Mutations in Hnrnpa1 cause congenital heart defects

Incomplete penetrance of congenital heart defects (CHDs) was observed in a mouse model. We hypothesized that the contribution of a major genetic locus modulates the manifestation of the CHDs. After genome-wide linkage mapping, fine mapping, and high-throughput targeted sequencing, a recessive frameshift mutation of the heterogeneous nuclear ribonucleoprotein A1 (Hnrnpa1) gene was confirmed (Hnrnpa1ct). Hnrnpa1 was expressed in both the first heart field (FHF) and second heart field (SHF) at the cardiac crescent stage but was only maintained in SHF progenitors after heart tube formation. Hnrnpa1ct/ct homozygous mutants displayed complete CHD penetrance, including truncated and incomplete looped heart tube at E9.5, ventricular septal defect (VSD) and persistent truncus arteriosus (PTA) at E13.5, and VSD and double outlet right ventricle at P0. Impaired development of the dorsal mesocardium and sinoatrial node progenitors was also observed. Loss of Hnrnpa1 expression leads to dysregulation of cardiac transcription networks and multiple signaling pathways, including BMP, FGF, and Notch in the SHF. Finally, two rare heterozygous mutations of HNRNPA1 were detected in human CHDs. These findings suggest a role of Hnrnpa1 in embryonic heart development in mice and humans.

control. Poor DNA samples (with < 50% of the markers having genotyped) or poor sequencing markers (with < 50% of samples having genotyped) were filtered from the downstream analysis.

Genotyping of mouse rs32183020 by Sanger sequencing
Two primers (rs32183020-F and rs32183020-R; Supplemental Table 5) were first designed and amplified in a 50 µ L reaction volume containing 10 x PCR buffer, 50 mM MgCl 2 , 10 mM dNTPs, 100-200 ng DNA, and 1 U Taq. The reaction was hot-started for 4 min and run through 35 cycles using the following settings: 94 °C for 30 s, 55 °C for 30 s, and 72 °C for 30 s. The forward primer (rs32183020-F) was used for sequencing by using the ABI 3730XL sequencer.
Targeted next-generation sequencing 3 µg DNA was randomly fragmented and the both ends of fragments were ligated to adapters. Extracted DNA was then amplified by ligation-mediated PCR (LM-PCR), purified, and hybridized to the NimbleGen 2 and then washed out. Both the non-captured and captured LM-PCR products were subjected to q-PCR to estimate the magnitude of enrichment. Each captured library was then loaded on Illumina Hiseq2000 sequencer, and we performed high-throughput sequencing for each captured library to ensure that each sample met the desired average sequencing depth.
All the short reads were aligned against Mus musculus reference genome NCBI37/mm9 chr15 using BWA (2) version 0.5.9. Illumina 1.5 encoding Phred quality scores were converted to Sanger scores by BWA using the '-I' option. The multi-threading mode was enabled with eight threads concurrently generating SA coordinates for each mate-read and with permission of up to two mismatches. Picard tools (http://picard.sourceforge.net) version 1.66 was applied to prepare the original aligned results from BWA for variants detection in the succeeding step. First, mate-pair information was fixed to ensure that it is synced between each read and its mate pair with FixMateInformation function. Fixed BAM file generated from the original SAM file was then reordered and sorted using ReorderSam and SortSam tools, respectively. ReorderSam guaranteed that contigs in reads file are in ordered as those in the provided reference files, whereas SortSam sorted the order of the BAM file by coordinates (or unsorted, queryname are used as alternates). Finally, duplicates in the alignments were marked but not removed by Picard with MarkDuplicates model. SAMtools (Sequence Alignment/Map) (3) version 0.1.17 with default parameters was used to detect SNPs and indels from the previous prepared BAM file. To avoid the possible overestimation of mapping qualities during BWA-short alignment, '-C50' option was enabled when we applied mpileup function in SAMtools. Results in bcf format from mpileup analysis were then translated into vcf file using bcftools. The maximum read depth (-D option) was set enlarged enough (2000) so as not to exclude any reads, as we would like to maximally cut down the false negatives. ANNOVAR (4) version 20111120 was adopted for the functional annotation of the variants detected using SAMtools. The refSeq Gene and UCSC known genes were used for the gene-based annotation, whereas phastCons30way, tfbs, band, and segdup databases were downloaded from UCSC to carry out region-based annotation. We also mapped the variants to mouse dbSNP128 to identify the variants which were not recorded.

Human subjects and HNRNPA1 variants screening
A total of 273 Chinese non-syndromic CHD patients and their parents with no reported cardiac phenotype and 225 sporadic Pakistani CHD patients were recruited for this study. Three hundred normal Chinese subjects were also recruited. All patients underwent an extensive clinical assessment, including standard physical examination, electrocardiogram (ECG), ultrasonic echocardiogram, and chest X-ray. We then classified them under different pathologic types according to the diagnosis mentioned above or open heart surgery. Familial histories and clinical information of all patients were reviewed.
Human genomic DNA was extracted from peripheral blood leukocytes using standard methods. The human HNRNPA1 gene is located on 12q13.13 and is encoded by eleven exons (NC_000012.12). The exons and nearby introns of HNRNPA1 were amplified by PCR using specific primers (as shown in Supplemental Table 4). All fragments were sequenced using the ABI 3730XL sequencer. Meanwhile, PCR products were sequenced using the Sanger method.
When any mutation was detected in human CHD patients, we first searched the NCBI-SNP database to see whether the mutation existed or not, then we compared the observation with 1000 genome database. Finally, we sequenced 300 normal individuals to ensure that the mutation was absent in the normal subjects.

Characterization of Hnrnpa1 mutant mice by histological and OPT analyses
The E13.5 embryos were sectioned and then stained with Hematoxylin and Eosin. Optical projection tomography (OPT) was performed for P0 mice (5). Hearts were dissected, placed in low-melt agarose, and exposed to benzyl alcohol/benzyl benzoate for imaging using a Bioptonics 3001 OPT Scanner (Edinburgh, UK). Images were visualized and evaluated using Volocity (Perkin Elmer, Waltham, MA).

Quantitative RT-PCR (qRT-PCR)
The whole embryos, pharyngeal tissues and heart tubes were dissected at E9.5 respectively and immediately frozen using liquid nitrogen. Then, the samples were stored at -80 °C until they were genotyped. The total RNA was isolated with TRIzol reagent (15596-026; Thermo Fisher Scientific) and standard protocols. First strand cDNA was synthesized using the SuperScript TM III first-strand synthesis system kit (18080-051; Thermo Fisher Scientific). Then, qRT-PCR was performed using TaKaRa SYBR Premix Ex Taq™ kit (RR420W; TaKaRa). For Hnrnpa1, six wild-type embryos, nine heterozygous, and eight homozygous mutants were used. Pharyngeal tissues (SHF) and heart tubes were isolated for analyses of SHF and heart tube specific genes (Supplemental Figure 7). For analysis of the SHF, eight wild-type littermate controls, five Hnrnpa1 +/ct heterozygotes and nine Hnrnpa1 ct/ct homozygotes were used for Fgf8, Fgf10, Isl1, Mef2c and Tbx1; eight wild-type littermate controls, five Hnrnpa1 +/ct heterozygotes and eight Hnrnpa1 ct/ct homozygotes were used for Nkx2.5; eight wild-type littermate controls, six Hnrnpa1 +/ct heterozygotes and nine Hnrnpa1 ct/ct homozygotes were used for Acvr1, Bmpr1a and Jag1. For analysis of the heart tube, eight wild-type embryos, eight heterozygous, and nine homozygous mutants were dissected for analysis of Mlc2a, Mlc2v and Nkx2.5; eleven wild-type embryos, nine heterozygous, and nine homozygous mutants were dissected for analysis of Mef2c; eleven wild-type embryos, eleven heterozygous, and nine homozygous mutants were dissected for analysis of Myocd and SRF. GAPDH was used as an internal control (9). qRT-PCR primers for Hnrnpa1(designed), Acvr1 (10), Mlc2v (15), Nkx2.5 (designed) and SRF (16) are listed in Supplemental Table 6.

Western blot analysis
Embryos were dissected at E9.5 and lysed with ice-cold radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 0.1% Triton-X-100) supplemented with protease and phosphatase inhibitors (Roche). The embryo lysates were then passed through 1 mL syringes with needles for around 20 times. The supernatant was obtained after centrifugation at 12,000 g for 10 min at 4 °C. Extracted protein samples were denatured in 1 × Laemmli buffer and subjected to SDS-PAGE. Total proteins were then transferred onto nitrocellulose membrane (GE Healthcare). The membrane was incubated with primary antibodies overnight at 4 °C after 1h blockage in BSA at room temperature. β-Actin was used as the loading control. Primary antibodies used: anti-mouse hnRNP A1 (sc-32301; Santa Cruz); anti-β-Actin (A2228; Sigma-Aldrich). The blots were detected using chemiluminescence.

In vitro differentiation of mESCs into cardiomyocytes CRISPR knockout of Hnrnpa1
CRISPR/Cas9 was used to generate Hnrnpa1 knockout Nkx2.5-EGFP-mESCs (17). Briefly, sgRNA (GGTCGTGGAGGCGGTTTTGG) was linked into the Cas9 plasmid (48139; Addgene) for lipofetamine 2000 (Life tech) induced transfection. Then, cells were cultured under fresh mESC medium until the colonies emerged after the three days of puromycin (Life Tech, 2.5 µg/mL) treatment. Two Hnrnpa1 −/− cell lines (KO1 and KO2) were utilized for further experiments, which were homozygous through genotype analysis. PCR and Sanger sequencing were performed to confirm the sequences in the knockout region and predicted off-target genomic loci (Supplemental Table 7).
Cardiomyocyte differentiation: mESCs were cultured and expanded with mESCs maintenance medium (15% FBS, 85% DMEM, L-Glutamine, NEAA, 0.1 mM 2-mercaptoethanol, and 10 ng/mL LIF), which was changed on a daily basis. When cell colonies cultured on fibronectin-coated plate reached about 80% confluence, the medium was removed. After washing by PBS, the cells were digested with 0.05% trypsin-EDTA for 3 minutes at 37 °C . After centrifugation, the cells were suspended in the differentiation medium (15% FBS, 85% DMEM, L-Glutamine, and NEAA). After determining the cell number, the cells were seeded at a density of 2 × 10 5 /mL into 6-well ultra-low attachment plates (2 mL/well). Cells were cultured for 2 days to allow their spontaneous aggregation. Differentiation medium was changed every day to allow growth of the EBs. At the end of day 6, EBs were transferred to a fibronectin-coated plate to allow them to adhere to the plate for further differentiation. The medium was changed every day until the end of the observation period (day 9).

Flow cytometry analysis of Nkx2.5-EGFP positive percentage during differentiation
During the differentiation of Nkx2.5-EGFP-mESCs, GFP was expressed when Nkx2.5 was expressed. To determine the percentage of Nkx2.5-EGFP cells during differentiation, the cells were collected and flow cytometry analysis was performed at day 7. Differentiated mESCs in each group were trypsinized with 0.05% trypsin for 5 minutes and then collected through centrifugation, respectively. Cells were washed with 1 x PBS, followed by fixation with 4% paraformaldehyde (PFA). The cells were kept in dark, in 4 °C refrigerator until they were analyzed. Data were analyzed by collecting 50000 events on Beckman Coulter FC500 using CXP Analysis 2.0 software. For each cell type, in vitro differentiation was performed three times. Cell samples for flow cytometry analysis were collected at day 7 of the differentiation period. Undifferentiated wild-type mESCs were collected as control.   Table 4 Primers for sequencing of human HNRNPA1 Genotyping Primer ID Sequence  Table 5 Primers for genotyping