Sharpin suppresses β1-integrin activation by complexing with the β1 tail and kindlin-1

Abstract

Background: Previously sharpin has been identified as an endogenous inhibitor of β1-integrin activation by directly binding to a conserved region in the cytoplasmic tails (CTs) of the integrin β1-associated α subunits.

Methods: Here we employed biochemical approaches and cellular analyses to evaluate the function and molecular mechanism of the sharpin-kindlin-1 complex in regulating β1-integrin activation.

Results: In this study, we found that although the inhibition of sharpin on β1-integrin activation could be confirmed, sharpin had no apparent effect on integrin αIIbβ3 activation in CHO cell system. Notably, a direct interaction between sharpin and the integrin β1 CT was detected, while the interaction of sharpin with the integrin αIIb and the β3 CTs were substantially weaker. Importantly, sharpin was able to inhibit the talin head domain binding to the integrin β1 CT, which can mechanistically contribute to inhibiting β1-integrin activation. Interestingly, we also found that sharpin interacted with kindlin-1, and the interaction between sharpin and the integrin β1 CT was significantly enhanced when kindlin-1 was present. Consistently, we observed that instead of acting as an activator, kindlin-1 actually suppressed the talin head domain mediated β1-integrin activation, indicating that kindlin-1 may facilitate recruitment of sharpin to the integrin β1 CT.

Conclusion: Taken together, our findings suggest that sharpin may complex with both kindlin-1 and the integrin β1 CT to restrict the talin head domain binding, thus inhibiting β1-integrin activation.

Keywords: Integrin; Kindlin-1; Sharpin; Talin.

Fig. 1

Sharpin suppresses integrin α5β1 activation but not integrin αIIbβ3 activation. a Sharpin (SH) was co-expressed in CHO cells together with DsRed-fused talin head (TH) and EGFP-fused kindlin-1 (K1) by transient transfection. Activation of endogenous integrin α5β1 in transfected cells were measured by the GST-Fn-III binding assay. b CHO cells that stably express αIIbβ3 (CHO-αIIbβ3) were used to express the indicated regulators. Their effects on integrin αIIbβ3 activation were evaluated by the PAC-1 antibody binding assay. c CHO cells were transfected with two different siRNA (siRNA1 and siRNA2) specifically targeting endogenous sharpin in CHO cells or non-targeting siRNA (NCS) as a control. 24 h after transfection, expression of sharpin protein in CHO cells was evaluated by immunoblotting. d, e NSC or two siRNA targeting hamster sharpin were co-transfected either in CHO cells (d) or CHO-αIIbβ3 cells (e) together with DsRed-fused talin head (TH) or TH plus EGFP-kindlin-1 (K1); their effects on integrin α5β1 activation in CHO cells or integrin αIIbβ3 activation in CHO-αIIbβ3 cells were evaluated by the GST-Fn-III binding assay and the PAC-1 antibody binding assay, respectively. f Sharpin (SH) was co-expressed in 3 T3 cells together with DsRed-fused talin head (TH) and EGFP-fused kindlin-1 (K1) by transient transfection. Activation of endogenous integrin α5β1 in transfected 3 T3 cells was measured by the 9EG7 antibody binding assay. g 3 T3 cells were transfected with three different siRNA (siRNAa, siRNAb and siRNAc) targeting endogenous sharpin in 3 T3 cells or non-targeting siRNA (NCS) as a control. 24 h after transfection, expression of endogenous sharpin protein in 3 T3 cells was evaluated by immunoblotting. h NSC or two different siRNA (siRNAb and siRNAc) were co-transfected in 3 T3 cells together with DsRed-fused talin head (TH) or TH plus EGFP-kindlin-1 (K1), and their effects on integrin α5β1 activation in 3 T3 cells were evaluated by the 9EG7 antibody binding assay. The results represent the mean ± SD of at least 3 experiments. (MFI: median of fluorescence intensity; ns, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001)

Fig. 2

Sharpin directly binds to the integrin β1 CT and inhibits the talin head domain binding. a Purified GST and GST-fused integrin CT, as indicated, were coupled to Glutathione Sepharose beads and used to incubate with his-tagged sharpin (His-SH). After incubation, the beads were extensively washed and proteins bound to the beads were eluted by boiling the beads in laemmli sample buffer. GST proteins loaded on the beads and co-precipitated His-SH were evaluated by SDS-PAGE followed by Coomassie blue (C. blue) staining and immunoblotting (IB). b The N-terminus (SH-N, 1–217 amino acids) and C-terminus (SH-C, 217–387 amino acids) of sharpin were expressed and purified with a his tag and used to test their binding to GST or GST-β1 CT, as described in (a). c Selected region of HSQC spectra of 50 μM ¹⁵N-labeled β1 CT in the absence (black) and presence (red) of 250 μM N-terminus of sharpin (SH-N). d Selected region of HSQC spectra of 50 μM ¹⁵N-labeled β3 CT in the absence (black) and presence (red) of 250 μM N-terminus of sharpin (SH-N). e, f Purified protein of the N-terminus of sharpin was immobilized on CM5 chip surfaces. Various concentrations (2.5 μM, 5 μM, 10 μM, 20 μM and 40 μM) of either the integrin β1 CT protein (e) or the integrin β3 CT protein (f) were injected and passed over the chips, and the binding curves were recorded on a Biacore 8 K instrument. g Purified GST and GST-β1 CT proteins were loaded onto Glutathione Sepharose beads which were then used for incubating with flag-fused talin head (Flag-TH) in the presence or absence of his-sharpin (His-SH) with different ratios. After incubation, the beads were washed and co-precipitated proteins were measured by SDS-PAGE followed by Coomassie blue (C. blue) staining and immunoblotting (IB). h Purified GST, GST-β1 CT and GST-β1 CT mutants that carry the NPIY/AAAA mutations or the KSAV/AAAA mutations were coupled to Glutathione Sepharose beads and used to incubate with His-SH, Flag-TH or kindlin-1 (K1) proteins, respectively. Binding of His-SH, Flag-TH or K1 to these GST proteins were evaluated by SDS-PAGE followed by immunoblotting (IB). Meanwhile, the loaded GST proteins on the beads were also measured by Coomassie blue (C. blue) staining

Fig. 3

Kindlin-1 directly interacts with sharpin (via its F0 subdomain) and recruits sharpin to inhibit β1-integrin activation. a The kindlin family members (K1, K2 and K3) were fused with the DNA-binding domain of Gal4 in pGBKT7 vector and sharpin (SH) was fused with the transcriptional activation domain of Gal4 in pGADT7 vector. The interaction between kindlin and sharpin was evaluated using the Matchmaker™ Gold yeast two-hybrid system by a serial dilution method on selection media. Two molecules known for interacting with each other (Bop1/Bop2) were used as a positive control and the empty vectors were used as a negative control. Growth of yeast cells on SD2 selection media indicates successful transformation; growth of yeast cells on SD4 selection media indicates a positive protein-protein interaction. b Glutathione Sepharose beads were loaded with GST and GST-fused sharpin (GST-SH) proteins and used to incubate with his-tagged kindlins (His-K1, His-K2 and His-K3). The loading of GST and GST-SH on the beads was measured by Coomassie blue (C. blue) staining. Binding of His-kindlins to GST proteins was analyzed by immunoblotting (IB). c Purified GST and GST-β1CT proteins were coupled to Glutathione Sepharose and used to incubate with flag-tagged talin head (Flag-TH) in the presence or absence of his-tagged sharpin (His-SH) and/or kindlin-1 (K1). After incubation, beads were extensively washed. The loading of GST proteins was evaluated by Coomassie blue (C. blue) staining. Precipitated protein samples on the beads, including Flag-TH, His-SH and K1, were evaluated by SDS-PAGE followed by immunoblotting (IB). d Glutathione Sepharose beads were loaded with GST and GST-fused kindlin-1 (GST-K1) proteins and used to incubate with his-tagged N-terminus (His-SH-N) or C-terminus (His-SH-C) of sharpin. The loading of GST and GST-K1 on the beads was measured by Coomassie blue (C. blue) staining. Binding of His-SH-N or His-SH-C to GST proteins was analyzed by immunoblotting (IB). e Kindlin-1 (K1) and its mutants, including K1ΔF0, K1ΔF1, the N-terminal fragment (F0 + F1) of kindlin-1 (K1N), and the kindlin-1 QW/AA mutant (K1AA), were expressed and purified with a his tag. Interaction of GST or GST-fused sharpin (GST-SH) with these kindlin-1 proteins were evaluated by pull-down assays followed by immunoblotting (IB). f Interaction of sharpin with kindlin-1 and its mutants were also evaluated using the Matchmaker™ Gold yeast two-hybrid system, same as described in (a). g The effects of kindlin-1 mutants, including K1AA and K1ΔF0, on integrin α5β1 activation in CHO cells were evaluated by co-transfection and the GST-Fn-III binding assay. The results represent the mean ± SD of at least 3 experiments. (MFI: median of fluorescence intensity; ns, not significant; *, p < 0.05; **, p < 0.01; ***, p < 0.001)

Fig. 4

Kindlin-1 suppresses β1-integrin-mediated cell adhesion and signaling. (A & B) CHO cells expressing the indicated regulators were used to incubate with coated fibronectin for 15 min. After washing and fixation, adherent cells were imaged (a) and counted (b). c CHO cells expressing the talin head (TH) plus empty vectors and TH plus kindlin-1 (K1) and sharpin (SH) were allowed to spread on coated fibronectin for 60 min and then fixed. Representative images of spreading cells were shown. Bar distance was 10 μm. d Spreading areas of cells were calculated with ImageJ software. e Adherent cells on fibronectin were directly lysed for immunoblotting using antibodies for focal adhesion kinase (FAK), tyrosine-phosphorylated FAK (Y-FAK) and actin. f The relative density of Y-FAK signals, as shown in (e), were quantified using ImageJ software. The results represent the mean ± SD of at least 3 experiments. (*, p < 0.05; **, p < 0.01; ***, p < 0.001)

Fig. 5

A model to show the different roles of sharpin in regulating integrin α5β1 and αIIbβ3 activation. a Sharpin directly interacts with the integrin β1 CT and kindlin-1 as well, thus being able to competitively inhibit the talin head binding to the β1 CT and leading to a negative regulation on integrin α5β1 activation. b Sharpin fails to interact with the integrin β3 CT, which is unlikely to affect the talin head binding to the β3 CT. In addition, kindlin-3, the dominant kindlin member in platelets to support integrin αIIbβ3 activation, also exhibits significantly compromised binding to sharpin