Neuroglian

DEVELOPMENTAL BIOLOGY

Embryonic

As early as six hours, the long alternatively spliced form of Neuroglian is expressed on the surface of specific CNS and PNS neurons, and a few PNS support cells. The longitudinal strip of expression found by seven hours in the CNS prefigures the location where the longitudinal axons will form, and appears even before the longitudinal glia have migrated into this same position (Hortsch, 1990).

Between 11 and 12 hours of development, the short alternatively spliced form of Neuroglian is expressed in glia, and a variety of non-neuronal tissues, including trachea, hindgut, salivary gland and muscle.

The longitudinal glia (LG), progeny of a single glioblast, form a scaffold that presages the formation of longitudinal tracts in the ventral nerve cord (VNC) of the Drosophila embryo. The LG are used as a substrate during the extension of the first axons of the longitudinal tract. The differentiation of the LG has been examined in six mutations in which the longitudinal tracts are absent, displaced, or interrupted to determine whether the axon tract malformations may be attributable to disruptions in the LG scaffold. Embryos mutant for the gene prospero have no longitudinal tracts, and glial differentiation remains arrested at a preaxonogenic state. Two mutants of the Polycomb group also lacked longitudinal tracts; here the glia fail to form an oriented scaffold, but cytological differentiation of the LG is unperturbed. The longitudinal tracts in embryos mutant for slit fuse at the VNC midline and scaffold formation is normal, except that it is medially displaced. Longitudinal tracts have intersegmental interruptions in embryos mutant for hindsight and midline. In hindsight, there are intersegmental gaps in the glial scaffold. In midline, the glial scaffold retracts after initial extension. LG morphogenesis during axonogenesis is abnormal in midline. Commitment to glial identity and glial differentiation also occurs before scaffold formation. In all mutants examined, the early distribution of the glycoprotein Neuroglian is perturbed. This is indicative of early alterations in VNC pattern present before LG scaffold formation begins. Therefore, some changes in scaffold formation may reflect changes in the placement and differentiation of other cells of the VNC. In all mutants, alterations in scaffold formation precedes longitudinal axon tract formation (Jacobs, 1993).

The lack of widespread axonal defects in the CNS of neurotactin mutants suggests that the function of Nrt in CNS morphogenesis might be largely replaced by functionally related molecules. If so, embryos lacking Nrt as well as one of these other molecules may display synergistic mutant phenotypes. To test this possibility, embryos lacking function of both nrt and one of several genes encoding neural CAMs were examined. Embryos of some double mutant combinations of neurotactin and other genes encoding adhesion/signaling molecules, including neuroglian, derailed , and kekkon1, display phenotypic synergy. This result provides evidence for functional cooperativity in vivo between the adhesion and signaling pathways controlled by neurotactin and the other three genes (Speicher, 1998).

Neuroglian (Nrg) is a Drosophila neural CAM related to several vertebrate CAMs, though most closely to mouse L1. Two forms of Nrg that differ in their cytoplasmic domains and patterns of expression are known . The long Nrg isoform is neural-specific; it is initially (early stage 12) found in a fraction of CNS neurons, but during stage 13 it can be detected in most (and probably all) differentiating neurons. The short Nrg isoform is expressed by glia, is widely expressed in other tissues, and is probably expressed throughout the entire CNS. nrg1, a loss-of-function mutation for both Nrg forms, is lethal and causes motor neuron pathfinding defects, but the overall CNS structure of mutant embryos looks normal. Furthermore, unlike nrt5 embryos, no defects are detected with mAb 1D4 in nrg1 embryos. In contrast, nrg1; nrt5 double mutant embryos have a severe CNS phenotype. With mAb BP102, thinning or complete interruption of longitudinal connectives, as well as fusion of commissures are observed. Fas II fascicles exhibit similar abnormalities as those observed in nrt5 embryos, albeit with a much higher expressivity and penetrance. Most notably, interruptions of the longitudinal axon bundles are frequent, as are misguidance phenotypes. Double mutant embryos, like single nrt- embryos, also show a local constriction of the ventral nerve cord with a variable expressivity. This defect may be a consequence of the impaired axogenesis and condensation of the nerve cord. No defects outside the CNS are evident in the double mutants (Speicher, 1998).

Using mAb 22C10 (see Futsch), which recognizes a subset of neurons, and mAb 1D4, the behavior of several identified pioneer axons were examined during early stages of axogenesis in nrg1;nrt5 embryos.The pioneer axon of the intersegmental nerve, aCC, as well as the pioneer axon of the segmental nerve, establish their correct pathways. Likewise, the axons of the U neurons follow the aCC pathway correctly. In contrast, in 37 of 128 cases (29%), the axons of the dMP2 and MP1 neurons, pioneers of the MP1 pathway, do not normally defasciculate from the aCC axon and turn to the posterior; instead, they either become and remain stalled or they delay their extension for a considerable time. Other axons showing misguidance phenotypes are those of the six ventral unpaired medial (VUM) neurons. In the wild type, the VUM axons initially fasciculate together before splitting into two fascicles that grow laterally on either side of the midline, passing the RP2 neuron and fasciculating with the corresponding anterior aCC axon. In 19 of 128 (15%) double mutant segments, the fascicle of VUM axons either does not split or splits into more than two fascicles, each joining a different aCC axon, including that of the same hemisegment. The first two axons of the vMP2 pathway, pCC (the pioneer) and vMP2, grow correctly in most hemisegments; only in 4 of 128 cases was a misrouted vMP2 axon observed. Anomalies in the trajectory of the SP1 axon are also observed, though rarely. nrg1; nrt5 embryos also display, due to slight mispositioning of cells, a somewhat irregular appearance of what is normally a highly stereotyped pattern of neurons. However, the relative positions of neurons are maintained (Speicher, 1998).

It seems most likely that the phenotypes of nrg1; nrt5 embryos result from a direct requirement for these two CAMs during axogenesis, and not as a secondary consequence of a previous requirement during neurogenesis. Thus, expression of the nuclear proteins Eve, Ftz, and En, markers of the specification of subsets of neurons that are arranged in characteristic patterns, is found to be normal in nrg1; nrt5 embryos between stages 12 and 16. This suggests that a failure of proper cell fate determination does not cause the axonal mutant phenotype. Likewise, glial cells expressing Repo, a specific marker for most of the CNS glia, form at the correct time and place and in normal numbers in nrg1; nrt5 embryos. The longitudinal glia (LG), which could provide a matrix for longitudinal axon extension, migrate and arrange normally in the double mutant, prefiguring the longitudinal connectives. It is from stage 14 onward, when the LG normally stretch in the anterior-posterior direction and enwrap the connectives, that gaps in the LG begin to appear, overlapping with gaps in the connectives. It is most likely, therefore, that this LG phenotype in late mutant embryos is a consequence, rather than the origin, of the interruptions observed along the axonal connectives (Speicher, 1998).

GPCR signaling is required for blood-brain barrier formation in Drosophila

The blood-brain barrier of Drosophila is established by surface glia, which ensheath the nerve cord and insulate it against the potassium-rich hemolymph by forming intercellular septate junctions. The mechanisms underlying the formation of this barrier remain obscure. The G protein-coupled receptor (GPCR) Moody, the G protein subunits Gαi and Galphao, and the regulator of G protein signaling Loco are required in the surface glia to achieve effective insulation. The data suggest that the four proteins act in a complex common pathway. At the cellular level, the components function by regulating the cortical actin and thereby stabilizing the extended morphology of the surface glia, which in turn is necessary for the formation of septate junctions of sufficient length to achieve proper sealing of the nerve cord. This study demonstrates the importance of morphogenetic regulation in blood-brain barrier development and places GPCR signaling at its core (Schwabe, 2005).

The Drosophila nerve cord is ensheathed by a thin single-layer epithelium, which in turn is surrounded by an acellular layer of extracellular matrix material. Ultrastructural analysis has revealed that septate junctions (SJs) between the epithelial cells are responsible for the insulation of the nerve cord. Fate-mapping studies have shown that the nerve cord is enveloped by glia expressing the glial-specific marker Repo, but to date there has been no direct proof that it is these surface glia that form intercellular SJs and thus the insulating sheath. Moreover, the time course for the formation of the sheath and of the SJ-mediated seal has not been established (Schwabe, 2005).

Several assays were developed to follow the morphogenesis of the surface glial sheath. Due to the onset of cuticle formation, immunohistochemistry becomes unreliable after 16 hr of development. Live imaging of GFP-tagged marker proteins was therefore used to visualize cell shapes, in particular the actin cytoskeleton marker GFP/RFP-Moesin and the SJ marker Neuroglian (Nrg)-GFP. Nrg-GFP expressed under its own promoter and RFP-Moesin driven by repo-Gal4 are colocalized in the same cells, establishing that the SJ-forming cells are repo positive and thus conclusively demonstrating the insulating function of the surface glia. To probe the permeability of the transcellular barrier, fluorescent dye was injected into the body cavity and dye penetration into the nerve cord was quantified by determining mean pixel intensity in sample sections (Schwabe, 2005).

The surface glia are born in the ventrolateral neuroectoderm and migrate to the surface of the developing nerve cord, where they spread until they touch their neighbors (17 hr of development). The glia then join to form a contiguous sheet of square or trapezoidal cells, tiled to form three-cell corners. SJ material is visible as a thin contiguous belt by 18 hr but continues to accumulate until the end of embryogenesis. Similar to other secondary epithelia, the surface glia do not form a contiguous adherens-junction belt (zonula adherens), but only spotty, inconsistent adherens junctions were seen, as visualized by Armadillo-GFP (driven by own promoter). At 16 hr, the fluorescent dye freely penetrates into the nerve cord, but by 20 hr the nerve cord is completely sealed. The completion of the seal thus coincides with the onset of visible movements in the late embryo (Schwabe, 2005).

To further gauge the dye-penetration assay, embryos mutant for known septate-junction components were examined: Neurexin IV, which is required for blood-nerve barrier formation in the PNS, Neuroglian, and the sodium-pump component Nervana2, for which only a role in the earlier formation of the ectodermal seal has been demonstrated. In all three mutants, severe penetration of dye was found, well after the nerve cord is sealed in wild-type (22 hr). These findings provide further evidence that the sealing of the nerve cord is achieved by SJs and suggest that the components of the ectodermal SJs are required for the function of surface glial SJs as well (Schwabe, 2005).

In a genome-wide screen for glial genes, using FAC sorting of GFP-labeled embryonic glia and Affymetrix microarray expression analysis, two novel GPCRs, Moody (CG4322) and Tre1 (CG3171: Trapped in endoderm-1) were identified. Both are orphan receptors belonging to the same novel subclass of Rhodopsin-family GPCRs. Their expression was examined by RNA in situ hybridization; different subtypes of glia in the embryonic nerve cord can be distinguished based on their position and morphology. In the CNS, moody is expressed in surface glia from embryonic stage 13 onward (10 hr); in addition to cells surrounding the nerve cord (subperineurial glia), this includes cells lining the dorsoventral channels (channel glia). moody is also expressed in the ensheathing glia of the PNS (exit and peripheral glia). Both CNS and Peripheral nervous system expression of moody are lost in mutants for the master regulator of glial fate, glial cells missing (gcm^N17), confirming that they are indeed glial. tre1 is expressed in all longitudinal glia and a subset of surface glia, as well as in cells along the midline. As expected, the (lateral) glial expression is lost in gcm mutants, while midline expression is not. Both moody and tre1 are also expressed outside the nervous system in a largely mutually exclusive manner, specifically in the germ cells, the gut, and the heart (Schwabe, 2005).

Several additional G protein signaling components are found in the surface glia. The six extant Gα genes show broad and overlapping expression in embryogenesis, with three of them (Go, Gq, and Gs) expressed throughout the nervous system and Gi expressed more specifically in surface glia. Gβ13F and Gγ1 are ubiquitously expressed during embryogenesis. Finally, the RGS loco is uniformly expressed in early embryos due to a maternal contribution but is then transcriptionally upregulated in surface and longitudinal glia, as well as in other tissues outside the nervous system. The nervous-system expression of loco is lost in gcm mutants. The presence of both Moody and Loco protein in the surface glia is confirmed using immunohistochemistry, but at 17 hr of development, when staining is feasible, the protein levels are still quite low (Schwabe, 2005).

In sum, the GPCR Moody, the RGS Loco, and Gi are differentially expressed in surface glia. This expression precedes and accompanies the morphogenesis and sealing of the surface glial sheath (Schwabe, 2005).

To examine protein expression and distribution of the GPCR signaling components in greater detail, third-instar larval nerve cords were examined. By this stage, the surface glia have doubled in size and show robust protein expression of GPCR signaling and SJ components (Schwabe, 2005).

Moody immunostaining is found at the plasma membrane, where it shows strong colocalization with the SJ marker Nrg-GFP. Loco immunostaining is punctate and more dispersed throughout the cytoplasm, with some accumulation at the plasma membrane, where it colocalizes with Moody. To avoid fixation and staining artifacts, fluorescent-protein fusions (Moody-mRFP; Loco-GFP) were generated and expressed using moody-Gal4, which drives weak surface glial expression. In the live nerve-cord preparations, Loco-GFP is much less dispersed and shows strong colocalization with Moody-mRFP at the plasma membrane (Schwabe, 2005).

In the absence of a known ligand, the coupling of G proteins to receptors is difficult to establish, but their binding to RGS proteins is readily determined. Loco physically binds to and negatively regulates Gi, and vertebrate Loco homologs (RGS12/14) have been shown to negatively regulate Gi/Go. In S2 tissue-culture assays, it was found that Loco binds to Gi and Go, but not to Gs and Gq. Double-label immunohistochemistry confirms that both Gi and Go are expressed in the surface glia (Schwabe, 2005).

Thus, Loco physically interacts with Gi and Go and shows subcellular colocalization with Moody, suggesting that the four signaling components are part of a common molecular pathway (Schwabe, 2005).

Using dye penetration as the principal assay, whether the GPCR signaling components that are expressed in surface glia play a role in insulation was examined. moody genomic (Δ17; Bainton, 2005) and RNAi mutants show similar, moderate insulation defects. The embryos are able to hatch but show mildly uncoordinated motor behavior and die during larval or pupal stages. The dye-penetration defect of moody^Δ17 is completely rescued by genomic rescue constructs containing only the moody ORF. Both moody splice forms (α and β; (Bainton, 2005) are able to rescue the defect independently, as well as in combination. tre1 genomic (Kunwar, 2003) and RNAi mutants show no significant dye-penetration defect and no synergistic effects when combined with moody using RNAi. Thus, despite the close sequence similarity of the two GPCRs and their partially overlapping expression in surface glia, only moody plays a significant role in insulation. Overexpression of moody causes intracellular aggregation of the protein (Schwabe, 2005).

loco is expressed both maternally and zygotically. loco zygotic nulls are paralytic and, on the basis of an ultrastructural analysis, a disruption of the glial seal, has been suggested. In a dye-penetration assay, loco zygotic null mutants show a strong insulation defect, which can be rescued by panglial expression of Loco in its wt or GFP-tagged form. The extant null allele of loco (Δ13) did not yield germline clones; therefore loco RNAi was used to degrade the maternal in addition to the zygotic transcript. In loco RNAi embryos, dye penetration is indeed considerably more severe. Overall, insulation as well as locomotor behavior is affected much more severely in loco than in moody and is close in strength to the SJ mutants. Overexpression of loco is phenotypically normal (Schwabe, 2005).

Thus, positive (moody) and negative (loco) regulators of G protein signaling show qualitatively similar defects in loss of function, suggesting that both loss and gain of signal are disruptive to insulation. Such a phenomenon is not uncommon and is generally observed for pathways that generate a localized or graded signal within the cell (Schwabe, 2005).

Both Gi and Go have a maternal as well as a zygotic component. Gi zygotic null flies survive into adulthood but show strong locomotor defects. In Gi maternal and zygotic null embryos show a mild dye-penetration defect, which is markedly weaker than that of moody, suggesting redundancy among Gα subunits. To further probe Gi function, the wt protein (Gi-wt) as well as a constitutively active version (Gi-GTP) were overexpressed in glia using repo-Gal4; such overexpression presumably leads to a masking of any local differential in endogenous protein distribution. Expression of Gi-wt results in very severe dye penetration, while overexpression of Gi-GTP is phenotypically normal. Only Gi-wt but not Gi-GTP can complex with Gβγ; overexpression of Gi-wt thus forces Gβγ into the inactive trimeric state. This result therefore suggests that the phenotypically crucial signal is not primarily transduced by activated Gi but rather by free Gβγ. Similar results have been obtained in the analysis of Gi function in asymmetric cell division (Schwabe, 2005).

Go null germline clones do not form eggs and do not survive in imaginal discs, indicating an essential function for cell viability (Katanaev, 2005). Therefore animals with glial overexpression of constitutively active (Go-GTP), constitutively inactive (Go-GDP), and wt (Go-wt) Go (Katanaev, 2005) were examined. Overexpression of Go-GDP, which cannot signal but binds free Gβγ, leads to severe dye penetration, again pointing to a requirement for Gβγ in insulation. However, Go-GTP and Go-wt show a moderate effect, suggesting that signaling by active Go does contribute significantly to insulation, in contrast to active Gi (Schwabe, 2005).

Overall, it was found that all four GPCR signaling components expressed in surface glia are required for insulation, further supporting the notion that the four components are part of a common pathway. The phenotypic data suggest that this pathway is complex: two Gα proteins, Gi and Go, are involved, but with distinct roles: activated Go and Gβγ appear to mediate most of the signaling to downstream effectors, while activated Gi seems to function primarily as a positive regulator of Gβγ. The loss of moody appears much less detrimental than the loss of free Gβγ (through overexpression of Gi-wt or Go-GDP); this is inconsistent with a simple linear pathway and points to additional input upstream or divergent output downstream of the G proteins. Finally, it was consistently observed that both loss (moody, Gi null, and Go-GDP) and gain (loco and Go-GTP) of signal are disruptive to insulation, suggesting that the G protein signal or signals have to be localized within the cell (Schwabe, 2005).

These complexities of G protein signaling in insulation preclude an unambiguous interpretation of genetic-interaction experiments and thus the linking of moody to Gi/Go/loco by genetic means. Double-mutant combinations between moody and loco were generated using genomic mutants as well as RNAi, with very complex results: in moody loco genomic double mutants, the insulation defect is worse than that of loco alone, while in moody loco RNAi double mutants the insulation defect is similar to that of moody alone. This strong suppression of loco by moody is also observed in the survival and motor behavior of the RNAi-treated animals. Thus the phenotype of the double-mutant combination is dependent on the remaining levels of moody and loco, with moody suppressing the loco phenotype when loco elimination is near complete (Schwabe, 2005).

To understand how the GPCR signaling components effect insulation at the cellular level, the distribution of different markers in the surface glia was examined under moody and loco loss-of-function conditions and under glial overexpression of Gi-wt. To rule out cell fating and migration defects, the presence and position of the surface glia were determined using the panglial nuclear marker Repo. In all three mutant situations, the full complement of surface glia is present at the surface of the nerve cord, with the positioning of nuclei slightly more variable than in wt (Schwabe, 2005).

In the three mutants, the SJ marker Nrg-GFP still localizes to the lateral membrane compartment, but the label is of variable intensity and sometimes absent, indicating that the integrity of the normally continuous circumferential SJ belt is compromised. Notably, the size and shape of the surface glia are also very irregular. While qualitatively similar, the phenotypic defects are more severe in loco and under Gi-wt overexpression than in moody, in line with the results of functional assays. When examining the three mutants with the actin marker GFP-Moesin, it was found that the cortical actin cytoskeleton is disrupted in varying degrees, ranging from a thinning to complete absence of marker, comparable to the effects observed with Nrg-GFP. However, GFP-positive fibrous structures are present within the cells, indicating that the abnormalities are largely restricted to the cell cortex. The microtubule organization, as judged by tau-GFP marker expression, appears normal in the mutants. The light-microscopic evaluation thus demonstrates that, in the GPCR signaling mutants, the surface glia are positioned correctly and capable of forming a contiguous epithelial sheet as well as septate junctions. Instead, the defects occur at a finer scale -- abnormally variable cell shapes and sizes, and irregular distribution of cortical actin and SJ material (Schwabe, 2005).

The changes in cell shape and actin distribution that were observed in the three mutants might simply be a secondary consequence of abnormalities in the SJ belt; to test this possibility, how a loss of the SJ affects the morphology and the actin cytoskeleton of the surface glia was examined. SJ components are interdependent for the formation and localization of the septa, and lack of a single component, such as Nrg, leads to nearly complete loss of the junction and severe insulation defects. In Nrg mutants, the surface glial cell shape and cortical actin distribution show only mild abnormalities. Thus, in contrast to the GPCR signaling mutants, the complete removal of the SJ causes only weak cytoskeletal defects, strongly arguing against an indirect effect. It is concluded that GPCR signaling most likely functions by regulating the cortical actin cytoskeleton of the surface glia, which in turn affects the positioning of SJ material along the lateral membrane (Schwabe, 2005).

More detailed insight into the nature of the defects in GPCR signaling mutants is afforded by electron microscopy. The surface glia in nerve cords of first-instar wild-type and mutant larvae were examined. Initially, dye penetration into the nerve cord was tested using ruthenium red. In wild-type, the dye diffuses only superficially into the surface glial layer, while in moody and loco mutants the dye penetrates deep into the nerve cord, in concordance with light-microscopic data. Tissue organization and SJ morphology were examined under regular fixation in randomly selected transverse sections. It has been reported that the surface glial sheath is discontinuous in loco mutant nerve cords, but this analysis was carried out at 16 hr of development, i.e., at a time when, even in wild-type, SJs are not yet established and the nerve cord is not sealed. In contrast to these findings, in the current study it was observed that, in loco as well as moody mutants, the glial sheath is in fact contiguous at the end of embryonic development. The ultrastructure of individual septa and their spacing also appear normal, indicating that moody and loco do not affect septa formation per se. However, the global organization of the junctions within the glial sheath appears perturbed: in wild-type, the surface glia form deep interdigitations, and the SJs are extended, well-organized structures that retain orientation in the same plane over long distances. In moody and loco mutants, the SJs are much less organized; they are significantly shorter in length and do not form long planar extents as in wild-type (Schwabe, 2005).

Taken together, the light- and electron-microscopic evaluations of the GPCR signaling mutants both show defects in the organization of the surface glial epithelium. The reduction in SJ length is consonant with the variability and local disappearance of the Nrg-GFP marker. Since the sealing capacity of the junction is thought to be a function of its length, the reduction in mean SJ length in the mutants provides a compelling explanation for the observed insulation defect (Schwabe, 2005).

Therefore, in addition to a reduction of the insulating SJs, this analysis of the GPCR signaling mutants revealed irregular cell shape and size, as well as weaker and variable accumulation of cortical actin in the surface glia. These data suggest that the primary defect in the mutants lies with a failure to stabilize the cortical actin, whose proper distribution is required for the complex extended morphology of the glia, which then affects SJ formation as a secondary consequence. Several lines of evidence exclude the reverse chain of causality, that is, a primary SJ defect resulting in destabilization of cortical actin and cell-shape change. Surface glia coalesce into a contiguous sheath and show strong accumulation of cortical actin before SJ material accumulates and sealing is completed. In the GPCR signaling mutants, there is misdistribution of SJ material along the cell perimeter, but the junctions do form. Finally, the GPCR signaling mutants show cell-shape and cortical actin defects that are much more severe than those observed in the near complete absence of SJ (Schwabe, 2005).

Compared to the columnar epithelia of the ectoderm and the trachea, the surface glial sheath is very thin. Compensating for their lack in height, surface glia form deep 'tongue-and-groove' interdigitations with their neighbors. This increases the length of the intercellular membrane juxtaposition and thus of the SJ, which ultimately determines the tightness of the seal. It is proposed that the surface glial interdigitations are the principal target of regulation by GPCR signaling. In GPCR signaling mutants, a loss of cortical actin leads to diminished interdigitation and thus to a shortening of the SJ, resulting in greater permeability of the seal. This model integrates all the observations at the light- and electron-microscopic levels (Schwabe, 2005).

Effects of Mutation

Drosophila Neuroglian (Nrg) and its vertebrate homolog L1-CAM are cell-adhesion molecules (CAM) that have been well studied in early developmental processes. Mutations in the human gene result in a broad spectrum of phenotypes (the CRASH-syndrome) that include devastating neurological disorders such as spasticity and mental retardation. Although the role of L1-CAMs in neurite extension and axon pathfinding has been extensively studied, much less is known about their role in synapse formation. A single extracellular missense mutation in nrg⁸⁴⁹ mutants disrupts the physiological function of a central synapse in Drosophila. The identified giant neuron in nrg⁸⁴⁹ mutants make a synaptic terminal on the appropriate target, but ultrastructural analysis reveals in the synaptic terminal a dramatic microtubule reduction, which is likely to be the cause for disrupted active zones. The results reveal that tyrosine phosphorylation of the intracellular ankyrin binding motif is reduced in mutants, and cell-autonomous rescue experiments demonstrate the indispensability of this tyrosine in giant-synapse formation. This function in giant-synapse formation is conserved in human L1-CAM but not in either human L1-CAM with a pathological missense mutation or in two isoforms of the paralogs NrCAM and Neurofascin. It is concluded that Nrg has a function in synapse formation by organizing microtubules in the synaptic terminal. This novel synaptic function is conserved in human L1-CAM but is not common to all L1-type proteins. Finally, the findings suggest that some aspects of L1-CAM-related neurological disorders in humans may result from a disruption in synapse formation rather than in axon pathfinding (Godenschwege, 2006).

The function of L1-type proteins in early neuronal development has been intensively studied, but less is known about their contribution to synapse formation. Recently, it has been shown that the L1-type protein Neurofascin (but not L1-CAM) is important in directing GABAergic innervation of the Purkinje axon initial segment and that this involves ankyrin. This study provides evidence that the ankyrin-independent form of Drosophila L1-type protein Nrg has a function in synapse formation and this novel function in synapse formation is conserved in the human L1-CAM protein but is not common to all L1-type isoforms (Godenschwege, 2006).

L1-type proteins are multifunctional proteins with many interaction partners, and the complete loss of function of Nrg protein results in defects in neurite outgrowth and axonal guidance as well as in lethality in Drosophila. In contrast, nrg⁸⁴⁹ flies carrying a single missense mutation in the second Ig domain are viable, but synapse formation was disrupted in virtually all animals and pathfinding was rarely affected. This suggests that the nrg⁸⁴⁹ missense mutation does not affect the overall function of Nrg but rather disrupts a subset of functions that is important for giant-synapse formation, but nonessential for the outgrowth or guidance of the GF axon. This implies that the functions of Nrg during axon pathfinding and synapse formation are distinct and separable (Godenschwege, 2006).

Earlier studies revealed that the extracellular domain of Nrg and L1-CAM functioning as a ligand is sufficient for many developmental processes including the stimulation of neurite outgrowth, sensory axon pathfinding, and eye development. For example, expression of Nrg^GPI is able to rescue certain aspects of pathfinding in a nrg loss-of-function background and is able to activate Echinoid, epidermal growth factor, (EGF), and fibroblast growth factor (FGF) receptor signaling. In addition, Nrg^GPI has normal homophilic binding properties and can induce ankyrin recruitment when binding to full-length Nrg. More recently, it has been shown that Nrg¹⁸⁰, human L1-CAM, rat-NrCAM, human-NCAM, and Drosophila FasII are able to rescue axon pathfinding defects of specific sensory neurons under nrg loss-of-function conditions, suggesting a functional overlap of these proteins in axonal guidance. However, although neuronal Nrg¹⁸⁰ and its human homolog L1-CAM were able to rescue the synaptic dysfunction in the nrg⁸⁴⁹ missense mutants, no functional overlap was found with Nrg^GPI, rat-NrCAM, NCAM, or FasII. This supports the hypothesis that Nrg function during synapse formation is distinct from its well-known function during neurite extension and axonal guidance and shows that the intracellular domain of Nrg is indispensable for synapse formation (Godenschwege, 2006).

The highly conserved tyrosine of the ankyrin binding site (FIGQY) in the intracellular domain is essential for Nrg function in giant-synapse formation, and the ultrastructural phenotypes suggest that one of its functions may be to stabilize microtubules in the synaptic terminal. L1-type proteins have been shown to generate different microdomains that are either ankyrin free or ankyrin containing, as determined by the phosphorylation status of the tyrosine in the ankyrin binding motif. The homophilic interaction of L1-type proteins in trans or heterophilic interaction in cis induces the recruitment of ankyrin to the unphosphorylated FIGQY, which in turn interacts with the spectrin cytoskeleton. Recently, it has been shown that the spectrin cytoskeleton is important for stabilization of the neuromuscular junction and that the loss of spectrin can induce synapse disassembly and retraction. Interestingly, synaptic boutons in the NMJ lacking spectrin exhibit ultrastructural phenotypes similar to the giant synapse of nrg⁸⁴⁹ mutants: they are devoid of microtubules, and the synaptic vesicle density is severely reduced. Hence, FIGQY-unphosphorylated Nrg may have an effect on the microtubule cytoskeleton by its connection to the spectrin cytoskeleton via ankyrin (Godenschwege, 2006).

Although a function cannot generally be excluded for ankyrin binding Nrg in giant-synapse formation, a disruption in this pathway seems not to be the primary cause for the phenotypes in nrg⁸⁴⁹ mutants. Strong evidence is provided that in nrg⁸⁴⁹ mutants, signaling via ankyrin-independent Nrg is disrupted and this affects giant-synapse formation. In nrg⁸⁴⁹ flies, tyrosine phosphorylation of the FIGQY motif is reduced. Furthermore, Nrg^Y1234F, which still has residual binding affinity for ankyrin, has no capacity to rescue mutant synapses but rather has a dominant-negative effect on synapse formation when expressed in a wild-type background. Interestingly, Nrg/L1 with a phosphorylated FIGQY motif is found at cell-cell contact sites in the nervous system and does not recruit ankyrin but has been proposed to bind to phospho-FIGQY-specific proteins. In vertebrates, one of these phospho-FIGQY-specific proteins has been identified as doublecortin, a neurogenesis-specific protein that stabilizes microtubules. Hence, it is proposed that the phosphorylated FIGQY Nrg has a function in giant-synapse formation possibly by anchoring of microtubules in the synaptic terminal via a protein similar to doublecortin (Godenschwege, 2006).

Nrg links the plasma membrane to the cytoskeleton, where it organizes and stabilizes the synaptic terminal, and when this function is disrupted, it may lead to the synaptic phenotypes seen in nrg⁸⁴⁹ animals. Hence, Nrg may be important for the stability of the active zones by providing a scaffolding function at the synapse that affects local signaling. Alternatively, nrg⁸⁴⁹ synaptic terminals may represent nascent synapses that have never matured because the dramatic cytoskeletal defects in large synaptic terminals may affect retrograde signaling. Finally, it should be noted that Nrg may not only have a scaffolding function that affects signaling important for synapse formation but may also signal itself. For example, vertebrate L1-CAM can activate mitogen-activated protein kinases (MAPKs) and extracellular-signal-regulated kinases 1 and 2, which are known to play a role in synaptic plasticity and memory formation (Godenschwege, 2006).

Unfortunately, the results do not allow unequivocal distinguishing of which interaction on the extracellular side is affected by the mutation in nrg⁸⁴⁹ flies. Missense mutations of a surface residue in the second Ig domain in L1-type proteins have been shown to affect homophilic interactions, heterophilic interactions, or both simultaneously. The finding that expression of Nrg¹⁸⁰ on both sides of the synapse has rescue capacity could be simply due to the fact that heterophilic interaction on both sides of the synapse may contribute to synapse formation. However, because Nrg^S213L protein has some residual function, it is also possible that homophilic interaction between wild-type and mutant Nrg protein could be the reason for a rescue capacity of Nrg¹⁸⁰ on either side of the synapse. Hence, it is suggested that the missense mutation in nrg⁸⁴⁹ flies probably disrupts both homophilic and a heterophilic interaction during giant-synapse formation and that the combination of these interactions results in signaling via the phosphorylated, ankyrin binding motif that is important for synapse assembly (Godenschwege, 2006).

The finding that vertebrate L1-CAM, but not the tested paralog isoforms of NrCAM and Neurofascin, rescue the synaptic defects in the Drosophila nrg⁸⁴⁹ mutants suggests that the synaptic function is conserved but is not a feature of all L1-type proteins and therefore is highly specific. A few of the over 140 identified pathological mutations in L1 have been studied in cell culture, and the results suggest that some pathological defects may be the result of disrupted neurite outgrowth, extension, and branching. However, until now, none of the pathological mutations have been characterized with respect to their affect on synapse formation. It is shown that human L1 with a corresponding pathological mutation (L1^H210Q) in the second Ig domain is not able not rescue the synaptic defects in the Drosophila nrg⁸⁴⁹ allele. Hence, it is possible that some of the clinical L1 phenotypes in humans may also be attributed to synaptic defects rather than axonal growth and pathfinding errors. Therefore, future studies of human L1 mutations in the GFS may not only give new insights into the endogenous role of Nrg and L1 in synapse formation but may also help to understand the pathology of L1-related neurological disorders (Godenschwege, 2006).

Genetic analysis of an overlapping functional requirement for L1- and NCAM-type proteins during sensory axon guidance in Drosophila

L1- and NCAM-type cell adhesion molecules represent distinct protein families that function as specific receptors for different axon guidance cues. However, both L1 and NCAM proteins promote axonal growth by inducing neuronal tyrosine kinase activity and are coexpressed in subsets of axon tracts in arthropods and vertebrates. The functional requirements for the Drosophila L1- and NCAM-type proteins, Neuroglian (Nrg) and Fasciclin II (FasII), have been studied during postembryonic sensory axon guidance. The rescue of the Neuroglian loss-of-function (LOF) phenotype by transgenically expressed L1- and NCAM-type proteins demonstrates a functional interchangeability between these proteins in Drosophila photoreceptor pioneer axons, where both proteins are normally coexpressed. In contrast, the ectopic expression of Fasciclin II in mechanosensory neurons causes a strong enhancement of the axonal misguidance phenotype. Moreover, these findings demonstrate that this functionally redundant specificity to mediate axon guidance has been conserved in their vertebrate homologs, L1-CAM and NCAM (Kristiansen, 2005).

This study presents an analysis of the requirements and the functional specificity of Drosophila L1- and NCAM-type proteins during the postembryonic development of the Drosophila peripheral sensory nervous system. The partially penetrant phenotypes, which have been reported for L1- and NCAM-LOF mutants in Drosophila and different vertebrate model systems, suggest that the requirement for these neural CAMs is not absolute and that the lack of either L1- or NCAM-type proteins during nervous system development can be partially compensated for by other gene products. Moreover, considering the unique specificity of L1's and NCAM's homo- and heterophilic adhesive interactions, a molecular redundancy between these protein families may be unexpected. The specificities of the homophilic adhesive interactions within the L1 and the NCAM protein families have undergone considerable evolutionary changes. Drosophila Nrg and FasII exhibit a very low cross-reactivity with their vertebrate homologs, L1-CAM and NCAM. Although only the neuronal isoforms of human (L1-CAM^RSLE+) and of Drosophila Neuroglian (Nrg¹⁸⁰) have been directly tested for their ability to interact with each other, these results indicate that the ability of vertebrate CAMs to rescue the Nrg LOF phenotype most likely relies on homotypic adhesion, rather than on an interaction with endogenous Drosophila CAMs. This conclusion is also supported by the observation that the GOF phenotype in the wing sensory nervous system is only observed when the vertebrate transgene is expressed in both the wing epithelium and the sensory neurons. In addition, endogenous Nrg expression is not required for the production of the GOF axonal misguidance phenotype in the Drosophila wing (Kristiansen, 2005).

Although axonal growth and guidance involve a large array of different neuronal adhesion molecules, there appears to be a limited number of signaling pathways that are shared among structurally different CAM families. The two major signaling pathways, which are triggered by Ig-CAMs, involve nonreceptor tyrosine kinases or receptor tyrosine kinases, such as FGFR and EGFR. Both of these signaling pathways may act synergistically or in a redundant manner. L1-CAM-, NCAM-, as well as N-cadherin-mediated neuronal cell adhesion all activate neuronal FGF receptors and thereby induce neurite outgrowth in vitro. This suggests that structurally different neural CAMs are capable of feeding into the same signaling pathway and that multiple adhesive specificities coordinately influence axonal growth and guidance (Kristiansen, 2005).

Axonal guidance in the Drosophila ocellar sensory system (OSS) and the wing sensory nervous system involves the Nrg-mediated activation of FGF and EGF receptors. Constitutive activation of FGFR or EGFR can rescue the nrg³ LOF phenotype in the OSS, and Nrg GOF axonal misguidance in the developing wing is reversed by a hypomorphic allele of the Drosophila EGF receptor. The two types of neurons in the Drosophila OSS, ocellar pioneer (OP) and bristle mechanosensory (BM) neurons, differ in their expression of Nrg and FasII protein and in their requirement for both proteins during axonal growth and guidance. Whereas the neuron-specific isoforms Nrg¹⁸⁰ and FasII^PEST+ are coexpressed in OP axons, BM axons only express Nrg¹⁸⁰, but not FasII. The surrounding epidermis, which interacts with BM but not with OP axons, expresses the nonneuronal Nrg¹⁶⁷ isoform (Kristiansen, 2005).

The nrg LOF rescue experiments reveal strikingly different requirements for Nrg and FasII protein in the two neuronal cell populations. The requirement for Nrg in OP axons can be sustained by either the neural Nrg¹⁸⁰ or FasII^PEST+, but not by the nonneuronal Nrg¹⁶⁷ isoform. The two Nrg protein isoforms have identical extracellular domains and only differ in the size of their respective cytoplasmic domain. The capacity of FasII to fulfil the Nrg¹⁸⁰ requirement in OP axon guidance suggests that these structurally different proteins share a redundant function in these axons. This conclusion is further supported by the observation that the partially penetrant nrg LOF OP axonal misguidance phenotype is significantly amplified by a reduction of the fasII gene dosage. Remarkably, ectopic FasII^PEST+ expression in BM neurons enhances the deleterious effect of the Nrg loss, a situation that fits within the concept of antiredundancy or opposing functional capacities (Kristiansen, 2005).

The scenario of cell-specific redundant functions of Nrg¹⁸⁰ and FasII^PEST+ is maintained by their vertebrate homologs L1-CAM/Nr-CAM and NCAM¹⁴⁰, respectively. This indicates that the redundant specificities of L1 and NCAM proteins in neuronal subsets and the corresponding molecular interactions have been conserved in both CAM families over a long evolutionary time period. However, in contrast to the nonneuronal Nrg¹⁶⁷ isoform, which exhibits an antiredundant capacity compared with Nrg¹⁸⁰, the nonneuronal (RSLE−) vertebrate L1-CAM isoform is able to rescue the Nrg deficiency in OP axons. In contrast to Drosophila Nrg, the two vertebrate L1-CAM isoforms differ by the inclusion or exclusion of two small exons. The insertion of the five additional amino acid residues, which are encoded by exon2, into the L1-CAM extracellular domain modifies the homo- and heterophilic functions of vertebrate L1-CAMs. The inability of the human L1-CAM^RSLE+ isoform to efficiently interact with Drosophila Neuroglian suggests that the L1-CAM^RSLE+ GOF phenotype is the result of homotypic molecular interactions. Moreover, the rescue of nrg³ OP axonal phenotype by L1-CAM^RSLE− occurs in an Nrg deficient background, suggesting that vertebrate L1-CAM^RSLE− proteins are able to engage in homotypic molecular interactions in Drosophila. Interestingly, the nonneuronal human L1-CAM^RSLE− protein, for which a lower homophilic interaction capacity has been postulated, causes a much weaker GOF phenotype than the neuronal mouse L1-CAM^RSLE+ isoform. Nevertheless, the results indicate that this lower homophilic binding activity of the RSLE− isoform is sufficient to support the functional replacement of Nrg¹⁸⁰ in OP axons in Drosophila (Kristiansen, 2005).

Inclusion of the cytoplasmic miniexon generates a tyrosine-based endocytosis signal (RSLEY) in the neuronal vetebrate L1-CAM isoform. The AP-2-mediated endocytosis of the neuronal L1-CAM^RSLE+ isoform appears to be an important step in the activation of the MAPK signaling cascade by L1-CAM. Since neither Drosophila Nrg isoform contains an equivalent endocytosis signal in their cytoplasmic domain, Drosophila Nrg function either does not involve endocytosis or uses a different type of sorting signal than vertebrate L1 proteins (Kristiansen, 2005).

Although the analysis of the two Nrg isoforms indicates a specific requirement for Nrg¹⁸⁰ in OSS neurons, analysis of GOF conditions in the wing peripheral nervous system reveals an underlying common ability to activate RTK signaling. Since the Nrg-mediated activation of EGFR kinase only requires the extracellular Nrg domain for its interaction with the EGFR, both Nrg isoforms are able to exhibit an identical RTK-dependent axonal misguidance GOF phenotype. The ability of homologous vertebrate L1- and NCAM proteins to elicit the same response in Drosophila sensory neurons indicates a common, conserved specificity to influence RTK activity and thereby to regulate axonal growth and guidance. However, the different ability of the neuronal versus the nonneuronal Nrg isoform to rescue the nrg LOF phenotype in the OSS indicates that Nrg-mediated axonal guidance is also regulated by cytoplasmic interactions (Kristiansen, 2005).

Since the separation of arthropods and chordates, there has been an enormous diversification in the size and organization of metazoan nervous systems. At the same time, there has also been an increase in the number of L1- and NCAM-type paralogous genes in vertebrates (but not in Drosophila), as well as structural divergence and acquisition of new specific functions within each protein family. Both types of proteins have conserved an average of 25%–30% amino acid identity between their vertebrate and Drosophila homologues. The two groups of genes are of roughly similar size, and both have undergone independent events that resulted in the generation of different tissue-specific isoforms in Drosophila and vertebrates. Although both the vertebrate and invertebrate proteins are normally coexpressed in specific axonal tracts, their respective realms of expression have shifted in insect versus vertebrate nervous systems. As a result, NCAM expression is more widespread than L1-CAM or Nr-CAM in vertebrates, while FasII is more restricted than Nrg in insects. Therefore, all these genes are evidently highly accessible to mutation and genetic drift, and the current situation most probably reflects a selective pressure to maintain NCAM- and L1-type protein coexpression in specific axonal tracts of the nervous system. Nevertheless, although both L1 and NCAM proteins have acquired many new functions in both arthropod and chordate species, it appears that they initially had at least partially overlapping roles in growth cone signaling during axon guidance. Both CAM families have apparently maintained some of these shared functions and a common specificity, including a basic function as activators of RTK signaling, over a long time period (Kristiansen, 2005).

Therefore, it seems that the functional redundancy between L1- and NCAM-type proteins could constitute an important evolutionary constraint. It prevents the drift of these molecules into completely different functional entities, while at the same time, it allows their structures to further diverge and acquire separate and additional specificities. It has been proposed that functional redundancy is one mechanism for the canalization (stability after developmental perturbation and during evolution) of developmental processes. The requirement for a shared specificity between L1- and NCAM-type proteins in the control of RTK signaling during axon guidance might therefore reflect a requisite for redundancy that is found in any complex communication process. Redundancy is an essential component in any communication process for ensuring reliability by compensating the naturally occurring perturbations. Neuronal wiring is a cell communication-driven process where a highly complex set of signaling systems operates in parallel. As the number of different signals involved in axon guidance enlarged concomitant with an increase in complexity during evolution, the system noise affecting growth cone signal integration during development also increased. Unspecific adhesive interactions may also constitute a major source of noise for navigating growth cones. Therefore, cooperative redundancy might contribute to establishing a “buffered” physiological context required for ensuring process fidelity. It is postulated that this is the reason why the ancestral functional redundancy between L1- and NCAM-type molecules has been conserved over the last 600 million years of evolution (Kristiansen, 2005).

Axonal ensheathment and septate junction formation in the peripheral nervous system of Drosophila

Axonal insulation is critical for efficient action potential propagation and normal functioning of the nervous system. In Drosophila, the underlying basis of nerve ensheathment is the axonal insulation by glial cells and the establishment of septate junctions (SJs) between glial cell membranes. However, the details of the cellular and molecular mechanisms underlying axonal insulation and SJ formation are still obscure. This study reports the characterization of axonal insulation in the Drosophila peripheral nervous system (PNS). Targeted expression of tau-green fluorescent protein in the glial cells and ultrastructural analysis of the peripheral nerves allow visualization the glial ensheathment of axons. Individual or a groups of axons are ensheathed by inner glial processes, which in turn are ensheathed by the outer perineurial glial cells. SJs are formed between the inner and outer glial membranes. Neurexin IV, Contactin, and Neuroglian are coexpressed in the peripheral glial membranes and these proteins exist as a complex in the Drosophila nervous system. Mutations in neurexin IV, contactin, and neuroglian result in the disruption of blood-nerve barrier function in the PNS, and ultrastructural analyses of the mutant embryonic peripheral nerves show loss of glial SJs. Interestingly, the murine homologs of Neurexin IV, Contactin, and Neuroglian are expressed at the paranodal SJs and play a key role in axon-glial interactions of myelinated axons. Together, these data suggest that the molecular machinery underlying axonal insulation and axon-glial interactions may be conserved across species (Banerjee, 2006a).

The localization of Nrx IV, Cont, and Nrg in the embryonic nervous system was studied. Nrx IV, Nrg, and Cont show colocalization at the nerve glial membranes. Nrx IV, Cont, and Nrg are interdependent for their epithelial SJ localization. Having established that Nrx IV, Cont, and Nrg colocalize in the peripheral nerves, whether these proteins are interdependent for their localization in the peripheral nerves was studied. The effect of the absence of each of these three proteins on the localization of the other two was studied in the embryonic peripheral nerves using nrx IV, cont, and nrg null mutants. nrx IV mutant embryos were stained with anti-Nrx IV, anti-Cont, and anti-Nrg. Cont and Nrg proteins showed a rather diffused localization when compared with their wild-type localization. In addition, Cont is present as puncta in the cytoplasm of the glial cells because of its failure to get targeted properly to the membrane in the absence of Nrx IV. Similarly, cont null mutant embryos also showed less defined distribution and reduction of Nrx IV and Nrg in the glial membrane. nrg null mutant embryos display significant reduction in Nrx IV and Cont localization in the glial membranes. These results demonstrate that Nrx IV, Cont, and Nrg are interdependent for their proper localization in the embryonic peripheral nerves (Banerjee, 2006a).

Ultrastructural analyses of the embryonic epithelia in nrx IV, cont, and nrg mutants showed that these genes are required for the formation and/or organization of epithelial SJs. Nrx IV, Cont, and Nrg are interdependent for their proper localization both in the epithelia and in the peripheral nerves. These phenotypic similarities raised an interesting possibility that these proteins are part of a macromolecular protein complex that exists in the nervous system. To determine whether these three proteins form a biochemical complex in the nervous system, Drosophila heads, which are a rich source of both neurons and glial cells, were used. Coimmunoprecipitation experiments using Nrx IV, Cont, and Nrg antibodies efficiently coprecipitated Nrx IV, Cont, and Nrg. Interestingly, both isoforms of Nrg (180 kDa neuronal and 167 kDa epithelial) were immunoprecipitated by Nrx IV and Cont antibodies, suggesting that both isoforms are part of a protein complex that includes Nrx IV and Cont. However, at this point, it is not possible to differentiate whether isoform-specific complexes are formed or whether both isoforms are in the same complex. In addition, sucrose density gradient analysis of the fly head lysates was performed to determine how Nrx IV, Cont, and Nrg distribute in buoyant density gradients. Nrx IV and Cont cosediment in overlapping fractions. Nrg shows distribution in the lighter sucrose density fractions that do not overlap with Nrx IV and Cont but partially overlaps with that of Nrx IV and Cont, indicating that these proteins are associated with subcellular structures of the same density and may associate into a biochemical complex that is partially maintained during subcellular fractionation (Banerjee, 2006a).

The immunofluorescence analysis of nrx IV, cont, and nrg null mutants showed that in each of these mutants, the other two proteins show qualitatively reduced fluorescence intensities under identical confocal settings. The possible explanation for reduced fluorescence intensity could be that loss of any of these proteins affects the localization or stability of the other proteins. Whether the levels of the other two proteins had changed in nrx IV, cont, and nrg null mutants was examined using immunoblot analysis. Nrx IV protein levels did not seem to be affected in cont and nrg mutants when compared with wild type. Cont protein levels were severely affected in nrx IV mutants compared with wild-type and nrg mutant embryos. Nrg protein levels were also affected in nrx IV mutants but showed no change in cont mutants. At this stage, it cannot be rule out whether the change in Cont and Nrg protein levels in nrx IV mutants are attributable to reduced stability or less synthesis of these proteins. Together, the biochemical data indicate that Nrx IV, Cont, and Nrg form a protein complex in the nervous system and that, in the absence of Nrx IV, the stability of Cont and Nrg is severely affected (Banerjee, 2006a).

Inner glial membrane processes are involved in the ensheathment of either an individual axon or a group of axons. This glial ensheathment not only provides insulation of axons but also generates unique junctions between either glial membranes or between axons and glial membranes. Nrx IV, Cont, and Nrg are involved in the establishment of glial–glial SJs. Although additional components involved in these interactions need to be identified, these findings provide a basis for additional analysis of neuronal SJs, which would be relevant to the understanding of their vertebrate counterparts: the paranodal axo-glial SJs (Bhat, 2003; Hortsch, 2003; Salzer, 2003; Banerjee, 2006a and references therein).

The fundamental basis of axonal ensheathment in any species is to faithfully transmit neuronal signals along the nerve fibers and optimize desired cellular responses. To maximize the speed of conduction and/or to minimize the loss of nerve signals, many species evolved mechanisms in which axonal lengths remained short (as seen in insects) by increasing the diameter of the axons or by clustering voltage-gated Na⁺ channels to discrete unmyelinated regions of the axon, the node of Ranvier, as seen in myelinated nerve fibers of vertebrates. Most invertebrate species use some type of glial cells to ensheath their axons without generating a myelin sheath. The insulation is contiguous without any breaks, suggesting that primitive nodal structures or clustering of voltage-gated Na⁺ channels may not exist in invertebrates. However, recent reports have challenged some of these notions. In copepod crustaceans, ultrastructural analysis of the first antenna and the CNS has revealed extensive myelination of sensory and motor axons. These studies have raised some fundamental questions about axonal insulation and origins of myelination (Banerjee, 2006a and references therein).

In invertebrates like Drosophila, two types of glial cells are involved in insulation. The inner peripheral glial cells are involved in axonal ensheathment, and the outer (perineurial) glial cells wrap around the inner glial cells to provide another level of ensheathment. This two-cell ensheathment in Drosophila peripheral nerves may be advantageous to ensure that high K⁺ containing hemolymph does not interfere with action potential propagation. Although the cellular aspects of axonal insulation are being unraveled, the molecular mechanisms underlying the axonal ensheathment remain to be investigated. Most importantly, what are the protein constituents of the insulating membranes, and whether some of the vertebrate myelin proteins have their homologs in invertebrates? A detailed molecular analysis of the nature of the glial cells that ensheath axons as in Drosophila or produce myelin-like structures and the type of myelin in copepods may provide insights into whether myelinating glial cells arose from a common ancestor. However, a genetic dissection of the axonal ensheathment in Drosophila will uncover some of the basic aspects of the neuron-glial interactions that lead to ensheathment of nerve fibers across species (Banerjee, 2006a).

Cell adhesion molecules play a pivotal role in establishing intercellular junctions [e.g., cadherins and associated catenins form a protein scaffold that establish adhesion contacts at the adherens junctions and link them to the actin cytoskeleton. Similarly, claudins and associated cellular scaffolding proteins are required for establishing TJs. In both of these examples, transmembrane proteins bring two opposing membranes together to establish junctions through homophilic and/or heterophilic interactions (Banerjee, 2006a and references therein).

The finding that nrg null mutant nerves display increased spacing between the outer and inner glial membranes suggest that Nrg may be involved in cell–cell interactions and cell–cell adhesion between glial membranes. In contrast, the observation that loss of nrx IV and cont does not affect the membrane spacing between the inner and outer glial membranes suggests that Nrx IV and Cont are not involved in membrane adhesion or bringing the glial membranes in close apposition. Together, these data suggest that Nrg is critical for both the adhesion and SJ formation, whereas Nrx IV and Cont are critical for the formation of the septa at SJs. In addition, the missing axonal fascicles in nrg mutant nerves could result from axon fasciculation defects or axonal degeneration as a secondary consequence resulting from the loss of glial support. Axonal fasciculation defects have been observed in nrg mutants. Alternatively, axonal loss in nrg mutants might result from a disruption in axonal cytoskeleton, because Nrg possesses domains that could potentially interact with and stabilize the axonal cytoskeleton. Based on Nrg expression in S2 cells, Nrg is predicted to recruit membrane skeleton assembly within specialized domains of neurons in response to cell adhesion. Both Nrg protein forms Nrg¹⁶⁷ and Nrg¹⁸⁰ contain a short cytoplasmic domain as a binding site for ankyrin. Ankyrins are linker proteins that connect various membrane proteins with the actin-spectrin network in the cell. Loss of axons observed in nrg mutants is clearly suggestive of axon-glial interdependence that may alter axonal survival. In vertebrates, axolemmal-myelin interactions are critical for the formation of the paranodal axo-glial SJs (see Garcia-Fresco, 2006). This raises an interesting possibility that axon-glial interactions in Drosophila may use similar interactive mechanism for establishing axo-glial SJs. Thus, the current studies on Nrx IV, Cont, and Nrg suggest that these proteins are critical for the formation and/or organization of the SJs between either glial cells and possibly between axons and glial cells, which remains to be further investigated (Banerjee, 2006a).

Invertebrate axons are insulated from their salt-rich environment through a glial-dependent BBB, which plays a crucial role in electrical and chemical insulation. Ultrastructural studies have demonstrated SJs between perineurial glial cells and inner glial cell membrane to form the structural basis of BBB of some insects. Absence of SJs in nrx IV, cont, and nrg mutants, and a compromised BNB in the PNS as evidenced by dye exclusion analyses, provide additional confirmation in support of SJs as a prerequisite for blood-nerve barrier (BNB) formation. Not just in the PNS, recent reports on the BBB formation and function also in the CNS of Drosophila underscore the importance of SJs in proper sealing and insulation of the nerve cord (Schwabe, 2005), thereby supporting that insulation and establishment of functional SJs in both PNS and CNS go hand in hand. Although, G-protein-coupled receptor signaling pathway members have been identified to establish the BBB in Drosophila CNS, the signaling mechanisms that operate in the PNS still remain to be established (Banerjee, 2006a).

The axo-glial SJs in the myelinated axons share many anatomical features similar to those of invertebrate SJs, especially the electron-dense ladder-like transverse septa (Bhat, 2003; Banerjee, 2006a). Ensheathment of Drosophila axons by perineurial glial cells in the absence of myelin-producing glial cells would have predicted that the molecular components of the invertebrate SJs would be different from those of the vertebrate axo-glial SJs. Surprisingly, the fly SJ molecular components are present at the vertebrate axo-glial SJs and not at the TJs, which serve similar functions. Thus, the molecular similarities reflect an evolutionarily conserved function of creating an ionic barrier both at Drosophila SJs and at axo-glial SJs in the paranodal region. More importantly, Drosophila nerves contain a large number of axons, which are collectively held together by glial membranes not only to maintain nerve fasciculation but also to maintain the insulation of individual axons. It would be of significant interest to establish whether axo-glial SJs are present in Drosophila nerves and to identify downstream components of these junctions in Drosophila that link these junctions to the glial or axonal cytoskeleton. Identification of such molecules will provide insights into whether these junctions play a much broader role in axon-glial signal transduction. In summary, the molecular and functional similarities between the Drosophila SJs and vertebrate axo-glial SJs should allow a genetic and molecular dissection to be undertaken of the formation and function of these junctions in Drosophila (Banerjee, 2006a).

The L1-type cell adhesion molecule Neuroglian is necessary for maintenance of sensory axon advance in the Drosophila embryo

Cell adhesion molecules have long been implicated in the regulation of axon growth, but the precise cellular roles played by individual cell adhesion molecules and the molecular basis for their action are still not well understood. The sensory system of the Drosophila embryo was used to shed light on the mechanism by which the L1-type cell adhesion molecule, Neuroglian regulates axon growth. A highly penetrant sensory axon stalling phenotype was found in neuroglian mutant embryos. Axons stalled at a variety of positions along their normal trajectory, but most commonly in the periphery some distance along the peripheral nerve. All lateral and dorsal cluster sensory neurons examined, except for the dorsal cluster neuron dbd, showed stalling. Sensory axons were never seen to project along inappropriate pathways in neuroglian mutants and stalled axons showed normal patterns of fasciculation within nerves. The growth cones of stalled axons possessed a simple morphology, similar to their appearance in wild type embryos when advancing along nerves. Driving expression of the wild type form of Neuroglian in sensory neurons alone rescued the neuroglian mutant phenotype of both pioneering and follower neurons. A partial rescue was achieved by expressing the Neuroglian extracellular domain. Over/mis-expression of Neuroglian in all neurons, oenocytes or trachea had no apparent effect on sensory axon growth. It is concluded that Neuroglian is necessary to maintain axon advance along axonal substrates, but is not required for initiation of axon outgrowth, axon fasciculation or recognition of correct growth substrates. Expression of Neuroglian in sensory neurons alone is sufficient to promote axon advance and the intracellular region of the molecule is largely dispensible for this function. It is therefore unlikely that Nrg acts as a molecular 'clutch' to couple adhesion of F-actin within the growth cone to the extracellular substrate. Rather, it is suggested that Neuroglian mediates sensory axon advance by promoting adhesion of the surface of the growth cone to its substrate. The finding that stalling of a pioneer sensory neuron is rescued by driving Neuroglian in sensory neurons alone may suggest that Neuroglian can act in a heterophilic fashion (Martin, 2008).

nrg LOF mutants show a highly penetrant sensory axon stalling phenotype. This defect is displayed by all neurons examined, except for the dorsal cluster neuron, dbd, which apparently leads other dorsal cluster axons as it grows along the ISN into the CNS. The fact that this pioneering neuron is unaffected by loss of nrg function may suggest that Nrg is necessary for axon advance only along axonal substrates. This idea is supported by the observation that lateral cluster axons in nrg mutants almost always advance successfully along their initial non-neuronal growth substrate, the spiracular branch of the trachea, and stall only after they have subsequently joined the ISN. As sensory axons in nrg mutants generally advance for some distance along nerves before stalling, Nrg is apparently required to maintain axon growth along axonal substrates, rather than to initiate it. Most axon stalls occur in the periphery, but their locations are quite variable and do not obviously coincide with particular structures, such as nerve branches, pointing to a stochastic component in the stalling process (Martin, 2008).

Stalled sensory axons in nrg mutants are always found in their correct nerve pathways. The absence of axon misprojection phenotypes shows that Nrg is not necessary for recognition of correct growth substrates by sensory axons. It was also observed that sensory axons in nrg mutants fasciculate normally with other axons. The simple growth cone morphology of the stalled axons -- club-shaped with few filopodia -- is similar to that seen in wild-type sensory axons when they are growing along nerves. These observations show that Nrg is not necessary for axon-axon fasciculation in this system. The Ig-CAM Fasciclin2 (Fas2) is a key regulator of motor axon fasciculation in the Drosophila embryo and also acts redundantly to Nrg in the regulation of OP axon growth in the adult OSS. However, Fas2 is unlikely to be acting redundantly with Nrg in regulating sensory axon fasciculation in the embryo, since embryonic sensory axons do not express Fas2. In addition, the pattern of 22C10 sensory nerve staining is normal in fas2 LOF mutants and in embryos over/misexpressing Fas2 in sensory neurons and their growth substrates. Whether Nrg acts redundantly with some other CAM to mediate sensory axon fasciculation remains to be determined (Martin, 2008).

The sensory axon defects seen in nrg mutant embryos are similar to those reported for motor axons, although motor axons stall at a later stage of growth, close to their synaptic targets, the body wall muscles, and unlike sensory axons, some motor axons misproject in nrg mutants. The nrg LOF phenotype in the embryonic sensory system is much more stereotypic than in the adult OSS. This difference may indicate that Nrg performs more diverse functions in the adult OSS, including regulation of axon fasciculation and guidance in addition to axon advance. While Nrg is expressed in both sensory axons and several of their growth substrates, including trachea and motor axons, mutant rescue experiments demonstrate that expression in sensory axons alone is sufficient to mediate their advance. It is unlikely, therefore, that the sensory axon defects are secondary to other known morphological defects in nrg mutants, such as motor axon stalling or failure of glial cell ensheathment of peripheral nerves (Martin, 2008).

An important, unresolved question is whether the effects of Nrg on axon growth are mediated by homophilic interactions between Nrg on the growth cone and its substrate or whether heterophilic interactions are involved. Substrate-bound Nrg has been shown to promote neurite extension from Nrg-expressing neurons in vitro, but it is unclear whether the same holds true in vivo. Expression of Nrg in sensory neurons completely rescues the axon stalling phenotype of lch5-1 seen in nrg mutants. In wild-type embryos and in nrg mutants, lch5-1 is the pioneer for the lateral cluster fascicle and associates with motor axons, rather than dorsal cluster sensory axons as it grows towards the CNS. Lch5-1 mutant rescue findings might suggest, therefore, that Nrg is acting in a heterophilic fashion to promote advance of this axon. This conclusion rests upon the assumption that, in nrg mutant embryos in which the lch5-1 stall phenotype has been rescued, the lch5-1 growth cone does not employ the dorsal sensory axons as a growth substrate as it advances along the ISN. Observations of the dynamics of lch5-1 growth cone activity, and its association with sensory and motor axons in the ISN as it advances along the nerve in these rescued embryos, would help to resolve this question (Martin, 2008).

Mutant rescue experiments show that the intracellular region of Nrg is largely dispensable for its role in promoting sensory axon advance. It has long been known that Nrg-mediated cell-cell adhesion does not require the intracellular region of the molecule. However, mutation of the intracellular ankyrin-binding domain of L1- type CAMs does have clear effects on their membrane mobility and their coupling to retrograde actin flow in growth cones. Given the linkage between ankyrins and the actin-spectrin cytoskeleton, this has led to suggestions that L1-type CAMs may transmit traction force generated by actin flow to the extracellular substrate, as posited by the 'molecular clutch' hypothesis. The recent finding that MAO kinase phosphorylation of the ankyrin-binding site regulates neurite growth from cerebellar granule neurons on a Ng-CAM coated substrate highlights the potential importance of the intracellular region of L1-type CAMs in mediating their effects on axon growth. However, this view is difficult to reconcile with the current results. One possible explanation is that the in vitro models used in most vertebrate studies to date do not accurately reflect the in vivo functions of L1- type CAMs. Alternatively, Nrg, the Drosophila L1 homologue, may function differently to vertebrate L1-type CAMs in promoting axon growth (Martin, 2008).

The finding that the intracellular region of Nrg is not essential for its ability to promote sensory axon advance raises the question of how Nrg-mediated adhesion of the growth cone to its substrate is coupled to retrograde F-actin flow. Current models assume that such a coupling provides the motive force for growth cone advance. One possibility is that Nrg interacts in cis with another receptor or CAM, which in turn is coupled to the cytoskeleton of the growth cone. A genetic screen, using the sensory axon stalling phenotype described in this study as an assay, would be one way of identifying such a Nrg-interacting molecule (Martin, 2008).

Taken together, these results suggest that Nrg may mediate sensory axon advance along nerves in the Drosophila embryo by promoting adhesion between the surface of the growth cone and its axonal substrates. This adhesive interaction is likely to involve a heterophilic interaction between Nrg on the growth cone and some other, as yet unidentified molecule, on the substrate. Intracellular signaling by the Nrg molecule does not appear to be essential for its function in this context. These findings for Nrg are mirrored by the recent discovery that the adhesive, rather than the signaling activity, of another CAM, Drosophila N-cadherin, is essential for its role in target selection of photoreceptor afferents (Martin, 2008).

Ango, F., et al. (2004). Ankyrin-based subcellular gradient of neurofascin, an immunoglobulin family protein, directs GABAergic innervation at purkinje axon initial segment. Cell 119(2): 257-72. 15479642

Bai, J., Chiu, W., Wang, J., Tzeng, T., Perrimon, N. and Hsu, J. (2001). The cell adhesion molecule Echinoid defines a new pathway that antagonizes the Drosophila EGF receptor signaling pathway. Development. 128(4): 591-601. 11171342

Banerjee, S., Pillai, A. M., Paik, R., Li, J. and Bhat, M. A. (2006a). Axonal ensheathment and septate junction formation in the peripheral nervous system of Drosophila. J. Neurosci. 26(12): 3319-29. 16554482

Banerjee, S., Sousa, A. D. and Bhat, M. A. (2006b). Organization and function of septate junctions: an evolutionary perspective. Cell Biochem. Biophys. 46(1): 65-77. 16943624

Bhat, M. A. (2003). Molecular organization of axo-glial junctions. Curr. Opin. Neurobiol. 13: 552-559. 14630217

Bieber, A.J., Snow, P.M., Hortsch, M., Patel, N.H., Jacobs, J.R., Traquina, Z.R. Schilling, J. and Goodman, C.S. (1989). Drosophila Neuroglian: a member of the immunoglobulin superfamily with extensive homology to the vertebrate neural adhesion molecule L1. Cell 59: 447-460

Cantera, R., et al. (2002). Mutations in spalt cause a severe but reversible neurodegenerative phenotype in the embryonic central nervous system of Drosophila melanogaster. Development 129: 5577-5586. 12421699

Charles, P., et al. (2002). Neurofascin is a glial receptor for the paranodin/Caspr-contactin axonal complex at the axoglial junction. Curr Biol. 12(3): 217-20. 11839274

Chen, C.-L., Lampe, D. J., Robertson, H. M. and Nardi, J. B. (1997). Neuroglian is expressed on cells destined to form the prothoracic glands of Menduca embryos as they segregated from surrounding cells and rearrange during morphogenesis. Dev. Biol. 181: 1-13

Davis, J. Q., McLaughlin, T. and Bennett, V. (1993). Ankyrin-binding proteins related to nervous system cell adhesion molecules: candidates to provide transmembrane and intracellular connections in adult brain. J. Cell Biol. 121: 121-33

Desai, C. J., Popova, E. and Zinn, K. (1994). A Drosophila receptor tyrosine phosphatase expressed in the embryonic CNS and larval optic lobes is a member of the set of proteins bearing the "HRP" carbohydrate epitope. J Neurosci 14: 7272-7283

Doherty, P. and Walsh, F. S. (1996). CAM-FGF receptor interactions: A model for axonal growth. Mol. Cell. Neurosci. 8: 99-111

Dubreuil, R. R., et al. (1996). Neuroglian-mediated cell adhesion induces assembly of the membrane skeleton at cell contact sites. J. Cell Biol. 133: 647-655

Einheber, S., et al (1997). The axonal membrane protein Caspr, a homolog of Neurexin IV, is a component of the septate-like paranodal junctions that assemble during myelination. J. Cell Biol. 139: 1495-1506. 9396755

Faivre-Sarrailh, C., et al. (2004). Drosophila contactin, a homolog of vertebrate contactin, is required for septate junction organization and paracellular barrier function. Development 131: 4931-4942. 15459097

Fitzli, D., et al. (2000). A direct interaction of Axonin-1 with NgCAM-related cell adhesion molecule (NrCAM) results in guidance, but not growth of commissural axons. J. Cell Biol. 149: 951-968.

Garcia-Alonso, L., Romani, S. and Jimenez, F. (2000). The EGF and FGF receptors mediate Neuroglian function to control growth cone decisions during sensory axon guidance in Drosophila. Neuron 28: 741-752

Garcia-Fresco, G. P., et al. (2006). Disruption of axo-glial junctions causes cytoskeletal disorganization and degeneration of Purkinje neuron axons. Proc. Natl. Acad. Sci. 103(13): 5137-42. 16551741

Garver, T. D., et al. (1997). Tyrosine phosphorylation at a site highly conserved in the L1 family of cell adhesion molecules abolishes ankyrin binding and increases lateral mobility of neurofascin. J. Cell Biol. 137(3): 703-14.

Genova, J. L. and Fehon, R. G. (2003). Neuroglian, Gliotactin, and the Na+/K+ ATPase are essential for septate junction function in Drosophila. J. Cell Biol. 161: 979-989. 12782686

Godenschwege, T. A., Kristiansen, L. V., Uthaman, S. B., Hortsch, M. and Murphey, R. K. (2006). A conserved role for Drosophila Neuroglian and human L1-CAM in central-synapse formation. Curr. Biol. 16(1): 12-23. 16401420

Hortsch, M., Bieber, A. J., Patel, N. H. and Goodman, C. S. (1990). Differential splicing generates a nervous system-specific form of Drosophila Neuroglian. Neuron 697-709

Hortsch, M., et al. (1995). The cytoplasmic domain of the Drosophila cell adhesion molecule Neuroglian is not essential for its homophilic adhesive properties in S2 cells. J. Biol. Chem. 270: 18809-18817

Hortsch, M. (1996). The L1 family of neural cell adhesion molecules: old proteins performing new tricks. Neuron 17: 587-593

Hortsch, M., et al. (1998a). A conserved role for L1 as a transmembrane link between neuronal adhesion and membrane cytoskeleton assembly. Cell Adhes. Commun. 5(1): 61-73.

Hortsch, M., et al. (1998b). Structural requirements for outside-in and inside-out signaling by Drosophila neuroglian, a member of the L1 family of cell adhesion molecules. J. Cell Biol. 142(1): 251-61.

Hortsch, M. and Margolis, B. (2003). Septate and paranodal junctions: kissing cousins. Trends Cell Biol 13: 557-561. 14573348

Huber, A. H., et al. (1994). Crystal structure of tandem type III fibronectin domains from Drosophila Neuroglian at 2.0 A. Neuron 12: 717-31

Islam, R., Wei, S. Y., Chiu, W. H., Hortsch, M. and Hsu, J. C. (2003). Neuroglian activates Echinoid to antagonize the Drosophila EGF receptor signaling pathway. Development. 130(10): 2051-9. 12668620

Jacobs, J. R. (1993). Perturbed glial scaffold formation precedes axon tract malformation in Drosophila mutants. J Neurobiol 24 (5): 611-626.

Jenkins, S. M., et al. (2001). FIGQY phosphorylation defines discrete populations of L1 cell adhesion molecules at sites of cell-cell contact and in migrating neurons. J. Cell Sci. 114(Pt 21): 3823-35. 11719549

Koushika, S. P., Lisbin, M. J. and White, K. (1996). ELAV, a Drosophila neuron-specific protein, mediates the generation of an alternatively spliced neural protein isoform. Curr. Biol. 6(12): 1634-41.

Kristiansen, L. V., et al. (2005). Genetic analysis of an overlapping functional requirement for L1- and NCAM-type proteins during sensory axon guidance in Drosophila. Mol. Cell. Neurosci. 28(1): 141-52. 15607949

Lisbin, M. J., Qiu, J. and White, K. (2001). The neuron-specific RNA-binding protein ELAV regulates neuroglian alternative splicing in neurons and binds directly to its pre-mRNA. Genes Dev. 15: 2546-2561. 11581160

Lustig, M., Sakurai, T. and Grumet, M. (1999). Nr-CAM promotes neurite outgrowth from peripheral ganglia by a mechanism involving axonin-1 as a neuronal receptor. Dev. Biol. 209(2): 340-51.

Lustig, M., et al. (2001). Nr-CAM and neurofascin interactions regulate ankyrin G and sodium channel clustering at the node of Ranvier. Curr. Biol. 11(23): 1864-9. 11728309

Martin, V., et al. (2008). The L1-type cell adhesion molecule Neuroglian is necessary for maintenance of sensory axon advance in the Drosophila embryo. Neural Develop. 3(1): 10. PubMed citation

Rawlins, E. L., White, N. M. and Jarman, A. P. (2003), Echinoid limits R8 photoreceptor specification by inhibiting inappropriate EGF receptor signalling within R8 equivalence groups. Development 130: 3715-3724. 12835388

Salzer, J. L. (2003). Polarized domains of myelinated axons. Neuron 40: 297-318. 14556710

Schaeren-Wiemers, N., et al. (2004). The raft-associated protein MAL is required for maintenance of proper axon--glia interactions in the central nervous system. J Cell Biol. 166(5): 731-42. 15337780

Schafer, D. P., Bansal, R., Hedstrom, K. L., Pfeiffer, S. E. and Rasband, M. N. (2004). Does paranode formation and maintenance require partitioning of neurofascin 155 into lipid rafts? J. Neurosci. 24(13): 3176-85. 15056697

Schwabe, T., Bainton, R. J., Fetter, R. D., Heberlein, U. and Gaul, U. (2005). GPCR signaling is required for blood-brain barrier formation in Drosophila. Cell 123: 133-144. 16213218

Siegler, M. V. and (1999). Engrailed negatively regulates the expression of cell adhesion molecules Connectin and Neuroglian in embryonic Drosophila nervous system. Neuron 22(2): 265-76. PubMed citation: 10069333

Speicher, S., et al. (1998). Neurotactin functions in concert with other identified CAMs in growth cone guidance in Drosophila. Neuron 20(2): 221-33. PubMed citation: 9491984

Spencer, S. A. and Cagan, R. L. (2003). Echinoid is essential for regulation of Egfr signaling and R8 formation during Drosophila eye development. Development 130: 3725-3733. 12835389

Tepass, U., Tanentzapf, G., Ward, R. and Fehon, R. (2001). Epithelial cell polarity and cell junctions in Drosophila. Annu. Rev. Genet. 35: 747-784. 11700298

Tepass, U., Tanentzapf, G., Ward, R. and Fehon, R. (2001). Epithelial cell polarity and cell junctions in Drosophila. Annu. Rev. Genet. 35: 747-784. 11700298

Woods, D. F., et al. (1996). Dlg protein is required for junction structure, cell polarity, and proliferation control in Drosophila Epithelia. J. Cell Biol. 134: 1469-1482. PubMed citation: 8830775

Yamamoto, M., Ueda, R., Takahashi, K., Saigo, K. and Uemura, T. (2006). Control of axonal sprouting and dendrite branching by the Nrg-Ank complex at the neuron-glia interface. Curr. Biol. 16(16): 1678-83. Medline abstract: 16920632

Zhao, G. and Hortsch, M. (1998). The analysis of genomic structures in the L1 family of cell adhesion molecules provides no evidence for exon shuffling events after the separation of arthropod and chordate lineages. Gene 215(1): 47-55. PubMed citation: 9666073

date revised: 25 May 2008