Zasp regulates integrin activation

Abstract

Integrins are heterodimeric adhesion receptors that link the extracellular matrix (ECM) to the cytoskeleton. Binding of the scaffold protein, talin, to the cytoplasmic tail of β-integrin causes a conformational change of the extracellular domains of the integrin heterodimer, thus allowing high-affinity binding of ECM ligands. This essential process is called integrin activation. Here we report that the Z-band alternatively spliced PDZ-motif-containing protein (Zasp) cooperates with talin to activate α5β1 integrins in mammalian tissue culture and αPS2βPS integrins in Drosophila. Zasp is a PDZ-LIM-domain-containing protein mutated in human cardiomyopathies previously thought to function primarily in assembly and maintenance of the muscle contractile machinery. Notably, Zasp is the first protein shown to co-activate α5β1 integrins with talin and appears to do so in a manner distinct from known αIIbβ3 integrin co-activators.

Fig. 1.

Human Zasp activates α5β1 integrins. (A) Schematic diagram of human Zasp variant1 (Zasp V1, 727 amino acids). The percentage amino acid identity between the three conserved domains of hZasp and Drosophila Zasp (Dm Zasp) is given below. (B) Flow cytometry data analysis of CHO cells transfected with DNA encoding GFP and/or DsRed-tagged talin head and Zasp V1 proteins. A gate was drawn to define a double positive (GFP and DsRed) population. Histogram plots from cells in this gate were generated to measure the geometric mean fluorescence intensity of GFP, DsRed, and FN9-11 or PB1. Activation of endogenous α5β1 integrin in CHO cells co-expressing GFP or GFP–talin head and DsRed or DsRed–HAZasp V1 was calculated as described in Materials and Methods. (C) Talin head and hZasp synergistically activate α5β1 integrin in CHO cells. Results represent mean ± s.e.m. (n≧3; **P<0.01 and ***P<0.001). Inset shows expression of transiently transfected proteins.

Fig. 2.

Drosophila Zasp and talin co-localize at myotendinous junctions. (A–C′) Zasp colocalizes with talin at myotendinous junctions. (A) Anti-Zasp antibody (red), (B) anti-talin antibody (green) and (C) merged image of stage 16 embryo. The boxed area in C is shown enlarged in A′–C′. Arrows indicate myotendinous junctions.

Fig. 3.

Integrins detach from the ECM in Zasp mutants. Absence of Zasp causes detachment of αPS2 integrin from the Drosophila ECM ligand, tiggrin. Comparison of Drosophila myotendinous junctions in lateral muscles of wild-type (wt) and Zasp mutant embryos. Triple staining with rat anti-αPS2 integrin (green), rabbit anti-talin (white), mouse anti-tiggrin (red), and merged images are shown. In wild-type embryos, tiggrin, αPS2 integrin, and talin colocalize tightly at myotendinous junctions. In contrast, in a Zasp mutant embryo, talin and αPS2 integrin (arrows) are partially separated from tiggrin.

Fig. 4.

Drosophila Zasp modulates integrin function. (A) FRAP of integrin–YFP at myotendinous junctions from wild-type and Zasp mutant embryos at different stages. Integrin mobility decreases only in wild type embryos during maturation of myotendinous junctions. (B) FRAP of integrin–YFP at myotendinous junctions from wild-type talin and talinR367A mutant stage 17 embryos. Integrin mobility is higher in talinR367A mutants. (C) Talin head, but not talin headR367A, can increase larval survival in Drosophila when expressed in muscles of Zasp mutants.

Fig. 5.

Human Zasp does not cooperate with full-length talin. CHO cells were co-transfected with GFP, GFP–talin head or full-length GFP–talin and DsRed or DsRed–HAZasp. Full-length talin and Zasp show similar levels of α5β1 integrin activation as full-length talin alone. Activation indices of doubly transfected cells (GFP and DsRed positive) were calculated and normalized for integrin expression. Results represent mean ± s.e.m. (n≧3; **P<0.01). Inset shows expression of transiently transfected proteins.

Fig. 6.

Human Zasp specifically activates α5β1 but not αIIbβ3 integrins and does not cooperate with kindlin in α5β1 integrin activation. CHO cells stably expressing αIIbβ3 integrin were co-transfected with GFP or GFP–talin head and DsRed, DsRed-HAZasp V1 or DsRed-kindlin-1 constructs. After detachment, cells from each transfection were incubated with either FN9-11 or PB1 antibody (α5β1 integrin activation) or with PAC-1 or D57 antibodies (αIIbβ3 integrin activation) with appropriate antagonists. Cells co-expressing similar amounts of GFP- and DsRed-tagged proteins were analyzed (see Materials and Methods). Activation indices were normalized for integrin expression. Results significantly different from DsRed and GFP–talin-head are indicated. Values are means ± s.e.m. (n≧3; *P<0.05, **P<0.01 and ***P<0.001). (A) Zasp V1, but not kindlin-1, cooperates with talin head to activate α5β1 integrins. FN9-11 binding in the presence or absence of either XP280 (αIIbβ3 antagonist) or 3F compound (α5β1 antagonist) was measured (see Materials and Methods). (B) Kindlin-1, but not Zasp, synergizes with talin head to activate αIIbβ3 integrins. Binding of the αIIbβ3-specific, ligand-mimetic, PAC-1 antibody was measured. Inset shows expression of transiently transfected proteins. (C) Zasp V1 does not cooperate with kindlin-1 to activate α5β1 integrins. Experimental procedure was conducted as in A. Inset shows expression of transiently transfected proteins.

Fig. 7.

Human Zasp binding to the β1 tail is not required for integrin activation and the activating function of Zasp is conserved between isoforms. (A) Zasp binding to the β1 tail is LIM-domain mediated. Pulldown assays using recombinant αIIb or β1 tail proteins were performed with transfected CHO cell lysates expressing full-length DsRed–HAZasp V1 or DsRed–HAZasp V1 fragments. Binding was quantified by densitometry and normalized to the lysate control (n≧3; **P<0.01, ***P<0.001). Upper right: schematic of the domain organization of Zasp fragments and isoforms. Lower right: representative pulldowns of overexpressed DsRed-HAZasp V1 constructs. Binding was assessed by western blotting (anti-HA tag). Loading of each tail protein was judged by protein staining. Loading control represents 2.5% of the starting lysate in the binding assay. (B,C) Zasp-mediated integrin activation requires Zasp V1 middle region (215–546) but does not rely on integrin binding. CHO cells were co-transfected with GFP or GFP–talin head and DsRed, DsRed–HAZasp V1 constructs. Activation indices of α5β1 integrin from cells co-expressing similar amounts of GFP- and DsRed-tagged proteins were calculated and normalized for integrin expression (see Materials and Methods). Values are means ± s.e.m. (n≧3). Results that are significantly different from DsRed and GFP–talin head (*P<0.05, ***P<0.001) are indicated. Insets show expression of transiently transfected proteins. (D) The skeletal-specific isoform Zasp V6 activates integrin. CHO cells were co-transfected with GFP or GFP–talin head and DsRed, DsRed–HAZasp V1 or V6 constructs. Experimental procedure was as in C. Inset shows expression of transiently transfected proteins.