Finding the weakest link: exploring integrin-mediated mechanical molecular pathways

Abstract

From the extracellular matrix to the cytoskeleton, a network of molecular links connects cells to their environment. Molecules in this network transmit and detect mechanical forces, which subsequently determine cell behavior and fate. Here, we reconstruct the mechanical pathway followed by these forces. From matrix proteins to actin through integrins and adaptor proteins, we review how forces affect the lifetime of bonds and stretch or alter the conformation of proteins, and how these mechanical changes are converted into biochemical signals in mechanotransduction events. We evaluate which of the proteins in the network can participate in mechanotransduction and which are simply responsible for transmitting forces in a dynamic network. Besides their individual properties, we also analyze how the mechanical responses of a protein are determined by their serial connections from the matrix to actin, their parallel connections in integrin clusters and by the rate at which force is applied to them. All these define mechanical molecular pathways in cells, which are emerging as key regulators of cell function alongside better studied biochemical pathways.

Fig. 1.

The actin–integrin mechanical connection. Known direct and some indirect connections are shown here. For FAK and p130Cas, possible links connecting them to integrins and actin are shown. Myosin X (MYO10, not discussed in the main text) also binds integrin-β tails. Rather than exerting contractile forces between integrins and actin, however, this myosin binds integrins to transport them to adhesion sites (Zhang et al., 2004). The steps indicated in roman numbers show proposed molecular mechanosensing mechanisms: force could be detected by (i) exposing cryptic domains in ECM molecules, (ii) increasing the lifetime of fibronectin–α5β1-integrin bonds (catch bonds) and activating integrins, (iii) increasing the level of phosphorylation of p130Cas substrate domain, (iv) releasing FAK auto-inhibition and increasing its tyrosine kinase activity, (v) exposing previously buried vinculin-binding sites in talin rod domain, (vi) exposing previously buried binding sites to integrin-β tails or FIP in filamin immunoglobulin-like domains, (vii) releasing MLCK auto-inhibition and activating myosin II, and (viii) increasing the lifetime of myosin-IIA–actin bonds (catch bonds).

Fig. 2.

Summary of mechanical factors in integrin adhesions. (A) Role of stiffness. Whether and where a single connection between integrins and actin unfolds or breaks under force will depend on the stiffness of the ECM molecule (K_E), integrin (K_I), and adaptor protein (K_A), and the maximum force that can be withstood by the ECM–integrin bond (F_EI), the integrin–adaptor bond (F_IA) and the adaptor–actin bond (F_AA). These stiffness values will generally increase with protein stretching, and change after unfolding or unbinding events. To our knowledge, integrin stiffness (K_I) or possible unfolding have not been described. The total stiffness of the chain (k₁) will be determined by the softest element in it, and will operate in parallel with other chains in the cluster (k₂, k₃). The rearward speed v of the actin layer, multiplied by k₁, will provide the rate of force loading in the chain, which will affect both individual rupture forces (F_EI, F_IA, and F_AA) and the effectiveness with which the different chains in the cluster cooperate to withstand force. The spacing between integrins (s) will also affect cluster formation. (B) Nascent focal adhesions. In nascent adhesions, adaptor proteins mediate the formation of initial integrin clusters {containing talin, kindlin and PtdIns(4,5)P₂}, and might promote actin polymerization from clusters (formins). Myosin II filaments exert local contractile forces between filaments (blue arrows). (C) Focal adhesion maturation. As adhesions mature and grow, they connect to actin fibers moving rearward towards the cell center due to global myosin contractility (blue arrows). Rearward moving actin pulls on bound adaptor proteins, which in turn pull on integrins. Because of transient stick-slip bonds, rearward speeds are reduced with respect to actin to a certain extent in adaptor proteins, and even further in integrins (black arrows).

Fig. 3.

Schematic illustration of the signaling processes downstream of force dectection. The upper panel shows RPTP-α-, Fyn- and talin-dependent signaling events. (i) Cell attachment, integrin clustering and deformation of the extracellular matrix result in RPTP-α activation. This allows the SFK Fyn to access the catalytic site. Fyn is dephosphorylated at tyrosine 531 and switches to the open conformation. (ii) Fyn subsequently phosphorylates Rac1 GEFs, such as Tiam1, that in turn activate Rac1. (iii) Rac1 activity induces actin polymerization from integrin adhesion sites by formin family proteins and/or the ARP2/3 complex. (iv) Rearward actin flow driven by myosin II contractility stretches talin to expose cryptic vinculin-binding sites, which leads to adhesion strengthening and cluster growth. (v) When high forces are generated by actin rearward flow and contractility, the bond between actin and talin slips and talin relaxes. The lower panel depicts FAK- and p130Cas-dependent downstream signaling. (vi) Downstream of integrin clustering, FAK is autophosphorylated at tyrosine 397. Phosphorylated FAK binds to talin and phosphorylates the Rac1 GEF β-PIX (encoded by ARHGEF7) (Karasawa et al., 1982). (vii) FAK also forms a complex with Src to activate p130Cas. p130Cas is hyperphosphorylated and binds to Crk and DOCK180 to activate small Rho GTPases, such as Rap. (viii) Rac1 induces actin polymerization and cell protrusion, whereas Rap enhances adhesion maturation (ix). (x) Additionally, FAK can activate p190GEF (Lim et al., 2008b) to enhance RhoA activity, stress fiber formation (xi) and adhesion turnover. (xii) Cross regulation can alter the region of activity of Rho GTPases.