Phosphoglycerate kinase 1 acts as a cargo adaptor to promote EGFR transport to the lysosome

Abstract

The epidermal growth factor receptor (EGFR) plays important roles in multiple cellular events, including growth, differentiation, and motility. A major mechanism of downregulating EGFR function involves its endocytic transport to the lysosome. Sorting of proteins into intracellular pathways involves cargo adaptors recognizing sorting signals on cargo proteins. A dileucine-based sorting signal has been identified previously for the sorting of endosomal EGFR to the lysosome, but a cargo adaptor that recognizes this signal remains unknown. Here, we find that phosphoglycerate kinase 1 (PGK1) is recruited to endosomal membrane upon its phosphorylation, where it binds to the dileucine sorting signal in EGFR to promote the lysosomal transport of this receptor. We also elucidate two mechanisms that act in concert to promote PGK1 recruitment to endosomal membrane, a lipid-based mechanism that involves phosphatidylinositol 4,5-bisphosphate [PI(4,5)P2] and a protein-based mechanism that involves hepatocyte growth factor receptor substrate (Hrs). These findings reveal an unexpected function for a metabolic enzyme and advance the mechanistic understanding of how EGFR is transported to the lysosome.

Fig. 1. PGK1 promotes EGFR transport to the lysosome.

Quantitative results are shown as mean ± s.e.m.; *p < 0.05, NS (not significant) p > 0.05, unpaired two-sided Student’s t test. a Colocalization of VSVG with a GM130 was performed, n = 10 cells examined over 3 independent experiments. Statistics is shown for the 0, 1, 2, and 5-min time point, P = 0.485, 0.685, 0.694, and 0.3059, respectively. b Colocalization of VSVG-KDELR with GM130 was performed, n = 10 cells examined over 3 independent experiments. Statistics is shown for the 0, 20, 40, and 60-min time point, P = 0.938, 0.238, 0.1303, and 0.061, respectively. c Colocalization of VSVG with TGN46 was performed, n = 10 cells examined over 3 independent experiments. Statistics is shown for the 0, 20, 40, and 60-minute time point, P = 0.741, 0.546, 0.1244, and 0.948, respectively. d Colocalization of CTxB with TGN46 was performed, n = 10 cells examined over 3 independent experiments. Statistics is shown for the 10, 20, and 30-min time point, P = 0.5119, 0.3485, and 0.3766, respectively. e Colocalization of EGF with EEA1 was performed, n = 10 cells examined over 3 independent experiments. Statistics is shown for the 1, 5, and 10-min time point, P = 0.808, 0.092, and 0.2744, respectively. f Colocalization of Tf with Rab11 was performed, n = 10 cells examined over 3 independent experiments. Statistics is shown for the 0, 10, 20, and 30-minute time point, P = 0.4079, 0.1493, 0.5788, and 0.3328, respectively. g Colocalization of EGF with LAMP1 was performed, n = 10 cells examined over 3 independent experiments. Statistics is shown for the 20, 40, 60, and 80-min time point, P = 0.1066, 8.4 × 10⁻⁴, 2.08 × 10⁻¹¹, and 6.9 × 10⁻⁵, respectively. h Colocalization of EGF with LAMP1 was performed, n = 10 cells examined over 2 independent experiments. Statistics is shown for the 20, 40, 60, and 80-min time point, P = 0.218, 0.0006, 0.00048, and 0.654, respectively. i Colocalization of CXCR4 with LAMP2 was performed, n = 15 cells examined over 3 independent experiments. Statistics is shown for the 20, 40, 60, and 90-min time point, P = 0.2962, 0.4448, 0.6859, and 0.9378, respectively. j Total TfR level was quantified, n = 4. Statistics was performed for the 0, 16, and 24-h time point, P = 0.8127, 0.203, and 0.1873, respectively.

Fig. 2. PGK1 promotes EGFR degradation.

Quantitative results are shown as mean +/- s.e.m.; *p < 0.05, NS p > 0.05, unpaired two-sided Student’s t test. a Time-course analysis of EGFR degradation upon EGF treatment in HeLa cells. The effect of siRNA against PGK1 is examined. Cell lysates were immunoblotted for proteins indicated, n = 3. Two different siRNAs against PGK1 show similar results, with a representative experiment shown. Sequence #1 was used for all other experiments that involve siRNA against PGK1. b Quantitation of EGFR level at the 120-min time point for the analysis above that used siRNA #1 to reduced PGK1 level. EGFR level was normalized to tubulin level, P = 4.27×10⁻⁵. c Quantitation of ERK activation at the 120-min time point for the analysis above that used siRNA #1 to reduced PGK1 level. Phosphorylated ERK level was normalized to total ERK level, P = 0.017. d Time-course analysis of EGFR degradation upon EGF treatment for indicated time point in HeLa cells, examining the effect of PGK1 overexpression. Cell lysates were immunoblotted for proteins indicated, n = 5. A representative result is shown. e Quantitation of EGFR level at the 60-min time point for the analysis above. EGFR level was normalized to tubulin level, P = 0.003. f Quantitation of ERK activation at the 60-min time point for the analysis above. Phosphorylated ERK level was normalized to total ERK level, P = 0.0039. g Time-course analysis of CXCR4 degradation upon SDF-1α treatment for indicated time point in HeLa cells, examining the effect of siRNA against PGK1. Cell lysates were immunoblotted for proteins indicated, n = 3. A representative result is shown above. Quantitation is shown below for the 3-h time point, P = 0.1703. CXCR4 level was normalized to tubulin level. h Assay for endocytic transport of EGFR to the lysosome, comparing the effect of expressing wild-type versus catalytic-dead form of PGK1. Colocalization of EGF with LAMP1 was performed, n = 30 cells examined over 3 independent experiments. Representative images are shown on left with EGF in red and LAMP1 in green, bar = 10 μm. Quantitation is shown on right with statistical analysis performed for the 90-min time point, P = 0.9868.

Fig. 3. PGK1 recognizes the dileucine sorting signal in EGFR.

Quantitative results are shown as mean +/- s.e.m.; *p < 0.05, unpaired two-sided Student’s t-test. a Co-precipitation analysis examining the association of PGK1 with EGFR upon EGF stimulation for 1 h, n = 3. b Pull-down analysis examining the interaction of PGK1 to different forms of EGFR fused to GST, n = 3. c Pull-down analysis examining the interaction of different PGK1 forms with the juxta-membrane region EGFR fused to GST, n = 3. d Assay for endocytic transport of EGFR to the lysosome, examining the effect of mutating the dileucine sorting signal in EGFR. Colocalization of EGFR-myc with LAMP1 upon EGF treatment for indicated time point was performed, n = 10 cells examined over 3 independent experiments. Representative images are shown above with EGFR-myc in red and LAMP1 in green, bar = 10 μm. Quantitation is shown below for the 40-min time point, P = 1.13 × 10⁻¹⁵. e Time-course analysis of EGFR degradation upon EGF treatment for indicated time point in HeLa cells, examining the effect of mutating the dileucine sorting signal in EGFR. Cell lysates were immunoblotted for proteins indicated, n = 3. f Quantitation of EGFR level at the 0, 60,120, and 180-min time point for the analysis above. EGFR level was normalized to tubulin level, P = 0.0033, 0.004, 3.41 × 10⁻⁴,and 0.0007, respectively. g Quantitation of ERK activation at the 0, 60,120, and 180-min time point for the analysis above. Phosphorylated ERK level was normalized to total ERK level, P = 0.00014, 0.0741, 1.26 × 10⁻⁴, and 0.023, respectively. h Pull-down analysis examining the interaction of PGK1 to various forms of EGFR juxta-membrane regions fused to GST as indicated. Purified components were used to examine direct interaction, n = 3. i Pull-down analysis examining the effect of titrating increased level of a peptide containing either the dileucine sorting signal of EGFR or this signal mutated on the direct interaction with PGK1, n = 3. j Co-precipitation analysis examining of the effect of mutating the dileucine sorting signal in EGFR on its association with PGK1. HeLa cells were stimulated with EGF for 1 h, n = 3.

Fig. 4. S203 phosphorylation of PGK1 promotes EGFR transport to the lysosome.

Quantitative results are shown as mean +/- s.e.m.; *p < 0.05, unpaired two-sided Student’s t test. a Fractionation of cells into membranes containing endosomes (P) and cytosol (S) upon EGF stimulation for 30 min, with or without the ERK inhibitor U0126, n = 3. Quantitation is shown below, P = 0.00014 (control versus EGF) and 0.00041 (EGF versus U0126). b Fractionation of cells into membranes containing endosomes (P) and cytosol (S) upon EGF stimulation for 30 min, n = 3. c Immunoblotting of HeLa cell lysates examining the efficacy of siRNA against PGK1 and the level of rescue with phospho-mutant forms of PGK1, n = 3. d Fractionation of cells into membranes containing endosomes (P) and cytosol (S) upon EGF stimulation for 30 min, n = 3. e Co-precipitation analysis examining the association of wild-type and phospho-mutants of PGK1 with EGFR upon EGF stimulation for 30 min, n = 3. f PLA analysis examining the association of wild-type and phospho-mutants of PGK1 with EGFR upon EGF stimulation for 30 min, n = 10 cells examined over 3 independent experiments. Quantitation is shown for a representative experiment, P = 1.1 × 10⁻⁷ (wild-type versus S203A) P = 2.6 × 10⁻⁷(S203A versus S203D). g Colocalization of EGF (red) with LAMP1 (green) upon EGF treatment for indicated time point was performed, n = 20 cells examined over 3 independent experiments, bar = 10 μm. Statistics is shown on right for the 40-min time point, P = 8.265 × 10⁻⁷ (wild-type versus S203A) 2.472 × 10⁻¹¹(S203A versus S203D). h Time-course analysis of EGFR degradation upon EGF treatment for indicated time point, examining the effect of expressing PGK1 wild-type and phospho-mutants, n = 3. i Quantitation of EGFR level at the 60, 120, and 180-min time point for the analysis above. EGFR level was normalized to tubulin level, P = 0.0192, 0.0485, and 0.0352, respectively. j Quantitation of ERK activation at the 60, 120, and 180-min time point for the analysis above. Phosphorylated ERK level was normalized to total ERK level, P = 0.5193, 0.0025, and 0.0364, respectively. k Pull-down analysis examining the interaction of PGK1 phospho-mutants with the juxta-membrane region of EGFR (wild-type or dileucine mutated) fused to GST, n = 3.

Fig. 5. Hrs is required for PGK1 to interact with EGFR.

Quantitative results are shown as mean +/- s.e.m.; *p < 0.05, NS p > 0.05, unpaired two-sided Student’s t test. a Co-precipitation analysis examining the effect of siRNA against PGK1 on the association of Hrs with EGFR upon EGF stimulation for 15 min, n = 3. b Co-precipitation analysis examining the effect of siRNA against PGK1 on the association of HD-PTP with EGFR upon EGF stimulation for 15 min, n = 3. c Co-precipitation analysis examining the effect of siRNA against PGK1 on the association of CHMP4B with EGFR upon EGF stimulation for 15 min, n = 3. d Co-precipitation analysis examining the effect of siRNA against Hrs on the association of PGK1 with EGFR upon EGF stimulation for 30 min, n = 3. e PLA analysis examining the effect of targeting against different proteins through siRNA treatment on the association of PGK1 with EGFR in HeLa cells upon EGF stimulation for 30 min, n = 15 cells examined over 3 independent experiments. Puncta tracks the association of PGK1 with EGFR by using primary antibodies directed against endogenous PGK1 and EGFR, bar = 10 μm. Quantitation is shown on right for a representative experiment, P = 5.93 × 10⁻¹¹ for si-Hrs, P = 0.215 for si-TSG101, P = 0.0765 for si-HD-PTP, and P = 0.423 for si-CHMP4B.

Fig. 6. Hrs promotes PGK1 recruitment to endosomal membrane.

Quantitative results are shown as mean +/- s.e.m.; *p < 0.05, unpaired two-sided Student’s t test. a Co-precipitation analysis examining PGK1 association with Hrs upon EGF stimulation at time points indicated, n = 3. Quantitation from three experiments is shown on right, P = 0.00222 for 15 min and 4.22 × 10⁻⁴ for 30 min. b Colocalization of PGK1 (red) and Hrs (green) upon EGF stimulation for 30 min, n = 15 cells examined over 3 independent experiments, bar = 10 μm. Quantitation is shown on right for a representative experiment, comparing control versus EGF stimulation, P = 2.387 × 10⁻¹¹. c Co-precipitation analysis examining PGK1 wild-type and phospho-mutants associating with Hrs upon EGF stimulation for 30 min, n = 3. d PLA analysis examining PGK1 wild-type and phospho-mutants associating with Hrs upon EGF stimulation for 30 min, n = 10 cells examined over 3 independent experiments. Puncta tracks the association of PGK1 with Hrs by using primary antibodies directed against myc epitope of transfected PGK1-myc and endogenous Hrs, bar = 10 μm. Quantitation is shown on left for a representative experiment, P = 5.7 × 10⁻⁸. e Colocalization of PGK1 (green) wild-type and phospho-mutants with Hrs (red) upon EGF stimulation for 30 min, n = 15 cells examined over 2 independent experiments, bar = 10 μm. Quantitation is shown on right, P = 2.816 × 10⁻⁸ for S203A and S203D. f Reconstitution of PGK1 recruitment to endosomal membrane, examining the effect of depleting Hrs from the membrane. The S203D mutant of PGK1 was incubated with endosomal membrane that had Hrs depleted followed by immunoblotting for PGK1 on membrane. Immunoblotting for EGFR and LAMP2 assessed the level of membrane examined, n = 3.

Fig. 7. PIP5K1A promotes PGK1 recruitment to endosomal membranes.

Quantitative results are shown as mean +/- s.e.m.; *p < 0.05, unpaired two-sided Student’s t test. a Co-precipitation analysis examining the effect of various siRNA treatments as indicated on PGK1 association with EGFR upon EGF stimulation for 1 h, n = 3. b Co-precipitation analysis examining the effect of EGF stimulation for 1 h on the association of PGK1 with EGFR and PIP5K1A, n = 3. c Colocalization of EGF (red) with LAMP1 (green) upon EGF stimulation for indicated time point was performed, n = 10 cells examined over 3 independent experiments, bar = 10 μm. Quantitation is shown on right for a representative experiment, with statistical analysis performed for the 60-min time point, P = 1.29 × 10⁻⁶. d Time-course analysis of EGFR degradation upon EGF stimulation for indicated time point, examining the effect of siRNA against PIP5K1A, n = 3. e Quantitation of EGFR level at the 0, 60, 120, and 180-min time point for the analysis above. EGFR level was normalized to tubulin level, *P = 0.502, 0.0145, 0.0085, and 0.0073, respectively. f Quantitation of ERK activation at the 0, 60, 120, and 180-min time point for the analysis above. Phosphorylated ERK level was normalized to total ERK level, *P = 0.113, 0.0139, 0.0121, and 0.0561, respectively. g Reconstitution of PGK1 recruitment to endosomal membrane, examining the effect of the S203D mutation on this recruitment. Recombinant forms of PGK1 were incubated with endosomal membrane followed by immunoblotting for PGK1 on membrane, n = 3. h Reconstitution of PGK1 recruitment to endosomal membrane, examining the effect of depleting PIP5K1A from membrane. The S203D mutant of PGK1 was incubated with endosomal membrane followed by immunoblotting for PGK1 on membrane, n = 3. i Reconstitution of PGK1 recruitment to endosomal membrane, examining the effect of delivering PI(4,5)P₂ versus PI(3,4)P₂ to endosomal membrane that was depleted of PIP5K1A. The S203D mutant of PGK1 was incubated with endosomal membrane followed by immunoblotting for PGK1 on membrane, n = 3. Quantitation of PGK1 membrane recruitment by normalizing to the level of EGFR from three experiments is shown. *P = 0.04003 and 0.0097 for 20 and 50 nM group comparisons, respectively.