BIOLOGICAL OVERVIEW

The diaphanous (dia) locus plays a critical role during cytokinesis: various combinations of dia mutations result in multinucleate spermatids, polyploid larval neuroblasts and adult follicle cells, and pupal lethality (Castrillon and Wasserman, 1994). Subcellular characterization of dia-deficient spermatids reveals that inactivation of Diaphanous results in defects in the interzonal microtubules, structures known to be crucial for the execution of cytokinesis, and also in the absence of an actomyosin contractile ring (Giansanti, 1998). Embryos deficient for dia were analyzed for three cell-cycle-regulated actin-mediated events during embryogenesis: formation of the metaphase furrow, cellularization and formation of the pole cells. In dia embryos, all three processes are defective. Diaphanous is a good candidate for a protein that functions at the cell cortex (the intracellular region just below the plasma membrane), through which microtubules signal actin organization (reviewed by Wasserman, 1998). Indeed, recent studies in budding yeast provide a clue to the potential position of Diaphanous in positioning and assembly of cytoskeletal structures. Diaphanous homologs, FH (formin homology domain) proteins Bni1p and Bnr1p, mediate cross-talk between cortical actin and the spindle structure; this cross-talk is necessary for correct positioning of the spindle in the bud neck (Lee, 1999; Miller, 1999). Since an interaction between the mitotic spindle and the cell cortex is also thought to determine the position of the animal cell contractile ring, Diaphanous could similarly act in a pathway linking microtubule and microfilament organization (Afshar, 2000).

FH proteins contain several conserved sequence elements distributed over the primary sequence (Wasserman, 1998). A centrally located, proline-rich FH1 domain in fission yeast has been demonstrated to serve as a binding site for the actin-binding protein profilin (Drosophila homolog: Chickadee) (Chang, 1997). Functional domain analysis of a fission yeast FH protein, Fus1, has identified an amino-terminal FH3 homology domain required for proper localization (Petersen, 1998). In addition, the amino-terminal domain of several FH proteins interacts with Rho-GTPase proteins (Evangelista, 1997; Imamura, 1997; Kohno, 1996; Watanabe, 1997). Much less is known about the function of the carboxyl-terminal halves of FH proteins, which include the highly conserved FH2 region and, typically, flanking coiled-coil domains (Afshar, 2000).

To investigate the role of Diaphanous in the actin-mediated processes of the syncytial blastoderm, the phenotype of dia-deficient embryos was examined during early development. To eliminate the maternal contribution of Diaphanous to the oocyte, homozygous dia mutant clones in the female germline were generated. Two alleles of dia were used for the generation of these clones. dia⁵ is a null allele that causes larval and pupal lethality. dia⁹ is weaker than dia⁵, with some homozygous mutant flies surviving to adulthood. Fertilized embryos produced by ovarian clones of dia mutant tissue exhibit severe developmental defects. Only about 3% of the embryos produced by dia⁵clones hatch, whereas about 20% of those from dia⁹clones survived through the larval and pupal stages. Although similar results have been obtained for both mutant alleles, the phenotype of the dia⁵embryos is more severe at all stages of embryogenesis. Therefore, all of the experimental results presented below are from embryos produced by dia⁵clones (hereafter referred to as dia mutant embryos) (Afshar, 2000).

Defects in dia mutant embryos first appear at nuclear cycle 11; earlier stages, as assessed by nuclear migration, division and organization, appear as wild-type. Abnormalities in nuclear and actin cytoskeletal organization affect almost two-thirds of all fertilized embryos at cycles 11-13 and a higher percentage at later stages. Among embryos of a similar stage, the surface area affected varies considerably, ranging from a set of small patches to the entire embryonic surface. This variability does not extend to all phenotypes, however, because 100% of the embryos fail to form any pole cells. Furthermore, greater than 95% of embryos are grossly defective at gastrulation, despite the fact that half receive a wild-type copy of dia paternally. Cuticle preparations of dia mutant embryos reveal a wide range of phenotypes, including failure in head involution, loss of head structures, reduction or absence of denticle bands and incomplete formation of the cuticle (Afshar, 2000).

To explore the nature of the defects seen in the absence of diaphanous function, wild-type and dia mutant embryos were stained at nuclear cycles 11-13 with the DNA dye DAPI and with an antibody directed against F-actin. In the wild type, nuclei are positioned at the embryo cortex at interphase of nuclear cycles 11-13; a structure referred to as the actin cap is situated between each nucleus and the plasma membrane. During the transition to prophase, filament reorganization results in a concentration of actin at the edge of the caps. At metaphase, the resulting rings of cortical actin, together with associated plasma membrane, invaginate to form metaphase furrows. As viewed from above, actin staining at these furrows appears as a hexagonal array over the embryonic surface. In the sagittal view, actin staining at the metaphase furrow appears as a line between the metaphase nuclei. In dia-deficient embryos, severe structural changes in the actin cytoskeleton are manifested after nuclear cycle 11. Formation of the hexagonal actin arrays is disrupted during prophase and metaphase and there is an absence of actin staining between the metaphase nuclei. Similar patterns of staining are obtained when dia embryos are stained with antibodies directed against anillin (Drosophila gene: Scraps) and Peanut, other components of the metaphase furrow. There is thus a failure in the formation of the metaphase furrow. Consistent with the known role of metaphase furrows in maintaining nuclear organization, the nuclei in dia mutant embryos frequently exhibit abnormal spacing and, in some cases, fuse in subsequent nuclear cycles. These irregularities are readily apparent in contrast to the uniform pattern observed in the wild type. In regions in which cortical actin staining is weak or absent, nuclei are frequently found displaced into the interior of the embryo, although the centrosomes remain at the surface (Afshar, 2000).

To investigate whether the absence of metaphase furrows results from a failure in membrane invagination, dia embryos were stained with antibodies directed against myosin (also known as Zipper). In wild-type embryos, myosin localizes to the embryonic cortex between the actin caps at each interphase, appears at the tip of the invaginating membrane at prophase and disappears at metaphase. In dia embryos, myosin staining, albeit very weak and irregular, is detected between the actin caps at the cortex during interphase. At prophase, myosin, where detectable, remains at the cortex, with no detectable membrane pinching or invagination. Therefore, despite the presence of myosin at the cortex between actin caps, the membrane invagination that precedes metaphase furrowing is absent in dia embryos (Afshar, 2000).

Immunolocalization was used to determine whether Diaphanous plays a role in the recruitment of anillin and Peanut, a Drosophila septin. In wild-type embryos both anillin and Peanut localize to the embryonic cortex, between the actin caps at interphase. During prophase and metaphase, they localize to the metaphase furrow and their pattern of staining is similar to that of actin. In dia embryos, the staining patterns of both anillin and Peanut are very weak during interphase. Similarly, in dia embryos the localization of both of these proteins is disrupted during prophase and metaphase, when the metaphase furrow is being formed in wild-type embryos. Diaphanous is thus required for recruitment and proper localization of anillin and Peanut as well as myosin to the regions of membrane invagination (Afshar, 2000).

Following the 13 syncytial nuclear division cycles, wild-type cellularization initiates. Actin-associated membrane invaginates between neighboring nuclei. This invagination is accompanied by the growth of microtubules, which extend from the pair of centrosomes above each nucleus, to form a basket-shaped structure that eventually surrounds each nucleus. The subsequent formation of individual cells involves further growth of the membrane and an actomyosin-mediated contractile event that pinches off the membrane at the bottom of each nucleus. In dia embryos, there is a variable defect in the organization of both actin- and microtubule-based structures during cellularization. In the least severe cases, the cellularization furrow is absent between some nuclei, without any noticeable defect in morphology or positioning of nuclei or microtubule structure. In more severely affected embryos, actin staining is absent at the furrow canals and irregular at some regions of the cortex. Surface regions that lack any organized actin display abnormalities in the positioning of both nuclei and microtubule baskets. Viewed from en face, such embryos display readily apparent irregularities in the hexagonal actin and microtubule arrays and nuclear positioning. These abnormalities are manifested as starburst arrays of nuclei and associated microtubules, with the nuclei tilted outward. In such regions, actin staining is faint or absent at the cortex, although some patchy staining is detected deeper in the embryo. In addition, gamma-tubulin staining reveals that centrosomal behavior is abnormal in these regions. This phenotype is similar to that observed in embryos treated with Cytochalasin D, in which the disruption of the cortical actin results in misorganization of the nuclei and mislocalization of microtubule baskets. In the most severely affected dia embryos the defective cortical actin and arrays of misoriented nuclei are observed over the entire embryonic surface. The localization of both anillin and Peanut in dia embryos is abnormal in regions of nuclear misorientation, but is wild-type elsewhere. For example, anillin, which can be detected in the nuclei of wild-type embryos only after cellularization, is present in the tilted nuclei in dia embryos during cellularization (Afshar, 2000).

Diaphanous is necessary for pole cell formation. During interphase of the wild-type cortical nuclear divisions, actin cap formation causes protrusion of the plasma membrane and cytoplasm around each nucleus, forming a cytoplasmic bud. At the posterior pole, in contrast to the rest of the embryo, these cytoplasmic buds grow extensively during nuclear cycle 10. Cytokinesis at the base of each bud results in the formation of a set of pole cells, progenitors of the adult germline. In dia mutant embryos, formation and growth of the cytoplasmic buds at the posterior pole is wild-type. Such buds, however, never cleave to produce pole cells. Rather, they regress in synchrony with the buds covering the rest of the embryonic cortex. Buds reform at the posterior pole at each nuclear cycle, but do not undergo cytokinesis. In some embryos, the number and size of the somatic buds are abnormal compared to wild-type embryos. Moreover, unlike the somatic nuclei, the posterior pole nuclei fail to initiate cellularization. To investigate the basis for the failure in pole cell formation, the presence was sought of an actomyosin contractile ring at the base of the posterior cytoplasmic buds. In contrast to wild-type embryos, dia embryos lack any concentration of actin or myosin at the base of the cytoplasmic buds, suggesting the contractile process does not start in the absence of Diaphanous (Afshar, 2000).

The phenotypic analysis of dia embryos indicates a requirement for Diaphanous for nuclear organization, metaphase furrow formation, cellularization and pole cell formation. To investigate the function of Diaphanous in these actin-mediated events, anti-Diaphanous sera was used to assay the spatial and temporal pattern of localization of Diaphanous during early embryonic development. During the first ten nuclear division cycles, no Diaphanous localization to specific structures was detected. Once the nuclei migrate to the cortex, however, Diaphanous staining appears as a hexagonal array at the surface of interphase and prophase embryos. At interphase, actin filaments remain in cap structures and Diaphanous localizes to the site of formation of the metaphase furrow prior to any significant redistribution of actin, which occurs at prophase. This is readily apparent in a sagittal view, revealing intense Diaphanous staining between the interphase actin caps. By prophase, Diaphanous is abundant at the tip of the metaphase furrow; this intense localization continues through metaphase. During cellularization, Diaphanous is enriched at the tip of the cellularization front and at the site of membrane invagination. As cellularization ends, Diaphanous localizes to the basal surface of each newly formed cell, where a contractile event pinches off the membrane at the base of the nuclei to produce individual cells (Afshar, 2000).

It is thought that specific signals, acting through Diaphanous, could assign actin cytoskeletal function for a specific cortical event. Different sets of Rho proteins and their regulators could recruit and/or activate Diaphanous in particular locations and at particular times. Diaphanous has a genetic interaction with Rho1 and the RhoGTPase exchange factor encoded by the Drosophila pebble gene. Pebble is strictly required for cytokinesis, but not for any membrane invagination events in the syncytial blastoderm. In addition, the immunolocalization of Diaphanous in dividing cells closely overlaps that of Pebble, being nuclear during telophase and interphase and located at the cleavage furrow at anaphase. Thus, one can imagine that Diaphanous function is controlled by Pebble at anaphase during cytokinesis, whereas other factors regulate Diaphanous function during early embryogenesis (Afshar, 2000).

Additional roles in cell regulation in the mouse have been revealed by studies of Diaphanous-related formin (DRF) Rho GTPase binding proteins mDia1 and mDia2. The DRF proteins couple Rho and Src during signaling and the regulation of actin dynamics. The DRFs are required for cytokinesis, stress fiber formation, and transcriptional activation of the serum response factor (SRF). 'Activated' mDia1 and mDia2 variants, lacking their GTPase binding domains, cooperate with Rho-kinase or ROCK to form stress fibers but independently activate SRF. Src tyrosine kinase associates and co-localizes with the DRFs in endosomes and in mid-bodies of dividing cells. Inhibition of Src also blocks cytokinesis, SRF induction by activated DRFs, and cooperative stress fiber formation with active ROCK (Tominaga, 2000).

The DRF proteins complex with multiple regulatory factors and bridge different signaling pathways. These multimeric complexes are likely to be highly dynamic, regulated by cycles of GTP hydrolysis and binding to signaling molecules such as Src. In yeast, deregulation of the yeast DRF Bni1 also triggers the formation of actin cables, suggesting that it is an active scaffold that recruits cellular machinery capable of polymerizing actin. In this regard, Bni1 or Bnr1 has been shown to bind multiple actin binding proteins, including elongation factor 1, profilin, and Bud6/Aip3p and appears to bridge two small GTPase regulated signaling pathways. Bni1 is found in high molecular weight complexes with docking components such as Spa2 that in turn bind constituent kinases from both the Cdc42/pheromone and Rho1/cell wall integrity pathways. These data suggest that the DRFs may bridge signals from the Rho GTPases and Src, which is known to contribute to multiple signaling pathways. Deregulation of the mouse DRFs results in activation of nuclear signaling and stress fiber formation, likely through obligate binding to Src. Recent results suggest that SRF activation can be triggered by depletion of G-actin pools. Is this mediated by the DRFs, profilin, or Src? Expression of an interfering mutant of profilin (H119E) that binds mDia1 and mDia2, but fails to bind actin, has no effect on mDia2 activation of SRF. This is a surprising result, although consistent with recently published data showing that this mutant profilin has no effect on activated RhoA or LPA induction of stress fibers. These data suggest that Src is the more important DRF-binding partner and begs the question: What is the role of profilin in Rho- or DRF-regulated actin dynamics? Profilin has been presumed to be the critical effector for the formin homology proteins in the regulation of actin polymerization. These findings indicate that this is not the case for the DRFs (Tominaga, 2000 and references therein).

Src and related family members have multiple roles in cell signaling and receptor trafficking. Src becomes activated in response to receptor-peptide growth factor binding and in response to ligands such as LPA that utilize G protein coupled receptors (GPCRs). Receptor clustering and the formation of clathrin-coated pits have been suggested to be essential for full activation of signaling by GPCRs. Sequestration of both EGF receptors and GPCRs appears to be dependent upon Src. Also, Rho has been shown to have a role in targeting Src. There is increasing evidence to suggest that actin dynamics and the Rho GTPases are involved in receptor trafficking and endocytosis. mDia1 has also been shown to associate with phagocytic cups. The roles of mDia1 and mDia2 in these processes are currently being evaluated. In complexes derived from brain extracts, profilin has been shown to be complexed with components of the endocytosis machinery clathrin and dynamin. The data clearly show an association of mDia1 and mDia2 with Src in endosomes. It will be important to explore the possibility that Src and the Rho GTPases form part of a complex that integrates components involved in both endocytosis and signaling. Do mDia1 and mDia2 bind other GTPases that regulate endocytosis or phagocytosis? RhoB, for example, has been shown to be targeted to endosomes or to have a role in vesicle dynamics. Other Rho-family members are being tested for their ability to utilize the DRFs as effectors (Tominaga, 2000 and references therein).

In addition to the DRFs, other Rho effectors localize to the cleavage furrow and have critical roles in the completion of cell division. The ROCK-related Citron kinase, for example, likely cooperates with mDia1 and mDia2 during cytokinesis in a manner analogous to the ROCK-DRF collaborative induction of stress fiber formation. But what is the importance of a DRF bridge between Src and Rho in the process of cytokinesis? Both Cdc42 and RhoA are required for cell division and RhoA has been shown to localize to the cleavage furrow. Does Rho localize Src via mDia1 or mDia2 in a manner similar to the observed localization of Src to the cell periphery? Consistent with the observation that Src co-localizes with the DRF proteins at the mid-body of dividing cells, microinjection of inhibitory anti-Src antibodies inhibits the completion of mitosis. Active Src localizes to whole spindle microtubules in addition to the mid-body. This indicates that Src is active prior to any association with the DRFs. During early stages of mitosis, Cdc2 kinase has been shown to phosphorylate Src on amino terminal residues, which may contribute to its activation. `Active Src then persists after metaphase when cyclin B is degraded and cdc2 kinase activity is low. Src then localizes with the DRF and is maintained in its active form as defined by antibody staining. It is hypothesized that the DRFs bind and stabilize Src and therefore protect it from inactivating kinases such as Csk. Src has known mitotic substrates such as Sam68, but their role(s) in cell division is not clear. Further studies will be needed to identify important Src substrates and mechanisms of Src activation, stabilization, or localization by DRFs during cytokinesis (Tominaga, 2000 and references therein).

Apical secretion in epithelial tubes of the Drosophila embryo is directed by the formin-family protein Diaphanous

Apical localization of filamentous actin (F-actin) is a common feature of epithelial tubes in multicellular organisms. However, its origins and function are not known. This study demonstrates that the Diaphanous (Dia)/Formin actin-nucleating factor is required for generation of apical F-actin in diverse types of epithelial tubes in the Drosophila embryo. Dia itself is apically localized both at the RNA and protein levels, and apical localization of its activators, including Rho1 and two guanine exchange factor proteins (Rho-GEFs), contributes to its activity. In the absence of apical actin polymerization, apical-basal polarity and microtubule organization of tubular epithelial cells remain intact; however, secretion through the apical surface to the lumen of tubular organs is blocked. Apical secretion also requires the Myosin V (MyoV) motor, implying that secretory vesicles are targeted to the apical membrane by MyoV-based transport, along polarized actin filaments nucleated by Dia. This mechanism allows efficient utilization of the entire apical membrane for secretion (Massarwa, 2009).

Apical localization of F-actin is a general feature of tubular epithelial structures. It has been observed in mammalian MDCK cells forming tubes in three-dimensional cell culture, in the cytoplasm underlying the apical membrane facing the lumen in mammalian secretory organs, such as the lacrimal gland, and in the different epithelial tubes of the Drosophila embryo. The lower level of gene duplication in Drosophila, and the ability to follow the consequences of targeted gene inactivation in the tubular structures, allowed identification of the mechanism responsible for nucleating the actin terminal web at the apical side of epithelial tube cells. This study has demonstrate that Dia, which is known to promote the formation of linear actin filaments, is responsible for producing this actin network in Drosophila embryonic tubular structures. Despite differences in the diameter and function of the different tubular organs, the polarized apical actin cables formed by Dia appear to have a common role in trafficking secretory vesicles to the apical tube surface (Massarwa, 2009).

While the role of Dia in promoting apical secretion spans the entire duration of tracheal morphogenesis, two other Formin-homology proteins act at very specific junctions of Drosophila tracheal morphogenesis. Formin 3 participates in the generation of a continuous dorsal trunk tube by promoting vesicular trafficking in the fusion cells of each metamer, perpendicular to the tube lumen. Another Formin domain protein, DAAM, promotes the organization of F-actin in rings around the circumference of the tracheal tube, at the final stages of tracheal morphogenesis (Massarwa, 2009).

It is likely that each of the three Formin domain proteins is regulated by distinct activators that are concentrated at different sites. The localized activation of Formin 3 may eventually lead to polarized vesicle movement, similar to Dia, but toward a different membrane. The activation of DAAM may be necessary for the localized synthesis of F-actin, which will modify the contours of the apical membrane, and thus define the shape of chitin layered on top. The function of Dia stands out, since it is required throughout tracheal development, and is also involved in morphogenesis of other tubular organs (Massarwa, 2009).

The mechanism of localized activation of Dia operates after apical-basal polarity of the cells has been established. Thus, no defects were observe in overall polarity in dia mutant embryos. It seems that the steps upstream to Dia activation utilize the existing polarity at multiple tiers in order to trigger Dia at a highly restricted position. The two Rho-GEF proteins, Gef2 and Gef64C, exhibit a tight apical localization in the cells forming the tubes. The single Rho1 protein, which is downstream to the Rho-GEFs, is again tightly localized to the apical surface in tubular structures. Binding of Rho1 to Dia leads to an opening of the autoinhibited form of Dia and to the formation of a Dia dimer representing the active form (Goode, 2007). Since GTP-bound Rho1 is the immediate activator of Dia, it is particularly important that Rho1 be embedded in the apical membrane, to ensure spatially restricted nucleation of actin polymerization. In C. elegans, a GEF and a Rho protein were shown to be essential for the development of the lumen of the excretory cell. It will be interesting to determine if a Dia-family protein is subsequently activated to promote secretion (Massarwa, 2009).

Dia is also apically localized, both at the mRNA and protein levels. Elimination of the dia 3'UTR demonstrated a persistence of apical protein localization, even when mRNA localization was lost, suggesting that there are two parallel and independent mechanisms for apical localization. The multiple tiers of apical localization assure that activated Dia will be highly restricted to the apical surface (Massarwa, 2009).

It is interesting to note that, while Gef2, Gef64C, Rho1, and Dia proteins are broadly expressed, partially due to maternal contribution of mRNA, they exhibit apical localization only in the tubular structures. This raises the possibility that genes that are specifically expressed in the tubular organs contribute to the apical localization. Alternatively, apical localization may rely directly on the specific phospholipid composition of the apical tube membranes. It will be interesting to determine if a common mechanism is responsible for the apical localization of the different proteins in the pathway, and if this mechanism relies on components that are restricted to the tubular organs (Massarwa, 2009).

The cellular machinery which dictates the apical localization of Rho-GEFs/Rho1/Dia appears to be in place early on. For example, expression of Dia-GFP in the trachea demonstrated apical localization of the protein already at the stage when the tracheal pits are formed. Yet, generation of the polarized actin cables by Dia, and their utilization for secretion, takes place at a later stage, and follows a stereotypic temporal order in the different tracheal branches. What triggers activation of Dia, following the apical localization of the different components? This study has demonstrated that both Gef2 and Gef64C are required to trigger Rho1, which activates Dia. While the activity of the two Rho-GEFs is similar, both have to accumulate to a critical level in order to activate the system. Thus, no secretion takes place when either of them is missing, or when each of them is present at half dose. The delay in activation of Dia and in secretion, may be explained by the time required to accumulate sufficient levels of Rho-GEF proteins. When the system was 'short circuited' by expression of activated Rho1, which was properly localized to the apical surface, Dia-dependent apical secretion was observed already at early stages of tracheal pit formation (Massarwa, 2009).

The results identify Drosophila MyoV (Didum) as a primary motor for apical trafficking of secretory vesicles along the polarized, Dia-nucleated actin cables in tubular organs. When the activity of MyoV was compromised in the tubular epithelia, apical secretion of cargos requiring Dia-generated actin cables was abolished. In contrast, since MyoV operates downstream to Dia, the actin cables themselves remained intact. An analogous role for MyoV has been recently demonstrated during trafficking of Rhodopsins to photoreceptor rhabdomers (Li, 2007). The functional link between the Dia pathway and MyoV was demonstrated by the ability of myoV RNAi to suppress constitutively activated Rho1 or Dia phenotypes. These results further support the direct link between Dia and apical secretion (Massarwa, 2009).

The polarized actin network formed via the nucleating activity of Dia can account for the final phase of secretory vesicle transport to the apical plasma membrane. Class V myosins, such as MyoV, are known to be involved in transfer of vesicles from microtubules to cortical actin networks, suggesting that polarized microtubule arrays may promote the long-range trafficking of the secretory vesicles from their sites of formation to the cell cortex. Consistent with this scenario, a polarized arrangement was demonstrated of microtubulesin tube epithelial cells, the minus ends of which are in close proximity to the apical membrane, which remains intact in the absence of Dia. The universality of this system is highlighted by similarities to polarized secretion in budding yeast, where Myo2p-mediated transport of secretory vesicles into the bud utilizes Dia-generated actin bundles as tracks, in order to deposit the compounds for polarized cell growth (Massarwa, 2009).

When early steps in the secretory pathway are compromised by reducing the activity of the COPII or COPI complexes, accumulation of cargo is observed within the cells and reduced amounts are detected in the lumen. Since these manipulations block an early and global process of secretion, all cargo vesicles are affected (Tsarouhas, 2007). However, after exit from the Golgi, it appears that distinct classes of vesicles are generated, each containing a different set of cargos, and trafficked by a distinct mechanism. One class of vesicles contains chitin-modifying enzymes (such as Verm or Serp), and is targeted to the septate junctions. When the structure of the septate junctions was compromised, these proteins failed to be secreted. Another class of vesicles may contain transmembrane proteins that are deposited in the apical membrane, such as Crb (Massarwa, 2009).

This study now uncovers a third class of cargo vesicles. Several distinct cargos that are secreted to the apical lumen rely on Dia for their secretion. These cargos include the 2A12 antigen, Pio, and the artificial rat ANF-GFP construct. In the absence of Dia, these proteins failed to be secreted to the lumen, but also did not accumulate within the epithelial tube cells. It is believed that when secretion is disrupted, the vesicles are efficiently targeted for lysosomal degradation, since a block of lysosomal targeting facilitated intracellular accumulation of vesicles that failed to be secreted. Inability to secrete Pio resulted in tracheal defects that were similar to pio mutant embryos. Additional defects of dia pathway mutant embryos, such as highly convoluted tracheal branches, may stem from the absence of additional, yet unknown, proteins in the lumen. The mechanisms underlying the incorporation of distinct cargos into different secretory vesicles, as well as the recognition of each vesicle type by different motors and trafficking scaffolds, remain unknown (Massarwa, 2009).

In conclusion, this work has uncovered a universal mechanism, which operates in very different types of tubular epithelial structures in Drosophila. The conserved feature of an apical F-actin network in tubular epithelia of diverse multicellular organisms, and the high degree of conservation of the different components generating and utilizing these actin structures, strongly suggests that this polarized secretion mechanism is broadly used across phyla. The ability to generate polarized actin cables that initiate at the apical membrane provides an efficient route for trafficking vesicles by MyoV, leading to their fusion with the apical membrane and secretion. It is likely that different pathological situations manifested in aberrant formation of epithelial structures, or their utilization for secretion once the tubular organ is formed, represent defects in different components of this pathway. For example, it was shown that mutations in MyoVa in humans disrupt actin-based melanosome transport in epidermal melanocytes (Massarwa, 2009).

Spontaneous recovery of reward memory through active forgetting of extinction memory

Learned behavior can be suppressed by the extinction procedure. Such extinguished memory often returns spontaneously over time, making it difficult to treat diseases such as addiction. However, the biological mechanisms underlying such spontaneous recovery remain unclear. This study reports that the extinguished reward memory in Drosophila recovers spontaneously because extinction training forms an aversive memory that can be actively forgotten via the Dia pathway. Manipulating Rac1 activity does not affect sugar-reward memory and its immediate extinction effect but bidirectionally regulates spontaneous recovery-the decay process of extinction. Experiments using thermogenetic inhibition and functional imaging support that such extinction appears to be coded as an aversive experience. Genetic and pharmacological inhibition of formin Dia, a downstream effector of Rac1, specifically prevents spontaneous recovery after extinction in both behavioral performance and corresponding physiological traces. Together, the data suggest that spontaneous recovery is caused by active forgetting of the opposing extinction memory (Yang, 2023).

Multiple lines of evidence from the current study support that the spontaneous recovery of extinguished reward memory is caused by active forgetting of the opposing extinction memory. First, the small G protein Rac1, which is reported to mediate active forgetting of aversive labile memory, is activated in a delayed manner after appetitive memory extinction. It specifically and bidirectionally regulates the decay process of extinction, also known as the process of spontaneous recovery, through its downstream effector Dia. However, it does not affect the acquisition and retention of appetitive memory at multiple time points as well as the early extinction effect within 3 h. Second, the extinction of appetitive memory requires dopaminergic neurons mediating aversive but not rewarding reinforcement. Third, the extinction of appetitive memory produces a physiological trace of aversive memory in a pair of MBON-γ1pedc>α/β neurons. Such an aversive memory trace correlates with the extinction effect at the behavioral level and can be prolonged by inhibiting the Rac1/Dia pathway, which prevents spontaneous recovery. Thus, the extinction of appetitive memory appears to be a kind of aversive labile memory that decays with time due to Rac1/Dia-mediated forgetting. Such time-based forgetting of extinction memory results in spontaneous recovery (Yang, 2023).

Spontaneous recovery, which was first discovered by Pavlov's experiments, has been proposed to occur because extinction is susceptible to forgetting. In recent years, several biological mechanisms of forgetting have been revealed; however, the link between these forgetting mechanisms and spontaneous recovery has never been established. Rac1-mediated forgetting was first discovered in Drosophila, and it regulates the decay of labile aversive memory through its downstream molecule Dia. In the current study, the data support that spontaneous recovery of the original reward memory happens readily because the extinction memory is actively forgotten via the Rac1/Dia pathway over time. Such finding is consistent with Rac1's role in mediating forgetting of aversive memory, given that appetitive memory extinction appears to be a kind of aversive learning. They share similar punishment-teaching DANs and aversive memory traces (Yang, 2023).

Together with previous studies, the current data support that Rac1-mediated time-based forgetting is a delayed-start, slow-completion process. Overexpressing Rac1-CA significantly accelerated the forgetting process of extinction memory at 6 h but not at 3 h, indicating that the upstream forgetting signal of Rac1 might emerge after 3 h and already start to function at 6 h. Consistently, endogenous Rac1 activity was observed to increase 8 h but not 3 h after extinction. Considering that the extinction effect in control flies was still significant at the behavioral level after 12 h, Rac1-mediated forgetting is a slow-completion process. Similarly, there was a significant increase in endogenous Rac1 activity 1 h after aversive learning, when forgetting was not significant at the behavioral level (Yang, 2023).

Rac1 also plays a conserved role in regulating forgetting of different forms of memory in vertebrates and is associated with autism and Alzheimer's disease. In mice, a single-trial contextual fear conditioning induces delayed Rac1 activation. Decreasing or increasing such Rac1 activation suppresses or accelerates fear memory forgetting, respectively. Thus, it is of interest to test whether reward memory extinction also produces a punishment memory and Rac1 also specifically regulates the spontaneous recovery in vertebrates (Yang, 2023).

Rac1 bidirectionally regulates the forgetting of aversive olfactory memory. The current data show that genetic manipulations of Rac1 in MB neurons do not affect the acquisition and retention of the sugar reward memory. These findings together raise an interesting question: is Rac1 specifically responsible for the forgetting of punishment memory but not reward memory? Presynaptic plasticity at KC-MBON synapses plays a key role in aversive learning and memory,while postsynaptic plasticity at KC-MBON synapses was recently found to underlie appetitive learning and memory. Based on these findings, it became easy to understand that the forgetting of punishment memory and reward memory may be achieved through different molecular mechanisms (Yang, 2023).

Spontaneous recovery of aversive memory after extinction is also reported in Drosophila. According to previous studies and the current findings, it is possible that aversive memory spontaneously recovers through forgetting of extinction memory with a positive valence. Up to now, no forgetting molecule of reward memory has been reported. Significant natural decay has been reported in non-nutritious sugar reward memory and water reward memory. It is interesting to find forgetting molecules underlying such short-lived reward memory and to determine whether there are forgetting molecules that could explain the spontaneous recovery of aversive memory (Yang, 2023).

Extinction is vulnerable to relapse. This relapse is a major obstacle to the successful treatment of clinical disorders including addiction and PTSD. Two strategies have been proposed to prevent relapse: enhancing the consolidation of extinction learning and disrupting the reconsolidation of the original memory. The current findings suggest that blocking specific forgetting mechanisms may contribute to the success of relapse prevention, indicating a new strategy for relapse prevention (Yang, 2023).

Rac1 is proposed to be an attractive therapeutic target in treating human diseases, including cancer, cardiovascular diseases, and neurodegenerative disorders. Genetic inhibition of Rac1 prevents spontaneous recovery of extinguished reward memory—one kind of relapse—suggesting that Rac1 may also be a potential target in treating diseases such as addiction. Rac1 can regulate multiple signaling pathways that involve diverse cellular functions such as actin remodeling, microtubule stability, gene transcription, and superoxide production. Formin Dia, a nucleator of linear actin polymerization, is now found to be a downstream effector of Rac1 in regulating the spontaneous recovery of extinguished reward memory. It might be a more specific drug target than Rac1 in preventing relapse in diseases like addiction. Together, these findings suggest a new strategy to prevent relapse and provide potential drug targets (Yang, 2023).

GENE STRUCTURE

cDNA clone length - 3.6 kb

PROTEIN STRUCTURE

Amino Acids - 1091

Structural Domains

The Diaphanous protein contains two domains shared by the formin proteins, encoded by the limb deformity gene in the mouse. The FH1 domain consists of an unusual proline-rich region near the middle of Diaphanous. This proline-rich domain contains six repeats of 5 to 8 consecutive prolines separated by short stretches rich in the amino acids glycine, methionine, alanine or arginine, with a few interspersed prolines. The FH2 domain, present in a region of about 130 amino acids, has been termed FH2. Both FH1 and FH2 domains are found in Bni1p, the product of a Saccharomyces cerevisiae gene required for normal cytokinesis in diploid yeast cells. It is striking that not only the sequence composition of the FH1 and FH2 domains, but also the spacing between these domains, is conserved among Diaphanous, Bni1p, and the formins (Castrillon, 1994).

Secondary structure analysis, using the algorithm devised by Stock and colleagues, has revealed two regions of the Diaphanous protein that are very likely to form coiled-coil domains. One of these, spanning amino acids 441 to 500, ends just at the beginning of the FH1 domain. The second, spanning amino acids 863 to 1053, begins 17 amino acids before the end of the FH2 domain, just after the region of highest sequence similarity. Thus it appears that coiled-coil domains flank the FH1 and FH2 domains of diaphanous. Moreover, sequence analysis also predicts the existence of coiled-coil domains at equivalent positions in mouse formin IV and Bni1p (Castrillon, 1994 and references therein).

date revised: 15 November 2022

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.