Fates of neurotrophins after retrograde axonal transport: phosphorylation of p75NTR is a sorting signal for delayed degradation

Abstract

Neurotrophins can mediate survival or death of neurons. Opposing functions of neurotrophins are based on binding of these ligands to two distinct types of receptors: trk receptors and p75NTR. Previous work showed that target-derived NGF induces cell death, whereas BDNF and NT-3 enhance survival of neurons in the isthmo-optic nucleus of avian embryos. To determine the fate of retrogradely transported neurotrophins and test whether their sorting differs between neurotrophins mediating survival- or death-signaling pathways, we traced receptor-binding, sorting, and degradation kinetics of target-applied radiolabeled neurotrophins that bind in this system to trk receptors (BDNF, NT-3) or only to p75NTR (NGF). At the ultrastructural level, the p75NTR-bound NGF accumulates with a significant delay in multivesicular bodies and organelles of the degradation pathway on arrival in the cell body when compared with trk-bound BDNF or NT-3. This delayed lysosomal accumulation was restricted to target-derived NGF, but was not seen when NGF was supplied to the soma in vitro. The kinase inhibitors K252a and Gö6976 alter the kinetics of organelle accumulation: phosphorylation of p75NTR is a sorting signal for delayed sequestering of p75NTR-bound NGF in multivesicular bodies and delayed degradation in lysosomes when compared with trk-bound neurotrophins. Mutagenesis and mass spectrometry studies indicate that p75NTR is phosphorylated by conventional protein kinase C on serine 266. We conclude that, in addition to the known phosphorylation of trks, the phosphorylation of p75NTR can also significantly affect neuronal survival in vivo by changing the intracellular sorting and degradation kinetics of its ligands and thus signaling duration.

Figure 1.

A–D, Receptor binding of retrogradely transported neurotrophins and expression of neurotrophin receptors in the ION of 14- to 16-d-old chick embryos. A, Immunoprecipitation of radiolabeled BDNF cross-linked to its receptors reveals a significant shift en route, from p75NTR (p75) in the chiasm to trkB on arrival in the ION. Data are expressed as a percentage of BDNF relative to all receptor-bound BDNF. B, NGF binds nearly exclusively to p75NTR (p75), consistent with the absence of trkA expression by the ION (see C). C, Expression of trks and p75NTRs (p75) by the ION, as measured by qualitative RT-PCR. Note the lack of trkA expression and the relative abundance of p75NTR. D, Quantitative RT-PCR of trk and p75NTR (p75) expression in the ION, confirming virtual lack of trkA expression. Error bars indicate SEM.

Figure 2.

A–E, Accumulation of retrogradely transported neurotrophins in MVBs and lysosomes in the ION of E15 chick embryos (A–D) and comparison of NGF and BDNF arrival and accumulation in the ION (E). A, Silver grain indicative of retrogradely transported neurotrophin in an MVB within the ION. B, Silver grain indicative of retrogradely transported neurotrophin in a lysosome within the ION. Scale bar: A, B, 200 nm. Inset, Lysosome showing dark precipitate after enzyme histochemistry for acid phosphatase. Scale bar, 200 nm. C, Quantification of accumulation of NGF, BDNF, and NT-3 in MVBs at 10, 20, and 32 h after intraocular injection by autoradiography. Data are based on two to three independent experiments. Error bars indicate SEM. Statistically significant differences in C–E (unpaired Student's t test) between adjacent bars are indicated by asterisks (*p < 0.05; **p < 0.025; ***p < 0.01). D, Quantification of accumulation of NGF, BDNF, and NT-3 in lysosomes at 10, 20, and 32 h after intraocular injection. Note the consistent difference between NGF and the other two neurotrophins (BDNF and NT-3) in the kinetics of both MVB and lysosome transition. E, Accumulation of radiolabeled NGF and BDNF in the ION after intraocular injection. Note that both neurotrophins (NGF and BDNF) arrive in the ION at 5–10 h, and their accumulations peak at 15–20 h. More NGF accumulates at 15–20 h, consistent with a delayed degradation on arrival compared with BDNF. Error bars indicate SEM. The number (n) of independent experiments is indicated on bars. Values of p are indicated by an asterisk (*) only when they reach statistical significance.

Figure 3.

A–L, Comparison of accumulation and transition kinetics in MVBs and lysosomes (Lyso) of the E15 ION between NGF (A–F) and BDNF (G–L) at 20 and 32 h after intraocular injection of neurotrophin only, or neurotrophin plus kinase inhibitors K252a or Gö6976 (Go). Data were obtained in two to three independent experiments with ∼150 silver grains scored for each condition to determine labeling densities. Error bars indicate SEM. Statistical significance (t test) of changes within panels is indicated by black asterisks above error bars. We indicate statistically significant differences compared with the neurotrophin-only condition by white asterisks within the black bars, using Bonferroni's corrections for multiple comparisons. Note that NGF kinetics are significantly reversed by K252a and/or Gö6976 coinjection for MVBs and lysosomes, but the same inhibitors have no effect on BDNF kinetics of organelle transition. NGF normally did not accumulate in the MVB at 20 h (LD ∼ 1.0), but when coinjected with either K252a or Gö6976, NGF accumulated significantly (p < 0.01 and p < 0.05) with LDs of 11–16. NGF normally accumulated in lysosomes at 20 h with a LD of ∼2.0, but K252a changed this significantly to an LD of 14 (p < 0.01). NGF normally accumulated at 32 h in lysosomes with an LD of ∼18, but with either K252a or Gö6976, this accumulation was significantly (p < 0.05) reduced to LDs of 4 (K252a) and 7 (Gö6976). Statistically significant differences are indicated by asterisks (*p < 0.05; **p < 0.025; ***p < 0.01; ****p < 0.005).

Figure 4.

A–C, Conventional protein kinase C isoforms (PKCα, -β, -γ) effectively phosphorylate p75NTR, and this phosphorylation is abolished by the kinase inhibitor K252a. A, Human p75NTR is phosphorylated in vitro by several kinases, notably by conventional and atypical PKC, CAMKII, and mitogen-activated protein kinase (p38β), but not by PKA, PKCδ, PKCε, or CSII (casein kinase II). B, K252a (50 nm) inhibits phosphorylation of p75NTR by PKCγ, but not by PKCζ or p38β. C, Evidence for direct interaction in vitro between purified p75NTR and purified PKCγ. Recombinant His-tagged PKCγ (∼80 kDa) is coimmunoprecipitated with the intracellular domain of p75NTR (p75-ID). Detection and visualization was performed with anti-His antibody and alkaline phosphatase-labeled secondary antibody. Lane 1, In the presence of all reaction components (purified PKCγ, purified p75-ID, and anti-p75 antibody), an intense PKCγ band is visible (∼80 kDa). Lane 2, When p75-ID is omitted, PKCγ band is barely detectable. Lane 3, Control reaction without anti-p75 antibody. PKCγ band is not present. Lane 4, Standard band of 200 ng of PKCγ (∼80 kDa).

Figure 5.

A–C, Mutagenesis analysis shows that phosphorylation of p75NTR by conventional PKC requires serine 277 in p75NTR. A, Wild-type (wt) and mutated versions of human p75NTR were used for in vitro kinase assay with active PKCγ and ³²P-labeled ATP. Mutant 1 (m-1) has serine 277 and threonine 358 mutated to alanine; mutant 2 (m-2) has serines 303, 305, 308, 311, and 313 mutated to alanines; mutant 3 (m-3) has threonine 293 and serines 354 and 425 mutated to alanines; mutant 4 (m-4) has exclusively serine 277 mutated to alanine. B, Coomassie-stained SDS-PAGE gel shows similar protein loading for wt and p75NTR mutants. C, Diagram [adapted from Roux and Barker (2002)] shows the structure of vertebrate p75NTR. The arrow indicates serine 277 (numbering according to human p75NTR).

Figure 6.

A–E, Accumulation of neurotrophins in lysosomes of E14 ION neurons in vitro and effects of NGF and kinase inhibitors on the number of dying neurons in vivo. A, Quantification of lysosomal accumulation of silver grains representing radiolabeled NGF and BDNF after 1–4 h incubation. Error bars indicate SEM. Data are based on three independent experiments with a total of 472 neurons assessed, and 1581 silver grains scored. B, Quantification of lysosomal accumulation of NGF in the presence of 40 ng constitutively active conventional PKC after 3 and 4 h. Note the shift in the peak from 3 to 4 h, significantly different from B at p < 0.05. Error bars indicate SEM. Data are based on three independent experiments with a total of 372 neurons assessed and a total of 913 silver grains scored. C, Inset, Silver grain over a lysosome after internalization of radiolabeled NGF by dendrites or soma at 3 h. Scale bar, 200 nm. D, Retina-derived NGF (1 μg per eye) promotes the death of developing ION neurons, as previously reported (von Bartheld et al., 1994; Janiga et al., 2000). Coinjection of either of the kinase inhibitors, K252a or Gö6976 (2 μl of 0.25 mm), together with NGF significantly decreases apoptosis of ION neurons. Injection of the kinase inhibitors alone has a significantly smaller effect on apoptosis. Each embryo was injected into the right eye, and the experimental (left) ION was compared with the right ION innervating the noninjected or sham-injected control eye. Error bars indicate SEM. E, Inset, Example of a pyknotic cell (arrow) in the ION. Scale bar, 10 μm.