BIOLOGICAL OVERVIEW

Many cell surface proteins function in communication both between cells and between cell exterior and interior. To what degree are such molecules evolutionarily conserved, particularly those that also function in neural cell adhesion? Neuroglian is a cell surface transmembrane protein that functions in adhesion. It is found on neuronal axons, but not cell bodies, and on non-neuronal tissues such as trachea, hindgut, salivary glands and muscle. Neuroglian has extensive homology to a vertebrate neural cell adhesion molecule. This homology includes an extracellular immunoglobulin motif and an intracellular domain involved in communication with the cytoskeleton.

The conservation of neural extracellular matrix and cell-cell recognition molecules should further an understanding of their function and development. In this case Neuroglian is used in several contexts, with alternatively spliced varients. Such a system prefigures the complexity of the human nervous system.

The cytoplasmic domain of Neuroglian is not required for homophilic adhesive qualities. An artificial Neuroglian protein form was constructed by substituting the Neuroglian transmembrane segment and cytoplasmic domains with the sugar-lipid attachment signal of Fasciclin I. In essence a Neuroglian protein lacking an intracellular domain was created. This artificial Neuroglian molecule retains the ability to induce homophilic cell aggregation when expressed in tissue culture cells, and is able to interact with naturally occurring Neuroglian proteins. These results demonstrate that Neuroglian mediates a calcium-independent, homophilic cell adhesion activity and that neither cytoplasmic Neuroglian domains nor a direct interaction with cytoskeletal elements is essential for adhesion (Hortsch, 1995).

Loss of Discs large (DLG), a protein required for septate junction structure, cell polarity, and proliferation control in Drosophila epithelia, affects the distribution of Fasciclin III and Neuroglian, two transmembrane proteins thought to be involved in cell adhesion. Fasciclin III is highly enriched at the septate junction and is present in lower amounts in the lateral cell membrane, but is excluded from the adherens junction. Neuroglian is enriched at the apical end of the cell, reduced in the septate junction, and again on the rest of the lateral cell membrane. Localization of Fas III and Neuroglian in both salivary glands and imaginal discs is dependent on DLG. When septate junctions are completely eliminated in dlg mutants, both proteins are found apparently unrestricted along the cell membrane. In fact Neuroglian appears to have an elevated level of expression compared with wild type, while FAS III levels are reduced (Woods, 1996).

Neuroglian can transmit positional information directly to Ankyrin and thereby polarize its distribution in Drosophila tissue culture cells. The accumulation of Ankyrin at cell contacts requires the presence of the cytoplasmic domain of Neuroglian. A direct interaction between Neuroglian and Ankyrin can be demonstrated using yeast two-hybrid analysis. Thus, Neuroglian appears to be activated by extracellular adhesion so that Ankyrin and the membrane skeleton selectively associate with sites of contact and not with other regions of the plasma membrane (Dubreuil, 1996).

Expression of Neuroglian in Drosophila S2 tissue culture cells results in a selective recruitment of ankyrin and spectrin to sites of cell contacts. Ankyrin recruitment is strictly limited to cell contacts, even though Neuroglian is abundantly expressed over the entire cell surface. Thus, neuroglian can function as a signaling molecule that transmits the positional value of cell adhesion to the cytoplasmic assembly of ankyrin and spectrin. This outside-in signaling function appears to be conserved among L1 family members, since expression of human L1 in S2 cells also results in the assembly of ankyrin at cell contact sites. The adhesion-induced rearrangement of ankyrin and spectrin can be conveyed to other membrane proteins that interact with ankyrin and spectrin and might thereby provide a mechanism for the assembly of unique plasma membrane subdomains. For example, the NaK-ATPase, which is known to interact with ankyrin in vertebrates, is found to accumulate along with spectrin and ankyrin at sites of neuroglian-mediated adhesion in S2 cells. Thus, L1-mediated adhesion events result in a reorganization and compartmentalization of the plasma membrane, which may constitute an important biological function of L1 family members (Hortsch, 1998a and references).

Expression of the Drosophila cell adhesion molecule Neuroglian in S2 cells leads to cell aggregation and the intracellular recruitment of ankyrin to cell contact sites. The region of Neuroglian that interacts with ankyrin has been localized and the mechanism that limits this interaction to cell contact sites has been investigated. Yeast two-hybrid analysis and expression of Neuroglian deletion constructs in S2 cells have identified a conserved 36-amino acid sequence that is required for ankyrin binding. Mutation of a conserved tyrosine residue within this region reduces ankyrin binding and extracellular adhesion. However, residual recruitment of ankyrin by this mutant Neuroglian molecule is still limited to cell contacts, indicating that the lack of ankyrin binding at noncontact sites is not caused by tyrosine phosphorylation. A chimeric molecule, in which the extracellular domain of Neuroglian is replaced with the corresponding domain from the adhesion molecule fasciclin II, also selectively recruits ankyrin to cell contacts. Thus, outside-in signaling by Neuroglian in S2 cells depends on extracellular adhesion, but does not depend on any unique property of its extracellular domain. It is proposed that the recruitment of ankyrin to cell contact sites depends on a physical rearrangement of neuroglian in response to cell adhesion, and that ankyrin binding plays a reciprocal role in stabilizing the adhesive interaction (Hortsch, 1998a).

Cell adhesion molecules (CAMs) implement the process of axon guidance by promoting specific selection and attachment to substrates. In Drosophila, loss-of-function conditions of either the Neuroglian CAM, the FGF receptor coded by the gene heartless, or the EGF receptor coded by Egfr display a similar phenotype of abnormal substrate selection and axon guidance by peripheral sensory neurons. Moreover, neuroglian loss-of-function phenotype can be suppressed by the expression of gain-of-function conditions of heartless or Egfr. The results are consistent with a scenario where the activity of these receptor tyrosine kinases is controlled by Neuroglian at choice points where sensory axons select between alternative substrates for extension (Garcia-Alonso, 2000).

The ocellar sensory system (OSS) offers a simple scenario for the study of axon guidance at the cellular level. During axon guidance, growth cones make decisions at choice points. In order to change trajectories at these choice points, it is assumed that signal transduction mechanisms should operate to transform specific extracellular information in the modulation of their actin cytoskeleton. In the OSS, the initial decision to attach or not to attach to the head epithelium appears to be a key choice (at the first choice point) for two types of sensory axons as they navigate to their respective targets in the brain. Due to the process of head eversion, ocellar pioneer (OP) axons must navigate in the extracellular matrix (ECM), free of adhesion to the underlying epidermis. Reciprocally, bristle mechanosensory (BM) axons should follow the epidermis before and after head eversion, since they do not reach the brain until this later stage. Should BM axons initially extend apart from the epithelium, they might possibly be unable to follow a physical substrate toward their brain targets after head eversion. Therefore, the process of head eversion establishes a constraint that prevents substrate redundancy between ECM and epithelium. OSS axons must make a second decision in order to leave the surface of the head and project to the brain (at the second choice point). OP axons leave the ECM surrounding the head capsule toward the brain before head eversion. In contrast, BM axons leave the head's internal epithelial surface after head eversion, when several BM axons have converged together. Each of these decision processes are abnormal in nrg, htl, and Egfr mutant individuals. OP axons can decide to attach to the epithelium, preventing them from reaching the brain. OP axons can fail to leave the internal surface of the head (even when extending free of epithelial attachment) and project to ectopic positions within the head after eversion. BM axons can be found extending, although they are abnormally separated from the epidermis after head eversion, suggesting a failure of attachment to the epithelium before head eversion. BM axons also can stall in the epidermis, suggesting that Nrg may also promote axon extension. Finally, BM axons can fail to leave the epidermis toward the brain, as they should, remaining instead within the epidermal layer, where they project in abnormal directions. In such cases, they can sometimes be observed to perforate the epidermis and project outside of the head, suggesting that BM axons navigate in the epithelial surface, using proteases to facilitate their movements. In contrast, when attached to the epithelium, OP axons seem to have difficulty extending properly. One possibility is that extension in the epithelium requires this perforating activity that BM axons have and that may be missing in OP axons that normally extend in the ECM (Garcia-Alonso, 2000).

Since nrg, htl, and Egfr mutants exhibit similar OSS axon phenotypes, it seems possible that they function in a common mechanism during OSS axon guidance. In order to test genetically if Nrg behaves as an upstream regulator of the Fgfr and the Egfr, different double mutant combinations were constructed between nrg and gain-of-function conditions for htl and Egfr (which should function independently of upstream regulation). Gain-of-function conditions of htl can partially suppress the nrg phenotype. Elp alleles behave as gain-of-function conditions of Egfr and behave as strong suppressors of nrg OSS axon phenotype. Therefore, although it is still possible that Nrg could also perform some role based on pure adhesion, the results are most consistent with the idea that both Htl and Egfr mediate Nrg function in OSS neurons (Garcia-Alonso, 2000).

Htl and Egfr exhibit some specificity in their effects on OSS axon guidance. Htl seems to be preferentially required by OP axons and BM axons to project to the brain, while Egfr seems to be more involved in BM axon attachment and extension in the epithelium. In contrast to in vitro studies in vertebrates, no evidence has been found that Htl promotes axon growth. Rather, in the Drosophila OSS, it seems that the Egfr is preferentially required for outgrowth. This discrepancy with the vertebrate in vitro studies could be explained if the Drosophila in vivo situation were more prone to the deployment of compensatory molecular interactions (which might mask a role of Htl in axon growth) than the in vitro situation. In such a case, a deficit of one RTK could be partially compensated by an increase in the activity of the other RTK. This could happen if, for example, different RTKs were regulated through a common negative feedback loop. This explanation would also help explain the presence of a mild phenotype in consititutively active lambda-Htl individuals and would account for those cases in which some nrg OSS alteration is enhanced by an increase or suppressed by a reduction in RTK activity. This explanation is also consistent with the lack of effect of the gain of function of one RTK over the loss of function of the other (since this condition would itself increment the activity of the former RTK) (Garcia-Alonso, 2000).

It is likely that the RTKs also mediate the function of signals other than Nrg during OSS axon guidance. This is suggested by the fact that the nrg phenotype is weaker than the RTK phenotype and by the fact that it can be enhanced by a reduction of 50% in the amount of Egfr. These signals could represent other CAMs or growth factor ligands diffusing from the brain. One suggestive possibility is that OP and BM axons depart from the head capsule in response to some as yet unidentified diffusible signal from their brain targets (acting in addition to Nrg signaling). Further studies will be necessary to evaluate this possibility (Garcia-Alonso, 2000).

What would be the signal that triggers Nrg-dependent RTK activation? Nrg, like its vertebrate homolog L1, can behave as a homophilic CAM. It has been proposed that the homophilic interaction of L1 would activate the function of the Fgfr. Therefore, it is proposed that the homophilic interaction between Nrg¹⁸⁰ (the neural-specific form) molecules would give a positive input on RTK activity in both OP axons (from the beginning of axon extension) and BM axons (after several mechanosensory axons have converged), which would signal the axons to lift off from the epidermis and project to brain targets. However, BM axons initially extend as single processes and interact with the Nrg¹⁶⁷ form in the epithelium, where this interaction would result in the activation of the RTKs, and RTK signaling would promote extension in the epidermis. This shift in the RTK's activity outcome might be caused by the involvement of some other molecule specifically interacting with Nrg¹⁶⁷. In agreement with this idea, rescue experiments of nrg using Nrg¹⁸⁰ reveal that BM axon extension on the epidermis cannot be implemented by this molecular form. Since homophilic binding between the different Nrg forms is possible, this high degree of specificity suggests the existence of additional molecules (specifically interacting with Nrg¹⁶⁷) that mediate BM axon association with the epithelium. Some observations are consistent with this model. (1) OP axons fasciculate with one another (50 or so per ocellus) from the very beginning of axon extension. These axons fasciculate together due to the presence of Neurotactin and other CAMs in the membrane. Defasciculation of OP axons (caused by a lack of Neurotactin) increases the chances that OP axons extend abnormally in the epidermis. These results suggest that fasciculation helps generate a robust process of axon guidance. In this model, defasciculation would reduce the probability of Nrg¹⁸⁰–Nrg¹⁸⁰ interactions between OP axons, producing a deficit of RTK activity, therefore, making them more likely to extend attached to the epidermis. (2) BM axons initially follow the epidermis in isolation from other axons but begin to converge as they approach the dorsal antennal field. After several BM axons have converged together, the BM fascicle lifts off from the epidermis. Thus, the signal for BM nerves to leave the epidermis might be a given threshold of Nrg¹⁸⁰–Nrg¹⁸⁰ interactions between the different BM axons. If this model is correct, the specificity of the OP and BM growth cone interaction with the epidermis would reside in the way Nrg¹⁶⁷ and Nrg¹⁸⁰ differ in their interactions with other molecule(s). Nrg¹⁶⁷ differs from Nrg¹⁸⁰ in the cytoplasmic domain. It has been previously shown that the cytoplasmic domain can regulate the adhesive properties of the extracellular part of the protein. Therefore, it is possible that some CAM molecule might specifically interact with Nrg¹⁶⁷ to help promote initial RTK activation and attachment to the epithelium in BM axons. This molecule could represent the Drosophila homolog of some of the known vertebrate heterophilic partners of L1 (Garcia-Alonso, 2000).

In summary, these results strongly support the idea that Neuroglian functions in axon guidance by regulating the activity of both the Fgfr and the Egfr and suggest a scenario where other CAMs and growth factor activities would act in concert with Neuroglian for RTK regulation. In addition, the possible existence of cross-regulatory circuits between RTKs would add another level of control that might help to understand the high degree of canalization displayed by the axon guidance process from Drosophila to mammals (Garcia-Alonso, 2000).

Endocytic pathways downregulate the L1-type cell adhesion molecule Neuroglian to promote dendrite pruning in Drosophila

Pruning of unnecessary axons and/or dendrites is crucial for maturation of the nervous system. However, little is known about cell adhesion molecules (CAMs) that control neuronal pruning. In Drosophila, dendritic arborization neurons, ddaCs, selectively prune their larval dendrites. This study reports that Rab5/ESCRT-mediated endocytic pathways are critical for dendrite pruning. Loss of Rab5 or ESCRT function leads to robust accumulation of the L1-type CAM Neuroglian (Nrg) on enlarged endosomes in ddaC neurons. Nrg is localized on endosomes in wild-type ddaC neurons and downregulated prior to dendrite pruning. Overexpression of Nrg alone is sufficient to inhibit dendrite pruning, whereas removal of Nrg causes precocious dendrite pruning. Epistasis experiments indicate that Rab5 and ESCRT restrain the inhibitory role of Nrg during dendrite pruning. Thus, this study demonstrates the cell-surface molecule that controls dendrite pruning and defines an important mechanism whereby sensory neurons, via endolysosomal pathway, downregulate the cell-surface molecule to trigger dendrite pruning (Zhang, 2014).

Endocytic pathways profoundly regulate turnover and homeostasis of various cell-surface adhesion proteins and guidance receptors in the developing nervous systems. Perturbation of endocytic pathways often leads to a variety of neurodegenerative diseases, such as frontotemporal dementia, amyotrophic lateral sclerosis, Alzheimer's disease, lysosomal storage diseases, and Niemann-Pick disease. In Drosophila, the endolysosomal pathway is activated in neighboring glia to engulf degenerating axon/dendrite fragments for their subsequent breakdown during pruning, suggesting a non-cell-autonomous role. This study reports that Rab5 and the ESCRT complexes, two key endocytic regulators, cell autonomously promote dendrite pruning in ddaC neurons. Consistent with these findings, the endocytic pathways also play a cell-autonomous role in axon pruning of MB γ neurons. How do Rab5/ESCRT-dependent endocytic pathways facilitate dendrite pruning in ddaC neurons at the cellular level? This study has identified a cell-surface adhesion protein, namely the L1-type CAM Nrg, as a target of Rab5/ESCRT-dependent endocytic pathways (Zhang, 2014).

Drosophila Nrg and the mammalian L1-type CAMs regulate axonal growth and guidanc, synaptic stability and function, and axon/dendrite morphogenesis. Mutations in the human L1 CAM gene have been reported to cause a broad spectrum of neuronal disorders. This study has identified the Drosophila L1-type CAM Nrg as the key cell-surface molecule that inhibits dendrite pruning in ddaC neurons. The extracellular domains of the L1-type CAMs can regulate cell-cell adhesion via homophilic and/or heterophilic interactions, whereas their intracellular domains can link the proteins with F-actin/spectrins to stabilize the cytoskeletal architecture. In C. elegans, a ligand-receptor complex of cell adhesion molecules containing the nematode Nrg homolog controls dendrite-substrate adhesion to stabilize and pattern dendritic arbors in certain sensory neurons. In Drosophila, Nrg-mediated cell adhesion plays an essential role in stabilizing synapse growth and maintenance at the larval neuromuscular junction. Likewise, Nrg may also mediate adhesion of the dendrites to their adjacent epidermis to stabilize the dendritic architecture in ddaC sensory neurons, whereas downregulation of Nrg may reduce dendritic adhesion/stability and result in disassembly of dendrites. Consistent with the potential adhesive role, structure-function analysis indicates that the ECD of Nrg is important for its function in stabilizing dendrite and/or inhibiting dendrite pruning in ddaC neurons. The fact that overexpression of the ICD-deleted Nrg protein partially rescued the nrg¹⁴ mutant phenotype suggests that, in addition to the adhesion function of the ECD, the ICD of Nrg may recruit cytoskeletal components to stabilize dendritic branches in ddaC neurons. The model is therefore favored that the adhesive role of Nrg is a potential mechanism for inhibiting pruning in ddaC sensory neurons (Zhang, 2014).

Another major class of CAMs, integrins, were shown to regulate dendrite-substrate interactions and anchor ddaC dendritic arbors to the extracellular matrix. However, unlike Nrg, integrins do not accumulate on enlarged endosomes in Rab5 or ESCRT ddaC neurons, implying that integrins are not regulated by Rab5/ESCRT-dependent endocytic pathways in ddaC neurons. Moreover, other cell-surface molecules Robo and N-Cad, albeit regulated by the endolysosomal pathway in motor neurons, photoreceptors, or sensory neurons, are dispensable for normal progression of dendrite pruning in ddaC neurons. Thus, this study highlights an important role of the L1-type CAM Nrg in inhibiting dendrite pruning of ddaC sensory neurons (Zhang, 2014).

Interestingly, loss of nrg function causes precocious dendrite pruning without affecting the axonal integrity and connectivity in ddaC neurons, underscoring a specific requirement of Nrg in stabilizing the dendrites, but not the axons. Downregulation of Nrg may reduce dendritic adhesive properties of ddaC sensory neurons and thereby make the dendritic architecture more susceptible to pruning. It is conceivable that Nrg-independent mechanisms may be utilized to protect the axonal structure from the pruning machinery in ddaC neurons. Moreover, both nrg loss of function and gain of function did not affect axon pruning in MB γ neurons (data not shown), further supporting the conclusion that Nrg plays a specific role in dendrite pruning in ddaC sensory neurons. Future studies may elucidate whether and how Nrg mediates its dendritic adhesive properties to inhibit dendrite pruning (Zhang, 2014).

In summary, this study shows that Rab5/ESCRT-dependent endocytic pathways facilitate dendrite pruning of ddaC neurons by downregulating the Drosophila L1-type CAM Nrg during early metamorphosis. This study also demonstrated the role of the cell-surface adhesion protein Nrg in inhibiting dendrite pruning in ddaC sensory neurons. Thus, this study opens the door for further studies of the functions of cell-surface molecules in the regulation of dendritic adhesion during neuronal remodeling (Zhang, 2014).

GENE STRUCTURE

Bases in 3' UTR - 117

PROTEIN STRUCTURE

Neuroglian RNA undergoes tissue specific alternative splicing. The alternative proteins differ in their cytoplasmic domains. The longer form is expressed on the surface of neurons in the CNS and some glial support cells in the PNS. The shorter form is widely expressed in other cells and tissues (Hortsch, 1990, and Bieber, 1989).

To identify neuronal cell surface glycoproteins in the Drosophila embryo, antisera against horseradish peroxidase (HRP) was used to recognize a carbohydrate epitope that is selectively expressed in the insect nervous system. A large number of neuronal glycoproteins (denoted "HRP proteins") apparently bear the HRP carbohydrate epitope. Polyclonal anti-HRP antibodies were used to purify these proteins from Drosophila embryos. Three major HRP proteins are Neurotactin, Fasciclin I, and an R-PTP, DPTP69D. Western blotting data suggest that Fasciclin II, Neuroglian, DPTP10D, and DPTP99A are also HRP proteins (Desai, 1994).

Amino Acids - 1302

Structural Domains

Neuroglian has six immunoglobulin C2-type domains followed by five fibronectin type III domains (Bieber, 1989).

There are two adjacent fibronectin type III repeats in Neuroglian. Each domain consists of two antiparallel beta sheets and is folded topologically identically to fibronectin type III domains of Tenascin and Fibronectin. The hydrophobic interdomain interface includes a predicted metal-binding site, presumably involved in stabilizing the relative orientation between domains. The predicted metal binding domain is also present in the vertebrate homolog molecule L1. The Neuroglian domains are related by a near perfect 2-fold screw axis along the longest molecular dimension (Huber, 1994).

date revised: 19 June 2024

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.