The collagen ColQ binds to LRP4 and regulates the activation of the Muscle-Specific Kinase–LRP4 receptor complex by agrin at the neuromuscular junction

Figures

• 

  Figure 1. ColQ and LRP4 coimmunoprecipitate in lysates from transfected heterologous cells and myotubes.A, HEK 293T cells were untransfected, cotransfected with either ColQ-Flag and LRP4-Myc or ColQ-Flag and LRP6-Myc, or transfected with each of these constructs alone. Immunoprecipitations were performed with anti-Flag antibodies for ColQ-Flag (ip) or with control nonimmune IgGs of the same isotype (c), and coprecipitates of LRP4-Myc or LRP6-Myc were analyzed by Western immunoblot using anti-Myc antibodies. After stripping, blots were also probed for Flag to ensure that equal amounts of ColQ-Flag were precipitated. Conversely, LRP4 and LRP6 were precipitated with anti-Myc antibodies (ip), and coprecipitated ColQ was detected with anti-Flag antibodies. ColQ clearly coprecipitated with LRP4 but not or very little with LRP6. Cell lysates (input) show similar expression of ColQ or of LRP4 and LRP6 in the different conditions. One Western immunoblot image representative of four independent experiments is shown. B, ColQ and LRP4 interact in myotubes. ColQ-Myc was expressed (+) or not (−) in ColQ−/− myotubes by viral infection. Immunoprecipitation was performed with polyclonal anti-LRP4…

• 

  Figure 2. ColQ binds directly to the extracellular domain of LRP4.A, pull-down assays. Magnetic beads were conjugated with ColQ-Flag (+) or not (−) and were subsequently incubated with conditioned media (CM) of HEK 293T cells containing equal amounts of Myc-tagged ectodomain of LRP4 (ectoLRP4-Myc) or LRP6 (ectoLRP6-Myc) or with control medium (HEK 293T cells transfected with an empty vector). Precipitated proteins were analyzed by Western immunoblotting with anti-Myc antibodies. ColQ-Flag interacted with ectoLRP4-Myc but not with ectoLRP6-Myc. Inputs show that comparable amounts of ectoLRP4-Myc and ectoLRP6-Myc were incubated with the ColQ-coated beads; n = 3. B, same experiment as in (A) except that higher concentrations of ectoLRP6-Myc than ectoLRP4-Myc were tested. Even in these conditions, ColQ bound only to ectoLRP4 and not to ectoLRP6. C, uncoated (−) or ColQ-coated (+) beads were incubated with CM expressing ectoLRP4-AP or with ectoLRP4-AP purified from the CM. Purified ectoLRP4-AP bound to ColQ-Flag, indicating a direct interaction between ectoLRP4 and ColQ. D, magnetic beads conjugated with equal amounts of ectoLRP4-Myc and ectoLRP6-Myc were incubated with the same amount of…

• 

  Figure 3. Surface plasmon resonance analysis of ColQ–LRP4 interaction.A, sensorgram of the binding of ColQ-Flag to a sensor chip covalently coated with anti-Flag antibodies to obtain a sensor surface covered with ColQ. Binding responses were measured using report points 30 s after the beginning of the dissociation phase and are indicated in RUs (resonance units). The level of immobilized ColQ reached about 3900 RU. Reference surface was obtained by injecting preparations from untransfected cells (CT) at the same protein concentration as ColQ-Flag preparations. B, 750 nM of purified ectoLRP4 (blue curve) or purified ectoMuSK (red curve) were passed over the ColQ-coated and reference surfaces. Sensorgrams represent the differential specific binding to ColQ after subtraction of the signals obtained on the reference surface. In contrast to ectoLRP4 (binding response of 23 RU), there was no binding of ectoMuSK to ColQ. C, sensorgram of the single-cycle kinetics. Five increasing (1.5-fold) concentrations (296, 444, 666, 1000, and 1500 nM) of purified ectoLRP4 were sequentially loaded on the ColQ-coated or reference surfaces without any regeneration step. Reference surface and drift signals were…

• 

  Figure 4. Comparison of ColQ binding to LRP4 and to MuSK.A, pull-down assays. Magnetic beads conjugated with ColQ-Flag (+) or not (−) were incubated with the same amounts of ectoLRP4-Myc or ectoMuSK-Myc as shown in the input. EctoLRP4-Myc but not ectoMuSK-Myc precipitated with ColQ-Flag; n = 3. B, same experiment as in (A) except that ColQ-coated beads were incubated with higher ectoMuSK-Myc than ectoLRP4-Myc concentrations. C, magnetic beads conjugated with equal amounts of ectoLRP4-Myc and ectoMuSK-Myc were incubated with the same amount of enzymatic activity of purified AChE–ColQ or AChE. Bound AChE–ColQ or AChE was quantified by measuring AChE activity. Results are expressed as the mean ± SEM percentage of the control value set as 100% (AChE–ColQ bound to CT ColQ-free beads). n = 4; ∗∗p < 0.01; ∗∗∗p < 0.001, using two-way ANOVA followed by Tukey's multiple comparison post hoc test. For interaction factor: F = 7.41, p = 0.0045; for CT versus ectoLRP4 versus ectoMuSK: F = 8.63, p = 0.0024; for AChE–ColQ versus AChE: F = 91.22, p < 0.0001. No significant binding of AChE–ColQ to ectoMuSK was detected, whereas AChE–ColQ bound to ectoLRP4. D, coimmunoprecipitation experiments. HEK 293T cells…

• 

  Figure 5. LRP4 potentiates ColQ–MuSK interaction.A, influence of LRP4 on ColQ–MuSK coimmunoprecipitation. HEK 293T cells were cotransfected with ColQ-Flag and MuSK-HA in the presence (+) or the absence (−) of cDNAs encoding LRP4-Myc. ColQ was immunoprecipitated (ip) with anti-Flag antibodies, and coprecipitated MuSK-HA was revealed using anti-HA antibodies. Control experiments (c) without adding anti-Flag antibodies to the cell lysates were performed in parallel. B, quantification of data in (A). The different band intensities were measured using ImageJ software. MuSK-HA signals were normalized to those of precipitated ColQ. The levels of MuSK coprecipitated with ColQ are expressed as the mean ± SEM fold change relative to the condition without LRP4 (n = 6; ∗p < 0.05, using one-sample t test). In the presence of LRP4, the level of MuSK coprecipitated with ColQ was increased by a factor of about fourfold. C, influence of MuSK on ColQ–LRP4 coimmunoprecipitation. Cells were cotransfected with ColQ-Flag and LRP4-Myc in the presence (+) or not (−) of MuSK-HA. ColQ was immunoprecipitated as aforementioned, and coprecipitated LRP4-Myc was evaluated by Western immunoblot. D, quantification of data…

• 

  Figure 6. Role of the ColQ C-terminal domain in the ColQ–LRP4 interaction.A, 25 or 50 μg of a Flag-tagged peptide corresponding to the last 27 amino acids of the ColQ C-terminal domain (Flag-ColQ Cter [425–451]) or of a Flag-control peptide corresponding to the scrambled 425 to 451 sequence were used to coat magnetic beads, which were subsequently incubated with purified ectoLRP4-Myc. Pulled down ectoLRP4-Myc was analyzed by Western immunoblot with anti-Myc antibodies. EctoLRP4-Myc interacts with the Flag-ColQ Cter peptide but not with the corresponding scrambled peptide. The image is representative of three independent experiments. B, magnetic beads coated with similar amounts of ColQ-Flag and ColQΔCt-Flag or ColQ-free beads as a control were incubated with conditioned media of HEK 293T cells containing equal amounts of ectoLRP4-Myc, as shown in inputs, or with control medium. C, quantification of ectoLRP4 bound to ColQ or ColQΔCt from data in (B) reveals a lower binding of ectoLRP4 to ColQΔCt than to ColQ, albeit statistically not significant. Results were normalized to precipitated ColQ or ColQΔCt and are expressed as the mean ± SEM of ColQ condition set as 100%; n = 5. ColQ, collagen…

• 

  Figure 7. Domains of LRP4 interacting with ColQ.A, schematic representation of ectoLRP4-AP and its deletion mutants. B, plate-binding assays. Same concentrations of ectoLRP4-AP and of the indicated deletion mutants were added to ColQ-coated or CT wells. Bound proteins were quantified by measuring AP activity. Results are expressed as the mean ± SEM percentage of the control value set as 100% (ectoLRP4-AP bound to CT wells). n ≥ 6; ∗∗p < 0.01; ∗∗∗∗p < 0.0001, using two-way ANOVA followed by Tukey's multiple comparison post hoc test. For interaction factor: F = 18.07, p < 0.0001; for CT versus ColQ: F = 162, p < 0.0001; for the comparison of the different ectoLRP4 mutants: F = 21.01, p < 0.0001. The N-terminal region of LRP4 plays a crucial role in ColQ–LRP4 interaction as its deletion (ectoLRP4Δ1-AP) compromised binding to ColQ. Conversely, the N-terminal region alone (ectoLRP4Δ234-AP) bound to ColQ at the same level as ectoLRP4-AP. C, ColQ-coated wells were incubated with 25 nM of ectoLRP4-AP in the presence or not of 500 nM purified recombinant neural agrin. Results are the mean ± SEM percentage of ectoLRP4-AP bound to ColQ wells in the absence of agrin (set as 100%; n = 9, ∗∗∗p < 0.001,…

• 

  Figure 8. ColQ regulates agrin-induced MuSK phosphorylation in myotubes.A, MuSK phosphorylation was analyzed in WT and ColQ−/− myotubes treated (+) or not (−) with 10 nM of neural agrin for 1 h before cell lysis. MuSK was immunoprecipitated, and the levels of MuSK tyrosine phosphorylation were assessed after Western blot by probing immunoprecipitated MuSK with anti-phosphotyrosine antibody 4G10. Immunoblots were then stripped and reprobed with anti-MuSK antibody to visualize precipitated MuSK. Agrin induced a strong tyrosine phosphorylation of MuSK in WT and ColQ−/− myotubes, whereas no phosphorylation was observed in untreated myotubes. In parallel, cell surface (cs) proteins of myotubes treated in the same way were isolated and probed with antibodies against MuSK or LRP4 to determine their cs levels. Blots were also probed with antibodies against the membrane transferrin receptor (TfR) used as a loading control to normalize the results. The cs levels of MuSK relative to TfR were decreased in the absence of ColQ, whereas those of LRP4 were increased. Total levels of MuSK and LRP4 were unchanged in WT and ColQ−/− myotubes (input). B, quantification of MuSK phosphorylation signals…

• 

  Figure 9. ColQ regulates agrin-induced MuSK phosphorylation in transfected heterologous cells.A, HEK 293T cells were cotransfected with MuSK-HA and LRP4-Myc and in addition with ColQ-Flag or an empty plasmid. After 48 h, the cells were treated or not with 10 nM of neural agrin for 1 h. MuSK was immunoprecipitated, and the levels of MuSK phosphorylation were assessed as aforementioned for myotubes. The blots were stripped and reprobed with anti-HA antibodies to visualize precipitated MuSK. Cell surface (cs) levels of MuSK and LRP4 were analyzed as aforementioned, using anti-HA and anti-Myc antibodies, respectively, for immunodetection. B, quantification of the p-MuSK/MuSK ratios for data in (A). Results are expressed as the mean ± SEM percentage of agrin-treated HEK cells expressing MuSK and LRP4 but not ColQ (n = 3; ∗p < 0.05, using one-sample t test). MuSK tyrosine phosphorylation in total MuSK precipitates was decreased by about 50% in the presence of ColQ. C, quantification of data in (A) where p-MuSK/MuSK ratios were normalized to cs MuSK levels before agrin treatment. When taking into account the cs levels of MuSK, the decrease of agrin-induced MuSK phosphorylation level reached about…

• 

  Figure 10. ColQ regulates agrin-induced AChR clustering.A, WT and ColQ−/− myotubes were treated or not with neural agrin (5 nM) for 16 h. AChR clusters were visualized by α-bungarotoxin staining. Scale bar represents 20 μm. B, the number of AChR clusters per myotube was quantified for each condition. At least, 60 myotubes from three independent experiments were analyzed for each condition. For quantification, a size threshold was applied such that only AChR clusters >5 μm2 were scored. The number of AChR clusters after agrin treatment was decreased by less than 15% in ColQ−/− compared with WT myotubes. Results are expressed as the mean ± SEM. ∗p < 0.05; ∗∗∗∗p < 0.0001, using two-way ANOVA followed by Tukey's multiple comparison post hoc test. For interaction factor: F = 13.35, p = 0.0003; for −Agrin versus +Agrin: F = 455.2, p < 0.0001; for WT versus ColQ−/−: F = 0.32, p = 0.57 with a significant difference in the presence of agrin (p = 0.012) and no significant difference in the absence of agrin treatment. AChR, acetylcholine receptor; ColQ, collagen Q.