Fig 1.
The TBX20Avi-BirA system for isolation of the TBX20 interactome.
(A) Schematic diagram of ESC transduction with Lenti-BirA and subsequent differentiation. (B) RT-PCR panel showing changes in gene expression that ESCs undergo during 8-day differentiation into iCMs. (C) Immunohistochemistry of iCMs stained with cardiomyocyte marker Myosin heavy chain (MHC) and counterstained with DAPI. (D) Schematic of BirA-dependent biotinylation and streptavidin affinity isolation of TBX20 from iCMs. (E) Steptavidin affinity isolation of TBX20Avi complexes confirmed in BirA-expressing iCMs by western blot analysis.
Fig 2.
(A) Functional interaction network of the top 50 most enriched TBX20 interaction candidates. Known functional relationships are based on the STRING mouse database. Nodes are labeled with the protein’s gene symbol and color-coded based on its primary UniProt annotated cellular role. (B) Proteomic detection of TBX20 and CASZ1 in affinity purifications. The number of assigned spectra between control (-BirA) and BirA-expressing (+BirA) iCMs, unique peptides, percent amino acid sequence coverage, and respective molecular weights are provided. (C) Co-immunoprecipitations from Xenopus laevis embryos ectopically expressing full-length TBX20-EGFP alone, or in combination with full-length CASZ1-V5.
Table 1.
Lethality in mutant and control mice at 4, 8, and 16 weeks old.
Fig 3.
Tbx20flox/+; Casz1flox/+ compound heterozygotes display decreased cardiac function.
(A-D) Echocardiography in mice aged 8–11 weeks, (A) left ventricular ejection fraction, (B) fractional shortening, (C) left ventricular inner diameter at diastole (LIVD-D), and (D) left ventricular inner diameter at systole (LIVD-S) was measured for mice in indicated cohorts. Each red circle represents one mouse analyzed within that cohort. *p<0.05, **p<0.005, ***p<0.0005. Statistical significance between pairs was calculated using Student’s t-test.
Fig 4.
Double heterozygous hearts undergo pathological remodeling.
(A-D) Hearts from mice aged 8–11 weeks. (A) Transverse sections of hearts stained with hematoxylin and counter-stained with eosin. (B) Transverse sections of hearts stained with Picrosirius red and fast green, and visualized using bright field microscopy. (C) Polarizing light microscopy of Picrosirius red-stained sections. Thin collagen fibers stain green to yellow, while thicker collagen fibers stain orange to red. (D) Transverse heart sections were immunostained with tropomyosin (red, cardiomyocytes) and WGA (green, cell membranes). Region of left ventricular free wall shown. (E) Quantification of cardiomyocyte cross-sectional areas, shown as mean ± SEM of 450+ cardiomyocytes per heart, n = 3 hearts per genotype. *p<0.05, **p<0.0005. Statistical significance between pairs was calculated using Student’s t-test. A-C, scale bar = 300 μm. D, scale bar = 20 μm.
Table 2.
The TBX20-CASZ1 interaction is required for maintaining cardiac homeostasis in young adult mice (4–7 weeks old).
Table 3.
The TBX20-CASZ1 interaction is required for maintaining cardiac homeostasis in 8-11wk old mice.
Fig 5.
TBX20 and CASZ1 interact through their DNA binding domains.
(A) (Top) Schematic of full-length TBX20 and truncations. NLS, Nuclear Localization Signal. Putative activation domain shown in green and putative repression domain in orange. (Bottom) Co-immunoisolations from X. laevis embryos expressing full-length CASZ1-V5 and either full-length HA-TBX20 or deletions shown in top panel. (B) (Top) Schematic of full-length CASZ1 and the truncations used in the co-immunoisolations. (Bottom) Co-immunoisolations from X. laevis embryos expressing full-length HA-TBX20 alone, or in combination with either full-length CASZ1-V5 or truncations shown in top panel. (C) (Top) Sequence alignment of TBX20 position 248–288 across 90 TBX20 orthologs. Height of letters is relative to conservation at that residue. (Bottom) Sequence alignment of CASZ1 at positions 601–650 across 90 CASZ1 orthologs.
Fig 6.
TBX20F245I mutant displays impaired interaction with CASZ1.
(A) (Top) Schematic of full-length TBX20 with location of F256I mutation shown. (Bottom) Co-immunoisolations of full-length wild-type CASZ1 with wild-type TBX20 or TBX20F256I. (B) Ribbon models of the average structures of TBX20 calculated from the 100 ns molecular dynamics simulations of the T-box domain. Left panel: starting structure of wild-type TBX20 (cyan) including DNA for reference. Right panel: an overlay of wild-type TBX20 (green) and TBX20F256I (magenta). F256 and I256 side chains are displayed in stick form. Regions designated as 1, 2, and 3 undergo mutation-induced conformational changes in the unbound form. (C) Enlargement of the F256I residue shown in (B) in a different side chain rotamer that also induces steric clashes. The I256 mutation is rendered in purple stick form, with small red discs indicating steric clashes between the side chain of I256 and T-box residues E258 and T259.
Fig 7.
Proteomic analysis of DCM mouse hearts reveals activation of complement cascades and decreased protein abundance involved in glycogen metabolic processes.
(A-B) Comparison of normalized protein abundances (black dots, n = 3164) quantified by TMT-based proteomics between biological replicates of (A) Nkx2.5Cre (Control, CTL) and (B) Tbx20flox/+; Casz1flox/+; Nkx2.5Cre (Mutant) mouse hearts. (C) Comparison of average normalized TMT protein abundances between Mutant and CTL mouse hearts (n = 2). Proteins classified as differentially abundant in the Mutant (TMT abundance ratios ≥ ±1/3-fold in both replicates) are indicated (red dots, n = 175). (D) Functional interaction network of differentially abundant proteins constructed using STRING interaction database. Proteins with up- and down-regulated abundances are represented by circle and square nodes, respectively, and labeled with primary gene symbols. Proteins that did not have connectivity to at least one other protein were excluded. Nodes are color-coded by cluster connectivity, which was assigned by the ReactomeF1 plug-in (see Methods). (See S7 Fig) Network edges reflect known STRING database relationships (confidence score >0.4). Edge thickness correlates with STRING confidence score (0.4–1). (E) Gene ontology (GO)-based network of selected over-represented Biological Processes represented by the differentially abundant proteins. Over-represented GO terms (p-value <0.05) are colored according to the proportion of the total differentially abundant proteins annotated to that term that were up- (yellow) or down- (blue) regulated. Protein to GO term assignments are indicated by network edges connecting the respective gene symbols (squares) with their assigned GO term.