Cancer Res 2006; 66: (7).April, pp 3603 - 3610

A Novel Endocytic Mechanism of Epidermal Growth Factor Receptor Sequestration and Internalization

James D. Orth, Eugene W. Krueger, Shaun G. Weller, and Mark A. McNiven

Cells form transient, circular dorsal ruffles or ‘‘waves’’ in response to stimulation of receptor tyrosine kinases, including epidermal growth factor receptor (EGFR) or platelet-derived growth factor receptor. These dynamic structures progress inward on the dorsal surface and disappear, occurring concomitantly with a marked reorganization of F-actin. The cellular function of these structures is largely unknown. Here we show that EGF-induced waves selectively sequester and internalize f50% of ligand-bound EGFR from the cell surface. This process requires receptor phosphorylation, active phosphatidylinositol 3-kinase, and dynamin 2, although clathrincoated pits or caveolae are not required. Epithelial and fibroblast cells stimulated with EGF sequestered EGFR rapidly into waves that subsequently generated numerous receptorpositive tubular-vesicular structures. Electron microscopy confirmed that waves formed along the dorsal membrane surface and extended numerous tubules into the cytoplasm. These findings characterize a structure that selectively sequesters large numbers of activated EGFR for their subsequent internalization, independent of traditional endocytic mechanisms such as clathrin pits or caveolae.

enlarge

Figure 1. EGFR is sequestered into forming waves and subsequently into dynamic tubular-vesicular structures that form at waves. Live PANC-1 cells expressing EGFR-GFP were stimulated with 30 ng/mL EGF and imaged. A to D, individual frames showed that EGFR-GFP was sequestered within newly forming waves in <5 minutes post-stimulation (B, arrows). By 15 minutes, the waves constrict and formed dynamic tubular-vesicular structures containing EGFR-GFP (C and D, arrows, and video 1). E to I, high magnification of a structure formed from a dorsal wave observed f15 minutes post-EGF treatment. Close examination revealed that 2- to 5-Am-long EGFR-GFP tubules (100-200 nm in diameter) extended toward the cell center (F-I, arrows, image inverted for clarity, video 2). F to I, nascent vesicles generated from the cluster (arrows, arrow with heavy border) trafficked toward the cell center. Associated vesicles were also found to coalesce as they moved toward the cell center (arrowhead).

enlarge

Figure 4. Waves represent a distended dorsal plasma membrane that is transformed into a vesicle-generating tubular network. Scanning electron microscopy (SEM), vertical section (X-Z) transmission electron microscopy (TEM), and serial section transmission electron microscopy identified the form of waves on the dorsal plasma membrane. A, scanning electron microscopy revealed that waves consist of a ridge of uniform width that is composed of numerous bumps (arrows). B, a vertical section through a wave identified numerous membrane tubules that extend within the ruffle and into the cytoplasm (arrows). C, transmission electron microscopy of an en face section through a wave region resolved numerous tubular invaginations along the plasma membrane. CV, four serial sections through the same region in (C) were pseudo-colored and aligned to provide greater visual depth. Large numbers of tubules are more easily resolved as an extension of the plasma membrane. Bar, 5 um (A), 1 um (B, inset, to CV).

enlarge

Figure 5. Waves assemble and function independently of clathrin to internalize EGFR. PANC-1 and NR6 cells were stimulated with EGF for 5 minutes and stained for clathrin or Eps15. A, stimulated PANC-1 cells labeled for Dyn2 (green) and clathrin (red; X22 monoclonal antibody) show no accumulation of clathrin within waves (arrows; lack of yellow overlap). Clathrin-GFP also did not label waves (not shown). B, colabeling of cortactin (green) and the clathrin pit-associated protein Eps15 (red) in NR6 showed no enrichment of Eps15 at waves (arrows; lack of yellow overlap). C to DV, rhodamine-EGF is sequestered (C, arrows) and internalized significantly (D, arrow, inset ) in cells devoid of CHC (CVand DV, asterisk ). E and EV, cells expressing the CME inhibitor AP180-C (EV, asterisk ) can also internalize EGFR significantly (E, arrow). F, inset, immunoblot analysis of total cell lysate from either mock treated cells or cells treated with siRNA to reduce CHC protein. Values adjacent to bands represent percent reduction of CHC levels as assessed by quantitative densitometry normalized to mock. CHC levels were reduced on average by 95% in HeLa cells (two experiments done) and 71% in NR6 (three experiments done). g-Tubulin loading control is for NR6; HeLa loading control is not shown. Mock represents transfection reagent alone. AP180-C, Dyn2K44A, and CHC reduction (z80%) potently blocked CME of transferrin receptor in HeLa (H) and NR6 (N) and 100 nmol/L wortmannin (wort) had no effect. G, GTPase-deficient Dyn2K44A and wortmannin inhibited wave formation by f50% and f70%, respectively, whereas wave formation was modestly increased in AP180-C-expressing or CHC knockdown cells. H, waves internalize significant levels of EGFR. AP180-Cblocked EGF internalization in HeLa (that do not make waves) byf80% but by onlyf20% in NR6. CHC knockdown in HeLa and NR6 yielded the same result as AP180-C, with an f80% block in EGFR internalization in HeLa and f20% in NR6. Wortmannin treatment had no effect on EGFR internalization in HeLa but did block f50% in NR6. The block in EGF internalization in NR6 was additive when cells were treated with both siCHC and wortmannin (f65%). Bar, 10 Am (A and C-E V), 5 Am (B). F to H, control is nontransfected or mock treated (DMSO for wortmannin). Columns, mean (n = 3, with z100 cells in each condition); bars, SE.