Protein Interactions

The LG4 (laminin G-like) domain of the laminin alpha4 chain is responsible for the significantly higher affinity of the alpha4 chain to heparin than found for other alpha chains; four basic residues were identified to be essential for this activity. By creating GST (glutathione S-transferase)-fused LG1, LG2, LG4 and LG5 proteins, it was found that only LG4 is active for the adhesion of human HT1080 cells, human umbilical vein endothelial cells and Drosophila haemocytes Kc167 with a half-saturating concentration of 20 microg/ml. Adhesion was counteracted by treatment of the cells with heparin, heparan sulphate and heparitinase I. Upon mutating the four basic residues essential for heparin binding within LG4, the adhesion activity was abolished. Pull-down experiments using glutathione beads/GST-fusion proteins indicate a direct interaction of LG4 with syndecan-4, which might be the major receptor for cell adhesion. Neither the release of glypican-1 by treating human cells with phosphatidylinositol-specific phospholipase C nor targeted knockdown of dally or dally-like protein impaired the cell-adhesion activity. Since the LG4-LG5 domain of the alpha4 chain is cleaved in vivo from the main body of laminin-8 (alpha4beta1gamma1), it is suggested that the heparan sulphate proteoglycan-binding activity of LG4 is significant in modulating the signalling of Wnt, Decapentaplegic and fibroblast growth factors (Yamashita, 2004).

Heparin-column chromatography and elastase-digestion of medium from hemocyte Kc167 gave Drosophila laminin alpha3/5betagamma trimer, alpha3/5LG2-3 and alpha3/5LG4-5 modules with eluting NaCl concentrations of 450, 280 and 450 mM, respectively. Kc167 cells bound dish surface with alpha3/5betagamma trimer or alpha3/5LG4-5, but not with alpha3/5LG2-3 modules. Cell binding was counteracted by treating with heparin or heparan sulfate. RNA interference of Syndecan in Kc167 cells impaired the binding, but that of dally or dally-like did not. Green fluorescent protein-expressing hemocytes also bound surface with alpha3/5betagamma trimer or alpha3/5LG4-5 module. Thus, Syndecan-dependent binding of hemocytes to laminin may have a potential role in sessile hemocytes islets formation in T2-A8 segments of Drosophila (Narita, 2004).

The heparan sulfate proteoglycan syndecan is an in vivo ligand for the Drosophila LAR receptor tyrosine phosphatase

Receptor tyrosine phosphatases (RPTPs) are essential for axon guidance and synaptogenesis in Drosophila. Each guidance decision made by embryonic motor axons during outgrowth to their muscle targets requires a specific subset of the five neural RPTPs. The logic underlying these requirements, however, is still unclear, partially because the ligands recognized by RPTPs at growth cone choice points have not been identified. RPTPs in general are still 'orphan receptors' because, while they have been found to interact in vitro with many different proteins, their in vivo ligands are unknown. This study uses a new type of deficiency screen to identify the transmembrane heparan sulfate proteoglycan Syndecan (Sdc) as a ligand for the neuronal RPTP LAR. LAR interacts with the glycosaminoglycan chains of Syndecan in vitro with nanomolar affinity. Genetic interaction studies using Sdc and Lar LOF mutations demonstrate that Sdc contributes to LAR's function in motor axon guidance. Overexpression of Sdc on muscles is shown to generate the same phenotype as overexpression of LAR in neurons, and genetic removal of LAR suppresses the phenotype produced by ectopic muscle Sdc. Finally, it is shown that there is at least one additional, nonproteoglycan, ligand for LAR encoded in the genome. Taken together, these results demonstrate that Sdc on muscles can interact with neuronal LAR in vivo and that binding to Sdc increases LAR's signaling efficacy. Thus, Sdc is a ligand that can act in trans to positively regulate signal transduction through LAR within neuronal growth cones (Fox, 2005).

Genetic removal of Sdc from embryos bearing Lar mutations increases the penetrance of the characteristic Lar ISNb bypass phenotype. This effect on penetrance is as large as those usually observed when a second RPTP is genetically removed from a single Rptp mutant (e.g., removal of DPTP69D from a Lar mutant). Removal of Sdc increases penetrance for both hypomorphic and (zygotic) null Lar mutations. The effect on the null penetrance is likely due to reduction of maternally contributed LAR function. However, it was found that Sdc mutations alone do not produce ISNb bypass at a significant frequency, even when both maternal and zygotic Sdc are removed. One explanation for this finding might be that Sdc is partially redundant with Dally and/or Dlp, since these are cell-surface HSPGs expressed in a similar pattern to Sdc (Fox, 2005).

To test this model, Dally and Dlp expression were reduced in an Sdc mutant background. It is difficult to assess the appropriate extent of reduction for this experiment. Glypicans cannot be completely removed, since this would produce embryos with severe early phenotypes due to loss of Hedgehog and Wingless signaling. Zygotic triple mutants (Sdc dally dlp) were generated, and, also, dally and dlp dsRNAs were injected into Sdc maternal/zygotic mutant embryos (dally/dlp RNAi would affect both maternal and zygotically contributed mRNAs). The genetic triple mutants had CNS phenotypes that are stronger than the Sdc phenotype but displayed few motor axon guidance errors. Dally/Dlp-injected Sdc mutant embryos had more severe phenotypes, but ISNb guidance was not selectively affected. Overall, the data suggest that Sdc is not redundant with Dally or Dlp and that its absence is likely to be compensated for by non-HSPG proteins. Perhaps the second LAR ligand detected by embryo staining is redundant with Sdc. Like Sdc, this ligand is expressed both on CNS axons and in lines in the periphery (Fox, 2005).

A genetic epistasis experiment demonstrates that Sdc acts in trans (as a ligand) to regulate LAR function. However, the data allow this conclusion to be reached only for SNa bifurcation, which is affected by LAR overexpression but not by loss of LAR. In its regulation of the decision of ISNb growth cones to enter the muscle field, Sdc could also act as a ligand, since it is expressed at the appropriate time on patches of cells near the muscle field entry site that could be contacted by LAR-expressing ISNb growth cones at this choice point. Alternatively, Sdc could act as a coreceptor at this choice point since it is expressed on the motor nerves together with LAR. Finally, it is not known if the Sdc that interacts with LAR during ISNb axon guidance is attached to the cell surface or has been shed by proteolytic cleavage. If released then Sdc is the essential ligand; Sdc could be expressed by either muscles or neurons and transported to the choice point (Fox, 2005).

These results show that the cell-surface HSPG Sdc is an in vivo ligand for LAR and indicate that it positively regulates LAR signaling during motor axon guidance. Sdcís GAG chains bind directly to LAR with high affinity, and this binding requires basic sequences in the first Ig domain of LAR. Further work will be required to determine whether binding to Sdc directly stimulates LARís phosphatase activity, relocalizes LAR within the growth cone,or facilitates LAR signaling by another mechanism (Fox, 2005).

The HSPGs Syndecan and Dallylike bind the receptor phosphatase LAR and exert distinct effects on synaptic development

The formation and plasticity of synaptic connections rely on regulatory interactions between pre- and postsynaptic cells. The Drosophila heparan sulfate proteoglycans (HSPGs) Syndecan (Sdc) and Dallylike (Dlp) are synaptic proteins necessary to control distinct aspects of synaptic biology. Sdc promotes the growth of presynaptic terminals, whereas Dlp regulates active zone form and function. Both Sdc and Dlp bind at high affinity to the protein tyrosine phosphatase LAR, a conserved receptor that controls both NMJ growth and active zone morphogenesis. These data and double mutant assays showing a requirement of LAR for actions of both HSPGs lead to a model in which presynaptic LAR is under complex control, with Sdc promoting and Dlp inhibiting LAR in order to control synapse morphogenesis and function (Johnson, 2006).

Converging lines of evidence suggest that membrane-associated HSPGs serve an important purpose in the assembly, function and plasticity of excitatory synapses. The ancient HSPG families of syndecans and glypicans are necessary for Drosophila to regulate distinct aspects of synaptic morphogenesis. Genetic and biochemical data indicate that Sdc and Dlp interact with LAR to control presynaptic properties. Because LAR-family RPTPs have been shown to control the formation of excitatory synapses in Drosophila, C. elegans, and mammals, these findings may represent a more general mechanism for regulating synaptic morphogenesis and function. Despite the importance of LAR-family RPTPs during cellular morphogenesis inside and outside of the nervous system, the lack of physiologically relevant extracellular binding partners has made it challenging to study this well-conserved group of receptors (Johnson, 2006).

The data show that Sdc promotes the formation of presynaptic boutons. This Sdc function appears to be mediated by LAR, as supported by parallel phenotypes, direct binding, in vivo colocalization, and three types of double mutant analysis between Sdc and LAR. Despite the fact that Sdc exhibits gain-of-function activity and endogenous expression on both sides of the synapse, neuronal and muscle-specific rescue experiments show that Sdc function is mainly presynaptic. Because LAR is required only in neurons to promote synapse growth, these findings support a model in which Sdc acts as a neuronal cell-autonomous agonist of LAR. This is somewhat surprising, given the fact that soluble forms of Sdc bind to LAR and that endogenous Sdc appears to be released from the presynaptic membrane to fill the subsynaptic reticulum. One way for Sdc to act presynaptically would be to bind to LAR even before the two proteins are presented on the neuronal surface. Because Dlp has a competitive advantage over Sdc for binding to LAR, a prebound complex of Sdc and LAR would have the ability to stimulate synapse growth before the phosphatase could be inhibited by Dlp. Such a mechanism could provide a time- and/or HSPG concentration-dependent switch from bouton addition to active zone assembly (Johnson, 2006).

Sdc could promote LAR activity in collaboration with an additional cell-type-specific membrane protein. Data from a parallel study has also identified Sdc-LAR interactions during embryonic motor axon guidance. However, in contrast to CNS pathfinding, complete loss of Sdc alone has no significant effect on motor pathfinding, suggesting that additional LAR ligands exist in the early embryo. Recent experiments with the vertebrate LAR ortholog PTP-s suggest that non-HSPG ligands may regulate its ability to promote retinal axon outgrowth. Although it remains a formal possibility, an additional ligand may not be needed to account for the NMJ growth-promoting activity of LAR because the larval synaptic phenotype in Sdc mutants is nearly as strong as the growth defect in LAR mutants (Johnson, 2006).

Active zone assembly is vital for neurotransmission at the synapse and has been proposed as a means to modulate synaptic function over time. Analysis of Dlp reveals that synaptic glypicans are required to regulate active zone morphology and function. Moreover, Dlp is limiting for active zone morphogenesis, consistent with an instructive role. Because the activities of Dlp appear opposite to those of LAR, it is proposed that the high affinity binding of Dlp to LAR induces an inhibition of receptor function. This hypothesis is supported by the double RNAi experiments showing that the LAR effect on Ena phosphorylation is epistatic to the effect of Dlp, indicating that Dlp acts upstream of LAR. Because loss of Dlp at the NMJ did not induce a significant change in the number of presynaptic boutons, the results lead to a model in which Dlp is specialized for control of active zone properties. Such a function might provide a means to independently regulate and spatially distinguish LAR inhibition from LAR activation. In any case, the presence of active zone phenotypes in dlp but not in Sdc reveals specialization among synaptic HSPGs (Johnson, 2006).

LAR regulates both NMJ growth and active zone morphogenesis. Thus, LAR appears to provide a link between two important synaptic properties that are regulated by different extracellular factors. A mechanism to couple bouton growth and active zone formation would make sense because active zones appear early in the nascent bouton. Because LAR catalytic activity is necessary for bouton addition, and yet LAR inhibition by Dlp appears necessary for proper active zone formation, LAR's role at the active zone may be primarily structural. For example, LAR may simply provide an anchorage point for synaptic components like the scaffolding protein Liprin-alpha that regulates active zone formation. Alternatively, LAR may exist in distinct yet active signaling states, one of which is dependent on PTP activity (promoting synapse growth), and one of which is dependent on recruitment of signaling molecules (controlling active zone assembly). Because loss of Dlp or LAR has opposite effects on quantal content at the NMJ, it is attractive to speculate that the Dlp-LAR pathway normally provides a means to modulate the strength of neurotransmission, either during NMJ growth or during synaptic plasticity. LAR PTPs are required for normal physiology and plasticity at mammalian hippocampal synapses (Johnson, 2006).

Sdc and Dlp are both HSPGs that bind to LAR and thus might be expected to act similarly, but the results show that their functions are different. One way to account for the specificity might be a difference in the effect of soluble versus cell-surface HSPGs on LAR. Some ligand molecules such as Ephrins function when clustered at high density (e.g., on a membrane surface) but fail to activate their receptors when presented in a soluble, monomeric form. Another possibility could be that LAR binding or signaling is differentially influenced by direct protein-protein interactions with the Sdc versus Dlp core proteins. The two HSPGs have very different core structures; Sdc is a transmembrane molecule with HS modification sites near the N terminus, whereas Dlp is a GPI-anchored protein with HS sites proximal to the membrane and a large disulphide bonded globular domain located more distally. It may also be relevant that Dlp consistently binds more effectively to LAR than Sdc, with KD measurements in solution showing an affinity approximately 2-fold higher. These results suggest a competition model in which Dlp displaces Sdc, possibly to favor the stabilization of active zones after new growth at the synapse. In this model, presynaptic growth would be initially promoted by Sdc and would then be limited or halted by Dlp binding after formation of close membrane contact between the nerve and muscle. Such a mechanism could insure a transition from growth to synapse stabilization and could participate in subsequent maintenance or plasticity of the synapse (Johnson, 2006).

Of course other molecules influence synapse size, and these might include coligands or coreceptors that may bind to Sdc, Dlp, and/or LAR. Potential candidates might include bone morphogenic protein (BMP), the type II BMP receptor Wishful thinking (Wit), or the Wnt ortholog Wingless (Wg), which have all been shown to regulate NMJ morphology in Drosophila. However, in addition to significant phenotypic differences compared to the HSPGs, overexpression studies indicate that neither BMP nor Wg are limiting for NMJ morphogenesis. In contrast, Sdc and Dlp are both limiting for different aspects of synapse development, consistent with an instructive role in this context. Consistent with this notion, Syndecan-2 is sufficient to promote dendritic spine maturation during hippocampal synaptogenesis in culture. Although vertebrate Syndecan-2 has yet to be tested at the synapse by loss of function, the colocalization and parallel biology of Synecan-2 and vertebrate LAR-family receptors strongly suggest conservation in the regulation of synaptic LAR (Johnson, 2006).

The genetic and biochemical studies described in this study have identified a partnership between HSPGs and LAR in Drosophila NMJ development that sets precedents for (1) the in vivo requirement for members of the syndecan and glypican families in synapse growth and electrophysiological function, (2) the specificity of HSPG function at the synapse, with distinct actions of Sdc and Dlp, and (3) biochemical identification of Sdc and Dlp as LAR binding partners, plus genetic evidence to place them in a pathway regulating biological function at the synapse (Johnson, 2006).

Embryonic

Syndecan is prominent in lymph glands (hematopoetic organs), in the PNS and CNS and along the basal surfaces of gut ectoderm (Spring, 1994).

The majority of cell surface HS polymers are carried by two classes of HSPGs: the transmembrane Syndecans and the glysosylphosphatidyl inositol (GPI)-linked Glypicans. While mammalian species express a number of genes in each of these classes, Drosophila has one Syndecan gene (sdc) and two Glypican genes (dally and dallylike). Although some degree of core-protein specificity has been observed in functional comparisons between different HSPGs during early pattern formation and cell fate determination, true loss-of-function mutations have yet to be analyzed for sdc and dallylike (dlp). Although anti-Sdc antibodies had been previously described, this study has developed methods to image Sdc protein at improved resolution and showed that the antibody recognizes Sdc and that Sdc localizes to both longitudinal and commissural axons pathways within the Drosophila embryonic CNS (Johnson, 2004).

Effects of Mutation or Deletion

The presentation of secreted axon guidance factors plays a major role in shaping central nervous system connectivity. Recent work suggests that heparan sulfate (HS) regulates guidance factor activity; however, the in vivo axon guidance roles of its carrier proteins (heparan sulfate proteoglycans, or HSPGs) are largely unknown. The HSPG Syndecan (Sdc) is critical for the fidelity of Slit repellent signaling at the midline of the Drosophila CNS, consistent with the localization of Sdc to CNS axons. sdc mutants exhibit consistent defects in midline axon guidance, plus potent and specific genetic interactions supporting a model in which HSPGs improve the efficiency of Slit localization and/or signaling. Testing this hypothesis, Slit distribution was shown to be altered in sdc mutants and Slit and its receptor were shown to bind to Sdc. However, when the function of the transmembrane Sdc was compared to a different class of HSPG that localizes to CNS axons (Dallylike), functional redundancy was found, suggesting that these proteoglycans act as spatially specific carriers of common HS structures that enable growth cones to interact with and perceive Slit as it diffuses away from its source at the CNS midline (Johnson, 2004).

The availability of two P-element insertions (P10608 and KG06163) into the sdc locus made it possible to address sdc function by making a small excision-induced deletion [Df(2R)48] removing the first two exons of sdc, including the promotor and 5′ untranslated region, the translational start codon, and the signal sequence. To confirm the prediction that this deletion eliminates Sdc protein expression, Df(2R)48 embryos were stained with an antibody that recognizes the extracelluar domain near the transmembrane region of Sdc (Spring, 1994). Only a minute residual signal was found that may represent either nonspecific background or limited perdurance of maternally loaded Sdc. Importantly, simultaneous staining of these mutants with anti-Robo antibodies revealed that although levels of Robo expression appeared normal in Df(2R)48, Robo-positive axons now crossed the CNS midline where Slit concentration is normally at its highest. It was also found that ventral muscles overshoot their insertion sites in strong sdc mutants, reminiscent of defects in slit mutants. To be certain that perturbation of the neighboring genes sara and FKB13 was not responsible for the guidance errors, mutations in each locus were examined and also a sara rescue construct was introduced into the Df(2R)48 background. It was found that these flanking genes do not contribute to the midline phenotype of Df(2R)48. Because HS is known to facilitate cell fate decisions in other contexts, both neuronal and midline glial patterning were examined, but no defects were found. Moreover, levels of Slit expression in midline glia appeared to be comparable to wild-type levels (Johnson, 2004).

To quantify the midline guidance defects in different sdc alleles, mAb 1D4 was used to visualize ipsilateral axon fascicles and scored for the frequency of ectopic midline crossing. An allelic series of phenotypic penetrance was found ranging from 5%-40%, suggesting that the efficiency of axon guidance depends on the amount of Sdc present. However, because Sdc family members are known to be proteolytically processed and released from the cell surface, it was important to determine whether Sdc functions autonomously in axons. A sdc cDNA under the control of the GAL4 upstream activating sequence was used, and its ability to rescue the Df(2R)48 phenotype under the control of either a midline glial-specific GAL4 source (slit-GAL4) or a postmitotic neuron-specific source (elav-GAL4) was examined. Only neuronal expression of Sdc rescues the guidance errors, suggesting that Sdc acts locally to increase growth cone sensitivity to Slit (Johnson, 2004).

Because sdc axon phenotypes suggest the Slit signaling system's failure to restrict midline crossing, it was asked if sdc mutations display specific genetic interactions with mutations in slit or its receptors. Using the same assay that identified Slit as the Robo ligand, embryos transheterozygous for sdc and mutations in several loci were compared. Highly significant (p < 0.005) interactions were found between sdc and both slit and robo in this assay. This interaction appears to be specific because no enhancement is observed when sdc is combined with a mutation in the receptor tyrosine phosphatase gene DPTP69D, which is known to contribute to midline guidance. Although no interaction is seen between sdc and single mutations in robo2 or robo3, crosses between sdc and double mutants removing robo and one of the other robo family genes (e.g., robo, robo2) reveal significant increases in the interaction when they are compared to robo alone. These genetic results suggest that Sdc acts in the Slit-Robo pathway (Johnson, 2004).

A biochemical assay was developed to determine whether Sdc binds to Slit and/or Robo in cellular extracts in which all three proteins are endogenously expressed. Immunoprecipitation of either Slit or Robo and subsequent detection with anti-Sdc antibodies reveals that Sdc associates with both Slit and its receptor, suggesting the possibility of a ternary complex. This association is specific because no Sdc is trapped by nonspecific IgG or N-Cadherin antibodies that successfully immunoprecipitate other signaling molecules. Thus, Sdc participates directly in a complex with Slit and Robo (Johnson, 2004).

Having gathered multiple lines of evidence revealing a role for Sdc in Slit signaling, popular models for the underlying mechanism could be tested. HS has been proposed to support the patterning activity of several secreted ligands by restricting their diffusion from a focal source. If this were the case for Sdc in Slit signaling, a change in the distribution of Slit at sites distant from the midline glia would be predicted. To examine this, an immunohistochemical method was optimized to allow visualization of Slit protein not only on the surface of midline glia but also within the CNS neuropil. No gross qualitative difference was found in the pattern of Slit distribution between Df(2R)48 and wild-type controls. However, rigorous quantitative analysis of confocal images with an independent axon surface marker (LAR) as an internal control to ensure comparable signal strength reveals that sdc mutants show a highly significant change in the pattern of Slit accumulation throughout the neuropil. Slit is still highly expressed on midline glia in sdc mutants, but the Slit signal is significantly reduced in the neuropil. These data are highly reproducible and are consistent with the ligand-trapping model anticipated from previous studies on HS (Johnson, 2004).

The striking difference in protein structure between Syndecans and Glypicans raises two related questions: (1) is there core-protein specificity to HSPG function on the growth cone surface, or are these proteins acting mainly as carriers for a common HS structure? and (2) are the cytoplasmic domains of Sdc, known to interact with signaling proteins inside the cell, essential for Sdc function during axon guidance? Both of these questions were answered by testing the specificity of Sdc relative to anther HSPG. Of course, it was important to compare Sdc to an HSPG normally localized to axons if possible. Using an existing antibody, it was possible to show that Glypican Dallylike (Dlp) is expressed on the surface of embryonic axons in a pattern nearly identical to that of Sdc. This antibody recognizes Dlp ectopically expressed by a UAS-dlp transgene. When UAS-dlp was expressed in a Df(2R)48 background under the control of a Slit-GAL4 driver, no significant rescue was found of the sdc midline phenotype. However, neuron-specific expression of Dlp generated a highly significant degree of functional rescue (p < 0.005). This clearly shows that an increase in Dlp expression can compensate for the loss of Sdc, consistent with the finding that Slit can bind to at least one mammalian Glypican. In addition, this experiment shows that the unique intracellular signaling motifs found in Sdc are not essential for Slit signaling (Johnson, 2004).

In conclusion, it has been found that Sdc localizes to developing axons, is required for accurate growth cone navigation at the CNS midline, and interacts genetically and physically with Slit and Robo. Although a full account of HSPG functional specificity awaits the analysis of mutations in dallylike, the fact that both Syndecan and Glypican can serve interchangeably to improve the efficiency of growth cone repulsion suggests a model in which the total amount of cell surface HS determines the sensitivity of Robo-expressing growth cones to the midline repellent. Of course, the possibility cannot be ruled out that the highly conserved cytoplasmic domains of Sdc play a more subtle modulatory function in this context or an essential function in some distinct context. This might explain the difference in the efficiency of Dlp and Sdc in the rescue of sdc guidance defects. These and other answers will come from future dissection of the Sdc mechanism and the action of HSPGs in neural development (Johnson, 2004).

The heparan sulfate proteoglycans Dally-like and Syndecan have distinct functions in axon guidance and visual-system assembly in Drosophila

Heparan sulfate proteoglycans (HSPGs), a class of glycosaminoglycan-modified proteins, control diverse patterning events via their regulation of growth-factor signaling and morphogen distribution. In C. elegans, zebrafish, and the mouse, heparan sulfate (HS) biosynthesis is required for normal axon guidance, and mutations affecting Syndecan (Sdc), a transmembrane HSPG, disrupt axon guidance in Drosophila embryos. Glypicans, a family of glycosylphosphatidylinositol (GPI)-linked HSPGs, are expressed on axons and growth cones in vertebrates, but their role in axon guidance has not been determined. This study demonstrates that the Drosophila glypican Dally-like protein (Dlp) is required for proper axon guidance and visual-system function. Mosaic studies reveal that Dlp is necessary in both the retina and the brain for different aspects of visual-system assembly. Sdc mutants also show axon guidance and visual-system defects, some that overlap with dlp and others that are unique. dlp⁺ transgenes are able to rescue some sdc visual-system phenotypes, but sdc⁺ transgenes are ineffective in rescuing dlp abnormalities. Together, these findings suggest that in some contexts HS chains provide the biologically critical component, whereas in others the structure of the protein core is also essential (Rawson, 2005).

The distribution of Dlp was examined in the developing visual system by using a monoclonal antibody that specifically recognizes Dlp in tissues. In Drosophila, the adult eye is comprised of approximately 800 sensory units, or ommatidia, each with eight distinct photoreceptors, R1-R8. In the eye imaginal disc, Dlp was found on photoreceptor cell bodies and on cells within the morphogenetic furrow. Axons from R1-R6 terminate in the lamina, the first optic ganglion, whereas those from R7 and R8 project to the medulla. In the optic lobe, Dlp is present on photoreceptor axons at the boundary between the lamina and adjacent tissues and along the lamina plexus. Dlp is also observed in the medulla neuropil, medulla glia, medulla neuropil glia, neuroblasts of the proliferative centers, and in the mushroom body neuropil (Rawson, 2005).

To evaluate the function of Dlp in axon guidance, a photoreceptor-specific monoclonal antibody (24B10) was used to visualize photoreceptor projections in dlp mutants. In 50% of dlp mutant hemispheres, the lamina plexus was irregular and thickened. Additionally, 80% of dlp mutant larvae had fibers that aberrantly crossed between ommatidial bundles and/or photoreceptor process expansions outside the normal termination zone of the lamina plexus. Examination of dlp mutant pupae revealed that 80% of optic lobes contain irregularities in the R7 and R8 medulla termini. Crossover of R7 axons to neighboring medulla cartridges was observed (~50%) and misrouting of R7/R8 axons (~20%) (Rawson, 2005).

Visual-system function was assessed in dlp mutants by recording electroretinogram (ERG) profiles in adult flies. A wild-type ERG is composed of the photoreceptor response generated by a light-induced depolarization of the photoreceptor neurons and of the two transient voltage changes, the 'on-transient' and the 'off-transient', resulting from currents related to synaptic transmission during the initiation and termination of the light stimulus. dlp mutants show statistically significant defects in the photoreceptor response and in both on- and off-transients, suggesting that Dlp is required for proper photoreceptor currents and synaptic transmission (Rawson, 2005).

Axon guidance in the visual system depends on photoreceptor specification, as well as on glial-cell migration and lamina-neuron differentiation. Although dlp mutants have reduced and roughened eyes, thin sections of dlp eyes demonstrate that all photoreceptors are present and retain proper polarity in each ommatidium. Staining with several photoreceptor-specific markers confirm that dlp mutant photoreceptors differentiate properly. Likewise, glial cells and lamina neurons were found in the correct number and location in dlp mutants, indicating that patterning defects of these critical cells cannot account for the observed axon-guidance defects (Rawson, 2005).

Because Dlp is expressed in several visual-system elements and cell types, somatic mosaic studies were conducted to determine which cells require dlp for visual-system assembly. Using a method that generates clones encompassing a majority of cells in the retina, crossover of axons between ommatidial bundles and photoreceptor process expansions outside the lamina plexus was observed in 67% of animals with dlp mutant photoreceptors projecting to a heterozygous brain. Conversely, R7/R8 termination defects were absent from the medulla of 40 hr pupae with dlp mutant retinas and dlp/+ optic lobes. These results indicate that Dlp is required in the eye to specify proper axon guidance to the lamina, but not to the medulla (Rawson, 2005).

The photoreceptor-specific requirement for Dlp was further evaluated via the mosaic analysis with a repressible cell marker (MARCM) technique to visualize axons from small dlp mutant clones. dlp mutant photoreceptors display ectopic axon outgrowths directed away from their proper targets in the lamina (37% of dlp/dlp axons). Ectopic axon processes were four times more prevalent on axon bundles near the boundaries of the lamina than on those in more-central regions. This suggests that expression of Dlp on photoreceptors may be important for the detection of repellant cues that prevent aberrant axon outgrowth (Rawson, 2005).

Finally, the tissue-specific requirement was evaluated for Dlp in physiological function of the visual system. Using a mosaic strategy that generated eyes composed solely of dlp mutant photoreceptors, no see statistically significant defects were seen in ERG recordings compared to controls. These results demonstrate that the ERG defects of dlp mutants are produced by loss-of-function in the optic lobe, not in the eye (Rawson, 2005).

Drosophila Syndecan (Sdc) represents another class of HSPG, and it is required for normal axon guidance in the embryo. Antibody specific for Sdc revealed that this proteoglycan is present on photoreceptors in the retina and on photoreceptor projections to the optic lobe. Like Dlp, Sdc is enriched throughout the lamina plexus and at the boundary between the lamina and adjacent tissues. Sdc immunoreactivity was also detectable at a low level on cells just medial to the lamina plexus, the medulla glia, but was absent from some Dlp-expressing cells such as the mushroom body and the medulla neuropil glia (Rawson, 2005).

Analysis of sdc null mutant larvae revealed that 50% of optic lobes had photoreceptor-projection abnormalities and/or lamina-plexus defects including gaps (29%) and gross disorganization (21%, n = 28). Compared to dlp mutants, sdc mutants showed a low penetrance of the lamina-thickening (4%) and lamina-axon-crossover phenotypes. Additionally, some sdc mutants had R7/R8-axon misrouting in the larval stage. sdc mutant pupae showed crossover of R7 axons between medullary cartridges (100%) and defective axon pathfinding to the medulla (86%) but a low penetrance of R7/R8-termini disruption (~10%), a phenotype common in dlp mutants (80%-100%). As is the case for dlp, sdc mutants do not show defects in the specification of photoreceptors, glia, or lamina neurons, confirming that the above phenotypes are not secondary to other overt patterning deficiencies. Overall, dlp and sdc mutants share similar axon-guidance phenotypes, but each has a largely distinct level of penetrance for a given defect (Rawson, 2005).

Electrophysiological analysis revealed that sdc mutants have grossly abnormal ERGs, with defective photoreceptor depolarization and complete absence of on- and off-transients. These ERG abnormalities -- particularly the virtual loss of photoreceptor depolarization -- are distinct from those found in dlp animals. In addition, whereas dlp null mutants have a reduced and roughened eye, sdc mutants do not. These findings are consistent with distinct molecular functions for Dlp and Sdc during visual-system assembly (Rawson, 2005).

Sdc and Dlp are both heparan sulfate-modified proteoglycans, and Dlp expression is capable of limited rescue of sdc axon-guidance abnormalities in the embryo. To determine whether these two cell-surface molecules have overlapping functions in the visual system, a series of rescue experiments was performed. With or without a GAL4 driver, dlp mutant animals bearing a UAS-dlp⁺ construct show complete rescue of pupal axon-pathfinding defects and ERG abnormalities, suggesting that low levels of Dlp expression are sufficient to provide the function necessary for normal axon guidance and visual-system assembly. Despite this, ubiquitous expression of Sdc was unable to rescue the axon-guidance abnormalities of dlp mutants, demonstrating that Dlp and Sdc are not functionally interchangeable during photoreceptor axon guidance (Rawson, 2005).

The ability of UAS-dlp⁺ transgenes to rescue axon-guidance phenotypes of sdc mutant larvae and pupae was tested. Pupal sdc mutant phenotypes, including misrouting of R7/R8 photoreceptors to the medulla and crossover of R7 to neighboring medulla cartridges, are largely rescued with neuronal-directed expression of Sdc. Conversely, neuron-specific expression of Sdc in sdc mutants is not sufficient to rescue photoreceptor projection defects to the larval lamina or the ERG abnormalities in adults, suggesting that expression of Sdc in additional cell types is critical for complete restoration of axonal patterning and visual-system function. In contrast to the inability of Sdc expression to rescue dlp mutant phenotypes, neuron-specific expression of Dlp restores proper R7/R8 photoreceptor projection to the medulla of sdc mutant pupae. This rescue demonstrates that a structurally distinct HSPG can provide some of the functions normally served by Sdc (Rawson, 2005).

In vertebrates, glypicans are expressed on axons and growth cones in the developing nervous system. Whereas previous findings have established a function for glypicans in growth-factor signaling, their role in axon guidance has not been reported. Consistent with the expression pattern of Dlp on axons and glial cells of the developing visual system, dlp mutants show defects in photoreceptor projections to the lamina and medulla. Mosaic analysis demonstrates that Dlp is required on photoreceptors for some, but not all, aspects of axon guidance, serving functions in both peripheral and central components of the visual system. dlp is not required in the retina for normal electrophysiological responses to light, indicating that the synaptic-transmission defects in dlp mutant adults are due to loss of Dlp activity in the optic lobes (Rawson, 2005).

Comparison of sdc and dlp mutants as well as transgene rescue experiments demonstrates that Sdc and Dlp have some overlapping functions in visual-system assembly but others that are unique. For example, photoreceptor misrouting to the medulla in sdc mutants can be rescued by neuron-specific expression of dlp⁺, consistent with the capacity of dlp⁺ to rescue midline-crossover defects in sdc mutant embryos. In contrast to the ability of dlp⁺ transgenes to rescue sdc mutants, UAS-sdc⁺ constructs are unable to rescue any of the dlp mutant phenotypes, even under conditions where very modest levels of Dlp rescues completely. These findings show that Dlp functions cannot be readily provided by another unrelated HSPG, and they argue that the conserved sequence elements of the Dlp core protein are critical for these unique functions (Rawson, 2005).

Effects of HSPGs on axon pathfinding have been considered principally in the context of classical axon-guidance pathways, Slit-Robo signaling in particular. Although there is genetic evidence from mouse, Drosophila, and C. elegans that HSPGs affect Slit-Robo signaling, there are many other possibilities. HSPGs have been extensively characterized as molecules that affect both morphogen signaling and distributions. Recent studies demonstrating that the classical morphogens Wnt, Hh, and BMP also play a bona fide role in axon guidance suggest the possibility that HSPGs govern axon guidance by affecting morphogen function or distributions during this process (Rawson, 2005).

Reference names in red indicate recommended papers.

Beauvais, D. M., Burbach, B. J. and Rapraeger, A. C. (2004). The syndecan-1 ectodomain regulates alphavbeta3 integrin activity in human mammary carcinoma cells. J. Cell Biol. 167(1): 171-81. 15479743

Burbach, B. J., Ji, Y. and Rapraeger, A. C. (2004). Syndecan-1 ectodomain regulates matrix-dependent signaling in human breast carcinoma cells. Exp. Cell Res. 300(1): 234-47. 15383330

Casar, J. C., et al. (2004). Heparan sulfate proteoglycans are increased during skeletal muscle regeneration: requirement of syndecan-3 for successful fiber formation. J. Cell Sci. 117(Pt 1): 73-84. 14627628

Chen, L., Klass, C. and Woods, A. (2004). Syndecan-2 regulates transforming growth factor-beta signaling. J. Biol. Chem. 279(16): 15715-8. 14976204

Cornelison, D. D., et al. (2004). Essential and separable roles for Syndecan-3 and Syndecan-4 in skeletal muscle development and regeneration. Genes Dev. 18(18): 2231-6. 15371336

Couchman, J.R. (2003). Syndecans: proteoglycan regulators of cell-surface microdomains?. Nat. Rev. Mol. Cell Bio. 4: 926-937. 14685171

Dealy, C. N., et al. (1997). FGF-stimulated outgrowth and proliferation of limb mesoderm is dependent on Syndecan-3. Dev. Biol. 184: 343-350.

Deepa, S. S., et al. (2004). Chondroitin sulfate chains on syndecan-1 and syndecan-4 from normal murine mammary gland epithelial cells are structurally and functionally distinct and cooperate with heparan sulfate chains to bind growth factors. A novel function to control binding of midkine, pleiotrophin, and basic fibroblast growth factor. J. Biol. Chem. 279(36): 37368-76. 15226297

Elenius. V., Gotte, M., Reizes, O., Elenius, K. and Bernfield, M. (2004), Inhibition by the soluble syndecan-1 ectodomains delays wound repair in mice overexpressing syndecan-1. J. Biol. Chem. 279(40): 41928-35. 15220342

Endo, K., et al. (2003). Cleavage of syndecan-1 by membrane type matrix metalloproteinase-1 stimulates cell migration. J. Biol. Chem. 278(42): 40764-70. 12904296

Ethell, I. M., et al. (2001). EphB/Syndecan-2 signaling in dendritic spine morphogenesis. Neuron 31: 1001-1013. 11580899

Fox, A. N. and Zinn, K. (2005). The heparan sulfate proteoglycan syndecan is an in vivo ligand for the Drosophila LAR receptor tyrosine phosphatase. Curr. Biol. 15(19): 1701-11. 16213816

Goutebroze, L., et al. (2003). Syndecan-3 and syndecan-4 are enriched in Schwann cell perinodal processes. BMC Neurosci. 4(1): 29. 14622446

Hsueh, Y. P., et al. (1998). Direct interaction of CASK/LIN-2 and syndecan heparan sulfate proteoglycan and their overlapping distribution in neuronal synapses. J. Cell Biol. 142(1): 139-51. 9660869

Hsueh, Y. P. and Sheng, M. (1999). Regulated expression and subcellular localization of syndecan heparan sulfate proteoglycans and the syndecan-binding protein CASK/LIN-2 during rat brain development. J. Neurosci. 19(17): 7415-25. 10460248

Hu, H. (2001). Cell-surface heparan sulfate is involved in the repulsive guidance activities of Slit2 protein. Nat. Neurosci. 4: 695-701. 11426225

Johnson, K. G., et al. (2004). Axonal heparan sulfate proteoglycans regulate the distribution and efficiency of the repellent Slit during midline axon guidance. Curr. Biol. 14: 499-504. 15043815

Johnson, K. G., et al. (2006). The HSPGs Syndecan and Dallylike bind the receptor phosphatase LAR and exert distinct effects on synaptic development. Neuron 49(4): 517-31. 16476662

Keum, E., et al. (2004). Syndecan-4 regulates localization, activity and stability of protein kinase C-alpha. Biochem. J. 378(Pt 3): 1007-14. 14670076

Kinnunen, T., et al. (1998). Cortactin-Src kinase signaling pathway is involved in N-syndecan-dependent neurite outgrowth. J. Biol. Chem. 273(17):10702-8.

Langford, J. K., et al. (2004). Identification of an invasion regulatory domain within the core protein of syndecan-1. J. Biol. Chem. 280(5): 3467-73. 15563454

Liang, Y., Annan, R. S., Carr, S. A., Popp, S., Mevissen, M., Margolis, R. K., and Margolis, R. U. (1999). Mammalian homologues of the Drosophila slit protein are ligands of the heparan sulfate proteoglycan glypican-1 in brain. J. Biol. Chem. 274: 17885-17892. 10364234

Lim, S. T., Longley, R. L., Couchman, J. R. and Woods, A. (2003). Direct binding of syndecan-4 cytoplasmic domain to the catalytic domain of protein kinase C alpha (PKC alpha) increases focal adhesion localization of PKC alpha. J. Biol. Chem. 278(16): 13795-802. 12571249

McQuade, K. J. and Rapraeger, A. C. (2003). Syndecan-1 transmembrane and extracellular domains have unique and distinct roles in cell spreading. J. Biol. Chem. 278(47): 46607-15. 12975379

Midwood, K. S., Valenick, L. V., Hsia, H. C. and Schwarzbauer, J. E. (2004). Coregulation of fibronectin signaling and matrix contraction by tenascin-C and syndecan-4. Mol. Biol. Cell. 15(12): 5670-7. 15483051

Minniti, A. N., Labarca, M., Hurtado, C. and Brandan, E. (2004). Caenorhabditis elegans syndecan (SDN-1) is required for normal egg laying and associates with the nervous system and the vulva. J. Cell Sci. 117(Pt 21): 5179-90. 15456854

Narita, R., et al. (2004). Syndecan-dependent binding of Drosophila hemocytes to laminin alpha3/5 chain LG4-5 modules: potential role in sessile hemocyte islets formation. FEBS Lett. 576(1-2): 127-32. 15474023

Okamoto, O., et al. (2003). Normal human keratinocytes bind to the alpha3LG4/5 domain of unprocessed laminin-5 through the receptor syndecan-1. J. Biol. Chem. 278(45): 44168-77. 12947106

Rawson, J. M., et al. (2005). The heparan sulfate proteoglycans Dally-like and Syndecan have distinct functions in axon guidance and visual-system assembly in Drosophila. Curr. Biol. 15(9): 833-8. 15886101

Reiland, J., et al. (2004). Heparanase degrades syndecan-1 and perlecan heparan sulfate: functional implications for tumor cell invasion. J. Biol. Chem. 279(9): 8047-55. 14630925

Rhiner, C., Gysi, S., Frohli, E., Hengartner, M. O. and Hajnal. A. (2005). Syndecan regulates cell migration and axon guidance in C. elegans. Development 132(20): 4621-33. 16176946

Saoncella, S., Calautti, E., Neveu, W. and Goetinck, P. F. (2004). Syndecan-4 regulates ATF-2 transcriptional activity in a Rac1-dependent manner. J. Biol. Chem. 279(45): 47172-6. 15371457

Schulz, J. G., et al. (2004). Syndecan 3 intramembrane proteolysis is presenilin/gamma-secretase-dependent and modulates cytosolic signaling. J. Biol. Chem. 278(49): 48651-7. 14504279

Shimo, T., et al. (2004). Indian hedgehog and syndecans-3 coregulate chondrocyte proliferation and function during chick limb skeletogenesis. Dev. Dyn. 229(3): 607-17. 14991716

Spring, J., Paine-Saunders, S. E., Hynes, R. O. and Bernfield, M. (1994). Drosophila syndecan: conservation of a cell-surface heparan sulfate proteoglycan. Proc. Natl. Acad. Sci. 91(8): 3334-8. 8159748

Steigemann, P., Molitor, A., Fellert, S., Jackle, H. and Vorbruggen, G. (2004). Heparan sulfate proteoglycan syndecan promotes axonal and myotube guidance by slit/robo signaling. Curr. Biol. 14(3): 225-30. 14761655

Steinfeld, R., Van Den Berghe, H. and David, G. (1996). Stimulation of fibroblast growth factor receptor-1 occupancy and signaling by cell surface-associated syndecans and glypican. J. Cell Biol. 133(2): 405-416.

Suzuki, N., et al. (2003). Syndecan binding sites in the laminin alpha1 chain G domain. Biochemistry 42(43): 12625-33. 14580209

Tkachenko, E. and Simons, M. (2002) Clustering induces redistribution of syndecan-4 core protein into raft membrane domains. J. Biol. Chem. 277(22): 19946-51. 11889131

Tkachenko, E., Lutgens, E., Stan, R. V. and Simons, M. (2004). Fibroblast growth factor 2 endocytosis in endothelial cells proceed via syndecan-4-dependent activation of Rac1 and a Cdc42-dependent macropinocytic pathway. J. Cell Sci. 117(Pt 15): 3189-99. 15226395

Yamashita, H., Goto, A., Kadowaki, T. and Kitagawa, Y. (2004), Mammalian and Drosophila cells adhere to the laminin alpha4 LG4 domain through syndecans, but not glypicans. Biochem J. 382(Pt 3): 933-43. 15182231

date revised: 20 September 2006

Home page: The Interactive Fly © 2003 Thomas B. Brody, Ph.D.

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