Fig 2: SorCS1b lacking VPS10 and SorCS1b are targeted to the axon surface to similar levels.(A) Representative images of cultured hippocampal neurons (DIV21) transfected with the SorCS1bΔVPS10-IRES-GFP expression vector followed by immunostaining of surface SorCS1 and MAP2 before and after cell permeabilization, respectively. The GFP and MAP2 signals were used to distinguish axons (GFP-positive but MAP2-negative neurites, arrowheads) from dendrites (GFP- and MAP2-positive neurites). Immunoreactivity for surface SorCS1 was detected in both axons and dendrites of neurons transfected with SorCS1b ΔVPS10-IRES-GFP. (B) Quantification of the average intensity of surface SorCS1 on axons of neurons transfected with SorCS1bΔVPS10-IRES-GFP and SorCS1b-IRES-GFP. Deletion of the VPS10 domain does not affect the surface expression level of SorCS1b. n = 30 cells for SorCS1bΔVPS10-IRES-GFP and 32 cells for SorCS1b-IRES-GFP from three independent experiments, t test, N.S., not significant. Data are presented as mean ± SEM. Scale bar: 30 µm.

Fig 3: Artificial synapse formation assays using NLGN1 ectodomain-coated beads in cultures of hippocampal neurons with SorCS1b overexpression and AβO treatment.Representative low-magnification images showing VGLUT1 accumulation induced by NLGN1-Fc–coated beads at contacting axons. NLGN1-coated breads were exposed to cultured hippocampal neurons transfected with IRES-GFP, SorCS1b-IRES-GFP, or SorCS1bΔVPS10-IRES-GFP with AβO treatment (500 nM monomer equivalent, 24 h) or with vehicle treatment. The neurons were immunostained for VGLUT1, an excitatory presynaptic maker, and MAP2, a dendrite marker. For quantification, the beads contacting axons (GFP-positive and MAP2-negative neurites) (arrowheads) were selected. At contacting axons of IRES-GFP–transfected neurons, NLGN1-Fc–coated beads significantly induced VGLUT1 accumulation, and AβO treatment seemed to suppress the VGLUT1 accumulation. In contrast, it seems that AβO treatment failed to suppress NLGN1-induced VGLUT1 accumulation at contacting axons transfected with SorCS1-IRES-GFP but not those transfected with SorCS1bΔVPS10-IRES-GFP. These qualitative results are consistent with the quantitative results shown in Fig 4B. Scale bar: 15 µm.

Fig 4: The SorCS1 ectodomain binds to NRX1β and 2β, depending on their N-terminal histidine-rich domain.(A) Representative images showing the results of cell surface–binding assay testing for interaction between SorCS1-Fc and known synaptic organizers. SorCS1-Fc (1 µM) was added to COS-7 cells expressing the indicated construct. Note that SorCS1-Fc binds to COS-7 cells expressing HA-NRX1βS4(−), but not to those expressing any of the other organizers. For the N-terminal extracellular HA-tagged constructs, surface HA was immunostained to verify the expression of the construct on the COS-7 cell surface. Scale bars: 30 µm. (B) Representative images showing the binding of SorCS1-Fc (1 µM) to COS-7 cells expressing the indicated isoform of extracellularly HA-tagged NRX constructs. S4(+) and S4(−) indicate with and without an insert at splicing site 4, respectively, and ΔHRD indicates lack of the N-terminal histidine-rich domain (HRD) of β-NRX. HA fluorescent signals correspond to surface HA. Scale bar: 30 µm. (C) Quantification of bound SorCS1-Fc for each NRX construct. n = 30 cells for each construct from three independent experiments, one-way ANOVA, P < 0.0001. ***P < 0.001 compared with HA-CD4 and §P < 0.001 between NRXβ and NRXβΔHRD by Tukey’s multiple comparisons test. Data are presented as mean ± SEM. (D) Pull-down assays of purified recombinant His-tagged SorCS1 ectodomain protein with Fc, NRX1βS4(−)-Fc, or NRX1βS4(−)ΔHRD-Fc proteins indicate that the SorCS1 ectodomain and the NRX1β ectodomain form a complex when the NRX1β HRD is present.

Fig 5: The raw full membrane images of Fig 1D Western blots.Immunoblotting with anti-His tag antibody (left) and with anti-Fc antibody (right) on the same blot. The boxed regions were cropped to prepare the representative images shown in Fig 1D. Note that in lane 6, the anti-His antibody not only detected the SorCS1-His band at around 150 kD, but also an extra-band at around 60 kD, which corresponds to the size of NRX1β-Fc (lane 6 in the right hand panel). On the other hand, the anti-His antibody did not detect any band between 50–75 kD in lane 7, which was loaded with NRX1βΔHRD-Fc. These results suggest that the anti-His antibody can recognize the HRD of NRX1β and confirm that the Fc protein loaded in the lane 7 is lacking the HRD. Unexpectedly, the migration of NRX1βΔHRD-Fc on SDS–PAGE was slower than that of NRX1β-Fc, presumably because of the phenomenon called as “gel shifting.”