A quantitative high-throughput in vitro splicing assay identifies inhibitors of spliceosome catalysis

Abstract

Despite intensive research, there are very few reagents with which to modulate and dissect the mRNA splicing pathway. Here, we describe a novel approach to identify such tools, based on detection of the exon junction complex (EJC), a unique molecular signature that splicing leaves on mRNAs. We developed a high-throughput, splicing-dependent EJC immunoprecipitation (EJIPT) assay to quantitate mRNAs spliced from biotin-tagged pre-mRNAs in cell extracts, using antibodies to EJC components Y14 and eukaryotic translation initiation factor 4aIII (eIF4AIII). Deploying EJIPT we performed high-throughput screening (HTS) in conjunction with secondary assays to identify splicing inhibitors. We describe the identification of 1,4-naphthoquinones and 1,4-heterocyclic quinones with known anticancer activity as potent and selective splicing inhibitors. Interestingly, and unlike previously described small molecules, most of which target early steps, our inhibitors represented by the benzothiazole-4,7-dione, BN82685, block the second of two trans-esterification reactions in splicing, preventing the release of intron lariat and ligation of exons. We show that BN82685 inhibits activated spliceosomes' elaborate structural rearrangements that are required for second-step catalysis, allowing definition of spliceosomes stalled in midcatalysis. EJIPT provides a platform for characterization and discovery of splicing and EJC modulators.

Fig 1

A high-throughput assay (EJIPT) utilizing EJC binding to spliced mRNAs to measure splicing in vitro. Splicing reaction mixtures containing biotin-labeled pre-mRNA, 20 μM each compound, splicing extract, and 0.5 mM ATP are assembled by liquid handler in 384-well plates and incubated for 1.5 h at 30°C. Reaction mixtures are then transferred to a NeutrAvidin-coated plate, which captures biotin-labeled RNAs/RNPs postsplicing. Anti-eIF4AIII (3F1) antibody is added to detect splicing-dependent EJC formation on spliced mRNAs. Following a washing step, incubation with an HRP-conjugated anti-mouse IgG secondary antibody, and then addition of chemiluminescent substrate, signals representing eIF4AIII/mRNA complexes are obtained by an automatic plate reader.

Fig 2

The analysis of in vitro splicing by the splicing gel method and by the EJIPT assay. (A) ³²P-labeled Ad2 pre-mRNA was spliced in the presence of no drug (Control), no extract, 10 μM control AMO, 10 μM U2 snRNA AMO, or 2.5 μM okadaic acid. RNAs after splicing were isolated and resolved on denaturing PAGE gels. Identities of splicing intermediates are depicted to the right of the gels, and quantifications of the splicing efficiency (mRNA/pre-mRNA ratio) by phosphorimaging from two independent experiments are shown below each lane. (B) The EJIPT assay was performed in 384-well microplates using biotin-labeled Ad2 pre-mRNA spliced under the same conditions as described for panel A.

Fig 3

Identification, validation, and selectivity of the hits from the EJIPT HTS. (A) RNA from splicing reactions was probed for pre-mRNA and mRNA by Northern dot blot analysis. Column A samples (black box; rows 1 to 4) consist of DMSO controls. Samples in row 5 (red box, columns A to D) represent splicing-inhibitory conditions, as follows: A, 5 mM AMP-PNP; B, no extract; C, 2.5 μM okadaic acid; D, 2.5 μM calyculin A. Row 6 samples represent a standard curve of mRNA from 12.5 to 0.2 ng in 1:4 serial dilutions. Row 7 samples represent a standard curve of pre-mRNA from 29 to 0.4 ng in 1:4 serial dilutions. Samples in columns B to D (rows 1 to 4) represent splicing reactions with candidate compounds at a 40 μM final concentration. (B) Ratios of mRNA to pre-mRNA from the Northern dot blot assay in panel A were plotted for DMSO samples (gray), compounds (blue), and inhibitory controls (red). Compounds outside the range of 3SD (red and green dotted lines) of the DMSO samples were tested on splicing gels, and percent splicing efficiencies are reported. (C) Dose-response curve of NSC95397 (compound B2 in panels A and B) determined by the EJIPT assay using anti-Y14 (4C4) antibody. The IC₅₀ is calculated from fitting the dose-response data using Sigma Plot software. (D) ³²P-labeled CDC14-15 pre-mRNA was spliced in the presence of DMSO or 20 or 80 μM NSC95397 and resolved on denaturing PAGE gels. Identities of splicing intermediates are depicted to the right of gels. (E) Biotin-labeled Ad2 pre-mRNA was spliced in the presence of DMSO or 20 μM NSC95397 and measured by the EJIPT assay with eIF4AIII (3F1) or Y14 (4C4) antibodies and the RNA-binding assays with hnRNP C1/C2 (4F4) or hnRNP A1 (4B10) antibodies. The effect of the compound in the TNT assay is also reported. Results are expressed as percentages relative to DMSO (100%) and are averages of three or more independent experiments. The error bars represent standard deviations of the repeats.

Fig 4

NSC95397 and related compounds contain 1,4-naphthoquinone and 1,4-heterocyclic quinone scaffolds that confer splicing inhibition in vitro. Compounds were added to splicing reaction mixtures with CDC14-15 pre-mRNA at a 20 or 80 μM final concentration. Radiolabeled RNA was purified and resolved by denaturing PAGE, and gels were autoradiographed. Structural similarity to NSC95397 and reported inhibition of CDC25 are listed above. Phosphorimaging quantification of mRNA is expressed as a percentage compared to DMSO after values were normalized to total counts in the lane (below).

Fig 5

BN82685 is a general splicing inhibitor that acts after the first catalytic step. (A) In vitro splicing reaction mixtures containing biotin-labeled Ad2, CDC14-15, or IgM C3-C4 pre-mRNA substrates in the presence of DMSO or 10 μM BN82685 were measured by the EJIPT assay using anti-Y14 antibody. (B) Mixtures described in panel A with ³²P-labeled pre-mRNAs spliced in vitro were analyzed on denaturing PAGE gels. The identities of splicing intermediates for Ad2 and CDC14-15 are shown on the left, and intermediates for IgM C3-C4 are shown on the right. (C) Compounds were added to splicing reaction mixtures with CDC14-15 pre-mRNA at the final indicated concentrations. Phosphorimaging quantification of each intermediate on the splicing gel is expressed as a percentage compared to DMSO after normalizing to total counts in the lane (below).

Fig 6

BN82685 stalls the progression of complex C. (A) DMSO or 20 μM BN82685 was added to splicing reaction mixtures containing radiolabeled Ad2 pre-mRNA and stopped by the addition of heparin (5 mg/ml) at the indicated time points (0 to 80 min). Positions of splicing complexes resolved on 3.5% native acrylamide gels are listed to the right. The arrow indicates the presence of a stalled complex in BN82685-treated reactions. (B) Size exclusion elution profiles of 20-min (left) and 90-min (right) reactions were generated from radioactivity counts (bottom), and even-numbered fractions were run on denaturing gels (top).

Fig 7

BN82685 freezes spliceosomes in a postactivation state. (A) Ninety-minute in vitro splicing reactions of biotin-labeled Ad2 pre-mRNA containing either DMSO or 20 μM BN82685 were incubated with streptavidin-coated beads. Precipitates were treated with (+) or without (−) RNase A, beads were washed, and RNAs were isolated and probed by Northern blotting. (B) For the reactions in panel A, coprecipitating proteins, as indicated to the right of the gels, were probed by Western blotting. (C) Streptavidin precipitates were treated as described in the Materials and Methods and analyzed by mass spectrometry. The number of peptide spectrums captured for individual proteins in BN82685 samples for three independent experiments are expressed as an average ratio compared to DMSO controls (set as 1). Error bars, standard deviation; *, P < 0.05 by paired-end Student t test.