Fig 1.
lon-2/glypican functions in the attractive unc-6/netrin guidance pathway.
(A) During the first larval stage of C. elegans, the pioneer neuron AVM extends ventrally along the body wall until it reaches the ventral nerve cord. Its migration results from the combined attractive response to UNC-6/netrin (secreted at the ventral midline) via the UNC-40/DCC receptor and the repulsive response to SLT-1/Slit (secreted by the dorsal muscles) via its SAX-3/Robo receptor. We visualized the morphology of the AVM axon using the transgene Pmec-4::gfp. (B) The heparan sulfate proteoglycans lon-2/glypican and sdn-1/syndecan cooperate to guide the axon of AVM, as their simultaneous loss enhances guidance defects. The role of lon-2/glypican in axon guidance is specific, as the loss of lon-2/glypican, but not the loss of the other C. elegans glypican, gpn-1, enhances the defects of sdn-1/syndecan mutants. (C) Complete loss of lon-2/glypican enhances the axon guidance defects resulting from disrupted slt-1/Slit signaling in mutants for slt-1/Slit or its receptor sax-3/Robo, as well as in animals misexpressing slt-1 in all body wall muscles (using a Pmyo-3::slt-1 transgene). Data for wild type and lon-2 are the same as in (B). (D) Complete loss of lon-2/glypican does not enhance the AVM guidance defects of unc-6/netrin mutants or of mutants for its receptor unc-40/DCC, suggesting that lon-2/glypican functions in the same genetic pathway as unc-6/netrin. Data for wild type and lon-2 are the same as in (B). (E) Loss of sdn-1/syndecan function does not enhance the defects of slt-1/Slit or sax-3/Robo mutants but enhances the defects of unc-40/DCC mutants. Data for wild type, sdn-1, slt-1, sax-3, and unc-40 are the same as in (B–D). Error bars are standard error of the proportion. Asterisks denote significant difference\: *** p ≤ 0.001,** p ≤ 0.01, and * p ≤ 0.05 (z-tests, p-values were corrected by multiplying by the number of comparisons). ns, not significant.
Fig 2.
lon-2/glypican functions in the repulsive unc-6/netrin guidance pathway.
(A) Schematics of the migration path of the DTCs in the wild type and examples of defective DTC migration in lon-2/glypican mutants (the anterior and the posterior DTCs exhibit similar defects). In wild-type animals, the DTCs migrate away from the vulva along the anteroposterior axis [1], then turn dorsally [2], and turn again to migrate towards the midbody region [3]. Loss of lon-2/glypican leads to defective DTC guidance, including a failure to migrate dorsally, premature dorsal turning, or a failure to remain dorsal. (B) Quantification of the DTC migration defects in lon-2/glypican mutants and rescue by lon-2(+) (see S7 Table). For each transgenic line, transgenic animals were compared to nontransgenic sibling controls. Complete loss of lon-2/glypican does not enhance the defects of the unc-6/netrin null mutants or those of the null mutants for unc-5/UNC5 and unc-40/DCC, suggesting that lon-2/glypican functions in the same guidance pathway as unc-5/UNC5, unc-40/DCC, and unc-6/netrin (see S6 Table). (C) The axons of the GABAergic motorneurons project dorsally from the ventral midline towards the dorsal nerve cord. unc-6/netrin, unc-5/UNC5, and unc-40/DCC are required for this dorsal guidance of GABAergic axons. Complete loss of lon-2/glypican does not enhance the partially penetrant defects of unc-40/DCC null mutants, suggesting that lon-2/glypican functions in the same pathway as unc-40/DCC and unc-6/netrin to guide axons dorsally (see S5 Table). Error bars are standard error of the proportion. Asterisks denote significant difference: *** p ≤ 0.001 and * p ≤ 0.05 (z-tests, p-values were corrected by multiplying by the number of comparisons). ns, not significant.
Fig 3.
unc-6/netrin signaling via the unc-5/UNC5 receptor requires lon-2/glypican.
(A) The axon of PVM normally migrates ventrally in the wild type, but it can be forced to migrate dorsally by misexpressing the repulsive receptor unc-5/UNC5. We quantified PVM since AVM could not be reliably identified (both AVM and neighboring ALMR axons project dorsally in Pmec-7::unc-5 transgenic animals.) (B) Upon misexpression of unc-5/UNC5 in PVM, using the transgene Pmec-7::unc-5, the axon of PVM projects dorsally in an unc-6/netrin-, unc-40/DCC-, and unc-34/enabled-dependent manner. Loss of lon-2/glypican partially suppresses this forced dorsal migration, indicating that unc-6/netrin signaling depends on lon-2/glypican. Scale bar, 5 μm. Error bars are standard error of the proportion. Asterisks denote significant difference: *** p ≤ 0.001 (z-tests, p-values were corrected by multiplying by the number of comparisons). ns, not significant. Wild type (without evIs25) is the same as in Fig 1B.
Fig 4.
lon-2/glypican functions in the epidermal cells underlying the developing axon.
(A) Epidermal expression of lon-2/glypican is sufficient for function. Providing wild-type lon-2(+) in the hypodermis (under the heterologous hypodermal promoters Pdpy-7 and Pelt-3) rescues the function of lon-2 in the double mutants lon-2 slt-1, as it brings the defects down to the level of slt-1 single mutants. In contrast, expression of lon-2(+) in other epidermal cells (seam cells), the migrating neuron AVM, the intestine, or the body wall muscles fails to rescue the function of lon-2. For each rescued transgenic line, transgenic animals were compared to nontransgenic sibling controls (see S2 and S3 Tables). Data for wild type, lon-2, slt-1, and lon-2 slt-1 are the same as in Fig 1B and 1C. (B) Expression of sdn-1/syndecan in the migrating neuron is sufficient for function. Providing wild-type copies of sdn-1(+) in AVM (expressed under the heterologous promoter Pmec-7) rescues the axon guidance function of sdn-1 in a lon-2 sdn-1 double mutant. We assayed rescue of sdn-1 function using the double mutant lon-2 sdn-1 since it is easier to rescue defects that are 33% penetrant (as in the double lon-2(e678) sdn-1(zh20)) than to rescue defects that are 12% penetrant (as in the single mutant sdn-1(zh20)). For each transgenic line, transgenic animals were compared to nontransgenic sibling controls (see S2 Table). Data for wild type, lon-2, sdn-1, and lon-2 sdn-1 are the same as in Fig 1B–1D. Scale bar, 5 μm. Error bars are standard error of the proportion. Asterisks denote significant difference: *** p ≤ 0.001, ** p ≤ 0.01 and * p ≤ 0.05 (z-tests, p-values were corrected by multiplying by the number of comparisons). ns, not significant.
Fig 5.
A secreted form of LON-2/glypican that lacks the heparan sulfate chain attachments is functional in axon guidance.
(A) The HSPG LON-2/glypican is composed of a core protein and three HS chains. The core protein is predicted to fold into a globular domain on its N-terminal region and to be GPI-anchored. (B) Schematics of the engineered forms of LON-2 that we used: LON-2ΔGAG, in which the HS chain attachment sites are mutated; LON-2ΔGPI, in which the GPI anchor is deleted; N-LON-2, in which the C-terminus is deleted; and C-LON-2, in which the N-terminal globular domain is deleted. Western blot analysis of protein extracts of worms expressing LON-2::GFP or LON-2ΔGAG::GFP confirms that deleting the HS attachment sites on LON-2 affects HS addition on LON-2. Protein extracts from wild type (N2) and an unrelated GFP strain (lqIs4) are negative controls. (C) A form of LON-2/glypican lacking HS chain attachment sites (LON-2ΔGAG) functions in axon guidance. LON-2ΔGAG rescues the AVM guidance defects of double mutants lon-2 slt-1 back to the level of slt-1 single mutants. Secreted globular LON-2/glypican is functional in axon guidance. LON-2/glypican was engineered to be secreted by deleting its GPI anchor (LON-2ΔGPI) or by deleting the C-terminus, thus lacking the GPI anchor and the HS attachment sites (N-LON-2). Both LON-2ΔGPI and N-LON-2 function in axon guidance, as assayed by their ability to rescue axon guidance defects of lon-2 slt-1 back down to the level of slt-1 single mutants. In contrast, a form of LON-2/glypican containing its C-terminus including the three HS attachment sites, but lacking its N-terminal globular domain (C-LON-2), is not functional (see S2 and S3 Tables), indicating that the N-terminal globular domain of the core protein is key to the function of LON-2/glypican in axon guidance. For each rescued transgenic line, transgenic animals were compared to nontransgenic sibling controls (see S2 and S3 Tables). Data for wild type, lon-2, slt-1, and lon-2 slt-1 are the same as in Fig 1B and 1C. (D) A form of LON-2/glypican lacking HS chain attachment sites is functional in DTC guidance. The DTC migration of lon-2 mutants carrying the transgene Plon-2::LON-2ΔGAG is rescued back to wild-type levels. Secreted N-terminus globular LON-2/glypican (N-LON-2) is functional in DTC guidance, as DTC guidance defects of lon-2 mutants are rescued by N-LON-2. Transgenic animals were compared to nontransgenic sibling controls (see S7 Table). Data for wild type and lon-2 are the same as in Fig 2B. Error bars are standard error of the proportion. Asterisks denote significant difference: *** p ≤ 0.001, ** p ≤ 0.01 (z-tests, p-values were corrected by multiplying by the number of comparisons).
Fig 6.
LON-2/glypican associates with UNC-40/DCC-expressing cells.
(A) Experimental design. Each construct was individually and transiently transfected in S2 cells. After 2 d, cells from independent single transfections were mixed and incubated overnight and then immunostained for the corresponding tags. HA::LON-2-conditioned medium was mixed with UNC-40::FLAG-expressing cells. (B) HA::LON-2 is released from cells that produce it and associates with UNC-40-expressing cells. HA::LON-2 fills the cytoplasm of the cells that produce it (indicated by an asterisk, see also S8 Fig). Notably, HA::LON-2 is observed decorating the outline of UNC-40::FLAG-expressing cells (experiments 1, 6, 7, and 8). HA::LON-2ΔGAG also associates with UNC-40::FLAG-expressing cells (experiment 2). Cells expressing UNC-40ΔNt::FLAG that lacks the extracellular domain do not have HA::LON-2 signal, indicating that the association of LON-2 with UNC-40-expressing cells requires the extracellular domain of UNC-40 (experiment 3). HA::LON-2-conditioned medium contains HA::LON-2 that associates with UNC-40::FLAG-expressing cells, indicating that HA::LON-2 is released from the cells that produce into a diffusible form that interacts with UNC-40::FLAG cells (experiment 8). HA::LON-2 does not associate with cells expressing SfGFP::UNC-6 (experiments 4, 6, and 7) or with untransfected cells. UNC-40-FLAG-expressing cells can simultaneously associate with HA::LON-2 and SfGFP::UNC-6 (experiment 6). HA::LON-2 associates with cells expressing a mutant form of UNC-40/DCC that is unable to bind SfGFP::UNC-6, as it lacks the Fn4/5 UNC-6 binding domains (UNC-40ΔFn4/5::FLAG; experiment 7). Scale bars, 10 μm. (C) Quantification of the association of HA::LON-2 (from expressing cells, from medium of expressing cells, or from cells expressing HA::LON-2ΔGAG) with cells expressing UNC-40::FLAG, UNC-40ΔNt::FLAG, SfGFP::UNC-6, or UNC-40ΔFn4/5::FLAG and untransfected cells. Ten different optical fields containing ~300 cells from three independent experiments were quantified and averaged. (D) Cells expressing UNC-40::FLAG can display irregular morphology, which is enhanced by the presence of HA::LON-2. Images of the different morphologies displayed by UNC-40::FLAG-expressing cells: with a smooth edge, with an irregular edge, or with membrane extensions. The morphology of S2 cells expressing mCherry alone or coexpressing UNC-40::FLAG and mCherry, which were mixed with control untransfected cells or with HA::LON-2-expressing cells, were quantified for irregular edges (grey bars) or membrane extensions (black bars). A higher percentage of UNC-40::FLAG-expressing cells show membrane extensions or irregular edges when mixed with HA::LON-2-expressing cells, as compared to when they are mixed with control mCherry cells. Error bars are standard error of the mean. Asterisks denote significant difference: *** p ≤ 0.001, * p ≤ 0.05. ns, not significant. In (D), significant differences in irregular cell shape are indicated by grey asterisks, and significant difference in membrane extensions is indicated by the black asterisk.
Fig 7.
A model for the role of LON-2/glypican in UNC-6/netrin-UNC-40/DCC-mediated axon guidance.
HSPG LON-2/glypican (red) is expressed from the hyp7 epidermal cells (pink) underlying the migrating growth cone of the AVM neuron (tan). LON-2/glypican is released from the hypodermal cell surface and may associate with the developing axon expressing the receptor UNC-40/DCC (blue), directly or indirectly interacting with UNC-40/DCC, to modulate UNC-6/netrin (green) signaling. A second HSPG, SDN-1/syndecan (black), acts in the SLT-1/Slit-SAX-3/Robo (grey) axon guidance pathway.