Endocytosis of activated TrkA: evidence that nerve growth factor induces formation of signaling endosomes

Abstract

The survival, differentiation, and maintenance of responsive neurons are regulated by nerve growth factor (NGF), which is secreted by the target and interacts with receptors on the axon tip. It is uncertain how the NGF signal is communicated retrogradely from distal axons to neuron cell bodies. Retrograde transport of activated receptors in endocytic vesicles could convey the signal. However, little is known about endocytosis of NGF receptors, and there is no evidence that NGF receptors continue to signal after endocytosis. We have examined early events in the membrane traffic of NGF and its receptor, gp140(TrkA) (TrkA), in PC12 cells. NGF induced rapid and extensive endocytosis of TrkA in these cells, and the receptor subsequently moved into small organelles located near the plasma membrane. Some of these organelles contained clathrin and alpha-adaptin, which implies that TrkA is internalized by clathrin-mediated endocytosis. Using mechanical permeabilization and fractionation, intracellular organelles derived from endocytosis were separated from the plasma membrane. After NGF treatment, NGF was bound to TrkA in endocytic organelles, and TrkA was tyrosine-phosphorylated and bound to PLC-gamma1, suggesting that these receptors were competent to initiate signal transduction. These studies raise the possibility that NGF induces formation of signaling endosomes containing activated TrkA. They are an important first step in elucidating the molecular mechanism of NGF retrograde signaling.

Fig. 4.

Strategy for cell fractionation experiments. NGF (1 nm) was bound to PC12 cells at 4°C for 1 hr. Cells were then either briefly washed in binding buffer at 4°C or not washed, and warmed at 37°C for 10 min. Cells were then chilled (4°C), washed, resuspended in a cytoplasm-like buffer, and permeabilized by passage through a ball homogenizer. The cell ghosts (P1) were separated from cytosol and organelles released from the cells by centrifugation at 1000 × g. Two alternative strategies were used to fractionate the membranes in the supernatant of the 1000 × g centrifugation. (1) They were layered over a 0.4 ml pad of 10% buffered sucrose and centrifuged at 100,000 × g (1 hr), forming a pellet (P2′) and the cytosol (S2′). (2) To separate large and small vesicles, membranes were centrifuged at 8000 × g for 35 min to pellet large vesicles (P2). The supernatant of the 8000 ×g spin was then layered over a 0.4 ml pad of 10% buffered sucrose and centrifuged at 100,000 × gfor 1 hr, which separated small vesicles (P3) from cytosol (S3).

Fig. 8.

TrkA and tyrosine-phosphorylated TrkA were detected in intracellular organelles. Top, PC12 cells incubated with or without NGF (1 nm) at 4°C were washed, warmed 10 min (37°C), chilled (4°C), permeabilized, and fractionated as in Figure 4. Equal amounts of cells were used to compare conditions. TrkA was immunoprecipitated with RTA from one-fifth of P1, one-half of P2, one-half ofP3, and one-tenth of S3. Shown is a Western blot of immunoprecipitates probed with RTA followed by HRP-conjugated anti-rabbit IgG. Chemiluminescence was used for detection. The bands for gp140^TrkA and gp110^TrkA are noted. Bottom, TrkA immunoprecipitates (as above) were Western-blotted and probed with anti-phosphotyrosine antibody (4G10) followed by¹²⁵I-labeled goat anti-mouse IgG. Data were taken directly from the PhosphorImager. The position of the tyrosine-phosphorylated 140 kDa band comigrated exactly with TrkA. Tyrosine-phosphorylated TrkA was present in P1, P2, andP3 in NGF-treated cells.

Fig. 9.

The “specific activity” of TrkA tyrosine phosphorylation increased after NGF treatment. A, Data for gp140^TrkA and gp110^TrkA from three experiments as in the top of Figure 8 were quantified by densitometry and plotted with error bars (±SEM). The proteins, conditions, and fractions are labeled at the left. B, Data for tyrosine-phosphorylated TrkA from three experiments as in the bottom of Figure 8 were quantified by PhosphorImaging or densitometry and plotted as in A. The values are reported as a percentage of total tyrosine-phosphorylated TrkA in NGF-treated cells.C, The specific activity of TrkA tyrosine phosphorylation was the ratio of the amount of tyrosine-phosphorylated TrkA to the amount of gp140^TrkA, plotted in arbitrary units. The average specific activity was calculated using data from four individual experiments. Differences between the P2and P3 fractions within a treatment group were not significant. When comparing fractions from control and NGF-treated cells (for example, P1-control vs P1-NGF-treated), significant differences, calculated using Student’s t test, are indicated by the probability value (P).

Fig. 10.

PLC-γ1 was bound to TrkA in intracellular organelles. PC12 cells incubated at 4°C for 1 hr with NGF (1 nm; lanes 2–4,6–8, 10) or without NGF (lanes 1, 5, 9) were warmed 10 min at 37°C. Cells were chilled (4°C), permeabilized, and fractionated as in Figure 4. P1, P2′, andS2′ were lysed and immunoprecipitated with anti-PLC-γ-1 (lanes 1, 2,5, 6, 9,10) or anti-TrkA (1088; lanes 4,8). In lanes 3 and 7, TrkA-immunoprecipitated lysates were subsequently immunoprecipitated with anti-PLC-γ1. Immunoprecipitates were Western-blotted with anti-phosphotyrosine antibody (4G10; A) and then stripped and reprobed with anti-PLC-γ1 (B) and, finally, with anti-TrkA (RTA; C). Chemiluminescence was used for detection: P1 and S2′, 1 min exposure; P2′, 2 hr exposure. The positions for PLC-γ1 and TrkA are indicated. NGF treatment resulted in association of tyrosine-phosphorylated TrkA and tyrosine-phosphorylated PLC-γ1 inP1 and P2′ (see text). Although anti-PLC-γ1 brought down TrkA in P1 andP2′ after NGF treatment (A,C, lanes 2, 6), PLC-γ1 was not reproducibly found on blots of TrkA immunoprecipitates (lanes 4, 8, A,B). Because TrkA immunoprecipitation clearly brought down PLC-γ1 in both P1 and P2′, the complexes present in the TrkA immunoprecipitates may have been unstable.

Fig. 1.

NGF induced trkA internalization.A, NGF treatment decreased cross-linking to surface TrkA receptors. PC12 cells, an equal number for each condition tested, were incubated with [¹²⁵I]NGF (1 nm) at 4°C for 2 hr and then warmed to 37°C for 0, 5, 10, or 30 min. Cells were chilled, and the membrane-impermeant cross-linker BS³ was added for 30 min at 4°C. Cell lysates were immunoprecipitated with 1088, an anti-Trk C-terminal antibody, before SDS-PAGE. The dried gel was exposed to x-ray film. The positions of molecular weight markers are indicated, as is a band corresponding to a complex containing an [¹²⁵I]NGF monomer cross-linked to a TrkA monomer. The more slowly migrating band marks a higher-molecular-weight complex containing [¹²⁵I]NGF and TrkA. There was no [¹²⁵I]NGF cross-linking in experiments in which unlabeled NGF (1 μm) was present during binding and cross-linking.B, Constitutive and NGF-induced internalization of TrkA: internalization of TrkA increased after warming in the presence of NGF (+NGF). Cells were incubated with unlabeled NGF (1 nm) for 30 min at 4°C (untreated). Control samples were handled identically except that no NGF was present. NHS-SS-biotin was added at 4°C to biotinylate cell surface proteins; cells were then either kept on ice or warmed 10 min at 37°C to allow for endocytosis. Glutathione was added to some samples to remove biotin on cell surface proteins. After samples were lysed and boiled in 0.5% SDS, TrkA was immunoprecipitated with 1088. After SDS-PAGE, proteins were transferred to nitrocellulose and blotted with [¹²⁵I]streptavidin. TrkA is marked by anarrow. Biotinylated TrkA was analyzed without warming and without glutathione (100%), without warming but with glutathione (bkg), or with warming and with glutathione (int). Biotin on TrkA receptors internalized during warming was protected from glutathione reduction.C, Quantitation of TrkA internalization. Experiments were performed as in B. The signals for biotinylated TrkA were normalized for TrkA protein. The percent internalization of TrkA was computed as described in Materials and Methods. Values are mean ± SEM from three separate experiments at 10 min and two at 20 min (open circles, untreated; open squares, NGF-treated). There was a low level of constitutive internalization. NGF markedly increased TrkA internalization.

Fig. 2.

NGF changed the distribution of TrkA immunostaining in PC12 cells. Cells were exposed to media at 37°C without NGF (a) or with NGF (2 nm) for 30 sec (b), 2 min (c), and 60 min (d). A Trk-specific antibody, sc11, was used to examine the distribution of TrkA. Most TrkA staining was intracellular. With NGF treatment there was an increase in bright punctate staining near the plasma membrane (e.g., small arrows inb). Note the marked increase with NGF treatment of TrkA staining in the juxtanuclear region at 60 min (large arrows in d).

Fig. 3.

NGF treatment resulted in TrkA and clathrin colocalization. PC12 cells were treated with NGF for 30 sec (A–C, G,H) or with vehicle alone for the same interval (D–F). After fixation and permeabilization, cells were immunostained using antibodies to Trk (sc11) and the clathrin heavy chain (X22). TrkA staining is shown ingreen; clathrin heavy chain staining is inred. In the absence of NGF, TrkA (D) and clathrin (E) staining is present diffusely in the cytosol and in the juxtanuclear region. There is little overlap in their distribution (F) except in the juxtanuclear region. In the presence of NGF, TrkA (A) and clathrin (B) staining is widely distributed in the cytosol; some TrkA staining is seen near the plasma membrane. Clathrin staining appears to be concentrated at the plasma membrane. C (at lower power) and G and H (at higher power; scale bar, 2 μm) show colocalization of TrkA and clathrin staining near the plasma membrane (arrows;yellow denotes colocalization). The organelles showing colocalization had the same distribution and size as those seen with increased frequency after NGF treatment (Fig. 2).

Fig. 5.

Electron micrographs of PC12 cells and vesicles released from these cells. Cell fractions were processed as indicated in Materials and Methods. A, A cell not permeabilized.B, P1: a cell after permeabilization. Note the marked decrease in the electron density of the cytoplasm. Numerous discontinuities were seen in the plasma membrane (arrow). C, Organelles in the P2 fraction. D, Organelles in the P3 fraction. Scale bars: A, B, 0.5 μm; C, D, 0.4 μm.

Fig. 6.

Sucrose gradient fractionation of internalized NGF. Cells incubated with [¹²⁵I]NGF (1 nm) for 1 hr at 4°C were washed, warmed 10 min, chilled (4°C), and then permeabilized. P2′ was applied to 10–40% sucrose gradients over a 50% sucrose pad and centrifuged at 100,000 × gfor 1 hr. Gradient fractions were collected from the bottom of the tube. [¹²⁵I]NGF was quantified in each fraction. Data are representative of two experiments.

Fig. 7.

TrkA was cross-linked to NGF in intracellular organelles. Cells were incubated with [¹²⁵I]NGF (1 nm), washed, warmed 10 min (37°C), chilled (4°C), and then permeabilized and fractionated as in Figure 4. The membrane-permeable cross-linking reagent DSS was added before fractionation. One-fifth of the cell ghost membranes (P1), the entire 8000 × g pellet (P2), the entire 100,000 × g pellet (P3), and one-tenth of the 100,000 ×g supernatant (S3) were immunoprecipitated with 1088 and analyzed by SDS-PAGE and autoradiography. The arrow marks the cross-linked complex containing TrkA and [¹²⁵I]NGF in P1, P2, and P3. There was no cross-linking when [¹²⁵I]NGF binding was carried out in the presence of unlabeled NGF (1 μm). The amount of [¹²⁵I]NGF cross-linked to TrkA was quantified by PhosphorImager.