BIOLOGICAL OVERVIEW
In the course of a genetic mosaic screen for new mutations affecting early eye development, alleles of a conserved gene were isolated and the gene was called capulet or act up (acu) for its effect on actin. Since the first full publication describing this protein refers to it as Act up, it will be so termed in this report, but 'Capulet' is the accepted FlyBase name. acu encodes a homolog of yeast cyclase-associated protein (CAP) that sequesters monomeric actin; acu is required to prevent actin filament polymerization in the eye disc. acu is required for cells to change shape in the morphogenetic furrow, and it also prevents premature neuronal differentiation in the eye disc, probably by restricting the movement of Hedgehog (Benlali, 2000).

The actin cytoskeleton has been implicated in mediating cell shape changes in many systems. Filamentous (F) actin is assembled from monomeric (G) actin subunits, with kinetics that favor addition of subunits to the barbed ends of the filaments and dissociation of subunits from the pointed ends. This process is regulated by a variety of proteins. Profilin (Drosophila's Chickadee) binds to actin monomers and was originally thought to sequester them, preventing polymerization. However, profilin has now been shown to acts as a nucleotide exchange factor for actin and promotes the elongation of actin filaments at their barbed ends. ADF/cofilin stimulates pointed end depolymerization, increasing the G actin pool and promoting filament turnover. Capping and severing proteins also limit the length of filaments. GTPases of the Rho family appear to regulate actin filament dynamics in response to extracellular signals (Benlali, 2000 and references therein).

Cyclase-associated protein (CAP) was first identified in yeast as a protein required for RAS activation of adenylate cyclase (Fedor-Chaiken, 1990; Field, 1990; Kawamukai, 1992). Loss of CAP was also shown to alter the actin distribution in yeast cells and to make them grow poorly on rich medium; these defects are caused by mutations in its C-terminal domain and could be suppressed by overexpression of profilin (Gerst, 1991; Vojtek, 1991). Yeast CAP and its homologs in other species contain a C-terminal actin monomer-binding domain that can inhibit actin filament polymerization (Gieselmann, 1992; Freeman, 1995; Gottwald, 1996; Zelicof, 1996). This function has been highly conserved throughout evolution, as mouse and human CAP homologs can complement C-terminal CAP mutations in yeast (Matviw, 1992; Vojtek, 1993). Localization of CAP to cortical actin patches is mediated by a proline-rich region of the protein that binds SH3 domains (Freeman, 1996) (Benlali, 2000 and references therein).

The Drosophila CAP homolog, acu, is required for the organized progression of photoreceptor differentiation. In the eye disc, clones of acu mutant cells overlapping the morphogenetic furrow express markers of photoreceptor differentiation, such as Neuroglian, Atonal (Ato), and the decapentaplegic marker dpp-lacZ more anteriorly than surrounding wild-type cells. As differentiation proceeds from posterior to anterior, this represents an acceleration of differentiation in acu mutant cells. However, many cells in more posterior acu clones fail to differentiate as neurons and later appear to die, leaving scars in the adult eye. Removal of acu function from the entire eye disc by inducing clones in a Minute background with FLP recombinase expressed from the eyeless promoter results in a severe disorganization of the pattern of differentiation in the eye disc. dpp-lacZ is normally expressed in the morphogenetic furrow, preceding expression of the neuronal protein Elav. In the absence of acu, dpp-lacZ is expressed in very abnormal patterns and Elav-expressing photoreceptors appear to be randomly scattered over the eye disc (Benlali, 2000).

The N-terminal domains of the S. cerevisiae and S. pombe CAP proteins bind to and activate adenylate cyclase (Gerst, 1991; Kawamukai, 1992). Mutations in the S. pombe N-terminal domain can be rescued by human CAP, but the N-terminal domain of the S. cerevisiae protein appears to be divergent, as its function cannot be rescued by S. pombe or mouse homologs (Kawamukai, 1992; Vojtek, 1993; Yu, 1994). The C-terminal domains of the yeast CAP proteins sequester actin monomers to inhibit actin filament polymerization, and this function is conserved in the mouse, human, pig, and Dictyostelium homologs (Vojtek, 1991; Gieselmann, 1992; Vojtek, 1993; Yu, 1994; Freeman, 1995; Gottwald, 1996; Zelicof, 1996). To test whether Drosophila CAP also controls actin filament polymerization, rhodamine-conjugated phalloidin was used to stain actin filaments in cells in the eye disc. Phalloidin outlines the apical surfaces of wild-type photoreceptor cells, revealing their constriction in the furrow. In acu mutant clones of cells, the intensity of phalloidin staining is greatly increased. Actin filaments appear to fill the cell bodies, particularly near the apical surface of the disc, rather than being restricted to a small region of each ommatidium. Thus, acu is required to prevent excessive actin filament polymerization in the eye disc (Benlali, 2000).

Overexpression of UAS-acu throughout the early eye disc using an eyeless-GAL4 driver has no effect on photoreceptor differentiation, suggesting that Acu activity is either regulated or not limiting. CAP activity is thought to be regulated by binding to phosphatidylinositol biphosphate (PIP₂), which inhibits CAP function. Deletion of the N-terminal and proline-rich regions of the Dictyostelium CAP protein protects it from inhibition by PIP₂, increasing its ability to sequester actin monomers in vitro (Gottwald, 1996). Transgenic flies expressing only the corresponding C-terminal domain of Acu under UAS regulatory elements (UAS-acuC) were generated to determine whether Acu would act as a constitutively active form of the protein. This truncated protein includes the region homologous to an SH3-binding domain that is required for yeast CAP localization to cortical actin patches (Freeman, 1996; Yu, 1999). However, expression of UAS-acuC in the eye disc using ey-GAL4 or ubiquitously using da-GAL4 has no apparent effect on development. Because the AcuC protein is functional, this may indicate that Acu is not regulated in the same way as other CAP proteins. The role of PIP₂ in eye development is not yet clear; a phospholipase C encoded by small wing appears to negatively regulate the formation of the R7 photoreceptor, and a PIP 5-kinase encoded by skittles is required for cell survival in the eye (Benlali, 2000).

Because the N-terminal domains of the yeast CAP proteins stimulate adenylate cyclase activity, it is possible that an effect on adenylate cyclase could contribute to the acu mutant phenotype. Cyclic AMP-stimulated protein kinase (PKA) has been shown to antagonize Hh signaling in the eye, and loss of PKA function leads to premature photoreceptor differentiation. A reduction in adenylate cyclase activity, leading to lowered cAMP levels, could perhaps have a similar effect. However, premature photoreceptor differentiation has been observed only in acu clones overlapping the morphogenetic furrow, suggesting that this differentiation is still dependent on signals from more posterior cells. In contrast, pka mutant clones entirely anterior to the furrow are able to differentiate as photoreceptors. To test the role of the N-terminal domain of Acu, attempts were made to rescue the acu mutant phenotype by constitutively expressing only the C-terminal domain. Expression of UAS-acuC under the control of da-GAL4 is sufficient to rescue acu^E593/acu^E636 transheterozygotes to viability and to restore normal photoreceptor differentiation. The C-terminal domain may be slightly less active than the full-length protein, because one of the UAS-acuC insertions tested could only rescue acu mutants to the pupal stage and did not restore normal phalloidin levels to their eye discs. The defects observed in acu mutants thus appear to be due only to loss of function of the C-terminal domain of Acu, since they can be rescued by restoring this domain. However, the possibility cannot be ruled out that the truncated proteins produced in these mutants retain some N-terminal domain activity. Additional phenotypes might be seen if the Acu protein were completely deleted. It is concluded that the affects af acu mutation on eye development are not due to acu involvement in hh signaling through PKA (an N-terminal domain function), but instead involve interference with actin polymerization (a C-terminal domain function) (Benlali, 2000).

Another protein implicated in the regulation of actin polymerization is profilin, encoded by the Drosophila gene chickadee (chic). In yeast, overexpression of profilin can rescue the defects caused by mutation of the C-terminal domain of CAP (Vojtek, 1991), suggesting that profilin may act to sequester actin monomers in yeast. However, in other systems, profilin appears to stimulate actin polymerization. The phenotype of chic clones were examined in the eye disc. Unlike acu clones, chic clones show greatly reduce phalloidin staining. Profilin and CAP thus appear to have opposite effects on actin polymerization in the eye disc. Clones of cells doubly mutant for acu and chic predominantly show the chic phenotype of lack of phalloidin staining, although some cells near the edges of the clone showed high levels of staining. Thus, profilin is required for normal actin polymerization even in the absence of monomer sequestration by CAP (Benlali, 2000).

Surprisingly, chic clones overlapping the morphogenetic furrow, like acu clones, disrupt the timing of photoreceptor differentiation, leading to early expression of markers such as dpp-lacZ, Ato, and Elav. Large chic clones made in a Minute background show a disorganization and reduction of photoreceptor differentiation, and an uncoupling of the expression patterns of dpp-lacZ and Elav. To determine whether this uncoordinated differentiation is due to lack of actin polymerization, GTPases Rac and Cdc42, which are thought to act through the kinase PAK to promote actin polymerization were used. The effect of expressing a dominant-negative form of the GTPase Drac1 in the eye disc using ey-GAL4 was examined. This method of inhibiting actin polymerization also results in uncoordinated differentiation, with an abnormal sequence of dpp-lacZ and Elav expression in some cells. A dominant negative form of Dcdc42 has milder effects, while expression of activated forms of either Drac1 or Dcdc42 with ey-GAL4 leads to an extreme reduction in the size of the eye disc (Benlali, 2000).

Since premature photoreceptor differentiation was observed both in acu mutant clones, in which actin filament levels are increased, and in chic clones, in which actin filament levels are decreased, an effect on cell shape might be common to both mutations. Using an antibody to the membrane-associated protein Armadillo (Arm), it was observed that acu mutant cells in the region of the morphogenetic furrow do not undergo the normal shape changes, and instead retain large apical profiles. The same phenotype was observed in chic mutant clones. This suggests both that coordinated alterations in actin filaments are required for apical constriction of cells in the morphogenetic furrow, and that this cell shape change is important in controlling the pattern and timing of differentiation (Benlali, 2000).

One hypothesis that might explain these results is suggested by the observation that Hh is a concentration-dependent regulator of ato. The extent of Hh diffusion is thus critical to determine the position, and therefore the developmental stage, at which photoreceptors differentiate. The Hh protein is modified by cholesterol addition and N-terminal acylation, and its movement through tissues appears to be a highly regulated process requiring the sterol-sensing-domain protein Dispatched and heparan sulfate proteoglycan synthesis by the product of the tout velu (ttv) gene. The rate-limiting step for Hh movement might therefore be transfer between cells rather than diffusion through the extracellular space. Constriction of the apical profiles of cells increases the density of their packing at the apical surface, where signaling is thought to occur; it would thus increase the number of such transfers required to travel a given distance (Benlali, 2000 and references therein).

Using an antibody to Hh, a tight band of Hh protein was indeed observed at the morphogenetic furrow in wild-type cells, and more anterior Hh staining in both large and small clones of acu mutant cells. To test whether this ectopic anterior Hh is functional, the accumulation of full-length Cubitus interruptus (Ci) protein was examined. Such accumulation is normally associated with a shift of Ci from its repressor to its activator form, and is the earliest response to Hh reception that can be visualized. In clones of cells mutant for acu, accumulation of Ci is observed more anteriorly than in surrounding wild-type cells. chic mutant cells do not themselves accumulate Ci, suggesting that actin filaments may be required for the response to Hh, but wild-type cells anterior to chic clones show precocious Ci accumulation. Ectopic Hh signaling has been shown to be sufficient to induce apical constriction of surrounding cells. Consistent with the presence of ectopic Hh in the clones, apical constriction of cells has been observed at the lateral edges of acu and chic mutant clones and anterior to smaller clones. It thus appears that apical constriction of cells in the morphogenetic furrow prevents long-range Hh movement, perhaps by increasing the number of rate-limiting active transport steps required (Benlali, 2000).

Although this effect on Hh movement could explain the premature photoreceptor differentiation observed in the absence of acu, several alternative hypotheses have been considered. One possibility is that free actin could exert a signaling effect in the eye disc; it has been reported to regulate the activity of serum response factor. The results with profilin mutants argue against this possibility. Because opposite changes in the extent of actin filament polymerization in acu and chic mutant cells cause a similar disruption of both cell shape and the timing of photoreceptor differentiation, their effects on differentiation are most likely to be due to the lack of coordinated changes in cell shape. Consistent with this model, a reduction in the amount of another cytoskeletal protein, myosin light chain, causes eye phenotypes normally associated with disruption of morphogenetic furrow progression (Benlali, 2000 and references therein).

The shape of embryonic mesenchymal cells has recently been shown to be critical to direct their differentiation into smooth muscle, perhaps because mechanical tensions activate certain signaling pathways. Thus, cell shape changes in the morphogenetic furrow might have a direct effect on photoreceptor differentiation. Loss of both Hairy (H) and Extramacrochaetae (Emc), two HLH proteins that inhibit Ato function, can produce similar precocious photoreceptor differentiation, and it is possible that cell shape might influence their expression. However, both H and Emc are still present, though H expression is slightly reduced, in acu mutant clones. This may indicate that sufficient levels of Hh to completely inhibit H expression are not attained in these cells. Consistent with this, the effects of acu on the direct Hh targets dpp and Ato are more robust than its effects on later markers of differentiation that require Ato activity for their expression (Benlali, 2000).

Another possibility is that actin-based projections similar to the cytonemes described for the Drosophila wing disc (RamirezWeber, 2000) could contribute to the reception of Hh or other signals and would be disrupted by changes in actin filament polymerization. However, no region of the eye disc appears to be able to induce cytoneme formation. Thus, an effect of cell shape changes on the rate of Hh transport remains the most likely explanation of these results. In the wing disc, expression of the Patched (Ptc) receptor is upregulated anterior to Hh-expressing cells; Ptc then sequesters Hh to restrict its movement. Ptc upregulation does not appear to occur in the eye disc; in a dynamic system like the eye, cell shape changes may be a more rapid way to restrict Hh movement than changes in ptc gene expression. Resolving the mechanisms by which cell shape can affect differentiation will integrate the morphogenetic and signaling processes that are critical for development (Benlali, 2000).

PROTEIN STRUCTURE

Amino Acids - 511

Structural Domains

The ACU transcript encodes a homolog of the cyclase-associated protein (Benlali, 2000) originally identified in yeast and subsequently found in other species (Gerst, 1991; Matviw, 1992; Vojtek, 1993).

date revised: 13 July 2000

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.