Supplemental Material

This article contains the following supporting material:

• Figure S1 - The transmembrane domain of paranodin does not induce ER retention N2a cells were transiently transfected with GFP-TM-NrCAM (A) or with GFP-TM-pnd (B). After fixation with 4 % paraformaldehyde and permeabilization with 0.1 % TX-100 in PBS, cells were immunostained with anti-GFP mAb. As control, GFP-TM-NrCAM, which contains the signal peptide of NrCAM upstream and the transmembrane domain of NrCAM downstream of GFP, is targeted to the cell membrane (A). Immunofluorescence staining for GFP-TM-pnd, which encompassed the signal peptide of NrCAM upstream of GFP fused with the transmembrane and intracellular domains of paranodin, shows that the construct is highly expressed at the cell membrane with no ER retention (B). Bar: 10 μm.
• Figure S2 - Contactin-GFP allows the cell surface expression of paranodin (A) COS-7 cells transfected with contactin (lane 1) or contactin-GFP (lanes 2 and 3) were lysed with NP-40 and analyzed by Western blotting using anti-contactin antibody (lanes 1 and 2) and anti-GFP mAb (lane 3). Contactin-GFP is detected with an apparent molecular weight of 170 kDa using both antibodies. COS-7 cells co-transfected with contactin-GFP and paranodin were lysed with NP40 (lane 4). Immunoprecipitation was performed using anti-paranodin anti-serum and the immunoblot revealed using anti-GFP mAb (lane 4). Contactin-GFP is co-immunoprecipitated with paranodin. (B) N2a cells were transfected with contactin-GFP and live cells were immunostained with anti-GFP. The strong immunostaining indicates that GFP is extracellular. (C) N2a cells co-transfected with paranodin and contactin-GFP were fixed and permeabilized with TX-100. Double-staining for paranodin (red) and GFP (green) indicate that the two proteins are co-localized at the cell membrane. Bar: 30 μm.
• Figure S3 - Structural prediction and molecular dynamics study of the PGY region of paranodin The structure of EGF-2 and LNG-4 was predicted by usual homology modeling, the structure of PGY using the Rosetta de novo structure prediction method, and the structure of the linker region was obtained from the Phyre web server of the Structural Bioinformatics Group at Imperial College London. The connection of the EGF-2 (magenta) + linker (red helix barrel) + PGY (green) + LNG-4 (orange) domains was achieved by a knowledge-based driven manual docking of the individual partners using the virtual reality equipment at LORIA (http://www.loria.fr/projets/docking/pages/). The stability of the resulting model was tested using a 4-ns MD simulation. Snapshots structures were recorded every 500 ps and superimposed (inset) to indicate predicted stability and flexibility.
• Table S1 - Quantitative analyses of the cell membrane expression of caspr2, contactin-GFP, paranodin, pnd-caspr2C and pndΔPGY2. Cells were fixed at various time after transfection (4, 6, 8, 10, 12, 14, 16, 24, 28 and 48 h) with methanol and processed for immunofluorescence staining of paranodin or caspr2. The cell surface staining of contactin-GFP was realized on live cells using the anti-contactin 24 antiserum and the internal pool of contactin-GFP was visualized by imaging the GFP fluorescence. The percentage of transfected cells with cell membrane expression of the recombinant molecules was estimated (more than 100 cells were examined under each condition).
• Supplementary Data