE brings chromatin together, allowing both positive and negative regulatory interactions between distantly located DNA fragments (Bieli, 2015).

The current results indicate that sequences around the transcription start of ap (apP) may serve a similar function. First, this element, when placed in cis with the ap CRMs (apE and apDV), maintains the ap expression pattern and keeps reporter gene expression off in cells where low or no activity of apDV and apE has been observed. Second, in the absence of trx, the expression of ap and apDV+E+P-lacZ is strongly reduced. All these data suggest that sequences within the apP integrate Trx input, thereby maintaining ap expression in a highly proliferative tissue such as the wing disc. Interestingly, trx mutant clones were not round and did not show ectopic wg activation, which is a hallmark of ap loss-of-function clones. This suggests that in trx mutant clones enough Ap protein is still present to maintain wg expression off. However, derepression of the ventral-specific integrin αPS2 was found in trx mutant clones in the wing pouch as previously described for ap mutant clones (Bieli, 2015).

It has been suggested that TrxG proteins could act passively antagonizing PcG silencing, rather than playing an active role as co-activators of gene transcription. For example, Ubx expression in the leg and haltere does not require Trx in the absence of Polycomb repression. These possibilities were tested and trx mutant clones were generated that were also mutant for the PcG member Sex combs on midlegs (Scm). Dorsally-located Scm- trx- double mutant clones still downregulate ap-lacZ expression while ventral-induced ones are unable to derepress ap-lacZ as was observed for Scm- single mutant clones. Therefore, the results suggest that TrxG maintains ap expression in dorsal cells, while ap expression is repressed in the ventral compartment by PcG proteins. Moreover, it has been shown that the sequences around the ap transcription start, including the PRE, are occupied by PcG complexes PRC1 and PRC2, as well as Trx (Bieli, 2015).

Enhancers-promoter interactions initiate transcription but their dynamics during development have remained poorly understood. A Chromosome conformation capture (3C) experiment provides evidence for the direct interaction between the ap CRMs apE and apDV with the maintenance element encoded by the apP. Beyond this, it was also found that these elements cooperate continuously during wing development. Flip-out experiments, in which the apDV and apE CRMs were removed at different time points, suggest that these elements need to be present continuously to ensure correct ap expression. Additionally, flies carrying apE only on one chromosome and apDV only on the homologue were unable to fully rescue wing development suggesting that these CRMs need to be in cis. It is conceivable that in cis configuration of the three ap CRMs facilitates and stabilizes enhancer-promoter looping. It could also help to rapidly establish relevant chromatin contacts after each cell division. These results are in accordance with previous observations, in which constant interactions between ap enhancers and promoter during embryogenesis have been described. The current results extend these observations to the wing disc, a highly proliferative tissue, where the expression of the trans-factors that regulate the activity of the apE and apDV is very dynamic. This raises the question on how this contact is re-assembled over many cell generations. It is possible that some epigenetic modifications are laid down in the activated apE and apDV CRMs, which are then inherited during cell divisions to ensure contact with apP. Studies of the chromatin status of these elements will be required to fully understand this process (Bieli, 2015).

A key question in developmental biology is how transcriptional regulation is coupled to tissue growth to precisely regulate gene expression in a spatio-temporal manner. For example, during Drosophila leg development, initial activation of the ventral appendage gene Dll by high levels of Wg and Dpp initiates a cascade of cross-regulation between Dll and Dachshund (Dac) and positive feedback loops that patterns the proximo-distal axis. Other mechanisms to expand gene expression patterns depend on memory modules such as PREs, as it is the case for the Hox genes or other developmental genes like hh. To direct wing formation, expression of ap in the highly proliferative tissue of the wing disc must be precisely induced to generate and maintain the D/V border. These in-depth analyses at the ap locus provide a functional and molecular explanation of how expression of this dorsal selector gene is initiated, refined at the D/V border, and maintained during wing disc development. It is proposed that this three-step mechanism may be common for developmental patterning genes to make the developmental program robust to perturbations (Bieli, 2015).

GENE STRUCTURE

Bases in 5' UTR - 941

Exons - six

Bases in 3' UTR - 1759; within the 3'UTR are several copies of ATTTA thought to result in transient instability.

PROTEIN STRUCTURE

Amino Acids - 469

Structural Domains

In addition to two Lim domains, Apterous contains a C-terminal homeodomain (Bourgouin, 1992).

The predicted Islet protein contains two LIM domains and a C-terminal homeodomain, with extensive homology to the vertebrate Islet-1 and Islet-2 proteins. The homology is highest in the homeodomain (95% identity to Islet 1 and 2) and somewhat lower in the LIM domains (85%). Overall the Drosophila and vertebrate proteins show 57% identity. A highly conserved 16 amino acids stretch located C-terminal to the homeodomain denotes the Islet-specific domain, and is found in all members of the Islet subfamily, but not in other LIM-HD proteins. The Drosophila Islet homolog shows greater similarity to the vertebrate Islet proteins than to other Drosophila LIM-HD proteins. For example, within the homeodomain, Drosophila Islet shows only 38% amino acid identity to Apterous, which is 92% identical to its vertebrate homolog LH-2. Mutations affecting the isl locus completely abolish immunoreactivity with antibodies that recognize both vertebrate Islet-1 and Islet-2, suggesting that only a single islet homolog exists in Drosophila (Thor, 1997).

The LIM domain is a cysteine-rich domain composed of 2 special zinc fingers joined by a 2-amino acid spacer. Some proteins are made up of only LIM domains, while others contain a variety of different functional domains. LIM proteins form a diverse group that includes transcription factors and cytoskeletal proteins. The primary role of LIM domains appears to be in protein-protein interaction, through the formation of dimers with identical or different LIM domains or by binding distinct proteins. In LIM homeodomain proteins, LIM domains seem to function as negative regulatory domains. LIM homeodomain proteins are involved in the control of cell lineage determination and the regulation of differentiation, and LIM-only proteins may have similar roles. LIM-only proteins are also implicated in the control of cell proliferation since several genes encoding such proteins are associated with oncogenic chromosome translocations. In analyzing sequence relationships among various LIM domains it is suggested that they may be arranged into 5 groups that appear to correlate with the structural and functional properties of the proteins containing these domains. All N-terminal LIM domains (LIM1) are segregated into cluster A, whereas all LIM2 domains of the same proteins constitute cluster B. This relationship suggests that the putative duplication leading to the LIM A and B domains is ancient, preceding their association with different structural motifs (e.g., homeodomains, kinases). Furthermore, the sequence relationships between the LIM domains (LIM1 and LIM2) in the same protein may be conserved by functional constraints based on cooperation between LIMA and B domains. In contrast, the two type C LIM (another LIM domain cluster) domains of some of the LIM-only proteins like CRP are more similar to one another, implying the possibility of a more recent duplication. Cluster D is a rather divergent set of LIM domains that includes the cytoskeletal proteins Zyxin and Paxillin. The closest homologs of Apterous, for both LIM1 and LIM2 domains, are human and rat LH2. The Islet LIM1 and LIM2 domains define Islet as a cohesive subfamily of LIM proteins (Dawid, 1995).

The presence of the LIM domain of mammalian Isl-1 (Drosophila homolog: Islet) inhibits binding of the homeodomain to its DNA target. This in vitro inhibition can be released either by denaturation/renaturation of the protein or by truncation of the LIM domains. A similar inhibition is observed in vivo using reporter constructs. LIM domains in a chimeric protein can inhibit binding of the Ultrabithorax homeodomain to its target. The ability of LIM domains to inhibit DNA binding by the homeodomain provides a possible basis for negative regulation of LIM-homeodomain proteins in vivo (Sanchez-Garcia, 1993).

date revised:  25 April 2024 

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

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