Lipid binding promotes oligomerization and focal adhesion activity of vinculin

Abstract

Adherens junctions (AJs) and focal adhesion (FA) complexes are necessary for cell migration and morphogenesis, and for the development, growth, and survival of all metazoans. Vinculin is an essential regulator of both AJs and FAs, where it provides links to the actin cytoskeleton. Phosphatidylinositol 4,5-bisphosphate (PIP2) affects the functions of many targets, including vinculin. Here we report the crystal structure of vinculin in complex with PIP2, which revealed that PIP2 binding alters vinculin structure to direct higher-order oligomerization and suggests that PIP2 and F-actin binding to vinculin are mutually permissive. Forced expression of PIP2-binding-deficient mutants of vinculin in vinculin-null mouse embryonic fibroblasts revealed that PIP2 binding is necessary for maintaining optimal FAs, for organization of actin stress fibers, and for cell migration and spreading. Finally, photobleaching experiments indicated that PIP2 binding is required for the control of vinculin dynamics and turnover in FAs. Thus, through oligomerization, PIP2 directs a transient vinculin sequestration at FAs that is necessary for proper FA function.

Figure 1.

PIP₂ directs vinculin oligomerization by unfurling the N terminus of Vt. (A) The R1060A mutation used for crystallization does not affect PIP₂ (oxygen, phosphor, and carbon atoms are shown as red, orange, and white sticks, respectively) binding. Vt α-helices are colored sequentially across the color spectrum from red (H1) to blue (H5). PIP₂ and residues 1,060 and 1,061 are depicted, and residues 1,060 and 1,061 and Vt α-helices H1–H5 are labeled. (B) PIP₂ directed dimerization of the vinculin five-helix bundle tail domain (α-helices are colored sequentially across the color spectrum from red (H1) to blue (H5), monomers are labeled as I and II) by binding to the α1-α2 loops and the C terminus of Vt. (C–E) Close-up view of the PIP₂-directed oligomerization of Vt (α-helices and bonds are colored sequentially across the color spectrum from red to blue). Residues Lys-944 and Arg-945 are contributed from a third (III) monomer within the trimer. (D and E) View of the second (D; shown in blue) and third (E; shown in green) PIP₂ in the asymmetric unit in the same orientation as in C, which indicates the specificity. (F) The PIP₂-directed Vt trimer. Vt α-helices are colored sequentially across the color spectrum from red to blue, PIP₂ is shown as spheres, and the three Vt molecules are labeled I–III. (G) With the exception of the loop connecting the last α-helix H5 to its C-terminal extended coil, unbound Vt (shown in cyan) superposes onto PIP₂-bound Vt with a root-mean-square deviation (RMSD) of <0.5 Å for ∼970 atoms (depending on the monomer in the asymmetric unit). However, the PIP₂ binding site coincides with the N-terminal extended coil (in particular residues 881–886, indicated by a red “N”), which is unfurled in the PIP₂-bound structure.

Figure 2.

Pull-down assay to analyze the lipid-binding capacity of vinculin. (A) The interaction of wild-type and mutant Vt with VH was analyzed by electrophoretic mobility shift assays on a native gel, incubating the respective Vt proteins with VH. All proteins formed a Vt–VH complex. Data shown are representative of three independent experiments. (B) Unilamellar vesicles of the phospholipids PC and 15% PIP₂ were incubated with wild-type and mutant Vt (top: K915Q, K924Q, K944Q/R945Q, K970Q, R1057Q, and K1061Q; bottom: K915Q/K1061Q, K944Q/R945Q/K1061Q, and K915Q/K944Q/R945Q) and sedimented. Supernatant (S) and pellet (P) fractions were detected on Coomassie-stained SDS-PAGE gel. Inactive and talin-VBS3 (residues 1,945–1,970) activated full-length (FL) vinculin were run as the nonbinding and binding controls, respectively. Densitometric values averaged over three experiments together with their standard deviations are shown above or below the respective experiment (gray bars, supernatant; white bars, pellet) and are as follows. (top) Wild type, 49.9% unbound in the supernatant with a standard deviation of 0.19; K915Q, 71.2%, 2.5; K924Q, 64.5%, 3.03; K944Q/R945Q, 89.3%, 1.35; K970Q, 48.7%, 2.5; R1057Q, 60%, 2.1; R1061Q, 96%, 0.72; (bottom left) wild type, 50%, 0.4; K915Q/K1061Q, 95.9%, 0.9; K944Q/R945Q/K1061Q, 97%, 0.9; K915Q/K944Q/R945Q, 93.7%, 1.1; (bottom right) full-length vinculin with VBS, 84.6%, 1.3; full-length vinculin alone, 99%, 0.1. The white line in one panel indicates the removal of unrelated lanes for presentation purposes. The corresponding full-length gels are shown in Fig. S1 B. Error bars indicate SD. (C) Previously identified residues on α-helix H3 reside near the PIP₂-induced Vt–Vt interface. Shown is a close-up view of the PIP₂-directed Vt (α-helices are colored sequentially across the color spectrum from red [H1] to blue [H5]) interface between subunits II and III. Residues Lys-952, Lys-956, and Arg-963, which were previously thought to bind to PIP₂, reside near or at the Vt–Vt interface, which might explain the decreased binding of the potentially oligomer-defective mutant (K952Q, K956Q, and R963Q) Vt compared with wild-type Vt.

Figure 3.

PIP₂ binding in the context of full-length vinculin. (A) The PIP₂ binding site is distinct from the head–tail interface. Superposition of Vt–PIP₂ onto full-length vinculin highlights the steric hindrance of the Vt N-terminal extended coil in the full-length vinculin–PIP₂ model (Vt α-helices are colored sequentially across the color spectrum from red [H1] to blue [H5]; Vh1, residues 1–258 are shown in dark gray; the remainder of VH, residues 259–843, is shown in light gray; PIP₂ is shown as spheres). The disordered regions (residues 844–876 and 1,063–1,066) are indicated. PIP₂ binding orders the C terminus, and the Vt–PIP₂ structure is the only vinculin structure that has interpretable electron density for the last residues. (B) PIP₂-directed oligomerization of full-length vinculin. Once vinculin is activated and its head and tail domain severed, PIP₂ directs vinculin oligomerization by binding to three Vt molecules. For clarity, only one PIP₂ molecule, interacting with Lys-944 and Arg-945 of vinculin (III), is shown, whereas the PIP₂ molecules bound to Lys-944 and Arg-945 of vinculin (I) and (II), which form a higher-order oligomer, are not shown. Vt α-helices are colored sequentially across the colorspectrum from red (H1) to blue (H5), VH are shown in gray, and PIP₂ is represented with spheres.

Figure 4.

The lipid-binding–deficient vinculin mutants bind and bundle F-actin. (A) Wild-type and mutant Vt (top: K915Q, K944Q/R945Q “944,” K1061Q, K915Q/K944Q/R945Q “915 944,” K915Q/K1061Q; K944Q/R945Q/K1061Q “944 1061”; bottom: K924Q, K970Q, and R1057Q) were cosedimented with F-actin and analyzed on Coomassie-stained SDS-PAGE. Full-length (FL) vinculin was run as the nonbinding control. S, supernatant; P, pellet. The white lines indicate the removal of unrelated lanes for presentation purposes. Data shown are representative of three independent experiments. The corresponding full-length gels are shown in Fig. S2 (A–C). (B) Confocal images of TRITC-phalloidin–labeled F-actin in the absence or presence of wild-type or mutant (K1061Q or K915Q/K944Q/R945Q/K1061Q) Vt domain. Data shown are representative of three independent experiments. Larger micrographs corresponding to these panels are shown in Fig. S3. Bars, 20 µm. (C) Model based on the superposition (not depicted) of the cryo-EM Vt–F-actin structure (Thompson et al., 2014; F-actin subunits are shown in gray or black) onto the PIP₂-bound Vt dimer (α-helices H1–H5 are colored sequentially across the color spectrum from red [H1] to blue [H5]; PIP₂ are represented as spheres), which suggests that vinculin has distinct F-actin and lipid binding sites. Residues Ile-997 and Val-1001 residing on α-helices H3 (green) that exhibit a significant decreased affinity for F-actin (Thompson et al., 2014) are also shown as spheres.

Figure 5.

Raver1 and PIP₂ binding to vinculin is mutually exclusive. (A) Superposition of the Vt–raver1 structure onto the Vt–PIP₂ structure. Vt α-helices are colored sequentially across the color spectrum from red (H1) to blue (H5), monomers are labeled I–III, and raver1 is shown in gray. Superposition of lipid-bound Vt (I) onto the raver1-bound Vt shows overlapping binding of Vt (III) and raver1. (B) Close-up view of some interface residues. In its raver1-bound state, vinculin residue Arg-945 engages in interactions with raver1 residues Glu-120 and Arg-117, whereas in its lipid-bound state, Arg-945 points toward PIP₂. (C) Size exclusion chromatogram (SEC) of Vt–raver1 (blue line; peak #4 eluting at 14.56 ml; apparent mol wt = 54.6 kD; polypeptide chain mol wt = 51.2 kD), Vt-K944Q/R945Q/raver1 (black line; peak #5, 15 ml; apparent mol wt = 45 kD; polypeptide chain mol wt = 51.2 kD), raver1 (dotted line; peak #3, 15.77 ml; apparent mol wt = 32.19 kD; polypeptide chain mol wt = 31.5 kD), and Vt (gray line; peak #1, 17.2 ml; apparent mol wt =17.22 kD; polypeptide chain mol wt = 19.7 kD). (D) SDS-PAGE of peak fractions from the SEC shown in the previous panel. The lane numbers correspond to the peak numbering in the previous panel. (E) Lipid cosedimentation assay of the Vt–raver1 complex. The Vt–raver1 complex does not cosediment with PIP₂ and remains in the soluble (S) fraction. P, pellet.

Figure 6.

PIP_2-directed oligomerization of vinculin contributes to the stabilization of FAs. Vinculin^−/− MEFs engineered to express wild-type GFP-vinculin or mutant GFP-vinculin fusions (K944Q/R945Q and R1061Q) were analyzed by confocal laser-scanning microscopy. MEFs were grown on gelatin-coated coverslips, fixed after 24 h, and stained with TRITC-phalloidin to visualize the actin cytoskeleton. Representative images, which define the localization of GFP-vinculin (green) at FAs decorating F-actin (red), along with the merged channels, are shown. Also shown are nuclei stained with DAPI. Data shown are representative of three independent experiments. Bars, 100 µm.

Figure 7.

MEFs expressing the PIP₂-binding–deficient mutant of vinculin are impaired in wound closure. Vinculin-null cells (left) that were engineered to express comparable levels of wild-type vinculin (center left) or K944Q/R945Q or K1061Q mutant vinculin (center right and right) were plated onto fibronectin-coated 12-well plates. Monolayers were then wounded, and the extent of cell migration into the wound was recorded at 0, 4, 8, and 12 h. Vinculin-null cells migrate in a rapid yet chaotic fashion, whereas wound closure in these cells expressing wild-type vinculin is an ordered process. Vinculin-null cells expressing K944Q/R945Q or K1061Q mutant vinculin had reduced migration and wound closure versus cells expressing wild-type vinculin. Data shown are representative of three independent experiments. Bars, 20 µm.

Figure 8.

FRAP analyses of wild-type and mutant GFP-vinculin in FAs. (A) Representative images of FRAP recovery. The arrows indicate the bleached FA. (B) FRAP recovery curves for the mean of at least eight independent measurements for each cell line are displayed. The t₅₀ values for wild-type (19 s) and the K1061Q mutant (7 s) were determined by fitting the curves to a single exponential. A t₅₀ value for the K944Q/R945Q mutant could not be determined due to the absence of significant recovery. The immobile fraction is represented by the difference between the normalized plateau value and the prebleach fluorescence, and was estimated to be <5% for the wild-type cells, ∼45% for the K1061Q mutant, and >90% for the K944Q/R945Q mutant. Error bars indicate SD.