Fig 2: Nidogens bind to the receptor-type protein tyrosine phosphatases (RPTPs) LAR and PTPRδ.(A) Schematic showing the domain organisation of the LAR family of RPTPs. Full-length LAR, PTPRδ and PTPRσ contain three extracellular immunoglobulin-like (Ig) and eight fibronectin III (FNIII) domains, as well as two intracellular protein tyrosine phosphatase (D1 and D2) domains. RPTPs undergo proteolytic processing between the FNIII and phosphatase domains; this generates an extracellular subunit that remains non-covalently bound to the intracellular phosphatase subunit. (B) Schematic diagram of human LAR, PTPRδ and PTPRσ chimeric proteins used for co-immunoprecipitations. The extracellular domain of each of these RPTPs was fused to the murine Igκ-chain leader sequence and haemagglutinin (HA) tag at the N-terminus; the C-terminus was fused to the platelet-derived growth factor receptor transmembrane (PDGFR TM) domain and a Myc tag. (C–E) Western blots showing the direct interaction of nidogen-1 and -2 with the extracellular domain of LAR in the presence of VSVG-HCT. Nidogen-1 (C) and nidogen-2 (D) were immunoprecipitated from N2a cell lysates, and co-immunoprecipitates were probed using an anti-HA antibody. Conversely, HA-eLAR-Myc was immunoprecipitated using an anti-HA antibody, followed by the detection of nidogens (E). Non-specific antibodies bound to beads and empty beads were used as controls; 5% input was loaded. (F–H) Western blots showing the interaction of PTPRδ with nidogen-1 and -2, in the presence of VSVG-HCT. Nidogen-1 (F), nidogen-2 (G) and HA-ePTPRδ-Myc (H) were immunoprecipitated from lysates of N2a cells, and co-immunoprecipitates were probed using an appropriate antibody (anti-HA for nidogen immunoprecipitations, and anti-nidogen for HA-eLAR-Myc immunoprecipitations). Non-specific antibodies bound to beads and empty beads were used as negative controls; 5% input was loaded.

Fig 3: Individual LAR fibronectin III domains display a limited effect in inhibiting the binding of the nidogen-HcT complex to endogenous LAR.(A) Representative immunofluorescence images of motor neurons upon internalisation of HCT-555 and nidogen-2 in the presence of 20 μM recombinant LAR FNIII2, FNIII4, FNIII5 or FNIII7 domains. Images in the top two panels have been colour mapped based on their intensities. Scale bar: 20 μm. (B, C) Graphs showing quantification of endocytosed nidogen-2 (B) and HCT-555 (C) shown in panel (A). Control refers to cultures treated with buffer alone. Data are presented as a percentage of internalised nidogen-2 or HCT in buffer-treated motor neurons (n = 3 independent experiments; error bars indicate s.e.m.). Results were tested for statistical significance using Kruskal–Wallis test (P = 0.0001), followed by Dunn’s post-hoc test (B) and one-way ANOVA (P = 0.0005), followed by Dunnett’s multiple comparisons test (C).

Fig 4: Endogenous LAR co-localises with the nidogen-HCT complex in signalling endosomes.(A) Representative immunofluorescence images of mouse motor neurons treated with HCT-647 and labelled with antibodies against internalised nidogen-2 and total LAR. Images have been pseudo-coloured in magenta (HCT-647), yellow (nidogen-2) and cyan (LAR). A selected region in the upper panel has been magnified in the lower panel. Scale bars: 20 μm (top panel) and 5 μm (bottom panel). (B) Graph showing overlapping intensity profiles of HCT-647, nidogen-2 and LAR in an axonal segment (boxed region in the lower panel of A). Empty arrowheads point to co-localised HCT and LAR organelles, arrowheads denote co-localised nidogen-2 and LAR puncta, while arrows represent puncta containing HCT, nidogen-2 and LAR. (C) Quantification of the neuronal correlation between HCT-647 and nidogen-2 with LAR using fluorescence intensities (n = 46 neurites; Spearman coefficient 0.743 and 0.497, and P < 0.0001 and 0.0004, for LAR-HCT and LAR-nidogen-2, respectively).

Fig 5: Nidogen-2 and LAR/PTPRδ fragments exhibit dose-dependent interactions in vitro.(A) Schematic of recombinant LAR FNIII fragments used for bacterial expression and purification. LAR FNIII1-4 and FNIII5-7 fragments were tagged with a 6×His tag at the N-terminus and a FLAG tag at the C-terminus. (B, C) Plots showing in vitro binding between purified nidogen-2 and bacterially expressed LAR FNIII1-4-FLAG (B) and LAR FNIII5-7-FLAG (C). Serial dilutions of each purified LAR fragment were added to a fixed amount of immobilised nidogen-2 (0.5 picomoles), followed by addition of an anti-FLAG antibody to reveal complex formation using ELISA. All datapoints were normalised to the absorbance obtained using 5 μM of LAR FNIII fragment (n = 3 independent experiments; error bars indicate s.e.m.). (D) Schematic of PTPRδ fragments used to identify the interacting domains between PTPRδ and nidogens. Truncated proteins were fused to the murine Igκ-chain leader sequence and an HA tag at the N-terminus; the C-terminus was fused to the PDGFR transmembrane domain and a Myc tag. (E, F) Schematic of recombinant PTPRδ fragments used for protein expression and purification. PTPRδ Ig1-3 was tagged with a 6×His tag at the C-terminus (E), while the FNIII5-7 fragment was tagged with a 6×His tag at the N-terminus and a FLAG tag at the C-terminus (F). (G, H) Plots showing in vitro binding between purified nidogen-2 and recombinant PTPRδ Ig1-3-His (G) and PTPRδ FNIII5-7-FLAG (H). Serial dilutions of each PTPRδ fragment were added to a fixed amount of immobilised nidogen-2 (0.5 picomoles), followed by addition of an anti-His or anti-FLAG antibody, respectively, to reveal complex formation using ELISA. All datapoints were normalised to the absorbance obtained using 5 μM of PTPRδ FNIII fragment (n = 3 independent experiments; error bars indicate s.e.m.).