Paxillin and Focal Adhesion Kinase (FAK) Regulate Cardiac Contractility in the Zebrafish Heart

Abstract

An orchestrated interplay of adaptor and signaling proteins at mechano-sensitive sites is essential to maintain cardiac contractility and when defective leads to heart failure. We recently showed that Integrin-linked Kinase (ILK), ß-Parvin and PINCH form the IPP-complex to grant tuned Protein Kinase B (PKB) signaling in the heart. Loss of one of the IPP-complex components results in destabilization of the whole complex, defective PKB signaling and finally heart failure. Two components of IPP, ILK and ß-Parvin directly bind to Paxillin; however, the impact of this direct interaction on the maintenance of heart function is not known yet. Here, we show that targeted gene inactivation of Paxillin results in progressive decrease of cardiac contractility and heart failure in zebrafish without affecting IPP-complex stability and PKB phosphorylation. However, we found that Paxillin deficiency leads to the destabilization of its known binding partner Focal Adhesion Kinase (FAK) and vice versa resulting in degradation of Vinculin and thereby heart failure. Our findings highlight an essential role of Paxillin and FAK in controlling cardiac contractility via the recruitment of Vinculin to mechano-sensitive sites in cardiomyocytes.

Fig 1. Targeted knock-down of Paxillin leads to heart failure in zebrafish.

(A-C) Co-immunostaining of Paxillin (A, red) and the known Z-disk protein Focal Adhesion Kinase (FAK) (B, green) on adult primary zebrafish cardiomyocytes revealed localization of Paxillin to sarcomeric Z-disks. Cell nuclei were counterstained with DAPI (C, blue). Scale bar 10 μm. (D-G) Lateral view of paxillin start MO (MO1-paxillin) (D) and paxillin 5bp-mismatch-MO (MO1-control) (E) injected embryos at 72 hpf. (F) 78% of embryos injected with 5.4 ng MO1-paxillin developed pericardial edema and blood congestion at the cardiac inflow tract (n = 3; *P<0.0001), whereas control-injected embryos developed no pathological phenotype. (G) Fractional shortening (FS) measurements of Paxillin morphant ventricles at 48, 72 and 96 hpf. FS of Paxillin morphant ventricles was slightly reduced to 69.6% ± 5.07% compared to corresponding 5bp-mismatch-MO injected embryos (FS: 71.6% ± 3.96%) at 48 hpf. At 72 hpf, FS in Paxillin morphants was reduced to 49.29% ± 4.39% compared to control morphants (FS: 69.3% ± 4.36%), whereas ventricular chambers of Paxillin morphants became almost silent (FS: 4.23% ± 6.17%) compared to controls (FS: 74% ± 4.96%) at 96 hpf (n = 6–8 individuals per time point). (H-K) Lateral view of MO2-paxillin (H) and MO2-paxillin+paxillin mRNA (I) injected embryos at 72 hpf. (J) Co-injection of 4.5 ng of MO2-paxillin and 1 ng wild-type paxillin mRNA rescued the Paxillin morphant heart failure phenotype (n = 3; *P = 0.0073). (K) Quantification of ventricular FS revealed that ectopic expression of paxillin mRNA significantly improved contractile function in Paxillin morphants at 72 hpf (n = 8–9 individuals; *P<0.0001). Error bars indicate s.d.

Fig 2. Paxillin does not regulate IPP-complex stability and PKB activation in zebrafish.

(A) Western blot analysis of Paxillin morphant protein lysates compared to lysates obtained from control-injected embryos with antibodies against ILK and PINCH. For each sample 50 embryos were pooled and 20 μg of protein lysate were loaded per lane. The figure shows one representative western blot out of three independent experiments. (B) Western blot analysis of protein levels of total PKB and serine 473-phosphorylated PKB in MO1-paxillin-injected embryos. For each sample 50 embryos were pooled and 20 μg of protein lysate were loaded per lane. The figure shows one representative western blot out of three independent experiments.

Fig 3. Knock-down of fak1a and fak1b phenocopies the Paxillin morphant phenotype.

(A) Western blot analysis of FAK protein levels in Paxillin-deficient zebrafish embryos compared to control-injected embryos. For each sample 50 embryos were pooled and 20 μg of protein lysate were loaded per lane. The figure shows one representative western blot out of three independent experiments. (B, C) Quantitative real time PCR of Paxillin-depleted and control-injected embryos showed no significant (ns) alteration of fak1a (n = 3; *P = 0.9017) (B) and fak1b (n = 3; *P = 0.8878) (C) mRNA expression. For statistical analysis student’s t-test was performed. (D-F) Lateral view of MO1-fak1a/fak1b (D) and 5bp-mismatch-MOs (MO1-control) (E) injected embryos at 72 hpf. Embryos co-injected with 2.7 ng of MO1-fak1a and 2.15 ng of MO1-fak1b showed a severe heart failure phenotype whereas injection of the same amount of fak1a and fak1b 5bp-mismatch MOs (MO1-control) caused no significant pathological cardiac phenotype (F) (n = 3; *P<0.0001). (G) FS measurements of FAK morphant ventricles at 36, 48 and 72 hpf. FS of FAK double-knock-down morphant ventricles was slightly reduced to 56.38% ± 5.85% compared to corresponding 5bp-mismatch-MO injected embryos (FS: 62.4% ± 5.77%) at 36 hpf. At 48 hpf, FS in FAK morphants was reduced to 63.57% ± 9.18% compared to control morphants (70.33% ± 3.72%), whereas ventricular chambers of FAK-deficient embryos became almost silent (FS: 2.6% ± 3.71%) compared to controls (FS: 69.33% ± 4.36%) at 72 hpf (n = 5–8 individuals per time point). (H) Western blot analysis of Paxillin levels in control-, MO1-fak1a/fak1b- and MO1-paxillin-injected embryos. Paxillin protein levels are severely reduced after targeted knock-down of FAK and Paxillin, respectively. For each sample 50 embryos were pooled and 20 μg of protein lysate were loaded per lane. The figure shows one representative western blot out of three independent experiments. (I) Quantitative real time PCR of FAK-depleted and control-injected embryos showed no significant (ns) alteration of paxillin mRNA expression (n = 3; *P = 0.0937). For statistical analysis student’s t-test was performed. (J) Western blot analysis of embryos co-injected with zebrafish paxillin mRNA and MO2-paxillin compared with embryos injected with MO2-paxillin or paxillin mRNA alone. Ectopic expression of zebrafish paxillin mRNA was able to restore FAK protein expression. For each sample 50 embryos were pooled and 20 μg of protein lysate were loaded per lane. Error bars indicate s.d.

Fig 4. Vinculin protein levels are severely decreased in Paxillin- and FAK-deficient zebrafish embryos.

(A) Western blot analysis of MO-mediated Paxillin inactivation (MO1-paxillin) led to severely reduced Vinculin protein levels in vivo. Similar to loss of Paxillin, targeted ablation of FAK (MO1-fak1a/fak1b) also resulted in degradation of Vinculin, whereas Vinculin levels in control-injected embryos were completely unaffected. For each sample 50 embryos were pooled and 20 μg of protein lysate were loaded per lane. The figure shows one representative western blot out of three independent experiments. (B) Western blot analysis of embryos co-injected with zebrafish paxillin mRNA and MO2-paxillin compared with embryos injected with MO-paxillin or paxillin mRNA alone. Ectopic expression of zebrafish paxillin mRNA was able to restore Vinculin protein levels. For each sample 50 embryos were pooled and 20 μg of protein lysate were loaded per lane. (C, D) Bar graphs compare average of vinculin mRNA expression in Paxillin- (n = 3; *P = 0.1179) (C) and FAK-depleted (n = 3; *P = 0.8427) (D) compared to control-injected embryos. For statistical analysis student’s t-test was performed.

Fig 5. Vinculin does not localize to focal adhesion sites in Paxillin- and FAK-deficient zebrafish embryos.

(A-C) Co-immunostaining of dissected embryonic (72 hpf) zebrafish hearts with antibodies against β-Catenin (red) and Vinculin (green) of control- (A), MO1-fak1a/fak1b- (B) and MO1-paxillin- (C) injected hearts. Arrow heads indicate focal adhesion sites. Scale bar 5μM.