Semaphorin-1a Acts in Concert With the Cell Adhesion Molecules Fasciclin II and Connectin to Regulate Axon Fasciculation in Drosophila

Corresponding author: Alex L. Kolodkin, Department of Neuroscience, The Johns Hopkins University School of Medicine, 725 N. Wolfe St./813 WBSB, Baltimore, MD 21205., kolodkin{at}jhmi.edu (E-mail)

Communicating editor: K. ANDERSON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Semaphorins comprise a large family of phylogenetically conserved secreted and transmembrane glycoproteins, many of which have been implicated in repulsive axon guidance events. The transmembrane semaphorin Sema-1a in Drosophila is expressed on motor axons and is required for the generation of neuromuscular connectivity. Sema-1a can function as an axonal repellent and mediates motor axon defasciculation. Here, by manipulating the levels of Sema-1a and the cell adhesion molecules fasciclin II (Fas II) and connectin (Conn) on motor axons, we provide further evidence that Sema-1a mediates axonal defasciculation events by acting as an axonally localized repellent and that correct motor axon guidance results from a balance between attractive and repulsive guidance cues expressed on motor neurons.

THE establishment of functional circuits in the mature nervous system requires a well-choreographed series of axon pathfinding and target recognition events during neural development. Neurons extend axons toward their targets through a variety of different environments that present temporally and spatially regulated cues. These cues in turn provide guidance information for discrete growth cone steering decisions. The molecular mechanisms that underlie these events are beginning to be elucidated, and it is clear that long- and short-range attractive and repulsive cues act to coordinate these complex events (TESSIER-LAVIGNE and GOODMAN 1996 ; MUELLER 1999 ). A large phylogenetically conserved gene family, the semaphorins, includes both secreted and transmembrane proteins that are expressed in a variety of neural and non-neural tissues during development. Several secreted and transmembrane semaphorins have been shown to function as repellents both in vivo and in vitro (MARK et al. 1997 ; WINBERG et al. 1998A , WINBERG et al. 1998B ; YU et al. 1998 ; ENCINAS et al. 1999 ; KIKUCHI et al. 1999 ; RAPER 2000 ; XU et al. 2000 ). While repulsive interactions resulting from axons avoiding semaphorin-expressing substrates in vivo or exhibiting growth cone collapse upon semaphorin exposure in vitro are easily interpretable, the situation is often more complex. For example, the transmembrane semaphorin Sema-1a is expressed on all motor axon pathways in Drosophila. Ectopic expression of Sema-1a on target muscles demonstrates that Sema-1a can function as a repellent. However, loss of Sema-1a function results in discrete defasciculation defects, suggesting that Sema-1a functions in axon-axon interactions (YU et al. 1998 ). Plexin A (Plex A), a member of the Plexin family of proteins, has been shown to function as a receptor for Sema-1a and is also expressed on motor axons (WINBERG et al. 1998B ). Interestingly, axon guidance defects observed in PlexA mutant embryos are very similar to those seen in Sema1a mutants. The motor axon defasciculation phenotypes observed in both Sema1a and PlexA mutants can be interpreted as resulting from the lack of a localized axonal repulsive signal that facilitates defasciculation of motor axons at specific choice points in their trajectories. Here, we focus on the role played by axonally localized Sema-1a in mediating defasciculation events essential for the generation of precise neuromuscular connectivity.

In both vertebrates and invertebrates, axons serving related functions often travel within fasciculated bundles, segregating from each other once they arrive at their appropriate targets. Temporal and spatial regulation of these segregation events is likely to involve a balance between attractive and repulsive guidance cues. Axonally expressed cues govern the degree of association among axons within any fascicle. Non-neuronal cues also regulate fasciculation by directing individual axons to avoid intermediate portions of their trajectories and continue to their ultimate targets or by attracting axons upon arrival at their targets and causing them to leave the fascicle. Therefore, either increasing attraction among axons that are part of the same nerve bundle or increasing repulsion between axons and the non-neuronal substrates they traverse should enhance axonal fasciculation. Conversely, increasing repulsion among axons that make up a fascicle or increasing attraction between axons and their intermediate or final targets should facilitate axonal defasciculation. These predictions are in large part supported by several studies on motor axon pathfinding and target recognition in both invertebrates and vertebrates.

Neural cell adhesion molecule (NCAM) and its Drosophila orthologue, fasciclin II (Fas II), are both cell adhesion molecules (CAMs) belonging to the immunoglobulin (Ig) superfamily of proteins and are expressed on the surface of vertebrate and invertebrate motor axons, respectively (LANDMESSER et al. 1988 ; GRENNINGLOH et al. 1991 ; VANVACTOR et al. 1993 ). Overexpression of Fas II on motor axons in Drosophila increases attraction among motor axons and results in hyperfasciculation (LIN and GOODMAN 1994 ). In vertebrates, removal of the highly positively charged polysialic acid moiety, which is found covalently attached to NCAM located on vertebrate motor axons in regions where defasciculation events occur, negatively regulates axon-axon adhesion and results in hyperfasciculation of motor axons and incorrect targeting events (TANG et al. 1994 ). In Drosophila the CAM connectin (Conn), a leucine-rich repeat (LRR)-containing protein that is expressed on a subset of motor axons during development, has also been shown to mediate homophilic adhesive events in vivo (NOSE et al. 1997 ). The importance of regulating Fas II- and Conn-mediated adhesion among motor axons is further demonstrated by genetic analysis of mutations in the beaten path (beat) gene (FAMBROUGH and GOODMAN 1996 ). Beat is a secreted Ig domain-containing molecule that is expressed and secreted by motoneurons. In beat mutants hyperfasciculation of motor axons is observed, and this defect can be suppressed by loss-of-function mutations in FasII and conn. Taken together, these studies show that axonally expressed CAMs play an important role in establishing correct motor axon pathways during development. Elevating or reducing levels of target-derived attractants [netrin A and B (NetA and NetB), Conn, fasciclin III (Fas III)], or repellents [semaphorin-2a (Sema-2a) or Net-B] on muscles in Drosophila also leads to defects in muscle innervation and either hyper- or hypofasciculation of motor axons (CHIBA et al. 1995 ; KOSE et al. 1997 ; NOSE et al. 1997 ; WINBERG et al. 1998A ). These studies in Drosophila serve to demonstrate the effects that altering the levels of target-derived guidance cues has on motor axon fasciculation. Further, the importance of the relative, and not the absolute, levels of both axonal and target-derived attractants and repellents has been demonstrated in Drosophila by experiments employing systematic alterations in the dosage of axonal and target-derived attractants and repellents (WINBERG et al. 1998A ).

To better understand how attractive and repulsive axonal cues mediate defasciculation events at pathfinding choice points during neural development, we have manipulated the expression levels of Sema-1a and two CAMs, Fas II and Conn, which are normally expressed on motor axons during Drosophila neural development. Our results, employing both gain-of-function and loss-of-function paradigms, show that Sema-1a controls axonal defasciculation decisions at discrete choice points by countering the attractive functions of Fas II and Conn. These results complement observations of WINBERG et al. 1998B showing suppression of PlexA motor axon defasciculation defects by mutations in FasII. They also further demonstrate a repulsive role in vivo for axonally expressed Sema-1a in regulating axonal bundling in fascicles, showing that a normal balance must obtain between attractive and repulsive forces on axons for correct axon pathfinding to occur.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Genetic crosses and immunohistochemistry:
All lethal lines were balanced with a marked second or X chromosome balancer, CyOlacZ or FM7clacZ, respectively. For overexpression experiments, a UAS-FasII transgenic line (LIN and GOODMAN 1994 ) and a pan-neural GAL4 driver line, C155 (LUO et al. 1994 ) were recombined into the Sema1a^P1 background. Sema1a^P1 homozygous embryos were identified by performing ß-galactosidase activity staining as described (YU et al. 1998 ). For suppression experiments, a null FasII allele, FasII^eB112, and a conn hypomorphic allele, conn^Fvex238, were also recombined into the Sema1a^P1 background. FasII^eB112/+; Sema1a^P1/Sema1a^P1 and FasII^eB112/+; Sema1a^P1/Sema1a^P1; +/conn^Fvex238 embryos were generated by crossing FasII^eB112/FM7clacZ; Sema1a^P1/CyOlacZ females with Sema1a^P1/CyOlacZ and Sema1a^P1/CyOlacZ; conn^Fvex238/conn^Fvex238 male flies. These embryos were identified by performing ß-galactosidase activity staining and selecting those embryos with the Sema1a-specific distribution of ß-galactosidase activity but without the ftz and pan-neuronal pattern of ß-galactosidase activity that are produced by the marked X and second balancer chromosomes, respectively. FasII^eB112/+; Sema1a^P1/Sema1a^P1; conn^Fvex238/conn^Fvex238 embryos were generated by crossing FasII^eB112/+; Sema1a^P1/CyOlacZ; +/conn^Fvex238 female flies with FM7clacZ/Y; Sema1a^P1/CyOlacZ; +/conn^Fvex238 males (these flies were collected from the cross above). The triple FasII; Sema1a; conn embryos were identified by performing ß-galactosidase activity staining and selecting those embryos with the ftz-specific distribution of ß-galactosidase activity but with no pan-neuronal pattern of ß-galactosidase activity. These triple-mutant embryos were further selected by using anti-Connectin (C1, 427) staining (MEADOWS et al. 1994 ) and scoring for a lack of ISNb phenotypes in using anti-Fas II (1D4) staining (VANVACTOR et al. 1993 ). Immunohistochemistry, staging, and scoring of Drosophila embryos were performed as described (YU et al. 1998 ). Chi-square analysis was used to determine the statistical significance of phenotypic differences observed among different genotypes.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The stereotypic pattern of embryonic neuromuscular connectivity in Drosophila provides an excellent model system to study the fasciculation and defasciculation events governed by a variety of attractive and repulsive guidance cues. These include semaphorins and the CAMs FasII and Conn (LIN et al. 1994 ; FAMBROUGH and GOODMAN 1996 ; NOSE et al. 1997 ; WINBERG et al. 1998A ; YU et al. 1998 ). In this study we focus our attention on a subset of motor axon pathways (see schematic in Fig 3A), including the intersegmental nerve (ISN), the intersegmental nerve b (ISNb), and the segmental nerve a (SNa). The axons of the ISN and ISNb form a single fused fascicle as they exit the central nervous system (CNS). After navigating into the periphery, motor axons of ISNb undergo a series of defasciculation events (Fig 1A and Fig 3A). First, the ISNb splits off from the ISN near ventral oblique muscles 15–17 and navigates among the ventral lateral muscles (VLMs). Within the VLMs, the RP3 motoneuron defasciculates from ISNb and participates in the formation of a synaptic arborization between VLMs 6–7. Other motoneurons, including RP1, 4, and 5, remain part of the ISNb and extend dorsally. Subsequent pathfinding events allow RP1, 4, and 5 to extend from the external surface of VLM 6 to the internal surface of VLM 13. RP1, 4, and RP5 then participate in the establishment of synaptic arborizations on VLMs 13 and 12, respectively. Motor axons of SNa also undergo a series of defasciculation events similar to the ISNb (Fig 2A and Fig 3A). Upon extending dorsally past VLM 12, the dorsal and lateral branches of SNa defasciculate from each other and navigate to different lateral muscles (LMs). Motor axons in the dorsal branch of the SNa further defasciculate between LMs 22 and 23 to innervate their LM targets (Fig 3A). In a series of loss-of-function and gain-of-function experiments, we investigated the dynamic interactions between repulsive Sema-1a signals and attractive forces provided by the CAMs Fas II and Conn in the generation of precise neuromuscular connectivity.

View larger version (123K): In this window In a new window Download PPT slide   Figure 1. Sema-1a and Fas II in ISNb motor axon guidance. Abdominal segments of a filleted late stage 16 wild-type embryo (A), a Sema1aP1 mutant embryo (B), a FasIIeB112/+; Sema1aP1 double-mutant embryo (C), and a pan-neurally expressed Fas II in Sema1aP1 mutant embryo (D). All panels show embryos stained with MAb 1D4 to reveal ISNb phenotypes. (A) The ISNb pathway in a wild-type embryo is shown, revealing the synaptic arborization on VLMs 6–7 elaborated by RP3 (lower arrow) and the synaptic arborization on VLM 12 made by RP5 (upper arrow). Numbers mark the location of VLMs 12, 13, 6, and 7. (B) In Sema1aP1 mutant embryos ISNb is most often stalled at VLMs 6–13 (arrowhead). RP3 and RP5 synaptic arborizations are absent (arrows). (C) Suppression of ISNb phenotypes is observed in FasIIeB112/+; Sema1aP1 double-mutant embryos. Synaptic arborizations on VLMs 12 and 6–7 appear normal (arrows). (D) Dramatically enhanced Sema1aP1 ISNb phenotypes were observed when Fas II was expressed on all neurons in a Sema1aP1 mutant background, including a failure of the ISNb to defasciculate from the ISN, which results in a fusion bypass with the ISN (arrowheads). ISNb was also observed to stall ventrally on VLMs 6–7 (asterisk). Bar, 12 µm.

View larger version (115K): In this window In a new window Download PPT slide   Figure 2. Sema-1a, Fas II, and Conn in SNa motor axon guidance. Abdominal segments of a filleted late stage 16 wild-type embryo (A), a Sema-1aP1 mutant embryo (B), a FasIIeB112/+; Sema1aP1; connFvex238 triple-mutant embryo (C), and a pan-neurally expressed Fas II in Sema1aP1 mutant embryo (D) stained with MAb 1D4 to reveal SNa phenotypes. (A) The SNa pathway in a wild-type embryo is shown. SNa bifurcates, forming dorsal (arrowhead) and lateral (asterisk) branches after passing VLM 12. The SNa dorsal branch further bifurcates on LMs 22–23 (arrow). (B) In Sema1aP1 mutant embryos, the dorsal SNa branch is most often stalled on LMs 22–23 (arrows) where normally it would bifurcate. (C) Rescue of Sema1aP1 SNa stall phenotype is observed in FasIIeB112/+; Sema1aP1; connFvex238 triple-mutant embryos. The SNa dorsal branches bifurcate at LMs 22–23 and extend dorsally (arrows). (D) Dramatically enhanced Sema1aP1 SNa phenotypes were observed when Fas II was expressed on all neurons in a Sema1aP1 mutant background that included failure of the SNa lateral branches to defasciculate from the SNa pathway, resulting in the lack of SNa lateral branches (asterisks). The stall phenotype of the dorsal SNa branch was also observed (arrowheads). Bar, 12 µm.

View larger version (25K): In this window In a new window Download PPT slide   Figure 3. Schematic representation of the interaction between Sema-1a, Fasciclin II, and Connectin in ISNb and SNa motor axon guidance. Schematics of motor axon fasciculation defects in various genetic backgrounds are shown using a lateral view of a subset of motor axon pathways and abdominal muscles (anterior is to the left). Arrows between panels indicate manipulations of Sema1a, FasII, and conn. Sema-1a and Fas II are expressed on all motor axons. Conn is expressed only on SNa axons. Representative pathways are shown (A–C) for the most abundant classes of phenotype observed. Filled circles denote synaptic arborizations. (A) In wild-type embryos, the SNa and the ISNb undergo a series of axonal defasciculation events to establish characteristic synaptic arborizations in VLMs and LMs. See text for details of the generation of normal motor axon trajectories in wild-type embryos. Wild-type SNa and ISNb pathways are often observed in FasII/+; Sema1a; conn triple-mutant embryos. Sema1a defasciculation defects in the SNa and the ISNb were rescued by removing the CAMs Fas II and Conn from these motor axons. (B) In wild-type embryos with pan-neural ectopic Fas II or in Sema1a mutant embryos, the ISNb axons most often fail to defasciculate at discrete choice points, resulting in the lack of RP3 innervation on VLMs 6–7 and a stall phenotype on VLMs 6–13. The dorsal SNa branch most often stalls on LMs 22–23 where normally it would bifurcate. (C) In Sema1a mutant embryos with ectopic expression of Fas II on all neurons, the ISNb most often does not enter the VLMs and instead exhibits a fusion-bypass phenotype. The lateral branch of the SNa often fails to defasciculate from the SNa pathway, which results in a loss of the entire SNa lateral branch.

Fasciculation and defasciculation in the ISNb pathway are regulated by a dynamic interaction between Sema-1a and Fas II:
In theory, either increasing attraction between motor axons or increasing repulsion between motor axons and their substrates should enhance axonal adhesion and favor axonal fasciculation. On the other hand, either increasing repulsion between motor axons or increasing attraction between motor axons and their substrates should discourage axon adhesion and facilitate axonal defasciculation.

Since Sema-1a regulates defasciculation of motor axons by acting as a repellent (WINBERG et al. 1998B ; YU et al. 1998 ), we explored whether the establishment of neuromuscular connectivity is further regulated by dynamic interactions between repulsive Sema-1a and attractive Fas II and Conn. First, we used the UAS-GAL4 system to overexpress Fas II on motor axons by crossing a UAS-FasII transgenic line to a pan-neural GAL4 line, C155 (BRAND and PERRIMON 1993 ; LUO et al. 1994 ). While single alleles of both the C155-GAL4 and UAS-FasII transgenes showed only modest hyperfasciculation defects (Table 1) when combined with a null mutation in Sema1a (Sema1a^P1; YU et al. 1998 ), dramatic enhancement of hyperfasciculation defects in the ISNb pathway was observed.

A failure of the ISNb to separate from the ISN (a "fusion bypass" phenotype) was increased 5-fold from what is observed in Sema1a^P1 embryos when Sema1a^P1 mutant embryos ectopically express Fas II on all motor axons (P < 0.001; Fig 1B and Fig D, and Table 1). Similarly, the percentage of hemisegments exhibiting aberrant or absent RP3 innervation of the VLMs 6 and 7 is increased 1.5-fold following these same genetic manipulations (P < 0.001; Fig 1B, and Fig D, and Table 1). A failure of RP1, 4, and 5 to defasciculate around the VLMs 13 and 6, a "stall" phenotype, was also increased in Sema1a^P1 embryos expressing pan-neural Fas II (P < 0.001; Fig 1B and Fig D, and Table 1). The dramatic enhancement of these hyperfasciculation phenotypes resulting from neuronal CAM overexpression suggests that the interplay between a repulsive Sema-1a cue and an attractive Fas II cue in part determines the state of axonal fasciculation. Interestingly, these enhanced ISNb hyperfasciculation defects showed Sema1a dosage dependence since an intermediate level of enhanced ISNb phenotypes was observed in heterozygous Sema1a^P1 embryos ectopically expressing Fas II on motor axons (Table 1).

Since increasing adhesion by overexpressing Fas II on neurons results in hyperfasciculation of motor axons in Sema1a embryos, reducing motor axon adhesion by removing Fas II should suppress fasciculation defects in Sema1a mutant embryos. Therefore, we combined a null allele of FasII mutation (FasII^eB112; GRENNINGLOH et al. 1991 ) into Sema1a^P1 mutant animals (FasII^eB112/FM7clacZ; Sema1a^P1/CyOlacZ) and then examined motoneuron phenotypes in double-mutant embryos. Since we used an antibody directed against Fas II (MAb 1D4; VANVACTOR et al. 1993 ) as a marker to identify motoneuron pathways, we examined motoneuron phenotypes in Sema1a embryos heterozygous for FasII (FasII^eB112/+; Sema1a^P1/Sema1a^P1). Embryos heterozygous for FasII^eB112 exhibited wild-type motor axon projections (data not shown). The aberrant RP3 arborization and ISNb phenotypes observed in Sema1a mutants were suppressed by 50% after removing one allele of FasII in homozygous Sema1a^P1 embryos (P < 0.001; Fig 1B and Fig C, and Table 1). Our results showing FasII suppression of Sema1a loss-of-function mutants are similar to, and expand upon, those of WINBERG et al. 1998B , where ISNb defects resulting from the loss of a Sema-1a receptor Plex A were observed to be suppressed by removing Fas II from motor axons (FasII^eB112/+; PlexA/PlexA). Taken together, these Sema1a and FasII loss-of-function and gain-of-function studies strongly suggest that the degree of adhesion among motor axons in the peripheral nervous system is regulated by a dynamic interaction between attractive and repulsive guidance cues and that this serves to modulate the state of fasciculation in the ISNb pathway.

Sema-1a, Fas II, and Conn regulate the formation of SNa motor axon trajectories:
The experiments described above show that overexpression of the CAM Fas II on motor axons significantly enhances Sema1a ISNb hyperfasciculation defects and that loss-of-function mutations in FasII can suppress Sema1a ISNb hyperfasciculation defects. Since both Sema-1a and Fas II are expressed on all motor neuron pathways, including the SNa pathway, we wondered whether Sema-1a and Fas II dynamically regulate fasciculation of this motor axon pathway.

We examined SNa phenotypes in Sema1a embryos ectopically expressing Fas II on motor axons. Similar to the enhanced ISNb hyperfasciculation defects we observed following these same genetic manipulations, we observed an 20-fold enhancement of a failure of the SNa lateral branch to separate from the SNa dorsal branch, resulting in a lack of a lateral SNa branch (P < 0.001; Fig 2B and Fig D, and Table 2). A failure of the dorsal SNa branch to defasciculate between the LMs 22–23 (SNa stall phenotype) was also significantly increased in Sema1a^P1 embryos ectopically expressing neuronal Fas II (P < 0.001; Fig 2B and Fig D, Table 2). These results show that, as was observed for the ISNb, Sema-1a and Fas II act in concert to regulate motor axon pathfinding in the SNa.

To test whether the Sema1a SNa stall phenotype, like the ISNb stall phenotype, could be suppressed by removing Fas II from motor axons, we examined Sema1a embryos heterozygous for FasII (FasII^eB112/+; Sema1a^P1/Sema1a^P1). The SNa defasciculation defects resulting from a loss of Sema1a, however, cannot be suppressed by reducing the dose of FasII on motor axons (Table 2). A similar observation was noted previously with respect to a lack of FasII-mediated suppression of PlexA SNa stall phenotypes (WINBERG et al. 1998B ). This failure to suppress SNa stall phenotypes could be due to an inability of the removal of one-half of the Fas II from the SNa pathway to compromise Fas II function in Sema1a embryos. This also raises the possibility that an additional CAM might work together with Sema-1a and Fas II to regulate fasciculation of the SNa pathway.

Since the CAM Conn is expressed on the SNa pathway and is important for its formation (FAMBROUGH and GOODMAN 1996 ; NOSE et al. 1997 ), we asked whether Conn also functions together with Sema-1a and Fas II to regulate SNa pathway fasciculation. We combined a null allele of FasII mutation (FasII^eB112), a hypomorphic allele of conn mutation [conn^Fvex238; 5% expression level of endogenous Conn (NOSE et al. 1994 )], and Sema1a^P1 and then examined SNa phenotypes in these triple-mutant embryos (FasII^eB112/+; Sema1a^P1/Sema1a^P1; conn^Fvex238/conn^Fvex238). We observed 65% suppression of SNa stall phenotypes following Conn removal (P < 0.001; Fig 2C and Table 2). This suppression cannot be achieved by removing Conn alone from motor axons in Sema1a^P1 embryos (Table 2). It should also be noted that both conn hypomorphic (conn^Fvex238) and null (conn^Flex14) alleles behave similarly with respect to an inability to suppress Sema1a SNa stall phenotypes (Table 2). Finally, conn mutations did not have similar effects on the ISNb, as expected, since Conn is not expressed on this motor axon pathway (NOSE et al. 1992 ). This genetic analysis using loss- and gain-of-function mutations in Sema1a, FasII, and conn demonstrates that Sema-1a, acting as a repellent, functions cooperatively with the CAMs Fas II and Conn to regulate fasciculation and defasciculation to establish correct neuromuscular connectivity in the SNa pathway.

CNS axon fasciculation is also regulated by an interaction between Sema-1a and Fas II:
WINBERG et al. 1998B showed that mutations in Fas II suppressed PlexA CNS defasciculation defects, which are observed in the most-lateral Fas II-expressing longitudinal connective in PlexA mutant embryos. Mutations in Sema1a also show discontinuities in the same connective (YU et al. 1998 ), suggesting that Sema-1a/PlexA-mediated repulsive interactions are essential for establishing the normal pattern of connectives in Drosophila. To further address this issue, we removed one allele of FasII⁺ in a homozygous Sema1a mutant background and asked whether this would suppress defects in the most-lateral Fas II-expressing longitudinal connective. Indeed, whereas Sema1a^P1/Sema1a^P1 mutant embryos exhibit these CNS defects in 31% of hemisegments, in FasII^eB112/+; Sema1a^P1/Sema-1a^P1 mutant embryos this phenotype is suppressed to a level of 10% (26/260 hemisegments). Therefore, as we have observed in embryonic motor axon projections, reduction in the level of the Ig CAM Fas II can suppress CNS fasciculation defects resulting from a loss of Sema-1a.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Adhesive and repulsive interactions determine whether or not fasciculation or defasciculation will occur at specific choice points along axonal trajectories. We previously reported that the transmembrane semaphorin Sema-1a is expressed on the surface of motor axons in Drosophila embryos and is required for their correct guidance to their target muscles (YU et al. 1998 ). The failure of axonal defasciculation observed in Sema1a mutant embryos is likely due to the lack of Sema-1a-mediated repulsion among motor axons along efferent trajectories. Here we show that reducing adhesion by removal of attractive cues, the CAMs Fas II and Conn, rescues characteristic hyperfasciculation defects (both ISNb and SNa phenotypes) of Sema1a mutants. On the other hand, increasing adhesion by overexpression of the CAM Fas II enhances the hyperfasciculation defects in the ISNb and SNa pathways in Sema1a mutant embryos. In addition, reduction in the level of Fas II will also suppress CNS fasciculation defects in Sema1a mutant embryos. These experiments complement those by WINBERG et al. 1998B showing that FasII loss of function can suppress defasciculation defects observed in PlexA mutants. Further, they show that mutations in genes encoding different classes of CAMs, including both Ig superfamily members and LRR-containing proteins, genetically interact with Sema1a mutants, suggesting that CAM-specific signaling events (WALSH and DOHERTY 1993 ) are not involved in this interaction. Taken together, these results demonstrate that Sema-1a regulates axonal fasciculation at specific choice points by countering the attractive functions of at least two CAMs, Fas II and Conn.

Although further analysis is needed, it appears that both Sema-1a and PlexA proteins are expressed in all motor axons and most CNS axons, suggesting that these axons are all competent to respond to Sema-1a (WINBERG et al. 1998B ; YU et al. 1998 ). How might Sema-1a modulate specific defasciculation events at choice points when it appears to be expressed along the entire motor axon? Sema-1a may serve to negatively regulate motor axon adhesion over the entire trajectory and thereby allow extending motor axon growth cones to respond to target recognition cues. A precedent for a reciprocal role for the CAM Fas II comes from a detailed ultrastructural analysis of FasII mutants, where it was observed that although a loss of Fas II does not compromise the extension of axons during early CNS development, it does result in a lack of fasciculation among individual neurons that normally comprise discrete axon bundles (LIN et al. 1994 ). On the basis of our analyses of Sema-1a and those of WINBERG et al. 1998B of Plex A, it seems likely that Sema-1a on CNS axons acts as a generally expressed repellent. The absence of the CAM Fas II, therefore, changes the balance Fas II and Sema-1a and results in defasciculation of CNS bundles. Therefore, a combination of attractants and repellents may serve to allow individual axons within a developing bundle to respond to cues that may reside either at choice points or in adjacent intermediate or final target regions. On the other hand, local regulation of the function of these axonally localized cues may also serve to modulate their function, a possibility supported by the observation that Beat, which negatively regulates Fas II in motor axons, appears to be localized at certain ventral motor axon choice points (FAMBROUGH and GOODMAN 1996 ). Both Conn and Fas II, in addition to being expressed on motor axons, are also expressed on embryonic muscles (NOSE et al. 1992 ; SCHUSTER et al. 1996 ). However, our interpretation that the genetic interactions we observe reveal a balance of repulsive and attractive axonal cues required to regulate motor axon fasciculation is likely not to be compromised by this muscle CAM expression for the following reasons. First, if loss of CAM expression on motor neurons and target regions was capable of altering axonal fasciculation through alteration of axon/target interactions, one would expect the effect to be a reduction in CAM-mediated motor axon extension on muscle and an enhancement of motor axon fasciculation—the opposite of what we observe. Second, the levels of Fas II, and to a lesser extent Conn, on embryonic muscles have been shown to be much lower than high axonal levels of these CAMS observed during the embryonic stages we have analyzed here for motor axon pathfinding (NOSE et al. 1992 ; SCHUSTER et al. 1996 ).

Previous studies support the idea that guidance cues act cooperatively to generate the precise pattern of neuromuscular connectivity in Drosophila. WINBERG et al. 1998A showed that netrins and a secreted semaphorin expressed on muscles and axonal Fas II function in a complementary fashion to regulate motor axon target selection. The dynamic balance of attractive cues (Net A and B, and Fas II) and repulsive cues (Sema-2a and Net B) was shown to govern the ability of motor axons to undergo normal target recognition and formation of synaptic arborizations. Indeed, alteration of the degree of axonal attraction mediated by Fas II directly affected the ability of motor axons to respond to target-derived attractants and repellents. Our results, in combination with those of WINBERG et al. 1998B on PlexA/FasII interactions, extend these observations to the axon bundle itself and illustrate the importance of the interplay between axonal attractants and repellents for complex axon guidance events. It should be noted that although our results support a repulsive role for Sema-1a in motor axon guidance, the related transmembrane semaphorin Sema-1a in the grasshopper has been shown to act as an attractive cue for peripheral afferents in the developing limb (WONG et al. 1999 ). In this situation, however, Sema-1a is target derived and not axonal, and it remains to be seen if transmembrane semaphorin-mediated attraction is also plexin dependent. Finally, the simultaneous expression of transmembrane semaphorins and CAMs on axon pathways undergoing complex pathfinding and defasciculation events is not restricted to invertebrates. For example, in the developing vertebrate olfactory system the Ig superfamily members NCAM and olfactory cell adhesion molecules show selective distribution on main and accessory olfactory neurons, suggesting that these CAMs play an instructive role in matching odorant receptor expression zones in the olfactory epithelium and in the olfactory bulb (reviewed in YOSHIHARA and MORI 1997 ). In addition, both the transmembrane semaphorin Sema6A and the plexin Plexin A1 are expressed in subsets of primary and accessory olfactory neurons (reviewed in YU and KOLODKIN 1999 ). The defasciculation events involved in directing these neurons to their unique glomerular targets in the main and accessory olfactory bulbs are reminiscent of the motor axon guidance events discussed here, suggesting that the maintenance of a balance on axons between repulsive transmembrane semaphorins and attractive CAMs is phylogenetically conserved.

	ACKNOWLEDGMENTS

We thank C. Goodman and R. White for MAbs and fly stocks, members of the Kolodkin laboratory for helpful discussions, and J. Terman for critical reading of the manuscript. This work was supported by a National Institutes of Health grant (NS-35165) to A.L.K. and by an undergraduate award from the Grass Foundation and Pomona College to A.S.H.

Manuscript received April 24, 2000; Accepted for publication June 19, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BRAND, A. H. and N. PERRIMON, 1993  Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118:401-415[Abstract].

CHIBA, A., P. SNOW, H. KESHISHIAN, and Y. HOTTA, 1995  Fasciclin III as a synaptic target recognition molecule in Drosophila.. Nature 374:166-168[Medline].

ENCINAS, J. A., K. KIKUCHI, A. CHEDOTAL, F. DE CASTRO, and C. S. GOODMAN et al., 1999  Cloning, expression, and genetic mapping of Sema W, a member of the semaphorin family. Proc. Natl. Acad. Sci. USA 96:2491-2496[Abstract/Free Full Text].

FAMBROUGH, D. and C. S. GOODMAN, 1996  The Drosophila beaten path gene encodes a novel secreted protein that regulates defasciculation at motor axon choice points. Cell 87:1049-1058[Medline].

GRENNINGLOH, G. E., J. REHM, and C. S. GOODMAN, 1991  Genetic analysis of growth cone guidance in Drosophila: fasciclin II functions as a neuronal recognition molecule. Cell 67:45-57[Medline].

KIKUCHI, K., A. CHEDOTAL, H. HANAFUSA, Y. UJIMASA, and F. DE CASTRO et al., 1999  Cloning and characterization of a novel class VI semaphorin, semaphorin Y. Mol. Cell. Neurosci. 13:9-23[Medline].

KOSE, H., D. ROSE, X. ZHU, and A. CHIBA, 1997  Homophilic synaptic target recognition mediated by immunoglobulin-like cell adhesion molecule fasciclin III. Development 124:4143-4152[Abstract].

LANDMESSER, L., L. DAHM, K. SCHULTZ, and U. RUTISHAUSER, 1988  Distinct roles for adhesion molecules during innervation of embryonic chick muscle. Dev. Biol. 119:645-670.

LIN, D. M. and C. S. GOODMAN, 1994  Ectopic and increased expression of fasciclin II alters motoneuron growth cone guidance. Neuron 13:507-523[Medline].

LIN, D. M., R. D. FETTER, C. KOPCZYNSKI, G. GRENNINGLOH, and C. S. GOODMAN, 1994  Genetic analysis of fasciclin II in Drosophila: defasciculation, refasiculation, and altered fasciculation. Neuron 13:1055-1069[Medline].

LUO, L., Y. J. LIAO, L. Y. JAN, and Y. N. JAN, 1994  Distinct morphogenetic functions of similar small GTPases: Drosophila Drac1 is involved in axonal outgrowth and myoblast fusion. Genes Dev. 8:1787-1802[Abstract/Free Full Text].

MARK, M. D., M. LOHRUM, and A. W. PUSCHEL, 1997  Patterning neuronal connections by chemorepulsion: the semaphorins. Cell Tissue Res. 290:299-306[Medline].

MEADOWS, L. A., D. GELL, K. BROADIE, A. P. GOULD, and R. A. WHITE, 1994  The cell adhesion molecule, connectin, and the development of the Drosophila neuromuscular system. J. Cell Sci. 107:321-328[Abstract].

MUELLER, B. K., 1999  Growth cone guidance: first steps towards a deeper understanding. Annu. Rev. Neurosci. 22:351-388[Medline].

NOSE, A., V. B. MAHAJAN, and C. S. GOODMAN, 1992  Connection: a homophilic cell adhesion molecule expressed on a subset of muscles and the motoneurons that innervate them in Drosophila.. Cell 70:553-567[Medline].

NOSE, A., M. TAKEICHI, and C. S. GOODMAN, 1994  Ectopic expression of connection reveals a repulsive function during growth cone guidance and synapse formation. Neuron 13:525-539[Medline].

NOSE, A., T. UMEDA, and M. TAKEICHI, 1997  Neuromuscular target recognition by a homophilic interaction of Connection cell adhesion molecules in Drosophila. Development 124:1433-1441[Abstract].

RAPER, J. A., 2000  Semaphorins and their receptors in vertebrates and invertebrates. Curr. Opin. Neurobiol. 10:88-94[Medline].

SCHUSTER, C. M., G. W. DAVIS, R. D. FETTER, and C. S. GOODMAN, 1996  Genetic dissection of structural and functional components of synaptic plasticity. I. Fasciclin II controls synaptic stabilization and growth. Neuron 17:641-654[Medline].

TANG, J., U. RUTISHAUSER, and L. LANDMESSER, 1994  Polysialic acid regulates growth cone behavior during sorting of motor axons in the plexus region. Neuron 13:405-414[Medline].

TESSIER-LAVIGNE, M. and C. S. GOODMAN, 1996  The molecular biology of axon guidance. Science 274:1123-1133[Abstract/Free Full Text].

VANVACTOR, D., H. SINK, D. FAMBROUGH, R. TSOO, and C. S. GOODMAN, 1993  Genes that control neuromuscular specificity in Drosophila. Cell 73:1137-1153[Medline].

WALSH, F. S. and P. DOHERTY, 1993  Factors regulating the expression and function of calcium-independent cell adhesion molecules. Curr. Opin. Cell Biol. 5:791-796[Medline].

WINBERG, M. L., K. J. MITCHELL, and C. S. GOODMAN, 1998a  Genetic analysis of the mechanisms controlling target selection: complementary and combinatorial functions of netrins, semaphorins, and IgCAMs. Cell 93:581-591[Medline].

WINBERG, M. L., J. N. NOORDERMEER, L. TAMAGNONE, P. M. COMOGLIO, and M. K. SPRIGGS et al., 1998b  Plexin A is a neuronal semaphorin receptor that controls axon guidance. Cell 95:903-916[Medline].

WONG, J. T., S. T. WONG, and T. P. O'CONNOR, 1999  Ectopic semaphorin-1a functions as an attractive guidance cue for developing peripheral neurons. Nat. Neurosci. 2:798-803[Medline].

XU, X. M., D. A. FISHER, L. ZHOU, F. A. WHITE, and S. NG et al., 2000  The transmembrane protein semaphorin 6A repels embryonic sympathetic axons. J. Neurosci. 20:2638-2648[Abstract/Free Full Text].

YOSHIHARA, Y. and K. MORI, 1997  Basic principles and molecular mechanisms of olfactory axon pathfinding. Cell Tissue Res. 290:457-463[Medline].

YU, H. H. and A. K. KOLODKIN, 1999  Semaphorin signaling: a little less per-Plexin. Neuron 22:11-14[Medline].

YU, H. H., H. H. ARAJ, S. A. RALLS, and A. L. KOLODKIN, 1998  The transmembrane semaphorin Sema I is required in Drosophila for embryonic motor and CNS axon guidance. Neuron 20:207-220[Medline].

This article has been cited by other articles:

				C. O. Law, R. J. Kirby, S. Aghamohammadzadeh, and A. J. W. Furley The neural adhesion molecule TAG-1 modulates responses of sensory axons to diffusible guidance signals Development, July 15, 2008; 135(14): 2361 - 2371. [Abstract] [Full Text] [PDF]

				C. M. Miller, A. Page-McCaw, and H. T. Broihier Matrix metalloproteinases promote motor axon fasciculation in the Drosophila embryo Development, January 1, 2008; 135(1): 95 - 109. [Abstract] [Full Text] [PDF]

				F. Moret, C. Renaudot, M. Bozon, and V. Castellani Semaphorin and neuropilin co-expression in motoneurons sets axon sensitivity to environmental semaphorin sources during motor axon pathfinding Development, December 15, 2007; 134(24): 4491 - 4501. [Abstract] [Full Text] [PDF]

				M. A. Wolman, A. M. Regnery, T. Becker, C. G. Becker, and M. C. Halloran Semaphorin3D Regulates Axon Axon Interactions by Modulating Levels of L1 Cell Adhesion Molecule J. Neurosci., September 5, 2007; 27(36): 9653 - 9663. [Abstract] [Full Text] [PDF]

				J. C. Ayoob, J. R. Terman, and A. L. Kolodkin Drosophila Plexin B is a Sema-2a receptor required for axon guidance Development, June 1, 2006; 133(11): 2125 - 2135. [Abstract] [Full Text] [PDF]

				P. Cafferty, L. Yu, H. Long, and Y. Rao Semaphorin-1a Functions as a Guidance Receptor in the Drosophila Visual System J. Neurosci., April 12, 2006; 26(15): 3999 - 4003. [Abstract] [Full Text] [PDF]

				J. C. Ayoob, H.-H. Yu, J. R. Terman, and A. L. Kolodkin The Drosophila Receptor Guanylyl Cyclase Gyc76C Is Required for Semaphorin-1a-Plexin A-Mediated Axonal Repulsion J. Neurosci., July 28, 2004; 24(30): 6639 - 6649. [Abstract] [Full Text] [PDF]

				Z.-Y. Chen, C. Sun, K. Reuhl, A. Bergemann, M. Henkemeyer, and R. Zhou Abnormal Hippocampal Axon Bundling in EphB Receptor Mutant Mice J. Neurosci., March 10, 2004; 24(10): 2366 - 2374. [Abstract] [Full Text] [PDF]

				A.E. Wiencken-Barger, J. Mavity-Hudson, U. Bartsch, M. Schachner, and V.A. Casagrande The Role of L1 in Axon Pathfinding and Fasciculation Cereb Cortex, February 1, 2004; 14(2): 121 - 131. [Abstract] [Full Text] [PDF]

				M. Alenius and S. Bohm Differential function of RNCAM isoforms in precise target selection of olfactory sensory neurons Development, March 1, 2003; 130(5): 917 - 927. [Abstract] [Full Text] [PDF]

				B. Key and J. St John Axon Navigation in the Mammalian Primary Olfactory Pathway: Where to Next? Chem Senses, March 1, 2002; 27(3): 245 - 260. [Abstract] [Full Text] [PDF]

				Z. He, K. C. Wang, V. Koprivica, G. Ming, and H.-J. Song Knowing How to Navigate: Mechanisms of Semaphorin Signaling in the Nervous System Sci. Signal., February 12, 2002; 2002(119): re1 - re1. [Abstract] [Full Text] [PDF]

				T. Kawasaki, Y. Bekku, F. Suto, T. Kitsukawa, M. Taniguchi, I. Nagatsu, T. Nagatsu, K. Itoh, T. Yagi, and H. Fujisawa Requirement of neuropilin 1-mediated Sema3A signals in patterning of the sympathetic nervous system Development, January 2, 2002; 129(3): 671 - 680. [Abstract] [Full Text] [PDF]