Global genetic analysis in mice unveils central role for cilia in congenital heart disease

Abstract

Congenital heart disease (CHD) is the most prevalent birth defect, affecting nearly 1% of live births¹, but the incidence of CHD is up to ten fold higher in human fetuses^2,3. A genetic contribution is strongly suggested by the association of CHD with chromosome abnormalities and high recurrence risk⁴. Here we report findings from a recessive forward genetic screen in fetal mice, showing the cilium and cilia transduced cell signaling play important roles in the pathogenesis of CHD. The cilium is an evolutionarily conserved organelle projecting from the cell surface with essential roles in diverse cellular processes. Using echocardiography, we ultrasound scanned 87,355 chemically mutagenized C57BL/6J fetal mice and recovered 218 CHD mouse models. Whole exome sequencing identified 91 recessive CHD mutations in 61 genes. This included 34 cilia-related genes, 16 genes involved in cilia transduced cell signaling, and 10 genes regulating vesicular trafficking, a pathway important for ciliogenesis and cell signaling. Surprisingly, many CHD genes encoded interacting proteins, suggesting an interactome protein network may provide a larger genomic context for CHD pathogenesis. These findings provide novel insights into the potential Mendelian genetic contribution to CHD in the fetal population, a segment of the human population not well studied. We note pathways identified show overlap with CHD candidate genes recovered in CHD patients⁵, suggesting they may have relevance to the more complex genetics of CHD overall. These CHD mouse models and >8,000 incidental mutations are sperm archived, creating a rich public resource for human disease modeling.

Inbred C57BL/6J mice mutagenized with ethylnitrosourea (ENU) were bred to recover recessive coding mutations (Extended Data Fig. 1). Phenotyping was conducted using noninvasive fetal echocardiography, an ultrasound imaging modality also employed clinically for CHD diagnosis. This unbiased study design allows detection of even those CHD mutants inviable to term⁶, a segment of CHD cases that would be largely clinically inaccessible for genetic analysis. Using combination of color flow, 2D, spectral Doppler and M-mode imaging, a wide spectrum of CHD phenotypes were recovered, including cardiac looping defects, OFT malalignment/septation defects, ventricular/atrioventricular septal defects, and various right/left heart obstructive lesions⁶ (Figure 1; Supplemental Movies S1-S3). The CHD diagnoses were confirmed with microCT/microMRI, necropsy, and histopathology (Extended Data Fig. 1). Heritability was ascertained with recovery of two or more G3 fetuses with the same CHD phenotype (n=4.6±2.6). Ultrasound scanning of 87,355 G3 fetuses from 2,651 mutant lines (defined by the G1 sire) recovered 218 CHD mutant lines. These were curated in the Mouse Genome Informatics Database (http://www.informatics.jax.org/searchtool/Search.do?query=b2b&submit=Quick%0D%0ASearch), with G1 sperm cryopreserved in the Jackson Laboratory (JAX^®Mice).

Figure 1. Ultrasound diagnoses of CHD and cilia defects in CHD mutants.

Vevo 2100 color flow imaging showed criss-crossing of blood flow indicating normal aorta (Ao) and pulmonary artery (PA) alignment (a, Supplemental Movie S1), confirmed by histopathology (b). E16.5 mutant (line b2b327) exhibited blood flow pattern indicating single great artery (PA) and ventricular septal defect (VSD) (c, Supplemental Movie S2), suggesting aortic atresia with VSD, confirmed by histopathology (d). Color flow imaging of E15.5 mutant (line b2b2025) with heterotaxy (stomach on right; Supplemental Movie S3c) had side-by-side Ao/PA with Ao emerging from right ventricle (RV), indicating DORV/VSD (e,f,Supplemental Movie S3a), and presence of AVSD (g,h,Supplemental Movie S3b,S3c). Histopathology also showed bicuspid aortic valve (BAV,i), interrupted aortic arch (IAA,j), and common AV valve (k). (l-n). Cc2d2a mutant exhibits dextrocardia with ventricular inversion (dextroversion) (m), and AVSD (l) with malformed AV cushions (n), but normal outflow cushions. (o-x). Confocal imaging of E12.5 Cc2d2a mutant (m/m) vs. wildtype (+/+) embryo sections showed no cilia in AV cushion (o,p), but normal ciliation in outflow cushion (q,r). Fewer and shorter cilia were observed in other mutant embryo tissues (s-x). Red:acetylated tubulin;green:IFT88.

Many mutant lines (30%) exhibited complex CHD associated with heterotaxy, the randomization of left-right patterning. These heterotaxy mutants often died prenatally. We note clinical studies show 16% of human fetuses with complex CHD have heterotaxy⁷, and these also exhibit high prenatal/intrauterine death (30-60%)^8,9. Also commonly observed in mutant lines with or without laterality defects are double outlet right ventricle (DORV) and atrioventricular septal defects (AVSD), phenotypes that usually resulted in prenatal lethality. We note DORV and AVSD are also observed with high prevalence in aborted/stillborn human fetuses³. Most heterotaxy mutant lines exhibited three distinct visceral organ situs phenotypes: normal situs solitus, mirror symmetric situs inversus, or randomization with heterotaxy, with complex CHD observed only with heterotaxy (Extended Data Fig. 2). These three phenotypes are found in patients and mice with mutations causing primary ciliary dyskinesia (PCD)^10,11, a sinopulmonary disease arising from immotile/dyskinetic respiratory cilia. This reflects the common requirement for motile cilia in airway clearance and left-right patterning^11,12. Hence, we conducted videomicroscopy of the tracheal epithelia to assess cilia motility as a proxy for the node.

Mutation recovery was conducted with whole exome sequencing analysis of a single mutant from each of 113 mutant lines. Cumulatively this yielded 7,996 coding mutations in 4,809 genes (Extended Data Fig. 3a). Genotyping analysis identified the pathogenic mutation as one consistently homozygous across all G3 mutants bearing the same phenotype. In mutant lines with multiple homozygous mutations on the same chromosome linked to the CHD phenotype, the pathogenic mutation was identified with segregation analysis of additional mutants, and/or based on phenotype comparison to existing knockout/mutant alleles. For five lines (b2b002/b2b012/b2b284/b2b1519/b2b1146), the mutation was confirmed with mapping analyses¹³.

In this manner, 91 pathogenic mutations were recovered in 61 genes (Supplemental Spreadsheet 1). The distribution of missense/splicing/nonsense mutations were similar to other ENU mutagenesis studies¹⁴ (Extended Data Fig. 3a). All mutations except one were in highly conserved amino acids (Supplemental Spreadsheet 2). Five of the 19 splicing mutations are >2-bases from the canonical splice donor-acceptor sites, the farthest being 10 bases apart (Dnah5^{c.13329-10T>A};b2b1537; validated by cDNA sequencing, Extended Data Fig. 4). Multiple alleles were recovered in 15 genes (Extended Data Table 1). Of 61 CHD genes recovered, 27 were not previously annotated to cause CHD (Extended Data Fig. 3b), either because no mutant/knockout mice were available, or the CHD phenotype was not investigated (Supplemental Spreadsheet 1). In some instances, the existing mutants died early, precluding CHD analysis, such as the Smarca4 knockout mouse with peri-implantation lethality¹⁵. A Hectd1 mutant was previously described only with neural tube defects¹⁶ versus our Hectd1^b2b327 mutant shown to have aortic atresia (Figure 1c,d). The identical Adamts6 mutation was found in 3 lines, shown to be a spontaneous mutation in the C57BL6/J support colony (Supplemental Spreadsheet 1).

Thirty-four of the 61 CHD genes are cilia-related (Figure 2), 23 in laterality mutant lines (Extended Data Table 2, Fig.3a,5). Twelve are required for motile cilia function, with 8 clinically known to cause PCD and 3 likely novel PCD genes (Ccdc151,Daw,Pkd1l1) (Extended Data Tables 2,3; Supplemental Spread Sheet 1). A mutation was recovered in Foxj1, a transcription factor known to regulate ciliogenesis. While Foxj1 knockout mice have no cilia in the airway¹⁷, our Foxj1^b2b774 mutant harboring a missense mutation in the DNA binding domain has ciliated trachea with robust ciliary motion (Supplemental Movie S4), but with no nodal flow (Supplemental Movie S5).

Figure 2. Congenital heart disease genes recovered from mouse mutagenesis screen.

Diagrams illustrate biological context of CHD gene function (color-highlighting indicate CHD genes recovered; asterisk denote CHD genes recovered from previous screen²⁷). For clarity, ciliome genes (Dctn5,Fuz) with endocytic function and/or involved in cilia-transduced signaling were not shown in the ciliogenesis panel.

Mutations in 22 ciliome genes affect primary cilia - 11 in laterality and 11 in nonlaterality mutant lines (Extended Data Table 2, Fig. 3a,5). Among nonlaterality mutants are Cep110 - centrosome component, Jbts17 - cilia transition zone protein, and four other genes (Fuz,Myh10,Dctn5,Lrp2) that also play important roles in vesicular/endocytic trafficking, a biological process essential for ciliogenesis^18,19 (Figure 2). Many genes affecting laterality are associated with non-PCD ciliopathies (Anks6,Bicc1,Tbc1d32, Cep290,Dync2h1,Ift74,Ift140,Nek8,Pcsk5,Pskh1,Cc2d2a,Tmem67; Extended Data Table 3)^20,21. Analysis of Cc2d2a mutant embryo by confocal imaging showed shorter and fewer cilia throughout the embryo (Figure 1s-x). In Cc2d2a mutant embryo, while cilia were abundant in outflow cushions, they were absent in AV cushions (Figure 1p,r). Consistent with this, Cc2d2a mutant embryos die prenatally with AVSD (Figure 1l,n), but with normal semilunar valves (Figure 1m,n).

Of the 26 non-ciliome CHD genes, 15 are known to mediate cell signaling, including SHH (Fuz,Kif7,Lrp2,Sufu,Tbc1d32), WNT/planar cell polarity (Ptk7,Prickle 1,Fuz), TGF-p/BMP (Cfc1,Ltbp1,Megf8, Pcsk5,Smad6,Tab1), and calcium signaling (Pkd1l1,Pkd1) (Figure 2). Five of these were recovered in the laterality mutant lines (Tbc1d32,Cfc1,Megf8,Pcsk5,Pkd1l1). These pathways are all transduced or modulated by the cilium, and play significant roles in cardiovascular development and/or left-right patterning.

We also recovered 10 CHD genes mediating vesicular/endocytic trafficking, including four (Lrp2,Myh10,Dctn5mFuz) that are ciliome related (Figure 2). The enrichment for endocytic genes may be mechanistically linked to cilia, as vesicular trafficking at the cilia base has important roles in both ciliogenesis and cilia transduced cell signaling^18,19,22. These endocytic genes caused a similar spectrum of OF malalignment and septation defects (Extended Data Fig. 6). Only Ap1b1, an adaptin subunit mediating endocytic protein sorting in the Golgi was observed to cause laterality defects (b2b1660; Extended Data Fig. 2). While Ap1b1 mutants showed normal airway and embryonic node cilia motility (Supplemental Movie S6,S7), Ap1b1 mutant embryonic fibroblasts exhibited ciliogenesis defect (65% vs. 85% ciliation in Ap1b1 mutant versus wildtype fibroblasts;p=5.90×10^-6). Ap2b1, an adaptin protein complex 2 subunit homologous to Ap1b1, functions in clathrin mediated endocytosis. Snx17, known to regulate endocytic receptor recycling, has a mutation in the cargo binding FERM domain, region known to interact with the NPXY motif of endocytic protein LRP1, which was also recovered in our screen (Extended Data Fig.3b).

Other CHD genes recovered included a novel zinc finger protein Zbtb14, and two genes modulating chromatin, Smarca4 and Prdm1. Also recovered were six extracellular matrix related genes (Acan,Adamts6, Frem2,Lox,Mmp21,Ndst1), one (Lox) identified as cilia-related, three novel genes (Cml5,Gm572,Mmp21) causing CHD with laterality defects, and two genes (Acan, Pkd1) causing biventricular hypertrophy without other structural heart defects.

Many of the CHD genes encode proteins known to physically interact or are in the same multiprotein complexes. For example, ANKS6 can bind BICC1 and also NEK8, with NEK8/ANKS6 colocalizing in the cilia inversin compartment^24,25. TMEM67, CEP290, JBTS17, and CC2D2A are colocalized in the NPHP1-4-8 multiprotein complex in the cilia transition zone²³. DNAH5, DNAH11, DNAI1 and DAW1 are associated with the outer dynein arm complex of motile cilia, while IFT140 and IFT74 are intraflagellar transport protein components that together with DYNC2H1 mediate protein transport in the cilium (Figure 2)^18,20. SNX17 can directly bind LRP1 to promote LRP1 cell surface recycling²⁶. These findings prompted an examination of the 113 exomes for the prevalence of additional mutations in the 61 CHD genes. This showed 58 additional incidental mutations in 32 of the CHD genes, a significant enrichment compared to random simulations (P=0.00114;Online Methods), suggesting possible modifier effects.

To further assess for potential protein-protein interactions (PPI), we constructed an interactome of CHD genes using known PPIs in the public databases and our own computationally predicted PPIs (Online Methods). This yielded an interactome of 778 genes, 292 are predicted PPIs (Supplemental Spreadsheet 3;Figure 3a). The average shortest distance between CHD genes is 4.3±2.5 edges, whereas random genes of matched degree distribution had a larger distance of 5.7±4.4 edges (P=0.01). Statistical analysis of the interactome revealed significant Go Ontology term enrichment related to development, including heart development terms (Supplemental Spreadsheet 3;Figure 3b). Pathways identified included “clathrin-mediated endocytosis signaling”, “TGFβ signaling”, “BMP signaling” and “axon pathfinding” (Supplemental Spreadsheet 3). Several CHD genes recovered are known to regulate axon guidance and neurogenesis, including Plxnd1, Robo1, and Cxcr4, while Pde2a regulates synaptic plasticity and long term potentiation, and Myh10 is required for synaptic membrane recycling (Extended Data Fig.7).

Figure 3. Interactome network of CHD genes with known and predicted interactors.

An interactome network was constructed (a) comprising the 61 CHD genes (magenta squares) and their known (blue edges) and predicted (red edges) interactions. The CHD interactome yielded significant enrichment of GO terms related to developmental processes. Circle size proportional to number of genes associated with GO terms.

Overall, our study provides compelling evidence that the cilium plays a critical role in CHD pathogenesis. The observed enrichment for cilia and cilia-related genes emerged from a phenotype driven forward genetic screen agnostic to the specific genes that may be recovered. The finding of ciliome mutations in CHD mutant lines with and without laterality defects would suggest a broad role for cilia in CHD pathogenesis. While our study is focused on monogenic recessive coding mutations, we noted striking overlap of genes recovered in the mouse screen versus heterozygous de novo coding mutations in CHD candidate genes identified in CHD patients⁵. Of 28 CHD genes identified in the human study - 11 are cilia-related, 7 are in TGFb/SHH/WNT signaling, and 2 in vesicular/endocytic trafficking (Supplemental Spread Sheet 4). Heterozygous mutation in PITX2, a gene essential for left/right patterning¹², was recovered in a non-heterotaxy CHD patient, suggesting a broader role for left-right patterning in CHD pathogenesis. The human study identified an important role for genes regulating chromatin. Consistent with this, our screen recovered two mouse CHD genes modulating chromatin, Smarca4 and Prdm1⁵. Hence, findings from our fetal mouse screen may provide insights not only on Mendelian recessive genetic contribution to CHD in the unborn, a segment of the CHD population not well studied, but these findings also may have relevance to the complex genetics of CHD overall. The CHD mouse models generated in this study are invaluable resources for interrogating CHD pathogenesis, with the cryopreserved mutant lines providing access to a large library of mutations, including null mutations for genes without KOMP alleles.

Online Methods

Congenital Heart Defect Phenotyping

C57BL/6J mice were chemically mutagenized using ethylnitrosourea and subject to a two generation backcross breeding scheme²⁷. To identify CHD mutants, G3 fetuses were ultrasound scanned with a two-tier ultrasound phenotyping strategy using the Acuson Sequoia and Vevo2100 ultrasound systems as previously described⁶. The fetal ultrasound imaging was conducted at E13.5-15.5, after completion of major cardiovascular developmental processes, including outflow tract septation and ventricular chamber septation. All of the CHD diagnoses were confirmed or refined by additional analysis using micro-CT/micro-MRI analyses and 3D histopathology using episcopic confocal microscopy^6,28 (Extended Data Figure 1). The summative CHD diagnosis and phenotype description are curated in the public Mouse Genome Informatics (MGI) database (www.informatix.jax.org; enter search box “b2b” will pull up all the lines or type in specific line ID or MGI ID to go directly to a specific mutant line). Included in each mutant line MGI record are sample images and videos collected from various different imaging modalities, including ultrasound images and videos, necropsy images, microCT/microMRI scans, episcopic confocal microscopy image stacks and 3D reconstructions, histology, and high speed videos for analysis of tracheal airway cilia motility.

Cilia Videomicroscopy

Trachea of newborn mice were harvested and mounted as previously described. For nodal cilia imaging, the node-containing region of E7.5-E7.75 embryos (3-6 somite stages) was dissected and mounted on a glass coverslip¹⁰. Videomicroscopy was carried out using a Leica DMIRE2 inverted microscope at 37°C with a 100× oil objective, and DIC optics. To examine flow, 0.35 μm microspheres were added to the medium bathing the ciliated tissue samples (∼1 × 10¹¹ particles/ml, Polysciences, Inc.). High-speed videos (200 fps) were collected using a Phantom v4.2 camera (Vision Research, NJ).

Mutation Recovery with Whole Exome Sequencing Analysis

Genomic DNA from CHD mutant mice were sequence captured using Agilent SureSelect Mouse All Exon kit V1, and sequenced using Illumina HiSeq 2000 with minimum average 50× target sequence coverage (BGI Americas). Sequence reads were aligned to C57BL/6J mouse reference genome (mm9) and analyzed using CLCBio Genomic Workbench and GATK software. To minimize false-negatives, variant calls were set at 5× minimum coverage and >20% alternate reads. Sequence variants were annotated with annovar (http://www.openbioinformatics.org/annovar/) and filtered against dbSNP128 and our in-house mouse exome databases with custom scripts. Intronic variants within 10bp from the splice junction were annotated as potential splicing mutation. All of the mutations recovered are searchable together with phenotype information via the public Bench to Bassinet Congenital Heart Disease Mouse Mutation Database (http://benchtobassinet.com/ForResearchers/BasicScienceDataResourceSharing/GeneDiscoveryinMouseModels.aspx). The exome datasets are available from the GNomEx CvDC Datahub (https://b2b.hci.utah.edu/gnomex/gnomexGuestFlex.jsp?topicNumber=67).

Ciliome Gene Annotation

Ciliome genes were derived from a compilation of multiple bioinformatics, genomic and proteomic studies in different species²⁹. This included a bioinformatics study of X-box containing genes in C. elegans and several proteomic and comparative genomic studies comprising ciliated organisms in several species (C. elegans, Drosophila, Human, Chlamydomonas, Plasmodium falciparum and Trypanosoma). In addition, we also included genes recently identified in the cilia transition zone complex²³ and genes identified to cause Primary Ciliary Dyskinesia (http://ghr.nlm.nih.gov/condition/primary-ciliary-dyskinesia/show/Related+Gene(s)).

Analysis of Splicing Defect Mutations

For analysis of splicing defect mutations, RNA was isolated from skin, trachea, or heart tissue and reverse transcribed to generate cDNA using a High Capacity RNA-to-cDNA Kit (Life Technologies, Inc.). PCR amplification of the cDNA was carried out with primers spanning the affected splice junction using AmpliTaq Gold™ DNA polymerase system (Life Technologies) (95°C for 5 min, 40 cycles at 95°C for 30 s, 58°C for 30 s at 72°C for 1 min, 72°C for 5 min). The amplified products were separated on a 3% agarose gel and the bands were excised using the QIAquick Gel Extraction Kit (Qiagen) and analyzed by Sanger capillary sequencing.

Simulation-based Mutation Enrichment Analysis

A simulation approach was used to test whether mutations in the 61 CHD genes are enriched in the mouse exome datasets generated from the 113 CHD mutant mouse lines. From the 113 exomes, we recovered 149 mutations in the 61 CHD genes, 91 of which were identified as pathogenic, yielding an excess of 58 incidental mutations. For this simulation, all the genes from the mouse genome were grouped in bins of 500 genes based on coding DNA sequence (CDS) length. Then 61 genes with the same size distribution as the observed 61 CHD genes were randomly sampled. For each bin that contained k of our 61 CHD genes, we randomly sampled k genes, and counted the number of mutations in these genes in the 113 mouse exomes. Based on 50,000 sets of 61 randomly sampled genes, we calculated a p-value based on the count of random gene sets that harbored 149 or more mutations. To take potential gene coverage bias into account, we limited our random gene selection to genes with at least one coding mutation in our CHD exomes, which yielded p-value of 0.00114. When we choose random genes from all genes, including genes not found in our exome dataset, p-value is <0.00002.

Interactome Analysis

To construct the interactome, each of the mouse CHD genes was mapped to its human ortholog, and PPIs of these genes were extracted from HPRD (www.hprd.org) and BioGRID (thebiogrid.org). The interactome is augmented with PPIs discovered computationally using the recently developed HiPPIP model (unpublished results). Pathway and Gene Ontology term enrichment analysis of the interactome were carried out using Ingenuity Pathway Analysis suite and BiNGO³⁰ plugin of Cytoscape (http://www.cytoscape.org/). P-value for each pathway and GO term was calculated using Fisher's exact test with Benjamini-Hochberg correction (Supplementary Spread Sheet 3). Shortest paths were computed using Dijkstra's algorithm implemented in the networkx package of python. Average length of the shortest paths between pairs of CHD genes was compared to that of 61 random genes, where random genes were sampled to have similar degree distribution as the 61 CHD genes. For this, all human genes were categorized into 7 bins based on their degree in the interactome, (0, 1-5, 6-25, 26-50, 51-75, 75-100, >100). As many random genes were selected from any bin as the number of CHD genes in that bin. This process was repeated 1,500 times, and average length of the shortest paths is computed each time. P-value is computed empirically to determine statistical significance.

Extended Data

Extended Data Table 1. CHD Genes with Multiple Alleles.

Gene Symbol (mouse)	Gene Symbol (human)	Mouse CDS (bp)	# Alleles Recovered
Daw1	DAW1	933	2
Dnai1	DNAI1	2,106	2
Drc1	DRC1	2,262	2
Pde2a	PDE2A	2,808	2
Ccdc39	CCDC39	2,814	3
Tmem67	TMEM67	2,988	2
Armc4	ARMC4	3,114	la
Adamts6	ADAMTS6	3,351	lb
Pcsk5	PCSK5	5,634	2
Plxnd1	PLXND1	5,778	2
Cep290	CEP290	7,440	2
Pkd1l1	PKD1L1	7,824	2
Megf8	MEGF8	8,370	2
Dnah11	DNAH11	13,467	8
Dnah5	DNAH5	13,866	9

^aIdentical mutation was recovered from two different mutant lines.

^bIdentical mutation was recovered from three different mutant lines, later confirmed to be a spontaneous mutation in the support colony.

Extended Data Table 2. Ciliome Mutations in Laterality and Non-Laterality Lines.

Laterality Mutant Lines (30)		Non-Laterality Mutant Lines (31)
Ciliome Genes*	Non-Ciliome Genes	Ciliome Genes	Non-Ciliome Genes
Anks6	Aplbl	CepllO	Acan
Armc4 (m)	Cfcl	Dctn5	Adamts6
Bice I	Cml5	Fuz	Ap2bl
Cc2d2a	Gm572	Jbtsll	Cxcr4
Ccdcl51 (m)	Meg/8	Kif7	Dnm2
Cede3 9 (m)	Mmp21	Lrp2	Frem2
Cep290	Pcsk5	Lox	Hectdl
Daw1 (m)	Total: 7	MyhlO	Lrpl
Dnaaf3 (m)		Pde2a	Ltbpl
Dnah5 (m)		Pkdl	Ndstl
Dnah11(m)		Sufu	Plxnd1
Dnail (m)		Total: 11	Prdml
Drc1 (m)			Prickle 1
Dync2hl			Ptkl
Dyxlcl (m)			Robol
Foxjl (m)			Smad6
Iftl40			Smarca4
Ift74			Snx17
Nek8			Tabl
Pkdlll (m)			Zbtbl4
Pskhl			Total: 20
Tbcld32
Tmem67
Total: 23 (12)

^*m: mutation affects motile cilia function

Extended Data Table 3. Mouse CHD genes and associated human diseases.

Ciliopathy Related
Armc4	Primary ciliary dyskinesia, Kartagener Syndrome
Ccdc39	Primary ciliary dyskinesia, Kartagener Syndrome
Dnaaf3	Primary ciliary dyskinesia, Kartagener Syndrome
Dnahl1	Primary ciliary dyskinesia, Kartagener Syndrome
Dnah5	Primary ciliary dyskinesia, Kartagener syndrome
Dnail	Primary ciliary dyskinesia, Kartagener syndrome
Drc1	Primary ciliary dyskinesia, Kartagener syndrome
Dyx1c1	Primary ciliary dyskinesia, Kartagener syndrome
Cc2d2a	Joubert Syndrome; COACH syndrome
Cep290	Joubert Syndrome 5; Senior-loken Syndrome6, Bardet Biedl Syndrome Leber Congenital Amaurosis 10, Meckel Syndrome-4
Jbts17	Joubert Syndrome
Kif7	Joubert Syndrome, Acrocallosal Syndrome, Hydrolethalus Syndrome
Tmem67	Joubert Syndrome; Nephronophthisis-11, Meckel Syndrome, COACH syndrome, Bardet Biedl Syndrome
Anks6	Polycystic kidney disease; nephronophthisis 16
Bicc1	Polycystic kidney disease; cystic renal dysplasia
Nek8	Polycystic kidney disease; nephronophthisis 9
Pkd1	Polycystic kidney disease
Pkd1l1	Kidney disease, diaphanospondylodysostosis
Dync2h1	Short rib Polydactyly; asphyxiating thoracic dystrophy
Ift140	Mainzer-Saldino Syndrome, short rib thoracic dysplasia
Other Human Diseases
Ap2b1	Ataxia telangiectasia, cerebellar degeneration
Cfc1	Conotruncal heart defects associated with heterotaxia
Cxcr4	WHIM Syndrome
Dnm2	Charcot-Marie-Tooth Syndrome, Centronuclear myopathy
Frem2	Fraser's Syndrome, Cryptophthalmos syndrome
Lrp1	Alzheimer Disease, Schizophrenia
Lrp2	Donnai-Barrow Syndrome; Facio-oculo-acoustico-renal (FOAR)
Ltbp1	Pseudoexfoliation Syndrome
Megf8	Carpenter's syndrome
Pcsk5	VACTERL/VATER syndrome
Prdm1	Devic disease, B cell lymphoma
Smad6	Aortic valve disease
Smarca4	Rhabdoid tumor predisposition syndrome type 2 (RTPS2), Coffm-Siris syndrome
Snx17	Cerebral cavernous malformation; cavernous malformation
Sufu	Medulloblastoma, basal cell nevus syndrome

Supplementary Material

Supplemental Spre Sheet 3

A. Extended Data Figure Legends

Extended Data Figure 1. Breeding, phenotyping, and mutation recovery pipeline for mouse forward genetic screen.

a. Two generation backcross breeding scheme used to generate G3 mutants with recessivemutations causing congenital heart defects, with all offspring from a single G1 male definedas a distinct pedigree or mutant line.

b. Pipeline for recovery and curation and cryopreservation of CHD mutant mouse modelsand the recovery of pathogenic CHD causing mutations.

Extended Data Figure 2. Situs anomalies and congenital heart defects in Ap1b1^b2b1660 mutants.

Mutants from line 1660, identified with an Ap1b1 mutation, exhibits situs solitus (a), situs inversus (b), or heterotaxy (c). Situs solitus, characterized by normal left-right visceral organ positioning, the heart apex (arrow) points to the left (levocardia), four lung lobes are on the right and one on the left, stomach is to the left, and the dominant liver lobe is on the right. With situs inversus, there is complete mirror reversal of organ situs, while with heterotaxy, visceral organ situs is randomized, such as dextrocardia with levogastria shown in (c). The heterotaxy mutant in (c) exhibit complex CHD with AVSD (d), ventricular septal defect (VSD) (e), duplicated inferior vena cava (IVC) (f), and left pulmonary isomerism with bilateral single lung lobes (g)

Extended Data Figure 3. Distribution of mutations recovered by whole mouse exome sequencing analysis.

a. Distribution of incidental coding mutations (left), pathogenic CHD causing mutations (middle) and ciliome-CHD genes (right) recovered from 113 CHD mutant mouse exomes.

b. Mouse CHD genes and the associated defect phenotypes.

Extended Data Figure 4. Pathogenic splicing mutations causing CHD

a. The 19 pathogenic splicing mutations recovered are shown, with mutations locatedbeyond the 2-base canonical splice junction highlighted in grey.

b. Schematic diagram showing the anomalous Dnah5^{c.133290-10T>A} mutant transcriptobserved, with the PCR primer location and anomalous PCR product size indicated.

c. Sanger sequencing profile showing point mutation in Dnah5^{c.133290-10T>A} mutant transcriptvs. that of wildtype.

d. PCR amplification of Dnah5^{c.133290-10T>A} heterozygous mutant (m/+) showed the expected 514 bp wildtype and 271 bp mutant PCR product, while only the 271 bp mutant product was observed in the homozygous mutant (m/m) sample.

Extended Data Figure 5. Ciliome mutations causing CHD with and without laterality defects

A flow chart showing the distribution of ciliome vs. non-ciliome CHD genes among laterality versus nonlaterality-CHD lines, and further stratification of of ciliome CHD genes affecting primary vs. motile cilia function.

Extended Data Figure 6: CHD phenotypes associated with mutations affecting endocytic trafficking

Outflow tract malalignment defects with double outlet right ventricle (DORV) and overriding aorta (Ao) were observed in Ap2b1 (a), Dnm2 (b), Snx17 (c) mutants, with Ap2b1 (a) mutant also showing anterior positioning of the aorta (Taussig-Bing type DORV). Snx17 mutant also has AVSD. Lrp2 (d) and Lrp1 (e) mutants both exhibited outflow tract septation defect with persistent truncus arteriosus (PTA). See Supplemental Movie S8 for 3D reconstruction of the Ap2b1 mutant heart shown in (a).

Extended Data Figure 7. CHD genes associated with axon guidance

Diagram illustrating the biological context of several CHD genes known to be involved in axonal guidance (color-highlighting indicates CHD genes recovered from present screen).

B. Supplemental Spreadsheets

Supplemental Data Spreadsheet 1: CHD mutations and comparison to other mutant alleles and human diseases

Supplemental Data Spreadsheet 2: Position and sequence conservation of CHD mutations

Supplemental Data Spreadsheet 3: Gene ontology and pathway analyses of CHD interactome genes

Supplemental Data Spreadsheet 4: Ciliome and pathway annotation for 28 CHD candidate genes recovered from Pediatric Cardiac Genomics Consortium human CHD patient exome sequencing analysis

C. Supplemental Movie Legends

Movie S1. Ultrasound imaging of normal fetus shown in Figure 1a.

Vevo 2100 color flow Doppler imaging in coronal view show normal alignment of the two great arteries with normal connection to the two ventricles.

Movie S2. Ultrasound imaging of mutant fetus shown in Figure 1c.

Vevo 2100 color flow Doppler imaging in coronal view show a single outflow tract originating from right ventricle (RV) indicating aortic atresia. Shunting of blood between LV and RV suggests ventricular septal defect, and biventricular hypertrophy.

Movie S3a. Ultrasound imaging of mutant fetus from line b2b2025 shows DORV and presence of a VSD.

Vevo2100 color flow imaging in coronal view of fetus in Figure 1e-k showed aorta and pulmonary artery side-by side, both emerging from the right ventricle (RV) indicating DORV, and a shunting of blood between the two ventricles indicating ventricular septal defect (VSD).

Movie S3b. Ultrasound imaging of mutant fetus from line b2b2025 shows AVSD and muscular VSD.

Vevo 2100 color flow imaging in transverse view of fetus in Figure 1e-k detected forward blood flow and regurgitation from a common atrioventricular valve suggesting atrioventricular septal defect. Also observed was a muscular ventricular septal defect.

Movie S3c. Ultrasound imaging of mutant fetus from line b2b2025 shows heterotaxy.

Vevo2100 2D imaging in coronal view of fetus in Figure 1e-k detected heart apex pointing to left suggesting levocardia, but stomach (Stom) located on right, which together indicated this fetus has heterotaxy.

Movie S4. Videomicroscopy of ciliary motion of tracheal epithelia of newborn Foxj1 mutant mice.

Tracheal airway epithelium in a newborn homozygous Foxj1^b2b774 mutant mouse shows normal ciliary motion.

Movie S5. Videomicroscopy of ciliary motion in embryonic node of Foxj1 mutant embryo. Cilia in the embryonic node of a homozygous Foxj1^b2b774 mutant embryo shows dyskinetic ciliary motion and no nodal flow.

Movie S6. Videomicroscopy of ciliary motion in tracheal epithelia of newborn Ap1b1 mutant.

Tracheal airway epithelium in a newborn homozygous Ap1b1^b2b1660 mutant mouse shows normal ciliary motion.

Movie S7. Videomicroscopy of ciliary motion in embryonic node of Ap1b1 mutant embryo. Cilia in the embryonic node of a homozygous Ap1b1^b2b1660 mutant fetus shows normal ciliary motion and flow.