A chromatin-focused CRISPR screen identifies USP22 as a barrier to somatic cell reprogramming

Abstract

Cell-autonomous barriers to reprogramming somatic cells into induced pluripotent stem cells (iPSCs) remain poorly understood. Using a focused CRISPR-Cas9 screen, we identified Ubiquitin-specific peptidase 22 (USP22) as a key chromatin-based barrier to human iPSC derivation. Suppression of USP22 significantly enhances reprogramming efficiency. Surprisingly, this effect is likely to be independent of USP22's deubiquitinase activity or its association with the SAGA complex, as shown through module-specific knockouts, and genetic rescue experiments. USP22 is not required for iPSC derivation or maintenance. Mechanistically, USP22 loss during reprogramming downregulates fibroblast-specific genes while activating pluripotency-associated genes, including DNMT3L, LIN28A, SOX2, and GDF3. Additionally, USP22 loss enhances reprogramming efficiency under naïve stem cell conditions. These findings reveal an unrecognized role for USP22 in maintaining somatic cell identity and repressing pluripotency genes, highlighting its potential as a target to improve reprogramming efficiency.

Fig. 1. EpiDoKOL screen identified USP22 as a barrier to human somatic cell reprogramming.

A Schematic representation of the screen outline. B Positively selected gRNAs during reprogramming identified by MAGeCK algorithm at the gene level. Created in BioRender. Onder, T. (2025) https://BioRender.com/j82z183. C Fold change in reprogramming efficiency upon gRNA expression. Triangular images above the bars are sections of representative Tra-1-60-stained wells. Error bars indicate the error of mean. n = 3, independent experiments were conducted for PAXIP1 gRNA expression and n = 12, independent experiments for USP22 gRNA expression. Two-sided t-test p-values 0.0001 and 0.0081 for USP22 g1 and g2, respectively, and 0.1685 and 0.1466 for PAXIP1 g1 and g2, respectively. D Western blot image showing USP22 protein level of USP22 targeting gRNAs in fibroblasts. Tubulin was used as a loading control. E Fold change in reprogramming efficiency upon USP22 targeting gRNA expression. Triangular images above the bars are sections of representative Tra-1-60-stained wells. Error bars indicate the error of mean. n = 11, independent experiments were conducted for USP22 sg1 expression, n = 10 for USP22 sg2 expression and n = 5 for USP22 sg3 expression. Two-sided t-test to compare g1, g2, and g3 to the gNT1 gives following p-values 0.0001, 0.0111, and 0.0013, respectively.

Fig. 2. Catalytic activity of USP22 is not a barrier to reprogramming.

A Diagram showing USP22 domains and mutated amino acid positions. B Western blot image showing USP22 protein levels after USP22 overexpressions in both non-targeting and USP22 targeting gRNA expressing fibroblasts. Actin was used as a loading control. C Western blot image showing H2Bub protein levels after USP22 overexpressions in both non-targeting and USP22 targeting gRNA expressing fibroblasts. Histones H3 and H2B were used as loading controls. D Fold change in reprogramming efficiency upon USP22 overexpressions in both wild-type and USP22 knockout background. Triangular images above the bars are sections of representative Tra-1-60-stained wells. Error bars indicate the error of mean. n = 3, independent experiments for KR mutations, n = 4 for CA and KQ mutations in USP22 knockout background, and n = 5 for other comparisons. Two-sided t-test was performed between sgNT1 and USP22 sg1 without any overexpression and p-value is 0.0153. Two-sided t-test was performed between USP22 sg1 expressing cells without overexpression and WT, C185A, K129Q, and K129R and p-value is 0.0129, 0.0054, 0.0264, and 0.0052, respectively. E Western blot image showing H2Bub and H2Aub protein levels after ATXN7L3 or ENY2 knockouts in fibroblasts. H2B was used as a loading control. F Fold change in reprogramming efficiency upon ATXN7L3, ENY2, USP27X, or USP51 knockouts. Triangular images above the bars are sections of representative Tra-1-60-stained wells. Error bars indicate the error of mean. n = 5, independent experiments. Two-sided t-test shows p-value smaller than 0.0001 for ENY2 sg1 and sg2 compared to sgNT1.

Fig. 3. USP22 loss results in differentiation defects.

A Schematic representation of generating USP22 knockout iPSC clones used in Fig. 3B. B Western blot image showing USP22 protein levels in iPSCs generated from Cas9 and USP22 targeting gRNA expressing fibroblasts. Tubulin was used as a loading control. C Western blot image showing USP22 protein levels in iPSC clones generated from Cas9 and USP22 targeting gRNA expressing iPSCs. Tubulin was used as a loading control. D Relative expression levels of lineage markers upon EB formation for NT1#2, USP22KO#9-10. E Hematoxylin and eosin-stained sections of teratomas of NT1#2, USP22KO#9-10 iPSCs. Arrowheads point to representative tissues from endoderm (glandular epithelium), ectoderm (neuroepithelium), and mesoderm (cartilage) lineages. 2 mice were injected per cell line.

Fig. 4. USP22 loss stabilizes pluripotency network during reprogramming.

A Volcano-plot of RNA-seq data comparing USP22 KO and control cells during reprogramming. B Gene ontology analysis for differentially regulated genes upon USP22 knockout. Top 5 upregulated and downregulated gene sets are shown. C GSEA p-value and normalized enrichment scores on pre-ranked gene lists according to log2FC value for comparisons of reprogramming fibroblasts expressing USP22 gRNA to NT1 for all gene sets available at The Molecular Signatures Database (mSigDB). D GSEA results for pluripotency- and fibroblast-related gene sets upon USP22 loss during reprogramming. E Relative expression levels of pluripotency-associated genes on day 9 of reprogramming in control (NT1), USP22 KO cells expressing either wild-type or catalytic dead mutant USP22 cDNAs. n = 3 biological replicates with two technical replicates each. F Number of Tra-1-60- and/or KLF17-positive colonies generated under naïve conditions (PGXL) from control or USP22 KO fibroblasts. Immunofluorescence images of KLF17 and TRA-1-60-stained cells are shown below. G Relative expression of naïve and primed makers in USP22 KO cells after reprogramming under naïve PSC culture conditions. Expression values are normalized to primed PSCs.