Conservation and divergence of protein pathways in the vertebrate heart

Abstract

Heart disease is the leading cause of death in the western world. Attaining a mechanistic understanding of human heart development and homeostasis and the molecular basis of associated disease states relies on the use of animal models. Here, we present the cardiac proteomes of 4 model vertebrates with dual circulatory systems: the pig (Sus scrofa), the mouse (Mus musculus), and 2 frogs (Xenopus laevis and Xenopus tropicalis). Determination of which proteins and protein pathways are conserved and which have diverged within these species will aid in our ability to choose the appropriate models for determining protein function and to model human disease. We uncover mammalian- and amphibian-specific, as well as species-specific, enriched proteins and protein pathways. Among these, we find and validate an enrichment in cell-cycle-associated proteins within Xenopus laevis. To further investigate functional units within cardiac proteomes, we develop a computational approach to profile the abundance of protein complexes across species. Finally, we demonstrate the utility of these data sets for predicting appropriate model systems for studying given cardiac conditions by testing the role of Kielin/chordin-like protein (Kcp), a protein found as enriched in frog hearts compared to mammals. We establish that germ-line mutations in Kcp in Xenopus lead to valve defects and, ultimately, cardiac failure and death. Thus, integrating these findings with data on proteins responsible for cardiac disease should lead to the development of refined, species-specific models for protein function and disease states.

Fig 1. Workflow for investigating multispecies cardiac proteomes.

(A) Heart tissue from M. musculus, S. scrofa, X. laevis, and X. tropicalis was collected and subjected to differential protein extraction as detailed. The resulting 3 extracts were fractionated by SDS-PAGE. (B) Number of proteins identified in the 5 gel fractions for each extract (extract 1–3 from panel A) per species. These numbers include proteins found in 2 or more extracts. (C) PCA of MS1 proteome data demonstrates that the mammalian samples separate from each other and the Xenopus samples but the Xenopus samples maintain a tight association. (D) Data analysis workflow. (E) Identified proteins were mapped to human accession numbers using Blast2GO. (F) The shared proteins in all species represent a core cardiac proteome that is enriched for a variety of pathways. Shown here in a tree map are the top 10 most enriched (adjusted p ≤ 0.05) GO biological process terms; box size scales with enrichment significance of the terms. (G) Examples of proteins detected in subsets of the species analyzed. See S8 Table for numerical data underlying figure. GO, Gene Ontology; Kcp, Kielin/chordin-like protein; LC-MS/MS, liquid chromatography coupled with tandem mass spectrometry;MS1, precursor ion.

Fig 2. Evolutionary comparison of protein complexes.

(A) Proteins detected in this study were mapped to known human protein complexes listed in CORUM. (B) A complex score was calculated from the MS1 peak area of individual protein components of each complex. These complexes were then clustered into 6 clusters and analyzed further. (C–F) Individual clusters are plotted along with example protein complexes, demonstrating the underlying abundance data that drove the complex clustering. (G) Three representative complexes for each cluster are listed. See S9 Table for numerical data underlying figure. CORUM, Comprehensive Resource of Mammalian Protein Complexes.

Fig 3. GSEA reveals the increased relative abundance in cell-cycle proteins in Xenopus laevis.

(A) Pairwise GSEA was carried out between X. laevis and the 3 other species, revealing an enrichment (adjusted p ≤ 0.05) in cell cycle related proteins in X. laevis. (B) All proteins found in any enriched cell cycle related category by GSEA were clustered using k means (k = 4). Relative protein abundance is shown. The proteins found in cluster 3 were further analyzed in Cytoscape and grouped by function to show the interrelated nature of the proteins enriched in this cluster. See S13 Table and S14 Table for numerical data underlying figure. GSEA, Gene Set Enrichment Analyses.

Fig 4. Validation of cell-cycle enrichment in X. laevis using targeted MS.

(A) Workflow of target protein and peptide selection for PRM MS–based validation. (B) Mean protein abundance with SEM in each species by PRM assay is shown. The data were grouped by functional classes for visualization. For all proteins in (B), Xenopus laevis was significantly (p ≤ 0.05) higher than at least one other species. Shown to the right of each graph are examples of the most enriched proteins in each functional class (**p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001). See S16 Table for numerical data underlying figure. DDA, data dependent analysis; MS, mass spectrometry; PRM, parallel reaction monitoring.

Fig 5. Association of cardiac protein abundance and model system for selected proteins with known roles in human cardiac disease.

A diagram of detected proteins with links to human heart disease reported in the KEGG database was created in Cytoscape showing relative protein abundance data across the 4 species examined here, as well as known protein–protein interactions between these heart disease–related proteins. See S8 Table for numerical data underlying figure. AFib, atrial fibrillation; ARVC, arrhythmogenic right ventricular cardiomyopathy; AVSD, atrioventricular septal defect; BAV, bicuspid aortic valve; BRS, brugada syndrome; CPVT, catecholaminergic polymorphic ventricular tachycardia; DCM, dilated cardiomyopathy; HCM, hypertrophic cardiomyopathy; LQTS, long QT syndrome; LVNC, left ventricular noncompaction; MVP, mitral valve prolapse; SVAS, congenital supravalvar aortic stenosis.

Fig 6. Kcp is essential for heart development and survival in X. laevis.

(A) PRM quantification of peptides specific for Kcp confirms that Kcp is enriched in Xenopus. (B) Human and Xenopus show chromosomal synteny as revealed by Metazome. KCP is indicated in black. Upstream and downstream genes are colored as indicated. (C) Schematic of Kcp showing relative position of TALEN-induced mutations and bottom predicted protein generated by TALEN mutagenesis. (D) Stiol images of ultrasound Doppler of living wild type and Kcp^{Δexon2/ Δexon2} at Stage 64 positioned with dorsal top, ventral bottom of image. Blood flow shown in red. (E) Contrast-enhanced CT imaging of wild-type and Kcp^{Δexon2/ Δexon2} hearts (Stage 64) with AVV highlighted in white, bottom panels show lateral and posterior views showing thicker and misshapen AVV in Kcp^{Δexon2/ Δexon2} hearts verses control. (F) The AVV are detached from the muscular walls in Kcp mutants. Transverse sections through Masson trichrome stained Stage 64 control hearts (heterozygous) at low (upper left panel) and high (upper right panel) magnification focused on the AVV. Similar sections through a Kcp null froglet in at low (lower left panel) and high (lower right panel) magnification. Note in lower left panel an enlarged AVV in the outer valve leaflet detaching from the muscle wall. (G) Histology of wild-type and Kcp^{Δexon2/ Δexon2} hearts with Alcian Blue staining shows massive accumulation of collagen (blue) in AVV. a* Denotes position of AVV. (H) AVVs in kcp null hearts show a decrease in Filamin A and a concomitant increase in Fibrilin in the outer valve leaflet. (1–8) Immunochemistry of Cardiac-Actin:GFP (marking cardiomyocytes with GFP) froglets stained for Filamin (red), and DAPI (blue) in (1 and 2) heterozygous controls and (3 and 4) Kcp null. (5 and 6) Immunochemistry of Cardiac-Actin:GFP (marking cardiomyocytes with GFP) froglets stained for Fibrillin (red) and DAPI (blue) in (5 and 6) heterozygous controls and (7 and 8) Kcp null. See S16 Table for numerical data underlying figure. a, atria; avv, atrioventricular valve; GFP, green fluorescent protein; Kcp, Kielin/chordin-like protein; PRM, parallel reaction monitoring; TALEN, transcription activator-like effector nuclease; v, ventricle.