InteractiveFly: GeneBrief

capricious : Biological Overview | Regulation | Developmental Biology | Effects of Mutation | References

Gene name - capricious

Synonyms -

Cytological map position - 70A2--3

Function - cell adhesion

Keywords - axon guidance, boundary formation, wing, CNS, mesoderm

Symbol - caps

FlyBase ID: FBgn0023095

Genetic map position -

Classification - leucine-rich repeats protein

Cellular location - surface

NCBI link: Entrez Gene
caps orthologs: Biolitmine
Recent literature
Kulkarni, A., Ertekin, D., Lee, C. H. and Hummel, T. (2016). Birth order dependent growth cone segregation determines synaptic layer identity in the visual system. Elife 5. PubMed ID: 26987017
The precise recognition of appropriate synaptic partner neurons is a critical step during neural circuit assembly. However, little is known about the developmental context in which recognition specificity is important to establish synaptic contacts. This study shows that in the Drosophila visual system, sequential segregation of photoreceptor afferents, reflecting their birth order, lead to differential positioning of their growth cones in the early target region. By combining loss- and gain-of-function analyses it was demonstrated that relative differences in the expression of the transcription factor Sequoia regulate R cell growth cone segregation. This initial growth cone positioning is consolidated via cell-adhesion molecule Capricious in R8 axons. Further, the initial growth cone positioning was shown to determine synaptic layer selection through proximity-based axon-target interactions. Taken together, this study demonstrates that birth order dependent pre-patterning of afferent growth cones is an essential pre-requisite for the identification of synaptic partner neurons during visual map formation in Drosophila.

Upon reaching the target region, neuronal growth cones transiently search through potential targets. Synaptic connections will only be formed with a subset of these targets. The capricious (caps) gene may regulate these processes in Drosophila. caps encodes a transmembrane protein with leucine-rich repeats (LRRs). During the formation of neuromuscular synapses, caps is expressed in a small number of synaptic partners, including muscle 12 and the motorneurons that innervate it. Loss-of-function and ectopic expression of caps alter the target specificity of muscle 12 motorneurons, indicating a role for caps in selective synapse formation (Shishido, 1998). In addition, Capricious, along with Tartan, another leucine-rich repeat protein, contributes to formation of the affinity boundary between dorsal and ventral compartments during Drosophila wing development. caps and tartan are expressed in the dorsal compartment during boundary formation. Caps and Tartan confer affinity for dorsal cells, assessed by sorting-out behavior. Caps supports boundary formation without conferring dorsal compartment signaling properties. Fringe, in contrast, confers dorsal signaling properties without affecting DV affinity. Therefore, Caps, Tartan, and Fringe have complementary roles in boundary formation (Milán, 2001).

The final step in formation of neural connectivity involves the recognition of target cells. Although earlier events of growth cone guidance greatly restrict the target region, neurons still have to choose a specific synaptic partner from among several potential targets. In each abdominal hemisegment of Drosophila larvae, ~40 motorneurons innervate 30 muscle fibers in a specific manner. Once a motor axon enters its target region during late embryogenesis, its growth cone searches over the surface of many muscles but withdraws from most of these contacts, forming stable synapses only with its own target or targets. caps regulates the formation of some of the selective synaptic connections in this system (Shishido, 1998).

A screen was performed for enhancer trap lines that express a reporter gene in specific muscle fibers during the establishment of motorneuron innervation. caps was identified by analysis of one such line, E2-3-27. In E2-3-27 embryos, the reporter (caps-LacZ) is expressed in four dorsal and six ventral muscles. caps-LacZ is also expressed in central nervous system (CNS) motorneurons that innervate caps-LacZ-positive muscles. caps-LacZ is not expressed in motorneurons that have been identified as innervating caps-negative muscles (for example, RP1, RP3, and RP4). Thus, the expression of caps-LacZ is correlated with neuromuscular specificity (Shishido, 1998).

To determine the function of caps in vivo, caps loss-of-function mutant alleles, which lack the first exon and do not express Caps protein, were generated. Most of the caps mutants die late in embryogenesis or soon after hatching, although a few survive to adulthood. Although no gross developmental defects are found in the CNS or musculature of caps mutant embryos and larvae, the target specificity of muscle 12 motorneurons is altered. In wild-type larvae, muscle 12 is innervated by the terminal branch of ISNb, including the RP5 axon, which projects to the boundary between muscles 12 and 13 and forms synaptic endings exclusively on muscle 12. In contrast, in caps mutant larvae, the terminal branch is often accompanied by additional varicosities on muscle 13, a neighboring caps-negative muscle. Thus, caps restricts arborization of the nerve terminal to muscle 12 (Shishido, 1998).

Ectopic overexpression of caps in all embryonic muscles by G14-GAL4 driver causes formation of more ectopic synapses. In ~70% of the hemisegments, the ISNb terminal formed one or more additional collaterals that formed more robust synaptic endings on muscle 13. The ectopic nerve endings contained type III boutons, which are typical of muscle 12 but not muscle 13 neuromuscular synapses. Since the ectopic synapses were present in the first-instar larvae, caps may function while the connections are being formed. This possibility is further supported by the absence of such ectopic endings when caps expression was induced after completion of synaptogenesis by Mhc82-GAL4 (Shishido, 1998).

It is proposed that caps mediates selective synapse formation. The loss-of-function phenotype may result from improper recognition of the target muscle, whereas the extra synapses on muscle 13 could reflect retention of inappropriate synaptic contacts. In contrast, the gain-of-function phenotype could indicate that the nerve terminal is attracted to muscle 13 and other muscles by ectopic caps. In both cases, however, muscle 12 motorneurons reach their target region normally and extend along muscle 12 before making ectopic synapses on muscle 13. Thus, caps may stabilize specific motorneuronal contacts during a late phase of target selection (Shishido, 1998).

The expression of caps on both sides of the synaptic partners suggests that caps functions homophilically, as has been proposed for the candidate target recognition molecules, Connectin and Fasciclin III. However, expression of caps in S2 cells did not promote cell aggregation. Thus, caps may mediate synaptic target recognition through cell-cell signaling rather than adhesion (Shishido, 1998).

In order to study the role of caps in neural recognition of muscle targets, Caps protein was expressed ectopically in all muscles. Panmuscle Caps alters the target specificity of muscle 12 motorneurons (MNs), indicating that Caps can function in muscles as a target recognition molecule. Panneural Caps alters the pathfinding of muscle 12 MNs. The defect appears to be caused by changes in the steering behavior of muscle 12 MNs at a specific choice point along their pathway to the target muscle. These results revealed a novel function of Caps in axon pathfinding. Deletion analyses of Caps were performed. Caps lacking the intracellular domain was expressed in all neurons or in all muscles, and the ability of altered Caps protein to induce the pathfinding and targeting phenotypes was examined. The function of muscularly expressed Caps in target recognition is intracellular domain dependent, whereas the function of neurally expressed Caps in pathfinding is not, suggesting that Caps may function in neurons and muscles in a different manner. The requirement of the intracellular domain for the function of muscularly expressed Caps suggests the presence of a signaling event within muscle cells that is essential for selective synapse formation (Taniguchi, 2000).

Ectopic and increased expression of Caps was induced in all neurons using the GAL4-UAS system. A GAL4 line, elav-GAL43E1, which expresses GAL4 in all neurons, was crossed with the UAS-caps-Ia+Ib line, which contains two copies of UAS-caps. The progeny (elav-GAL43E1/UAS-caps-Ia+Ib) was raised at 29°C during embryonic and larval development to induce maximal ectopic expression. As expected, Caps was ectopically expressed on all neurons, starting from embryonic stage 12, in elav-GAL43E1/UAS-caps-Ia+Ib individuals. Caps protein was detected in all major axon tracts in the CNS and in the periphery, suggesting that ectopically expressed Caps was properly transported to axons (Taniguchi, 2000).

To analyze the effect of panneural Caps expression on the formation of the nervous system, motor neuron axonal processes were examined in the third-instar larvae by monoclonal antibody 1D4 (anti-Fasciclin II) and monoclonal antibody 22C10 staining. No gross morphological defects were seen in the CNS and musculature, suggesting that their overall development proceeded normally. However, a highly specific change was detected in the trajectory of motoneurons that innervate muscle 12 (muscle 12 MNs). In wild-type larvae, axons of muscle 12 MNs, that fasciculate to form the terminal branch of the intersegmental nerve b (ISNb), project along the internal surface of muscle 13 before reaching their final target, muscle 12. In contrast, in elav-GAL43E1/UAS-caps-Ia+Ib larvae, they pass along the exterior of muscle 13 in 29% of segments. Such a phenotype is not observed in control larvae (elav-GAL43E1/+ raised at 29oC). Despite the abnormality in their trajectory, the axons of muscle 12 MNs reach their normal target by turning interiorly at the cleft between muscles 12 and 13, and establish normal synapses on muscle 12. Thus, ectopic and increased Caps expression on all neurons affects axon pathfinding of muscle 12 MNs, but not their synapse formation. The penetrance of the misrouting phenotype is dependent on the level of expression of ectopic Caps on neurons (Taniguchi, 2000).

In larvae that express Caps panneurally, defects were also observed in the formation of the transverse nerve. The nerve was often split and the cell body of a neuron associated with the nerve (lateral bipolar cell) was occasionally mislocated. With the exception of the transverse nerve phenotype, the effect of ectopic Caps on the formation of motor nerves is highly specific to muscle 12 MNs. Notably, the subbranches of ISNb that terminate on ventral muscles other than muscle 12 displayed no abnormalities in their trajectory or targeting. The other motor nerves, the intersegmental nerve (ISN), intersegmental nerve d (ISNd), segmental nerve a (SNa), and segmental nerve c (SNc) also retained their normal morphology, although it remained possible that there were subtle abnormalities that could not be detected with the marker used (Taniguchi, 2000).

To determine when the axon guidance errors of muscle 12 MNs begin to be seen in individuals that panneurally express Caps, the development of the motor nerves in the embryos was examined by mAB 1D4 staining. In wild-type embryos, motor axons exit the CNS through the ISN or segmental nerve (SN) roots. They then divide into five peripheral motor branches (ISN, ISNb, ISNd, SNa, and SNc) that project to different groups of muscle fibers. The axons of muscle 12 MNs follow the ISNb pathway. During embryonic stage 15, the axons of ISNb, including those of muscle 12 MNs, separate from the ISN and enter the ventral muscle field at muscle 28. They then extend between the external surface of muscles 6 and 7 and the internal surface of muscle 14. At a choice point near muscle 30, muscle 12 MNs and motoneurons that innervate muscle 13 shift their trajectory by turning to a more internal muscle layer. Muscle 12 MNs then extend along the internal surface of muscle 13, forming the terminal branch of ISNb. They reach their final target (muscle 12) by late stage 16 and begin to form functional synapses by early stage 17 (Taniguchi, 2000).

In elav-GAL43E1/UAS-caps-Ia+Ib embryos, no abnormality is seen in the development of ISNb until midstage 16; it defasciculates from ISN at the normal branching point, extends along the internal surface of muscle 14, and reaches muscle 30. However, specific defects in the ISNb trajectory are observed at its distal edge during late stage 16 to early stage 17. The terminal branch of ISNb formed by muscle 12 MNs normally extends along the internal surface of muscle 13 and starts to establish synaptic contacts with muscle 12 by this stage. In 18% of hemisegments in elav-GAL43E1/UAS-caps-Ia+Ib embryos, a misrouting phenotype similar to that seen in the larvae is observed. Instead of traveling along the internal surface of muscle 13, the terminal branch of ISNb takes an abnormal path along the external surface of muscle 13 to reach muscle 12. In 13% of hemisegments in elav-GAL43E1/ UAS-caps-Ia+Ib embryos, the ISNb stalled near muscle 30, failing to extend to more interior muscle layers. Occasionally, thin axonal processes were seen to emanate from the stalled nerve terminal, which stopped prematurely and failed to innervate muscle 12. These results suggest that ectopic expression of Caps in all neurons changes the behavior of muscle 12 MNs at their specific choice point near muscle 30 (Taniguchi, 2000).

Capricious is a transmembrane protein with a short intracellular domain that has no known functional motif. To examine the function of Caps in neurons and muscles, additional ectopic expression experiments were performed using modified forms of the Caps protein. Two Caps deletion constructs were generated: CapsID, which lacks the entire intracellular domain, and CapsED, which lacks most of the extracellular domain extending from the second LRR domain to the amino acid just preceding the transmembrane domain (amino acids 72-449). Ectopic expression of these deletion constructs was induced by the GAL4-UAS system. The UAS lines, UAS-CapsID-4 and UAS-capsED-20, were used to ectopically express CapsID and CapsED, respectively. When crossed with GAL4 drivers, these lines induced similar levels of ectopic Caps expression to the UAS-caps-Ia5 line with intact Caps (Taniguchi, 2000).

The UAS lines were first crossed with elav-GAL4 line to study the ability of the deleted forms of Caps to induce the pathfinding defects of muscle 12 MNs. Ectopically expressed CapsID and CapsED were detected on all major nerve tracts, indicating that the modified forms of Caps are processed and transported properly (Taniguchi, 2000).

The effects of panneuronal expression of CapsID and CapsED were studied in third-instar larvae. Misexpression of CapsED causes no defects compared to controls in the trajectory of muscle 12 MNs, indicating that the extracellular domain of Caps is essential for the induction of the misrouting phenotype. However, when CapsID was expressed, the misrouting phenotype of muscle 12 MNs was observed to the same extent as when the intact Caps was expressed (19% in elav-GAL43E1/UAS-CapsID-4 compared to 23% in elav-GAL43E1/UAS-caps-Ia5). Thus, the intracellular domain of Caps is not required for the ability of neurally expressed Caps to cause pathfinding defects of muscle 12 MNs (Taniguchi, 2000).

The effects of panmuscle expression of the Caps deletion constructs were examined. When intact Caps is ectopically expressed on all muscles, muscle 12 MNs extend their axons to and establish aberrant synapses with the neighboring muscle 13, indicating that Caps can function on muscles as a target recognition molecule. The requirement of the extracellular and intracellular domains of Caps in this process was examined by crossing the UAS lines for CapsID and CapsED with 24B-GAL4 drivers. Panmuscle expression of CapsID and CapsED was confirmed by staining with antibody against Caps. When intact Caps is ectopically expressed in all muscles, the protein is not uniformly distributed on the muscle surface but is strongly concentrated at neuromuscular synaptic sites. CapsID and CapsED are similarly localized at synaptic boutons when expressed in muscles, indicating that the deleted regions play no role in the synaptic localization of the protein (Taniguchi, 2000).

Whether the deletion constructs can cause the mistargeting phenotype of muscle 12 MNs was examined by staining the third-instar larvae of 24B-GAL4/UAS-capsID-4 and 24B-GAL4/UAS-CapsED-20. In the control larvae that misexpressed intact Caps at a similar level, terminals of muscle 12 MNs formed ectopic synapses on muscle 13 in 40% of segments. In contrast, in 24B-GAL4/UAS-CapsID-4 and 24B-GAL4/UAS-CapsED-20 larvae, the extent of targeting errors by muscle 12 MNs was dramatically decreased to the background level. Thus, both extracellular and intracellular domains are required for muscle Caps to mediate target recognition. This is in contrast with the observations in relation to panneurally expressed Caps, where the intracellular domain is not essential for inducing the misrouting phenotype. These results suggest that neurally expressed Caps and muscularly expressed Caps may function in different manners (Taniguchi, 2000).

By what mechanism does panneurally expressed Caps alter the behavior of muscle 12 MNs at this specific choice point? Axons have to complete two steps -- defasciculation from the main nerve tract and steering into the specific target region -- to change their trajectory successively at discrete choice points along the motor pathway. The defasciculation event would mainly relate to changes in axon-axon interactions, whereas the steering event relates to changes in axon-target region interactions. Although Caps does not promote cell adhesion when expressed on S2 cells, it is conceivable that overexpressed Caps on motor axons changes their behavior by increasing axon-axon interactions. However, this is considered to be unlikely. If pan-neurally expressed Caps increases the axon-axon interactions in a general manner, one would expect to see defasciculation defects at many other choice points along the motor pathway. Instead, the phenotype specificity in individuals that express Caps panneurally is highly restricted to the most distal branch of the ISNb. Furthermore, although the stall phenotype was seen in some of the ISNb terminals in the embryos, these axons appear eventually to separate from the main ISNb. These observations suggest that panneurally expressed Caps affects the pathway choice of muscle 12 MNs by influencing their proper steering events rather than defasciculation (Taniguchi, 2000).

In this model, ectopically expressed Caps on muscle 12 MNs affects their behavior by changing their affinity for the neighboring cells on their pathway. Since Caps is normally expressed on muscle 12 MNs, this effect is likely to be due to an increase in the quantity of Caps protein on the motoneurons. This notion is supported by the dose dependency of the phenotype. During normal development, Caps on muscle 12 MNs could function as a receptor that interacts with molecular cues expressed on their pathway along the external muscle layer (e.g., muscles 28, 14, and 30). Overexpression of Caps on the motoneurons may increase their affinity for the external muscle layer so as to prevent the growth cone from navigating to more interior muscle layers (e.g., muscles 12 and 13). Since Caps is normally expressed on muscles 14 and 28, this may involve a homophilic interaction between the Caps on motoneurons and on muscles. Alternatively, Caps on the motoneurons may interact with an unknown ligand expressed on the muscles. In any event, the results indicate that the steering of muscle 12 MNs at the choice point near muscle 30 is highly sensitive to the amount of Caps protein on the motoneurons (Taniguchi, 2000).

The data from the deletion analysis shows that panneural expression of CapsID is as potent as that of the intact Caps in inducing the pathfinding defects, indicating that the intracellular domain is not required for this function of the molecule. Thus, if Caps functions as a receptor in muscle 12 MNs as proposed above, it must transduce the signal by interacting with other molecules on the membrane. When Caps is expressed in all muscles, muscle 12 MNs form ectopic synapses on muscle 13 (Shishido, 1998). Since Caps is normally expressed on muscle 12 but not on muscle 13, it has been proposed that Caps functions as a specific molecular label on muscle 12 to be recognized by muscle 12 MNs. Three other candidate target recognition molecules -- Connectin, Fasciclin III, and Toll -- are similarly expressed on the surface of a subset of muscles. How do these molecules mediate selective synapse formation? Are they simply presented by muscles as ligands that are passively recognized by motoneurons? Or do they play more active roles in the initiation of synapse formation? To obtain insights on Caps function in selective synapse formation, the effect of deletion of its intracellular domain was studied. The ability of panmuscularly expressed Caps to induce targeting errors by muscle 12 MNs is completely abolished when the intracellular domain is deleted. This is not due to nonspecific conformational change of the molecule, because the CapsID retains its activity when ectopically expressed on neurons. Furthermore, panmuscularly expressed CapsID is found to be present at synaptic sites as is intact Caps, suggesting that the deletion does not affect the transportation or localization of the protein. Taken together, the results suggest that some molecular event(s) in muscles mediated by the intracellular domain of Caps are required for the ectopic synapse formation (Taniguchi, 2000).

In what way might the intracellular domain of Caps be critical for its function? One possible role played by the intracellular domain could be its interaction with cytoplasmic components necessary for synaptic localization. Such a role was shown for the intracellular domain of Fasciclin II, a homophilic cell adhesion molecule essential for the growth and maintenance of neuromuscular junctions. Indeed, there is evidence for the existence of synaptic targeting mechanisms for Caps. The Caps protein is normally localized on newly formed synaptic boutons in subsets of muscles in the first-instar larvae. Ectopically expressed Caps is concentrated at neuromuscular synaptic sites of third-instar larvae. However, the intracellular domain appears to be unnecessary for this process, since CapsID is normally concentrated at synaptic sites. Thus, the essential role played by the intracellular domain in selective synapse formation is not in the synaptic targeting of the molecule. Interestingly, CapsED, which lacks most of the extracellular domain, is also normally localized at synapses. Thus, the functional domain necessary for synaptic targeting may be in the transmembrane domain. Alternatively, there may be multiple targeting systems that act both through the intracellular and extracellular domains of Caps (Taniguchi, 2000).

Cell biological evidence is accumulating to support the notion that postsynaptic cells play active roles in synapse formation. During axodendritic synaptogenesis in vertebrates, dendritic filopodia and their protrusive motility are proposed to be essential for the initiation of synaptic contacts. Similarly, during neuromuscular synaptogenesis in Drosophila, muscles extend fine processes (called myopodia) that actively interact with motoneuronal growth cones. It is an interesting possibility that Caps on muscles participates in such processes by interacting with receptors on specific motoneurons via its extracellular domain and by transmitting the signal into muscles via its intracellular domain. Although the exact nature of the signal mediated by the intracellular domain of Caps is unknown, the present study supports the notion that molecular events occurring in postsynaptic cells are essential for the formation of specific synapses (Taniguchi, 2000).


Coordinate control of synaptic-layer specificity and rhodopsins in photoreceptor neurons: Senseless regulates capricious

How neurons make specific synaptic connections is a central question in neurobiology. The targeting of the Drosophila R7 and R8 photoreceptor axons to different synaptic layers in the brain provides a model with which to explore the genetic programs regulating target specificity. In principle this can be accomplished by cell-type-specific molecules mediating the recognition between synaptic partners. Alternatively, specificity could also be achieved through cell-type-specific repression of particular targeting molecules. This study shows that a key step in the targeting of the R7 neuron is the active repression of the R8 targeting program. Repression is dependent on NF-YC (CG3075), a subunit of the NF-Y (nuclear factor Y) transcription factor (Mantovani, 1999). In the absence of NF-YC, R7 axons terminate in the same layer as R8 axons. Genetic experiments indicate that this is due solely to the derepression of the R8-specific transcription factor Senseless (Sens) late in R7 differentiation. Sens is sufficient to control R8 targeting specificity and Sens directly binds to an evolutionarily conserved DNA sequence upstream of the start of transcription of an R8-specific cell-surface protein, Capricious (Caps) that regulates R8 target specificity. R7 targeting requires the R7-specific transcription factor Prospero (Pros) in parallel to repression of the R8 targeting pathway by NF-YC. Previous studies demonstrated that Sens and Pros directly regulate the expression of specific rhodopsins in R8 and R7. It is proposed that the use of the same transcription factors to promote the cell-type-specific expression of sensory receptors and cell-surface proteins regulating synaptic target specificity provides a simple and general mechanism for ensuring that transmission of sensory information is processed by the appropriate specialized neural circuits (Morey, 2008).

The compound eye comprises about 750 simple eyes (ommatidia), each containing a cluster of eight photoreceptor neurons (R1-R8). These neurons form synaptic connections in two regions of the optic lobe, the lamina and the medulla. The R1-R6 neurons innervate the lamina; the R7 and R8 neurons form connections in the M6 and M3 medulla layers, respectively. Genetic studies have led to the identification of cell-surface proteins regulating R7 and R8 target specificity. Notably, mis-targeting mutant R7 neurons terminate selectively in M3, the layer in which wild-type R8 axons terminate, suggesting a close relationship between the genetic programs controlling R7 and R8 target specificity. This study describes transcriptional regulatory pathways that control the differential targeting specificity of these neurons (Morey, 2008).

In a screen for R7 targeting mutants, a strong loss of function mutation was identified in the NF-YC gene, which encodes a subunit of NF-Y, an evolutionarily conserved heterotrimeric transcription factor. Although NF-Y function has not been studied extensively in the fly, it has been shown to act as both an activator and a repressor in other organisms. The targeting of visual-system neurons was assessed in mosaic animals to generate large patches of mutant retinal tissue early in development. About 75% of NF-YC mutant R7 axons terminated in M3, the same layer as wild-type R8 axons. This phenotype was fully rescued by an NF-YC complementary DNA. In contrast with the marked effect of NF-YC mutations on R7, targeting of R8 to the M3 layer and targeting of R1-R6 to the lamina were unaffected (Morey, 2008).

To assess whether NF-YC is required in a cell-autonomous fashion in R7 neurons, mosaic flies were generated in which a fraction of R7 neurons was rendered mutant and labelled with green fluorescent protein (GFP), whereas the remaining R7 neurons and all the R8 neurons were wild-type and unlabelled. About 17% of the mutant R7 neurons (n = 144 of 807) mis-targeted to M3. The decrease in penetrance of the phenotype, in comparison with mutant R7 neurons generated by mitotic recombination induced earlier in the eye primordium, probably reflects perdurance of NF-YC protein present in precursor cells. NF-YC is therefore required autonomously for R7 targeting but not for the targeting of other classes of photoreceptor neurons. As NF-YC is expressed in all R cells, NF-YC must function in combination with other factors or signals selectively activating NF-YC function in R7 (Morey, 2008).

Given that NF-YC is part of a transcription factor complex and is expressed in the nucleus of R7 neurons, it is likely that the change in targeting specificity reflects a change in gene expression. Wild-type R7 neurons initially target to the temporary R7 layer in the medulla and then, during mid-pupal development, extend to their final target. Targeting of NF-YC mutant R7 neurons to the temporary layer is indistinguishable from the wild type. Extension to the final target layer at 70% after puparium formation (APF) is frequently disrupted, with many R7 neurons terminating in the layer within which R8 terminates. Consistent with this finding was the observation that NF-YC mutant R7 neurons expressed all five early R7 markers tested in wild-type patterns. It was reasoned, then, that NF-YC might repress a subset of R8-specific genes in the R7 neuron that later in development control final target layer selection. Indeed, the R8-specific transcription factor Sens was expressed ectopically in NF-YC mutant R7 neurons (Morey, 2008).

sens is a key regulator of R8 development. In wild-type larval eye discs, Sens is expressed in two or three cells that have the potential to become R8 before becoming restricted to a single differentiating R8 neuron. Sens remains expressed in R8 into the adult. It is required at a very early stage of eye development to regulate R8 specification and, much later during pupal development and in the adult, to regulate the transcription of R8-specific rhodopsins directly. In NF-YC mutant larval eye discs, Sens expression in R8 begins before overt R8 differentiation as in the wild type. By contrast, Sens mis-expression in mutant R7 neurons was first observed 15-20 h after the onset of differentiation as assessed by the expression of the R7-specific marker pros. Expression of Sens in mutant R7 neurons persists throughout pupal development and into the adult and is cell-autonomous. As Sens mis-expression occurs after the onset of R7 differentiation and NF-YC mutant R7 neurons mis-target to the M3 layer during the late phase of R7 targeting, sens may promote an R8 targeting program that is distinct from the role of sens in cell fate earlier in development (Morey, 2008).

If upregulation of Sens in NF-YC mutant R7 neurons is responsible for targeting to the M3 layer, removal of Sens from NF-YC mutant cells should suppress the targeting defect. To test this, mitotic recombination was induced on two different chromosomes (namely chromosomes X and 3) to generate R7 neurons that were simultaneously mutant for both NF-YC and sens, and their targeting was assessed in an otherwise wild-type background. Removing sens from NF-YC mutant R7 neurons completely suppresses the mis-targeting phenotype. Thus, during wild-type development the NF-YC mediated repression of sens in R7 is necessary to prevent inappropriate targeting to M3 (Morey, 2008).

To test whether Sens is sufficient to implement an R8 targeting program, sens was mis-expressed in R7 neurons. Under these conditions about 25% of the R7 neurons were redirected to the M3 layer, thus phenocopying NF-YC loss-of-function mutations. Additional experiments using the method in which Sens was provided conditionally early in development to promote R8 cell fate, but removed later, support the view that Sens functions at later stages of R8 development to promote targeting. Taken together, these data raise the possibility that Sens could directly control the expression of cell-surface proteins regulating R8 target specificity (Morey, 2008).

Caps is the only cell-surface molecule that has been shown to be both specifically expressed in the R8 neuron and required for R8 targeting and it is therefore an excellent candidate for direct regulation by Sens. Indeed, like Sens, Caps is expressed ectopically in R7 in NF-YC mutants. Expression of Caps, as detected with an enhancer trap, is specifically activated in NF-YC mutant R7 neurons about 9 h after the onset of Sens expression. Furthermore, a previous study showed that ectopic expression of Caps in R7 respecified their connections to the R8 layer. Both NF-YC mutant R7 neurons and R7 neurons mis-expressing Caps initially target correctly but then select the inappropriate M3 layer during mid-pupal development. Taken together, these observations indicate that caps could be a downstream target of Sens (Morey, 2008).

Examination of the DNA sequences 1 kilobase upstream of caps and within the first large intron led to the identification of four and three putative Sens-binding sites, respectively. An evolutionarily conserved Sens-binding site was identified 500 base pairs upstream of the caps transcriptional start site. Sens protein binds specifically to this site in gel-shift assays, making it likely that caps is a direct target of Sens. However, Sens must regulate R8 target specificity by controlling the expression of other genes in addition to caps, because loss of caps does not suppress the NF-YC mutant phenotype. This is consistent with the finding that loss of caps, in an otherwise wild-type background, results in targeting defects in about 50% of the R8 neurons. Together, these data suggest that Sens directly regulates the expression of Caps, a cell-surface protein controlling R8 target specificity, and must also regulate the expression of other genes involved in this process (Morey, 2008).

Specific repression of sens in R7 neurons could arise through interactions between NF-YC and the R7-specific transcription factor Pros. Like NF-YC, Pros is also required for R7 target specificity. It is expressed in R7 from an early stage of its development through to the adult in a similar fashion to Sens expression in R8. About 20% of the pros-null mutant R7 neurons terminate in M3. Two lines of evidence support the view that Pros works in parallel with NF-YC: first, the loss of pros in R7 neurons does not lead to ectopic expression of Sens, and second, the frequency of mis-targeting R7 axons in single pros-null mutant cells is markedly increased by removing NF-YC. Thus, Pros could either promote R7 targeting directly or, like NF-YC, act to repress an R8 targeting program, or both (Morey, 2008).

Thus, R7 targeting requires NF-YC and, in parallel, Pros, whereas R8 targeting relies on Sens-dependent regulation of caps and other genes. Mutations in many other genes required for R7 targeting cause R7 neurons to mis-target to the M3 layer specifically rather than terminating promiscuously in the medulla. This underscores a tight inter-relationship between the mechanisms regulating targeting to these two layers. On the basis of the strong M3 mis-targeting phenotype of NF-YC mutant R7 neurons and complete suppression of the phenotype by the removal of sens, a key mechanism regulating R7 layer specificity is repression of an R8 targeting program. More generally, repression of inappropriate pathways may promote differential targeting in closely related neurons (Morey, 2008).

The roles of Pros and Sens in target layer selection are analogous to their function in controlling the expression of R7-specific and R8-specific rhodopsins. R7 and R8 neurons express different rhodopsins and hence detect different wavelengths of light. In R8, Sens directly represses the transcription of R7 rhodopsins and directly activates the transcription of an R8 rhodopsin. In the R7 neuron, Pros binds to an upstream regulatory sequence in the R8 rhodopsin genes and represses their expression. NF-YC mutant R7 neurons no longer express R7 rhodopsins, and all express R8 rhodopsins. This is consistent with the finding that NF-YC mutant R7 neurons in adults express Sens but no longer express Pros. Thus, transcription of both R8-specific rhodopsins and an R8-specific targeting protein Caps is directly regulated by Sens (Morey, 2008).

These observations suggest a simple solution to the mechanisms by which sensory neurons connect to the neural circuits specialized for the reception of different sensory stimuli (for example, different wavelengths of light or different odours). Although the molecular basis of this coupling is understood in considerable detail for vertebrate olfactory neurons, in which odorant receptors have a direct function in controlling target specificity, little is known about the coupling in other sensory systems. Coupling is likely to be regulated in a different fashion in other neurons, because even in the fly olfactory system, for example, targeting is independent of sensory receptor expression. On the basis of these studies on Sens it is proposed that the same transcription factors directly control both rhodopsin expression and the cell-surface proteins that control target layer specificity. More generally, it is speculated that in many sensory neurons a common set of transcription factors may directly control, and thereby coordinate, the expression of cell-surface proteins regulating target specificity and the receptors detecting specific sensory stimuli (Morey, 2008).

Modulation of Drosophila retinal epithelial integrity by the adhesion proteins Capricious and Tartan

The development of the Drosophila eye imaginal disc requires complex epithelial rearrangements. Cells of the morphogenetic furrow are apically constricted and this leads to a physical indentation in the epithelium. Posterior to the furrow, cells start to rearrange into distinct clusters and eventually form a precisely patterned array of ommatidia. These morphogenetic processes include regulated changes of adhesion between cells. Two transmembrane adhesion proteins, Capricious and Tartan, have dynamic and complementary expression patterns in the eye imaginal disc. Novel null mutations in capricious and double null mutations in capricious and tartan are described. They are shown to have redundant functions in regulating the architecture of the morphogenetic furrow and ommatidial spacing. It is concluded that Capricious and Tartan contribute to the adhesive properties of the cells in the morphogenetic furrow and that this regulated adhesion participates in the control of spacing ommatidial clusters (Mao, 2008).

The development of the Drosophila compound eye is a complex process involving the interplay of many signalling pathways. The Drosophila eye is composed of a regular hexagonal lattice of about 800 individual facets known as ommatidia. Each ommatidium consists of a unit of eight photoreceptor neurons (R1-R8) and four cone cells, and is surrounded by pigment cells. The eye develops from a monolayer epithelium known as the eye-antennal imaginal disc. At the start of the third larval instar, the cells in the imaginal disc start to differentiate. This differentiation starts at the posterior of the disc and sweeps anteriorly, preceded by a physical indentation known as the morphogenetic furrow (MF). Developing rows of ommatidia are left in its wake, and this progressive development implies that there is a gradient of developmental stages in a single disc, with the most mature being at the posterior (Mao, 2008).

Most of the cells in the eye disc have a columnar epithelial morphology, but in the morphogenetic furrow they become apically constricted. As a result of this constriction, these cells change from being columnar to bottle-shaped and the consequent change in epithelial packing produces the indentation of the furrow itself. Immediately after the passage of the furrow, and therefore posterior to it, cells begin to rearrange, developing from random packing into first lines of cells, then arcs, and finally morphologically distinct clusters within the epithelium. This process depends on myosin II contractility but presumably also requires precise changes in the adhesive properties of cells as the clusters separate from their neighbours. In fact, adhesive changes can be directly observed-the clusters show increased levels of apical Armadillo/β-catenin, a key component of the adherens junctions, a phenomenon dependent on Atonal and the epidermal growth factor receptor (EGFR) pathway. Beyond this increase in adherens junctions, little is known about the adhesion processes that participate in the clustering process (Mao, 2008).

Capricious (Caps) and Tartan (Trn) are highly similar transmembrane proteins with multiple extracellular leucine rich repeats (LRRs) and shorter intracellular domains. They share 67% protein sequence identity in their extracellular domains, which consist of 14 LRR repeats, but only 15% overall identity in their intracellular domains, including a conserved motif of 31 amino acids adjacent to the membrane. Since they lie within 115 kb of each other in the genome, it is likely that they represent a relatively recent gene duplication event. Although their exact molecular function is not well characterised, they can act as homotypic adhesion proteins in cell culture and at least in some contexts their intracellular domains are dispensable supporting the idea that their primary roles are in cell adhesion. Consistent with this, their functions have mostly been associated with their adhesion properties. Caps is required for targeting a subset of embryonic motor neurons to their specific muscles during embryonic development and in targeting R8 photoreceptor axons to the appropriate layers of the optic lobe. Caps and Trn have also been implicated in the formation of affinity boundaries between dorsal and ventral compartments in the developing wing imaginal disc. They have halso been shown to have overlapping functions in adhesion of cells in the developing leg imaginal disc (Mao, 2008 and references therein).

Morphological plasticity in the developing eye involves precisely ordered remodelling of epithelial cell contacts. This study describes the specific and complementary expression patterns of Caps and Trn in the imaginal eye disc and their redundant roles in regulating aspects of epithelial organisation in the morphogenetic furrow and the spacing of developing ommatidia (Mao, 2008).

caps and trn have developmentally regulated expression patterns in the eye. In third instar eye imaginal discs, caps-lacZ is expressed in all cells in the morphogenetic furrow and at a lower level in cells just posterior to the furrow before becoming restricted to single photoreceptor cells. By simultaneous staining with the R8 photoreceptor marker Senseless, it was shown that the single cells eventually expressing caps-lacZ are the R8 cells, the founders of ommatidial development. This result is consistent with the expression pattern reported by Shinza-Kameda (2006), although that study confined description of expression to the later stages when caps is restricted to R8. trn is a close sequence relative of caps and in the wing imaginal disc they are believed to act in partnership as adhesion proteins that regulate cell affinity at compartment borders. Therefore the expression pattern in the eye of trn-lacZ was examined. Interestingly, trn-lacZ is also expressed dynamically, initially in all cells in the furrow, then at a lower level in cells just posterior to the furrow, before becoming restricted to a non-overlapping subset of photoreceptor precursors from caps. trn-lacZ colocalised with R1 and R6 markers anti-BarH1. In summary, both caps and trn are widely expressed in the morphogenetic furrow, and each then becomes restricted to non-overlapping subsets of photoreceptors. These complementary expression patterns in the eye suggested that the Caps and Trn proteins might have a previously unrecognised function in eye development (Mao, 2008).

Specific antibodies against Caps and Trn were raised to examine their expression pattern in more detail. Unfortunately the Caps antiserum did not reliably detect the endogenous level of Caps protein in eye discs. In contrast, the Trn antibody successfully recognised Trn protein in wild-type discs. Its specificity was confirmed by the loss of signal in clones of cells mutant for Trn but not in clones of cells mutant for Caps. The antibody staining pattern confirmed the caps-lacZ expression pattern: Trn is expressed broadly in the morphogenetic furrow and in subsets of ommatidial cells after the furrow. Since photoreceptor specific markers are almost all nuclear, and Trn is membrane localised, overlapping staining patterns cannot readily be used to confirm the identity of the specific ommatidial cells stained posterior to the furrow. However, the expression pattern is fully consistent with that of the trn-lacZ line, that is, the staining is localised in the expected location of R1, 6 and 7 but not of R8, 2, 5, 3, 4. Z-sections along the anterior-posterior axis of the disc revealed that it is expressed mostly in the apical membrane of photoreceptor cells but is also visible in some basolateral membranes. Interestingly, Tartan is only expressed in the anterior half of the furrow (Mao, 2008).

Since caps and trn single mutants did not affect eye development, and since Caps and Trn are highly related proteins, it was asked whether they might act redundantly in eye development. To address this, a caps trn double null mutation, capsDel1 trn28.4, was generated by using P-element induced male recombination, to simultaneously delete caps and recombine the new mutation onto the existing trn28.4 null allele. This double capsDel1 trn28.4 null retains the intervening genes CG33262 and CG11281, so represents a ‘clean’ removal of the two related proteins. Mitotic clones of the capsDel1 trn28.4 mutation (marked by lack of GFP, green) showed subtle but consistent defects. Within the morphogenetic furrow, the mutant cells showed normal levels of apical constriction, and accumulated high levels of Armadillo/β-catenin indistinguishably from the wild type. However, at the clone border between mutant and wild-type cells, there was a consistent reduction in the apical constriction of cells and their Armadillo accumulation in adherens junctions. This phenotype is fully penetrant but appears more pronounced when the clone boundary is perpendicular to the morphogenetic furrow (Mao, 2008).

The apical constriction of morphogenetic furrow cells generates the indentation of the furrow itself. Sagittal sections in the Z-axis of capsDel1 trn28.4 clones along the furrow, showed that the cells at the clone boundary with the enlarged apical profiles were also taller than their neighbours, that is, their apical surfaces were elevated, thereby disrupting the furrow itself (Mao, 2008).

These related phenotypes of relaxation of apical constriction and increase in apical-basal height of the cells were a non-autonomous effect: they were observed in both mutant and wild-type cells at the clone boundary. The range of the phenotype was only 2-3 rows of cells beyond the clone border, and in some cases this non-autonomy was predominantly in the wild-type, and sometimes predominantly in the mutant territory. Given the subtle nature of the effects, the borders of clones of single mutants for caps or trn were reexamined, an it was confirmed that they were never visibly affected (Mao, 2008).

A second phenotype associated with the capsDel1 trn28.4 double null clones (marked by lack of GFP, green) was a perturbation in ommatidial spacing in third instar eye discs. Again, this was apparent only at the boundaries between mutant and wild-type cells. Ommatidia close to these boundaries were often clearly displaced from their normal positions but there was no obvious change to their individual morphology, nor was the total number of ommatidia obviously affected. 25 individual eye discs containing clones were analysed, and the number of ommatidia adjacent to clonal boundaries was counted, along with the number of these ommatidia that were displaced from their normal position. In total, of 846 ommatidia at boundaries, 187 (i.e., 22%) were displaced. This phenotype is also non-autonomous, with both mutant and wild type ommatidia showing mis-positioning. These spacing defects remain later, at pupal stages of eye development, and they are made more apparent by the fusion of neighbouring ommatidia that are abnormally close to each other. These fusions are observed in about 5% of ommatidia adjacent to clone boundaries. Very occasional defects were observed in the normal number of cone cells. As with the third instar eye disc, the pupal phenotypes were not observed in clones mutant for caps or trn alone (Mao, 2008).

Since defects were observed only at the boundaries between capsDel1 trn28.4 mutant and wild type tissue, it was wondered whether the sudden step-like changes in Caps and Trn levels were more important than the overall levels of these adhesion proteins. Therefore clones were made over-expressing Caps and Trn but they did not show any visible furrow or ommatidial spacing defects at the clone boundaries. This implies that the boundary effects seen in clones of capsDel1 trn28.4 cells are caused by the juxtaposition of cells expressing Caps and Trn with cells not expressing them. In summary, it is concluded that there is a redundant function for Caps and Trn in controlling aspects of cell morphology and ommatidial spacing in Drosophila retinal development (Mao, 2008).

One of the main tissues in which Caps and Trn have been studied is the developing wing imaginal disc. Despite evidence that caps and trn have an important function in maintaining compartment borders in the wing, previously studied mutants in these genes have not affected the dorsal-ventral boundary. Therefore advantage was taken of having made a previously unavailable double null mutation to look at DV border formation in the wing. Clones of capsDel1trn28.4 double null (marked by lack of GFP) did not perturb or cross the boundary, as marked by staining with an antibody against Senseless. This is consistent with earlier data, where clones of cells simultaneously null for trn and hypomorphic for caps did not cause defects at the DV boundary. The fact that complete loss of both proteins does not affect DV boundary formation or maintenance suggests that Caps and Tartan are not essential for compartmentalising cells in this part of the wing (Mao, 2008).

This study has shown that the related adhesion proteins Capricious and Tartan have redundant functions in the remodelling of epithelial cell contacts that occur during the early stages of Drosophila eye development. Each is expressed in a two phase pattern, first broadly in the furrow and then later in non-overlapping subsets of photoreceptors: Caps in R8 and Trn in R1, 6 and 7. A null mutation was made of caps and also a double null, in which both caps and trn are absent. Analysis of these mutations shows that while removal of either gene alone has no phenotype, the loss of both leads to subtle but reproducible defects in retinal development. The earliest phenotype is a reduction of apical constriction and accumulation of Armadillo in the morphogenetic furrow. Slightly later, displacement of ommatidia from their normal very precise array was seen. Finally, this displacement leads to occasional fusion of neighbouring ommatidia and other minor defects in the pupal retina. Intriguingly, all the defects observed are limited to clone boundaries; cells fully within the mutant clones appear normal (Mao, 2008).

The capsDel1trn28.4 phenotypes in the eye are relatively minor. They are nevertheless reproducible and quite penetrant. Essentially all clones that cross the furrow perpendicularly show a reduction in apical constriction and Armadillo staining at their clone boundaries, and 22% of ommatidia that lie at the clone boundary are detectably misplaced; ommatidial fusion defects in the pupal retina are rarer, at about 5%. It is proposed that these phenotypes are all a consequence of the initial furrow defects, and that these are caused by loss of the furrow expression of Caps and Trn. This implies that the later, photoreceptor-specific expression of Caps and Trn does not participate in the observed phenotypes. This proposal is based on the following logic: (1) the redundant function of the two proteins is difficult to reconcile with non-overlapping expression: if they are in different cells, how can they replace each other's function? Although it would be possible to imagine a scenario where this could occur, a more parsimonious explanation is that the redundant phenotype depends on their function where they are co-expressed, in the furrow. (2) The expression of Caps in R8 is already known to have a quite separate function, in the targeting of the R8 axon growth cones to the appropriate layer of the optic lobe (Shinza-Kameda, 2006). The R8 cell bodies are in the retina, which is why caps-lacZ expression there is seen there, but the protein must be transported to the axon terminals. The discovery of an equally specific but non-overlapping expression of Trn, suggests that it too might have an analogous function in axon targeting, although this prediction has not been tested (Mao, 2008).

The idea that the later, photoreceptor specific expression of Caps and Trn is responsible for axonal guidance defects, but not retinal patterning, appears inconsistent with the protein expression of Trn that is see at the apical surface of the photoreceptors, i.e. in the retina, distant from the axon terminals. Unfortunately, the wild-type protein expression of Caps, which is know to be involved in axonal guidance, could not be detected so it is possible that Caps protein is localised very differently from Trn-only in the axons. Although a better anti-Caps antibody could resolve this fully, on balance it is suspected that the apical expression of Trn, and possibly Caps, either reflects a function distinct from the retinal defects report in this study and also from axonal guidance; or that it is a non-functional consequence of the intracellular trafficking pathways that transport the functional pool to the axon terminal (Mao, 2008).

The third reason (for suggesting that the functions this study has uncovered are dependent on Caps and Trn in the furrow, and that the later defects in spacing are secondary consequences of a primary furrow defect) is that this is consistent with the furrow acting to organise epithelial packing. Detailed inspection of cells in the furrow and immediately after they emerge from it, shows profound rearrangement that starts with straight lines of cells, evolving into arcs and finally into morphologically distinct clusters. Adhesion defects in the furrow may disrupt this process such that ommatidial clusters and their spacing become less ordered. It is not understood why these phenotypes manifest only at clone boundaries, but it is presumed to be a consequence of a discontinuity in adhesive properties. Similarly, the short range non-autonomy of the phenotype is probably due to local cell packing problems caused by adhesion anomalies at the boundaries of wild-type and mutant tissue. Another possible explanation for the non-autonomous effects is that changes in cell shape and epithelial morphology in the furrow could affect the range or efficiency of intercellular signalling molecules, thereby affecting normal retinal development. Little is known about how epithelial characteristics can modulate secreted signals and this will be a fruitful area for future study (Mao, 2008).

A very recent paper (Sakurai, 2007) has analysed the functions of Caps and Trn in the developing leg disc. That study also shows a completely redundant function caused by rather subtle adhesion defects. Leg disc development is, however, very different from eye development and the developmental consequences are therefore distinct. In the leg, the sharpening of a progressive border that develops between tarsus 5 and the pretarsus segment was compromised in double mutants. By analysing cell movement within the developing leg disc, Sakurai proposed that Caps and Trn expression allows cell mobility within the epithelium: their downregulation coincides with reduced mobility, while their overexpression leads to cell invasion into inappropriate territories. In the eye, there is no evidence for significant mixing of cells within the epithelium and, the eye model suggests a different use of a rather similar function for these adhesion proteins. In both cases, however, Caps and Trn appear to regulate the ability of cells within an epithelium to reorganise with respect to their neighbours (Mao, 2008).

In summary, the results are interpreted to imply that Caps and Trn expressed in the morphogenetic furrow participate in modulating the adhesivity of epithelial cells. At this stage in development, they are beginning to undergo complex and coordinated rearrangements, with concomitant adhesion changes with their neighbours. Even quite minor disruption of this process leads to alterations in epithelial packing that can have consequent effects on the spacing of ommatidia. The relatively minor retinal phenotype of loss of Caps and Trn implies that other adhesion proteins contribute to the overall regulation of this process. For example, Drosophila E-cadherin, an essential component of adherens junctions, is necessary for epithelial maintenance, and mutant (hypomorph) clones fail to form adherens junctions and lose their epithelial integrity completely. It is suspected that complex regulation of adhesion may require the action of several adhesion systems. The data also leads to the tentative suggestion that Trn may, like Caps, have a later function in photoreceptor neuron development, for example in axon targeting. Finally, and on a separate tack, construction of a double null mutation for caps and trn unambiguously shows that neither are essential for the normal formation of the dorsal-ventral boundary of the wing imaginal disc, a process that overexpressed Caps and Trn can disrupt (Mao, 2008 and references therein).


The caps-LacZ reporter is expressed in four dorsal (1, 2, 9, and 10) and six ventral (12, 14 to 17, and 28) muscles. In muscle 12, caps-LacZ and caps RNA are expressed in a single nucleus of the syncytial muscle, near the contact site of the motorneuronal growth cone. caps-LacZ is also expressed in central nervous system (CNS) motorneurons aCC, RP2, RP5, and the most medial U, all of which innervate caps-LacZ-positive muscles (Shishido, 1998).

Caps protein is expressed on the surface of developing motor axons. In first-instar larvae, Caps protein was detected in the mature synaptic sites of all caps-positive muscles (Shishido, 1998).

A screen of cell-surface molecules identifies leucine-rich repeat proteins as key mediators of synaptic target selection

In Drosophila embryos and larvae, a small number of identified motor neurons innervate body wall muscles in a highly stereotyped pattern. Although genetic screens have identified many proteins that are required for axon guidance and synaptogenesis in this system, little is known about the mechanisms by which muscle fibers are defined as targets for specific motor axons. To identify potential target labels, a screen was performed of 410 genes encoding cell-surface and secreted proteins, searching for those whose overexpression on all muscle fibers causes motor axons to make targeting errors. Thirty such genes were identified, and a number of these were members of a large gene family encoding proteins whose extracellular domains contain leucine-rich repeat (LRR) sequences, which are protein interaction modules. By manipulating gene expression in muscle 12, it was shown that four LRR proteins participate in the selection of this muscle as the appropriate synaptic target for the RP5 motor neuron (Kurusu, 2008).

A database was assembled of 976 Drosophila cell-surface and secreted (CSS) proteins that are likely to be involved in cell recognition events during development. EP lines were found that allowed expression of 410 of these genes in muscles. All the genes were defined that altered presynaptic NMJ terminal patterning and structure without visibly affecting the muscles themselves. The screen identified 30 genes that cause mistargeting of axons within the VLM field with a penetrance of ~30% and 55 genes that produce major alterations in synaptic boutons or the structures of NMJ arbors with ~60% penetrance (Kurusu, 2008).

LRR genes represented 5 of the 12 mistargeting genes of interest. To evaluate the roles of the LRR genes in synaptic targeting, the genes of LOF mutants (and also knocked down or overexpressed mutants) were examined using both panmuscle drivers and drivers expressed only in specific muscle fibers. For mechanistic analysis, focus was placed primarily on results obtained by driving targeting genes or RNAi constructs in muscle 12 only and examining the consequences for innervation of muscle 12 by RP5 and the 1s neuron. These results are interpretable because the axons are known and only the last stage of their targeting should be affected. By contrast, mistargeting phenotypes observed with panmuscle expression could result from errors at any point along the axonal trajectory, and the axons that mistargeted cannot be identified in many cases. There may also be targeting errors that cannot be detected using mAb 1D4, because it labels all motor axons. Only NMJs that display morphological abnormalities can be seen and it was possible to miss phenotypes in which an axon from one motor neuron is replaced by another one, if it forms a similar NMJ. Ideally, these experiments should be performed using reagents that label single identified motor axons, but these are not yet available (Kurusu, 2008).

LRRs are ~24 aa protein domains that can be found outside the cell or in cytoplasmic proteins. A chain of LRRs forms a concave binding surface that is used for interactions with other proteins. The LRR domains of Trn and Caps are interchangeable, suggesting that they can interact with a common receptor (Milán, 2005). Trn and Caps are involved in cell-cell interactions in tracheae and imaginal discs, and Caps regulates layer-specific targeting in the optic lobe (Kurusu, 2008).

Within the VLM field, Caps is expressed on muscle 12, while Trn appears to be expressed on all VLMs, but with higher levels on muscles 6 and 7. trn caps double mutant embryos have stronger motor axon phenotypes than trn single mutants, and they exhibit ISNb terminal loop phenotypes that are suggestive of RP5 mistargeting (Kurusu, 2008).

Because muscle 12 NMJs send loopback branches onto muscle 13 when Caps is expressed in all muscles, it was proposed to be an attractive cue which facilitates targeting of RP5 to muscle 12. However, the actual situation may be more complex, because selective overexpression of either Caps or Trn on muscle 12 produces phenotypes in which muscle 12-destined axons either stall on muscle 13, so that muscle 12 remains uninnervated, or grow under muscle 13 rather than over it to reach muscle 12. These apparently nonautonomous effects (alteration of axonal extension over an adjacent muscle) might be explained by Trn- or Caps-induced alterations in the pattern of myopodia, projections from the muscle that reach out to contact innervating axons and direct their growth. Myopodia can extend over distances similar to the width of a muscle fiber. Perhaps when Trn or Caps is expressed on muscle 12, the myopodia extend under muscle 13 rather than over it. If RP5 axons contact these aberrant myopodia, they may grow under 13 to reach 12; if they fail to contact them, they may stall on muscle 13 (Kurusu, 2008).

CG14351, which is denoted in this study as Haf, is a large protein (1316 aa), and LRRs occupy only aa ~100-350 of the XC domain. Haf has a signal sequence, a single transmembrane region, and a large cytoplasmic domain (~500 aa). It appears to be expressed by all VLMs (Kurusu, 2008).

The embryonic motor axon phenotypes observed in the haf insertion mutant and in haf RNAi x pan-muscle-GAL4 embryos indicate that ISNb cannot innervate any of the VLMs in a normal manner if Haf is not expressed in muscles. Only 40%-50% of ISNbs have a normal morphology. The remainder bypass onto the ISN or SNa or follow abnormal trajectories within the VLM field, sometimes contacting inappropriate targets. The phenotypes are highly variable, suggesting that many different kinds of errors are produced by loss of Haf (Kurusu, 2008).

It is suggested that in the embryo Haf is a permissive muscle factor that is required for target selection by all muscles within the VLM field but does not define the identities of specific fibers. This model is also consistent with the larval phenotypes that result from knocking down Haf expression on muscle 12 only. In this case, the RP5 and 1 s axons seem to reach muscle 12 and form NMJs in a normal manner, but they also extend further and form ectopic synapses on lateral muscles 5 or 8. This phenotype suggests that a stable NMJ on muscle 12 sometimes cannot form when Haf is knocked down, and in these cases the axons (or NMJ branches) continue to grow until they reach the lateral muscles (Kurusu, 2008).

CG8561 is a 1092 aa protein that has a signal sequence but lacks a transmembrane region. Its C-terminal sequence is characteristic of proteins that are attached to membranes by glycosyl-phosphatidylinositol anchors. It appears to be expressed by all muscles (Kurusu, 2008).

CG8561 muscle RNAi and muscle overexpression produce mistargeting and NMJ arbor phenotypes in larvae. When CG8561 RNAi is expressed in muscle 12 only, the muscle 12 NMJ sends loopback branches to muscle 13 in 70% of affected hemisegments. This implies that CG8561 confers a preference for the RP5 and 1 s axons to choose muscle 12, and in its absence these axons do not strongly prefer muscle 12 to the adjacent muscle 13 (Kurusu, 2008).

CG8561 has a vertebrate ortholog, the acid-labile subunit (Als) of the IGF-1 binding complex. CG8561 mRNA is expressed in the larval fat body (FB) and in a group of neurosecretory cells (NSCs) that express insulin-like peptides. Starvation causes downregulation of the mRNA in the FB and NSCs. These data suggest that CG8561 (dAls) may be involved in insulin/IGF-1 signaling. Interestingly, the single fly insulin/IGF-1 receptor, InR, is expressed in neurons and is required for guidance of photoreceptor axons into the optic lobe (Kurusu, 2008).

Identification of 30 mistargeting genes among the 410 CSS genes that were screened suggests that there may be ~70 mistargeting genes in the entire cell-recognition database, and perhaps twice that many if genes with lower mistargeting percentages are included. If it is assumed that the screen is capable of identifying all CSS proteins involved in targeting, the results seem inconsistent with a simple version of a lock-and-key model, because mRNAs or proteins within the mistargeting set were not found that were expressed in small subsets of muscles. Also, the four LRR proteins examined in this paper have complex effects on targeting that are not explainable by a simple model (Kurusu, 2008).

Fifty-three LRR proteins were screened, and 16 of these produced either mistargeting, NMJ phenotypes, or both. If the failure of an LRR protein to produce a phenotype when overexpressed in muscles in the course of the screen indicates that it is not involved in targeting or synapse development, then there are 48 signal sequence-containing LRR proteins that remain to be examined for expression patterns, GOF phenotypes, and LOF phenotypes. A comprehensive analysis of this large family may help to define mechanisms involved in target selection and synaptic growth (Kurusu, 2008).

Assessing the role of cell-surface molecules in central synaptogenesis in the Drosophila visual system
A hallmark of the central nervous system is its spatial and functional organization in synaptic layers. During neuronal development, axons form transient contacts with potential post-synaptic elements and establish synapses with appropriate partners at specific layers. These processes are regulated by synaptic cell-adhesion molecules. In the Drosophila visual system, R7 and R8 photoreceptor subtypes target distinct layers and form en passant pre-synaptic terminals at stereotypic loci of the axonal shaft. A leucine-rich repeat transmembrane protein, Capricious (Caps), is known to be selectively expressed in R8 axons and their recipient layer, which led to the attractive hypothesis that Caps mediates R8 synaptic specificity by homophilic adhesion. Contradicting this assumption, the current results indicate that Caps does not have a prominent role in synaptic-layer targeting and synapse formation in Drosophila photoreceptors, and that the specific recognition of the R8 target layer does not involve Caps homophilic axon-target interactions. Flies were generated that express a tagged synaptic marker to evaluate the presence and localization of synapses in R7 and R8 photoreceptors. These genetic tools were used to assess how the synaptic profile is affected when axons are forced to target abnormal layers by expressing axon guidance molecules. When R7 axons were mistargeted to the R8-recipient layer, R7s either maintained an R7-like synaptic profile or acquired a similar profile to r8s depending on the overexpressed protein. When R7 axons were redirected to a more superficial medulla layer, the number of presynaptic terminals was reduced. These results indicate that cell-surface molecules are able to dictate synapse loci by changing the axon terminal identity in a partially cell-autonomous manner, but that presynapse formation at specific sites also requires complex interactions between pre- and post-synaptic elements (Berger-Muller, 2013).

The axon mistargeting experiments shed light on the relationship between axon targeting and synaptogenesis. Since these two processes are intimately linked, it is difficult to assess how axon guidance can influence synapse formation. This paper shows that that Caps does not have a role in synaptogenesis in photoreceptors, but Caps overexpression can be used to retarget R7 axons to an incorrect synaptic layer. Thus, Caps overexpression is a useful tool to assess the effect of axon guidance on synaptic specificity. When R7 axons were redirected to the R8-recipient layer by Caps overexpression, the pattern of synaptic connections was different from R8 axons. This result indicates that synaptic specificity does not depend solely on axon targeting. It also suggests that although Caps expression can confer the 'R8 identity' to R7 axon terminals in terms of axon targeting, it does not have a role in determining the R8 synaptic profile on photoreceptor axon shaft. To note, the rhodopsin promoter Rh4, which was used to visualize synapses in mistargeted R7, was activated, suggesting that photoreceptor identity was unchanged (Berger-Muller, 2013).

On the contrary, when R7 axons were misguided to the R8-recipient layer by Gogo and Fmi overexpression, the synaptic profile was similar to R8 photoreceptors. This indicates that, endogenously, Gogo and Fmi could control synapse formation at subcellular level in R8 photoreceptor axons, directly or indirectly. It also suggests that Gogo and Fmi co-expression can dictate photoreceptor neurons to transform their axonal identity to that of R8s (Berger-Muller, 2013).

A coordinated control of transcriptional factors, including Prospero and NF-YC, is known to control R7 cell identity, initially its axon targeting and later the rhodopsin expression during development. Thede data indicates that the synaptic profile is likely part (or downstream) of such cell fate determination, and that Gogo and Fmi can be important factors of the R8 genetic program (Berger-Muller, 2013).


Capricious and Tartan, two transmembrane proteins with leucine-rich repeats, contribute to formation of the affinity boundary between dorsal and ventral compartments during Drosophila wing development. Engrailed/Invected expression confers posterior (P) identity and Apterous (Ap) confers dorsal (D) identity in the wing disc. P compartment cells lacking engrailed/invected activity do not respect the anterior-posterior boundary. Likewise, dorsal cells lacking ap activity fail to respect the dorsal-ventral (DV) boundary in the wing disc. Modulation of Notch signaling has been implicated in DV boundary formation. Fringe acts as a glycosyltransferase to modify the receptor protein Notch in the dorsal compartment. Fringe activity makes D cells more sensitive to Delta, a ligand expressed by V cells and less sensitive to Serrate, the ligand expressed by D cells. Consequently, signaling by each ligand is limited to nearby cells on the opposite side of the boundary, with the result that high levels of Notch activity are limited to a narrow band of cells along the DV boundary. Although altering the signaling properties of cells by modulation of Fringe activity has been shown to allow cells to cross the boundary, Fringe activity has been shown to be insufficient to support boundary formation. This observation, together with the fact that Notch signaling is activated symmetrically has suggested that other Apterous-dependent cell interactions might be needed for formation of the DV affinity boundary. Evidence suggests that capricious and tartan are targets of Apterous that contribute to DV boundary formation in the wing disc. caps and tartan are expressed in the D compartment during boundary formation. Caps and Tartan confer affinity for D cells, assessed by sorting-out behavior. Caps supports boundary formation without conferring D signaling properties. Fringe, in contrast, confers dorsal signaling properties without affecting DV affinity. Thus, Caps, Tartan, and Fringe have complementary roles in boundary formation (Milán, 2001).

In second and early third instar wing discs, caps-lacZ is expressed in the D compartment. At this stage, expression of Caps protein coincides with that of Apterous, and caps-lacZ expression depends on Apterous activity. Expression of the Apterous inhibitor, dLMO, under control of patched-Gal4 represses caps-lacZ expression. During third instar, dorsal expression of caps-lacZ decreases and new lateral expression domains arise. These domains are initially stronger in the D compartment but become symmetric in D and V compartments in mature third instar discs. tartan expression was monitored using a lacZ reporter gene and antibody to Tartan protein, and is similar to caps expression at all stages. The dynamics of these expression patterns suggested that Caps and Tartan proteins might mediate cell interactions during early DV patterning and subsequently during medial-lateral patterning of the wing (Milán, 2001).

To assess the roles of Caps and Tartan in DV boundary formation, use was made of a rescue assay in which the Gal4-UAS system was used to restore Caps and Tartan expression in D cells of apterous mutant wing discs. apGal4/ap- mutant discs are not able to form a smooth DV boundary and fail to induce Wg expression uniformly along the interface between D and V cell populations. Expression of Caps in D cells under apGal4 control restores a smooth interface between D and V cells in the mutant discs, but does not restore Wg expression along the boundary. Tartan is considerably less effective at producing a smooth interface between D and V cells. Connectin, a GPI-anchored membrane protein that is related to Caps and Tartan in the LRR domains, was tested because Connectin has been shown to mediate homophilic cell adhesion. Connectin was ineffective in the boundary rescue assay. Fasciclin II, an unrelated adhesion protein, was also unable to restore the DV boundary. These observations suggest that Caps expression produces an affinity boundary between D and V cells by a mechanism that is not simply due to increased cohesion among dorsal cells (Milán, 2001).

To test the requirement for Caps and Tartan in boundary formation, clones of cells mutant for caps or tartan were produced. Single mutant clones did not produce observable alterations in the wing disc. Clones simultaneously mutant for caps and tartan do not cause defects at the DV boundary, but do perturb medial-lateral cell interactions. In wild-type discs, loss of caps and tartan activity may be compensated for by other proteins. It was therefore asked whether reduced levels of caps and tartan activity would cause defects when DV boundary formation was compromised by reduction of Ap activity. Two sensitized genetic backgrounds were examined. Bx1 produces a wing scalloping phenotype that is very sensitive to the level of expression of other genes involved in DV patterning. Bx1 is a dominant mutation that overexpresses the dLMO protein. dLMO competes with Ap for binding to its cofactor Chip and thereby reduces Ap activity. The second sensitized genotype was provided by a mutant with reduced expression of the Ap cofactor Chip. Chipe5.5 was selected because it is less sensitive to modification than Bx1 but shows specific genetic interactions with ap, dLMO, Serrate, and fringe. Interactions were scored on the basis of dominant wing scalloping phenotypes in flies heterozygous for Chipe5.5 or Bx1. Deletions in 18 different genes dominantly enhanced both phenotypes. Nine of these uncover genes with known roles in DV patterning, including ap, vestigial, cut, and Serrate. Df(3L)C190, a deletion that removes the caps and tartan genes, enhances the Bx1 and Chipe5.5 phenotypes. The caps tartan double mutant chromosome used for the clonal analysis also enhances both phenotypes. The contributions of caps and tartan were then tested individually. A caps lack-of-function mutant enhances both phenotypes. A tartan mutant lack-of-function mutant enhances Bx1, but does not produce a phenotype in the less sensitive Chipe5.5 background. These observations indicate that reduced caps and tartan activity causes wing defects when the system is sensitized by reduction of Ap activity. These observations suggest that Caps and Tartan contribute to DV boundary formation (Milán, 2001).

Use was made of the flip-out Gal4 system to produce clones of Gal4-expressing cells in the wing disc to examine effects of ectopic Caps and Tartan expression. Of 178 clones examined, 23 contacted the boundary on the D side compared to 25 on the V side. Sixty-four clones were located internally in the D compartment, compared to 66 in the V compartment. Caps-expressing clones differ in two respects from control clones. Fewer Caps-expressing clones were recovered in the V compartment (V/D ratio ~0.7). Nonetheless, twice as many V clones were recovered at the boundary as would be expected if Caps had no effect on their distribution. Comparable results were obtained for clones expressing Tartan or Caps and Tartan together. These observations suggest that clones expressing Caps or Tartan survive poorly in the V compartment. To ask whether poor survival of V clones could be responsible for their accumulation at the DV boundary, Caps and Tartan were co-expressed with the viral apoptosis inhibitor p35. p35 expression suppresses the loss of V clones. Yet, V clones were still overrepresented by ~2-fold at the DV boundary. Control clones expressing GFP and p35 were evenly distributed between D and V compartments. These observations suggest that V clones expressing Caps or Tartan survive poorly if they fail to contact D cells, and that Caps or Tartan expression causes V cells to sort-out toward the D compartment (Milán, 2001).

Caps- and Tartan-expressing clones of V compartment origin sort-out toward the DV boundary but remain in the V compartment. Although these clones do not cross the boundary, many of them appear to push the Wg stripe toward dorsal. To examine this behavior more closely, clones were produced in early 2nd instar discs, before the DV boundary forms. Some of these clones were bisected by the nascent DV boundary so that they contributed to both compartments (referred to as D+V clones). Control D+V clones expressing GFP or lacZ reporter genes have no effect on the shape of the Wg stripe, and most are of similar size in both compartments. These clones were generally elongated in shape and had irregular borders where they contacted neighboring wild-type cells. In contrast, D+V clones expressing Caps or Tartan are more compact in shape, have smoother borders, and tend to be considerably smaller in the V compartment. Many of these clones distorted the Wg stripe where they crossed the boundary. D+V clones expressing Caps and Tartan together had similar effects (Milán, 2001).

The effects of smaller clones on the shape of the DV boundary were examined, using expression of an ap-lacZ reporter gene to mark dorsal cells. GFP-expressing clones that contact the DV boundary have no effect on the ap-lacZ border or on the Wg stripe. In contrast, ventral Caps- or Tartan-expressing clones often displace both the ap-lacZ border and the Wg stripe toward dorsal. In one case, a ventral Caps-expressing clone was observed that had separated a group of D cells from the rest of the D compartment. The effects of clones expressing Tartan or Caps and Tartan together were similar to those of Caps-expressing clones. D compartment clones had no effect (Milán, 2001).

Although it is not possible to observe how these distortions of the DV boundary arise, it is tempting to speculate that they result from V cells attempting to sort-out into the D compartment. Sorting out could be caused by increased affinity for D compartment cells or by repulsion by V compartment cells. Either mechanism could provide a force to push the clones into the D compartment and displace D cells and the Wg stripe. Since Caps and Tartan are expressed by D cells, it was asked whether V clones are attracted to the D compartment by homophilic cell adhesion mediated by Caps and Tartan. To measure homophilic adhesion, use was made of a cell aggregation assay that involved the use of cultured S2 cells. S2 cells expressing Caps and Tartan do not aggregate more than control cells. Likewise, binding of a secreted Caps-Alkaline Phosphatase fusion protein to cells expressing Caps, Tartan, or both could not be detected. Thus, the sorting-out behavior of Caps- and Tartan-expressing clones is unlikely to depend on homophilic cell adhesion mediated by these proteins. It is proposed that Caps and Tartan interact with other surface proteins expressed in the D compartment (Milán, 2001).

To examine how Caps and Tartan induce sorting behavior, clones that had sorted toward the DV boundary were observed using confocal microscopy. Membrane-bound cellular processes were observed extending from Caps-expressing cells toward cells in the D compartment. Caps protein outlined the cell surface and appeared in bright spots that may be intracellular vesicles. In more apical sections, Caps protein was located on thin structures that extended into the dorsal compartment. In cross-section, these structures can be seen to project from V cells over the apical surface of nearby D cells. Since Caps is a membrane protein, it is inferred that these are membranous cellular processes, perhaps filopodia. Processes were also observed projecting toward the D compartment from V clones that were not in contact with the boundary. Similar structures extending between closely spaced clones within the V compartment were not observed. All projections were oriented toward the D compartment. These observations support the idea that Caps may interact with another cell surface protein in the D compartment (Milán, 2001).

Tartan-expressing clones also sort-out toward the D compartment. Clones of Tartan-expressing cells were examined: cellular projections could not be visualized with anti-Tartan antibody. As an alternative, use was made of a transgene expressing cytoplasmic ß-Gal to mark cellular processes when coexpressed with Tartan. Ventral Tartan-expressing clones also extend cytoplasmic processes toward D cells. Projections produced by cells expressing Caps were similar in appearance when visualized using cytoplasmic ß-Gal. Cellular processes expressing Caps and Tartan may help ventral cells to sort toward the D compartment. These observations suggest that the behavior of Caps- and Tartan-expressing V clones is guided by increased affinity for D cells (Milán, 2001).

Why are V clones expressing Caps and Tartan unable to cross the Wg stripe. Caps and Tartan are not able to restore Notch signaling and Wg expression when expressed in the D compartment of ap mutant discs. Likewise, clones of cells expressing Caps or Tartan do not induce Wg expression in adjacent V cells. Instead, Wg is expressed normally where Tartan or Caps-expressing V cells contacted D cells. Thus, ventral Caps- and Tartan-expressing clones retain ventral signaling properties. In this respect, they differ from ventral Fringe-expressing clones, which acquire the signaling properties of D cells (Milán, 2001).

Fringe acts as a glycosyltransferase enzyme to modify Notch and make it differentially sensitive to its ligands. Consequently, ventral clones of Fringe-expressing cells induce ectopic expression of Wg where they contact other V cells. In cases where ventral Fringe-expressing clones contact the DV boundary, Wg is induced at the interface with other V cells, but not at the interface with D cells. The change in signaling properties of these cells results in relocation of the stripe of Wg expression to the interface between the clone and other V cells. Consequently, Fringe-expressing clones cross the boundary defined by the Wg stripe. Caps- and Tartan-expressing clones retain ventral signaling properties and so cannot reposition the Wg stripe (Milán, 2001).

The behavior of Fringe-expressing clones differs in a second respect. Fringe-expressing clones are not lost from the V compartment and do not accumulate at the DV boundary. Thus, Fringe-expressing clones do not acquire the ability to sort-out toward D cells, an ability that is conferred by Caps or Tartan expression. These properties are reflected in the different abilities of Fringe and Caps to rescue the DV affinity boundary in ap mutant discs. Expression of Fringe restores Notch signaling and induces Wg expression, but is unable to produce a smooth DV affinity boundary. In contrast, Caps produces a smooth boundary but does not restore Notch signaling. Coexpression of Caps and Fringe restores Wg expression and produces a normal DV affinity boundary. Likewise, clones expressing Caps, Tartan, and Fringe sort-out toward the DV boundary and cross into the D compartment. These observations suggest that Fringe and Caps/Tartan play distinct but complementary roles in boundary formation (Milán, 2001).

It is proposed that the activities of Caps and Tartan, as well as those of the Notch ligands and Fringe, are required for DV boundary formation. Apterous controls expression of Serrate and Fringe as well as Caps and Tartan in dorsal cells during boundary formation. The ligands for Notch are transmembrane proteins. Therefore, ligand-receptor binding may contribute to adhesion between D and V cells at the boundary while inducing signaling. By increasing the affinity for Delta, Fringe may promote binding between Notch on D cells and Delta on V cells. Likewise, by reducing the affinity of D cells for Serrate, Fringe may promote binding between Serrate on D cells and Notch on V cells. Increased binding between oppositely specified cells is likely to help to stabilize the interface between the two cell populations, but seems unlikely to help drive the initial segregation of the populations needed to establish a smooth boundary. Indeed, restoring Fringe and Serrate expression in apterous mutant discs is not sufficient to restore a normal DV boundary. It is proposed that the transient expression of Caps and Tartan in D cells initiates the segregation of the two cell populations. Once they are separated, Fringe-dependent cell interactions may stabilize the boundary. Fringe has also been implicated in boundary formation in vertebrate limbs. It will be of interest to learn whether Caps and Tartan homologs play comparable roles in DV boundary formation in vertebrates (Milán, 2001).

Caps and Tartan expression induce the formation of cellular processes that projected from V cells toward D cells. Cytonemes and similar structures have been proposed to mediate long-range cell interactions in imaginal discs. The structures observed appear to differ from cytonemes in that they project across the signaling center into the opposite compartment, rather than projecting toward the signaling center. Filopodia have been implicated in guiding morphogenetic movements in epithelial sheets. Filopodia expressing E-Cadherin have been implicated in the formation of adhesive zippers between epithelial cells, which serve as nucleation centers for reorganization of the cytoskeleton. It is proposed that imaginal disc cells use filopodia that express cell-surface proteins, including Caps and Tartan, to assess the identity of nearby cells and to control cell behavior. Caps and Tartan do not appear to mediate homophilic adhesion. This suggests that dorsal cells express another cell surface protein that is able to bind the LRR domains of Caps and Tartan. Expression screening and systematic searches for membrane proteins expressed on D cells may help to identify the Caps/Tartan binding partner (Milán, 2001).

The mechanisms by which tissue boundaries form are not well understood. Differences in cell adhesion can contribute to tissue boundary formation. Sorting-out of cell populations can be guided by both the amount and types of adhesion proteins that cells express. A different view comes from studies on Ephrins and Eph receptors, which suggest that repulsion or deadhesion can promote segregation of cell populations. Many adhesion proteins form regulated connections with the cytoskeleton and participate in contact-mediated signaling. A useful distinction can be made between initial cell-cell contacts, which may be transient, and formation of stable contacts that may involve substantial reorganization of the cytoskeleton. If signaling promotes reorganization of the cytoskeleton, cell interactions might be destabilized. Repeated cycles of deadhesion and readhesion could lead to sorting out behavior. It is possible therefore that adhesive differences and differences in cell behavior both contribute to forming tissue boundaries. At present, it is not clear which type of explanation best describes formation of the compartment boundaries in imaginal discs. Caps and Tartan are cell surface proteins that mediate cell interactions. These findings suggest that contacts with appropriately specified cells mediated by Caps and Tartan might be stabilized, whereas contacts with inappropriate cells might be destabilized. A deeper understanding of these processes awaits identification of the cell surface proteins with which Caps and Tartan interact (Milán, 2001).

Axon guidance of Drosophila SNb motoneurons depends on the cooperative action of muscular Krüppel and neuronal capricious activities

The body wall musculature of the Drosophila larva consists of a stereotyped pattern of 30 muscles per abdominal hemisegment that are innervated by about 40 distinct motoneurons. Proper innervation by motoneurons is established during late embryogenesis. Guidance of motor axons to specific muscles requires appropriate pathfinding decisions as they follow their pathways within the central nervous system and on the surface of muscles. Once the appropriate targets are reached, stable synaptic contacts between motoneurons and muscles are formed. Recent studies have revealed a number of molecular components required for proper motor axon pathfinding and demonstrate specific roles in fasciculation/defasciculation events, a key process in the formation of discrete motoneuron pathways. The gene capricious (caps), which encodes a cell-surface protein, functions as a recognition molecule in motor axon guidance, regulating the formation of the selective connections between the SNb-derived motoneuron RP5 and muscle 12. Krüppel (Kr), best known as a segmentation gene of the gap class, functionally interacts with caps in establishing the proper axonal pathway of SNb including the RP5 axons. The results suggest that the transcription factor Krüppel participates in proper control of cell-surface molecules that are necessary for the SNb neurons to navigate in a caps-dependent manner within the array of the ventral longitudinal target muscles (Abrell, 2001).

Use has been made of a Kr gain-of-function mutation, termed Irregular facets (If), to identify modifiers of Kr activity during eye morphogenesis. One suppressor of ectopic Kr activity in the eye, the P-element insertion l(3)02937, was found to reside within the gene caps. caps encodes a CAM of the 'leucine-rich repeat' family and has been shown to be required for proper pathfinding and synapsing of the RP5 motoneuron with muscle 12. In addition to l(3)02937, a second P-element insertion, l(3)05121, was identified that resides in the first exon of the caps gene. Genetic analyses showed that the two P-element insertions failed to complement each other and the previously identified caps65.2 null-mutation, indicating their allelism to caps. Furthermore, precise excision of the P-element insertion l(3)05121 results in a reversion of the caps mutant phenotype to wildtype, indicating that the P-element insertion is the cause of the mutation. Moreover, the phenotype caused by the newly identified caps alleles is indistinguishable from the caps65.2 loss-of-function phenotype, suggesting that they represent either strong hypomorphic or lack-of-function caps mutations (Abrell, 2001).

While the innervation of muscles occurs during embryogenesis, previous studies on caps function have focussed on mutant defects that were observed in the motoneuronal pattern of third instar larvae, showing that Caps is necessary for proper pathfinding of the RP5 axons. RP5 axons are part of the SNb fascicle. During embryogenesis, the SNb enters the ventral muscle field between muscle 15 and 28 (choice point 'target entry'). Its axons pass the ventral oblique muscle field towards the ventral longitudinal muscle targets in the most internal muscle layer. Close to the second choice point of the SNb nerve, the RP3 axon separates from SNb to target the cleft between muscles 7 and 6. The SNb continues along muscle 14 and enters the ventral longitudinal muscle field. At a position close to muscle 30, the RP1 and RP4 axons separate to target muscle 13. The remaining RP5 bypasses muscle 13 to synapse with muscle 12. This muscle as well as the RP5 axons are characterized by caps expression both in the embryo and in larvae (Abrell, 2001).

The early route of SNb pathfinding and the defasciculation patterns of RP1, RP3 and RP4 are not affected in caps homozygous larvae, whereas the pathway of RP5 is altered. RP5 loses its target specificity and synapses with both muscle 12 and 13 instead of muscle 12 only. Furthermore, overexpression of caps in all muscles causes the same phenotype. These results were taken to indicate that caps plays an important role in selective target recognition and synapse formation by the motoneuron RP5 (Abrell, 2001).

In order to see whether and to what extent caps affects neuromuscular connectivity in the embryo, the phenotypes of the different caps mutants were examined. Of a total of 111 homozygous caps65.2 mutant individuals examined, all embryos developed into normal looking larvae. However, the majority of the caps65.2 mutant larvae (62%) failed to hatch. Out of 81 hatched larvae examined, only two individuals survived to pupal stages and developed into adults. Similar results were obtained with homozygous capsl(3)02937 (64% unhatched larvae; 3% adults) and capsl(3)05121 (73% unhatched larvae; 3% adults) mutants. The two newly identified caps alleles are therefore similar in strength to caps65.2, previously shown to be a lack-of-function mutation. Moreover, these data indicate that caps functions primarily during embryogenesis and that previous studies on caps function in larvae have been performed with the fraction of mutants that develop into larvae (Abrell, 2001).

SNb development was examined in homozygous mutant embryos using monoclonal antibodies directed against the CAM FasII. In all three caps mutants, the SNb enters the ventral muscle field and the RP axons defasciculate normally. Furthermore, RP1, RP3 and RP4 properly synapse with their respective target muscles, whereas the RP5 axons stall, show enlarged growth cone-like structures and fail to contact the target muscle 12. Instead, the RP5 axons are split and found also in direct contact with the transversal nerve (TN), a link never observed in wildtype. These phenotypes suggest that pathfinding of RP5 axons during embryogenesis is retarded and in cases where the RP5 axons defasciculate and elongate, they show erratic targeting and synapsing, as has been described for third instar larvae (Abrell, 2001).

The specific SNb defect of the caps mutants on RP5 pathfinding and its failure to synapse exclusively with muscle 12 correlates with the observation that both RP5 neurons and their target muscle are characterized by caps expression both in embryos and larvae. In addition, overexpression of caps in all muscles causes a phenotype similar to loss-of-function mutations, suggesting that relative levels of Caps are important for pathfinding. Thus, both lack-of-function and gain-of-function studies indicate a specific role for caps in selective target recognition and synapse formation. The results show that this function of caps is required during embryogenesis and that the previously reported phenotype observed in third instar larvae represents only a weak phenotype common to escapers. In all other cases, impairment of caps activity causes the RP5 axons to stall immediately after defasciculation from the SNb fascicle. This observation indicates that caps is a critical component specifically required for RP5 pathfinding after the defasciculation of the RP5 axons from SNb has occurred (Abrell, 2001).

caps is expressed in both motoneuron RP5 and its synaptic target muscle 12 in third instar larvae. In view of the embryonic caps mutant phenotypes, it was asked whether caps is also expressed during early embryogenesis and whether Kr (which is expressed in neuroblasts and in a distinct subset of embryonic muscles) could be responsible for the control of caps expression (Abrell, 2001).

caps is indeed expressed in neuroblasts. However, this aspect of the caps expression pattern was not altered in Kr lack-of-function mutant embryos except in the central region which is distorted due to the earlier segmentation function of Kr. caps expression was monitored in Krres mutant embryos in which the early segmentation function of Kr is specifically rescued. No difference was observed in the caps expression patterns of Krres mutant and wildtype embryos and the Kr neural expression pattern was unchanged in caps mutant embryos, indicating that Kr and caps activities are independently controlled. Since caps acts as a dose-dependent modifier of ectopic Kr activity during eye formation, it was next asked whether later aspects of the expression patterns, caps expression in RP5 neurons and their target muscles and Kr expression in a subset of muscles along the SNb pathway, might reflect the need of the two gene activities for the proper guidance of SNb axons (Abrell, 2001).

Previous work has shown that in the absence of Kr activity, the SNb stalls at the second choice point and RP axons fail to defasciculate. This observation is consistent with the proposal that the two genes act in the same genetic pathway. In order to test this proposal, genetic interaction studies were performed using capsl(3)05121 and Krres mutant combinations, asking whether a reduction of Kr and caps activities causes defects in neuromuscular connectivity (Abrell, 2001).

Heterozygous Krres/+ or capsl(3)05121/+ single mutant embryos developed a normal motoneuron pattern. Each of the RP axons was properly connected to its target muscle as revealed by anti-FasII antibody staining. Thus, a reduction of either Kr or caps activity has no effect on motoneuron development and pathfinding. In contrast, double heterozygous Krres/+;capsl(3)05121/+ embryos, where the dosage of both genes was reduced at the same time, develop a specific SNb nerve phenotype without affecting the ISN and SNa. In about one-third of the cases, the SNb stop along ventral longitudinal muscles, ending with a large growth cone-like structure. In addition, properly defasciculated RP axons fail to continue along their normal paths; a portion of them elongate and stall either in a position very close to the TN or is directly connected to it. Double homozygous Krres;capsl(3)05121 mutants develop an even stronger phenotype; the SNb is absent in most of the double mutants analysed or does not extend beyond its second choice point close to muscle 28. In only few cases, the SNb stalls in the ventral muscle field as had been described for homozygous Krres embryos (Abrell, 2001).

These phenotypes were also obtained with similar frequencies in double homozygous Krres;caps65.2 mutants, indicating that the phenotype is not dependent on a particular caps allele. The results show in addition that the defects are stronger and more pronounced in double mutant embryos than those obtained with single mutant embryos (Abrell, 2001).

The failure to detect the SNb nerve in the majority of homozygous Krres;capsl(3)05121 mutant embryos correlates with a thickening of the ISN. In order to test whether SNb might have lost its identity due to a transformation into ISN identity, the SNb-derived RP neurons were labelled by virtue of the transgenic islH-tau-myc marker gene. The results show that the RP axons were present. However, the SNb fails to separate from the ISN or in the cases separation occurs, it stalls shortly after the defasciculation. These observations indicate that the SNb is not transformed into ISN identity and suggest that the SNb has lost the capability to respond to guidance cues such as CAMs and repellents. Moreover, the results indicate that Kr and caps activities cooperate in a synergistic fashion necessary for proper defasciculation of the SNb axons at the exit junction and for RP axon guidance in the ventral muscle field (Abrell, 2001).

Previous studies have shown that overexpression of caps in all neurons causes a specific misrouting or stalling of RP5 at the second choice point near muscle 30 in about one-third of the embryos. This effect of panneural caps expression is dependent on the extracellular domain of Caps, suggesting that Caps functions as a cell-adhesion component which participates in the guidance of the SNb at the specific choice point near muscle 30. In contrast, panmuscular expression of caps has no effect on SNb guidance and pathfinding, but severely interfers with synapsing of RP5 resulting in connections being formed not only with muscle 12, but also with the neighboring muscle 13 (Abrell, 2001).

In order to investigate whether misexpression of Kr can interfere with SNb formation, Kr was ectopically expressed in all motoneurons or muscles using the Gal4/UAS system. To achieve this, the ftzNG-Gal4 ('panmotoneuronal expression’) and the 24B-Gal4 driver lines ('panmuscular expression') were used in combination with one and two copies of UAS-Kr transgenes. Panmotoneural Kr expression from one transgene in wildtype embryos results in a minor phenotype of the SNb, in which the distal RP axons fail to reach their target muscles and maintain growth cone-like structures at their ends. Panmotoneural Kr expression from two transgenes causes a stronger phenotype. In 36% of the cases, the most distal RP axons stalls and the RP5 axon does not innervate the target muscle 12, whereas in all other cases, the SNb stalls at the second choice point, a phenotype that is similar to the Krres homozygous mutant phenotype. This observation indicates that the Kr overexpression phenotype is dosage-dependent and that both the lack-of-function and gain-of-function effects of Kr interfere with SNb development. Since Kr is a cell-autonomous transcription factor, it is likely that it is required for and can interfere with the transcription of motoneuronal genes necessary for proper motoneuronalguidance (Abrell, 2001).

In the wildtype embryos, Kr is not only expressed in the nervous system but also in specific subsets of muscle founder cells and muscles. Relevant sites of Kr expression during the formation of neuromuscular connectivity are the ventral oblique muscles 14 and 16, the ventral acute muscle 27 and the ventral longitudinal muscles 6, 7 and 13. Heterozygous Krres mutant embryos develop a normal muscle pattern, whereas in homozygous Krres mutant embryos muscle 27 is transformed into a duplicated muscle 26. The other muscles that normally express Kr either appear to be normal or develop a variably altered morphology. Overexpression of four copies of Kr in all muscles leads to the reverse result, i.e. muscle 26 is transformed into a second muscle 27. Thus, Kr is necessary to determine muscle identity and can alter muscle fate upon ectopic expression in muscles that normally do not express the gene (Abrell, 2001).

Pan muscular expression of Kr in response to one or two copies of the 24B-driven Gal4/UAS-Kr cDNA transgenes does not disturb the muscle pattern. However, it has severe consequences for the formation and defasciculation of SNb. With one copy of Kr, the RP3 axon separates properly from SNb and succeeds in finding the cleft between target muscles 7 and 6, whereas the remaining RP axons fail to defasciculate. In the majority of cases, however, the SNb passes the ventral oblique muscles and enters the ventral longitudinal muscles normally, but it stalls at the second choice point and RP axons fail to defasciculate. Two copies of transgene-derived Kr expression in all cases cause stronger defects that the SNb stalled at the position where RP3 would normally defasciculate. The same Kr-dependent phenotypes are found in response to a different panmuscular Gal4 driver, namely the twi Gal4 line. Thus, in contrast to panmuscular expression of caps, panmuscular expression of Kr appears to interfere with a muscle-specific program that regulates defasciculation of the RP axons and/or the elongation of SNb after the second choice point has been reached. The finding is consistent with the argument that Kr determines the spectrum of molecules expressed in muscles that are used to transmit signals to other cells, namely cell-surface and secreted molecules (Abrell, 2001).

In order to test whether panmuscular Kr expression interferes with SNb guidance in a Kr- and caps-dependent manner, the phenotype was examined of double homozygous Krres; capsl(3)05121 mutant embryos that receive panmuscular Kr expression. Upon panmuscular Kr expression, the SNb is absent in 70% of the cases. In the other cases, the SNb had formed and separated from the ISN, but stalled in the region of the ventral longitudinal muscles. The observed phenotype is reminiscent of the phenotype obtained with Krres;capsl(3)05121 double homozygous mutants, but a higher proportion of axons defasciculate in the exit junction region. This observation indicates that panmuscular Kr expression can partially rescue the SNb phenotype of the double homozygous mutants. However, defasciculation from the stalled SNb occurs in an erratic manner, implying that axon guidance is still strongly impaired. The lack of a better rescue is likely to be due to the fact that Kr has not been expressed in its wildtype muscular pattern and/or that the correct levels and neuronal Kr activity are not provided through panmuscular expression. Nevertheless, the results are consistent with the proposal that Kr regulates a muscular programme which in turn regulates SNb axon guidance along the muscles. This proposal is supported by the notion that the phenotype is reminiscent of mutants affecting CAMs in motoneurons such as FasII (Abrell, 2001).

The similarity of motoneuronal phenotypes of Kr and CAM mutants, and the interaction between Kr and caps activity suggests that Kr might also interact genetically with additional CAMs which are necessary for proper path finding. This proposal was tested using mutant combinations of Krres and the loss-of-function mutant FasIIeb112. FasII is a more general CAM than Caps. It is expressed on all motoneurons during late embryonic stages and is necessary to maintain adhesion between the axons (Abrell, 2001).

Heterozygous Krres/+ or FasIIeb112/+ embryos develop a normal SNb pattern and all RPs are properly connected to their target muscles. In double heterozygous FasIIeb112/+;Krres/+ mutants, however, the SNb enters the ventral muscle field normally in all cases, but the nerve stops at the second choice point by forming a growth cone-like structure. No individual RP axons could be observed. This phenotype is very similar to the one observed with homozygous Krres mutant embryos, implying that the two gene activities cooperate to allow for proper SNb development (Abrell, 2001).

In addition to position, size and morphology, the innervation of muscles by specific motoneurons represents a diagnostic feature for the determination of muscle identity. Previous results have shown that Kr is expressed in a specific subset of muscle progenitors and is necessary for the acquisition of a specific muscle fate as shown by muscle transformations that occur in response to gain-of-function and lack-of-function Kr mutations. Muscle 27 is transformed into a duplicated muscle 26 in homozygous Kr mutant embryos whereas high level overexpression (four copies of Kr) in all muscles leads to the reverse transformation. The data suggest that Kr contributes not only to identifying characteristics of muscle 27 but also provides adhesion properties to other Kr-expressing muscles along the SNb pathway, i.e. muscles 14 and 16, the ventral acute muscle 27, and the ventral longitudinal muscles 6, 7 and 13. The genetic interactions between Kr and the CAMs FasII and Caps support the hypothesis and imply that the adhesive properties of motoneurons and/or muscles are established in such a way that a concomitant reduction in adhesion of Krres/+ ; caps/+ or Krres/+; FasIIeb112/+ mutants results in a situation which no longer provides sufficient information to allow accurate axonal pathfinding and innervation. This proposal is also consistent with the finding that the presence of one muscle in an otherwise muscle-depleted embryo can be sufficient for the defasciculation of nerve bundles (Abrell, 2001).

In muscles 6, 7 and 13, Kr is known to maintain the expression of a direct target gene, knockout (ko). ko mutant embryos display a Kr-like motoneuron phenotype, suggesting that the gene, which encodes a novel protein with unknown biochemical characteristics, plays a key role in SNb defasciculation and RP pathfinding by acting downstream of Kr. However, in contrast to Kr, ectopic misexpression of ko does not affect SNb branching and synaptic targeting of RP neurons, and no genetic interaction as observed between ko and caps could be found. It is therefore likely that Kr transmits its signal not only via ko, but also through other factors that are still to be identified. It was also found that in contrast to caps, ectopic panmotoneural expression of Kr causes defects similar to the Kr lack-of-function mutation, and a reduction of combined caps or FasII and Kr activities. It is therefore speculated that Kr activity can also directly interfere with the spectrum of CAMs in motoneurons, resulting in non-compatible cell-surface characteristics between axons and muscles, which in turn interfere with neuromuscular connectivity (Abrell, 2001).

Short-range cell interactions and cell survival in the Drosophila wing

During development of multicellular organisms, cells are often eliminated by apoptosis if they fail to receive appropriate signals from their surroundings. Short-range cell interactions support cell survival in the Drosophila wing imaginal disc. Evidence is presented showing that cells incorrectly specified for their position undergo apoptosis because they fail to express specific proteins that are found on surrounding cells, including the LRR transmembrane proteins Capricious and Tartan. Interestingly, only the extracellular domains of Capricious and Tartan are required, suggesting that a bidirectional process of cell communication is involved in triggering apoptosis. Evidence showing that activation of the Notch signal transduction pathway is involved in triggering apoptosis of cells misspecified for their dorsal-ventral position (Milán, 2002).

In second instar wing discs, the LRR proteins Caps and Tartan are expressed in cells of the dorsal compartment. During third instar, dorsal expression of Caps and Tartan decreases and new lateral expression domains arise. The region of low Caps and Tartan expression in the center of the mature third instar wing disc coincides with the domain in which Dpp signaling induces Spalt expression. The reciprocity of Spalt and Caps/Tartan expression in third instar wing discs suggested that Spalt might repress Caps/Tartan at this stage. spalt mutant clones located medially show ectopic expression of Caps protein and a tartan-lacZ reporter gene. Ubiquitous expression of Spalt in the wing pouch reduces the levels of expression of Caps and Tartan in the lateral wing disc. These results indicate that Spalt restricts expression of caps and tartan to lateral cells in third instar wing discs (Milán, 2002).

To determine whether apoptosis might be a general response of cells unable to engage in normal interactions with their neighbors, the effects caused by producing cells with inappropriate dorsal-ventral compartment identity were examined. Clones of cells expressing Apterous (Ap) were produced to examine the survival of D cells in the V compartment. Fewer than 20% of surviving Ap-expressing clones were of V compartment origin. Half of these had sorted out into the D compartment and so were in contact with other Ap-expressing cells. The remaining ~10% of clones were recovered in the V compartment. Ventral Apterous-expressing clones were round in shape and induced Wg expression at their borders. Expression of the Apterous inhibitor dLMO was used to produce cells with V identity in the D compartment. Only 30% of dLMO-expressing clones were of D compartment origin. Most of these had sorted out into the V compartment. Fewer than 5% of dLMO-expressing clones were recovered in the D compartment. These were round in shape and induced Wg expression at their borders. These observations suggested that dLMO-expressing clones are preferentially lost from the D compartment if they are unable to make contact with V cells. Likewise, Ap-expressing clones are preferentially lost from the V compartment if they are unable to make contact with D cells. Loss of the inappropriately specified cells was suppressed by coexpression of p35. Under these conditions 48% of dLMO and p35-expressing clones were of D origin, and 51% of clones expressing Ap and p35 were of V origin. This indicates that inappropriately positioned cells are lost by apoptosis. Apoptosis of these cells occurs when clones were induced in second instar. Clones induced during third instar survive equally in both compartments. Caps and Tartan are expressed in D cells under Ap control in second instar wing discs. Ectopic expression of Caps or Tartan cause clones to sort out toward the D compartment, suggesting that these proteins may confer a preferential affinity for D compartment cells. To test whether loss of Caps or Tartan expression contributes to the poor survival of dorsal dLMO-expressing clones, the recovery was measured of clones coexpressing dLMO with Caps or with Tartan. When coexpressed with Caps, 58% of dLMO-expressing clones were of dorsal origin and were recovered in the D compartment, compared to 30% when dLMO was expressed alone. Coexpression with Tartan yielded 54% dorsal dLMO-expressing clones. Expression of CapsDeltaC and TrnDeltaC is able to support survival of dLMO-expressing clones in the D compartment almost as effectively as the full-length proteins (Milán, 2002).

Although Caps and Tartan are able to support the survival of dLMO-expressing clones in the D compartment, the reverse is not true. Caps and Tartan expression are induced by Apterous but obviously cannot support the survival of Ap-expressing clones in the V compartment. Moreover, ectopic expression of Caps or Tartan cause loss of clones by apoptosis in the V compartment. In these experiments ventral Caps- or Tartan-expressing clones that made contact with cells in the D compartment either by sorting out or by sending long cytoplasmic extensions are able to survive. The remaining clones are lost by apoptosis but can be rescued by coexpression of p35. Thus, Caps and Tartan do not appear to function as general survival factors. Their ability to support cell survival depends on the developmental context. Caps and Tartan can mediate cell interactions that prevent apoptosis of misspecified cells in areas where these proteins are expressed by the surrounding cells (Milán, 2002).

One of the most striking observations is that the extracellular domains of Caps and Tartan are sufficient to convey these cues in two developmental contexts. By analogy to the dual roles of EphB and EphrinB proteins as both ligands and receptors, it is suggested that Caps and Tartan proteins serve as ligands to identify cells to their neighbors, perhaps by engaging a cell surface receptor. Failure to receive this signal may cause neighboring cells to elicit a signal that triggers apoptosis of the mispositioned cell. In this context it is worth noting that Caps and Tartan can perform this function in situations where the surrounding cells also express these proteins but not where the surrounding cells don't express them, for example, in the early V compartment. Thus, their ability to mediate cell interactions that support the survival of misspecified cells is context specific. They do not appear to act as general survival factors (Milán, 2002).

Cells in the wing disc often die in small groups, raising the possibility that death signals may not be targeted precisely at the defective cell. Although the nature of the proposed death signal is not known, the results have implicated activation of the Notch signaling pathway in elimination of cells mispositioned with respect to DV identity. Blocking Notch activation in these cells using the dominant-negative NotchECD receptor or using a dominant-negative form of the Notch effector Mastermind is sufficient to prevent removal of these cells by apoptosis. This indicates that loss of cells is due to activation of the conventional Notch signaling pathway. There is a similar requirement for Notch activation in programmed cell death in the eye imaginal disc. It is clear that Notch signaling is not dedicated to elimination of cells. On the contrary, wing disc cells unable to transduce the Notch signal are lost. Thus, it is evident that Notch signaling is used to cause apoptosis in a specific context, in conjunction with other signals. Cells may die when they receive a combination of signals that indicate incorrect position. Dorsal cells expressing dLMO lack Caps and Tartan, which mediate dorsal cell interactions, as well as Serrate and Fringe, which influence Notch signaling. Restoring either category of cell interaction is sufficient to suppress apoptosis of these cells (Milán, 2002).

In the second larval instar, cells in the wing disc assess their DV position. Cells that are misspecified with respect to DV compartment identity tend to sort out into the appropriate compartment. dLMO-expressing cells sort-out into the V compartment. Ap-expressing cells sort-out into the D compartment. If mispositioned cells are able to contact similarly specified cells, they can survive. Although Caps and Tartan are able to trigger cell interactions that provide dLMO-expressing cells with the information that they need to survive in the D compartment, Caps and Tartan are not essential for survival of dorsal cells. Clones simultaneously mutant for both genes survive equally well in D and V compartments. Caps and Tartan are also used to generate a difference in medial-lateral cell affinity during the third instar. Lateral Spalt-expressing cells are eliminated by apoptosis. Caps or Tartan can suppress loss of these cells, but they are not required for survival of lateral cells. These observations indicate that there must be additional cell surface proteins that are capable of mediating the cell interactions that are needed in the dorsal compartment and in the lateral region of the wing disc to support cell survival. In this context it is worth noting that medial spalt mutant cells are not eliminated by apoptosis. Survival of medial cells may depend on the activity of a second Dpp target gene, optomoter blind (omb). Cells lacking omb in the center of the disc are lost, and large mutant clones produce extensive loss of wing tissue. This type of strong wing-scalloping phenotype is typically associated with massive cell death and activation of the JNK pathway (Milán, 2002).

The findings indicate that apoptosis of misspecified cells is associated with loss of expression of specific cell surface proteins. This correlates with alterations in cell affinity. However, alteration in cell affinity per se does not appear to be sufficient to drive cells into apoptosis. Removal of Caps and Tartan laterally caused affinity differences without compromising clone survival. Likewise, spalt mutant clones caused ectopic expression of Caps and Tartan and became rounded due to differences in cell affinity in the medial part of the wing disc. Caps- and Tartan-expressing clones survived normally and remained well integrated in the disc epithelium, despite these differences in affinity. Loss of spalt mutant clones during pupal stages is due to sorting out of the clones from the epithelial sheet to form vesicles of mutant tissue, and not due to apoptosis. Taken together, these observations suggest that alterations in cell affinity are not the cause of apoptosis of mispositioned Spalt-expressing cells. Apoptosis of these clones appears to be due to the absence of specific cell surface cell interactions that can be mediated by Caps and Tartan proteins (Milán, 2002).

Boundary formation in the Drosophila wing: functional dissection of Capricious and Tartan

Cells in multicellular organisms often do not intermingle freely with each other. Differential cell affinities contribute to organizing cells into different tissues. Drosophila limbs and rhombomeres of the vertebrate hindbrain are subdivided into compartments. Cells in adjacent compartments do not mix. The wing primordium is subdivided into dorsal (D) and ventral (V) compartments by the activity of the LIM-homeodomain protein Apterous in D cells. The leucine-rich repeats transmembrane proteins Capricious and Tartan have been shown to contribute to formation of the affinity boundary between dorsal and ventral compartments. This paper describes a structure-function analysis of Capricious and Tartan. Evidence is presented that both the extracellular and intracellular domains are required for the establishment of a DV affinity boundary. The data suggest that the extracellular domains of Capricious and Tartan may work as ligands of an unknown D cell surface protein. Their intracellular domains may be required to transduce a signal necessary for the establishment of the DV boundary (Milán, 2005).

Although Caps and Tartan confer a D-specific affinity, expression of Caps or Tartan in S2 cells is not able to promote cell aggregation and soluble forms of the extracellular domains of Caps and Tartan are not able to bind to full-length forms of Caps and Tartan expressed in S2 cells. In a sorting assay, membrane-bound cellular processes extending from ventral Caps- or Tartan-expressing cells toward cells in the D compartment were observed. Similar cellular extensions were not found between closely spaced clones within the V compartment. These observations support the idea that Caps and Tartan may be interacting with another cell surface protein in the D compartment. To further test this, the ability of Caps and Tartan to rescue the DV affinity boundary in the complete absence of Apterous activity was tested. Caps and Tartan were able to partially rescue the DV affinity boundary, suggesting that Caps and Tartan may confer, independently of any other protein expressed in D cells, an affinity difference. However, the DV boundary was not always smooth, and cells tended to violate it, suggesting that Caps and Tartan may require the binding to another cell surface protein in D cells to generate an affinity boundary. Coexpression of Caps and Tartan gave the same results. Thus, in the strong but not complete loss-of-function background for Apterous activity, the unknown dorsal partner of Caps and Tartan may be expressed at low but sufficient levels to interact with Caps and Tartan and generate a DV affinity boundary. In the complete absence of Apterous activity, their partner may not be expressed at all (Milán, 2005).

The extracellular domain of Caps is required for this interaction to take place. A truncated form of Caps lacking the extracellular domain (Caps-ED) is not able to make V cells sort toward the DV boundary, and its expression in D cells in the partial absence of Apterous activity is not able to rescue the DV affinity boundary. Its subcellular distribution and relative amount of overexpression was comparable to the full-length forms of Caps (Milán, 2005).

It was asked if the intracellular domains of Caps and Tartan are required to induce segregation of D and V cell populations. For this purpose, truncated forms of Caps and Tartan lacking the intracellular domain were tested in the sorting and rescue assays. Of interest, they make V cells sort toward the DV boundary, the same way as full-length forms of Caps and Tartan, suggesting that the lack of intracellular domains is not compromising Caps/Tartan activity in this assay. Clones induced in early second instar discs, before the DV boundary forms, were bisected by the nascent DV boundary so that they contributed to both compartments. Control D+V clones had no effect on the shape of the Wg stripe and were of similar size in both compartments. In contrast, D+V clones expressing truncated forms of Caps or Tartan tended to be considerably smaller in the V compartment, the same way as full-length forms of Caps and Tartan. Many of these clones distorted the Wg stripe where they crossed the boundary. These results suggest the extracellular domains of Caps and Tartan are not only required but also are sufficient to make V cells sort toward the DV boundary probably through interaction to an unknown D partner. Expression in D cells of truncated forms of Caps and Tartan lacking the intracellular domain in the partial absence of Apterous activity is able to partially rescue the DV affinity boundary. However, the DV boundary is not always smooth, and cells violate it. Are the rescue and sorting assays qualitatively or just quantitatively different tests of function? That the truncated forms behave quantitatively as full-length forms in the sorting assay and the subcellular distribution of both types of proteins were comparable suggests the inability of the truncated forms to rescue the DV affinity boundary and reflects a qualitative requirement of the intracellular domain in this process (Milán, 2005).

Fusion proteins consisting of the extracellular and transmembrane domains of Caps and Tartan and the β-galactosidase protein, a protein well-known to oligomerize, in the intracellular domain were not able to completely rescue the DV affinity boundary in the partial absence of Apterous activity, suggesting the intracellular domains of Caps and Tartan are not require for oligomerization. Their capability to partially rescue the DV affinity boundary was comparable to Caps-ID and Trn-ID, suggesting the β-galactosidase protein is not interfering with the activity of Caps and Tartan. Because β-galactosidase protein makes a tetramer, the possibility cannot be ruled out that Caps and Tartan may need to be arranged in a higher order multimer to be functional. Even though the intracellular domains of Caps and Tartan are quite different, they are completely interchangeable. Protein swaps consisting of the extracellular and transmembrane domains of Caps or Tartan and the intracellular domains of Tartan or Caps, respectively, were able to completely rescue the DV affinity boundary in the rescue assay (Milán, 2005).

With all these results, it is proposed that Caps and Tartan may work both as ligands and receptors to generate a stable DV affinity boundary in the wing primordium. In the sorting assay, Caps and Tartan may work as ligands of an unknown D cell surface protein. The extracellular domains of Caps and Tartan are necessary and sufficient to make V cells sort toward the DV boundary, probably being attracted by a D receptor. The extracellular domains of Caps and Tartan contain 14 LRRs and are 67% identical in terms of amino acid sequence, suggesting that they may bind to a common protein. In the rescue assay, both the extracellular and intracellular domains of Caps and Tartan are required. The extracellular domains of Caps and Tartan partially rescue the DV affinity boundary, suggesting Caps and Tartan are behaving as ligands in this assay. In the presence of the intracellular domains, the rescue is complete. Their intracellular domains would then either transduce a signal or interact with the cytoskeleton, thus generating a strong DV affinity difference (Milán, 2005).

Sorting-out of cell populations can be guided by both the amount and types of adhesion proteins that cells express. A different view comes from studies on Ephrins and Eph receptors, which indicate that a signaling cascade downstream of both ligands and receptors is required to induce repulsion or de-adhesion, thus promoting segregation of cell populations. Many adhesion proteins form regulated connections with the cytoskeleton and participate in contact-mediated signaling. The data suggest that the establishment of the DV affinity boundary in the Drosophila wing relies not only on the types of cell adhesion proteins the cells express but also on cell communication downstream of these proteins (Milán, 2005).

Regulation of layer-specific targeting by reciprocal expression of a cell adhesion molecule, Capricious

Layer-specific innervation is a major form of synaptic targeting in the central nervous system. In the Drosophila visual system, photoreceptors R7 and R8 connect to targets in distinct layers of the medulla, a ganglion of the optic lobe. Capricious (CAPS), a transmembrane protein with leucine-rich repeats (LRRs), is a layer-specific cell adhesion molecule that regulates photoreceptor targeting in the medulla. During the period of photoreceptor targeting, caps is specifically expressed in R8 and its target layer but not in R7 or its recipient layer. caps loss-of-function mutations cause local targeting errors by R8 axons, including layer change. Conversely, ectopic expression of caps in R7 redirects R7 axons to terminate in the CAPS-positive R8 recipient layer. CAPS promotes homophilic cell adhesion in transfected S2 cells. These results suggest that CAPS regulates layer-specific targeting by mediating specific axon-target interaction (Shinza-Kameda, 2006).

Different classes of R-cell axons are targeted to distinct layers of the optic lobe. The present study demonstrates that CAPS is expressed in R8 and in the medulla layer M3 where R8 axons terminate. In caps mutant animals, R8 does not precisely target to this region in the medulla. Ectopic expression of CAPS in R7 redirects R7 terminals to the CAPS-positive M3 layer. In vitro, CAPS mediates homophilic cell adhesion. These observations are consistent with the idea that CAPS mediates homophilic interaction between pre- and post-synaptic cells during the formation of layer-specific neural connections (Shinza-Kameda, 2005).

Its expression in specific synaptic partners provided the first clue to the function of CAPS in the visual system. Previous studies showed that during the development of the neuromuscular system, CAPS expression is closely correlated with synaptic connectivity: caps-positive motoneurons innervate caps-positive muscles, whereas CAPS-negative motoneurons connect with caps-negative muscles. Genetic analysis suggested that caps functions as an attractive target cue that guides motoneurons to the correct muscles. Thus, in the two Drosophila systems in which neural connectivity has been well characterized, caps is expressed in both partners of specific synaptic pairings and appears to mediate the interaction between innervating axons and their targets. This idea is further supported by the in vitro cell aggregation assay, which showed that CAPS binds homophilically. However, it should be noted that caps may regulate target specificity through signal transduction rather than by simple adhesion. During the formation of neuromuscular connectivity, motor axons choose among individual muscle cells present in the target region. In contrast, during R-cell targeting in the medulla, afferent axons establish specific synaptic connections by selecting the correct target layer. CAPS thus can mediate cell-to-cell specificity in two different types of targeting system (Shinza-Kameda, 2006).

Analysis of the effects of caps loss-of-function and misexpression in R cells supports the notion that CAPS functions as an attractive target recognition molecule in the visual system. In caps mutants, R8 axons reach the medulla and are initially guided to the correct layer. However, R8 axons then fail to stabilize in the target region and instead extend abnormally to nearby nontarget regions. The onset of the phenotype corresponds to the developmental stage when CAPS expression in specific layers becomes eminent in the medulla. These observations suggest that caps is required for the stabilization of the contact between R8 and the medulla layer. The same phenotype was observed in the mosaic animals, in which caps mutant R cells innervated largely wild-type medulla and in MARCM mutant clones. Thus, CAPS is required in the R cells where it is specifically expressed in R8 for proper targeting. Because CAPS is also expressed in the R8 recipient layer in the medulla, it is likely that CAPS also functions in the medulla to mediate afferent-target interaction. However, because of technical limitations, it was not possible to make large caps mutant clones that include all or most of the medulla target neurons. Thus, it remains to be determined if caps is also required in the medulla for correct targeting of R8 cells (Shinza-Kameda, 2005).

Misexpression experiments further support the notion that CAPS mediates specific interaction between R cells and medulla layers. CAPS expression on R7 axons is sufficient to redirect them to terminate in a CAPS-positive M3 layer. Although previous loss-of-function analyses have implicated several other cell surface receptor and adhesion molecules in layer-specific targeting in this system, this is the first molecule whose misexpression was shown to switch the target specificity. Two lines of evidence suggest that the misexpression phenotype is caused by the alteration in axon-target recognition rather than that in axon-axon interaction, such as defasciculation defects. (1) In GMR-GAL4/UAS-caps animals, R7 axons initially do defasciculate from R8 to reach their target region before retracting back to the R8 recipient layer. (2) CAPS misexpression only in R7 is sufficient to induce the mistargeting phenotype. Taken together, loss-of-function and gain-of-function analyses support the notion that CAPS mediates interaction between specific R-cell growth cones and specific layers in the medulla neuropil, both of which express the molecule (Shinza-Kameda, 2006).

Layer-specific targeting is widely observed in vertebrate and invertebrate brain and appears to provide a major means to segregate synaptic connections between specific pre- and post-synaptic cells. Although several axon guidance molecules have been shown to be expressed in particular layers, their roles in layer-specific targeting in vivo remain largely unknown. The best-characterized cues expressed on specific layers are Sidekicks, cell adhesion molecules belonging to the immunoglobulin superfamily, which are expressed both pre- and post-synaptically in specific synaptic partners that project to the same layers in the retina (Yamagata, 2002). When misexpressed on the presynaptic cells, Sidekicks can change their target-layer specificity. The loss-of-function and gain-of-function analyses presented in this study show that CAPS functions in a similar manner to promote layer-specific connectivity in the Drosophila visual system. These examples provide evidence for the role of local cues in the formation of layer-specific connectivity. In contrast, recent studies using the Drosophila visual system have revealed the roles of several receptors and cell adhesion molecules, such as N-cadherin, LAR, and Flamingo, in target layer selection. Although their expression is not layer specific, they contribute to specific aspects of R-cell targeting. Interestingly, the gain-of-function phenotype of CAPS is similar to the of loss of function for N-cadherin or DLAR, suggesting that they may participate in the same molecular mechanism that determines the target specificity in this system. Understanding how such general and cell-specific molecules cooperate to establish layer-specific connectivity will be an important subject in future studies (Shinza-Kameda, 2006).

Distinct functions of the leucine-rich repeat transmembrane proteins Capricious and Tartan in the Drosophila tracheal morphogenesis

A key step in organogenesis of the Drosophila tracheal system is the integration of isolated tracheal metameres into a connected tubular network. The interaction of tracheal cells with surrounding mesodermal cells is crucial in this process. In particular, single mesodermal cells called bridge-cells are essential for the guided outgrowth of dorsal trunk branches to direct formation of the main airway, the dorsal trunk (Wolf, 2000). This study presents evidence that the two leucine-rich repeat transmembrane proteins Capricious and Tartan contribute differently to the formation of branch interconnections during tracheal development. Capricious is specifically localized on the surface of bridge-cells and facilitates the outgrowing dorsal trunk cells of adjacent metameres toward each other. Capricious requires both extracellular and intracellular domains during tracheal branch outgrowth. In contrast, Tartan is expressed broadly in mesodermal cells and exerts its role in tracheal branch outgrowth through its extracellular domain. It is proposed that Capricious contributes to the instructive role of bridge-cells whereas Tartan provides permissive substrate for the migrating tracheal cells during the network formation (Krause, 2006).

The initial outgrowth of tracheal cells from ectodermal clusters is triggered by branchless (bnl), a gene encoding a Drosophila FGF homolog (dFGF/Bnl). Bnl is expressed dynamically in small groups of cells surrounding the tracheal primordia and acts as a chemoattractant that guides outgrowth of primary tracheal branches. In addition to the Bnl-signaling, a cellular guidance mechanism is essential for normal outgrowth of dorsal trunk branches. This guidance requires single mesodermal cells called bridge-cells, which are positioned posteriorly next to the tracheal cell clusters and serve as guidance posts for the outgrowing dorsal trunk branches. Initially, filopodia-like extensions from anterior and posterior dorsal trunk cells get in touch with bridge-cells, slide along the bridge-cell surface and contact the opposite extensions. The dorsal trunk fusion process then starts with deposition of the epidermal adhesion protein DEcadherin (DEcad) at the contact point between two fusion cells. Localized at the adherens junctions, DEcad interacts with a and β-catenin (Armadillo) that subsequently bind the actin cytoskeleton. A cytoskeleton-associated plakin Short Stop (Shot), which interacts with both actin and microtubules, is required for DEcad accumulation at the fusion site. Finally, each of fusion cells forms an intracellular tube and the two lumens fuse and become continuous (Krause, 2006).

The molecular mechanisms underlying the initial cell-to-cell contacts between the extending dorsal trunk cells and the guiding bridge-cells are not yet known. The transcription factor Hunchback (Hb) was shown to play a key role in bridge-cell differentiation, which in turn is necessary for dorsal trunk fusion. In addition, the transcription factors Extradenticle (Exd) and Homothorax (Hth) are also expressed in the bridge-cells. However, neither of these transcription factors nor Bnl may be directly involved in the recognition of bridge-cells by tracheal extensions. Thus, it has been speculated that this process is mediated by extracellular matrix and adhesion molecules that are expressed on the surfaces of tracheal and bridge cells (Krause, 2006).

Caps is specifically expressed in the bridge-cells. When caps is lacking, the bridge-cells may not bind the dorsal trunk cells effectively and thus fail to provide sufficient guidance. Consequently, discontinuous dorsal trunks are formed. Since only dorsal trunk cells require bridge-cell's guidance, other tracheal branches are not affected by lack of caps activity (Krause, 2006).

What is the function of Caps in the bridge-cells? Caps might be involved in directly mediating communication between the bridge-cells and the tracheal cells, since it represents a transmembrane protein with 14 leucine-rich repeats (LRR) in the extracellular domain. LRRs are arranged consecutively and parallel to a common axis so that the conformation resembles a horseshoe, which is ideal for mediating cell-to-cell interactions. Possibly, the extracellular domain of Caps assumes the horseshoe-like conformation as well and binds molecules localized on the surface of dorsal trunk cells. Thus, Caps may provide an important link between the guiding bridge-cells and the outgrowing tracheal cells (Krause, 2006).

When caps is ectopically expressed in mesodermal cells, discontinuous dorsal as well as lateral trunks are formed. Moreover, some dorsal trunk cells grow dorsally or ventrally rather than anteriorly towards their adjacent targets. The outgrowing tracheal cells may adhere to ectopic Caps on the surface of mesodermal cells, which normally do not express Caps, rather than binding endogenous Caps on the bridge-cells. As a result, these tracheal cells may become disoriented and extend in abnormal directions. This ability of ectopic Caps to stall normal development of dorsal trunk branches suggests that selective expression of Caps in the bridge-cells is important for the specific local guidance for the outgrowing dorsal trunk cells. However, more severe dorsal trunk defects were observed in hb mutant embryos, which lack the bridge-cells (Wolf, 2000). This observation suggests that additional components besides Caps are involved in the bridge-cell's guidance. Such components may include the chemoattractant Bnl, which is also expressed in the bridge-cells (Krause, 2006).

Previous results demonstrated that Caps plays an essential role during pathfinding of motorneurons, layer-specific targeting in the visual system and boundary formation in wing imaginal discs. Normally, Caps is expressed in a subset of CNS neurons including aCC, RP2, RP5 and U motorneurons. When Caps is overexpressed in all neurons, the axons of muscle 12 motorneurons (MNs) become misrouted. Similar results are obtained when CapsEd lacking the intracellular domain is misexpressed in neurons. However, when CapsId lacking the extracellular domain is misexpressed, no defects can be observed in pathfinding of muscle 12 MNs. These results clearly indicate that a neural expression of CapsEd is sufficient to misroute muscle 12 MNs. During boundary formation of the wing discs, cells incorrectly specified for their position undergo apoptosis because they fail to express Caps. Again, exclusive expression of CapsEd is sufficient to prevent apoptosis of misspecified cells. During these developmental events, Caps is thought to function as a cell adhesion molecule providing specific affinity between different cells and thus, may require only its membrane-anchored extracellular domain. In contrast, muscularly expressed Caps requires extracellular and intracellular domain to establish aberrant synapses of muscle 12 MNs. Similarly, only embryos misexpressing complete Caps reveal interruptions in dorsal and lateral trunks. These results suggest that Caps requires its intracellular domain both for its function during the establishment of motorneuron synapses and during tracheal morphogenesis. This hypothesis was further strengthened by investigations using Caps and Trn hybrid proteins in ectopic expression assays. Misexpression of TrnEdCapsId in mesodermal cells results in discontinuous tracheal branches. Moreover, misexpression in trn mutants aggravates the tracheal defects. These observations suggest that the TrnEdCapsId hybrid protein functions similar to Caps even though it contains only the intracellular domain of Caps. Thus, it is the intracellular domain that determines specificity of Caps function during tracheal morphogenesis (Krause, 2006).

It has been postulated that Caps might interact with receptors on specific motorneurons via its extracellular domain and transmit the signal into muscles via its intracellular domain. Likewise, Caps may function as a signal transmitter between the tracheal cells and the bridge-cells. Interestingly, proteins containing LRRs are predicted to undergo conformational changes upon binding the ligand or other proteins. These changes do not involve the usual movement of separate domains relative to each other, but rather an elastic alteration of the entire structure. Such conformational changes may subsequently induce Caps to interact with other intracellular proteins involved in signaling pathways or regulation of cytoskeletal structures. It appears that depending on the cell type, Caps may act as a cell adhesion molecule or as a receptor that relays signals from the outside to the inside of cells (Krause, 2006).

Sequence alignment of Caps from Drosophila and Anopheles reveals evolutionary conservation of three putative motifs within the intracellular domain: a putative tyrosine phosphorylation site, a predicted PDZ binding motif and a conserved RHR motif. Site-directed mutagenesis of these putative functional motifs and in vivo analysis shows that only the RHR motif is essential for Caps function during tracheal formation. This RHR motif is not yet recognized as a functional motif by protein databases. Therefore, no information regarding other proteins containing such a motif or prediction about its putative function is known. Located immediately after the transmembrane domain, the RHR motif may contribute to conformational changes, which enable Caps to transmit extracellular signals to small membrane-associated proteins that bind components of signaling pathways or cytoskeleton (Krause, 2006).

During the Drosophila wing development, Caps and Trn share redundant function and contribute evenly to the formation of affinity boundary between the dorsal and ventral compartments. This result is rather expected since only the extracellular domain of Caps or Trn is required during the establishment of boundary in wing discs and the extracellular domains of Caps and Trn are 65% identical. Similarly, Trn requires only its extracellular domain to mediate its function during the establishment of tracheal network. The current results demonstrate that the extracellular domain of either Trn or Caps can rescue the trn tracheal phenotype. Thus, the extracellular domains of both Caps and Trn provide the mesodermal cells with a substrate that mediates a normal tracheal branch progression in trn mutants (Krause, 2006).

In contrast, Trn and Caps are expressed in divergent patterns and contribute differently to the formation of continuous tracheal branches. Whereas caps is expressed selectively in the bridge-cells, trn is detected in broad subsets of mesodermal cells excluding the bridge-cells. Furthermore, ectopic expression of Caps in mesodermal cells disrupts formation of normal tracheal interconnections while ectopic expression of Trn in mesodermal cells does not affect tracheal development. Rather, it rescues the tracheal defects of trn mutant embryos. Finally, although Caps requires both its extracellular and intracellular domain for proper function during the formation of tracheal branches, Trn needs only its extracellular domain (Krause, 2006).

Based on these observations, the following model is proposed for Caps and Trn functions during tracheal development: Caps is important for the bridge-cell, which provides instructive cues for the extending dorsal trunk cells, while Trn contributes to permissive matrix function of mesodermal cells for normal tracheal branch outgrowth. Localized at the surface of bridge-cells, Caps may bind to cell surface molecules on tracheal cells and allow the tracheal cells to extend along the bridge-cells so that they can find the correct targets. In the absence of caps, the bridge-cells cannot mediate their local guidance as effectively and consequently, the dorsal trunk cells fail to interconnect to their targets. In contrast, when caps is expressed in additional mesodermal cells besides the bridge-cells, then the nearby tracheal cells can also adhere to these mesodermal cells through Caps and extend in unspecified directions. Thus, disconnected tracheal branches are formed. Previous studies indicate that specific interactions between cell surface proteins on the tracheal cells and the surrounding mesodermal cells are crucial for migration of tracheal cells. The mesodermal cells may serve as a matrix enabling (and facilitating) the tracheal cells to recognize the correct path and to migrate efficiently. These cell-to-cell interactions might involve transmembrane proteins such as Trn. Localized at the surface of mesodermal cells, Trn may interact directly with other molecules on the tracheal cells and thereby support the tracheal cells to extend across the mesodermal sheet. When trn is absent, the navigation of outgrowing tracheal cells is partially hampered and they cannot migrate along their paths as efficiently. Consequently, they fail to connect to their targets. Overexpression of Trn in mesodermal cells mimics the wild-type expression of Trn and does not affect the progression of tracheal cells (Krause, 2006).

These results suggest that migrating tracheal branches require an adhesive substrate provided by the broad expression of Trn in the immediate proximity. The major airway, the dorsal trunk, relies additionally on instructive guidance by the extending bridge-cells expressing Caps, which may bind to surface molecule(s) on the progressing tracheal cells. The intracellular domain of Caps might induce signaling leading to cytoskeletal changes that generate a 'pulling' force of the bridge-cells on the migrating tracheal cells. Identification of extracellular binding partners on tracheal cells and intracellular interaction partners of Caps may elucidate molecular mechanisms underlying transmission of external cues in the bridge-cells that induce intracellular events leading to cellular guidance (Krause, 2006).

Target recognition at the tips of postsynaptic filopodia: accumulation and function of Capricious

While much evidence suggests that postsynaptic dynamism contributes to the formation of synapses, few studies have addressed its possible role in target selection. Do postsynaptic motile structures seek specific synaptic partner cells, as does the presynaptic growth cone? This study examined the dynamics of myopodia, postsynaptic filopodia in Drosophila muscles, and the role of Capricious (CAPS) during the process of synaptic matchmaking. CAPS is a target recognition molecule with an extracellular domain containing leucine-rich repeat sequences. It is expressed in specific subsets of embryonic/larval body wall muscles, including muscle 12 (M12). Evidence is provided that implicates the tips of myopodia as loci of initial neuromuscular recognition: (1) CAPS, expressed as a GFP-fusion protein in M12, accumulated at the tips of myopodia; and (2) simultaneous live imaging of presynaptic motoneurons and postsynaptic myopodia revealed that initial neuromuscular contacts occur at the tips of myopodia. The live imaging also showed that individual postsynaptic myopodia appear to be able to discriminate partner and non-partner presynaptic cells: whereas many myopodial contacts with the partner motoneurons are stabilized to form synapses, those with non-partner neurons are retracted. In caps mutants, or in double mutants lacking both CAPS and the closely related protein Tartan, fewer contacts were observed between myopodia of M12 and the presynaptic growth cones during the process of initial neuromuscular interaction. The nascent synaptic sites of M12 were also reduced. These results provide evidence for the sensing function of postsynaptic filopodia, and implicate Caps-mediated recognition at the tips of myopodia in synaptic matching (Kohsaka, 2009).

Filopodia are thin cellular protrusions that are thought to function as sensors of the local environment. Sensory function of filopodia has been best illustrated in neuronal growth cones. For example, a single filopodium of a neuron in culture is able to steer the entire growth cone when it contacts a more adhesive substrate. Growth cones of grasshopper pioneer axons in vivo change their direction upon the contact of a single filopodium with guidepost cells. Filopodia are also found both pre- and postsynaptically during the process of synaptogenesis and are believed to actively participate in cellular interactions. Although there is evidence for the sensory function of presynaptic filopodia, few studies have addressed the question of whether postsynaptic filopodia are able to select specific partners. This study took advantage of the fact that highly specific and stereotyped synaptogenesis occurs during a relatively short period of time at the neuromuscular junction of Drosophila. Several lines of evidence are presented that support the active role of myopodia, postsynaptic filopodia of muscles, in synaptic partner recognition (Kohsaka, 2009).

A striking concentration of CAPS was found at the tips of myopodia. If myopodia have a passive role and merely help increase the target area, then it would seem to be more efficient for CAPS to be distributed uniformly along the length of myopodia. There would be more chance for the target marker to be presented to the presynaptic cells. Concentration of CAPS at the tips instead supports the idea that myopodia play an active role in directing growth cones to their targets. Filopodia and other cellular extensions have been observed in a variety of cells that send and/or receive signals, and implicated in contact-mediated long-distance communication. By presenting molecules such as CAPS at the tips, myopodia might be able to efficiently send signals to presynaptic motoneurons. At the same time, myopodia might receive signals from the presynaptic cells by using molecules such as CAPS as sensors. Such bidirectional signaling occurring at the tips of myopodia might be a trigger for synaptic matchmaking. As CAPS is also expressed in MN12s and can function as a homophilic cell adhesion molecule in vitro, the signaling might be mediated by homophilic interaction between CAPS on pre- and postsynaptic cells (Kohsaka, 2009).

It has been postulated that the tips of filopodia are the signaling center that regulate filopodial extension, retraction and adhesion from the following reasons. (1) It is often the distal part of the filopodium that makes contact with the environmental cues. (2) Actin monomers are added to the barbed ends of filaments at the tips of filopodia. (3) Several signaling and adhesion molecules, including Mena, integrins and the tyrosine-phosphorylated proteins, are concentrated at the tips of filopodia. Therefore, filopodia tips are well situated to link information from the environment to the dynamics of filopodia. Concentration of CAPS suggests a possibility that signaling events at the tips of myopodia are crucial in selective synapse formation. Consistent with this idea, this study found, by live imaging of neuromuscular interaction in vivo, that many of the initial contacts between motoneuronal growth cones and muscles do occur at the tips of myopodia. By tracing the dynamics and final fate of individual myopodia, it was also found that some of these contacts are stabilized to form the synaptic site, whereas others are eliminated. These results suggest that protrusive activity of myopodia actively contributes to neuromuscular interaction. Furthermore, evidence is provided that the behavior of myopodia differs depending on whether or not contact was made with a partner motoneuron. These results suggest that myopodia search for appropriate synaptic partners. Taken together with the localization of CAPS, these observations provide strong evidence for the sensory function of myopodia, and also highlight the role of myopodial tips as a possible signaling center for synaptic matchmaking. As the dynamic behavior of the postsynaptic cell is just as important as that of the presynaptic cell, future studies of axon guidance and target recognition during the formation of synapses should focus on the postsynaptic cell as much as the presynaptic cell (Kohsaka, 2009).

What kind of signaling is taking place at the tips of myopodia? How is the signal transmitted to other regions of muscles to affect the process of synaptogenesis? Recognition by cell adhesion molecules such as CAPS might allow for stabilization of specific myopodial contacts. Signaling at the tips might be transmitted to neighboring regions of the muscle to induce myopodia clustering and postsynaptic differentiation. These data indicate that signaling in muscles through the cytoplasmic domain of CAPS is likely to be required for target recognition, as CAPS-ID, which lacks the cytoplasmic domain, acts as a dominant negative when expressed in muscle (Kohsaka, 2009).

The reduction in the number of contacts between myopodia and growth cones and size of the nascent synaptic sites observed in caps and caps;tartan double mutants are consistent with this model. Myopodia-filopodia interaction appears also to be important for the differentiation of the presynaptic terminals. It was observed that presynaptic terminals form while interacting with the surrounding myopodia. Reduction in the size of nerve terminals observed in caps and caps;tartan double mutants also suggests that stabilization of contacts between growth cones and myopodia is crucial for presynaptic differentiation. As myopodia-like structures and their clustering have been reported in vertebrates, signaling events that regulate myopodia-growth-cone interaction might be a common mechanism for neuromuscular synaptogenesis. Future studies on the molecular events occurring at the tips of myopodia might shed light on the very beginning of synaptogenesis (Kohsaka, 2009).

Functions of Capricious orthologs in other species

Deletion of LRRTM1 and LRRTM2 in adult mice impairs basal AMPA receptor transmission and LTP in hippocampal CA1 pyramidal neurons
Leucine-rich repeat transmembrane (LRRTM; see Drosophila Toll-6 & Toll-7 and Capricious) proteins are synaptic cell adhesion molecules that influence synapse formation and function. This study took advantage of the generation of a LRRTM1 and LRRTM2 double conditional knockout mouse (LRRTM1,2 cKO) to examine the role of LRRTM1,2 at mature excitatory synapses in hippocampal CA1 pyramidal neurons. Genetic deletion of LRRTM1,2 in vivo in CA1 neurons dramatically impaired long-term potentiation (LTP), an impairment that was rescued by simultaneous expression of LRRTM2, but not LRRTM4. Mutation or deletion of the intracellular tail of LRRTM2 did not affect its ability to rescue LTP, while point mutations designed to impair its binding to presynaptic neurexins (see Drosophila Neurexin) prevented rescue of LTP. These proteins at mature synapses also caused a decrease in AMPA (see Drosophila GluR1A) receptor-mediated, but not NMDA receptor-mediated, synaptic transmission and had no detectable effect on presynaptic function. Imaging of recombinant photoactivatable AMPA receptor subunit GluA1 in the dendritic spines of cultured neurons revealed that it was less stable in the absence of LRRTM1,2. These results illustrate the advantages of conditional genetic deletion experiments for elucidating the function of endogenous synaptic proteins and suggest that LRRTM1,2 proteins help stabilize synaptic AMPA receptors at mature spines during basal synaptic transmission and LTP (Bhouri, 2018).

LRRTMs Organize Synapses through Differential Engagement of Neurexin and PTPsigma

Presynaptic neurexins (Nrxs) and type IIa receptor-type protein tyrosine phosphatases (RPTPs) organize synapses through a network of postsynaptic ligands. This study shows that leucine-rich-repeat transmembrane neuronal proteins (LRRTMs; see Drosophila Capricious) differentially engage the protein domains of Nrx (see Drosophila (see Drosophila Neurexin)) but require its heparan sulfate (HS) modification to induce presynaptic differentiation. Binding to the HS of Nrx is sufficient for LRRTM3 and LRRTM4 to induce synaptogenesis. Mammalian Nrx1gamma was identified as a potent synapse organizer, and LRRTM4 is revealed as its postsynaptic ligand. Mice expressing a mutant form of LRRTM4 that cannot bind to HS show structural and functional deficits at dentate gyrus excitatory synapses. Through the HS of Nrx, LRRTMs also recruit PTPsigma to induce presynaptic differentiation but function to varying degrees in its absence. PTPsigma forms a robust complex with Nrx, revealing an unexpected interaction between the two presynaptic hubs. These findings underscore the complex interplay of synapse organizers in specifying the molecular logic of a neural circuit (Roppongi, 2020).


Search PubMed for articles about Drosophila capricious

Abrell, S. and Jäckle, H. (2001). Axon guidance of Drosophila SNb motoneurons depends on the cooperative action of muscular Krüppel and neuronal capricious activities. Mech. Dev. 109: 3-12. 11677048

Berger-Muller, S., Sugie, A., Takahashi, F., Tavosanis, G., Hakeda-Suzuki, S. and Suzuki, T. (2013). Assessing the role of cell-surface molecules in central synaptogenesis in the Drosophila visual system. PLoS One 8: e83732. PubMed ID: 24386266

Bhouri, M., Morishita, W., Temkin, P., Goswami, D., Kawabe, H., Brose, N., Sudhof, T. C., Craig, A. M., Siddiqui, T. J. and Malenka, R. (2018). Deletion of LRRTM1 and LRRTM2 in adult mice impairs basal AMPA receptor transmission and LTP in hippocampal CA1 pyramidal neurons. Proc Natl Acad Sci U S A 115(23): E5382-E5389. PubMed ID: 29784826Chang, Z., Price, B. D., Bockheim, S., Boedigheimer, M. J., Smith, R. and Laughon, A. (1993). Molecular and genetic characterization of the Drosophila tartan gene. Dev. Biol. 160: 315-332. 8253267

Kohsaka, H. and Nose, A. (2009). Target recognition at the tips of postsynaptic filopodia: accumulation and function of Capricious. Development 136(7): 1127-35. PubMed Citation: 19270171

Krause, C., Wolf, C., Hemphala, J., Samakovlis, C. and Schuh, R. (2006). Distinct functions of the leucine-rich repeat transmembrane proteins Capricious and Tartan in the Drosophila tracheal morphogenesis. Dev. Biol. 296(1): 253-64. Medline abstract: 16764850

Kurusu, M., Cording, A., Taniguchi, M., Menon, K., Suzuki, E. and Zinn, K. (2008). A screen of cell-surface molecules identifies leucine-rich repeat proteins as key mediators of synaptic target selection. Neuron 59(6): 972-85. PubMed Citation: 18817735

Mantovani, R. (1999). The molecular biology of the CCAAT-binding factor NF-Y. Gene 239: 15-27. PubMed Citation: 10571030

Mao, Y., Kerr, M. and Freeman, M. (2007). Modulation of Drosophila retinal epithelial integrity by the adhesion proteins Capricious and Tartan. PLoS ONE 3(3): e1827. PubMed Citation: 18350163

Milán, M., et al. (2001). The LRR proteins Capricious and Tartan mediate cell interactions during DV boundary formation in the drosophila wing. Cell 106: 785-794. 11572783

Milán, M., Pérez, L. and Cohen, S. M. (2002). Short-range cell interactions and cell survival in the Drosophila wing. Dev. Cell 2: 797-805. 12062091

Milán, M., Pérez, L. and Cohen, S. M. (2005). Boundary formation in the Drosophila wing: functional dissection of Capricious and Tartan. Dev. Dyn. 233(3): 804-10. PubMed Citation: 15830355

Morey, M., et al. (2008). Coordinate control of synaptic-layer specificity and rhodopsins in photoreceptor neurons. Nature 456: 795-799. PubMed Citation: 18978774

Roppongi, R. T., Dhume, S. H., Padmanabhan, N., Silwal, P., Zahra, N., Karimi, B., Bomkamp, C., Patil, C. S., Champagne-Jorgensen, K., Twilley, R. E., Zhang, P., Jackson, M. F. and Siddiqui, T. J. (2020). LRRTMs Organize Synapses through Differential Engagement of Neurexin and PTPsigma. Neuron 106(1): 108-125 e112. PubMed ID: 31995730

Sakurai, K. T., Kojima, T., Aigaki, T. and Hayashi, S. (2007). Differential control of cell affinity required for progression and refinement of cell boundary during Drosophila leg segmentation. Dev. Biol. 309: 126-136. PubMed Citation: 17655839

Shinza-Kameda, M., et al. (2006). Regulation of layer-specific targeting by reciprocal expression of a cell adhesion molecule, Capricious. Neuron 49: 205-213. 16423695

Shishido, E., Takeichi, M. and Nose, A. (1998). Drosophila synapse formation: Regulation by transmembrane protein with leu-rich repeats, CAPRICIOUS. Science 280: 2118-2121. 9641918

Taniguchi, H.,, et al. (2000). Functional dissection of Drosophila capricious: its novel roles in neuronal pathfinding and selective synapse formation. J. Neurobiol. 42: 104-116. 10623905

Wolf, C. and Schuh, R. (2000). Single mesodermal cells guide outgrowth of ectodermal tubular structures in Drosophila, Genes Dev. 14: 2140-2145. Medline abstract: 10970878

Yamagata, M., Weiner, J. and Sanes, J. R. (2002). Sidekicks: synaptic adhesion molecules that promote lamina-specific connectivity in the retina. Cell 110: 649-660. 12230981

Biological Overview

date revised: 5 October 2020

Home page: The Interactive Fly © 2011 Thomas Brody, Ph.D.