Structural determinants of the integrin transmembrane domain required for bidirectional signal transmission across the cell membrane

Abstract

Studying the tight activity regulation of platelet-specific integrin α_IIbβ₃ is foundational and paramount to our understanding of integrin structure and activation. α_IIbβ₃ is essential for the aggregation and adhesion function of platelets in hemostasis and thrombosis. Structural and mutagenesis studies have previously revealed the critical role of α_IIbβ₃ transmembrane (TM) association in maintaining the inactive state. Gain-of-function TM mutations were identified and shown to destabilize the TM association leading to integrin activation. Studies using isolated TM peptides have suggested an altered membrane embedding of the β₃ TM α-helix coupled with α_IIbβ₃ activation. However, controversies remain as to whether and how the TM α-helices change their topologies in the context of full-length integrin in native cell membrane. In this study, we utilized proline scanning mutagenesis and cysteine scanning accessibility assays to analyze the structure and function correlation of the α_IIbβ₃ TM domain. Our identification of loss-of-function proline mutations in the TM domain suggests the requirement of a continuous TM α-helical structure in transmitting activation signals bidirectionally across the cell membrane, characterized by the inside-out activation for ligand binding and the outside-in signaling for cell spreading. Similar results were found for α_Lβ₂ and α₅β₁ TM domains, suggesting a generalizable mechanism. We also detected a topology change of β₃ TM α-helix within the cell membrane, but only under conditions of cell adhesion and the absence of α_IIb association. Our data demonstrate the importance of studying the structure and function of the integrin TM domain in the native cell membrane.

Keywords: cell adhesion; integrin; platelet; talin; transmembrane domain.

Figure 1

Proline scanning of α_IIbβ₃TM domains.A, sequence alignment of human integrin transmembrane domains. The predicted TM C-terminal boundary is marked with a red dashed line. The conserved small amino acids are highlighted in yellow. Shown in red are the residues that inhibit integrin inside-out activation when mutated to proline. B, backbone structure of α_IIbβ₃ TM domain. Backbone hydrogen bonds are shown as dashed lines in yellow. Selected residues are marked as spheres of Cα atoms. C and D, ligand mimetic PAC-1 binding to α_IIbβ₃ integrin with indicated proline substitutions. HEK293FT cells were transfected with indicated α_IIbβ₃ constructs. The ligand-mimetic mAb PAC-1 binding was performed in the buffer containing 1 mM Ca²⁺/Mg²⁺ and detected by flow cytometry. The level of PAC-1 binding was normalized to α_IIbβ₃ surface expression detected by mAb AP3 and shown as mean ± SD (n = 3). Unpaired two-tailed t test was performed between the control group without proline mutation and the group with proline mutation. Only p values less than 0.05 are shown. E, structural illustration of α_IIbβ₃ TM domain within the cell membrane in the absence and presence of selected proline mutations. The proline mutations were introduced in silico to the TM structure using PyMOL. The proline-induced broken of a rigid α-helical structure was indicated. The interfacial residues are shown as sticks or Cα spheres. MP, membrane proximal.

Figure 2

Proline scanning of α_IIbβ₃TM domains identified proline mutations that hindered activation.A and B, PAC-1 binding of β₃ TM proline mutations on the background of the α_IIb cytoplasmic mutations, α_IIb-R995D or α_IIb-F993A that mimics inside-out activation of α_IIbβ₃ integrin. C, PAC-1 binding of α_IIb TM proline mutations on the background of the β₃ cytoplasmic mutations, β₃-F993A that mimics inside-out activation of α_IIbβ₃ integrin. D, PAC-1 binding of selected β₃ TM proline mutations on the background of the α_IIb mutations, α_IIb-G976L or α_IIb-GAAKR that mimics inside-out activation of α_IIbβ₃ integrin. HEK293FT cells were transfected with indicated α_IIbβ₃ constructs. The PAC-1 binding was performed in the buffer containing 1 mM Ca²⁺/Mg²⁺ and detected by flow cytometry. The level of PAC-1 binding was normalized to α_IIbβ₃ surface expression detected by mAb AP3 and shown as mean ± SD (n = 3). Unpaired two-tailed t test was performed between the control group without proline mutation and the group with proline mutation.

Figure 3

Effect of disturbing the rigidity of TM and CT domains on talin1-head induced α_IIbβ₃activation.A, effect of the selected α_IIbβ₃ TM proline mutations on EGFP-talin-head (EGFP-TH) induced PAC-1 binding to α_IIbβ₃ integrin. B, the β₃-GS or β₃-GGGS mutation was generated by the insertion of GS or GGGS sequence before the conserved β₃-D723 residue. C, effect of β₃-GS and β₃-GGGS mutations on EGFP-TH induced PAC-1 binding to α_IIbβ₃ integrin. HEK293FT cells were transfected with indicated α_IIbβ₃ plus EGFP or EGFP-TH constructs. The PAC-1 binding was performed in the buffer containing 1 mM Ca²⁺/Mg²⁺ and detected by flow cytometry. The PAC-1 binding of EGFP and α_IIbβ₃ double-positive cells was normalized to α_IIbβ₃ surface expression detected by mAb AP3 and shown as mean ± SD (n ≥ 3). Unpaired two-tailed t test was performed between the control group without proline mutation and the group with proline mutation. D, interaction of β₃-GS and β₃-GGGS mutants with talin1-head domain. HEK293FT cells were transfected with α_IIb WT and indicated β₃ plus EGFP-TH constructs. Total cell lysates were immunoprecipitated with anti-EGFP antibody and immunoblotted with anti-β₃ antibody.

Figure 4

Effect of TM proline mutations on the activation of β₂and β₁integrins.A, effect of TM proline mutations of β₂ integrin on ICAM-1 binding to α_Lβ₂ integrin. The ICAM-1 binding to α_Lβ₂-transfected HEK293FT cells was performed in the presence of 1 mM CaCl₂ plus 1 mM MgCl₂ (Ca/Mg). B and C, effect of TM proline mutations of β₂ integrin on the binding of activation-dependent mAbs, KIM127 and m24. The mAb binding to α_Lβ₂-transfected HEK293FT cells was performed in the presence of 1 mM Ca/Mg or 1 mM CaCl₂ plus 1 mM MnCl₂ (Ca/Mn). D, effect of TM proline mutations of β₁ integrin on talin-head (TH)-induced fibronectin (Fn) binding to α₅β₁ integrin. α₅β₁-KO HEK293FT cells were transfected with α₅β₁ plus EGFP or EGFP-TH constructs. The binding of Alexa Fluor 647-labled human fibronectin (Fn) was performed in the presence of 1 mM Ca/Mg. Both the ligand and the LIBS mAb binding were normalized to integrin expression. Data are mean ± SD (n ≥ 3). Unpaired two-tailed t test was performed between the control group without proline mutation and the group with proline mutation.

Figure 5

Effect of TM proline mutations on α_IIbβ₃-mediated cell spreading.A–C, HEK293 cells stably expressing α_IIb/β₃ WT, α_IIb/β₃-L705P-A711P, or α_IIb-I982P/β₃ spread on immobilized PAC-1. D and E, HEK293 cells stably expressing α_IIb/β₃ WT or α_IIb/β₃-L705P-A711P spread on immobilized fibrinogen. Cells were allowed to adhere on platelets coated with 5 μg/ml PAC-1 or 25 μg/ml fibrinogen at 37 °C for 1 h and then fixed and stained with Alexa-Fluor-488-labeled AP3 for β₃ detection and Alexa-Fluor-564-labeled phalloidin for F-actin detection. Nuclei were stained with DAPI in panels A–C. F, quantification of cell spreading areas in panels A–E. The averaged cell areas were calculated based on 30 to 50 cells for each of three independent repeats. Data are mean ± SD. Unpaired two-tailed t test was performed between the WT and the mutant cells.

Figure 6

BM-labeling of α_IIbβ₃TM cysteine substitutions before and after soluble ligand binding.A, the cysteine scanning accessibility method for determining the burial/exposure status of integrin TM domain. A single cysteine mutation was introduced into the α or β TM domain. The cysteine located outside of the cell membrane can be labeled by biotin-maleimide (BM), while the cysteine buried in the cell membrane is not accessible to the labeling. B–E, the BM labeling of substituted cysteines in the β₃ (B and D) and α_IIb (C and E) TM domains. HEK293FT cells were transfected with wild-type α_IIb plus β₃ containing a single-cysteine mutation or wild-type β₃ plus α_IIb containing a single-cysteine mutation. The cells were treated with or without the high-affinity ligand eptifibatide (Ept) and labeled with BM on ice. The α_IIbβ₃ was immunoprecipitated with α_IIbβ₃-complex specific mAb 10E5 and immunoblotted with indicated detection reagents. F–I, quantification of western blot data as described in panels B–E. Data are mean ± SD (n ≥ 2).

Figure 7

BM labeling of the cysteine substitutions in the membrane-proximal (MP) region of α_IIbcytoplasmic tail.A, the introduction of a disulfide bond that cross-links the α_IIb and β₃ subunit at the N-terminus of TM domain. HEK293FT cells were transfected with β₃ WT or β₃-P688C mutant and the α_IIb MP cysteine mutants in the presence of L959C mutation. The α_IIbβ₃ proteins were immunoprecipitated with mAb 10E5 and subjected to immunoblot with anti-β₃ mAb AP3 under nonreduced condition. B, disulfide bond formation between α_IIb-L959C and β₃-P688C reversed the α_IIbβ₃ activation induced by the cysteine substitutions at the α_IIb MP region. HEK293FT cells were transfected with indicated α_IIb constructs plus β₃-WT or β₃-P688C mutant. PAC-1 binding to the transfected cells was performed in the presence of 1 mM Ca/Mg. Data are mean ± SD (n ≥ 2). C, BM labeling of α_IIb MP cysteine mutants in the absence and presence of the disulfide bond formed between α_IIb-L959C and β₃-P688C. D, quantification of western blot data as described in panel C. Data are mean ± SD (n ≥ 2).

Figure 8

BM labeling of the cysteine substitutions in the β₃TM domain upon cell adhesion or EGFP-TH-induced α_IIbβ₃activation.A, HEK293FT cells were transfected with α_IIb WT plus indicated β₃ cysteine mutants. The transfected cells were kept in suspension or allowed to adhere on fibrinogen-coated surface and labeled with BM. B, quantification of western blot data of panel A. C, HEK293FT cells were transfected with EGFP-TH and α_IIb-R995A plus indicated β₃ cysteine mutants. The cells were labeled with BM in suspension. D, quantification of western blot data of panel C. Data are mean ± SD of three independent experiments.

Figure 9

BM labeling of the cysteine substitutions in the β₃TM domain that was fully separated from the α_IIbsubunit.A, design of β₃-tail-TMCT construct. A protein C tag was added to the N-terminus of β₃-tail-TMCT construct. B and D, BM labeling of the cysteine substitutions in the TM domain of full-length β₃ coexpressed with α_IIb in HEK293FT cells. C and E, BM labeling of the cysteine substitutions in the TM domain of β₃-tail-TMCT expressed in HEK293FT cells. F, quantification of western blot data of panels B and C. G, quantification of western blot data of panels D and E. Data are presented as BM signal as a percentage of total β₃ signal.