Genoprime-Perturb-seq: Functional dissection of GATA2 variants via simultaneous Genotyping-coupled Prime-editing Perturb-seq

Abstract

Abstract The master transcription factor GATA2 is indispensable for hematopoietic stem cell homeostasis and the integrity of multilineage blood development. Germline heterozygous mutations in GATA2 result in GATA2 haploinsufficiency, leading to GATA2 deficiency syndrome—a complex, pleiotropic disorder characterized by profound immunodeficiency and a high predisposition for progression to MDS and AML. Despite its clinical significance, current limitations in public variant annotation hinder the ability to assess individual GATA2 variants in clinical practice. For example, in the ClinVar database, 89.9% of GATA2 single nucleotide variants (SNVs) are classified as variants of uncertain significance (VUS) or as those with conflicting interpretations (accessed July 31, 2025), underscoring the urgent need for alternative approaches to assess their pathogenicity. While multiple methods combining CRISPR perturbations with scRNA-seq have been reported, to our knowledge, no existing technology enables the systematic introduction of any desired missense mutation alongside simultaneous assessment of the resulting genotype and transcriptomic changes at single-cell resolution. Here, we developed GenoPrime-Perturb-seq (simultaneous Genotyping-coupled Prime-editing Perturb-seq), a novel platform for high-throughput prime editing integrated with single-cell genotyping and transcriptomic profiling. This approach, for the first time, enables the functional interrogation of individual GATA2 variants at single-cell resolution. GATA2 is a critical regulator of basophil lineage commitment, and its loss has been shown to skew hematopoietic progenitors toward neutrophil and/or eosinophil lineages. We found that the KU812 line, an immature basophilic leukemia cell line derived from a CML patient, showed upregulation of key granulocytic transcription factors such as CEBPA following GATA2 depletion. To validate the capacity of GenoPrime-Perturb-seq, we first applied it to KU812 cells by introducing either a canonical loss-of-function (R330*) or gain-of-function (L359V) GATA2 variant. We confirmed efficient on-target editing of each variant with high-fidelity and detected transcriptomic changes at single-cell resolution: cells harboring the R330* and L359V mutations segregated into distinct clusters, as expected. For example, the R330*-enriched cluster exhibited a transcriptome similar to that of neutrophil- and/or eosinophil-like lineages, marked by the upregulation of CEBPA and CEBPE. To assess the functional significance of individual GATA2 VUS within the dual zinc-finger domain, where most reported variants concentrated, we performed GenoPrime-Perturb-seq targeting 149 GATA2 SNVs listed in the ClinVar and MGeND databases. Strikingly, GenoPrime-Perturb-seq demonstrated high classification accuracy, achieving an AUC >0.9 in ROC analysis, with both sensitivity and specificity exceeding 0.9 at the Youden index threshold, when ClinVar-designated pathogenic variants and bootstrapped wild-type data were used as positive and negative labels, respectively. Furthermore, GenoPrime-Perturb-seq showed significantly superior performance in detecting cells differentiating towards the neutrophil- and eosinophil-like lineages compared to a method combining scRNA-seq with prime editing guide RNA capture. Finally, to validate the results of GenoPrime-Perturb-seq with an orthogonal system, we performed a bulk prime editing screen on 149 variants using an iPSC-to-hematopoietic progenitor cell (HPC) differentiation model, in which GATA2 knockout has been reported to block differentiation toward HPCs. As expected, the majority of the variants classified as pathogenic by GenoPrime-Perturb-seq resulted in impaired HPC formation. However, unexpectedly, some “loss-of-function” variants—defined by the KU812 screen—enhanced HPC development, suggesting that certain GATA2 variants may exert distinct functions in a differentiation stage-specific manner. Notably, these variants exhibited a pronounced block in neutrophil maturation in the KU812 assay and were clustered within the ZF1 domain. In conclusion, we have developed GenoPrime-Perturb-seq, a novel platform that enables high-throughput, high-fidelity functional assessment of variants. This system successfully evaluated the functional significance of individual GATA2 variants and may help bridge the gap between genomic variation and clinical phenotype, offering a promising approach for resolving a wide range of uncertain variants.