The adaptor protein GULP promotes Jedi-1-mediated phagocytosis through a clathrin-dependent mechanism

Abstract

During the development of the peripheral nervous system, the large number of apoptotic neurons generated are phagocytosed by glial precursor cells. This clearance is mediated, in part, through the mammalian engulfment receptor Jedi-1. However, the mechanisms by which Jedi-1 mediates phagocytosis are poorly understood. Here we demonstrate that Jedi-1 associates with GULP, the mammalian homologue of CED-6, an adaptor protein required for phagocytosis mediated by the nematode engulfment receptor CED-1. Silencing GULP or mutating the NPXY motif in Jedi-1, which is required for GULP binding, prevents Jedi-1-mediated phagocytosis. How GULP promotes engulfment is not known. Of interest, we find that Jedi-1-induced phagocytosis requires GULP binding to clathrin heavy chain (CHC). During engulfment, CHC is tyrosine phosphorylated, which is required for Jedi-mediated engulfment. Both phosphoclathrin and actin accumulate around engulfed microspheres. Furthermore, knockdown of CHC in HeLa cells prevents Jedi-1-mediated engulfment of microspheres, and knockdown in glial precursors prevents the engulfment of apoptotic neurons. Taken together, these results reveal that Jedi-1 signals through recruitment of GULP, which promotes phagocytosis through a noncanonical phosphoclathrin-dependent mechanism.

FIGURE 1:

Jedi-1 interacts with GULP through its NPXY motif. (A) Endogenous Jedi was immunoprecipitated from lysates of E13.5 dorsal root ganglia and spinal cord, and the proteins were separated by SDS–PAGE and immunoblotted with an antibody to Jedi-1 or CED-6 (GULP). Representative blot of three experiments. (B) Jedi-1-GFP and Jedi-1-GFP with the NPXY motif mutated to APXA or the four ITAM tyrosines mutated to phenylalanine were transiently transfected into control MEF cells or MEF cells stably expressing GST-GULP. GST-GULP was pulled down with glutathione beads, and anti-GFP was used to detect coimmunoprecipitation of Jedi-1 by Western blot (n = 3).

FIGURE 2:

The NPXY motif of Jedi-1 is required for engulfment. Jedi-1-GFP, Jedi-1-GFP with the NPXY mutated to NPXF or APXA, or GFP only was transfected into HeLa cells, and 2 d later, fluorescent microspheres were added for 2 h before rinsing and fixing the cells. The cells were imaged by confocal microscopy to ensure complete internalization. (A) Confocal images with Jedi-1-GFP– or GFP-expressing cells (green) with microspheres (red). Cell nuclei are shown in blue. Scale bars, 10 μm. (B) Expression of Jedi-1-GFP and mutant Jedi-1-GFP confirmed by immunoblotting. (C) Percentage of transfected HeLa cells engulfing at least one microsphere was quantified (Student's t test, p = 0.00015 for Jedi-1 relative to GFP; p = 0.0003 for Jedi-1 NPXF and p = 0.0004 for Jedi-1 APXA relative to wild type Jedi-1; n = 3). (D) GFP, Jedi-1-GFP, or APXA mutant Jedi-1-GFP was transfected into glial cells in cocultures of E13.5 DRG neurons and glia in the presence of NGF. NGF was withdrawn to induce neuronal death, and 48 h later the cultures were fixed and immunostained with anti-GFP and nuclei labeled with TO-PRO-3. The percentage of transfected glia (GFP positive) engulfing at least one apoptotic body (based on condensed TO-PRO-3 staining) was quantified by confocal analysis (Student's t test, p = 0.002 for Jedi-1 relative to GFP, p = 0.007 for Jedi-1-APXA relative to wild-type Jedi-1; n = 3).

FIGURE 3:

GULP is required for Jedi-1–mediated engulfment. Control MEFs, MEFs with GULP stably knocked down (psiGULP), or MEFs with endogenous GULP silenced but stably transfected with a knockdown-resistant GULP (psiGULP+GST-GULP) were transfected with GFP, Jedi-1-GFP, or the NPXY/APXA mutant of Jedi-GFP. (A) The cells were lysed, and the proteins were separated by SDS–PAGE and Western blotted for GFP (to detect Jedi-1) or tubulin. (B) Fluorescent microspheres were added to the cells, and after 2 h the cells were fixed and engulfment of microspheres was analyzed by confocal microscopy and quantified to determine the percentage of transfected cells (GFP positive) engulfing at least one microsphere (by two-way analysis of variance with a Bonferroni posthoc analysis; p < 0.001 for MEFs with wild-type Jedi-1 relative to MEFs with GULP knockdown–expressing wild-type Jedi-1; p < 0.001 for MEFs with GULP knocked down with wild-type Jedi-1 relative to MEFs with GULP knocked down and rescued with GST-GULP with wild-type Jedi).

FIGURE 4:

Knockdown of GULP reduces the engulfment of neurons by glial precursors. (A) Nontargeting or GULP shRNA was transfected into 3T3 cells. Lysates were subjected to Western blotting with antibodies to GULP and tubulin. (B) Knockdown of GULP was also verified by immunostaining glial cell precursors transfected with GULP shRNA for GFP to identify transfected cells (green) and GULP (red). Nuclei were stained with TO-PRO-3. (C) Representative confocal images of cocultures of E13.5 DRG neurons and glial precursor cells transfected with nontargeting or GULP shRNA (coexpressing GFP). On the day of transfection, neuronal death was induced by removing NGF. After 48 h, the cultures were fixed and immunostained with anti-GFP and nuclei labeled with TO-PRO3. Transfected glia are depicted in green and nuclei in red. Scale bars, 20 μm. The cross-hair in the orthogonal view of the cell transfected with scrambled shRNA is positioned on an engulfed apoptotic nuclear fragment, and the cross-hair in the GULP shRNA–transfected image is positioned on the cell body of a glial cell containing no apoptotic nuclear fragments. (The rectangle on top of the images depicts the xz-plane, and to the right is the yz-plane.) (D) Images from transfected cocultures were analyzed by confocal microscopy. The percentage of transfected glia engulfing TO-PRO-3 stained apoptotic nuclear fragments was quantified (p = 0.0005 for GULP shRNA relative to GFP only, by Student's t test; n = 3).

FIGURE 5:

The NPXY motif and GULP are required for internalization of Jedi-1. Time course for internalization of Jedi-1-GFP (A) or Jedi-1-GFP APXA mutant (B) after addition of microspheres. HeLa cells were transfected with wild-type or APXA mutant Jedi-1-GFP. Surface proteins were biotinylated with EZ-Link Sulfo-NHS-SS-Biotin at 4°C and then exposed to fluorescent microspheres and warmed to 37°C for the time indicated. The biotinylation of surface proteins was then reversed using the reducing agent DTT. The cells were lysed, and internalized, biotinylated Jedi-1 or mutant Jedi-1 was pulled down with avidin-conjugated agarose beads and detected by immunoblotting with a GFP antibody. Total levels of Jedi-1 are shown in lysates (n = 3). (C) Control MEFs or MEFs with GULP knocked down (MEF psiGULP) were transfected with Jedi-GFP. The cells were then exposed to microspheres or left untreated for 90 min, and the internalized Jedi-1 was detected as in A and B. Levels of Jedi-1–GFP and GULP are shown in the lysates.

FIGURE 6:

GULP interacts with clathrin, and this interaction is required for Jedi-1–mediated engulfment. (A) HeLa cells were transfected with pcDNA3, HA-GULP, or HA-GULP lacking the C-terminal 71 amino acids (G10). After 2 d, the cells were lysed, immunoprecipitated with HA antibody, and subjected to Western blotting with antibodies to HA or CHC. Endogenous GULP levels were analyzed by immunodepleting the lysates of HA-GULP using two subsequent rounds of immunoprecipitation with HA antibody and blotting the supernatant lysate with an antibody that recognizes endogenous GULP. A representative blot (n = 3). (B) HeLa cells transfected with Jedi-1-GFP and pcDNA3, HA-GULP or HA-G10 were lysed and immunoprecipitated with an antibody to HA and blotted with an antibody to GFP to detect coimmunoprecipitated Jedi-1. The lysates were subjected to Western blotting using antibodies to HA (to detect GULP) or GFP (to detect Jedi-1). (C) Control MEFs or MEFs with GULP constitutively knocked down (MEF psiGULP) were transfected with GFP, Jedi-1–GFP, or Jedi-1–GFP and HA-GULP or HA-G10. Two days later, microspheres were added for 2 h. The cells were then fixed, and the engulfment of the microspheres was quantified by confocal microscopy (p = 0.0001 for MEFs with GFP relative to MEFs expressing Jedi-1; p = 0.0009 for MEFs expressing Jedi-1 relative to MEFs expressing Jedi-1 with GULP knocked down; p = 0.0009 for MEFs expressing Jedi-1 with GULP knocked down and resistant GULP added back relative to MEFs expressing Jedi-1 with GULP knocked down and transfected with G10, by Student's t test).

FIGURE 7:

Clathrin is required for Jedi-1–mediated engulfment. (A) Nontargeting or CHC shRNA were transfected into HeLa cells. Lysates were subjected to Western blotting with antibodies to CHC and tubulin. (B) Knockdown of CHC was also verified by immunostaining glial cell precursors transfected with CHC shRNA for GFP to identify transfected cells (green) and CHC (red). Nuclei were stained with TO-PRO-3. (C) HeLa cells were cotransfected with Jedi-1–FLAG and GFP, nontargeting shRNA, or CHC shRNA that coexpresses GFP. After 48 h, fluorescent microspheres were incubated with the HeLa cells for 2 h before rinsing and fixation. The percentage of GFP positive cells that had engulfed one or more spheres was determined by confocal microscopy (n = 3; p = 0.006 for CHC shRNA relative to Jedi-1-FLAG–only control by Student's t test). (D) Confocal images depicting cocultures of E13.5 neurons and glial precursor cells transfected with GFP or CHC shRNA coexpressing GFP. On the day of transfection, neuronal death was induced by removal of NGF, and after 48 h the cocultures were fixed and immunostained with antibodies to GFP and nuclei were labeled with TO-PRO-3. Transfected glia are depicted in green and nuclei in red. The cross-hair in the orthogonal view of the cell transfected with scrambled shRNA is positioned on an engulfed apoptotic nuclear fragment, and the cross-hair in the CHC shRNA–transfected image is positioned on the cell body of a glial cell containing no apoptotic nuclear fragments. (The rectangle on top of the images depicts the xz-plane, and to the right is the yz-plane.) Scale bars, 20 μm. (E) Images from transfected cocultures were analyzed by confocal microscopy, and the percentage of GFP-positive glia that were engulfing apoptotic nuclear fragments was quantified (n = 3; p = 0.0002 for CHC shRNA relative to nontargeting control, by Student's t test).

FIGURE 8:

CHC is phosphorylated during Jedi-1–mediated engulfment in a GULP-dependent manner. Control MEFs, MEFs with GULP stably knocked down (psiGULP), or MEFs with endogenous GULP silenced but stably transfected with a knockdown-resistant GST-GULP (psiGULP+GST-GULP) were transfected with pcDNA3 or Jedi-1–GFP and then treated with or without fluorescent microspheres for 90 min. (A) The cells were lysed and immunoprecipitated with anti-CHC. The protein was subjected to Western blotting using anti–phospho-CHC and total immunoprecipitated CHC (n = 3). (B) Densitometry of phospho-CHC divided by the total CHC signal is shown n = 3 experiments (p = 0.02 for MEFs transfected with Jedi-1 with bead stimulation relative to Jedi-1–transfected MEFs with no bead stimulation, and p = 0.0009 for MEFs with GULP knockdown and GST-GULP rescue with Jedi-1 and bead stimulation relative to MEFs with GULP knockdown and GST-GULP rescue with Jedi-1 and no bead stimulation, by Student's t test). (C) The experiment in A was repeated, but total CHC was immunoprecipitated from the lysates and Western blotted for phosphotyrosine (n = 3).

FIGURE 9:

Phospho-CHC and actin colocalize with Jedi-1 and accumulate around engulfed microspheres. (A) MEFs transfected with Jedi-1–GFP (green) and treated with fluorescent microspheres (red) for 2 h were fixed and immunostained with anti–phospho-CHC (blue). The cells were imaged using confocal microscopy to ensure full internalization of microspheres. Note the accumulation of Jedi-1 and phospho-CHC around the internalized microsphere. (B) MEFs transfected with Jedi-1–FLAG (blue) were incubated with fluorescent microspheres and then fixed and stained with anti-FLAG and Alexa 488–phalloidin (green) to label actin. Note the ring of actin and Jedi-1–FLAG around the internalized microsphere. (C) Jedi-1–transfected MEF cells were incubated with fluorescent microspheres (red) and then fixed and stained with an antibody recognizing all CHC and phalloidin. Note the accumulation of phalloidin and CHC near the internalized microsphere. Scale bars, 5 μm. The regions boxed in white in A–C are enlarged and shown to the right of the merged image (only the blue and red channels are shown). (D) Table representing results of quantification of accumulation of actin or phospho-CHC within 0.5 μm of internalized beads and surrounding them by at least 50%. Distal beads are beads >10 μm from the nearest edge of the nucleus, and proximal beads are <10 μm from the nucleus. Representative images used for this quantification are shown to the right of the table. The nucleus is in blue and outlined with a dotted white line, and anti-phospho-CHC or phalloidin labeling is depicted in green, as indicated. Note that microspheres distal to the nucleus (arrows) have more phosphoclathrin or actin staining than beads that are proximal to the nucleus (arrowheads).

FIGURE 10:

Clathrin is required for recruitment of actin to the phagocytic cup. HeLa cells were transfected with Jedi-GFP and control siRNA or CHC siRNA. Cells were incubated for 45, 60, or 90 min with fluorescent microspheres. Localization of Jedi-GFP (green) and actin (blue) was analyzed using confocal microscopy. Scale bars, 5 μm (low-magnification images), 2 μm (enlarged images). Regions boxed in white are enlarged and shown to the right of the original image. Far right, only the blue and red channels. Note that beads are still in contact with Jedi-GFP even with CHC siRNA treatment, but the recruitment of actin is delayed and less robust. Representative images from three experiments.

FIGURE 11:

Phospho-CHC is required for Jedi-1–mediated engulfment of microspheres. (A) HeLa cells were transfected with Jedi-FLAG, control or CHC targeting siRNA, and GFP, WT-CHC-GFP, or phosphomutant CHC-GFP and verified for expression and knockdown by immunoblotting. WT-CHC-GFP and phosphomutant-CHC-GFP are siRNA resistant. (B) Cells were plated for a microsphere engulfment assay, and the number of GFP/FLAG double-positive cells engulfing at least one microsphere was determined (n = 3; p = 0.006 for Jedi-FLAG control siRNA vs. Jedi-FLAG with CHC siRNA; p = 0.04 for Jedi-FLAG CHC siRNA WT-CHC-GFP vs. Jedi-FLAG CHC siRNA phosphomutant CHC-GFP transfected cells by Student's t test).