Impaired human hematopoiesis due to a cryptic intronic GATA1 splicing mutation

Abstract

Studies of allelic variation underlying genetic blood disorders have provided important insights into human hematopoiesis. Most often, the identified pathogenic mutations result in loss-of-function or missense changes. However, assessing the pathogenicity of noncoding variants can be challenging. Here, we characterize two unrelated patients with a distinct presentation of dyserythropoietic anemia and other impairments in hematopoiesis associated with an intronic mutation in GATA1 that is 24 nucleotides upstream of the canonical splice acceptor site. Functional studies demonstrate that this single-nucleotide alteration leads to reduced canonical splicing and increased use of an alternative splice acceptor site that causes a partial intron retention event. The resultant altered GATA1 contains a five-amino acid insertion at the C-terminus of the C-terminal zinc finger and has no observable activity. Collectively, our results demonstrate how altered splicing of GATA1, which reduces levels of the normal form of this master transcription factor, can result in distinct changes in human hematopoiesis.

Figure 1.

Identification of a GATA1 intronic mutation in two unrelated patients with dyserythropoietic anemia. (A) Images of bone marrow aspirates from patients 1 and 2 are shown as images taken with 40× and 60× objectives, respectively. Notable features in both patients include mild to moderate dyserythropoiesis and other alterations, including hypolobated dysplastic megakaryocytes. (B) The thrombin-induced expression of CD62 and CD63 on platelets from patient 1 or a healthy control was detected by flow cytometry (shown as a percentage). There is defective expression of these antigens, indicating an α-/δ-granule deficiency. (C) Sequencing chromatograms of mutation in GATA1 (chrX:48,652,176 C>T in hg19) from a healthy control, the two patients, and their mothers. Scale bars: 10 µm.

Figure 2.

Decreased canonical splicing and intron retention due to a pathogenic GATA1 mutation. (A) Schematic of the minigene assay involving exons 5 and 6 and the intervening intron of GATA1. The minigene vector, pSpliceExpress, includes two exons from rat insulin (Rat Ins Ex) as a control for splicing, a SV-40 promoter (Prom), and a polyadenylation site (PolyA). (B) Representative RT-PCR from the minigene assay in the presence or absence of the mutation shows the reduction of canonical splicing and presence of an intron retention event, which was confirmed through cloning and Sanger sequencing of the resultant products. (C) Semiquantitative RT-PCR analysis of the minigene assay using 28, 30, 32, and 34 PCR cycles. The bar graph (below) depicts quantified levels of the various products, including the total amount of GATA1 (red), the amount of the wild-type band (blue), or the amount of the mutant (Mut) band in the setting of the mutation involving an intron retention event (***, P < 0.001; n = 4). (D) RT-PCR analysis of GATA1 exon 5 and exon 6 from control and patient peripheral blood mononuclear cells. (E) Representative RT-PCR analysis of GATA1 exons 5 and 6 from human HSPCs undergoing erythroid differentiation. (F) Schematic demonstrating alternative splicing in the fifth intron of GATA1, including the location of the intronic mutation and the nucleotides involved in intron retention (red). The splicing patterns are depicted above with a red dashed (alternative) or solid black (canonical) line. (G) RT-PCR analysis of the minigene assay performed as above or with increased expression of wild-type SF3B1 or the K700E SF3B1 mutated protein in 293T cells.

Figure 3.

Rare intron retention events during physiological human erythroid differentiation. (A) Sashimi plots showing splicing events in the GATA1 locus across the indicated stages of human erythropoiesis as determined by RNA-seq analysis. (B) Quantification of rare events of intron retention at indicated stages as counts of junction-spanning reads with the percentage of total transcript reads this represents shown in parentheses. Alternate and canonical splice acceptors are highlighted in pink. One representative example shown per stage of erythropoiesis.

Figure 4.

GATA1 variant produced through the intron retention event is stably expressed, but shows no function. (A) Schematic shows insertion of five amino acids as a result of the intron retention event. (B) A histogram plot (left) shows GFP levels indicating robust expression and equivalent transfection efficiency (GFP linked to GATA1 cDNA by internal ribosomal entry site). Mutant and wild-type GATA1 mRNA are expressed at equivalent levels (right). (C) A representative Western blot of GATA1 levels from exogenous cDNA expression in 293T cells. Loading control is GAPDH. (D) Western blot of GATA1 levels from exogenous cDNA expression in erythroid G1E cells (three biological replicates per indicated construct). Loading control is GAPDH. (E) Representative flow cytometric assessment (of three replicates) of erythroid differentiation (using CD235a, CD71, and CD49d surface markers) on day 7 and day 12 of culture of human HSPCs upon exogenous expression of wild-type GATA1 or the GATA1 mutant protein. (F) Representative histogram plots (of three replicates) assessing the marker of mouse erythroid differentiation, Ter119, in G1E cells infected with the empty vector (control), GATA1 wild type, or GATA1 mutant cDNAs vectors after 48, 72, or 96 h.

Figure 5.

GATA1 variant shows complete loss of transcriptional activity. (A) Volcano plot showing log₂ fold change (FC) of differentially expressed genes between GATA1 wild type–infected and control cells at indicated P values. (B) Volcano plot showing log₂ FC of differential genes between GATA1 mutant infected and control cells at indicated P values. (C) Dot plot showing equivalent expression of GATA1 wild type and mutant based on RNA-seq for three biological replicates per condition. cpm, counts per million. (D) RNA-seq coverage of the region containing the intron event showing GATA1 wild-type cDNA and GATA1 mutant cDNA (biological replicate 1 for each condition shown). (E) Principal component plot of RNA-seq replicates (symbol) by condition (color) showing PC1 and PC2 and the percent variance explained by each. This reveals that the GATA1 mutant clusters with the control, while the GATA1 wild type clusters separately. (F) Heatmap showing clustering for the top genes (n = 232) differentially expressed between GATA1 wild type and empty HMD control (log₂FC > 2, p.adj < 0.01), which shows three distinct clusters. (G) Depiction of GATA1 chromatin occupancy in 60-bp bins in a 6-kb window surrounding the TSS for the principal isoform of the indicated genes as in F shows an enrichment in cluster K1 genes. The darker red color in this plot indicates greater chromatin occupancy by GATA1 in the G1E cells. For RNA-seq experiments, three biological replicates shown unless otherwise indicated.