FIGURE SUMMARY
Title

A systems genetics approach identifies roles for proteasome factors in heart development and congenital heart defects

Authors
Farr, G.H., Reid, W., Hasegawa, E.H., Azzam, A., Young, I., Li, M.L., Olson, A.K., Beier, D.R., Maves, L.
Source
Full text @ PLoS Genet.

Identifying new candidate CHD genes.

A. Correlation of high shet genes with genes exhibiting damaging de novo variants in CHD cases (black bars; P = 1.09E-19). No correlation is observed between shet deciles and damaging de novo variants observed in controls (grey bars; P = 0.7). shet deciles are from Cassa et al.’s Supplementary Table 1 [23]. de novo variant data are from Jin et al.’s Supplementary Data Set 9 (CHD cases) and Supplementary Data Set 10 (controls) [18]. B. Correlation of high shet genes with known CHD genes, defined by [22] (S1 Table). shet deciles are from [23]. C. Venn diagram showing overlap of genes with high shet values in pink (3190 genes; S2 Table; top 2 deciles in dataset from [23]) and genes expressed in the mouse cardiac muscle cell lineage in green (1119 genes; S3 Table; datasets from [33] and [34]). The intersection is 245 candidate genes for CHDs (S4 Table). D. Venn diagram showing overlap of our Candidate CHD gene set in maroon (245 genes; S4 Table) and known CHD genes in purple (132 genes; S1 Table; [22]). The intersection is 11 genes, shown in the box. E. Gene Ontology (GO) term enrichment for candidate gene set, obtained using DAVID. The number of genes (out of 245) represented by each term is shown on top of each bar. Bonferroni-adjusted P values are shown.

PPI networks reveal connections between known and candidate CHD genes.

A. A Protein-Protein Interaction (PPI) network showing the evidence-based interactions among the proteins encoded by our 245 candidate CHD genes (yellow halos; S4 Table), 132 known CHD genes (green halos; S1 Table), and candidate genes that are also known CHD genes (blue halos, 11 genes from Fig 1D). Interactions between proteins/nodes are indicated by a single grey line, whose thickness corresponds to the strength of the supporting data. This network consists of 365 proteins/nodes, contains 980 edges, and is associated with a PPI enrichment P value < 1.0E-16. Purple nodes (Proteasome cluster genes) indicate the 8 factors (NF1, ODC1, POMP, PSMA6, PSMA7, PSMD3, PSMD6, and UBQLN2;) identified in the “Regulation of ornithine decarboxylase, and proteasome assembly” local cluster analysis (S7 Table). B. PPI subnetwork showing the interactions among the Proteasome cluster genes (purple nodes) and their direct connections from the larger PPI network in A. The node color code scheme is the same as in A. The lines between nodes are color coded to indicate the type of evidence supporting the protein-protein interaction. This subnetwork consists of 23 nodes, contains 58 edges, and has a PPI enrichment P value < 1.0E-16.

G0 CRISPR targeting of Proteasome cluster genes leads to heart defects in zebrafish embryos and larvae.

A.-C.hand2 G0 CRISPR leads to cardia bifida. A. 2 dpf control embryo showing myl7:EGFP heart. V, ventricle. A, atrium. Scale bar = 200 μm. B. 2 dpf hand2 G0 CRISPR embryo showing two domains (cardia bifida) of myl7:EGFP (arrows). C. Graph showing frequencies of different heart defect categories observed in G0 CRISPR embryos. Cardia bifida and heart looping defects were scored at 2 dpf. Hearts were again scored at 4 dpf for defects arising between 2 dpf and 4 dpf and were divided into two classes: lateral chamber arrangement and elongated chamber arrangement. Examples of the elongated chamber arrangement are illustrated in panels Q, S, T and V, and an example of the lateral chamber arrangement is shown in panel U. D. Graph showing frequencies of embryos with heart defects (at 4 dpf; all categories combined). Each dot represents an experimental replicate batch of embryos. Bars represent mean +/- SEM. N = 4-6 replicates per gene of interest; N = 24 control replicates. Each replicate batch consists of 7-75 embryos (mean = 36±16). The total number of embryos scored is given above the x-axis. These total numbers also apply to the corresponding columns in Panel C. ** P < 0.01. **** P < 0.0001. E.-O. Lateral views of 4 dpf larvae, showing representative G0 CRISPR phenotypes. The bubble (asterisk in E) is the swim bladder, an indicator of healthy larvae [40]. The arrowhead in I points to severe heart cavity swelling/edema. Scale bar = 500 μm. P.-V. Ventral views of hearts labeled with myl7:EGFP in 4 dpf larvae with pigment inhibited. Representative G0 CRISPR heart phenotypes are shown for each gene. Scale bar = 200 μm. V, ventricle. A, atrium.

pomp and psmd6 mutant zebrafish larvae show cardiac and extracardiac phenotypes.

A. Schematic illustrating pomp gene structure and CRISPR deletion. pomp exons are shown as boxes and the start codon is labeled. Guide RNA target sites are labeled in green and orange. DNA sequencing of the pompscm41 strain showed the pomp gene sequence between the two guide RNA sites has been deleted. B. Schematic illustrating psmd6 gene structure and CRISPR deletion. psmd6 exons are shown as boxes and the start codon is labeled. Guide RNA target sites are labeled in blue and red. DNA sequencing of the psmd6scm40 strain showed the psmd6 gene sequence between the two guide RNA sites has been deleted. C.-D. qRT-PCR analysis of (C) pomp and (D) psmd6 expression in 4 dpf phenotyped pomp and psmd6 control (presumed +/+ and +/-) or mutant (mut, presumed -/-) larvae. Each dot represents a replicate batch of embryos. Bars represent mean +/- SEM. N = 5 (control) or 4 (mutant) replicates. Each replicate consists of 10-20 embryos. **** P < 0.0001. E. Western analysis using anti-Ubiquitin and anti-beta Tubulin (as a loading control). Larvae from pomp and psmd6 clutches were phenotyped as control (presumed +/+ and +/-) or mutant (presumed -/-) larvae at 4 dpf and collected for lysates. Lysates were prepared from 3 replicate pools of animals for each condition, with n = 20 animals per replicate. Larvae from bortezomib or DMSO control treatments were collected at 4 dpf, with n = 20 animals per lysate. Molecular weight markers are shown. F.-K. Images of live 4 dpf larvae. The asterisk in F marks the swim bladder. Arrowheads in H, K point to heart cavity swelling/edema. Black arrows in H, K point to reduced craniofacial structures. White arrows in H, K point to areas of blood pooling. Embryos were genotyped. N = 3 pomp+/+, 12 pomp+/-, 9 pomp-/-, 4 psmd6+/+, 8 psmd6+/-, 6 psmd6-/-. Scale bars = 500 μm. L.-Q. Ventral views of hearts labeled with myl7:EGFP in live 4 dpf larvae with pigment inhibited. R.-S. Ventral views of hearts labeled with myl7:EGFP in formaldehyde-fixed 4 dpf larvae with pigment bleached. Scale bars = 100 μm. V, ventricle. A, atrium. T. Graph showing frequencies of heart defect categories observed in 4 dpf larvae. The total number of larvae scored is given at the base of the columns. Embryos were genotyped. U. Survival plot of two independent clutches each of pomp and psmd6. Dead embryos were collected and fixed over the course of the experiment; all embryos were genotyped at the conclusion of the experiment. Control (+/+ and +/-) curves are shown for only one clutch of each line; similar results were seen with the second clutch of controls for each line. Numbers of animals per clutch: pomp clutch 1: 40 +/+ , 100 +/-, 52 -/-; pomp clutch 2: 44 +/+ , 93 +/-, 61 -/-; psmd6 clutch 1: 60 +/+ , 107 +/-, 55 -/-; psmd6 clutch 2: 27 +/+ , 61 +/-, 22 -/-.

Early heart development appears normal in pomp and psmd6 mutants.

A.-C. Dorsal views of ALPM labeled for myl7 in blue in 18 hpf embryos. Normal heart morphology: 6/6 pomp+/+, 7/7 psmd6+/+, 11/11 pomp-/-, 8/8 psmd6-/-. Scale bar = 100 μm. D.-F. Dorsal views of hearts labeled for myl7 in blue in 24 hpf larvae. Normal heart morphology: 8/8 pomp+/+, 8/8 psmd6+/+, 6/6 pomp-/-, 10/10 psmd6-/-. Scale bar = 100 μm. G.-I. Ventral views of hearts labeled with myl7:EGFP in 48 hpf embryos. Normal heart morphology: 13/16 pomp+/+, 11/12 psmd6+/+, 14/14 pomp-/-, 11/12 psmd6-/-. Scale bar = 50 μm. V, ventricle. A, atrium. J.-L. Ventral views of hearts labeled with myl7:EGFP in 72 hpf larvae. Normal heart morphology: 35/38 pomp+/+, 22/24 psmd6+/+, 39/45 pomp-/-, 25/29 psmd6-/-. All embryos were genotyped. Scale bar = 50 μm. V, ventricle. A, atrium.

pomp and psmd6 mutants exhibit myocardial cell blebbing.

A.-C. Ventral views of hearts as maximum intensity projections (MIPs) of the entire myocardium, labeled with myl7:EGFP in 4 dpf larvae. Scale bar = 100 μm. V, ventricle. A, atrium. D.-F. Ventral views of hearts as MIPs of 2-7 slices (1.6-5.6 μm total section thickness) through the ventral ventricular wall, labeled with myl7:EGFP in 4 dpf larvae. Arrows in E and F indicate examples of rounded cardiomyocytes in the interior wall of the ventricle. Hearts in D - F are from different larvae than those in A - C. Presence of rounded trabecular cardiomyocytes: 0/17 pomp+/+, 1/15 psmd6+/+, 25/29 pomp-/-, 24/27 psmd6-/-. All embryos were genotyped. Scale bar = 50 μm. G. Graph showing numbers of protruding rounded cardiomyocytes per heart at 4 dpf. Only cells clearly projecting outward from the surface of the heart were counted. Each dot represents a heart/larva. Bars represent mean +/- SD. N = 9-13 per genotype. **** P < 0.0001.

Cardiac function deficits are observed by 4 dpf in pomp and psmd6 mutants.

A.-F. Brightfield views of hearts taken from videos of 3 dpf and 4 dpf larvae. A’-F’ panels show ventricles outlined in red. Arrowheads in E’ and F’ point to edema. Scale bars = 100 μm. G.-I. Graphs showing ventricular area, heart rate, and ventricular fractional area change (FAC) measurements at 3 dpf. Each dot represents the average of 3 measurements from a single heart/larva. Bars represent mean +/- SEM. Animals were genotyped after video recordings. N = 13 for pomp control (+/+). N = 7 for psmd6 control (+/+). N = 7 for pomp-/-. N = 8 for psmd6-/-. * P < 0.05. J.-L. Graphs showing ventricular area, heart rate, and ventricular FAC measurements at 4 dpf. Each dot represents the average of 3 measurements from a single heart/larva. Bars represent mean +/- SEM. Animals were genotyped after video recordings. N = 12 for all conditions: pomp control (+/+ and +/-), psmd6 control (+/+ and +/-), pomp-/-, and psmd6-/-. ** P < 0.01; **** P < 0.0001.

Myocardial sarcomere and myofibril formation appear normal in pomp and psmd6 mutants.

A.-C. Ventral views of ventricles labeled with α-actinin (A-C) in 4 dpf larvae. Scale bar = 25 μm. D.-E. Magnified views of α-actinin stain. Encircled green lines illustrate sarcomere lengths, with corresponding measurements shown. Encircled magenta lines illustrate myofibril widths, with corresponding measurements shown. Scale bar = 10 μm. F.-G. Graphs showing (F) sarcomere length and (G) myofibril width measurements at 4 dpf. Each dot represents the average of 9-14 measurements (one per myofiber) from a single heart/larva. Bars represent mean +/- SEM. All animals were genotyped. N = 7 pomp+/+ and psmd6+/+. N = 8 pomp-/- and psmd6-/-. No significant differences were observed between wild-type and mutant hearts.

pomp and psmd6 mutants exhibit reduced outflow tracts.

A.-F. Ventral views of outflow tracts, labeled with elnb (green), in (A-C) 72 hpf larvae and (D-F) 96 hpf larvae. Myocardium is labeled with myl7 (purple). G.-H. Graphs showing (G) the internal width of the elnb domain, measured at the maximum width position, and (H) the length of the elnb domain. Each dot represents an individual heart/larva. Bars represent mean +/- SD. All animals were genotyped. N = 7-9 per condition. * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001. Scale bar = 25 μm.

Acknowledgments
This image is the copyrighted work of the attributed author or publisher, and ZFIN has permission only to display this image to its users. Additional permissions should be obtained from the applicable author or publisher of the image. Full text @ PLoS Genet.