Workflow of the CNV study and additional individuals/families with ZFHX4 variants.

A Workflow of the CNV study. CNVs were called from an existing dataset including exome sequencing data for 50 nsCL/P trios (Ishorst et al. 2022). CNV filtering was performed as follows. (I) The CNV frequency threshold was set to ≤10% among all 50 nsCL/P trios. (II) De novo CNV calling was performed using PLINK/Seq. (III) Only CNVs that were extracted using either CoNIFER and/or EXCAVATOR2 in addition to XHMM were retained in the analysis. CNVs that (IV) did not span RefSeq Genes, (V) overlapped by >50% with regions of segmental duplication or >80% with genes in difficult to analyze regions (Segmental Dups track, UCSC Genome Browser on Human Feb. 2009 (GRCh37/hg19) Assembly), and (VI) had population frequencies ≥0.01 were excluded. B Pedigree of index trio with heterozygous de novo CNV in ZFHX4. C Exon and protein domain structure of human ZFHX4. Exons are colored in alternating white and black, and positions of the start codon (ATG) and stop codon (TAG) are marked. ZFHX4 is a transcription factor with four homeodomains (orange) and 23 zinc finger domains (blue). Protein domains are positioned in proportion to the corresponding exons. Red bar, heterozygous ZFHX4 de novo deletion in the index patient with nsCLP. P1-P3, position of primer pairs spanning the CNV for qPCR verification; black arrows, positions of additional ZFHX4 variants including loss-of-function variants from targeted-sequencing and a homozygous missense variant from a collaboration with Girisha & colleagues. D Pedigrees of four additional families (targeted-sequencing study and collaboration with Girisha & colleagues). E Tyr653 is one out of seven conserved residues (drawn bold) in the fold of a canonical zinc-finger motif. F Model of the zinc-finger structure in ZFHX4 proposes the formation of a hydrogen bond between Tyr653 and His662. Mutation of Tyr653 to histidine could change this conformational arrangement of side chains and thus impair zinc-binding. CNV copy number variation, nsCL/P nonsyndromic cleft lip with/without cleft palate, qPCR quantitative polymerase chain reaction, ES exome sequencing.

Expression pattern of Zfhx4 in mouse embryos.

A Heatmap of mean replicates in Reads Per Kilobase Million (RPKM) of Zfhx4 and the established orofacial clefting genes (Cdh1, Irf6, and Grhl3) from bulk RNA-Seq data for mouse secondary palatal shelves (embryonic days E10.5–E14.5). Dot plots of gene expression from single-cell RNA-Seq data for the B whole mouse embryo and C mouse embryo face at E11.5. The color of the dots corresponds to the average scaled expression level. The size of the dots corresponds to the percentage of cells that expressed the gene in the respective cell type.

Effect of F0-generation knock-out (F0-KO) and translation-blocking morpholino knock-down (MO-KD) of zfhx4 in zebrafish larvae (zfl).

A Ventral view of whole-mount Alcian blue-stained zfl at 4 dpf. For F0-KO (left) and MO-KD (right), the spectrum of the phenotypes is shown in blue (– = no phenotype, + = mild phenotype, ++ = strong phenotype). Scale bar = 113 µm. B Survival curves for each group (N = 3). Survival of F0-KO did not differ significantly from that of the Ctrl-Scr group, whereas survival was lower in MO-KD than in Ctrl-MO. Comparisons were performed using Mantel-Cox tests. Survival rates at 4 dpf: UI_CRISPR: 81.1%, Ctrl-Scr: 77.2%, F0-KO: 75.7%, UI_MO: 89.6%, Ctrl-MO: 84.4%, MO-KD: 78.1%. C Percentages of surviving zfl at 4 dpf with craniofacial anomalies (phenotype) and no phenotype in different groups (N = 3). No distinction between phenotype severities was made. F0-KO and MO-KD injection resulted in craniofacial anomalies in 19.13% and 38.17% of zfl, respectively, compared with rates of anomalies of 3.09% in Ctrl-Scr zfl and 0% in UI_CRISPR/MO and Ctrl-MO zfl. Comparisons were performed using two-way analysis of variance (ANOVA) and Tukey’s multiple comparisons tests. Data are presented as means with standard error of the mean (SEM). D PCR validation of the disruption of zfhx4 exon 3 in F0-KO zfl at 4 dpf. Primer binding sites are located on adjacent introns. Expected wt band size, 720 bp. E Liquid chromatography mass spectrometry (LC-MS) analysis of protein lysates from heads of 3 dpf zfl (N = 2). Depicted are Zfhx4 fold changes relative to the 60S ribosomal protein L4 (Rpl4) abundance in each sample. Zfhx4 levels were decreased by 77.1% in F0-KO compared with levels in Ctrl-Scr and by 41.6% in MO-KD compared with levels in Ctrl-MO. Individual data for LC-MS run1 (stars) and run2 (dots) are plotted. Comparisons were performed using two-way analysis of variance (ANOVA) and Tukey’s multiple comparisons tests. Data are presented as means with standard deviation (SD). ns, p > 0.05, **p ≤ 0.01, ***p ≤ 0.001; dpf days post-fertilization, UI_CRISPR/MO uninjected wild-type zfl, Ctrl-MO Control MO-injected zfl, Ctrl-Scr Scrambled control-injected zfl, wt wild-type.

Craniofacial outcomes of zfhx4 F0-generation knock-out (F0-KO) and translation-blocking morpholino knock-down (MO-KD) in zebrafish larvae (zfl).

A Schematic overview of the human primary (green) and secondary palate (light blue) (ventral view). B Schematic overview of the zfl neurocranium with medial (green) and lateral ethmoid (light blue) (ventral view). Both structures are homologous structures to the human palate as indicated by the corresponding colors. C Ventral view of the zfl cartilage structures of the neurocranium (green and light blue) and viscerocranium (dark blue). The eyes are outlined in light gray. D Ventral view of the craniofacial region of whole-mount Alcian blue stained zfl at 4 dpf as represented schematically in C (for a different zfl than that shown in Fig. 3A). Comparison of craniofacial phenotypes between F0-KO zfl against controls (UI_CRISPR and Ctrl-Scr). Scale bar = 200 µm. E Ventral view of the craniofacial region of whole-mount Alcian blue-stained zfl at 4 dpf as represented schematically in C. Comparison of craniofacial phenotypes between MO-KD-injected zfl against controls (UI_MO and Ctrl-MO). Scale bar = 200 µm. F Length of Meckel’s cartilage and ethmoid plate measured as indicated by the pink lines of 4 dpf old, Alcian blue-stained F0-KO- and Ctrl-Scr-injected zfl (mean length Meckel’s: F0-KO: 96.9 μm, Ctrl-Scr: 107.7 μm, difference between means: 10.8 ± 5.4 μm; mean length ethmoid: F0-KO: 103.4 μm, Ctrl-Scr: 108.9 μm, difference between means 5.5 ± 3.1 μm) and MO-KD- and Ctrl-MO-injected zfl (mean length Meckel’s: MO-KD: 79.5 μm, Ctrl-MO: 102.9 μm, difference between means: 23.5 ± 6.3 μm; mean length ethmoid: MO-KD: 91.1 μm, Ctrl-MO: 101.5 μm, difference between means: 10.4 ± 3.2 μm). Comparisons were performed using unpaired, two-tailed Welch’s t tests. G Measurement of the Meckel’s-palatoquadrate (M-PQ) angle as represented by the pink angle, which is a reliable parameter to assess craniofacial outcomes as suggested by Raterman et al. [43]. Comparison of M-PQ angles in 4 dpf old, Alcian blue-stained F0-KO- and Ctrl-Scr-injected zfl (mean M-PQ angle: F0-KO: 45.6°, Ctrl-Scr: 43.1°, difference between means 2.5° ± 1.1°) and MO-KD- and Ctrl-MO-injected zfl (mean M-PQ angle: MO-KD: 58.8°, Ctrl-MO: 46.4°, difference between means 12.5° ± 2.0°). Comparisons were performed using unpaired, two-tailed Student’s t tests. ns, p > 0.05, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. m Meckel’s, pq Palatoquadrate, ch Ceratohyal, hs Hyosympletic, dpf days post-fertilization, UI_CRISPR/MO uninjected wild-type zfl, Ctrl-MO Control MO-injected zfl, Ctrl-Scr Scrambled control-injected zfl.