Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2020 Mar 10;9(3):227.
doi: 10.3390/antiox9030227.

Extracellular Redox Regulation of α7β Integrin-Mediated Cell Migration Is Signaled via a Dominant Thiol-Switch

Affiliations
Free PMC article

Extracellular Redox Regulation of α7β Integrin-Mediated Cell Migration Is Signaled via a Dominant Thiol-Switch

Lukas Bergerhausen et al. Antioxidants (Basel). .
Free PMC article

Abstract

While adhering to extracellular matrix (ECM) proteins, such as laminin-111, cells temporarily produce hydrogen peroxide at adhesion sites. To study the redox regulation of α7β1 integrin-mediated cell adhesion to laminin-111, a conserved cysteine pair within the α-subunit hinge region was replaced for alanines. The molecular and cellular effects were analyzed by electron and atomic force microscopy, impedance-based migration assays, flow cytometry and live cell imaging. This cysteine pair constitutes a thiol-switch, which redox-dependently governs the equilibrium between an extended and a bent integrin conformation with high and low ligand binding activity, respectively. Hydrogen peroxide oxidizes the cysteines to a disulfide bond, increases ligand binding and promotes cell migration toward laminin-111. Inversely, extracellular thioredoxin-1 reduces the disulfide, thereby decreasing laminin binding. Mutation of this cysteine pair into the non-oxidizable hinge-mutant shows molecular and cellular effects similar to the reduced wild-type integrin, but lacks redox regulation. This proves the existence of a dominant thiol-switch within the α subunit hinge of α7β1 integrin, which is sufficient to implement activity regulation by extracellular redox agents in a redox-regulatory circuit. Our data reveal a novel and physiologically relevant thiol-based regulatory mechanism of integrin-mediated cell-ECM interactions, which employs short-lived hydrogen peroxide and extracellular thioredoxin-1 as signaling mediators.

Keywords: cell migration; extracellular thioredoxin-1; integrin α7β1; laminin binding; redox regulation; redox signaling; thiol-switch.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Figure A1
Figure A1
In a dose-dependent manner, reduced Trx1 promotes chemohaptotactic migration of non-transfected HT1080 cells toward poly-l-lysine in an integrin-independent manner. The bottom face of a CIM-plate was coated with 0.01 poly-l-lysine at 4 °C overnight. A group of 5 × 104/well non-transfected HT1080 cells was allowed to migrate in the presence of different concentrations of reduced Trx1. The Δcell index values were monitored. From them, relative ΔΔcell index values were calculated. Means ± SD of triplicates from a representative experiment out of three experiments are shown. Significance levels: ** p ≤ 0.01; *** p ≤ 0.005
Figure A2
Figure A2
Mass-spectrometric analysis of proteins trapped by Trx1-trap mutant. The Cys-X-X-Ser mutant of Trx1 was applied in an intermediate trapping approach to trap Trx1 substrates on α7wt-HT1080 cells. The trapped proteins were analyzed after tryptic digestion, using liquid chromatography coupled with mass spectrometry. Among the trapped proteins, integrin α7 was identified with a MaxQuant score of 22.9 with three individual peptides (marked in red). All three peptides were identified at a 2+ charge state, with individual MaxQuant scores of 76.5 (AIDLEQPNCADGR, PEP 1.1 E-7), 44.9 (LIPEVVLSGER, PEP 0.0027) and 36.7 (SEELSFVAGAPR, PEP 0.0055). The two cysteines of the hinge motif are highlighted with black boxes.
Figure 1
Figure 1
Schematic overview of integrin α7 constructs (a) and quantification of their transcription (b) and translation (c) in transfected HT1080 cells. (a) Redoxmodifiable cysteines of the integrin α7, located within the hinge region (C604, C610) and calf-2 domain (C862, C916, C923, C928) (yellow circles), were replaced by alanine residues (gray diamonds). HI, hinge (knee) domain; TM, transmembrane domain; CP, cytoplasmic domain; fos, heterodimerizing fos-zipper motif. (b) The qRT-PCR analysis of transfected integrin α7 constructs in HT1080 cells. To obtain relative values, the mRNA amounts of the α7 mutants were normalized to the α7wt value (dotted line). Means ± SD are shown for three independent experiments. Numbers (#1, #2 and #3) for α7ca2 and α7hi-ca2 indicate three independently transfected cell populations. (c) Sandwich-ELISA of whole HT1080 cell lysates (protein concentration: 1 mg/mL) performed with anti-integrin α7 mAb 3C12 and rabbit anti-integrin β1 serum as capturing and detecting antibody.
Figure 2
Figure 2
Subunits α7wt and the α7hi mutant, but not the α7ca2 and α7hi-ca2 mutants, are expressed as heterodimeric integrin on the surface of transfected HT1080 cells and induce morphological changes of cells after adhesion to laminin-111 and collagen-I. First column: surface exposure of integrin α7β1 wild type and mutants on transfected HT1080 cells was quantified by flow cytometry independently with two monoclonal antibodies, 3C12 and mAb3518 (clone #334908). Second to fourth columns: immunofluorescence of adherent transfectants on laminin-111 and collagen-I, either with (second column) or without (third and fourth column) permeabilization, and stained with an anti-integrin α7 antibody 3C12 (green), with phalloidin (red) and Hoechst dye (blue). Scale bars = 20 µm. Representative pictures of at least two independent staining experiments with at least 10 cells per condition are shown.
Figure 3
Figure 3
Cell morphology of HT1080 cells expressing α7wt and α7hi in life cell microscopy. Integrin α7hi, but not α7wt, induces bleb-like membrane protrusions in transfected HT1080 cells when plated on laminin-111, but not on collagen I. LifeAct-GFP-transduced HT1080 cells were monitored by life cell microscopy on µ-slides coated with laminin-111 or rat tail collagen I (ColI). Images were taken with confocal microscopy with 40× oil objective at 37 °C. Scale bar = 20 µm. Maximum projection of z-stacks was done with FiJi. Representative images from at least 30 cells for each condition are shown.
Figure 4
Figure 4
Adhesion and migration of HT1080 cells expressing integrin α7wt and α7hi. (a) Adhesion to laminin-111of naïve and transfected HT1080 cells in an E-plate of the xCELLigence DP device. Cells expressing α7wt or α7hi adhered faster and with a higher saturation signal than the naïve HT1080 cells. Adhesion via α6β1 integrin was challenged by adding the antibody GoH3. (b) Maximum adhesion signals and (c) maximum slope, indicating the maximum adhesion rate, were compared for at least three independent experiments. Means ± SD are shown. (d) Haptotactic migration toward laminin-111 of HT1080 cells was recorded in a CIM plate of the xCELLigence DP device. Cells expressing α7hi moved much slower than α7wt transfectants and naïve cells. (e) The migration rates were determined as change of Δcell index values over time, between 30 and 60 min. Experiments were performed in Tyrode’s solution. Means ± SD of a quadruplet determination of one out of three experiments are shown. The data in (a) and (d) were compared for the effect of GoH3 by two-way-ANOVA with multiple comparison correction via the Holm–Sidak method. Time periods with significant differences are marked in the corresponding color of the cells. As only the naïve HT1080 showed significant changes upon GoH3 treatment after 1 h, the comparison bars for comparing the GoH3-free vs. GoH3-treated samples are omitted in (b), (c) and (e). In the same plots, comparison bars are only shown for the comparison of α7wt vs. α7hi transfectants and are based on pairwise comparisons with Student’s t-test. Significance levels: * p ≤ 0.05; ***. p ≤ 0.005; ****, p ≤ 0.001.
Figure 5
Figure 5
Effect of H2O2 on integrin α7-dependent migration. Migration on laminin-111 of HT1080 cells expressing integrin α7wt (a) or α7hi (b) was monitored on CIM-plates in an xCELLigence DP device, in the absence and presence of 10 µM H2O2. The experiments were carried out in Tyrode’s solution and in the presence of 2 µg/mL GoH3. Means ± SD. of Δcell index values, measured in 5 min intervals, from at least four independent measurements are shown. The Holm-Sidak method revealed a significant difference (p < 0.05) for α7wt-, but not α7hi-transfectants, after the first 60 min of H2O2 treatment. (c) The average migration rates between 30 and 60 min after migration start are shown from one of three experiments with quadruplet determination with means ± SD. Significance level: ** p ≤ 0.01. (d) Relative ΔΔcell index values were calculated at each time point for HT1080 cells expressing α7wt or α7hi to compare H2O2-treated and nontreated samples. Means ± SD are shown (n = 4 for α7wt and n = 6 for α7hi).
Figure 6
Figure 6
Surface-exposed α7β1 integrin containing the α7wt, but not the α7hi mutant, preferentially takes the active/extended conformation with an accessible 9EG7 epitope. (a,b) Representative flow cytometric histograms, such as the ones shown here for α7wt transfected HT1080 cells, reveal (a) all β1 integrin heterodimers, irrespective of their conformational state, after staining with anti-β1 mAb MEM-101A and Alexa Fluor 405-conjugated secondary antibodies. No cross reaction was observed for 9EG7 as primary antibody. (b) Integrin β1 molecules in an extended conformation stained with biotinylated 9EG7 and secondarily with phycoerythrin (PE)-conjugated NeutrAvidin. No cross reaction was detected for MEM-101A or 12G10 as primary antibodies. To prove 9EG7 functionality, β1 activating antibody 12G10 was added along with 9EG7, representing the maximum signal reachable with 9EG7. (c) All surface-exposed β1-integrins were detected with MEM-101A antibody in a conformation-independent manner. The quantity of β1-integrins was lower in HT1080 cells transfected with the α7wt construct. (d) 9EG7 detected the portion of active conformation of integrin β1 relative to the maximum reachable signal in presence of 12G10. Box plots indicate medians, 25th and 75th percentile, and Tukey whiskers. In total, 12,720, 9117 and 8349 events were evaluated for naïve, α7wt and α7hi, respectively. ANOVA with Tukey’s multiple-comparisons test was performed. Significance levels are indicated by asterisks (*** p ≤ 0.005; **** p ≤ 0.001).
Figure 7
Figure 7
Ligand-binding activity and binding force of α7 integrins toward laminin-111. (a) Soluble integrins, α7wtβ1 and α7hiβ1, pretreated with or without 100 µM H2O2, were added to immobilized laminin-111. Means (±SEM) of triplicates of a representative experiment are shown. At least six titration curves for each condition were fitted according to [35], to obtain KD values. (b) AFM-based force measurements on single molecule interactions between integrins, α7wtβ1 and α7hiβ1, and laminin-111. After the cantilever with immobilized integrin was approached to the laminin-111 coated surface, it was retracted with a velocity of 300 nm/s. Force curves for α7wt and α7hi were measured two times, with three loading rates. Rupture force of the last contact was evaluated for 331 and 168 significant force–distance curves for α7wt and α7hi, respectively. For the histogram, the results were binned with a force width of 3 pN.
Figure 8
Figure 8
Extracellular thioredoxin-1 (Trx1) reduces cell migration and binding to laminin-111, mediated by α7wtβ1, but not by α7hiβ1 integrin. (a) Migration of HT1080 cells toward laminin-111. HT1080 cells expressing integrin α7wt (blue and purple lines) or α7hi (red and orange lines) were pretreated without and with 10 µM H2O2 for 30 min and washed. Migration was monitored on CIM-plates in the xCELLigence DP device, in the absence (full line) and presence (dashed line) of 10 µg/mL reduced Trx1 in Tyrode’s solution. Means of triplicate Δcell index values from at least three independent measurements are shown. The variation coefficients varied around 10–20% on average. (b) For every time point, relative ΔΔcell index values were calculated from the data in (a), revealing the Trx1-dependence of the measured values. Means of triplicate values are shown for one out of three representative experiments. (c) Immunoblot detection of β1 integrins in the eluate of a Trx1-trap mutant affinity column. The Trx1 trap mutant with its Cys-X-X-Ser active site retained Trx1-bound proteins from a cell lysate of α7wt-transfected HT1080 cells. In parallel, the eluate contained inter alia the integrin α7 subunit, as proven by mass spectrometry (Appendix A, Figure A2). (d) Titration of laminin-111 with soluble α7wtβ1 integrin, pretreated without and with 100 µM H2O2 (blue and purple lines). After removal of H2O2, binding was tested in the presence (dashed lines) and absence (full line) of 10 µg/mL Trx1, including a subsequent NEM-treatment, to prevent thiol oxidation. One triplicate set of two independent titration experiments are shown. (e) KD-values, and (f) saturation binding signals of six independent titration curves (as shown in (d)) were evaluated, and their means ± SD are shown. Significance levels are indicated by asterisks: * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.005.
Figure 9
Figure 9
Electron microscopy of recombinant soluble α7β1 integrin after negative staining. (a) The electron micrographs show representative fields of α7wtβ1 (left column) and α7hiβ1 (right column) molecules (scale bar: 20 nm). The tadpole-like complexes with one or two tails exhibit both bent (arrowheads) and extended (arrows) integrin forms. The percentage of molecules in different conformations was determined from at least 500 molecules under each condition (top row: no additives; middle row: 10 µM H2O2; bottom row: 10 µg/mL Trx1). The percentage of bent conformation is printed in each image. (b) Quantification of conformations of integrin molecules from electromicrographs, such as in (a), including a differentiation of the extended conformation into an extended close and extended open conformation. (c) Scheme of the thiol-switch mediated conformational changes of α7β1 integrin. The integrin hinge region is indicated as the loop in the central pivot region of the integrin molecule. Rotation around the hinge ensures the conformational equilibrium between bent and extended form. The pair of cysteines within the α7 hinge in the bent conformation is oxidized with H2O2 to mono-cysteine-sulfenic acid derivatives, which form a disulfide bond. This stabilizes the extended conformations. The disulfide-bonded form of the thiol-switch can be reduced by extracellular Trx1. This is accompanied with transition into the bent conformation. Mutation of both cysteines for alanines (upper left panel) represents the loop structure of the less active α7hi. The thiol-switch-dependent conformational changes correlate with changes in ligand binding, in 9EG7 epitope recognition and in cell migration.

Similar articles

See all similar articles

References

    1. Aumailley M. The laminin family. Cell Adh. Migr. 2013;7:48–55. doi: 10.4161/cam.22826. - DOI - PMC - PubMed
    1. Rodriguez M.L., Graham B.T., Pabon L.M., Han S.J., Murry C.E., Sniadecki N.J. Measuring the contractile forces of human induced pluripotent stem cell-derived cardiomyocytes with arrays of microposts. J. Biomech. Eng. 2014;136:051005. doi: 10.1115/1.4027145. - DOI - PMC - PubMed
    1. Campbell I.D., Humphries M.J. Integrin structure, activation, and interactions. Cold Spring Harb. Perspect. Biol. 2011;3 doi: 10.1101/cshperspect.a004994. - DOI - PMC - PubMed
    1. Humphries J.D., Byron A., Humphries M.J. Integrin ligands at a glance. J. Cell Sci. 2006;119:3901–3903. doi: 10.1242/jcs.03098. - DOI - PMC - PubMed
    1. Su Y., Xia W., Li J., Walz T., Humphries M.J., Vestweber D., Cabanas C., Lu C., Springer T.A. Relating conformation to function in integrin alpha5beta1. Proc. Natl. Acad. Sci. USA. 2016;113:E3872–E3881. doi: 10.1073/pnas.1605074113. - DOI - PMC - PubMed
Feedback