Molecular basis for the sorting of the SNARE VAMP7 into endocytic clathrin-coated vesicles by the ArfGAP Hrb

Abstract

SNAREs provide the specificity and energy for the fusion of vesicles with their target membrane, but how they are sorted into the appropriate vesicles on post-Golgi trafficking pathways is largely unknown. We demonstrate that the clathrin-mediated endocytosis of the SNARE VAMP7 is directly mediated by Hrb, a clathrin adaptor and ArfGAP. Hrb wraps 20 residues of its unstructured C-terminal tail around the folded VAMP7 longin domain, demonstrating that unstructured regions of clathrin adaptors can select cargo. Disrupting this interaction by mutation of the VAMP7 longin domain or depletion of Hrb causes VAMP7 to accumulate on the cell's surface. However, the SNARE helix of VAMP7 binds back onto its longin domain, outcompeting Hrb for binding to the same groove and suggesting that Hrb-mediated endocytosis of VAMP7 occurs only when VAMP7 is incorporated into a cis-SNARE complex. These results elucidate the mechanism of retrieval of a postfusion SNARE complex in clathrin-coated vesicles.

Figure 1

Identifying Hrb as a Binding Partner for VAMP7 (A) Schematic representations of VAMP7 and Hrb. VAMP7 contains a putative longin domain and a SNARE helix preceding its transmembrane helix (TM). Hrb has a predicted ArfGAP domain at its N terminus with the remainder of the protein predicted to possess very little secondary structure. This region contains a predicted clathrin-binding motif, three FXXFXXF motifs predicted to bind to AP2 appendage domains, and four NPF motifs that are able to bind EH domain-containing proteins. The full sequence of Hrb is shown in Figure S1A. (B) Yeast two-hybrid analysis indicates that VAMP7 longin domain (LD) but not Sec22b or Ykt6 longin domains (LDs) interacts with full-length Hrb, as indicated by growth of diploid strains containing the indicated plasmids on medium lacking adenine. The Hrb-DNA binding-domain construct autoactivates the reporter. Ctrl, empty binding- and activation-domain plasmids. (C) GST pull-down experiments indicate that VAMP7 longin domain but not Sec22b longin domain or GST alone binds full-length Hrb from A431 cell cytosol shown by immunoblotting with an anti-Hrb antibody. (D) GST-VAMP7 longin domain but not GST pulls down His₆-tagged Hrb as indicated by immunoblotting with an anti-His₆ antibody. (E) The binding of various C-terminally His₆-tagged Hrb constructs to GST-VAMP7 longin domain demonstrates that the Hrb binding site for VAMP7 longin domain is located between residues 136 and 176.

Figure 2

In Vivo Localization of VAMP7 and Hrb (A) NRK cells were double labeled with anti-lgp120 (a), anti-Hrb (e), or anti-VAMP7 (b and f) followed by fluorescently labeled secondary antibodies. Panels c and g are merged images of panels a and b and panels e and f, respectively. Panels d and h are higher magnifications of the regions boxed in panels c and g. (B) NRK cells double labeled with goat polyclonal anti-Hrb (panels a, e, i, and m) and mouse anti-lgp120 (b), mouse anti-clathrin (f), rabbit anti-MPR (j), and mouse anti-α-adaptin (n) followed by fluorescently labeled secondary antibodies indicate that Hrb is mainly found in plasma membrane-associated CCPs/CCVs. Panels c, g, k, and o are merged images and panels d, h, l, and p are higher magnifications of the regions boxed in panels c, g, k, and o. The higher-magnification images in panels h and p were analyzed using ImageJ software with a Pearson-Spearman correlation colocalization plug-in to perform quantitative statistical colocalization on the two-color confocal images as described by French et al. (2008). In both images, strong positive correlations for colocalization were indicated by r_P values of 0.51 (h) and 0.62 (p), and r_S values of 0.49 (h) and 0.57 (p). All images are confocal maximum-intensity z projections. The scale bars represent 10 μm.

Figure 3

Structure of the VAMP7 Longin Domain/Hrb_136–176 Complex (A) Schematic representation of the construct used to produce crystallizable VAMP7 longin/Hrb_136–176. The GST tag, which is proteolytically cleaved off, and regions of the protein not visible in the electron density are colored gray. Those parts visible in the structure are colored as in (B) and (C), with VAMP7 in purple and Hrb in gold. (B–E) Structure of Hrb_136–176 VAMP7 longin domain in ribbon and surface representations with VAMP7 in purple and Hrb in gold with (B) and (C), and (D) and (E), in the same orientations. (F) VAMP7 longin domain residues involved in Hrb binding are shown in purple on a surface representation with a Hrb fragment superimposed. Side chains that are important in binding to VAMP7 longin domain are indicated. (G) The same view as in (F) with Hrb removed. The side chains of VAMP7 involved in Hrb binding can be seen through the semitransparent VAMP7 surface.

Figure 4

Mutation of Key Residues in the VAMP7 Longin Domain/Hrb Interface Inhibits Their Binding (A and B) Details of side chains involved in the interaction between VAMP7 longin domain and Hrb, with carbon atoms and ribbons colored as in Figure 3. Mutated residues studied in (C) and (D) are boxed. (C and D) The inhibitory effects of indicated mutations on VAMP7/Hrb binding as shown by ITC. The K_D for the interaction increases from approximately 10 μM to unmeasurable levels > 300 μM for the four mutations tested (see also Figure S2).

Figure 5

Effect of Disrupting the VAMP7 Longin Domain/Hrb Interaction In Vivo (A) NRK cells stably expressing VAMP7-HA transfected with siRNA against Hrb (oligos 1 and 2) showed an increase in VAMP7-HA at the cell surface that was not observed following mock transfection (see also Figure S3A). Cells were fixed with paraformaldehyde, and an aliquot was permeabilized with Triton X-100. Cells were then incubated with [¹²⁵I]mouse monoclonal anti-HA. For each cell preparation the amount of VAMP7-HA on the cell surface, as a percentage of the total amount of VAMP7-HA expressed, was calculated (histogram, n = 3 ± SEM; ^∗p < 0.05 versus mock). A representative western blot shows the levels of Hrb in mock and Hrb knockdown (KD) cells. The VAMP8 blot is a control for protein loading. (B) NRK cells stably transfected with ΔpMEP4 containing either wild-type (WT) (a and b) or L43S/Y45S mutant (c and d) VAMP7-HA were treated with CdCl₂ to induce expression. They were then either fixed with paraformaldehyde (for surface staining of HA [a and c]) or fixed and then permeabilized with Triton X-100 (for total staining of HA [b and d]). For an expanded figure, see Figure S6A. Cells stably expressing L43S/Y45S mutant VAMP7-HA have increased VAMP7-HA at the cell surface. The total amount of WT or L43S/Y45S mutant VAMP7-HA expressed in the transfected cells was the same (Figure S6B), but the amount of HA tag was reduced when the WT construct was exposed to the proteolytic environment of lysosomes. (C) NRK cells transiently expressing chimeric ΔpMEP4 constructs encoding CD8 (a) or CD8 followed by a flexible linker consisting of amino acids GSGGSG (as depicted in [b]) connecting it to either the wild-type (c) or L43S/Y45S mutant (d) longin domains. For an expanded figure, see Figure S7. The amount of CD8 on the surface and in internal structures was visualized by immunofluorescence using mouse anti-CD8. Images are confocal maximum-intensity z projections. The scale bars represent 20 μm.

Figure 6

VAMP Analogous Binding Sites Exist on VAMP7 and Sec22b Longin Domains (A) Superposition of Cα traces of the longin domains of VAMP7 (purple) and Sec22b (cyan) with the Hrb peptide (gold) and Sec22b SNARE peptide (dark blue) can be seen to have the same conformations and superimpose well. (B and C) Enlargements of the binding sites showing molecular detail for the hydrophobic side-chain interactions that mediate the binding of VAMP7 LD/Hrb peptide and Sec22b LD/Sec22b SNARE peptide colored as in (A). (D) Structure-based alignment of the longin domains of Sec22b and VAMP7 and their respective binding partners the intramolecular SNARE helix of Sec22b and HrbVAMP. Key residues on the longin domains involved in their interactions are boxed in green and varied numbers of intervening residues are indicated by dots. Key residues on Hrb and the Sec22b SNARE helix along with those proposed to be important for VAMP7 SNARE helix binding are colored red. The arrows indicate the short β strand in the two determined structures.

Figure 7

Hrb Competes for the SNARE Helix Binding Site on VAMP7 Longin Domain (A) Structure-based alignment as in Figure 6D with the putative “bind-back” sequence of VAMP7 also included. Residues mutated to abolish the proposed interaction of VAMP7 longin domain with its SNARE helix are indicated by an asterisk. (B) Western blotting with anti-His₆ antibody of GST pull-down experiment for the interaction of various VAMP7 constructs with Hrb_1–176 His₆. The GST-VAMP7 longin domain (LD) and the GST-VAMP7 whole cytoplasmic domain (I139S/M140S/I144S) triple mutant (GST-VAMP7FL[SSS]) bound Hrb, but the wild-type construct (GST-VAMP7FL) did not. (C and D) Coomassie blue staining (C) and western blotting with anti-His₆ antibody (D) of GST pull-down experiment for the interaction of VAMP7 longin domain, and the whole cytoplasmic domain of VAMP7 on its own and in an endocytic SNARE complex, with His₆syntaxin4 (201–266) and SNAP23 (1–210). As in (B), the longin domain of VAMP7 bound Hrb_1–176 His₆, whereas the full-length cytoplasmic domain of VAMP7 did not. Formation of a VAMP7/SNARE complex restored Hrb_1–176 His₆ binding to the levels displayed by the longin domain alone.