BRAT1 links Integrator and defective RNA processing with neurodegeneration

Abstract

Mutations in BRAT1, encoding BRCA1-associated ATM activator 1, have been associated with neurodevelopmental and neurodegenerative disorders characterized by heterogeneous phenotypes with varying levels of clinical severity. However, the underlying molecular mechanisms of disease pathology remain poorly understood. Here, we show that BRAT1 tightly interacts with INTS9/INTS11 subunits of the Integrator complex that processes 3' ends of various noncoding RNAs and pre-mRNAs. We find that Integrator functions are disrupted by BRAT1 deletion. In particular, defects in BRAT1 impede proper 3' end processing of UsnRNAs and snoRNAs, replication-dependent histone pre-mRNA processing, and alter the expression of protein-coding genes. Importantly, impairments in Integrator function are also evident in patient-derived cells from BRAT1 related neurological disease. Collectively, our data suggest that defects in BRAT1 interfere with proper Integrator functions, leading to incorrect expression of RNAs and proteins, resulting in neurodegeneration.

Fig. 1. BRAT1 interacts with the Integrator catalytic cleavage heterodimer.

a A volcano plot of proteins enriched in BRAT1 immunoprecipitates from U2OS cells. The –Log₁₀ (Student’s t-test FDR) is plotted against the difference of mean intensities between wild-type and BRAT1^−/− samples. Red dots indicate significantly enriched proteins (intensity difference >1; FDR < 0.05). b, c Levels of BRAT1, INTS11 and INTS9 in BRAT1 (b) and INTS11 (c) immunoprecipitates from wild-type (U2OS wt) and BRAT1^-/- (clone #8) cells, measured by Western blotting. The black asterisk marks a nonspecific band. d Levels of BRAT1, INTS11 and INTS9 in BRAT1 immunoprecipitates from BRAT1^−/− (clone #8), U2OS wt, and U2OS cells transiently transfected with siINTS9 or siINTS11 as indicated. The black asterisk marks a nonspecific band. b–d The experiment was performed three times with similar results. e Immunoblot of indicated proteins and quantification in U2OS wt and BRAT1^−/− cells (clones #8, #6, and #16). Represented as the mean ± SD (n = 3 for INTS1, INTS3, INTS4 and INTS9; n = 4 for INTS11). Statistical significance was determined by a one-sided paired Student’s t-test (*p < 0.05, **p < 0.01, ***p < 0.001, ns not significant). Samples derived from the same experiment and blots were processed in parallel. Uncropped and unprocessed scans are provided in the Source Data file.

Fig. 2. Loss of BRAT1 impairs 3′ end processing of capped UsnRNAs and snoRNAs.

a Schematic representation of binding sites for primers used for RT-qPCR (arrows), canonical Integrator cleavage sites (scissors), and a downstream 3´box. b RT-qPCR analysis of unprocessed RNU1-1 transcripts in U2OS and BRAT1^−/− (clone #8) cells transfected with control non-targeting siRNA (siNT) or siINTS11, as indicated. Data are represented as the mean ± SD (n = 3). Statistical significance was determined by a one-sided paired Student’s t-test (*p < 0.05, **p < 0.01). c RT-qPCR analysis of unprocessed RNU2-1 or RNU7-1 transcripts in wild-type (U2OS wt) and BRAT1^−/− (clone #8) cells. Data are represented as the mean ± SD (n = 3). Statistical significance was determined by a one-sided paired Student’s t-test (*p < 0.05, **p < 0.01). d RNA-seq data for genes coding U1 snRNA transcripts in wild-type U2OS (wt #1 and wt #2) and BRAT1^−/− (clone #6 and #8) cells. Read coverage is shown on the left and metaplot analysis of library-normalized averaged samples on the right. TSS transcription start site. See also Supplementary Fig. 3a. e Total RNA was isolated from U2OS and BRAT1^−/− (clone #8) cells transfected with control non-targeting siRNA (siNT) or siINTS11, as indicated, resolved on urea-PAGE, Northern blotted and probed for levels of U1, U2, U7 snRNA, and 5S rRNA. Representative blots and quantification are shown. Data are represented as the mean ± SD (n = 4 for U1 and U2; n = 3 for U7). Statistical significance was determined by a one-sided paired Student’s t-test (*p < 0.05, **p < 0.01, ns not significant). f Total RNA was isolated from wild-type (U2OS wt) and BRAT1^−/− (clone #8) cells, resolved on urea-PAGE, and silver-stained. The experiment was performed three times with similar results. Uncropped and unprocessed scans are provided in the Source Data file. g RNA-seq data for genes coding U3 snoRNA transcripts presented similarly as in d.

Fig. 3. BRAT1 deletion disrupts the structural integrity of Cajal bodies.

a Immunofluorescence staining of subnuclear structures in U2OS and BRAT1^−/− (clone #8) cells transfected with control non-targeting siRNA (siNT) or siINTS11, as indicated. Cajal bodies were visualized by immunostaining of coilin (green), nucleoli by B23 (red), and nuclei by DAPI (blue). b Quantification of normalized coilin fluorescence inside the nucleolus using ScanR software; representative pictures are shown in a. Data are represented as the mean ± SD (n = 3). Statistical significance was determined by a one-sided paired Student’s t-test (*p < 0.05, ***p < 0.001). See also Supplementary Fig. 3b, c.

Fig. 4. BRAT1 deletion alters biogenesis of replication-dependent histone mRNAs and transcriptional regulation of protein-coding genes.

a RNA-seq data for genes coding histones H2A, H2B, H3, or H4 in wild-type U2OS (wt #1 and wt #2) and BRAT1^−/− (clone #6 and #8) cells, as indicated. Read coverage and metaplot analysis of averaged samples and histone coding region, 5’UTR, 3’UTR with a cleavage site (scissors) are shown. TSS transcription start site. See also Supplementary Fig. 4. b Schematic representation indicating pairs of primers (arrows) and RT-qPCR analysis of c-FOS mRNA and total RNA (pre-mRNA and mRNA) in wild-type (U2OS wt) and BRAT1^−/− (clone #8) cells. The efficiency of splicing was determined by the ratio between the expression of pre-mRNA and mRNA. Data are represented as the mean ± SD (n = 6 for mRNA and total transcription; n = 3 for splicing). Statistical significance was determined by a one-sided paired Student’s t-test (**p < 0.01, ***p < 0.001, ns not significant). c A volcano plot of BRAT1^−/− (clone #8 and #6) vs. wild-type U2OS (wt #1 and wt #2) transcription profiles showing differentially expressed genes (DEGs) and differentially expressed lncRNAs. Log₂ ratio for upregulated genes ≥1.0 (FDR < 0.05) and for downregulated genes ≥ −1.0 (FDR < 0.05), n = 27,006. d Venn diagrams demonstrating the overlap in identified differentially expressed genes and lncRNAs among cells depleted of BRAT1 (BRAT1^−/−), INTS11, INTS9, INTS4, or INTS8 (shRNA transfected cells). Statistical significance of the overlap was determined by a hypergeometric test (p-values are indicated).

Fig. 5. The V62E mutation in BRAT1-mutated patient cells impairs BRAT1 interaction with INTS11 and INTS9.

a Western blot of BRAT1, INTS11 and INTS9 levels in BRAT1 immunoprecipitates from control, parent, BRAT1 patient-derived fibroblasts and lymphoblastoid cell lines (LCLs), as indicated. Note that 2-fold more lysate from patient cells was employed for loading (input) and immunoprecipitation to ensure equivalent levels of INTS11. The black asterisk marks a nonspecific band. b BRAT1^−/− (clone #8) cells transiently transfected with FLAG-tagged BRAT1^WT and BRAT^V62E constructs were fixed 24 h post-transfection and stained with antibody against FLAG-tag (green) and DAPI (blue). c Levels of BRAT1, INTS11, INTS9 and FLAG-tagged BRAT1 in BRAT1 (left) and FLAG (right) immunoprecipitates from BRAT1^−/− (clone #8) cells transiently transfected with FLAG-BRAT1^WT or FLAG-BRAT^V62E, measured by Western blotting. a–c The experiment was performed three times with similar results. Uncropped and unprocessed scans are provided in the Source Data file.

Fig. 6. Altered Integrator function in BRAT1-mutated patient cells.

a RNA-seq coverage plots for genes coding U1, U4, U5, or U6 snRNA in unaffected parents (Mother and Father) and BRAT1 patient-derived (Patient 1 and Patient 2) lymphoblastoid cell lines (LCLs) are shown, as indicated. b Metaplot analysis of mean reads coverage of individual U1, U4, U5, or U6 snRNA transcripts from unaffected parents (Mother and Father) and BRAT1 patient-derived (Patient 1 and Patient 2) LCLs. Data are presented as the smoothed mean coverage of the selected genes (n = 4) averaged across samples (n = 2) in the respective group with 95% confidence interval of the mean shown (ribbon). TSS transcription start site. c RT-qPCR analysis of total and unprocessed RNU1-1 transcripts in control, parent and BRAT1 patient-derived fibroblasts (top) and LCLs (bottom). Data are represented as the mean ± SD (n = 3). Statistical significance was determined by a one-sided paired Student’s t-test (*p < 0.05, ***p < 0.001).