Drosophila Robo homologs

Axonal growth cones require an evolutionary conserved repulsive guidance system to ensure proper crossing of the CNS midline. In Drosophila, the Slit protein is a repulsive signal secreted by the midline glial cells. It binds to the Roundabout receptors, which are expressed on CNS axons in the longitudinal tracts but not in the commissural tracts. An analysis of the genes leak (referred to in much of the literature as roundabout 2) and kuzbanian is presented: both genes are involved in the repulsive guidance system operating at the CNS midline. Mutations in leak, were first recovered by Nusslein-Volhard (1984) based on defects in the larval cuticle. Analysis of the head phenotype suggests that slit may act as an attractive guidance cue while directing the movements of the dorsal ectodermal cell sheath. kuzbanian regulates midline crossing of CNS axons. It encodes a metalloprotease of the ADAM family and genetically interacts with slit. Expression of a dominant negative Kuzbanian protein in the CNS midline cells results in an abnormal midline crossing of axons and prevents the clearance of the Roundabout receptor from commissural axons. These analyses support a model in which Kuzbanian mediates the proteolytic activation of the Slit/Roundabout receptor complex (Schimmelpfeng, 2001).

Seven independent mutations resulting in a roundabout-like axon phenotype were initially identified based on their abnormal commissure phenotype, as revealed by a BP102 staining. In addition, the longitudinal connectives appear reduced. Complementation analysis and genetic mapping experiments show that five of these mutations indeed represent new roundabout alleles. In roundabout mutant embryos, Fasciclin II expressing CNS axons repeatedly cross the CNS midline, most frequently the formation of the medial-most fascicle, the pCC/MP2 fascicle, is affected. In older embryos, a collapse of this fascicle along the CNS midline can be observed (Schimmelpfeng, 2001).

The two remaining mutations (H1-3 and S4-14) complement roundabout (meaning they are not robo mutants) but fail to complement each other. Using recombination and deficiency mapping, the corresponding gene was placed in the chromosomal interval 21D1-22A2. The allele H1-3 is a temperature sensitive allele and leads to a stronger phenotype at 298C. The embryonic CNS phenotype of homozygous mutant S4-14 embryos is comparable to those carrying deficiencies of the interval 21D1-22A2 in trans to S4-14, suggesting that S4-14 represents an amorphic allele (Schimmelpfeng, 2001).

In contrast to the wild type, Fasciclin II positive axon fascicles frequently cross the CNS midline in S4-14 mutant embryos. Compared to roundabout mutant embryos, however, the phenotype appears more pronounced. It is not confined to the pCC/MP2 fascicle but also affects the lateral and medial Fasciclin II positive axon bundles. In addition, the formation of the longitudinal connectives is reduced. This phenotypic trait is again more pronounced than in roundabout mutant embryos. In S4-14 mutant but not in H1-3 mutant embryos muscle fibers frequently cross the CNS on the dorsal surface of the CNS, which they never do in the wild type. The formation of the peripheral nervous system as judged by the expression of the 22C10 marker appears normal (Schimmelpfeng, 2001).

Besides the embryonic CNS phenotype, a larval cuticle defect is observed. S4-14 leads to embryonic lethality; the homozygous mutant larvae show a typical cuticle phenotype. The head skeleton is defective and the lateral-Gräten and the H-piece do not form. A similar mutant cuticle phenotype has been described for mutant leak alleles, which have been mapped meiotically to the same genomic interval (Nusslein-Volhard, 1984). Subsequent complementation analyses have shown that H1-3 and S4-14 both are new leak alleles. Furthermore, leak1 mutant embryos display a similar but weaker mutant embryonic CNS phenotype. Thus, it can be concluded that leak is required to prevent crossing of longitudinal axon tracts similar to roundabout (Schimmelpfeng, 2001).

Complementation analysis reveals that leak is indeed allelic to robo2, indicating that leak encodes robo2. Mutations in the slit gene, as with leak, were initially identified based on their common head phenotype (Nusslein-Volhard, 1984). The head defects of leak/robo2 are indicative for abnormal head involution. In wild type embryos this process starts during stage 11/12 when the brain neuroblasts invaginate into the interior of the embryo. Subsequently, a complex set of morphogenetic movements results in a closure of the gap dorsal to the forming brain lobes. The posterior dorsal ectoderm continues to extend anteriorly and together with the adverse movement of the dorsal clypeolabrum leads to the formation of the dorsal pouch. At the end of embryogenesis the dorsal ectoderm has enclosed the anterior tip of the embryo. slit and leak are expressed in different ectodermal domains of the head. Digoxygenin labeled anti-sense slit RNA was used to probe slit expression during head development. slit expression can be observed in the ectoderm just ventral to the invaginating foregut as soon as the CNS midline expression of slit can be detected. In addition, slit expression is detected in the clypeolabrum. In stage 12 embryos three expression domains can be recognized. By stage 13 the middle domain resolves into two distinct domains. Domain 'd' has broad lateral extension. In stage 16 embryos, domains 'a-c' are found inside the embryo. Also, ectodermal slit expression in domain 'd' is found dorsal to the pharynx musculature in the dorsal pouch (Schimmelpfeng, 2001).

Leak protein expression was monitored using a polyclonal antiserum. In wild type stage 12 embryos high levels of Leak protein are found ventrally to the foregut adjacent to but in parts overlapping with the slit-expressing domain. This expression mode is different from the ventral nerve cord where Leak protein is excluded from the midline. However, during stage 13/14 Leak expression is lost in the ectodermal midline cells in the head. During head involution, the dorsal ectoderm slides over the forming brain lobes and over the developing foregut region. High levels of Leak expression are found in the leading edge cells of the migrating dorsal ectoderm. In stage 17 embryos Leak expression prefigures the forming mouth hooks (Schimmelpfeng, 2001).

Within the ventral nerve cord of slit mutant embryos, the ventral expression domain of Leak shifts toward the midline. Interestingly, Leak expression in the anterior ventral ectoderm appears to remain unchanged. In the dorsal expression domain, leak expressing ectodermal cells do not migrate as fast toward the anterior as they do in the wild type. In fact, in slit mutant embryos the dorsal ectoderm fails to migrate toward the anterior tip of the embryo. In addition, Leak expressing cells do not properly prefigure the mouth hooks. The medial connection is not formed but rather one can observe that the Leak expressing cells fan out laterally (Schimmelpfeng, 2001).

In summary, Leak appears to act as a Slit receptor during head involution. During this process it controls morphogenetic movements rather than directed growth of axons or filopodia. The observed migration defects in slit mutants may indicate that Slit/Leak interaction does not necessarily lead to a repulsive signaling mechanism (Schimmelpfeng, 2001).

To determine whether leak and roundabout generally act in the same genetic pathway controlling axon guidance at the embryonic CNS midline, the CNS phenotype of robo;leak double mutant embryos was examined. Mutations in either roundabout or leak result in relatively mild CNS defects and axons crossing the midline appear to be able to grow away from the midline. Opposed to that, however, in leak;roundabout double mutant embryos, most CNS axons collapse at the CNS midline, resulting in a severe, almost slit-like mutant CNS phenotype. No cuticle phenotype is found in roundabout mutant larvae and roundabout;leak double mutant larvae display a leak-like cuticle phenotype, which is comparable to the slit cuticle phenotype. Thus, in contrast to the CNS, where both Roundabout and Leak appear to act as Slit receptors, migration of muscle cells or ectodermal cells during head involution requires leak but not roundabout (Schimmelpfeng, 2001).

Thus, in addition to its role during axon guidance, leak also functions during cell migration. About 20 years ago, mutations in leak were identified based on their characteristic head defects that suggested a function during head involution, a process that requires morphogenetic cell movements. The role of Roundabout receptors in cell migration is not surprising since mutations in the Roundabout ligand Slit result in defects in the migration of mesodermal cells. slit mutants also exhibit similar defects in head involution as seen in leak mutant embryos (Nusslein-Volhard, 1984). Similarly, slit function in vertebrates has been associated with cell migration. The phenotypic analysis of the slit/leak mutant phenotypes may, however, indicate that Slit does not act as a chemorepellent during head involution. slit is expressed in the dorsal pouch along which leak expressing cells migrate during head involution. In slit mutants, migration of this cell sheath is reduced. A simple explanation of this phenotype could be that the slit expressing dorsal pouch cells are attractive for the leak expressing ectodermal cells -- alternatively one would have to postulate a more caudally located slit source driving the ectodermal cells toward the anterior of the embryo by a repulsive mechanism. However, no such slit expressing cells were detected, suggesting that Slit, as Netrins, can act in both ways: as repulsive and as attractive guidance cues (Schimmelpfeng, 2001).

Another complementation group consisting of five alleles that leads to abnormal commissure formation maps to the genomic interval 34C/D. Subsequent complementation analyses indicates that these alleles are EMS-induced alleles of kuzbanian. In kuzbanian mutant embryos, many Fasciclin II positive axons appear to cross the CNS midline. In fact, it appears as if many, if not all, commissural axons abnormally express the Fasciclin II antigen. As in roundabout or leak, the longitudinal axon tracts of kuzbanian mutant embryos are reduced in size but they are generally found in more lateral positions (Schimmelpfeng, 2001).

The kuzbanian phenotype suggests that kuzbanian might be involved in the repulsive signaling system operating at the CNS midline. Other work has also suggested that Kuzbanian participates in the processing of the secreted Slit protein (Brose, 1999). A possible genetic interaction of the two mutations was tested by first analyzing the CNS phenotype of slit kuzbanian/++ embryos. When heterozygous, both mutations do not lead to a detectable embryonic CNS phenotype. In contrast to this, in 12% of the slit/kuzbanian double heterozygote embryos (or in 6% of the neuromeres out of 650 neuromeres counted), a roundabout-like mutant CNS phenotype was found. In embryos lacking zygotic kuzbanian function, in addition to being heterozygous for slit, a kuzbanian-like CNS phenotype emerged. The position of the longitudinal connectives, however, is shifted toward the CNS midline. Removal of one copy of kuzbanian in a slit mutant background does not enhance the slit phenotype (Schimmelpfeng, 2001).

This could be interpreted such that the Kuzbanian protease is required to activate the Slit protein, which serves as a ligand of the Roundabout receptors. Different allelic combinations of roundabout and kuzbanian were analyzed. roundabout;kuzbanian double heterozygote embryos appear wild type. When one copy of kuzbanian is removed in a roundabout mutant background, a roundabout CNS phenotype developes. The phenotype revealed by Fasciclin II staining may be slightly more extreme compared to the roundabout phenotype, since the lateral Fasciclin II positive fascicles are also affected. The overall axon pattern of embryos homozygous for roundabout and kuzbanian appears to be a slightly more severe phenotype compared to the roundabout phenotype, since axons running along the CNS midline were frequently detected. This is also evident following a Fasciclin II staining (Schimmelpfeng, 2001).

slit and kuzbanian also genetically interact. Since slit is required for head formation as well, the cuticle phenotype of kuzbanian mutant larvae was examined. Removal of the zygotic expression did not lead to a head defect as observed for slit or leak mutant larvae. It is presumed that maternal kuzbanian function may mask a head phenotype. Removal of maternal and zygotic kuzbanian function leads to a neurogenic phenotype with no recognizable head structures (Schimmelpfeng, 2001).

In kuzbanian mutants Leak expression is unaffected. To address the question whether Kuzbanian affects the expression of Roundabout an anti-Robo antibody was used. In wild type embryos, the Roundabout protein is always found on the longitudinal connectives. No Roundabout protein is expressed on commissural tracts. In kuzbanian mutant embryos, however, Roundabout can be detected on commissural axons crossing the CNS midline. Thus, Kuzbanian may function to clear Roundabout/Slit receptor-ligand complex from axons crossing the CNS midline (Schimmelpfeng, 2001).

During development, kuzbanian is expressed ubiquitously and most, if not all, CNS neurons appear to express the gene. Expression of a dominant negative Kuzbanian protein in all CNS neurons using the elav promoter leads to a kuzbanian-like CNS axon phenotype. To better analyze which cells in the developing CNS have to express the Kuzbanian protein in order to prevent the inappropriate crossing of the CNS midline by navigating axons, the GAL4 system was used. Using a single-minded GAL4 driver strain, a dominant negative Kuzbanian protein was expressed in all CNS midline cells. In 18% of the embryos no gross CNS defects were observed; in the remaining embryos, inappropriate midline crossing of Fasciclin II positive axons was found. 40% of these embryos displayed an almost roundabout-like CNS axon phenotype. Additionally, the formation of the lateral Fasciclin II positive axon tracts was affected, pointing toward a non-cell-autonomous function of the dominant negative Kuzbanian protein when expressed in the CNS midline. This is also suggested by the observation that in embryos expressing the dominant negative Kuzbanian protein, muscle fibers traverse the CNS dorsally as observed in slit or leak mutant embryos. It was next of interest to find out whether the expression of dominant negative Kuzbanian in the CNS midline cells also affects the clearance of the Roundabout protein from commissural axons. As observed in kuzbanian mutant embryos, Roundabout is found on commissural axon tracts, suggesting that Kuzbanian participates in the down-regulation of Roundabout expression on commissural axons. This also shows that axons can cross the midline despite the expression of Roundabout (Schimmelpfeng, 2001).

This phenotypic analysis has indicated that in kuzbanian mutant embryos, longitudinally projecting axons expressing Fasciclin II inappropriately cross the midline. These commissural axons also ectopically express the Roundabout receptor, which they never do in wild type embryos. Together with the data obtained from analyzing kuzbanian function during Delta-Notch signaling, this may be interpreted either as a requirement of kuzbanian in the CNS midline to generate the fully active Slit protein or it may indicate that Kuzbanian activates one or more Roundabout receptors in the lateral CNS (Schimmelpfeng, 2001).

Therefore, to further elucidate the function of kuzbanian, a dominant negative version was expressed in all CNS midline cells. Interestingly, this results in a mutant CNS phenotype combining phenotypic traits of roundabout and leak: axons and muscle fibers cross the CNS midline. This may indicate a function of Kuzbanian during Slit processing, possibly influencing the formation of the Slit gradient in the developing Drosophila nervous system. Alternatively, it may be interpreted in such a way that the cleavage of the Roundabout/Slit complex is required to facilitate growth cone retraction at the midline. This latter possibility is reminiscent of recent findings regarding the function of Ephrin signaling. The membrane-anchored Ephrin ligands constitute a large family of axon guidance molecules, which upon binding to Eph-receptor tyrosine kinases frequently mediate repulsive signals. In order to form the receptor-ligand complex, which triggers signaling to the cellular cytoskeleton, the two cells involved must adhere and subsequently cannot pull their plasma membranes apart. Only after the membrane anchored Ephrin ligand is cleaved by the Kuzbanian metalloprotease is the retraction process initiated (Schimmelpfeng, 2001).

In wild type embryos, Commissureless is expressed in the CNS midline cells and down-regulates the expression of Roundabout on commissural axons. It has been suggested that this down-regulation is required in order to allow crossing of contralateral projecting axons. In roundabout mutant embryos, axons frequently cross or even recross the CNS midline since the midline repellent Slit cannot be perceived. Conversely, high levels of roundabout expression result in a commissureless mutant phenotype. In kuzbanian mutant embryos, axons are able to cross the midline despite the expression of Roundabout on commissural axons. The local expression of a membrane-bound dominant negative Kuzbanian protein in the CNS midline mimics this phenotype. Thus, kuzbanian functions at the CNS midline in the clearance of the Roundabout receptor from commissural axons. This process may involve calmodulin and Sos. In the slit;kuzbanian double mutant, too, axons cross the midline and Roundabout protein is found on the surface of commissural axons. Furthermore, in kuzbanian mutants, the function of roundabout appears to be reduced since axons cross the midline, indicating that only the activated Slit/Roundabout complex can induce its clearance from commissural axons (Schimmelpfeng, 2001).

Drosophila sensory neurons form distinctive terminal branch patterns in the developing neuropile of the embryonic central nervous system. A genetic analysis of factors regulating arbor position shows that mediolateral position is determined in a binary fashion by expression (chordotonal neurons) or nonexpression (multidendritic neurons) of the Robo3 receptor for the midline repellent Slit. Robo3 expression is one of a suite of chordotonal neuron properties that depends on expression of the proneural gene atonal. Different features of terminal branches are separately regulated: an arbor can be shifted mediolaterally without affecting its dorsoventral location, and the distinctive remodeling of one arbor continues as normal despite this arbor shifting to an abnormal position in the neuropile (Zlatic, 2003).

Two characteristics of sensory neurons that contribute to the final specification of their terminal arborization can be distinguished: position and modality. This study deals specifically with the question of modality-dependent arborization and, as a model for understanding the mechanisms involved, uses the distinctively different projections formed by the ch and md neurons, respectively. The complexity of individual arbors can be reduced by focusing separately on the essential features that give each arbor its characteristic identity; this study initially concentrated on arbor positioning within the mediolateral axis of the CNS. This work was carried out within a framework provided by a prominent set of axon fascicles that are labeled by antibodies to the cell adhesion molecule Fasciclin II (FasII), and this enables the detection of small changes in the position and/or structure of terminal arbors. The axons of the ch neurons branch and terminate on an intermediate axon fascicle, while those of the md neurons terminate on a medial fascicle. There are additional distinguishing features in the two projections, such as their positions in the dorsoventral axis and the late embryonic remodeling of the arbor formed by the dorsal bipolar dendrite (dbd) neuron, an identified member of the md class (Zlatic, 2003).

Using this system, the following four questions were asked about the way in which cell type-specific arborization is patterned in the developing CNS. (1) Since it is known that the Slit/Robo system positions axon fascicles in the mediolateral axis of the CNS, it can be asked whether the same system directly or indirectly controls the positions at which terminal arbors are formed in this axis. (2) Are different aspects of the same arbor, such as dorsoventral as opposed to mediolateral positioning, coordinately controlled or independently specified? (3) Will the likely targets of terminating neurons influence the position at which termination actually occurs? (4) What are the genetic controls that specify the different characteristics of a terminal arbor? It was found that arbor position in the mediolateral axis is determined in a binary fashion by the expression (ch neurons) or nonexpression (md neurons) of the Robo3 receptor for the midline repellent Slit. Robo3 expression is one of a suite of properties characteristic of ch neurons that depends on the expression of atonal in the precursors of these cells. Thus, while ectopic expression of Robo3 in md neurons switches their terminal arbor from a medial to an intermediate fascicle in the mediolateral axis, other properties of the arbor (such as late remodeling of the dbd arbor) are unaffected. In contrast, ectopic expression of the proneural gene ato in the precursors of md neurons transforms all aspects of their central projection to form a ch-like arborization (Zlatic, 2003).

The terminal arbors of sensory neurons in the Drosophila embryo form within a developing neuropile that consists largely of commissural and longitudinal axon fascicles together with the intervening branches and dendrites of other neurons. This mass of axons and dendrites is an organized three-dimensional structure within which there are clear signs of a conserved architecture that recurs throughout the insects and probably more widely in the arthropod phylum. At its most fundamental level, there is a subdivision in which the principal arborizations of the motoneurons form dorsally, whereas sensory terminals are generally found more ventrally. The endings of sensory neurons themselves terminate in distinctive regions of the more ventral neuropile that have an overall similarity across the insects -- the mechanosensory endings in the most ventral sector and the proprioceptive endings more dorsally (Schrader, 2000). This suggests that connections form in a structured environment that imposes an order on the developmental process. It is the nature of this ordering process with which this paper is concerned. The analysis focuses on the distinctive terminal arborizations formed by two different types of sensory neurons, the ch neurons and the dbd neurons (Zlatic, 2003).

The first detailed maps of sensory terminals in the developing CNS of the Drosophila embryo and larva revealed that the arborizations of the sensory axons are as characteristic of particular sense organs as the structures of the sense organs themselves (Merritt, 1995; Schrader, 2000). For example, ch organs have arbors that form at an intermediate and fairly ventral level in the neuropile. In contrast, the arbors of most md neurons develop at a more medial and slightly more ventral location. Within the md class, however, four neurons have specialized projections. In particular, the dbd neuron has a dorsal projection that is uniquely remodeled toward the end of embryogenesis. Interestingly, dbd and the other neurons with specialized arbors correspond to identifiable subsets of cells within the md class that are distinguished both by their peripheral structures and by patterns of gene expression (Zlatic, 2003).

The first finding is that in wild-type embryos, md and ch arborizations are closely aligned with elements of the FasII-positive axon tracts that form a conspicuous set of longitudinal pathways in the embryonic and larval nervous systems. These consistent patterns of termination allow the precise definition, in two dimensions (the dorsoventral and mediolateral axes), of the different coordinates at which ch and dbd terminals will normally form in developing neuropile (Zlatic, 2003).

How are these coordinates specified for each neuron type? Previous work in Drosophila has shown that the Robo receptors for the repellent Slit are responsible for positioning CNS axon fascicles in the medial, intermediate, and lateral tracts on either side of the midline. Robo has also been shown to influence dendritic growth and synaptic connectivity in the giant fiber system of the adult fly. At the time of sensory neuron branch formation, Robo is expressed in dbd/md and ch neurons and acts to confine both kinds of projections to the ipsilateral CNS. In a robo mutant, both sets of neurons form bilateral projections, but the distinctive locations of the terminals persist on both sides of the midline. This distinction depends on the expression of Robo3, which is present in ch neurons but not in dbd/md neurons. When Robo3 is absent or downregulated, ch terminals form at the more medial position characteristic of db/dmd neurons. Similarly, if robo3 is ectopically expressed in dbd/md neurons, their terminals are shifted into the ch region of the neuropile. The action of Robo3 on mediolateral positioning is a direct one for two reasons: (1) either ectopic expression of Robo3 or downregulation of Robo3 function (by expression of comm) in sensory neurons but not their central targets selectively induces lateral shifts of sensory arbors and (2) ch projections can form in an intermediate position independently of their normal substrate for arborization. In robo3 mutants, the intermediate FasII fascicle on which ch arbors normally form shifts medially, as do the ch projections. However, when Robo3 function is selectively restored in the sensory system of robo3 mutants, the ch projections form in the intermediate region of the neuropile, as in wild-type (Zlatic, 2003).

These experiments also show that the locations at which terminals form in the mediolateral and dorsoventral axes are separately specified and one can be manipulated independently of the other. Thus, in all cases where dbd/md or ch terminals are shifted in the mediolateral axis by changing levels of Robo3, the dorsoventral position remains unchanged: a medially shifted ch terminal is ventral; a laterally shifted dbd terminal is dorsal. This finding reinforces the distinction between two alternative ways of determining the position of sensory axon terminals in the embryonic neuropile. The simplest model would suggest that growing axons are attracted to a specific central target: in this case, the position of the target (however determined) fixes the three-dimensional location of the terminal arbor of the sensory neuron. These results suggest that this model does not apply: in these experiments, the position of the terminal arbor is regulated by factors that separately specify its position in two dimensions. It is the response of the sensory axon to these different factors rather than to their targets that determines where the terminal arbor will form (Zlatic, 2003).

It is concluded that there are several properties of ch neurons that dictate the position and shape of their terminal arbors, of which Robo3 expression is one. Robo3 expression is specified by the proneural gene ato. Thus, when ato is expressed in other embryonic sensory neurons, these cells too express Robo3. In addition, misexpressing ato can completely transform dbd projections into ch-like projections (although the cell body remains dbd-like). In such cases, both mediolateral and dorsoventral positions are changed and the late stage remodeling does not take place. Thus, ato not only regulates Robo3 expression but also controls the whole suite of properties that dictates the position-specific termination of ch neurons. The mechanism of this regulation, at least for Robo3 expression, is likely to be an indirect one. (1) Transformation (including Robo3 expression) of md into ch-like neurons by ectopic ato requires that ato be expressed early, during the period of sensory organ specification and not as the arbors form. (2) Only primary (not secondary) ch organs express ato, yet all ch neurons express Robo3 and project their terminal arbors into a common region of the neuropile (Zlatic, 2003).

Previous studies have shown that ectopic ato can transform adult es organs into ch organs and that those transformations are achieved by an active downregulation of cut, which is required for es organ formation but is also expressed in a subset of md neurons. If ato upregulated Robo3 by downregulating Cut, then in cut mutant embryos all normally cut-expressing md neurons should be found to express Robo3. This is not the case and in cut mutants Robo3 is expressed only in the ch neurons, as in wild-type. It is also found that md neurons of the dorsal cluster, which normally express Cut, retain their wild-type projections in 15 hr cut mutant embryos (Zlatic, 2003).

While ato regulates ch-like characteristics, what specifies the unique features of the dbd arbor? Alone among md neurons, the dbd neuron projects to the dorsomedial FasII fascicle and undergoes late stage remodeling. The md neurons are a heterogenous population with respect to their expression of proneural transcription factors and identity genes such as cut. Interestingly, the SOPs of the dbd (and ddaE) neurons do not express cut or the AS-C but a different proneural gene, amos. By analogy with the role of ato in determining the ch-like projection pattern, amos may determine aspects of the dbd-specific projection pattern. It is also observed that ectopic expression of ato in dbd can generate intermediates, halfway transformations between dbd and ch, both with respect to mediolateral positioning and late stage remodeling. These intermediates may reveal competitive interactions between ch- and dbd-specific transcription factors. Thus, proneural genes could conceivably act in a combinatorial fashion to specify diverse shapes and positions of terminal branches (Merritt, 1995). Such a view is reinforced by the fact that it is those sensory neurons, which express a combination of proneural genes in their SOPs, that have distinctive terminal arbors (Merritt, 1995). However, there are likely to be additional factors that contribute to specify individual neuronal arbors, in particular in those instances in which terminal projection varies as a function of position. For example, in the cockroach the transcription factor Engrailed controls axon projections and synaptic choice of identified sensory neurons. Similarly, the prepattern genes araucan and caupolican play a role in establishing the difference in the projection pattern between lateral and medial es neurons in the adult Drosophila notum (Zlatic, 2003 and references therein).

Working from first principles, it might be supposed that different kinds of sensory neurons would locate the distinctive positions appropriate for their terminal arbors by seeking out their target interneurons in the developing CNS. However, in its simplest form, this model must be incorrect: in one axis at least, terminating axons detect and are positioned by their response to a cue that is produced not by their targets but by cells on the midline. This might suggest that sensory terminals are located at appropriate positions within the neuropile by factors that are quite independent of the neurons with which they will ultimately form connections. Indeed, an early experiment in the zebrafish embryo showed that if the target Mauthner cell is removed, the conspicuous cap of commissural axon terminals that normally encloses the Mauthner axon hillock continues to form at its proper location in the neuropile. It cannot necessarily be concluded from these experiments that sensory terminals in Drosophila are similarly positioned by factors entirely independent of the target. In the periphery, motoneuron axons are guided to their proper muscles by a hierarchy of cues, starting with a transcription factor code that delivers them to particular regions of the muscle field within which they then seek out and synapse with their targets. In the case of the embryonic sensory neurons, a hierarchy of cues, including target-derived signals, could also be envisaged that would contribute to the final projection pattern and certainly to synaptogenesis (Zlatic, 2003).

When approaching the apparently complex issue of how appropriate connections are formed between neurons within the meshwork of alternatives presented in the developing neuropiles, there may be two simplifying processes that should be considered. The first is that initial patterns of termination may be determined by a combination of cues that separately dictate distinct features of the arbor. The second is that the combination of such cues coupled with specific patterns of receptor expression may lead to the local arborization of pre- and postsynaptic neurons on a common substrate. Such patterns of growth would build a coherent platform on which detailed connectivity could then be established (Zlatic, 2003).

robo2 and robo3 interact with eagle to regulate serotonergic neuron differentiation

The function of the central nervous system (CNS) depends crucially upon the correct differentiation of neurons and formation of axonal connections. Some aspects of neuronal differentiation are known to occur as axonal connections are forming. Although serotonin is a highly conserved neurotransmitter that is important for many CNS functions, little is known about the process of serotonergic neuron differentiation. In Drosophila, expression of the serotonin transporter (SerT) is both temporally and physically related to midline crossing. Additionally, the axon guidance molecules roundabout2 and roundabout3 (robo2/3) are necessary for serotonergic neuron differentiation and function independently of their ligand, slit. Loss of robo2 or robo3 causes a loss of SerT expression in about half of neurons, and resembles the phenotype seen in mutants for the transcription factor eagle (eg). A direct relationship is shown between robo2/3 and eg: robo2/3 mutants lose Eg expression in serotonergic neurons, and robo2 and eg interact genetically to regulate SerT expression. It is proposed that post-midline expression of Robo2/3 is part of a signal that regulates serotonergic neuron differentiation and is transduced by the transcription factor Eg (Couch, 2004).

Thus SerT expression in the fly is temporally and physically tied to axon guidance across the midline. The data further indicate that the axon guidance molecules robo2 and robo3 (robo2/3) positively regulate serotonergic neuron differentiation; a loss of robo2/3 function causes a loss of SerT expression in ~50% of neurons. A robo2/3 loss of function closely resembles an eg mutant phenotype. Finally, the data show a dose-sensitive relationship between Robo2/3, Eg and SerT expression, suggesting that they function in the same genetic pathway to control serotonergic neuron differentiation. This interpretation is supported by the fact that loss of robo2 or robo3 causes a loss of Eg expression, and by genetic rescue experiments (Couch, 2004).

By visualizing serotonergic axonal projections with tau-lacZ, it was determined that SerT expression begins at the end of stage 15, just after growth cones complete midline crossing and reach the contralateral side. This temporal correlation between midline crossing and SerT induction suggests that the midline is important for serotonergic neuron differentiation in the fly, as it is in the grasshopper (Condron, 1999). Further evidence for the importance of the midline comes from data showing that in wild-type cords, axons physically separated from the midline fail to express SerT. These results recapitulate similar experiments in the grasshopper. Additionally, when the repulsive axon guidance receptor Unc5 is expressed in serotonergic neurons, a partial loss of SerT expression is observed. Although these results suggest a role for the midline in serotonergic neuron differentiation, it remains unclear whether this role is temporally restricted as it is in the grasshopper, and, additionally, what factors act as the presumptive midline signal. FGF signaling in the grasshopper is crucial for SerT induction (Condron, 1999), and plays a role in the differentiation of vertebrate serotonergic neurons. In the fly, experiments indicate that FGF signaling also appears to be important for SerT regulation (Couch, 2004).

One problem with interpreting the role for the midline is the lack of an abnormal serotonergic phenotype in mutants for the master regulatory gene sim, where midline cells fail to properly differentiate. It is difficult to speculate about what factors may allow normal differentiation in the absence of normal midline cells, since there are many changes in gene regulation throughout sim mutants. Although the results suggest a role for the midline in serotonergic neuron differentiation, it is likely to be more complicated than a simple switch acting to induce differentiation (Couch, 2004).

The data show that a loss of robo2 or robo3 causes a loss of SerT expression, suggesting that Robo2/3 function positively to regulate serotonergic neuron differentiation. A positive role for Robo2 is further supported by results showing that overexpression of Robo2 prevents a loss of SerT in neurons physically separated from the midline. Possibly, Robo2 functions downstream of the midline signal required for SerT induction and thus allows differentiation to proceed in the absence of such a signal. An alternative hypothesis suggests that Robo2/3 function indirectly to induce SerT, by guiding serotonergic axons to an unknown signal in the contralateral neuropil. Such indirect signaling occurs in the developing vertebrate CNS where trophic support is required by commissural axons at the floorplate, an intermediate axonal target. Although the possibility that Robo2/3 act indirectly to regulate serotonergic neuron differentiation cannot be ruled out at this time, several lines of evidence suggest a more direct role. These data shows that overexpression of Robo2 not only spares SerT loss following a midline cut but also rescues an eg hypomorph, and furthermore, that an Eg gain of function rescues a robo2 loss of function; these results strongly suggests that Robo2 functions autonomously in the serotonergic neurons. Additionally, SerT loss is not seen in other guidance mutants that disrupt midline crossing or cause general disorganization of the CNS. However, it is difficult to clearly resolve the presence of Robo2/3 protein specifically in the serotonergic neurons because of the broad distribution of neuronal processes and the fact that serotonergic neuron branching does not correspond simply to any Fas2 pathway where axons are known to express Robo2/3 (Couch, 2004).

Robo2 and Robo3 appear to act as overall regulators of differentiation rather than specific regulators of SerT, since robo2/3 mutants not only lose SerT expression (mRNA and reuptake activity) but also have defects in serotonin synthesis later in development. Thus, the role of Robo2/3 in serotonergic neuron differentiation parallels that of other genes, including eg and the LIM-homeodomain transcription factor islet, that cause both a loss of SerT as well as serotonin synthesis when disrupted. The data further indicate that robo2/3 are not required in the formation of the serotonergic neurons from their progenitor neuroblast 7-3. All serotonergic neurons in a robo2/3 mutant express eg-lacZ, even those with a loss of SerT expression. This may at first appear to contradict the result showing a loss of Eg at stage 16 in these mutants, since eg expression must have occurred in order to produce lacZ. It is hypothesized that lacZ staining in stage 16 robo2 mutants is likely to be due to a lengthy persistence of lacZ rather than continued expression of eg, since eg mRNA is not detectable by in situ hybridization after stage 14. Most probably, eg-lacZ expression in robo2/3 mutants occurs in the progenitors of serotonergic neurons when other factors, such as engrailed (en), are known to control eg expression. Even in eg mutants, all serotonergic neurons continue to express eg-lacZ, despite a disruption in SerT, serotonin synthesis and expression of Ddc. Thus, a robo2/3 mutant, like an eg mutant, does not affect the early specification of serotonergic neurons, including early eg expression, but instead affects later maturation (Couch, 2004).

Interestingly, an effect is observed of robo2/3, but not robo, on serotonergic neuron differentiation. Disparities between Robo and Robo2/3 function have been previously observed in the lateral positioning of axons where only Robo2/3 appear to play a role, and in dendritic guidance, synapse formation and midline crossing, where all three Robo receptors have separable functions. Furthermore, Robo2 and Robo3 show greater homology to each other than to Robo. Robo2/3 have cytoplasmic domains that diverge from Robo, and lack two motifs considered important for Robo signaling. Possibly, Robo2 and Robo3 regulate a Robo-independent signaling cascade that is critical for serotonergic neuron differentiation. Additionally, a loss of slit, the ligand for all three Robo receptors, does not perturb SerT expression, indicating that either another ligand exists or the function of Robo2/3 in serotonergic neuron differentiation is ligand independent. In C. elegans, some activities of the Robo homolog SAX-3 are thought to be Slit independent (Couch, 2004).

The transmembrane protein Comm has been shown to negatively regulate the levels of all three Robo receptors. After midline crossing, Comm expression decreases and Robo levels increase in order to prevent inappropriate midline crossing. In serotonergic neuron differentiation, Comm may play a role in regulating Robo2/3, such that levels of both Robo2/3 increase following midline crossing and thereby permit differentiation to proceed. To test this possibility, Comm was expressed using egGal4 to specifically induce a loss of Robo2/3 in the serotonergic neurons. In these experiments, expression of Comm causes a loss of SerT activity in only a few cells and with low penetrance. This is believed to be due to expression of Comm at levels insufficient for total loss of Robo2/3. Alternatively, other regulators of Robo2/3 may exist. However, neither a loss of Comm nor overexpression of Robo2/3 results in precocious serotonergic neuron differentiation, indicating a requirement for other signals (Couch, 2004).

In both a robo2 and a robo3 loss-of-function mutant, expression of the zinc-finger transcription factor eg is lost in the same cells that lose SerT expression. Additionally, overexpression of Robo2 rescues the loss of SerT observed in an eg hypomorph in a dose-sensitive manner. Finally, Eg gain of function rescues the SerT loss seen in robo2 loss-of-function mutants. These results indicate that Robo2/3 function in the same genetic pathway as Eg to control serotonergic neuron differentiation. Although these results suggest that Robo2/3 regulate Eg in stage 16 embryos, other genes such as en and hunchback (hb) also have an established role in regulating Eg during serotonergic neuron differentiation. At present, it remains unclear if Robo2/3 cooperate with these genes to regulate Eg expression (Couch, 2004).

Additionally, in both robo2 and robo3 loss-of-function mutants only a percentage of neurons lose SerT expression (and serotonin synthesis), indicating the presence of a redundant mechanism for serotonergic neuron differentiation. The pattern of SerT and serotonin loss in robo2/3 mutants appears random and differs between nerve cords. At this point, it remains unclear why differentiation is affected in only some cells and not others, or what factors allow remaining cells to maintain normal SerT expression. One possibility is that cells must maintain a threshold level of Eg expression to differentiate properly. This is supported by differences in the degree of SerT loss according to the severity of the mutation in eg, since a hypomorphic allele displays a loss of SerT in ~30% of hemisegments while a null allele displays closer to 80% loss of SerT. Many studies have also suggested that a combinatorial code of transcription factors act to specify serotonergic properties: (1) loss-of-function mutations in several genes required for differentiation, including eg, en and hb show an incomplete loss of SerT phenotype; (2) if Eg is inappropriately expressed throughout the nervous system, only a few ectopic serotonin positive cells appear. These ectopic serotonergic cells always express the transcription factor hkb. Robo2 and Robo3 may also function redundantly. Further studies should indicate the relationship of Robo2/3 to other genes involved in serotonergic neuron differentiation, and the mechanism by which Robo2/3 regulate Eg expression (Couch, 2004).

One question that readily follows from these observations: how does Robo2 influence Eg expression? Robo2 and Robo3 are cell-surface axon guidance receptors, while Eg is a transcription factor. It is likely that other factors interact with both Robo2/3 and Eg to mediate their roles in serotonergic neuron differentiation. Although their relationship remains obscure, data indicate that Robo2 may regulate Eg post-transcriptionally. In a series of real-time RT-PCR experiments, no difference in eg mRNA levels was detected when EP2582 (UAS-robo2) was expressed using egGal4, scabrousGal4 or elavGal4, suggesting that Robo2 is insufficient to induce ectopic Eg expression. However, when Robo2 is overexpressed, a rescue of Eg protein expression is seen in egMz360 hypomorphs. Through the same series of PCR experiments, it was discovered that the egMz360 allele produces mRNA, although no Eg staining is observed. This suggests that the Gal4 insertion responsible for the egMz360 allele affects Eg protein expression, which in turn causes a disruption in SerT expression. Thus, expression of Robo2 appears to somehow rescue Eg protein expression in an egMz360 hypomorph sufficiently to rescue SerT activity. At this point, the mechanism of such a post-transcriptional rescue is unclear. Identifying the genetic and intracellular links between Robo2, Robo3 and Eg with more molecular approaches such as RNAi studies will probably reveal how Robo2/3 regulate not only Eg but eventually serotonergic neuron differentiation as well (Couch, 2004).

Slit/Robo-mediated axon guidance in Tribolium and Drosophila: Divergent genetic programs build insect nervous systems

As the complexity of animal nervous systems has increased during evolution, developmental control of neuronal connectivity has become increasingly refined. How has functional diversification within related axon guidance molecules contributed to the evolution of nervous systems? To address this question, the evolution of functional diversity was explored within the Roundabout (Robo) family of axon guidance receptors. In Drosophila, Robo and Robo2 promote midline repulsion, while Robo2 and Robo3 specify the position of lon- gitudinal axon pathways. The Robo family has expanded by gene duplication in insects; robo2 and robo3 exist as distinct genes only within dipterans, while other insects, like the flour beetle Tribolium castaneum, retain an ancestral robo2/3 gene. Both Robos from Tribolium can mediate midline repulsion in Drosophila, but unlike the fly Robos cannot be down-regulated by Commissureless. The overall architecture and arrangement of longitudinal pathways are remarkably conserved in Tribolium, despite it having only two Robos. Loss of TcSlit causes midline collapse of axons in the beetle, a phenotype recapitulated by simultaneous knockdown of both Robos. Single gene knockdowns reveal that beetle Robos have specialized axon guidance functions: TcRobo is dedicated to midline repulsion, while TcRobo2/3 also regulates longitudinal pathway formation. TcRobo2/3 knockdown reproduces aspects of both Drosophila robo2 and robo3 mutants, suggesting that TcRobo2/3 has two functions that in Drosophila are divided between Robo2 and Robo3. The ability of Tribolium to organize longitudinal axons into three discrete medial-lateral zones with only two Robo receptors demonstrates that beetle and fly achieve equivalent developmental outcomes using divergent genetic programs (Evans, 2012).

This work represents the first investigation of axon guidance in the Tribolium embryonic CNS. While there is a high degree of conservation in Slit/Robo-mediated guidance between these two insects, some important differences are noted. First and most obvious, Tribolium has only two Robo receptors (TcRobo and TcRobo2/3), while Drosophila has three (Robo, Robo2, and Robo3). Despite this divergent complement of axon guidance receptors, the architecture of the embryonic CNS in Drosophila and Tribolium, including the number and position of individual axon pathways, is remarkably similar (Evans, 2012).

Secondly, the results also suggest that the regulation of Slit/Robo signaling differs between Drosophila and Tribolium. In flies, Commissureless (Comm) is essential for preventing premature response to Slit by preventing Robo receptor molecules from reaching the growth cone surface in pre-crossing commissural axons. Tribolium Robos appear insensitive to regulation by Comm when expressed in Drosophila neurons, and the Tribolium genome does not appear to encode a Comm ortholog. How, then, is Slit responsiveness suppressed in pre-crossing commissural axons in Tribolium? A close examination of Robo receptor expression (at both mRNA and protein levels) and localization in different classes of Tribolium neurons may provide insight into the regulation of Slit/Robo signaling in beetles. Novel negative regulators of Slit/Robo signaling in Tribolium could be uncovered by forward genetic screens to identify genes required for midline crossing, similar to the screen that identified Drosophila comm (Evans, 2012).

The findings of this study demonstrate that at least two of the axon guidance functions performed by Drosophila Robo receptors -- midline repulsion and lateral position -- are shared by Robos in the beetle Tribolium. As in Drosophila, TcRobo (the ortholog of Drosophila Robo) is dedicated to midline repulsion, and does not appear to play a role in lateral positioning of longitudinal axon tracts. Similarly, TcRobo2/3 (orthologous to the ancestor of Drosophila Robo2 and Robo3) controls longitudinal pathway formation and also contributes to midline repulsion. Unlike in Drosophila robo2 mutants, however, no ectopic midline crossing could be detected in TcRobo2/3 knockdown embryos. While this may indicate that the RNAi-based approach only achieves a partial reduction of TcRobo2/3 function, the much more severe midline collapse phenotype seen in TcRobo+TcRobo2/3 double knockdowns compared to TcRobo single knockdowns suggests that indeed TcRobo2/3's midline repulsive activity (even at one half the dose of the single knockdown) was strongly reduced or eliminated. Indeed, no increase was observed in severity of TcRobo2/3's lateral positioning phenotype, or an appearance of ectopic midline crossing, when the concentration of TcRobo2/3 dsRNA was increased nearly ten-fold. The lack of ectopic midline crossing in TcRobo2/3 single knockdowns is interpreted as evidence that TcRobo2/3's role in midline repulsion is largely or completely redundant with that of TcRobo (Evans, 2012).

The finding that TcRobo2/3 performs a dual role in positioning both intermediate and lateral pathways indicates that this activity predates the divergence of Robo2 and Robo3, and that these two receptors in flies now subdivide a role that is performed by a single ancestral receptor in other insects. Thus Robo receptor control of lateral position of longitudinal pathways is likely to be widespread among insects. As Robo2/3 appears to be unique to insects, when did this activity arise? Two possibilities are envisioned: (1) lateral positioning activity was present in the ancestral Robo receptor, and retained by Robo2/3 after the gene duplication event that separated it from Robo; (2) Robo2/3 gained lateral positioning activity after it separated from Robo. Examining the roles of Robo and Robo2/3 in axon guidance in additional insects, or related animals that lack a distinct Robo2/3 receptor (e.g. crustaceans) may distinguish between these possibilities (Evans, 2012).

The third, and least well understood, function of Robos in axon guidance in the Drosophila embryonic CNS is Robo2's role in promoting midline crossing. Commissural axons fail to cross the midline when robo2 and the attractive ligand netrin or its receptor frazzled/DCC (fra) are removed in combination, and pan-neural misexpression of Robo2 can cause ectopic midline crossing. This activity is restricted to Robo2, as neither Robo nor Robo3 produces the same effect in gain of function assays, and neither robo nor robo3 can rescue the pro-crossing function of robo2 in fra,robo2 robo1 or fra, robo2robo3 mutants. No detect was detected in commissure formation after TcRobo2/3 knockdown in Tribolium embryos, and pan-neural misexpression of TcRobo2/3 did not produce ectopic midline crossing in Drosophila embryos. Two direct ways to test whether TcRobo2/3 shares Drosophila Robo2's pro-midline crossing activity would be to examine the effect of simultaneously knocking down TcFra and TcRobo2/3 on commissure formation in the Tribolium embryonic CNS, or by replacing Drosophila robo2 with TcRobo2/3 and asking whether it can promote commissure formation in a fra mutant background (Evans, 2012).

The coordinated targeting of axons and dendrites to specific regions within the three-dimensional network of neurites in the CNS is necessary for the proper assembly of motor and sensory circuits. In the Drosophila embryonic ventral nerve cord, the Slit/Robo and Semaphorin (Sema)/Plexin signaling pathways regulate mediolateral and dorsoventral positioning of neurites, respectively. Robo2 is required for the formation of longitudinal axon pathways in the lateral region of the neuropile, while Robo3 directs axons to join pathways in intermediate regions. Do the same mechanisms regulate longitudinal pathway choice in other animals? An antibody was generated against the Tribolium ortholog of FasII (TcFasII) in order to determine the number and location of FasII-positive axon pathways in the beetle embryonic CNS. Surprisingly, it was found that both the number and position of FasII-positive longitudinal pathways are conserved in the fly and beetle embryonic CNS, despite the fact that only two Robo receptors exist in Tribolium. Thus, in these two insect species, divergent genetic programs lead to equivalent developmental outcomes. Further, Drosophila's system of three Robo receptors specifying three medial-lateral zones in the neuropile is not typical of insects, as most insects (like Tribolium) have only two Robos (Evans, 2012).

Gene replacement experiments in Drosophila demonstrate that Robo2 and Robo3 are not functionally equivalent, yet a single Robo2/3 receptor performs both roles in Tribolium. How is this achieved? The data suggest that the ancestor of Robo2 and Robo3 was able to specify both intermediate and lateral pathway formation. Subsequent to the gene duplication that produced robo2 and robo3 in dipterans, Robo3 apparently lost the ability to promote lateral pathway formation. In contrast, Robo2 retains the ability to specify both intermediate and lateral pathways, but its expression was lost in neurons whose axons join the intermediate pathways. This model makes a number of testable predictions: TcRobo2/3 should be expressed on axons found in the intermediate and lateral zones of the axon scaffold in Tribolium embryos, similar to Robo3 expression in Drosophila; TcRobo2/3 should be able to substitute for both robo3 and robo2 to promote intermediate and lateral pathway formation in Drosophila; and Drosophila robo2, but not robo3, should be able to substitute for TcRobo2/3 to promote lateral pathway formation in Tribolium (Evans, 2012).

One possibility is that TcRobo2/3 expression in the neurons that make up the intermediate and lateral pathways may be differentially regulated, such that TcRobo2/3 is expressed at different levels or at different times in these two neuronal classes, and that these expression differences might be important for the sequential formation of distinct axon pathways. It is noted that embryogenesis in Tribolium takes around six times as long as Drosophila at 25°C, perhaps allowing temporal regulation of gene expression to play a more important role in specifying distinct guidance outcomes than in the rapidly developing fly embryo. Direct examination of the expression patterns of both beetle Robos will provide insight into how the regulation of Robo expression differs between Drosophila and Tribolium, and may also explain why the lateral position of dorsal and ventral lateral pathways exhibit a differential requirement for TcRobo2/3 activity (Evans, 2012).

This phylogenetic analysis indicates that the Robo family has expanded by gene duplication independently in a number of animal groups, including vertebrates, crustaceans, chelicerates and insects. In contrast, nematodes have retained a single Robo receptor. Notably, expansion of the Robo family by gene duplication does not appear to correlate with whole genome duplication events in insects, and does not appear to be common among insect axon guidance genes. Only a single ortholog of Slit was found in each of the analyzed insect and nematode genomes; similarly, the number of Plexin (two) and Frazzled/DCC (one) receptors is the same in all insects and C. elegans. The pattern of distribution of Robo orthologs together with sequence comparisons therefore suggests that a single Robo receptor was present in the last common ancestor of protostomes and deuterostomes. This ancestral Robo was almost certainly involved in Slit-dependent midline repulsion, but it is unclear whether it would have possessed any of the other known Robo axon guidance functions. An intriguing possibility is that additional activities were acquired by Robo paralogs subsequent to gene duplication in the lineages leading to vertebrates and insects, and that this functional diversification may have contributed to increases in nervous system complexity in these groups (Evans, 2012).

What molecular mechanisms have facilitated increases in nervous systemcomplexity over evolutionary time? The duplication and functional diversification of axon guidance pathway components is one possible mechanism providing increasingly precise control of axon guidance. This appears to be the case with the Robo family of axon guidance receptors, whose members exhibit functional diversity in a number of animal groups with complex nervous systems. Importantly, one key aspect of nervous system patterning and circuit formation in insects, specifying mediolateral targeting of neurites in the CNS neuropile, is controlled by Robo orthologs that appear to be restricted to insects (Robo2/3 and its descendants, Robo2 and Robo3). Correspondingly, insect nervous systems display a level of precision in axon guidance that is not present in close relatives outside the arthro. Elucidating the axon guidance roles of Robo receptors in additional insects and other animals (for example, crustaceans and onychophorans) should provide further insight into the origins of functional diversity in this important family of axon guidance molecules, and the role they have played in the evolution of animal nervous systems (Evans, 2012).

C. elegans Robo homolog

The C. elegans ventral nerve cord also has left and right longitudinal bundles, but unlike the vertebrate and Drosophila midlines, the C. elegans ventral nerve cord is asymmetric. Although most of the neurons that contribute longitudinal axons to the nerve cord have cell bodies in bilaterally symmetric pairs, the right ventral nerve cord contains about 40 axons, while the left ventral nerve cord in the central body contains only four axons. These two axon bundles are separated by an epidermal protrusion called the hypodermal ridge. Thus, most neurons whose cell bodies are on the left have axons that cross over to the right side, while axons from most cell bodies on the right remain ipsilateral. In C. elegans, midline crossing is only observed when an axon first joins the ventral nerve cord, usually at the anterior or posterior end of the nerve cord. In sax-3 mutants, most longitudinal axons in the ventral cord extend to their full length, but the left-right asymmetry of the ventral cord is disrupted so that many longitudinal axons extend on the incorrect side of the nerve cord. A single mutant axon can cross the midline many times along its trajectory. Thus, sax-3 acts to establish the asymmetry of the ventral nerve cord. In addition, sax-3 activity is required to guide axons to the ventral midline; in mutant animals, some ventral cord axons are found in aberrant lateral positions. These functions of sax-3 allow it to act in concert with the unc-6/netrin pathway to recruit axons to the ventral nerve cord. The sax-3 gene encodes a predicted transmembrane protein with five immunoglobulin domains and three fibronectin type III repeats that is closely related to Drosophila Robo (Zallen, 1998).

Multiple roles of sax-3 in axon guidance are seen for two serotonergic HSN motor neurons. In the wild type, each HSN motor neuron sends an axon ventrally to the midline and then anteriorly to the head without crossing the midline. In sax-3 mutants, the HSN axons are able to cross and recross the ventral midline. The HSN axon also has defects in the initial ventrally directed component of its outgrowth in sax-3 mutants. In wild-type animals, the HSN axon grows ventrally to the nerve cord either immediately at the cell body or shortly anterior to the cell body. In sax-3 mutants, a high proportion of HSN axons travel laterally for long distances before reaching the ventral nerve cord, either in an anterior direction or in an aberrant posterior direction. Thus sax-3 plays two roles in the guidance of the HSN motor neurons: it is required for growth along the epidermis to the ventral nerve cord and for selection of the ipsilateral nerve cord during anterior growth (Zallen, 1998).

Over half of the neurons in Caenorhabditis elegans send axons to the nerve ring, a large neuropil in the head of the animal. Genetic analysis of axon guidance in the nerve ring was initiated by focusing a subset of nerve ring sensory neurons. The primary nematode sensory organ, the amphid, consists of twelve bilaterally symmetric amphid neuron pairs that function to detect chemicals, touch and temperature. Each amphid neuron has a single axon that extends in the nerve ring and a ciliated sensory dendrite that connects to an opening in the anterior cuticle. Most amphid axons reach the nerve ring after first growing ventrally in the amphid commissure, while one pair projects directly to the ring in a lateral position. The navigation of sensory axons in the nerve ring is largely completed by the end of embryogenesis. However, nerve ring axons continue to grow as the animal grows during the larval stages. During larval development, nerve ring axons increase 2.5-fold in length, while retaining the spatial organization established during embryogenesis. In the nerve ring, amphid sensory axons contact sensory and interneuron targets, establishing neural circuits essential to sensation and behavior (Zallen, 1999 and references).

Genetic screens in animals that express the green fluorescent protein in a subset of sensory neurons have identified eight new sax (for sensory axon defects) genes that affect the morphology of nerve ring axons. To visualize nerve ring axons in living animals, the ceh-23 cell-specific promoter was used to express the green fluorescent protein (GFP). The ceh-23::gfp fusion was expressed in nine pairs of neurons in the head, including seven amphid neuron pairs, the BAG sensory neurons and the AIY interneurons, as well as the CAN neuron pair in the central body. The ceh-23::gfp transgene does not appear to disrupt the position, morphology or function of these neurons. sax-3/robo mutations disrupt axon guidance in the nerve ring, while sax-5, sax-9 and unc-44 disrupt both axon guidance and axon extension. Axon extension and guidance proceed normally in sax-1, sax-2, sax-6, sax-7 and sax-8 mutants, but these animals exhibit later defects in the maintenance of nerve ring structure. The functions of existing guidance genes in nerve ring development were also examined, revealing that SAX-3/Robo acts in parallel to the VAB-1/Eph receptor and the UNC-6/netrin, UNC-40/ DCC guidance systems for ventral guidance of axons in the amphid commissure, a major route of axon entry into the nerve ring. In addition, SAX-3/Robo and the VAB-1/ Eph receptor both function to prevent aberrant axon crossing at the ventral midline. Together, these genes define pathways required for axon growth, guidance and maintenance during nervous system development (Zallen, 1999).

The requirement for SAX-3/Robo in ventral guidance is consistent with two models: SAX-3 could act cell autonomously as a receptor on growing axons, or non-autonomously, either as a receptor in another pioneer axon or as a ligand on the substratum. Since amphid axons grow out in the amphid commissure bundle, their ventral guidance may be achieved through the combined action of axon-axon and axon-substratum interactions. Like amphid neurons, the AVM mechanosensory neuron has a cell body that is located on the lateral hypodermis and an axon that grows ventrally to the ventral midline; however, the AVM axon travels independently using only axon-substratum interactions. For this reason, the question of sax-3 autonomy was examined in the AVM neuron. AVM, like amphid neurons, relies on sax-3 for its ventral axon guidance as well as unc-6 and unc-40. To determine the site of sax-3 action for AVM ventral guidance, the sax-3 cDNA was expressed from the mec-7 promoter, which drives expression in six mechanosensory neurons, including AVM. The mec-7:: sax-3 fusion rescues the AVM defects of sax-3 mutants in two independent transgenic lines, indicating that SAX-3 can function cell autonomously in AVM ventral guidance. This result does not rule out the possibility that SAX-3 could behave non-autonomously in other cell contexts, such as the nerve ring (Zallen, 1999).

Several more highly related 5+3 proteins (proteins possessing 5 Ig domains and 3 FN domains) have been identified in vertebrates; these are likely to represent true Robo homologs. A human expressed sequence tag shows high homology to the second Ig domain of Robo and was used to probe a human fetal brain cDNA library. The clones recovered correspond to a human gene with five Ig and three FN domains. The homology is particularly high in the first two Ig domains (58% and 48% amino acid identity, respectively, compared to 26% and 30% for the same two Ig domains between Drosophila Robo1 and CDO). These data, together with the overall identity throughout the extracellular region and the presence of three conserved cytoplasmic motifs has led to the designation of this protein as the human roundabout 1 gene (H-robo1). Database searching reveals a nucleotide sequence nearly identical to H-robo1 in the database DUTT1. There are differences in the signal sequence, suggesting alternative splicing. There are also seven single base pair changes, presumably polymorphisms and a 9 bp insertion. Five ESTs show high sequence similarity to the cytoplasmic domain of H-robo1. Partial sequencing of cDNAs isolated using one of these ESTs as a probe confirms that there is a second human roundabout gene (H-robo2) (Kidd, 1998a).

Robo receptors interact with ligands of the Slit family. The nematode C. elegans has one Robo receptor (SAX-3) and one Slit protein (SLT-1), which act together to direct ventral axon guidance and guidance at the midline. In larvae, slt-1 expression in dorsal muscles repels axons to promote ventral guidance. SLT-1 acts through the SAX-3 receptor, in parallel with the ventral attractant UNC-6 (Netrin). Removing both UNC-6 and SLT-1 eliminates all ventral guidance information for some axons, revealing an underlying longitudinal guidance pathway. In the embryo, slt-1 is expressed at high levels in anterior epidermis. Embryonic expression of SLT-1 provides anterior-posterior guidance information to migrating CAN neurons. Surprisingly, slt-1 mutants do not exhibit the nerve ring and epithelial defects of sax-3 mutants, suggesting that SAX-3 has both Slit-dependent and Slit-independent functions in development (Hao, 2001).

The predicted SLT-1 protein consists of 1410 amino acids and shares a common domain structure and high sequence similarity with Drosophila Slit (41% amino acid identity) and all three mammalian Slit family members (39%-41% identity). All Slit proteins contain a putative signal sequence, four tandem arrays of leucine-rich repeats (LRRs), seven to nine EGF repeats, and a cysteine knot. C. elegans Slit has extensive similarity to Drosophila Slit and mammalian Slit2 in LRR-2 and LRR-4 (52%-60% identity); less similarity in LRR-1 and LRR-3 (25%-40% identity), and substantial similarity in all EGF repeats (28%-63% identity). C. elegans Slit and Drosophila Slit possess seven EGF repeats each, whereas nine EGF repeats are present in all three vertebrate Slits. The absence of hydrophobic anchor sequences in all Slits suggests that they encode large secreted proteins (Hao, 2001).

How are the trajectories of the twenty pharyngeal neurons of C. elegans are established? In this study focus was placed on the two bilateral M2 pharyngeal motorneurons, each of which has its cell body located in the posterior bulb and sends one axon through the isthmus and into the metacorpus. A GFP reporter was used to visualize these neurons in cell-autonomous and cell-non-autonomous axon guidance mutant backgrounds, as well as other mutant classes. The main findings are: (1) mutants with impaired growth cone functions, such as unc-6, unc-51, unc-73 and sax-3 (Drosophila homolog: Robo), often exhibit abnormal terminations and inappropriate trajectories at the distal ends of the M2 axons, i.e. within the metacorpus, and (2) growth cone function mutants never exhibit abnormalities in the proximal part of the M2 neuron trajectories, i.e. between the cell body and the metacorpus. These results suggest that the proximal and distal trajectories are established using distinct mechanisms, including a growth cone-independent process to establish the proximal trajectory. Five novel mutants were isolated in a screen for worms exhibiting abnormal morphology of the M2 neurons. These mutants define a new gene class designated mnm (M neuron morphology abnormal) (Mörcka, 2003).

Despite increasing evidence for transcriptional control of neural connectivity, how transcription factors regulate discrete steps in axon guidance remains obscure. Projection neurons in the dorsal spinal cord relay sensory signals to higher brain centers. Some projection neurons send their axons ipsilaterally, whereas others, commissural neurons, send axons contralaterally. This study shows that two closely related murine LIM homeodomain proteins, Lhx2 and Lhx9, are expressed by a set of commissural relay neurons (dI1c neurons) and are required for the dI1c axon projection. Midline crossing by dI1c axons is lost in Lhx2/9 double mutants, a defect that results from loss of expression of Rig-1 (Robo3) from dI1c axons. Lhx2 binds to a conserved motif (TAATTA) in the Rig-1 gene, suggesting that Lhx2/9 regulate directly the expression of Rig-1. These findings reveal a link between the transcriptional programs that define neuronal subtype identity and the expression of receptors that guide distinctive aspects of their trajectory (Wilson, 2008).

Whereas many molecules that promote cell and axonal growth cone migrations have been identified, few are known to inhibit these processes. In genetic screens designed to identify molecules that negatively regulate such migrations, CRML-1,the C. elegans homolog of CARMIL, also known as Lrrc16a, an actin-uncapping protein, was identified. Although mammalian CARMIL acts to promote the migration of glioblastoma cells, this study found that CRML-1 acts as a negative regulator of neuronal cell and axon growth cone migrations. Genetic evidence indicates that CRML-1 regulates these migrations by inhibiting the Rac GEF activity of UNC-73, a homolog of the Rac and Rho GEF Trio. The antagonistic effects of CRML-1 and UNC-73 can control the direction of growth cone migration by regulating the levels of the SAX-3 (a Robo homolog) guidance receptor. Consistent with the hypothesis that CRML-1 negatively regulates UNC-73 activity, these two proteins form a complex in vivo. Based on these observations, a role is proposed for CRML-1 as a novel regulator of cell and axon migrations that acts through inhibition of Rac signaling (Vanderzalm, 2009).

Changes in axon outgrowth patterns are often associated with synaptogenesis. Members of the conserved Pam/Highwire/RPM-1 protein family have essential functions in presynaptic differentiation. This study shows that C. elegans RPM-1 negatively regulates axon outgrowth mediated by the guidance receptors SAX-3/robo and UNC-5/UNC5. Loss-of-function rpm-1 mutations cause a failure to terminate axon outgrowth, resulting in an overextension of the longitudinal PLM axon. PLM overextension observed in rpm-1 mutants is suppressed by sax-3 and unc-5 loss-of-function mutations. PLM axon overextension is also induced by SAX-3 overexpression, and the length of extension is enhanced by loss of rpm-1 function or suppressed by loss of unc-5 function. Loss of rpm-1 function in genetic backgrounds sensitized for guidance defects disrupts ventral AVM axon guidance in a SAX-3-dependent manner and enhances dorsal guidance of DA and DB motor axons in an UNC-5-dependent manner. Loss of rpm-1 function alters expression of the green fluorescent protein (GFP)-tagged proteins, SAX-3::GFP and UNC-5::GFP. RPM-1 is known to regulate axon termination through two parallel genetic pathways; one involves the Rab GEF (guanine nucleotide exchange factor) GLO-4, which regulates vesicular trafficking, and another that involves the F-box protein FSN-1, which mediates RPM-1 ubiquitin ligase activity. glo-4 but not fsn-1 mutations affect axon guidance in a manner similar to loss of rpm-1 function. Together, the results suggest that RPM-1 regulates axon outgrowth affecting axon guidance and termination by controlling the trafficking of the UNC-5 and SAX-3 receptors to cell membranes (Li, 2008).

Although protein quality control (PQC) is generally perceived as important for the development of the nervous system, the specific mechanisms of neuronal PQC have remained poorly understood. This study reports that C. elegans Elongin BC-binding axon regulator (EBAX-1), a conserved BC-box protein, regulates axon guidance through PQC of the SAX-3/Robo receptor. EBAX-1 buffers guidance errors against temperature variations. As a substrate-recognition subunit in the Elongin BC-containing Cullin-RING ubiquitin ligase (CRL), EBAX-1 also binds to DAF-21, a cytosolic Hsp90 chaperone. The EBAX-type CRL and DAF-21 collaboratively regulate SAX-3-mediated axon pathfinding. Biochemical and imaging assays indicate that EBAX-1 specifically recognizes misfolded SAX-3 and promotes its degradation in vitro and in vivo. Importantly, vertebrate EBAX also shows substrate preference toward aberrant Robo3 implicated in horizontal gaze palsy with progressive scoliosis (HGPPS). Together, these findings demonstrate a triage PQC mechanism mediated by the EBAX-type CRL and DAF-21/Hsp90 that maintains the accuracy of neuronal wiring (Wang, 2013).

Robo and Ror function in a common receptor complex to regulate Wnt-mediated neurite outgrowth in Caenorhabditis elegans

Growing axons are exposed to various guidance cues en route to their targets, but the mechanisms that govern the response of growth cones to combinations of signals remain largely elusive. This study found that the sole Robo receptor, SAX-3 (see Drosophila Robo), in Caenorhabditis elegans functions as a coreceptor for Wnt/CWN-2 (Drosophila homolog: Wnt5) molecules. SAX-3 binds to Wnt/CWN-2 and facilitates the membrane recruitment of CWN-2. SAX-3 forms a complex with the Ror/CAM-1 receptor and its downstream effector Dsh/DSH-1, promoting signal transduction from Wnt to Dsh. sax-3 functions in Wnt-responsive cells and the SAX-3 receptor is restricted to the side of the cell from which the neurite is extended. DSH-1 has a similar asymmetric distribution, which is disrupted by sax-3 mutation. Taking these results together, it is proposed that Robo receptor can function as a Wnt coreceptor to regulate Wnt-mediated biological processes in vivo (Wang, 2018).

Robo function in planarians

The process by which the proper pattern is restored to newly formed tissues during metazoan regeneration remains an open question. This study provides evidence that the nervous system plays a role in regulating morphogenesis during anterior regeneration in the planarian Schmidtea mediterranea. RNA interference (RNAi) knockdown of a planarian ortholog of the axon-guidance receptor roundabout (robo) leads to unexpected phenotypes during anterior regeneration, including the development of a supernumerary pharynx (the feeding organ of the animal) and the production of ectopic, dorsal outgrowths with cephalic identity. Smed-roboA RNAi knockdown disrupts nervous system structure during cephalic regeneration: the newly regenerated brain and ventral nerve cords do not re-establish proper connections. These neural defects precede, and are correlated with, the development of ectopic structures. It is proposed that, in the absence of proper connectivity between the cephalic ganglia and the ventral nerve cords, neurally derived signals promote the differentiation of pharyngeal and cephalic structures. Together with previous studies on regeneration in annelids and amphibians, these results suggest a conserved role of the nervous system in pattern formation during blastema-based regeneration (Cebria, 2007).

Fish Robos

Roundabout (Robo) receptors and their secreted ligand Slits have been shown to function in a number of developmental events both inside and outside of the nervous system. Zebrafish robo orthologs have been cloned to gain a better understanding of Robo function in vertebrates. Further characterization of one of these orthologs, robo3, has unveiled the presence of two distinct isoforms, robo3 variant 1 (robo3var1) and robo3 variant 2 (robo3var2). These two isoforms differ only in their 5′-ends with robo3var1, but not robo3var2, containing a canonical signal sequence. Despite this difference, both forms accumulate on the cell surface. Both isoforms are contributed maternally and exhibit unique and dynamic gene expression patterns during development. Functional analysis of robo3 isoforms using an antisense gene knockdown strategy suggests that Robo3var1 functions in motor axon pathfinding, whereas Robo3var2 appears to function in dorsoventral cell fate specification. This study reveals a novel function for Robo receptors in specifying ventral cell fates during vertebrate development (Challa, 2005).

robo1, 2, and 3 isoforms in zebrafish are expressed both maternally and during early embryogenesis when the dorsoventral axis is forming. The early functions of Robo proteins have not been reported thus far. It is interesting that although both robo3 isoforms are expressed early, only Robo3var2 seems to effect early embryonic development. One explanation for this difference may be that Robo3var2, with its unique 5′ end lacking a classical signal peptide, functions differently than other Robo proteins. Robo1, 2, 3var1 may be able to substitute for each other, whereas Robo3var2 may perform a distinct function in early development. Further analysis of the mouse Rig-1, which also lacks a canonical signal peptide, may lend support to this hypothesis. slit genes are also expressed during gastrulation. slit2 is expressed in the anterior margin of the neural plate and overexpression of slit2 causes defects consistent with perturbations of convergent-extension movements during gastrulation. Dorsoventral cell fates were not examined in this study, thus it cannot be conclude that the actions of Robo3var2 in dorsoventral patterning are Slit dependent (Challa, 2005).

Members of the Slit family of secreted ligands interact with Roundabout (Robo) receptors to provide guidance cues for many cell types. For example, Slit/Robo signaling elicits repulsion of axons during neural development, whereas in endothelial cells this pathway inhibits or promotes angiogenesis depending on the cellular context. This study shows that duplicated miR-218 genes are intronically encoded in slit2 and slit3 and that the two mir-218 microRNAs suppresses Robo1 and Robo2 expression. The data indicate that miR-218 and multiple Slit/Robo signaling components are required for heart tube formation in zebrafish and that this network modulates the previously unappreciated function of Vegf signaling in this process. These findings suggest a new paradigm for microRNA-based control of ligand-receptor interactions and provide evidence for a novel signaling pathway regulating vertebrate heart tube assembly (Fish, 2011).

Distinct roles for Robo2 in the regulation of axon and dendrite growth by Xenopus retinal ganglion cells

Guidance factors act on the tip of a growing axon to direct it to its target. What role these molecules play, however, in the control of the dendrites that extend from that axon's cell body is poorly understood. Slits, through their Robo receptors, guide many types of axons, including those of retinal ganglion cells (RGCs). This study assesses and contrasts the role of Slit/Robo signalling in the growth and guidance of the axon and dendrites extended by RGCs in Xenopus laevis. As Xenopus RGCs extend dendrites, they express robo2 and robo3, while slit1 and slit2 are expressed in RGCs and in the adjacent inner nuclear layer. Interestingly, functional data with antisense knockdown and dominant negative forms of Robo2 (dnRobo2) and Robo3 (dnRobo3) indicate that Slit/Robo signalling has no role in RGC dendrite guidance, and instead is necessary to stimulate dendrite branching, primarily via Robo2. In vitro culture data argue that Slits are the ligands involved. In contrast, both dnRobo2 and dnRobo3 inhibited the extension of axons and caused the misrouting of some axons. Based on these data, it is proposed that Robo signalling can have distinct functions in the axon and dendrites of the same cell, and that the specific combinations of Robo receptors could underlie these differences. Slit acts via Robo2 in dendrites as a branching/growth factor but not in guidance, while Robo2 and Robo3 function in concert in axons to mediate axonal interactions and respond to Slits as guidance factors. These data underscore the likelihood that a limited number of extrinsic factors regulate the distinct morphologies of axons and dendrites (Hocking, 2010).

Mammalian Robo homologs

Degenerate PCR primers based on conserved sequences between H-robo1 and Drosophila robo1 were used to isolate a PCR fragment from a rat embryonic E13 spinal cord cDNA library. The fragment was used to probe an E13 spinal cord cDNA library, resulting in the isolation of a full-length rat robo gene (robo1). The predicted protein shows high sequence identity (>95%) with H-robo1 over the entire length. The 5' sequences of different rat robo1 cDNA clones suggest that this gene is alternatively spliced in a similar fashion to H-robo1/DUTTI. A similar approach was used to isolate rat robo2 cDNA clones. The mouse EST vi92e02 is highly homologous to the cytoplasmic portion of H-robo1. The C. elegans sax-3 gene is also a robo homolog (Zallen, 1998). A second Drosophila robo gene (D-robo2) is also predicted from analysis of genomic sequence in the public database. Taken together, these data indicate that Robo is the founding member of a novel Ig subfamily with at least one member in nematode, two in Drosophila, two in rat, and two in humans (Kidd, 1998a).

The isolation of several vertebrate Robo homologs suggests that Robo may play a role in orchestrating midline crossing in the vertebrate nervous system, similar to its role in Drosophila. In the vertebrate spinal cord, the ventral midline is comprised of a unique group of cells called the floor plate. As in the Drosophila nervous system, the vertebrate spinal cord contains both crossing and noncrossing axons. Spinal commissural neurons are born in the dorsal half of the spinal cord; commissural axons project to and cross the floor plate before turning longitudinally in a rostral direction. In contrast, the axons of two other classes of neurons, dorsal association neurons and ventral motor neurons, do not cross the floor plate. To address the possibility that Robo may play a role in organizing the projections of these spinal neurons, the expression of rat robo1 was examined by RNA in situ hybridization. At E13, when many commissural axons will have already extended across the floor plate, rat robo1 is expressed at high levels in the dorsal spinal cord in a pattern corresponding to the cell bodies of commissural neurons. Rat robo1 is also expressed at lower levels in a subpopulation of ventral cells in the region of the developing motor column. Interestingly, this expression pattern is similar to and overlaps partly with the mRNA encoding DCC (Drosophila homolog: Frazzled), another Ig superfamily member that is also expressed on commissural and motor neurons and encodes a receptor for Netrin-1 (Keino-Masu, 1996 ). Rat robo1 is not, however, expressed in either the floor plate or the roof plate of the spinal cord or in the dorsal root ganglia. This is in contrast to rat cdo, which is strongly expressed in the roof plate. Therefore, like its Drosophila homolog, rat roboI RNA appears to be expressed both by neurons with crossing axons and neurons with noncrossing axons, suggesting that it may encode the functional equivalent of Drosophila Robo1 (Kidd, 1998a).

Many neurons in both vertebrates and invertebrates innervate multiple targets by sprouting secondary axon collaterals (or branches) from a primary axon shaft. To begin to identify molecular regulators of axon branch initiation or extension, the growth of single sensory axons was studied in an in vitro collagen assay system and an activity was identified in extracts of embryonic spinal cord and of postnatal and adult brain that promotes the elongation and formation of extensive branches by these axons. Biochemical purification of the activity from calf brain extracts led to the identification of an amino-terminal fragment of Slit2 as the main active component and to the discovery of a distinct activity that potentiates its effects. These results indicate that Slit proteins may function as positive regulators of axon collateral formation during the establishment or remodeling of neural circuits (Wang, 1999).

To begin to address the in vivo function of Slit2, its sites of expression were examined by in situ hybridization at the time of sensory axon ingrowth in rats. At E13, a stage before sensory collaterals innervate the spinal cord, Slit2 mRNA is present mainly in floor plate and motor neurons. Then, starting from E14, its expression expands dorsally in a diffuse pattern. At E17, a stage when NGF-dependent sensory axon collaterals are growing into the dorsal spinal cord, Slit2 expression is highest in restricted regions of the ventral spinal cord, intermediate in level in the middle two-thirds, and present but lower in the dorsal-most regions. Expression of Slit2 is also detectable in a punctate pattern in the dorsal route ganglia (DRG) at E17, in contrast to its absence at earlier stages. The distributions of Slit1 and Slit3 mRNAs were also examined, since the other two Slit proteins might also affect sensory axons. Slit1 is expressed in dorsal spinal cord and the DRG at both E14 and, at higher levels, E17. Slit3 is expressed most highly in restricted regions of the ventral spinal cord at both E14 and E17, with lower level diffuse expression in DRG and little expression in dorsal spinal cord. The presence of Slit1 and Slit2 mRNAs in dorsal spinal cord raises the possibility that one or both of these proteins functions as a dorsal spinal cord-derived promoter of elongation or branching of sensory axons in that region. The presence of mRNAs for all three Slit genes in DRG also raises the possibility that one or more Slit proteins derived from DRG may function in autocrine/paracrine fashion. Robo proteins bind Slit proteins, and in Drosophila, Robo1 is implicated in mediating the repellent actions of Drosophila Slit. Robo2 is expressed at high levels in DRG at both E14 and E17, whereas Robo1 is not expressed at significant levels in DRG at these stages. Thus, if a Robo protein mediates the effects of Slit2-N, Robo2 would be the better candidate than Robo1 for the Slit2 receptor (Wang, 1999).

The purified amino-terminal fragment Slit2-N, but not the full-length molecule Slit2, possesses elongation- and branch-promoting activity, indicating that cleavage is important for bioactivity. It is possible that in the absence of cleavage the conformation of the N-terminal moiety is different and fails to lead to receptor binding or activation. Alternatively, the C-terminal fragment may have a dominant inhibitory activity on the N-terminal fragment when presented in cis within the molecule. The full-length molecule may actually function as an antagonist of Slit2-N. More quantitative studies involving mixing purified proteins will be required to address this issue. In addition, a full understanding of Slit2-N function will require identification of the receptor(s) on sensory axons that mediate its elongation- and branch-promoting activity. It is possible that Robo2, whose mRNA is expressed by sensory neurons, is involved in transducing the Slit2-N effect, either alone or in combination with a coreceptor. If so, it will be of interest to determine how Robo receptors elicit positive and negative responses in different neurons. It is also possible, however, that the positive effect of Slit2-N on sensory neurons is transduced by a receptor mechanism not involving Robo proteins (Wang, 1999).

The results raise the possibility that Slit2-N or the amino-terminal fragment of another Slit protein is involved in directing the formation, stabilization, or ingrowth of sensory axon collaterals into the dorsal spinal cord. Sensory axons do not project directly into the spinal cord, but rather first grow to the dorsal root entry zone and bifurcate, sending axons rostrally and caudally parallel to the spinal cord for several days. The ingrowth of sensory axon collaterals is preceded by the development of swellings ('nodes') along the entire intraspinal extent of sensory axons in the dorsal root entry zone, and collaterals develop interstitially from these nodes and project directly to their target laminae inside the gray matter of the spinal cord, never branching into inappropriate laminae en route. The amino-terminal fragment of one or more Slit proteins may play a causal role in directing this process either by stimulating the development of nodes or the subsequent stabilization or ingrowth of collaterals. Slit1 and Slit2 are both expressed in the dorsal spinal cord, increasing in expression level between E14 and E17, suggesting a possible involvement of either or both of these proteins. In addition, mRNAs for all three Slit proteins, but most prominently Slit1 and Slit2, are expressed in the dorsal route ganglia themselves at E17, suggesting that one or more sensory neuron-derived Slit protein could act in autocrine/paracrine fashion to contribute to these processes. In this case, however, the Slit proteins would then presumably not confer directionality on the branches. Perturbation of Slit protein function in vivo will be required to assess these possibilities (Wang, 1999).

Extending axons in the developing nervous system are guided in part by repulsive cues. Genetic analysis in Drosophila (Kidd, 1999) identifies the Slit protein as a candidate ligand for the repulsive guidance receptor Roundabout (Robo). The characterization of three mammalian Slit homologs is described and it is shown that the Drosophila Slit protein and at least one of the mammalian Slit proteins, Slit2, are proteolytically processed and show specific, high-affinity binding to Robo proteins. Furthermore, recombinant Slit2 can repel embryonic spinal motor axons in cell culture. These results support the hypothesis that Slit proteins have an evolutionarily conserved role in axon guidance as repulsive ligands for Robo receptors (Brose, 1999).

Several human and mouse expressed sequence tags have been identified that exhibited high homology to Drosophila Slit and these were used to probe both a human fetal brain library and an embryonic day 13 (E13) rat spinal cord library. The clones recovered corresponded to three distinct genes, all with high amino acid identity to Drosophila Slit. The human and rat cDNAs were designated hSlit1, -2, and -3, and rSlit1, -2, and -3, respectively. Recently, three other groups have also reported the cloning of human (hSlit1/MEGF4, hSlit2, and hSlit3/MEGF5) and rat (MEGF4 and MEGF5) slit family members have been identified. The nomenclature of Itoh (1998) was adopted for clarity. All mammalian Slit proteins share a common domain structure and high sequence homology with Drosophila Slit (43.5%, 44.3%, and 41.1% between dSlit and Slit1, -2, and -3, respectively), as well as with one another (60%-66% overall). All of the predicted Slit proteins contain a putative signal peptide, four tandem arrays of leucine rich repeats (LRRs) (which are flanked by conserved amino and carboxy-terminal sequences), a long stretch of EGF repeats, an Agrin-Laminin-Perlecan-Slit (ALPS) conserved spacer motif, and a cysteine knot (a dimerization motif found in several secreted growth factors). Like their Drosophila counterparts, the mammalian Slit proteins lack any hydrophobic sequences that might indicate a transmembrane domain and are thus predicted to encode secreted extracellular proteins. These conserved motifs are also found in a number of other proteins and have been implicated in mediating protein-protein and ECM-protein interactions. Slit1, -2, and -3 differ from Drosophila Slit in that they contain an additional LRR in the third tandem LRR array and two additional EGF repeats. A previously published Drosophila Slit sequence (Rothberg, 1990) lacks an LRR in the first tandem LRR array that is present in vertebrate Slits and in a novel Drosophila Slit sequence isolated by Kidd (1999). Interestingly, the C. elegans Slit homolog, like Drosophila Slit, also contains this additional LRR in LRR-1, has only seven EGF repeats, and is also lacking the extra LRR found in LRR-3 in vertebrates. This suggests that the addition of the EGF repeats occurred after the divergence of the chordate lineage from arthropods and nematodes, during metazoan evolution but prior to the triplication of the ancestral slit gene in the vertebrate lineage (Brose, 1999).

Expression of Slit1, Slit2, and Slit3 mRNAs in comparison to Robo1 and Robo2 mRNAs was determined by in situ hybridization in the developing rat spinal cord at embryonic stages E11-E13, i.e., when commissural axons are migrating to the midline. Rat Robo1 expression overlaps with that of Dcc and TAG-1, known markers for commissural (dorsal) and motor (ventral) neurons at these stages. Beginning at E11, rat Robo2 is also expressed ventrally in the region of the developing motoneuron cell bodies, but in a pattern distinct from that of rat Robo1. In contrast to rRobo1, rRobo2 is expressed in the dorsal root ganglia (DRG) by E13 but is not detected in the region of the commissural neurons overlapping with Dcc and TAG-1. Expression of rRobo2 is, however, detected dorsally in a lateral region of the spinal cord that may comprise the cell bodies of distinct subpopulations of commissural or association neurons (Brose, 1999).

Like Drosophila slit, which is expressed by cells at the midline of the Drosophila nervous system, the mRNAs for all three rat Slit proteins are expressed by floor plate cells at the ventral midline of the spinal cord. From E11-E13, Slit1 is expressed at high levels in the floor plate and at lower levels in other regions of the spinal cord, including the region of commissural and association neurons and the motor column. Likewise, Slit2 is also expressed by the floor plate, but in a different pattern: whereas Slit1 is expressed broadly by both floor plate cells and cells in the ventral portion of the ventricular zone, Slit2 is restricted to the most basal and medial region of the floor plate. Slit2 is expressed at high levels in the developing motor column and the roof plate, as well as very weakly in the region of the commissural neurons in the dorsal spinal cord. Like Slit1 and -2, Slit3 is also expressed by the floor plate but at significantly lower levels than Slit1 and Slit2. In addition, whereas the spatial distributions of Slit1 and Slit2 in the spinal cord remain largely constant between E11 and E13, Slit3 expression is dynamic: at E11 and E12 Slit3 is expressed in the motor column, but by E13, the stage at which most commissural axons have crossed the floor plate and have turned rostrally, Slit3 expression in the motor column is significantly diminished and expression is largely restricted to the floor plate. It is striking that, in addition to having specific additional sites of expression, the three mammalian Slit genes, like their counterpart in Drosophila, are all expressed by cells at the ventral midline of the nervous system. This expression is consistent with a potential role for Slit proteins in repulsive axon guidance at the midline mediated by Robo proteins (Brose, 1999).

hSlit2 protein can be detected on the surface of living cells with an antibody directed against either the N- or C-terminal tags, indicating that hSlit2 is secreted but remains associated with cell surfaces. Western blots of conditioned media and high salt (1 M NaCl) extracts of membranes from transfected cells reveal a band migrating at 190 kDa, which is slightly higher than the predicted size for hSlit2 and presumably reflects a glycosylated form of the protein. In addition to this 190 kDa isoform, two additional bands are detected: a 140 kDa protein that comprises the amino terminus of hSlit2 and a 55-60 kDa protein that comprises the carboxyl terminus. Since the molecular masses of these two proteins roughly add up to that of full-length hSlit2, it is presumed that they arise from proteolytic cleavage of the full-length protein. The amino- and carboxy-terminal fragments will be referred to as Slit2-N and Slit2-C, respectively. Full-length Slit2, Slit2-N, and Slit2-C exhibit different cell association characteristics. The majority (>90%) of the 190 kDa full-length hSlit2 is found associated with cell surfaces but can be readily extracted from membranes with either high salt (1 M NaCl) or heparin, suggesting its association with cell surfaces via heparan sulfate proteoglycans or other negatively charged moieties. hSlit2-C is more diffusible, partitioning roughly equally between the conditioned medium and cell surfaces. In contrast, hSlit2-N is largely absent from the conditioned medium and is found to be tightly cell associated and more resistant to heparin extraction, requiring several consecutive high salt washes to be fully released. Taken together, these results suggest that mammalian tissue culture cells contain one or more Slit-cleaving proteases, which appear to be absent from or inactive in Drosophila S2 cells. As observed for the mammalian protein, this processing of dSlit also seems to occur in vivo, as the antibody against dSlit detects two similarly sized bands of 190 kDa and 55-60 kDa on Western blots of Drosophila embryo extracts. To begin to identify the cleavage site, recombinant Slit2-C was purified and its amino terminus microsequenced by Edman degradation. The resulting sequence, TSPCDNFD, is found at the beginning of the 6th EGF repeat of hSlit2 and is consistent with the peptide sequences derived from microsequencing of the amino terminal fragment of bovine Slit2 purified from calf brain, all of which map to regions amino-terminal to this site. This sequence is at least partially conserved among Drosophila and mammalian Slit family members, suggesting that the cleavage site may also be conserved between insects and vertebrates. This sequence is not, however, well conserved in the C. elegans Slit homolog (Brose, 1999).

To test whether a mammalian Slit protein can act as a repellent, aggregates of hSlit2-expressing cells were cocultured either in contact with or at a distance from explants of ventral spinal cord from E11 rat embryos. Tissues were cultured in a matrix consisting of a mixture of collagen and matrigel (a partially purified extracellular matrix containing collagen and laminin). In this environment, unlike in a collagen matrix, motor axons grow out of the explants profusely, presumably stimulated by a component in matrigel. When ventral explants are cultured either at a distance (75-200 ┬Ám) from or in contact with aggregates of hSlit2-expressing cells, a clear repulsion of motor axons is observed. In many cultures, axons that are originally directed toward the COS cell aggregate appear to turn away. These results indicate that, at least under these culture conditions, hSlit2 can function as a diffusible repellent for developing motor axons (Brose, 1999).

The Slit gene encodes a secreted molecule essential for neural development in Drosophila embryos. Three Slit homologs have been identified in the mouse. The mouse SLIT1 protein can bind ROBO1, a transmembrane receptor implicated in axon guidance. Both whole-mount and section in situ hybridization studies reveal unique and complementary patterns of expression of the three mouse Slit genes and of Robo1, both within the central nervous system and in other developing tissues. The complementary expression patterns of Slit and Robo1 and their in vitro interaction suggest a ligand-receptor relationship. The expression of all three Slit genes in the floor plate suggests that they are likely to share the same functional properties with their Drosophila homolog in midline neural development and axon guidance. The complementary expression of Slit and Robo1 in different subdivisions of the somites suggests their possible function in axon pathfinding and neural crest cell migration. The unique expression pattern in limb and other organs indicates the potential for additional functions for the Slit gene family (Yuan, 1999).

In mammals, the environmental light/dark cycle strongly synchronizes the circadian clock within the suprachiasmatic nuclei (SCN) to 24 hr. It is well known that not only photic but also nonphotic stimuli can entrain the SCN clock. Actually, many studies have shown that a daytime injection of 8-hydroxy-2-(di-n-propylamino) tetralin (8-OH DPAT), a serotonin 1A/7 receptor agonist, as a nonphotic stimulus induces phase advances in hamster behavioral circadian rhythms in vivo, as well as the neuron activity rhythm of the SCN in vitro. Recent reports suggest that mammalian homologs of the Drosophila clock protein Period are involved in photic entrainment. Therefore, an examination was made to determine whether phase advances elicited by 8-OH DPAT are associated with a change of Period mRNA levels in the SCN. In this experiment, partial cDNAs were cloned encoding hamster Per1, Per2, and Per3 and both circadian oscillation and the light responsiveness of Period were observed. The inhibitory effect of 8-OH DPAT on hamster Per1 and Per2 mRNA levels in the SCN occurs only during the hamster's mid-subjective day, but not during the early subjective day or subjective night. The present findings demonstrate that the acute and circadian time-dependent reduction of Per1 and/or Per2 mRNA in the hamster SCN by 8-OH DPAT is strongly correlated with the phase resetting in response to 8-OH DPAT (Horikawa, 2000).

The ventral midline of the nervous system is an important choice point at which growing axons decide whether to cross and project contralaterally or remain on the same side of the brain. In Drosophila, the decision to cross or avoid the CNS midline is controlled, at least in part, by the Roundabout receptor on the axons and its ligand, Slit, an inhibitory extracellular matrix molecule secreted by the midline glia. Vertebrate homologs of these molecules have been cloned and have also been implicated in regulating axon guidance. Using in situ hybridization, the expression patterns have been determined of robo1,2 and slit1,2,3 in the mouse retina and in the region of the developing optic chiasm, a ventral midline structure in which retinal ganglion cell (RGC) axons diverge to either side of the brain. The receptors and ligands are expressed at the appropriate time and place, in both the retina and the ventral diencephalon, to be able to influence RGC axon guidance. In vitro, slit2 is inhibitory to RGC axons, with outgrowth of both ipsilaterally and contralaterally projecting axons being strongly affected. When presented alone in vitro, Slit2 does not have a differential effect on ipsilaterally and contralaterally projecting RGC axons; both are strongly inhibited. One possibility is that Slit2 alone is not enough to direct divergence but that additional factors are required. Other axon guidance molecules, such as Nr-CAM and Eph/ephrin receptors and ligands are expressed on the glial cells present at the ventral midline of the mouse diencephalon. In the future, it will be important to determine whether these molecules can synergize with Slit2 and thereby control RGC axon divergence at the midline. Overall, these results indicate that Robos and Slits alone do not directly control RGC axon divergence at the optic chiasm and may additionally function as a general inhibitory guidance system involved in determining the relative position of the optic chiasm at the ventral midline of the developing hypothalamus (Erskine, 2000).

In addition to acting as a chemorepellant for cortical axons, Slit1 regulates dendritic development. Slit1 is expressed in the developing cortex, and exposure to Slit1 leads to increased dendritic growth and branching. Conversely, inhibition of Slit-Robo interactions by Robo-Fc fusion proteins or by a dominant-negative Robo attenuates dendritic branching. Stimulation of neurons transfected with a Met-Robo chimeric receptor with Hepatocyte growth factor leads to a robust induction of dendritic growth and branching, suggesting that Robo-mediated signaling is sufficient to induce dendritic remodeling. These experiments indicate that Slit-Robo interactions may exert a significant influence over the specification of cortical neuron morphology by regulating both axon guidance and dendritic patterning (Whitford, 2002).

It is interesting to note that all three Slits have layer-restricted expression during postnatal development and in adulthood. The layer-specific expression during early postnatal development may be important in regulating layer-specific apical dendrite collateral branching or layer-specific axon terminations. For example, slit3 is expressed at high levels in most layers at P10 except layers IV and VI, which are the recipient layers for thalamic axons. It will be interesting to know whether a repulsive effect of Slit3 prevents innervation of layers II/III and V by these axons. It is also striking that all three slits as well as robos are expressed at high levels in the adult cortex, which suggests that Slit-Robo interactions must influence cellular events other than axon guidance and dendritic patterning. Although not much is known about how Slits are secreted, based on the in situ hybridization studies and the results of Robo-Fc blocking experiments in essentially pure neuronal cultures, it would appear that Slit proteins are secreted by neurons (Whitford, 2002).

To address how the highly stereotyped retinotectal pathway develops in zebrafish, fixed-tissue and time-lapse imaging was used to analyze morphology and behavior of wild-type and mutant retinal growth cones. Wild-type growth cones increase in complexity and pause at the midline. Intriguingly, they make occasional ipsilateral projections and other pathfinding errors, which are always eventually corrected. In the astray/robo2 mutant, growth cones are larger and more complex than wild-type. astray axons make midline errors not seen in wild-type, as well as errors both before and after the midline. astray errors are rarely corrected. The presumed Robo ligands Slit2 and Slit3 are expressed near the pathway in patterns consistent with their mediating pathfinding through Robo2. Thus, Robo2 does not control midline crossing of retinal axons, but rather shapes their pathway, by both preventing and correcting pathfinding errors (Hutson, 2002).

Commissural axons in vertebrates and insects are initially attracted to the nervous system midline, but once they reach this intermediate target they undergo a dramatic switch, becoming responsive to repellent Slit proteins at the midline, which expel them onto the next leg of their trajectory. A divergent member of the Robo family, Rig-1 (or Robo3), has been unexpectedly implicated in preventing premature Slit sensitivity in mammals. Expression of Rig-1 protein by commissural axons is inversely correlated with Slit sensitivity. Removal of Rig-1 results in a total failure of commissural axons to cross. Genetic and in vitro analyses indicate that Rig-1 functions to repress Slit responsiveness similarly to Commissureless (Comm) in Drosophila. Unlike Comm, however, Rig-1 does not produce its effect by downregulating Robo receptors on precrossing commissural axon membranes. These results identify a mechanism for regulating Slit repulsion that helps choreograph the precise switch from attraction to repulsion at a key intermediate axonal target (Sabatier, 2004).

FLRT3 is a Robo1-interacting protein that determines Netrin-1 attraction in developing axons

Guidance molecules are normally presented to cells in an overlapping fashion; however, little is known about how their signals are integrated to control the formation of neural circuits. In the thalamocortical system, the topographical sorting of distinct axonal subpopulations relies on the emergent cooperation between Slit1 and Netrin-1 guidance cues presented by intermediate cellular targets. However, the mechanism by which both cues interact to drive distinct axonal responses remains unknown. This study shows that the attractive response to the guidance cue Netrin-1 is controlled by Slit/Robo1 signaling and by FLRT3, a novel coreceptor for Robo1. While thalamic axons lacking FLRT3 are insensitive to Netrin-1, thalamic axons containing FLRT3 can modulate their Netrin-1 responsiveness in a context-dependent manner. In the presence of Slit1, both Robo1 and FLRT3 receptors are required to induce Netrin-1 attraction by the upregulation of surface DCC through the activation of protein kinase A. Finally, the absence of FLRT3 produces defects in axon guidance in vivo. These results highlight a novel mechanism by which interactions between limited numbers of axon guidance cues can multiply the responses in developing axons, as required for proper axonal tract formation in the mammalian brain (Leyva-Diaz, 2014).

Robo mutation in mammals

Chromosome 3 allele loss in preinvasive bronchial abnormalities and carcinogen-exposed, histologically normal bronchial epithelium indicates that it is an early, possibly the first, somatic genetic change in lung tumor development. Candidate tumor suppressor genes have been isolated from within distinct 3p regions implicated by heterozygous and homozygous allele loss. It is proposed that DUTT1, nested within homozygously deleted regions at 3p12-13, is the tumor suppressor gene that deletion-mapping and tumor suppression assays indicate is located in proximal 3p. The same gene, ROBO1, was independently isolated as the human homolog of the Drosophila gene, roundabout. The gene, coding for a receptor with a domain structure of the neural-cell adhesion molecule family, is widely expressed and has been implicated in the guidance and migration of axons, myoblasts, and leukocytes in vertebrates. A deleted form of the gene, which mimics a naturally occurring, tumor-associated human homozygous deletion of exon 2 of DUTT1/ROBO1, was introduced into the mouse germ line. Mice homozygous for this targeted mutation, which eliminates the first Ig domain of Dutt1/Robo1, frequently die at birth of respiratory failure because of delayed lung maturation. Lungs from these mice have reduced air spaces and increased mesenchyme, features that are present some days before birth. Survivors acquire extensive bronchial epithelial abnormalities including hyperplasia, providing evidence of a functional relationship between a 3p gene and the development of bronchial abnormalities associated with early lung cancer (Xian, 2001).

In Drosophila, Slit at the midline activates Robo receptors on commissural axons, thereby repelling them out of the midline into distinct longitudinal tracts on the contralateral side of the central nervous system. In the vertebrate spinal cord, Robo1 and Robo2 are expressed by commissural neurons, whereas all three Slit homologs are expressed at the ventral midline. Previous analysis of Slit1;Slit2 double mutant spinal cords failed to reveal a defect in commissural axon guidance. When all six Slit alleles are removed, many commissural axons fail to leave the midline, while others recross it. In addition, Robo1 and Robo2 single mutants show guidance defects that reveal a role for these two receptors in guiding commissural axons to different positions within the ventral and lateral funiculi. These results demonstrate a key role for Slit/Robo signaling in midline commissural axon guidance in vertebrates (Long, 2004).

Based on genetic studies in Drosophila and C. elegans and biochemical studies in vertebrates, receptors of the Robo family have been implicated in sensing Slit ligands as repellents. In vertebrates, several such receptors have been identified: Robo1, Robo2, and Rig-1. Rig-1 functions as an inhibitor of Slit responsiveness in commissural axons prior to crossing the floor plate. Since both Robo1 and Robo2 are also expressed by commissural neurons in the developing spinal cord, they are the primary candidate receptors for mediating repulsion by midline Slit proteins. Such a role is supported by the localization of both Robo1 and Robo2 to the postcrossing portion of commissural axons. Antibodies directed against the Robo1 extracellular domain appear to label both the ventral and lateral funiculi, in which commissural axons course longitudinally toward their final targets in the brain. Interestingly, Robo2-positive axons are found primarily in the lateral funiculus. Although some of these may be the axons of association neurons, it appears that many Robo2-expressing axons are commissural, as assessed by expression of the LacZ reporter in the ventral commissure under the floor plate when expressed from the Robo2 locus in Robo2 heterozygous animals. Based on the hypothesis that Robo receptors are required to expel commissural axons out of the floor plate once they have crossed, it is predicted that Robo mutants, like Slit triple mutants, would exhibit stalling or recrossing phenotypes. Indeed, in transverse sections of E11.5 Robo1 mutant embryos, as in the Slit triple mutants, L1-positive but TAG-1-negative axons are observed growing aberrantly into the dorsal region of the floor plate. As for the Slit triple mutant, the fact that these wandering axons express L1 but not TAG-1 suggests that these might be stalled or recrossing commissural axons. This conclusion is further strengthened by DiI analysis, which has revealed an increased number of stalled axons in the floor plate of Robo1 mutant E11.5 embryos, similar to but less penetrant than the stalling observed in Slit triple mutants. These results support a model in which Slit proteins in the floor plate mediate repulsion of commissural axons at least partly through the receptor Robo1 (Long, 2004).

This study provides evidence that Slit repellents in the floor plate act through Robo1 and Robo2 to guide commissural axons in the spinal cord. However, despite the severity of the Slit1;Slit2;Slit3 triple mutant phenotype, a significant number of commissural axons are observed that exhibit no obvious axon guidance phenotype. This is in contrast to what has previously been described in Drosophila, in which the removal of Slit leads to the collapse of both commissurally and longitudinally projecting axons into the ventral midline. This result suggests that in vertebrates, other repulsion systems beside Slit/Robo are involved in guiding commissural axons out of the floor plate and beyond. One likely system is provided by Semaphorins acting through Neuropilin receptors. Whereas in Drosophila Semaphorins are not required for commissural axon guidance in the CNS, in vertebrates Sema3B is expressed by floor plate cells and has been implicated in expelling postcrossing commissural axons from the midline via Neuropilin-2. Eph/Ephrin signaling may also contribute to guiding postcrossing axons, since several EphrinB proteins are expressed in the floor plate and dorsal spinal cord, and B class Eph receptors are expressed in the postcrossing segment of commissural axons. Despite this apparent redundancy between distinct repellent systems in the floor plate, removal of the Slit proteins is sufficient to severely disrupt midline axon guidance. The commissural axon phenotypes observed in the Slit triple mutant will serve as a useful baseline with which to compare future mutants where multiple repellent systems have been inactivated. These studies will be required to characterize the relative roles of each guidance system in directing commissural axons to leave the floor plate (Long, 2004).

During development, precerebellar neurons migrate dorsoventrally from the rhombic lip to the floor plate. Some of these neurons cross the midline while others stop. A role has been identified for the slit receptor Rig-1/Robo3 in directing this process. During their tangential migration, neurons of all major hindbrain precerebellar nuclei express high levels of Rig-1 mRNA. Rig-1 expression is rapidly downregulated as leading processes of these neurons cross the floor plate. Interestingly, most precerebellar nuclei do not develop normally in Rig-1-deficient mice, since they fail to cross the midline. In addition, inferior olivary neurons, which normally send axons into the contralateral cerebellum, project ipsilaterally in Rig-1 mutant mice. Similarly, neurons of the lateral reticular nucleus and basilar pons are unable to migrate across the floor plate and instead remain ipsilateral. These results demonstrate that Rig-1 controls the ability of both precerebellar neuron cell bodies and their axons to cross the midline (Marillat, 2004).

The mechanisms controlling axon guidance are of fundamental importance in understanding brain development. Growing corticospinal and somatosensory axons cross the midline in the medulla to reach their targets and thus form the basis of contralateral motor control and sensory input. The motor and sensory projections appear uncrossed in patients with horizontal gaze palsy with progressive scoliosis (HGPPS). Identified in patients affected with HGPPS, were mutations in the ROBO3 gene, which shares homology with roundabout genes important in axon guidance in developing Drosophila, zebrafish, and mouse. Like its murine homolog Rig1/Robo3, but unlike other Robo proteins, ROBO3 is required for hindbrain axon midline crossing (Jen, 2004).

Dyslexia, or specific reading disability, is the most common learning disorder with a complex, partially genetic basis, but its biochemical mechanisms remain poorly understood. A locus on Chromosome 3, DYX5, has been linked to dyslexia in one large family and speech-sound disorder in a subset of small families. The axon guidance receptor gene ROBO1, orthologous to the Drosophila roundabout gene, is disrupted by a chromosome translocation in a dyslexic individual. In a large pedigree with 21 dyslexic individuals genetically linked to a specific haplotype of ROBO, the expression of ROBO1 from this haplotype is absent or attenuated in affected individuals. Sequencing of ROBO1 in apes revealed multiple coding differences, and the selection pressure is significantly different between the human, chimpanzee, and gorilla branch as compared to orangutan. Novel exons and splice variants of ROBO1 were detected that may explain the apparent phenotypic differences between human and mouse in heterozygous loss of ROBO1. It is concluded that dyslexia may be caused by partial haplo-insufficiency for ROBO1 in rare families. Thus, the data suggest that a slight disturbance in neuronal axon crossing across the midline between brain hemispheres, dendrite guidance, or another function of ROBO1 may manifest as a specific reading disability in humans (Hannula-Jouppi, 2005).

To investigate Slit signalling in forebrain development, Robo1 knockout mice were generated by targeted deletion of exon 5 of the Robo1 gene. Homozygote knockout mice died at birth, but prenatally displayed major defects in axon pathfinding and cortical interneuron migration. Axon pathfinding defects included dysgenesis of the corpus callosum and hippocampal commissure, and abnormalities in corticothalamic and thalamocortical targeting. Slit2 and Slit1/2 double mutants display malformations in callosal development, and in corticothalamic and thalamocortical targeting, as well as optic tract defects. In these animals, corticothalamic axons form large fasciculated bundles that aberrantly cross the midline at the level of the hippocampal and anterior commissures, and more caudally at the medial preoptic area. Such phenotypes of corticothalamic targeting were not observed in Robo1 knockout mice but, instead, both corticothalamic and thalamocortical axons aberrantly arrived at their respective targets at least 1 day earlier than controls. By contrast, in Slit mutants, fewer thalamic axons actually arrive in the cortex during development. Finally, significantly more interneurons (up to twice as many at E12.5 and E15.5) migrated into the cortex of Robo1 knockout mice, particularly in both rostral and parietal regions, but not caudal cortex. These results indicate that Robo1 mutants have distinct phenotypes, some of which are different from those described in Slit mutants, suggesting that additional ligands, receptors or receptor partners are likely to be involved in Slit/Robo signalling (Andrews, 2006).

Upon arriving at their targets, developing axons cease pathfinding and begin instead to arborize and form synapses. To test whether CNS arborization and synaptogenesis are controlled by Slit-Robo signaling, single retinal ganglion cell (RGC) arbors were followed over time. ast (robo2) mutant and slit1a morphant arbors had more branch tips and greater arbor area and complexity compared to wild-type and concomitantly more presumptive presynaptic sites labeled with YFP-Rab3. Increased arborization in ast was phenocopied by dominant-negative Robo2 expressed in single RGCs and rescued by full-length Robo2, indicating that Robo2 acts cell-autonomously. Time-lapse imaging revealed that ast and slit1a morphant arbors stabilized earlier than wild-type, suggesting a role for Slit-Robo signaling in preventing arbor maturation. Genetic analysis showed that Slit1a acts both through Robo2 and Robo2-independent mechanisms. Unlike previous PNS studies showing that Slits promote branching, these results show that Slits inhibit arborization and synaptogenesis in the CNS (Campbell, 2007).

Pioneer longitudinal axons navigate using floor plate and Slit/Robo signals

Longitudinal axons transmit all signals between the brain and spinal cord. Their axon tracts through the brain stem are established by a simple set of pioneer axons with precise trajectories parallel to the floor plate. To identify longitudinal guidance mechanisms in vivo, the overall role of floor plate tissue and the specific roles of Slit/Robo signals were tested. Ectopic induction or genetic deletion of the floor plate was shown to divert longitudinal axons into abnormal trajectories. The expression patterns of the diffusible cues of the Slit family were altered in the floor plate experiments, suggesting their involvement in longitudinal guidance. Genetic tests of Slit1 and Slit2, and the Slit receptors Robo1 and Robo2 were carried out in mutant mice. Slit1;Slit2 double mutants had severe longitudinal errors, particularly for ventral axons, including midline crossing and wandering longitudinal trajectories. Robo1 and Robo2 were largely genetically redundant, and neither appeared to specify specific tract positions. However, combined Robo1 and Robo2 mutations strongly disrupted each pioneer tract. Thus, pioneer axons depend on long-range floor plate cues, with Slit/Robo signaling required for precise longitudinal trajectories (Farmer, 2008).

Alternative splicing of the Robo3 axon guidance receptor governs the midline switch from attraction to repulsion

Alternative splicing provides a means to increase the complexity of gene function in numerous biological processes, including nervous system wiring. Navigating axons switch responses from attraction to repulsion at intermediate targets, allowing them to grow to each intermediate target and then to move on. The mechanisms underlying this switch remain poorly characterized. The Slit receptor Robo3 is required for spinal commissural axons to enter and cross the midline intermediate target. Two functionally antagonistic isoforms of Robo3 possess distinct carboxy termini arising from alternative splicing. Robo3.1 is deployed on the precrossing and crossing portions of commissural axons and allows midline crossing by silencing Slit repulsion. Robo3.2 becomes expressed on the postcrossing portion and blocks midline recrossing, favoring Slit repulsion. The tight spatial regulation of opponent splice variants helps ensure high-fidelity transition of axonal responses from attraction to repulsion at the midline (Chen, 2008).

It is intriguing to consider how the two Robo3 isoforms are confined to appropriate axonal segments. The simplest mechanism would be for there to be a temporal switch in splicing, with the Robo3.1 transcript made first, followed at an appropriate time by the Robo3.2 transcript. However, by quantitative RT-PCR, a roughly constant ratio was found of the two transcripts in the spinal cord during the period of commissural axon growth to and across the midline. In situ hybridization studies similarly suggest that the two transcripts appear in parallel in commissural neurons, arguing against a temporal switch (Chen, 2008).

Alternative mechanisms to a temporal switch include one or more of the following: (1) constitutive translation of the two transcripts followed by selective trafficking to (or removal from) the two axon segments, (2) a temporal switch in translation of the two transcripts in the cell bodies (with Robo3.1 translated while the axon is extending to the midline and Robo3.2 after it has crossed), (3) transport of the transcripts into axons followed by selective translation of Robo3.1 in the precrossing axon segment and Robo3.2 in the postcrossing segment. As a variation on the third possibility, the Robo3.2 transcript could be transported down the axons but spliced selectively in the precrossing portion to yield Robo3.1, followed by local translation (this mechanism would require splicing within the axons, which has not so far been demonstrated). At present, it is not possible to distinguish fully between these mechanisms, although the absence of detectable transcripts in the axons tends to argue against models involving local translation (Chen, 2008).

Dendrite self-avoidance requires cell-autonomous slit/robo signaling in cerebellar purkinje cells

Dendrites from the same neuron usually develop nonoverlapping patterns by self-avoidance, a process requiring contact-dependent recognition and repulsion. Recent studies have implicated homophilic interactions of cell surface molecules, including Dscams and Pcdhgs, in self-recognition, but repulsive molecular mechanisms remain obscure. This study report a role for the secreted molecule Slit2 and its receptor Robo2 in self-avoidance of cerebellar Purkinje cells (PCs). Both molecules are highly expressed by PCs, and their deletion leads to excessive dendrite self-crossing without affecting arbor size and shape. This cell-autonomous function is supported by the boundary-establishing activity of Slit in culture and the phenotype rescue by membrane-associated Slit2 activities. Furthermore, genetic studies show that they act independently from Pcdhg-mediated recognition. Finally, PC-specific deletion of Robo2 is associated with motor behavior alterations. Thus, this study uncovers a local repulsive mechanism required for self-avoidance and demonstrates the molecular complexity at the cell surface in dendritic patterning (Gibson, 2014).

Cross GTPase-activating protein (CrossGAP)/Vilse links the Roundabout receptor to Rac to regulate midline repulsion

The regulators of the Rho-family GTPases, GTPase-activating proteins (GAPs) and guanine exchange factors (GEFs), play important roles in axon guidance. By means of a functional genomic study of the Rho-family GEFs and GAPs in Drosophila, a Rho-family GAP, CrossGAP (CrGAP), has been identified that is involved in Roundabout (Robo) receptor-mediated repulsive axon guidance. CrGAP physically associates with the Robo receptor. Too much or too little CrGAP activity leads to defects in Robo-mediated repulsion at the midline choice point. The CrGAP gain-of-function phenotype mimics the loss-of-function phenotypes of both Robo and Rac. Dosage-sensitive genetic interactions among CrGAP, Robo, and Rac support a model in which CrGAP transduces signals downstream of Robo receptor to regulate Rac-dependent cytoskeletal changes (Hu, 2005; full text of article).

Gbx2 regulates thalamocortical axon guidance by modifying the LIM and Robo codes

Combinatorial expression of transcription factors forms transcriptional codes to confer neuronal identities and connectivity. However, how these intrinsic factors orchestrate the spatiotemporal expression of guidance molecules to dictate the responsiveness of axons to guidance cues is less understood. Thalamocortical axons (TCAs) represent the major input to the neocortex and modulate cognitive functions, consciousness and alertness. TCAs travel a long distance and make multiple target choices en route to the cortex. The homeodomain transcription factor Gbx2 is essential for TCA development, as loss of Gbx2 abolishes TCAs in mice. Using a novel TCA-specific reporter, this study has discovered that thalamic axons are mostly misrouted to the ventral midbrain and dorsal midline of the diencephalon in Gbx2-deficient mice. Furthermore, conditionally deleting Gbx2 at different embryonic stages has revealed a sustained role of Gbx2 in regulating TCA navigation and targeting. Using explant culture and mosaic analyses, it was demonstrated that Gbx2 controls the intrinsic responsiveness of TCAs to guidance cues. The guidance defects of Gbx2-deficient TCAs are associated with abnormal expression of guidance receptors Robo1 and Robo2. Finally, Gbx2 was demonstrated to control Robo expression by regulating LIM-domain transcription factors through three different mechanisms: Gbx2 and Lhx2 compete for binding to the Lmo3 promoter and exert opposing effects on its transcription; repressing Lmo3 by Gbx2 is essential for Lhx2 activity to induce Robo2; and Gbx2 represses Lhx9 transcription, which in turn induces Robo1. These findings illustrate the transcriptional control of differential expression of Robo1 and Robo2, which may play an important role in establishing the topography of TCAs (Chatterjee, 2012).

Robo1 and 2 Repellent Receptors Cooperate to Guide Facial Neuron Cell Migration and Axon Projections in the Embryonic Mouse Hindbrain

The facial nerve is necessary for our ability to eat, speak, and make facial expressions. Both the axons and cell bodies of the facial nerve undergo a complex embryonic developmental pattern involving migration of the cell bodies caudally and tangentially through rhombomeres, and simultaneously the axons projecting to exit the hindbrain to form the facial nerve. The goal of this study was to test the functions of the chemorepulsive receptors Robo1 and Robo2 (see Drosophila Robo) in facial neuron migration and axon projection by analyzing genetically marked motor neurons in double-mutant mouse embryos through the migration time course, E10.0-E13.5. In Robo1/2 double mutants, axon projection and cell body migration errors were more severe than in single mutants. Most axons did not make it to their motor exit point, and instead projected into and longitudinally within the floor plate. Surprisingly, some facial neurons had multiple axons exiting and projecting into the floor plate. At the same time, a subset of mutant facial cell bodies failed to migrate caudally, and instead either streamed dorsally toward the exit point or shifted into the floor plate. It is concluded that Robo1 and Robo2 have redundant functions to guide multiple aspects of the complex cell migration of the facial nucleus, as well as regulating axon trajectories and suppressing formation of ectopic axons (Gruner, 2019).

Non-cell autonomous control of precerebellar neuron migration by Slit and Robo proteins

During development, precerebellar neurons migrate tangentially from the dorsal hindbrain to the floor plate. Their axons cross it but their cell bodies stop their ventral migration upon reaching the midline. It has previously been shown that Slit chemorepellents and their receptors, Robo1 and Robo2 (see Drosophila Robo), might control the migration of precerebellar neurons in a repulsive manner. This study used a conditional knockout strategy in mice to test this hypothesis. The targeted inactivation of the expression of Robo1 and Robo2 receptors in precerebellar neurons does not perturb their migration, and they still stop at the midline. The selective ablation of the expression of all three Slit proteins in floor-plate cells has no effect on pontine neurons and only induces the migration of a small subset of inferior olivary neurons across the floor plate. Likewise, this study shows that the expression of Slit proteins in the facial nucleus is dispensable for pontine neuron migration. Together, these results show that Robo1 and Robo2 receptors act non-cell autonomously in migrating precerebellar neurons and that floor-plate signals, other than Slit proteins, must exist to prevent midline crossing (Dominici, 2020).

Structural Principles in Robo Activation and Auto-inhibition

Proper brain function requires high-precision neuronal expansion and wiring, processes controlled by the transmembrane Roundabout (Robo) receptor family and their Slit ligands (see Drosophila Robo and Slit). Despite their great importance, the molecular mechanism by which Robos' switch from "off" to "on" states remains unclear. This study reports a 3.6 Å crystal structure of the intact human Robo2 ectodomain (domains D1-8). Robo cis dimerization via D4 is conserved through hRobo1, 2, and 3 and the C. elegans homolog SAX-3 and is essential for SAX-3 function in vivo. The structure reveals two levels of auto-inhibition that prevent premature activation: (1) cis blocking of the D4 dimerization interface and (2) trans interactions between opposing Robo receptors that fasten the D4-blocked conformation. Complementary experiments in mouse primary neurons and C. elegans support the auto-inhibition model. These results suggest that Slit stimulation primarily drives the release of Robo auto-inhibition required for dimerization and activation (Barak, 2019).

Slit and Roundabout involvement in olfactory sensory neuron and olfactory bulb axon guidance

Continued: roundabout Evolutionary homologs part 2/2

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

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

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