Non-coding cause of congenital heart defects: Abnormal RNA splicing with multiple isoforms as a mechanism for heterotaxy

Figures

• 

  Figure 1. Intronic variant in ZIC3 identified by WGS of an X-linked heterotaxy pedigree (A) A pedigree with four heterotaxy-affected males (black squares) displaying an X-linked recessive inheritance pattern. Generations are labeled I–IV while individuals within each generation are numbered from 1 to 20. Deceased individuals are denoted with a diagonal line with the cause of death indicated as “d.” when available. X-exome sequencing (dashed orange outer frame) performed on two separate trios did not identify a coding variant in ZIC3, while WGS (solid blue outer frame) completed on two males with heterotaxy (IV-1 and IV-18) revealed a ZIC3 c.1224+3286A>G intronic variant. A + W, alive and well; CHD, congenital heart defect; CL, cleft lip; CP, cleft palate; E, encephalocele; GA, gestational age; IUFD, intrauterine fetal demise; IVF, in vitro fertilization; MAB, missed abortion; P, pregnancy; SAB, spontaneous abortion; SAB 2/2 PA, spontaneous abortion secondary to placental abruption; VSD; ventricular septal defect; wk, week. (B) Schematic diagram of IV-1 and IV-18 WGS and variant filtering steps (Figure S1) identifying ZIC3 c.1224+3286A>G as a plausible disease-causing variant. (C) Sanger…

• 

  Figure 2. The ZIC3 c.1224+3286A>G variant acts as a 3′ splice acceptor between exon 2 and the predicted P1 in a minigene assay (A) Illustrative representation of ZIC3 showing the ZIC3 c.1224+3286A>G variant predicted to result in a 3′ cryptic splice acceptor that causes the inclusion of P1. (B) ZIC3 encodes two isoforms: isoform 1 formed by exons 1-2-3 (dominant isoform), and isoform 2 encoded by exons 1-2-4. The ZIC3 predicted isoform is expected to be encoded by exons 1-2-P1. (C) Minigene plasmids contain a 1,289-bp minigene construct composed of a CMV promoter (brown and white checkerboard pattern), exon 2 (final 40-bp portion, green), intron (140 bp, gray), P1 (255 bp, predicted coding region of P1, red; predicted 3′ UTR region of P1, red and white checkerboard pattern), a 2-bp barcode sequence, and an SV40 poly(A) tail signal sequence (dark yellow and white checkerboard pattern). (Ci) The exon 2 to P1 control construct contains the reference ZIC3 c.1224+3286A sequence predicted to not alter splicing resulting in full intron retention. (Cii) The exon 2 to P1 construct containing the ZIC3 c.1224+3286A>G variant is predicted to generate a 3′ splice acceptor and remove the intron. (D)…

• 

  Figure 3. ZIC3 c.1224+3286A>G variant reduces level of ZIC3 isoform 1 and disrupts RNA splicing (A) Illustrative representation of ZIC3 showing the immunogen region of the ZIC3 polyclonal antibody (pAb) used in (B), which corresponds to the terminal 42 amino acids of exon 3 and the location of forward and reverse primers (blue and red arrows, respectively) used for preliminary splicing analysis in (C) and (D). AA, amino acids. (B) Immunoblotting image of ZIC3 isoform 1 (∼55 kDa) detected in cell lysates from H1-OCT4-eGFP human embryonic stem cells (hESCs). The ZIC3 c.1224+3286A>G clones (ZIC3 AtoG_C1 and C2, respectively), as well as the ZIC3 knockout clones (ZIC3 KO_C1 and C2, respectively) were generated by CRISPR-Cas9 technology. The ZIC3 WT denotes non-edited H1-OCT4-eGFP hESC. Based on sequence homology, the bands at ∼60–65 kDa and ∼45 kDa might correspond to ZIC2 and ZIC4, respectively. GAPDH served as loading control. (C) Electrophoretogram of amplicons obtained by RT-PCR from cDNA amplification of ZIC3 WT, ZIC3 AtoG_C1, and ZIC3 AtoG_C2 cells. Forward and reverse primers are in exon 1 and P1, respectively (blue and red arrows in A). NTC, no-template control. The asterisk denotes a…

• 

  Figure 4. Unique isoforms generated by splicing events in ZIC3 AtoG_C1 cells (A and B) Integrative Genomics Viewer image of the chromosome X range 137,566,444–137,578,187 (GRCh38/hg38) showing Sashimi plots for ZIC3 generated from short-read (Ai and Bi) and long-read (Aii and Bii) RNA-seq analysis from ZIC3 WT (green) and ZIC3 AtoG_C1 (red) cells. The junction coverage minimum was set to 20 and 10 for short and long reads, respectively. The ZIC3 c.1224+3286 genomic position is denoted with an arrow. (C) Illustrative diagram depicting splicing events caused by the ZIC3 c.1224+3286A>G variant. Exons generated by the ZIC3 c.1224+3286A>G variant are illustrated as follows: exon 3T(170) (orange, a 170-bp truncated exon 3), exon 3A(163) (gold, a 163-bp alternative exon located in the 3′ UTR of exon 3), exon 3A(227) (dark blue, a 227-bp alternative exon located in the 3′ UTR of exon 3), P1(57) (red, a 57-bp P1 where the 3′ splice acceptor is caused by the ZIC3 c.1224+3286A>G variant), P2(151) (pink, a 151-bp pseudoexon 2), and exon 4L(1792) (dark pink, a 1,792-bp longer version of exon 4).

• 

  Figure 5. Differential expression analysis suggests ZIC3 AtoG_C1 and ZIC3 KO_C1 cells have similar gene profiles (A) Multidimensional scaling plot of undifferentiated ZIC3 WT (n = 3; blue), ZIC3 AtoG_C1 (n = 3; green), and ZIC3 KO_C1 (n = 3; red) H1-OCT4-eGFP cells. Each data point represents one RNA-seq sample, while the distance between any two samples corresponds to the leading logFC (base 2 logarithm of fold change, the average of the largest absolute logFC). (B and C) Volcano plots of DE genes between (B) undifferentiated ZIC3 WT vs. ZIC3 AtoG_C1 cells and (C) undifferentiated ZIC3 WT vs. ZIC3 KO_C1 cells. Blue and red dots denote downregulated and upregulated genes, respectively. An FDR-adjusted p-value cutoff of 0.01 was used to denote DE genes and the total numbers of downregulated and upregulated genes are shown. (D) Venn diagram of the total number of DE genes between each comparison. (E) Heatmap of the 40 genes that were DE in both ZIC3 AtoG_C1 cells and ZIC3 KO_C1 cells relative to ZIC3 WT cells.

• 

  Figure 6. Cellular localization of ZIC3 isoforms (A–G) Representative images of HeLa cells transfected with HA-tagged ZIC3 plasmids encoding either (A) WT, (B and C) previously published single-nucleotide variants (p.H286R and p.T323M), or (D–G) the coding sequence of ZIC3 isoforms generated by the ZIC3 c.1224+3286A>G variant. Cells were incubated with phalloidin (cytoplasmic marker, green), a rabbit α-HA tag (anti-HA, red), and DAPI (nuclear marker, blue). Merged images display nuclear localization of the HA-tagged ZIC3 isoforms in light purple color. (H) The cellular localization was classified as either nuclear (white), cytoplasmic (black), or mixed (gray, both nuclear and cytoplasmic) and the results are presented as percentages. Transfections were performed in n = 3 separate experiments and at least 100 cells were imaged for each transfection each time. Images were randomized and deidentified for unbiased scoring and statistical analysis was conducted using a Kruskal-Wallis test followed by a Dunn’s test for multiple comparisons. ns, not significant.

• 

  Figure 7. ZIC3 isoforms display differential expression and SV40 promoter activity (A) Western blot image of HEK293 cells transfected with HA-tagged ZIC3 plasmids encoding ZIC3 isoform 1 (ZIC3 WT), the coding sequence of four ZIC3 isoforms generated by the ZIC3 c.1224+3286A>G variant, or two previously published ZIC3 single-nucleotide missense variants (p.T323M and p.H286R). An untransfected control was also included. HA-tagged ZIC3 was detected using an antibody against the HA tag (α-HA). GAPDH served as a loading control. (B) pGL3-SV40 firefly (SV40) luciferase reporter activity in HEK293 cells transfected with HA-tagged pHM6 plasmids encoding ZIC3 isoform 1 (WT), the two ZIC3 missense variants described above, or the coding sequence of abnormal ZIC3 isoforms. The pHM6-empty and pGL3-Basic without the SV40 promoter (no SV40) vectors served as controls. Results are presented as the mean of relative luminescence units (Firefly/Renilla) ± standard error of the mean (SEM) from n = 3 independent experiments. Statistical analysis was conducted using ANOVA followed by Tukey’s test for multiple comparisons. ∗p < 0.05; ∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns, not significant.

• 

  Figure 8. Situs abnormalities in X. laevis embryos injected with abnormal ZIC3 isoforms (A–D) Representative images of X. laevis tadpoles (stage 47) that received at the two-cell stage one of the following: no injection (uninjected control [uninj. ctrl]) (E) in vitro synthesized mRNA encoding the coding sequence of HA-tagged ZIC3 isoform 1 (ZIC3 WT, 50 pg/cell; 100 pg/embryo), in vitro synthesized mRNA encoding the coding sequence of one of the four HA-tagged ZIC3 c.1224+3286A>G isoforms (50 pg/cell; 100 pg/embryo). Situs defects were assessed by the position of the heart, gallbladder, and gut and categorized into one of four groups. (A) Normal situs tadpoles display normal heart looping (green dashed line), normal right gut origin and counterclockwise gut coil (yellow dashed line), and normal position of the gallbladder on the right (red dashed line). (B) Situs inversus tadpoles exhibit reversed heart looping, left gut origin with clockwise gut coil, and leftward gallbladder. (C) Isolated situs anomaly tadpoles have one organ defect (right-origin gut coil with clockwise rotation), while (D) heterotaxy tadpoles have two or more organ defects (reversed heart looping, a left gallbladder…

• 

  Figure 9. ZIC3 is partially rescued in ZIC3 AtoG_C1 cells by splicing blocking vivo-morpholino (MO) (A) Schematic diagram showing the splice-blocking (SB) vivo-MO sequence and mRNA-binding site overlapping the 3′ splice acceptor generated by the ZIC3 c.1224+3286A>G variant. The mutated “g” is shown in bold red, while the “ag” cryptic splice acceptor caused by the variant is underlined. (B) Immunoblotting image of ZIC3 (∼55 kDa) detected in cell lysates from ZIC3 WT and ZIC3 AtoG_C1 cells. ZIC3 AtoG_C1 cells received either no treatment (NT), SB vivo-MO, or a scramble (SCR) vivo-MO for 24 or 48 h. Based on sequence homology, the bands at ∼60–65 and ∼45 kDa might correspond to ZIC2 and ZIC4, respectively. GAPDH served as loading control. (C) Relative levels of ZIC3 from ZIC3 AtoG_C1 cells that received NT (light gray), an SCR vivo-MO (dark gray), or an SB vivo-MO (black) for either 24 or 48 h. compared to ZIC3 from ZIC3 WT cells (white bar). Levels of ZIC3 were normalized to their respective GAPDH levels from n = 1 experiment.