RNA-coupled CRISPR Screens Reveal ZNF207 as a Regulator of LMNA Aberrant Splicing in Progeria

Abstract

Despite progress in understanding pre-mRNA splicing, the regulatory mechanisms controlling most alternative splicing events remain unclear. We developed CRASP-Seq, a method that integrates pooled CRISPR-based genetic perturbations with deep sequencing of splicing reporters, to quantitively assess the impact of all human genes on alternative splicing from a single RNA sample. CRASP-Seq identifies both known and novel regulators, enriched for proteins involved in RNA splicing and metabolism. As proof-of-concept, CRASP-Seq analysis of an LMNA cryptic splicing event linked to progeria uncovered ZNF207, primarily known for mitotic spindle assembly, as a regulator of progerin splicing. ZNF207 depletion enhances canonical LMNA splicing and decreases progerin levels in patient-derived cells. High-throughput mutagenesis further showed that ZNF207's zinc finger domain broadly impacts alternative splicing through interactions with U1 snRNP factors. These findings position ZNF207 as a U1 snRNP auxiliary factor and demonstrate the power of CRASP-Seq to uncover key regulators and domains of alternative splicing.

Keywords: Alternative pre-mRNA splicing; BuGZ; CRASP-Seq; Hutchinson-Gilford Progeria Syndrome (HGPS); Nonsense-mediated mRNA decay (NMD); PKM; RNA-coupled CRISPR screen; SRSF7; ZNF207; pre-mRNA processing.

Figure 1:. Overview of the CRASP-Seq platform for identifying alternative splicing regulators.

(A) Schematic of the CRASP-Seq vector, which integrates constitutive Cas9 and Cas12a hgRNA expression under a U6 promoter for genome editing. A doxycycline (DOX)-inducible TRE (Tetracycline Response Element) promoter controls the expression of a splicing minigene reporter. The reverse tetracycline transactivator (rtTA) is fused to a puromycin resistance gene via a T2A cleavage sequence, allowing selectable, inducible reporter expression. An RNA polymerase II termination signal ensures efficient expression. BGH: bovine growth hormone polyadenylation signal. (B) RT-PCR validation of alternative splicing outcomes for FAS exon-6 and EZH2 exon-14 using CRASP-Seq vectors in HAP1 cells co-expressing the indicated hgRNAs. (C) Workflow of the CRASP-Seq platform, illustrating the generation of splicing-derived mRNA products and the subsequent preparation of Illumina libraries for paired-end sequencing. This approach enables the precise association of each hgRNA within the library with quantified splicing outcomes. (D) Gene Ontology (GO) enrichment analysis of regulators identified by CRASP-Seq for the FAS and EZH2 splicing events. Only significantly enriched biological processes are shown. All the terms are listed in Figure S1G.

Figure 2:. Genome-wide identification of splicing regulatory factors associated with disease-related events.

(A) Schematic representation of the five alternative splicing events profiled using the CRASP-Seq platform in this study. (B) Heatmap of ΔPSI values for CRASP-Seq screen hits across the five alternative splicing events profiled in HAP1 and RPE1 cell lines. Functional annotations of genes are shown above the heatmap. (C) Flower plot depicting the number of splicing regulators identified for each alternative splicing event. Regulators shared by at least two events are displayed at the center of the plot. (D) Distribution of splicing regulatory hits across the five profiled events. Shared regulators, defined as hits observed in two or more events, are highlighted (n = 103). (E) Heatmap of pairwise correlations of ΔPSI values for the five splicing reporter events in HAP1, RPE1, and HepG2 cells. Sub-clusters are annotated with molecular signatures database (MSigDB) terms, with adjusted p-values for enrichment shown. ****p-value < 0.0001; Individual genes within sub-clusters are shown on the right.

Figure 3:. CRASP-Seq identifies ZNF207 as a regulator of progerin aberrant splicing.

(A) Heatmap showing ΔPSI values of mutant LMNA CRASP-Seq screen hits in HAP1, RPE1, and HepG2 cells. (B) Schematic representation of the ZNF207 protein structure, highlighting the C2H2 zinc finger (ZnF) domain, the charged helical structure, low-complexity regions, and the Gle2-binding-sequence (GLEBS) motif. (C) RT-PCR analysis of aberrant LMNA splicing in RNA from HGPS patient-derived immortalized fibroblasts treated with three indepen-dent siRNAs targeting ZNF207. Quantifications of PSI values from three independent experiments are shown below the gel. Data are presented as mean ± standard deviation (SD). ****p-value < 0.0001; Dunnett's multiple comparisons test following one-way ANOVA. (D) RT-PCR analysis of LMNA splicing in HGPS patient fibroblasts treated with ZNF207-targeting siRNA and/or ectopically expressing a ZNF207 ORF resistant to siRNA for nine days. Quantification of PSI values from three independent experiments are shown below the gel. Data are presented as mean ± SD. ****p-value < 0.0001; Dunnett's multiple comparisons test following one-way ANOVA. (E) Western blot analysis of lamins in HGPS patient-derived immortalized fibroblasts transduced with either a siRNA-resistant 3×FLAG C-terminal-tagged ZNF207 ORF or an empty vector. Cells were treated with either non-targeting control siRNA or siRNA targeting endog-enous ZNF207 (siZNF207) for 16 days. Blots were probed with antibodies against lamins, ZNF207, and GAPDH (loading control). Relative progerin expression was quantified and normalized to β-actin, with values further normalized to the corresponding siControl-treated sample. Quantifications from three independent experiments are presented on the right. Data are presented as mean ± SD. *p-value < 0.05, ****p-value < 0.0001; Uncorrected Fisher's LSD following one-way ANOVA.

Figure 4:. ZNF207 depletion induces widespread alternative splicing changes.

(A) Distribution of alternative splicing event types (CE: cassette exons, Alt3: alternative 3′ splice sites, Alt5: alternative 5′ splice sites, IR: retained introns) detected by RNA-seq following ZNF207 knockdown for 48 hours with three independent siRNAs in HEK293T cells. Events with significant splicing changes (|ΔPSI| ≥ 10; probability ≥ 0.95) upon siZNF207 knockdown are included. (B) Western blot analysis of ZNF207 in HEK293 Flp-In cells expressing doxycycline-inducible, siRNA-resistant 3×FLAG C-terminal tagged ZNF207 ORF. Cells were treated with control siRNA or siRNA targeting endogenous ZNF207 (siZNF207) for 48 hours. Blots were probed with antibodies against ZNF207, FLAG, and GAPDH (loading control). (C) RNA-seq analysis of splicing changes (ΔPSI) upon ZNF207 knockdown in HEK293 Flp-In cells. Comparisons include knockdown of ZNF207 relative to non-targeting siRNA treatment without induction (siZNF207) and ZNF207 ORF rescue expression in siRNA-treated cells compared to expression of empty vector (siZNF207 + rescue). Events with significant splicing changes (|ΔPSI| ≥ 10; probability ≥ 0.95) upon siZNF207 knockdown and rescue by ZNF207 ORF expression are highlighted. (D) RT-PCR analysis of ACLY (left) and SRRM1 (right) alternative splicing in HEK293 Flp-In cells treated with ZNF207-targeting siRNA and/or expressing siRNA-resistant C-terminally 3xFLAG-tagged ZNF207 ORF. Quantification of PSI values from three independent experiments are shown below the gel. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ****p < 0.0001; Dunnett's multiple comparisons test following one-way ANOVA. (E) Gene expression analysis (Z score normalized) from RNA-seq profiling of HEK293 Flp-In cells treated with ZNF207-targeting siRNA and/or rescued with siRNA-resistant ZNF207 ORF expression. Only genes with significant expression changes (adjusted p < 0.05; |log2FC| ≥ 0.35) upon siZNF207 and rescued by ZNF207 expression are shown. (F) RT-PCR analysis of LUC7L, LUC7L2, SNRNP70, and RBM3 alternative splicing in HEK293 Flp-In cells treated with ZNF207-targeting siRNA and/or expressing siRNA-resistant ZNF207 ORF. Experiments were conducted with and without the NMD inhibitor SMG1i (0.5 μM, 6 hrs). Quantification of PSI values from three independent experiments are shown below the gel. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; Šídák's multiple comparisons test following one-way ANOVA.

Figure 5:. ZNF207 associates with the splicing machinery and binds to regulated splicing targets.

(A) Overview of the TurboID experimental pipeline. HEK293 Flp-In cells expressing C-terminal miniTurbo-tagged ZNF207 were induced for 24 hours and incubated with biotin for 2 hours. Proximal biotinylated proteins were purified and analyzed via mass spectrometry. (B) Proximal interaction network of ZNF207 identified by TurboID. Only significant interactors consistently detected across all three replicates are shown (Table S7). (C) Gene Ontology (GO) enrichment analysis of biological processes among ZNF207 TurboID proximal interactors with FDR < 0.05. (D) Western blot analysis of total cell lysates (input) treated with benzonase, and FLAG immunoprecipitates (IP: FLAG-M2) from HEK293 Flp-In cells expressing 3×FLAG-ZNF207. Blots were probed with antibodies against FLAG, SNRPA, SNRNP70, RBM25, SF3B1, SNRNP200, PRPF6, SNRPF, and GAPDH (negative control). (E) Bar plot showing the distribution of ZNF207 eCLIP peaks across RNA biotypes. Only biotypes with at least 10 ZNF207 peaks are included. eCLIP reads were normalized to input, and peaks were called using the DEWSeq pipeline (see Methods for details). (F) Stacked bar chart showing the percentage of intronic and exonic ZNF207 binding peaks. (G) Average eCLIP signal profiles of C-terminal FLAG-tagged ZNF207 for ZNF207-regulated cassette exons (n = 715) and unchanged alternative exons (n = 3,745). Profiles include the negative control GFP and N-terminal FLAG-tagged ZNF207 (devoid of splicing activity) for the same exon subsets.

Figure 6:. High-Throughput mutagenesis reveals a critical residue required for ZNF207-dependent splicing regulation.

(A) Schematic of the CRASP-Seq base editor tiling approach. A lentiviral vector enables constitutive expression of sgRNAs while a doxycycline-inducible TRE promoter drives the expression of a mutant LMNA minigene reporter. Splicing-derived mRNA products are analyzed via Illumina paired-end high-throughput sequencing. (B) ΔPSI scores from gene-level analysis of all guides in the base editor library, including those targeting splice sites, across the 39 profiled genes. Intergenic and non-targeting controls are also included. The genes are arranged in ascending order of their ΔPSI scores, from left to right. (C) ΔPSI scores for base editor sgRNAs targeting ZNF207 coding sequence, excluding splice-site-overlapping guides. Average ΔPSI scores for individual sgRNAs within 10-amino-acid windows are plotted, with the scores of individual sgRNAs depicted as black dots. The ZNF207 protein is illustrated below, showing ZnF domains (blue), GLEBS motif (orange), and low-complexity regions (purple). (D) Schematic representation of ZNF207 constructs: full-length wild-type ZNF207 (top), full-length K42E point mutant (middle), and the ΔZnF truncation mutant lacking the zinc finger domain (bottom). (E) RT-PCR analysis of ACLY (left) and SRRM1 (right) alternative splicing in HEK293 Flp-In cells treated with ZNF207-targeting siRNA and/or expressing siRNA-resistant wild-type or truncation mutant ZNF207 ORFs. Quantification of ΔPSI values from three independent experiments are shown below the gel. Data are presented as mean ± SD. Statistical comparisons were performed relative to the siZNF207 sample without ZNF207 ORF expression. ****p < 0.0001; Dunnett's multiple comparisons test following one-way ANOVA. (F) Western blot analysis of total cell lysates (input) treated with benzonase, and FLAG immunoprecipitates (IP: FLAG-M2) from HEK293 Flp-In cells expressing 3×FLAG-tagged ZNF207 variants. Blots were probed with antibodies specific for FLAG, SNRPA, SNRNP70, RBM25, SF3B1, and GAPDH (negative control). (G) Quantification of co-immunoprecipitated spliceosome components, normalized to FLAG-ZNF207 pulldown efficiency and wild-type (WT) ZNF207 levels. Data are presented as mean ± SD. Statistical comparisons were performed relative to WT-ZNF207. ***p < 0.001, **p < 0.01, *p < 0.05; One-way ANOVA with Dunnett’s multiple comparisons correction.

Figure 7:. Model of ZNF207 as a regulator of alternative splicing.

Illustration of the proposed mechanism by which ZNF207 regulates alternative splicing. ZNF207 interacts with U1 snRNP and other core spliceosome components, binding directly to target pre-mRNAs. These interactions influence splice site selection, promoting the activation of specific exons.