Patient recruitment and DNA extraction
A group of 333 BAV index cases (249 males, 84 females) were prospectively recruited from La Timone Hospital, Marseille in strict compliance with all relevant ethical regulations (approved by the Marseille ethic committee n°13.061). All DNA and research protocols were collected in compliance with the Institutional Review Board after informed consent was obtained. Patients underwent clinical evaluation that included family history, physical examination and 2D-echocardiography.
Genomic DNA was extracted from peripheral blood leukocytes using the FlexiGene DNA kit (Qiagen) following the manufacturer’s protocol. DNA purity and quantity were assessed with Nanodrop spectrophotometer (Thermo Scientific).
The two exons and intronic HOXA1 regions were sequenced bidirectionally to search for sequence variations. Primers were designed according to HOXA1 genomic reference sequence NM_005522. PCR products were sequenced using BigDye Terminator Cycle Sequencing kit (Applied Biosystems) and run-on automatic sequencer ABI 3130XL (Applied Biosystems). Patient sequences were aligned to the HOXA1 reference sequence using Seqscape software (Applied Biosystems).
Variant pathogenicity was estimated using UMD-Predictor (http://umd-predictor.eu)56 and Human Splicing Finder57. Allele frequencies for each variant were obtained from the Genome Aggregation Database (gnomAD)58, which provides high-quality genetic data on 141,456 individuals (125,748 exomes and 15,708 whole genomes).
All mouse experiments were carried out using protocols approved by the “comité d’éthique pour l’expérimentation animale” (Marseille ethical committee, Protocol N°32-08102012). The null allele of Hoxa1neo (hereafter referred to as Hoxa1-) has been described previously59. All genotypes were observed at the expected Mendelian ratios (supplementary material, Table S2). Hoxa1-enhIII-Cre, Wnt1-Cre, RosaLacZ, and Gt(ROSA)26Sortm9(CAG-tdTomato)Hze (RosatdTomato) transgenic lines have been previously described32,60–62. Mice were genotyped by PCR and embryos staged taking the morning of the vaginal plug as embryonic day (E) 0.5.
Plasmids pCS2-Prep1 , pCMV-PBX1a , pML-EPHA2-r4-Luc  pGL4.74 [from Promega] have been described elsewhere. Plasmids coding for wild type HOXA1 (human HOXA1) and hHOXA1 mutants 9A, 9G and 7G were generated with the Gateway® cloning system. Gateway® entry vectors were obtained using the following primers for PCR delineating the hHOXA1 sequence: 5’-GGGGACAACTTTGTACAAAAAAGTTGGCATGGACAATGCAAGAATGAACTCC-3’ and 5’-GGGGACAACTTTGTACAAGAAAGTTGGGTAGTGGGAGGTAGTCAGAGTGTC-3’.
For plasmids coding for the hHOXA1 mutants +1His and +1HisArg, mutagenesis was carried out by NEBaseChanger using the following mutagenic primers forward 5’- CACCATCACCACCCCCAGCCG -3’, reverse 5’-GTGGTGGTGGTGGTGGTG-3’
For plasmids coding for the hHOXA1 variants -1His, -1HisArg and -3HisArg, mutagenesis was carried out by overlapping triple PCRs using the above primers as well as the following mutagenic primers: +1His forward 5’-CAGATCGGTTCGCCCCACCACCACCACCACCACCATCACCACCCCCAGCCGGCTACC-3’, reverse 5’-GGTAGCCGGCTGGGGGTGGTGATGGTGGTGGTGGTGGTGGTGGGGCGAACCGATCTG-3’; -1His, forward 5’-CAGATCGGTTCGCCCCACCACCACCACCACCACCATCACCACCCCCAGCCGGCTACC-3’, reverse 5’-GGTAGCCGGCTGGGGGTGGTGATGGTGGTGGTGGTGGTGGTGGGGCGAACCGATCTG-3’; -3His forward 5’-CAGATCGGTTCGCCCCACCACCACCACCATCACCACCCCCAGCCGGCTACC-3’, reverse 5’-GGTAGCCGGCTGGGGGTGGTGATGGTGGTGGTGGTGGGGCGAACCGATCTG-3’. BP-clonase® reactions were achieved using pDONR223 vector and amplified cDNA to generate Gateway® entry pEnt plasmids. These pEnt plasmids were involved in LR-clonase® reactions with pDest-FLAG N-ter destination vector  to generate pExp-FLAG-hHOXA1 expression vectors for the WT and mutant hHOXA1 sequences.
Human embryonic kidney cells, HEK293T, were maintained in culture flasks at 37 °C under 5% of CO2. D-MEM medium (Gibco, Hek293T: #61965-059; COS7: #31885-023) was supplemented with 10% fetal albumin serum (Invitrogen, #10270-106) and 100 U/ml penicillin-streptomycin (Gibco, #15140-122), 1 mM sodium pyruvate (Gibco, #11360-070). Embryonic carcinoma cells, P19, were maintained in culture flasks at 37°C under 5% of CO2. D-MEM medium (Gibco, Hek293T: #61965-059; COS7: #31885-023) was supplemented with 10% fetal albumin serum (Invitrogen, #10270-106) and 100 U/ml penicillin-streptomycin (Gibco, #15140-122).
Luciferase reporter assays
For Luciferase reporter assays, HEK293T cells were seeded in 96-well plates (1.4 × 104 cells per well) and transfected after 24 hours using JetPrime transfection reagent (Polyplustransfection, #114-07), with 65 ng of target luciferase reporter plasmid (pML-EPHA2-r4-Luc), 25 ng of each FLAG-hHOXA1 expression plasmids (WT, +1His, +1HisArg, -1His, or -3HisArg), 12.5 ng of PREP and PBX expression plasmids each, and 5 ng of constitutive reporter plasmid pGL4.74. Empty vector was added when needed so that each transfection included a total amount of 120 ng of DNA. Forty-Eight hours after transfection, cell lysis and enzymatic activity analysis were performed with Dual-Glo® Luciferase Assay System (Promega E2920) following manufacturer’s instructions. For each transfection, the constitutively active pGL4.74 reporter was used as an internal standard for target reporter activity normalization. The relative luciferase activity was established as the ratio between target and constitutive luciferase activities.
Protein half-life analysis
HEK293T cells were plated on 6-well plates, 700 000 cells per well. The cells were transfected with 200 ng of specific pExp plasmids coding for FLAG-tagged WT or mutant HOXA1 proteins using JetPrime transfection reagent (Polyplustransfection, #114-07). For translation inhibition, 24 h after transfection, cells were treated with 200 μg/ml of cycloheximide (Sigma #01810). For both translation and proteasome inhibition, cells were pre-treated 20h after transfection with 7 μM of MG132 (Merck Milipore #474790) and 24 h after transfection were treated with 200 μg/ml of cycloheximide and with 7 μM of MG132. HEK293T cells were lysed with IPLS buffer (20 mM TrisHCl pH 7.5, 120 mM NaCl, 0.5 mM EDTA, 0.5% NP40, 10% glycerol) supplemented with Complete™ protease inhibitor (Roche, #11873580001) for 20 min under agitation on ice. Cell lysates were centrifuged 5 min at 16,000 g at 4°C. 50 μl of each lysate was mixed to Laemmli 6×, boiled 5 min at 95°C and run on a 10% SDS-PAGE. Western blotting was used to visualize FLAG-tagged proteins (α-Flag: Sigma-Aldrich #F1804; α-mouse: Santacruz #sc-516102), as well as ACTIN (α-β-actin-HRP, Sigma-Aldrich #A3854). ImageJ software was used to estimate protein abundance.
AB wild-type (ABwt) zebrafish were maintained under standardized conditions and experiments were conducted in accordance with the European Communities council directive 2010/63. Morpholino oligonucleotides (MOs) were obtained from Gene Tools (Philomath, OR, USA) and injected into one-cell stage embryos. The sequence of the injected MOs are the following:
hoxa1a MO: 5′-CTAAGAATGTGCTCATTGTGTCATC-3’, 7.5 ng
hoxb1b MO: 5′-ATTGCTGTGTCCTGTTTTACCCATG-3’, 1 to 2 ng
Rescue experiments were performed by co-injecting hoxa1a MO (7.5 ng) or hoxb1b MO (1.5 ng) with 25 pg of human wild-type HOXA1 (WT) mRNA. Injections of human HOXA1 mRNA alone were performed with either 5pg (WT, -1His, -3HisArg) or 25 pg (WT, +1His) of in vitro transcribed 5’ capped sense RNAs synthesis using the mMessage mMachine kit (Ambion).
For aortic valve imaging, larvae were incubated one day prior imaging in 0.2 μM BODIPY-FL Ceramide (Invitrogen D3521) in Embryo medium + PTU (0.003% 1-phenyl-2-thiourea). 7 dpf larvae were then anesthetized with Tricaine (0.16 g/L) and mounted in low melting agarose. Imaging was performed with a Zeiss LSM710 two-photon microscope coupled to a Ti:sapphire laser (Spectra-Physics, Santa Clara, CA, USA) and a water immersion X objective.
Histological, immunohistochemistry and X-gal staining
Standard histological procedures were used63. Heart tissues from Hoxa1-/- and littermate controls were paraffin-embedded and cut at 8 μm per tissue sections. Sections were stained with H&E (Sigma) according to the manufacturer’s instruction. CD31 (PECAM) (1:100, Pharmingen) immunohistochemistry were performed using 4% paraformaldehyde fixed tissue and Alexa fluorescent-conjugated antibodies (Life Technologies) were used at 1:500. X-gal staining and in situ hybridization were performed on 12 μm thick frozen sections of hearts from E13.5 and E18.5 as described previously63. For each experiment a minimum of 3 embryos of each genotype was scored. Embryos were examined using an AxioZoom.V16 (Zeiss) and photographed with an Axiocam digital camera (Zen 2011, Zeiss).
RNA-FISH was performed according to the protocol of the RNAscope Multiplex Fluorescent v2 Assay (Acdbio; cat. no.323110), which detects single mRNA molecules. In briefly, E9.5 embryos were fixed for 20-30 h in 4% paraformaldehyde and then dehydrated in methanol. Whole-mount RNA-FISH was performed as previously described64. Embryos were imaged using an AxioZoom.V16 microscope (Zeiss). The following probes were used: mm-Sox10-C2 (Acdbio; cat n° 43591-C2), mm-Robo2-C1 (Acdbio; cat n° 475961).
Total RNA was isolated from the pharyngeal region and sorted cells with NucleoSpin RNA XS (Macherey-Nagel) following the protocol of the manufacturer. RNA-seq libraries were created on RNA samples (500 ng) with RNA integrity number (RIN) > 7 using ‘kapa mRNA HyperPrep Kit’ (Roche). Final cDNA libraries were checked for quality and quantified using 2100 Bioanalyzer (Agilent Technologies). The libraries were loaded in the flow cell at 1.8 pM concentration and clusters were generated in the Cbot and sequenced in the Illumina NextSeq 500 as 75 bp pair-end reads following Illumina's instructions. Reads were mapped onto the GRCh38 assembly of the Mus musculus genome using STAR v2.5.3a. Aligned reads in BAM format were annotated against the protein-coding mRNA regions. Stringtie v.1.3.1c platform was utilized to visualize the annotated/mapped read, quantify data (using corrected log2 transformed Reads Per Million Values of nonstrand specific, unmerged isoform), and perform percentile normalization. An intensity difference analysis method was used to identify differentially expressed genes (DEGs) on the normalized quantification data.
Three-dimensional (3D) reconstructions
Fiji software was used to make the 3D reconstructions presented in our manuscript. At E13.5, E15.5 and E18.5, images of 20-30 8 μm paraffin sections were manually aligned and taken to generate 3D reconstruction.
All parametric data are expressed as mean ±SEM. Statistic was determined using the Student’s t test to compare variances. For non-parametric data, the Mann-Whitney test was used to calculate significant between the medians. A p value of less than 0.05 was considered significant.