Nogo-A inhibits neurite outgrowth and cell spreading with three discrete regions

Abstract

Nogo-A is a potent neurite growth inhibitor in vitro and plays a role both in the restriction of axonal regeneration after injury and in structural plasticity in the CNS of higher vertebrates. The regions that mediate inhibition and the topology of the molecule in the plasma membrane have to be defined. Here we demonstrate the presence of three different active sites: (1) an N-terminal region involved in the inhibition of fibroblast spreading, (2) a stretch encoded by the Nogo-A-specific exon that restricts neurite outgrowth and cell spreading and induces growth cone collapse, and (3) a C-terminal region (Nogo-66) with growth cone collapsing function. We show that Nogo-A-specific active fragments bind to the cell surface of responsive cells and to rat brain cortical membranes, suggesting the existence of specific binding partners or receptors. Several antibodies against different epitopes on the Nogo-A-specific part of the protein as well as antisera against the 66 aa loop in the C-terminus stain the cell surface of living cultured oligodendrocytes. Nogo-A is also labeled by nonmembrane-permeable biotin derivatives applied to living oligodendrocyte cultures. Immunofluorescent staining of intracellular, endoplasmic reticulum-associated Nogo-A in cells after selective permeabilization of the plasma membrane reveals that the epitopes of Nogo-A, shown to be accessible at the cell surface, are exposed to the cytoplasm. This suggests that Nogo-A could have a second membrane topology. The two proposed topological variants may have different intracellular as well as extracellular functions.

Figure 1.

Nogo-A, -B, and –C isoforms, main studied fragments, and antibodies raised against Nogo-A sequence and their specificity on oligodendrocytes. A, Several monoclonal antibodies and polyclonal rabbit antisera (AS) were raised against Nogo-A. AS 472 (Chen et al., 2000) and mAb 11C7 were raised against the same 18 aa peptide. The other three mAbs, 11A8, 7B12, and 3D11, were raised against bacterially expressed NiR-G. AS Bianca was raised against bacterially produced NiR. AS 922 was raised against the Nogo-66 region between the two hydrophobic domains, and AS 294 was raised against the C terminus of Nogo. AS 922 and AS 294 recognize all Nogo isoforms. B, Western blot of lysates of cultured oligodendrocytes. The blot was incubated with different anti-Nogo-A antibodies. All antibodies recognize the 190 kDa Nogo-A band (arrow) and mAb 3D11, AS 922, and AS 294 stain additional bands at 60 and 80 kDa, presumably Nogo-A breakdown products. AS Bianca, AS 294, and AS 922 also recognize Nogo-B at ∼55 kDa (arrowhead).

Figure 2.

Analysis of the inhibitory properties of various Nogo fragments on 3T3 fibroblast spreading. The main inhibitory activity resides in two regions of the Nogo-A protein. Nogo-A deletion products were tested for their relative inhibitory activity (percentage inhibition) on fibroblast spreading (5 μg/100 μl, 1 μg/100 μl, 0.5 μg/100 μl, and 0.2 μg/100 μl of fragments coated per square centimeter). Mean of spreading on plastic is indicated by the dotted line. For each fragment at least three independent assays with proteins from at least two separate purifications were performed. The fragments that are most inhibitory for fibroblast spreading are highlighted. Control peptides (5 μg/100 μl coated per square centimeter) did not exhibit inhibitory properties except for l-laminin.

Figure 6.

Binding of amino terminal Nogo fragments to 3T3 cells and cortical membranes. A, Binding of NiR-G (aa 1–979) to 3T3 cells. 3T3 fibroblasts were incubated with a preformed complex consisting ofα-myc antibody (9E10), FITC-conjugated F(ab)₂ goatα-mouse IgG fragments, and myc-tagged amino-terminal fragment of Nogo-A (NiR-G). Binding of the complex to 3T3 fibroblasts was analyzed by flow cytometry. Unstained 3T3 fibroblasts were used as negative control. Further negative controls include 3T3 cells incubated with a complex consisting of the α-myc antibody and the FITC-conjugated α-mouse Ab or with the FITC-conjugated α-mouse Ab alone. B, Binding of NiR-G to 3T3 cells is protease sensitive: trypsinization of 3T3 fibroblasts before incubation with Nogo completely abolishes Nogo-binding to their surface. C, Western blot of purified, myc-epitope-tagged NiR-G. D, Silver-stained gel of IMAC-purified NiG-Δ20. M, Molecular weight standard. E, Binding of 1.3 nm [¹²⁵I]-NiG-Δ20 to rat brain cortical membranes and competition by increasing concentration of unlabeled NiG-Δ20. □ indicates the values obtained after incubation of 1.3 nm [¹²⁵I]-NiG-Δ20 in the absence of cortical membranes, i.e., nonspecific binding to the tubes. Values are represented as mean ± SE (**p < 0.01; ***p < 0.001; Student's t test for competition; n = 3).

Figure 3.

Analysis of the inhibitory activity of Nogo-A on neurite outgrowth. A, B, Nogo-A is an inhibitory substrate for E7–E9 RGC neurite outgrowth in a stripe assay. Stripes of Nogo/laminin (arrowhead) versus laminin-only stripes (arrow) were compared. A, The axons (stained with anti-neurofilament mAb) growing from chicken retina explants avoid the Nogo/laminin stripes (stained with AS Laura; arrowhead). This avoidance is concentration dependent and accompanied by strong axonal fasciculation. Control β-galactosidase/laminin stripes (stained with α-T7 mAb) are permissive substrates for RGC axons. B, Inhibition score for RGC outgrowth on different Nogo fragments. Cultures were evaluated by giving a score of 5 for striped neurite outgrowth with no fibers crossing the Nogo containing stripes, 4–2 for striped neurite outgrowth with increasing numbers of crossing fibers, 1 for random outgrowth with tendency to grow in the direction of the stripes, and 0 for complete random neurite outgrowth. The dotted line indicates the mean of all β-galactosidase control experiments. Values are represented as mean ± SE. The groups have been compared with the scores for Nogo-C at the same coated protein concentration (*p < 0.05; **p < 0.01; Mann–Whitney U test). C, Examples of PC12 neurite outgrowth on different substrates. On Nogo-A and its fragment NiG-Δ20, the number of neurites is reduced and they are shorter compared with cells grown on Nogo-C and NiR. D, Quantification of outgrowth of PC12 neurites grown on different Nogo substrates (scores from 0 = no outgrowth to 5 = long, branched neurites). Values are represented as mean ± SE. The groups have been compared with the scores for Nogo-C at the same coated protein concentration (*p < 0.05; **p < 0.01; Mann–Whitney U test). E, Primary rat cerebellar granule cells were plated on increasing amounts of coated NiG. The inhibition of neurite outgrowth and cell adhesion by NiG is dose dependent.

Figure 4.

Analysis of the collapsing activity of different Nogo-A fragments on DRG growth cones. A, Time course of the growth cone collapse of dissociated rat P6 DRG neurons induced by preclustered NiG-Δ20-AP, Nogo-66-AP, or Sema3A. After the indicated time periods, the cultures were fixed and doubly labeled by anti-tubulin antibody and tetramethylrhodamine B isothiocyanate-phalloidin. B, Dose-dependent growth cone collapse of chicken E13–E15 DRG explants treated with preclustered NiG-Δ20-AP, Nogo-66-AP, and AP-Sema3A. All three proteins induce growth cone collapse, but Nogo-66-AP is more potent than NiG-Δ20-AP. Both Nogo-AP fragments have a much weaker collapse-inducing activity than AP-Sema3A.

Figure 5.

Activity of NgR-expressing CHO cells versus wild-type cells on Nogo fragments. A, An mAb against the V5-epitope stains the surface of a CHO cell line stably transfected with rat NgR. B, On Western blot, no NgR protein is detectable in lysates of wild-type CHO cells, whereas high amounts of NgR are detectable in lysates of the CHO–NgR cell line. C, Both wild-type and NgR-transfected CHO cells are strongly inhibited in spreading by the Nogo-A-specific region NiG, whereas they are unresponsive to Nogo-66.

Figure 7.

Nogo-A is present at the cell surface of cultured oligodendrocytes. Live, unpermeabilized oligodendrocytes (3 d in culture) were incubated with anti-Nogo-A antibodies, fixed, and visualized with secondary antibodies conjugated with alkaline phosphatase. All antibodies against the N-terminus of Nogo-A stain the cell surface of oligodendrocytes (mAb 11C7, mAb 11A8, mAb 3D11, mAb 7B12, AS Bianca), although more weakly than an antibody against the surface glycoprotein MAG (shorter development; see Materials and Methods). The control mouse IgG (d) and an antibody against the intracellular oligodendrocyte protein CNPase (e) do not stain oligodendrocytes. Preincubation of 11C7 with the immunogen, peptide P472, reduced the staining to background levels (b), whereas preincubation with an unspecific peptide (Px) did not result in a reduction in staining intensity (c). Furthermore, AS 922 against the C-terminal Nogo-66-region stains the cell surface of living oligodendrocytes (l), whereas the corresponding preimmune serum (pre-922) gives only background staining (m). Staining of co-cultures of oligodendrocytes with CHO cells expressing rat Nogo-A with mAb 11C7 shows that although intact oligodendrocytes are stained, Nogo-A produced by CHO cells (arrow) is not detectable at the cell surface (o). Both cell types are stained with 11C7 after permeabilization (p). Scale bar, 25μm.

Figure 8.

Precipitation of Nogo-A from cell-surface biotinylated oligodendrocytes. Living cultures of oligodendrocytes were incubated with a cell-impermeable NHS-biotin analog. The biotinylated proteins were precipitated with streptavidin-coated beads, and the precipitate (Ppt) and supernatant (Sup) were analyzed by Western blot. All of the pellet and 1/10 of the supernatant were loaded. AS 472 showed the presence of Nogo-A in the precipitated sample, whereas the intracellular proteins β-tubulin and BiP were not present in the precipitated material. Intracellular Nogo-A, β–tubulin, and BiP were found in the supernatant.

Figure 10.

Topology of Nogo-A in cultured oligodendrocytes. Oligodendrocytes (3–5 d in culture) were either fixed and completely permeabilized with Triton X-100 (Tx-100) or only plasma membrane permeabilized with digitonin (DIG). Cells were incubated with different α-Nogo-A antibodies. All α-Nogo-A Abs, mAb 11C7, AS Bianca, and AS 922 specifically recognize oligodendrocytes in dissociated rat optic nerve cultures (left row). After selective permeabilization with DIG, mAb 11C7, AS Bianca, and α-actin IgM mAb stain Nogo in the cytoplasm of oligodendrocytes (right row), whereas AS 922 does not. Antibodies against the luminal ER protein BiP were used as a control for the selective permeabilization. In DIG-permeabilized cells, no α-BiP staining could be detected, whereas all cells were strongly stained in the Tx-100-permeabilized cultures. Scale bar, 30 μm.

Figure 9.

Nogo-A-specific epitopes are detected at the cell surface of 3T3 fibroblasts. 3T3 fibroblasts were detached with Cell Dissociating Buffer (Invitrogen) and incubated with 11C7 (α-Nogo-A) at a concentration of 0.1 μm for 1 hr. As an isotype control, the α-FLAG-M2 antibody is used at a concentration of 0.1 μm. After washing twice with PBS, the cells were incubated with a goat α-mouse FITC-labeled antibody (1:100) and analyzed by flow cytometry. Additional negative controls include unstained 3T3 fibroblasts and the secondary antibody alone. In all experiments dead cells were detected by Via-Probe (BD-PharMingen) and are excluded from the analysis. Note that the 11C7 antibody detects the Nogo-A-specific region on intact, unpermeabilized 3T3 fibroblasts.

Figure 11.

Intracellular localization of Nogo-A in cultured oligodendrocytes. Fixed oligodendrocytes permeabilized with Triton X-100 were double stained with mAb 11C7 (A, red) and the ER marker α-calnexin (B, green) or AS 472 (D, red) and α-MG160 outlining the Golgi complex (E, green) and analyzed by confocal microscopy in single optical sections. Nogo-A colocalizes with calnexin (C, yellow) as well as with the Golgi-marker (F, yellow). The magnification in C is taken from a different cell. The planes of the optical sections were chosen for optimal visualization of the colocalization of Nogo with the marker proteins in the cell body and the main processes. Both mAb 11C7 and AS 472 stain the entire cell with all the processes. Note the regions in which the Nogo-A immunoreactivity does not completely overlap with the ER-marker (C, arrowheads) or Golgi-marker (F, arrowheads). Scale bar: (in F) A–C, 20 μm; D–F, 15 μm.