Talin depletion reveals independence of initial cell spreading from integrin activation and traction

Abstract

Cell spreading, adhesion and remodelling of the extracellular matrix (ECM) involve bi-directional signalling and physical linkages between the ECM, integrins and the cell cytoskeleton. The actin-binding proteins talin1 and 2 link ligand-bound integrins to the actin cytoskeleton and increase the affinity of integrin for the ECM. Here we report that depletion of talin2 in talin1-null (talin1(-/-)) cells did not affect the initiation of matrix-activated spreading or Src family kinase (SFK) activation, but abolished the ECM-integrin-cytoskeleton linkage and sustained cell spreading and adhesion. Specifically, focal adhesion assembly, focal adhesion kinase (FAK) signalling and traction force generation on substrates were severely affected. The talin1 head domain restored beta1 integrin activation but only full-length talin1 restored the ECM-cytoskeleton linkage and normal cytoskeleton organization. Our results demonstrate three biochemically distinct steps in fibronectin-activated cell spreading and adhesion: (1) fibronectin-integrin binding and initiation of spreading, (2) fast cell spreading and (3) focal adhesion formation and substrate traction. We suggest that talin is not required for initial cell spreading. However, talin provides the important mechanical linkage between ligand-bound integrins and the actin cytoskeleton required to catalyse focal adhesion-dependent pathways.

Figure 1

Talin2 depletion in talin1^−/− cells does not affect initial spreading on fibronectin but causes cell rounding. (a) GFP and talin2 siRNA co-transfected cells rounded 90 min after plating on fibronectin-coated coverslip. GFP and control siRNA or GFP–talin1 and talin2 siRNA co-transfected cells showed a normal elongated fibroblast-like morphology. (b) The decrease in talin2 levels ranged from 48–68%, depending on the transfection efficiency of talin2 siRNA. (c, d) Time-lapse sequential images of the talin1^−/− control cell (c, Supplementary Information, Movie 1) and the talin-deficient cell (d, Supplementary Information, Movie 2) spreading on fibronectin. (e) Summary of early spreading behaviours of control and talin-deficient cells. (f) Time lapse between initiation of spreading and complete rounding up of talin-deficient cells. Scale bars are 50 μm (a), 10 μm (c, d).

Figure 2

Spread talin-deficient cells show defects in focal adhesion formation and adhesion to substrates. (a–h) Talin-deficient cells spread without assembling focal adhesions. Control siRNA (a, c, e, g) and talin2 siRNA (b, d, f, h) transfected cells were plated on fibronectin for 20 min before fixation and were stained for the indicated proteins and F-actin. (i, j) Talin-deficient cells spread with defective adhesion to the substrate. GFP and control siRNA (i) or GFP and talin2 siRNA (j, Supplementary Information, Movie 3) co-transfected cells were plated on fibronectin. DIC and TIRF images of live cells were taken at 10 s per frame. Scale bars are 5 μm (c) and 10 μm (a, e, g, i)

Figure 3

Talin1 head activates β1 integrin but does not rescue focal adhesion formation. (a) The spread phase on fibronectin was prolonged in cells expressing the talin1 head (talin1H) but not the talin1 rod (talin1R). Talin1^−/− cells co-transfected with GFP-talin1H and talin2 siRNA or with GFP-talin1R and talin2 siRNA were plated on fibronectin for 90 min before fixation. (b) Summary of early spreading behaviour of talin1H-expressing cells. (c) Sequential time-lapse images of the talin1H-expressing cell spreading on fibronectin (Supplementary Information, Movie 4). (d, e) Focal adhesion formation was defective in talin1H-expressing cells. 90 min after plating, cells were fixed and stained for vinculin (d) and paxillin (e). (f–h) Talin, through its head domain, activated and colocalized with β1 integrin at the leading edge of spread cells. GFP–talin1-restored cells (f), GFP–talin1H-expressing cells (g) and talin-deficient cells (h) were plated on fibronectin for 20 min, 90 min and 20 min, respectively before fixation. Scale bars are 50 μm (a), 5 μm (e) and 10 μm (c, d, f–h).

Figure 4

Talin depletion impairs FAK-Tyr 397 phosphorylation upon adhesion to fibronectin. (a) Western blots showed defects in adhesion-dependent FAK-Tyr 397 phosphorylation in talin-deficient cells. Control, talin2 siRNA, and talin1H and talin2 siRNA co-transfected cells were plated on fibronectin-coated culture dishes for the indicated time before lysis. (b) Profile of FAK-Tyr 397 phosphorylation levels with adhesion to fibronectin. pTyr 397: FAK ratio of control cells at 120 min was set at 1 arbitrary unit (AU) and the values of other time points were adjusted proportionally. Data are mean ± s.e.m., n = 3. (c) Activation profile of SFK with adhesion to fibronectin. pTyr 416: Src ratio increase of 10 min above the suspension sample in control cells was set at 1 AU and other time points were adjusted proportionally. Data are mean ± s.e.m., n = 3. (d–f) Immunostaining for phospho-FAK (pTyr 397) in control (d), talin2-deficient (e) and talin1H-expressing cells (f). Cells were fixed 20 min (d, e) and 90 min (f) after plating. The pTyr 397 signal was lost in talin2-deficient and talin1H-expressing cells. Scale bars are 10 μm.

Figure 5

Talin couples actomyosin contractility to the substrate and facilitates the distribution of microtubules. (a –d) Cells without full-length talin could not confine fast α-actinin rearward flow to the leading edge. Talin1^−/− cells co-transfected with α-actinin–GFP and control siRNA (a, Supplementary Information, Movie 6) or α-actinin–GFP, DsRed-talin1H and talin2 siRNA (c, Supplementary Information, Movie 7) were plated on fibronectin. DIC, TIRF or epifluorescence (epi) sequential images were taken at 10 s per frame. Kymographs (b, d) were generated from Supplementary Information, Movies 6 and 7, respectively. There was partial loss of the TIRF signal in d caused by ventral membrane detachment. Regions with fast rearward flow and the flow rate were determined by kymographs (averaged from three measurements for each cell). Scale bars are 5 μm (a, c) and 1 min (horizontal), 2 μm (vertical) (b, d). (e, f) Blebbistatin treatment (Bb, 50 μM) slowed the rearward flow in talin1H-expressing cells, as revealed by the kymograph of DIC (f, Supplementary Information, Movie 8). Cell flattening was observed immediately after drug perfusion. Images before and 2.5 min after treatment were shown (e). Scale bars are 10 μm (e), 1 min and 5 μm (f). (g, h) Coupling of actomyosin contractility to the substrate was affected in the absence of full-length talin. GFP–talin1 and talin2 siRNA or GFP–talin1H and talin2 siRNA co-transfected cells were plated on fibronectin and stained for phospho-MLC. (i–k) Actomyosin activity restricted microtubules in talin1H-expressing cells. Talin1H-expressing cells (k) were plated in the same way as talin1-restored (i) and talin1H-expressing (j) cells, but were treated with 50 μM blebbistatin for 2.5 min before fixation. Scale bars are 10 μm (g–k).