Novel obscurins mediate cardiomyocyte adhesion and size via the PI3K/AKT/mTOR signaling pathway

Abstract

The intercalated disc of cardiac muscle embodies a highly-ordered, multifunctional network, essential for the synchronous contraction of the heart. Over 200 known proteins localize to the intercalated disc. The challenge now lies in their characterization as it relates to the coupling of neighboring cells and whole heart function. Using molecular, biochemical and imaging techniques, we characterized for the first time two small obscurin isoforms, obscurin-40 and obscurin-80, which are enriched at distinct locations of the intercalated disc. Both proteins bind specifically and directly to select phospholipids via their pleckstrin homology (PH) domain. Overexpression of either isoform or the PH-domain in cardiomyocytes results in decreased cell adhesion and size via reduced activation of the PI3K/AKT/mTOR pathway that is intimately linked to cardiac hypertrophy. In addition, obscurin-80 and obscurin-40 are significantly reduced in acute (myocardial infarction) and chronic (pressure overload) murine cardiac-stress models underscoring their key role in maintaining cardiac homeostasis. Our novel findings implicate small obscurins in the maintenance of cardiomyocyte size and coupling, and the development of heart failure by antagonizing the PI3K/AKT/mTOR pathway.

Keywords: Intercalated disc; Obscurin; PI3K/AKT/mTOR; Phosphatidyl-inositol bisphosphates; Pleckstrin homology domain.

Fig. 1

Small obscurins are expressed in developing and mature hearts, localize at cardiomyocyte junctions, and are enriched in the ID fraction. (A) Domain architecture of canonical obscurin (obsc-A) and two novel variants (obsc-80 and obsc-40) illustrating their structural and signaling motifs. The epitopes for the obscurin antibodies used in this study are highlighted; please see key for notations. (B) Immunoblots of protein lysates prepared from developing and adult mouse hearts were stained for obscurins with the α-ObscCOOH antibody. GAPDH antibody staining of the same blot following stripping indicates equal loading of lysates. (B′) Quantification of the relative expression levels of obsc-80 and obsc-40 in developing and adult mouse myocardia (n = 5 hearts per age); the expression of obsc-80 and obsc-40 was normalized to that of GAPDH. The levels of obsc-80 plus obsc-40 in developing cardiac tissue were calculated as a percentage of the total amount in mature tissue, which was set at 100%. Although obsc-80 and obsc-40 are expressed in varying amounts throughout development and at maturity, their total levels remained unaltered. Student's t-test was performed to determine significance between the levels of obsc-80 and obsc-40 at different ages, p < 0.05. (C–E″) Transiently expressed GFP-obsc-80 (C) and GFP-obsc-40 (D), but not control GFP-protein (E), co-localize with N-cadherin (C′, D′, and E′; staining in red) at cell-cell junctions in primary RNCM. Scale bar 20 µm. (F–G) Subcellular fractionation of adult mouse whole heart (WH) lysates probed with α-ObscCOOH antibody for obsc-80 (F), obsc-40 (F), and giant obscurins (G). Results indicate that obsc-80 and obsc-40 are enriched in the ID fraction, compared to the cytosolic (C) and particulate (P) fractions; and giant obscurins are enriched in the cytosolic (C) fraction. N-cadherin, dystrophin, and GAPDH (F) were used as loading and purity controls for the ID, particulate, and cytosolic fractions, respectively.

Fig. 2

Small obscurins −80 and −40 localize to distinct ID subdomains. Cryosections of mouse hearts from embryonic day 17 (A–A″), postnatal day 3 (B–B″), and mature adult myocardium (C–E″) (n = 5 for each group) were sectioned and processed for immunofluorescent staining using α-ObscCOOH (green) and α-N-cadherin (red) antibodies; inset shows zoomed-in images of the boxed areas. Antibodies to the COOH-terminus of obscurins label the ID, coincident with N-cadherin at developmental and mature stages. (C–E″) Notably, we observed two staining patterns for obscurins at the mature ID: i. a broad band (D, arrow) coinciding with the ID membrane where N-cadherin (D′) also resides, and ii. a doublet flanking the ID membrane (C, arrow) on the edges of N-cadherin staining (D′). Scale bar 20 µm for A″–B″, 10 µm for a″, b″, C″, D″ and E″, and 5 µm for c″, d″ and e″. (F–H) Ultrathin cryo-sections of adult mouse cardiac muscle were labeled with antibodies specific for the COOH-terminus of obscurins, α-obscCOOH (F) and α-obscABD (G). Epitopes recognized by the α-obscCOOH antibody are located in close proximity to the ID membrane at the peaks of the ID folds (red arrows) and within the transition zone between the ID membrane and neighboring sarcomeres (yellow arrows). The α-obscABD antibody, however, only labels close to the ID membrane at the peaks of the ID folds (red arrows). Scale bar 500 nm. (H) The distance of each gold particle from the center of the ID membrane was measured. Measurements are represented as a histogram illustrating the bimodal and single Gaussian distributions of measurements for the α-obscCOOH and α-obscABD antibodies, respectively. The measurements for the α-obscCOOH antibody yielded populations with ~0.5 µm and ~0.02 µm mean distances from the ID membrane, whereas the measurements for the α-obscABD antibody yielded a single population with ~0.02 µm mean distance from the ID membrane (N= 2 hearts and n= 120 IDs per antibody).

Fig. 3

Obsc-40 and obsc-80 are in a complex with other ID proteins. Immunoprecipitates using adult mouse lysates and antibodies to desmoglein-2 (A), desmin (B), connexin43 (C), N-cadherin (D), ankyrinG (E), and vinculin (F) indicated that both obsc-80 and obsc-40 exist in complexes with those proteins. Representative blots from 3 independent experiments are shown.

Fig. 4

The obscurin PH-domain interacts with select PIP2s via residues R80 and R85. Lipid dot-blots using recombinant GST-ObscPH (A) and GST-protein (B) as control indicated that the PH-domain of obscurins interacts directly and strongly with PI(3,4)P₂ and PI(4,5)P₂ and weakly with PI(3,4,5)P₃, PI(3)P, PI(4)P, and PI(5)P. (C–D) Molecular modeling predicted two Possible Lipid Binding Sites (PLBS) on the surface of the obscurin PH-domain, PLBS 1 and PLBS 2. (E–F) Concurrent mutation of residues R40 and R41 within PLBS 1 to alanine (A) failed to alter the lipid binding capacity of the obscurin PH-domain. However, simultaneous mutation of R80 and R85 within PLBS 2 to alanine nearly eliminated binding to PI(3,4)P₂ and PI(4,5)P₂. (G–H) Single mutations of R80 or R85 to alanine within PLBS 2 resulted in loss of binding to PI(4,5) and PI(3,4), respectively. (I) The fold-change of the binding capacity of mutant R40A/R41A or R80A/R85A obscurin PH-domain relative to wild-type protein was calculated from 5 independent experiments. Student's t-test was performed to determine significance, p < 0.05 compared to wild-type control. (J–K) Dynamic molecular modeling was used to observe PLBS 2 with bound PI(3,4)P₂ or PI(4,5)P₂; dotted lines indicate electrostatic interactions.

Fig. 5

Overexpression of obsc-40, obsc-80, and obscurin PH-domain in cardiac cells leads to accumulation of PIP2s, and negatively regulates the PI3K/AKT/mTOR pathway. Obscurin GFP-PH-domain, GFP-obsc-40, GFP-obsc-80, and GFP-protein alone were transiently expressed in H9C2 cardiac-derived cells. (A) Protein lysates prepared from each cell transfection condition were probed with an α-GFP antibody, and showed sufficient overexpression of the indicated transgenes. Following insulin or DMSO vehicle treatment, lysates were prepared and either spotted on a membrane and probed with antibodies to PIP2, PIP3, and HSP90 (loading control) (B) or processed for immunoblotting with the indicated antibodies (C); Hsp90 served as loading control. Overexpression of the obscurin PH-domain, obsc-40, obsc-80, but not GFP alone, resulted in a significant increase in the amounts of PIP2s compared to PIP3 (B′), and a marked decrease in the phosphorylation levels of major components of the PI3K cascade (C′). The relative expression levels of the indicated proteins were calculated from 5 independent experiments. Two-way ANOVA was performed to determine significance, p < 0.05; *significant compared to GFP control transfected myocytes, ^#significant compared to DMEM corresponding controls. Dashed lines denote discontinuity in the data due to the different order that the samples were loaded in the gel and presented herein.

Fig. 6

Overexpression of obsc-40, obsc-80, and obscurin PH-domain in cardiac cells results in reduced cell adhesion and size. (A–A′) Confluent monolayers of primary RNCM after transient transfection with obscurin GFP-PH-domain, GFP-obsc-40, GFP-obsc-80, or GFP-protein were treated with vehicle DMSO or insulin, and subjected to a dispase mechanical dissociation assay. (A) Representative images of RNCM following dispase mechanical dissociation assay. (A′) Ectopic expression of obscurin PH-domain, obsc-40 and obsc-80 resulted in significantly increased fragmentation compared to control GFP-protein in both DMSO treated and insulin-stimulated cells. Each assay was done in triplicate, 5 independent times, and two-way ANOVA was used to determine significance, p < 0.05. (B) Representative overlay images of transmitted light and GFP-fluorescent transfected primary RNCM stained for α-actinin (red) overexpressing control GFP-protein, obscurin PH-domain, obsc-40, or obsc-80 subjected to DMSO vehicle or insulin stimulation (n= 5 independent repeats with 25 cells analyzed for each condition for each repeat). Cellular outline (marked in white) was used to measure cellular area. Scale bar 20 µm. (B′) Quantification of the cellular area indicated that overexpression of the obscurin PH-domain, obsc-40 or obsc-80, but not GFP-alone, led to significant decrease in cell size. Two-way ANOVA was performed to determine significance, p < 0.05, using measurements from five independent experiments; * indicates significance compared to GFP control transfected myocytes, ^# indicates significant compared to DMEM corresponding controls.

Fig. 7

Acute and chronic stress heart models exhibit decreased levels of obscurins and increased levels of active PI3k/AKT/mTOR. (A–A′) Lysates of adult mouse hearts subjected to MI, PO, or sham surgery (n = 5 per group) were analyzed via immunoblotting using antibodies to the COOH-terminus of obscurins. (A′) Quantitation of giant and small obscurins' relative expression in MI and PO cardiac models relative to sham controls demonstrated that the levels of obsc-80 and obsc-40, but not giant obsc-A and obsc-B, are significantly reduced in both stress models. (B–B′) The phosphorylation but not total levels of key members of the PI3K/AKT/mTOR cascade were significantly increased in MI and PO hearts compared to sham controls. (C–C′) A significant shift in the ratio of endogenous PIP2:PIP3 toward PIP3 was observed in MI- and PO-stressed hearts compared to controls. The relative expression levels of the indicated proteins were calculated from 5 independent experiments. One-way ANOVA was performed to determine significance, p < 0.05. HSP90 was used as loading and normalization control following stripping of the corresponding blot.

Fig. 8

Model of the obscurin PH-domain contributing to the regulation of the PI3K/AKT/mTOR pathway. (A) Baseline activation of the PI3K/AKT/mTOR pathway ensures a balanced level of endogenous PIP2s and PIP3s at the membrane. Once activated, phosphorylated PI3K (phosphorylation shown by a yellow sphere) converts PIP2 to PIP3 via the addition of a phosphate group. PIP3s then recruit PDK1 and AKT to the membrane where PDK1 phosphorylates and activates AKT. Activated AKT has a multitude of downstream targets, including mTOR that is a major regulator of cell growth and adhesion. (B) The obscurin-PH domain fine tunes this pathway by acting as a “sink” of PIP2s, limiting their availability for PI3K-mediated phosphorylation, and thus precluding the aberrant activation of the PI3K/AKT/mTOR pathway, which in turn results in regulated cell size and adhesion.