RNA preparation and labeling
Total RNAs were prepared from HeLa cells using Trizol (Invitrogen, Grand Island, NY). RNA standards (U11 and U12) were transcribed using T7 MegaScript Kit (Ambion, Grand Island, NY) and gel purified. A modified protocol of the ULYSIS nucleic acid labeling kits (Invitrogen) was used to label RNAs with Alexa546 or Alexa647. Total RNA (1 μg) was combined with 3 pmol each of standard E. coli tRNALys, Escherichia coli tRNAVal, and Saccharomyces cerevisiae tRNAPhe and resuspended in ULYSIS labeling buffer (5 mM Tris, 1 mM EDTA, pH 8.0). The RNAs were denatured at 95°C for 5 min, added to either Alexa546 or Alexa647 ULYSIS dye and incubated at 90°C for 10 min. G25 spin columns (GE Healthcare, Pittsburgh, PA) were used to remove excess dye. The labeled RNAs were then precipitated with ethanol for use in microarray hybridization.
Microarray design, hybridization, and analysis
Two to four 60-mer DNA probes complementary to different regions of the ncRNAs were designed with a GC content between 35% and 55% and limited secondary structure. The probes were spotted on glass slides in quadruplicate at 200 μM on Matrix II slides (Full Moon Biosystems, Sunnyvale, CA) and cross-linked by UV irradiation. Before use slides were pre-hybridized in 2X SSC, 0.2% SDS, 0.1% BSA for 20–30 min, rinsed in water twice and then dried.
Labeled total RNA samples were resuspended in Oligo Hyb Buffer (Full Moon Biosystems, Sunnyvale, CA) containing 10 μg polyA RNA and 20 μg salmon sperm DNA. Samples were applied to the hybridization chamber of the HS4800 Pro hybridzation station (Tecan, Morrisville, NC). The following program for hybridization was used: Wash 1 (54°C, wash 30 s, soak 30 s, 3 runs), two-step Sample Injecion (54°C), Hybridization (54°C, 16 hr), Wash 1 (54°C, wash 30 s, soak 30 s), Wash 1, 2 and 3 (23°C, wash 30 s, soak 30 s), Slide drying (30°C, 5 min). Wash buffers are Wash 1 (2X SSC, 0.2% SDS), Wash 2 (2X SSC), and Wash 3 (0.2X SSC).
Microarray slides were imaged using GenePix 4000b scanner (Axon Instruments, Molecular Devices, Sunnyvale, CA). The fluorescence intensities were quantified and background was subtracted using the GenePix Pro 6.0 software. Values for the 4 replicate spots on each slide were averaged.
Several hybridization temperatures were tested using total RNA and in vitro transcribed snRNAs showing reliable detection of 5 ng of spiked-in U11 and U12 snRNAs without significant cross-hybridization at temperatures ranging from 48°C to 54°C (). Fluorescence from non-specific probes (red dots) was <10% of the signal of the specific probes, demonstrating the specificity of this array. Similar results were obtained for the measurements of other in vitro transcribed ncRNAs, including U1, U2, U4, U5, U6, U7, U85, U90, U93 and hTR (human telomerase RNA).
The reproducibility between biological replicates was determined on total RNA isolated from three separate HeLa cell cultures, which were labeled with both Alexa546 and Alexa647 and hybridized on the array. For those probes with detectable signals at least twofold above background, the intensity for each probe normalized to the total signal on the chip did not show significant variation among the three biological replicates (, error bars), indicating that the array can reliably discern ncRNA abundance between replicate samples. The dynamic range of the array was determined by testing the linearity of the signals from the different probes with varying amounts of RNA. shows a range of 50–2500 ng of total RNA labeled with Alexa647 that was hybridized in the presence of 200 ng of Alexa546-labeled total RNA. Most probes exhibited linearity over the range in which experiments are typically performed (100–500 ng total RNA).
Quantitative real-time PCR
snRNA-specific primers and 5S rRNA-specific primers (endogenous control) were used to generate cDNAs using Transcriptor First-strand cDNA synthesis kit (Roche Applied Sciences, Indianapolis, IN) from an input of 100 ng of total RNA according to manufacturer’s instructions. Two and a half percent of the cDNA generated was used for each qPCR reaction (Applied Biosystems 7500 Fast Real-time PCR system) using SYBR Green dye chemistry. The same reverse primers were used for both RT and qPCR. Each reaction was performed in triplicate. For real-time qPCRs and RT-PCRs used to measure splicing of minor and major introns (), 1 μg total RNA was converted to cDNA using the VILO kit (Invitrogen) according to manufacturers recommendations. cDNA was then diluted to 10 ng/μl and 20–50 ng cDNA was then used as input for qPCR (with SYBR Green as described above) or regular PCR using platinum Taq (Invitrogen). For qPCR, all the primer sets for various genes span the splice junction and thus measure only spliced mRNAs. For RT-PCR, the primers are located in the surrounding exons and thus amplify both spliced and unspliced mRNAs.
Cell treatments, library preparations and RNA-seq
HeLa cells were grown in DMEM supplemented with 10% fetal bovine serum (FBS), 1% antibiotics and maintained at 37°C in a 5% CO2 humidified atmosphere. Actinomycin D (5 μg/ml), cycloheximide (10 μg/ml), DRB (100 μM), nocodazole (200 nM), or anisomycin (1 μg/ml) were added to the media for the indicated time, followed by RNA extraction as indicated above.
Luciferase splicing reporters were generated as previously described (Younis et al., 2010
). Briefly, the minor intron of the CHD4 gene was inserted at nucleotide position 571 of firefly luciferase gene. The luciferase protein is destabilized using both a PEST protein degradation sequence as well as a CL1 sequence, and mRNA is destabilized by adding five tandem AUUUA repeats into the 3′UTR. Both intron-containing and intronless luciferase were transcribed from a CMV promoter.
Control (5′-CCT CTT ACC TCA GTT ACA ATT TAT A-3′) and U6atac (5′-AAC CTT CTC TCC TTT CAT ACA ACA C-3′) antisense morpholinos (AMOs) were transfected into cells using the Neon system (Invitrogen). 6 hr post transfection, cells were labeled with 200 μM 4-thiouridine (4-SU) for an additional 2 hr. Total RNA was extracted using Trizol (Invitrogen) followed by extraction of polyA-containing RNA on oligo-dT columns using Oligotex kit (Qiagen, Germantown, MD). For isolation of 4-SU-labeled nascent transcripts, EZ-Link biotin-HPDP (Thermo Scientific, Waltham, MA) was reacted with the 4-SU and purified on streptavidin Dynabeads (Invitrogen) as previously described (Dolken et al., 2008
Nascent transcripts prepared as described above were used to prepare cDNA libraries using Encore NGS Library System I (Nugen, San Carlos, CA) according to the manufacturer’s recommendations. Briefly, 100 ng RNA was converted to cDNA and amplified using the Ribo-SPIA technology. The cDNA was then fragmented, end repaired, ligated to Illumina adaptors and bead purified according to instructions in the Ovation RNA-seq System (Nugen). Sequencing was performed at the University of Pennsylvania Core Facility on Illumina Hi-Seq2000 platform to generate single end 100 base reads. Raw and processed data is available on GEO under the accession number GSE48263. This link can be used to retrieve the data: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE48263
Data analysis and statistics
RNA-seq reads were aligned to the reference genome (UCSC, hg19) using Tophat by default parameters (Trapnell et al., 2009
). Uniquely mapped reads and the best of multiple hits were kept for downstream analysis. BEDTools software (Quinlan and Hall, 2010
) was used to generate reads per exon, from which FPKM (read fragments per kilobase of exon model per million mapped reads) values were calculated (Mortazavi et al., 2008
). The analysis of differential gene expression was performed using Cufflinks (Trapnell et al., 2010
). Significant genes were detected based on FDR level at 0.05 using Benjamini-Hochberg correction for multiple-testing (Trapnell et al., 2012
To identify differentially expressed splicing isoforms across samples and quantify the expression level of those alternative spliced genes, MISO (version 0.4.1 with default parameters) was applied to the aligned RNA-seq data (Katz et al., 2010
). Briefly, MISO estimates expression at alternative splicing event level by computing PSI (Percent Spliced Isoform) and measures the differential expression by Bayes factors. The MISO results were filtered for alternative splicing events using the following criteria: (a) at least 15 inclusion read, (b) 15 exclusion read, such that (c) the sum of inclusion and exclusion reads is at least 30, and (d) the ΔΨ is at least 0.30, and (e) the Bayes factor is at least 10, and that (a)–(e) are true in one of the samples.
RNase H protection assay
Total cell extract was prepared from AMO-transfected cells using 10 mM Tris-HCl pH 7.5, 2.5 mM MgCl2, 100 mM NaCl and 0.1% NP-40. RNase H along with 5 μM antisense DNA oligonucleotide for U6atac (5′-TCA TAC AAC AC-3′) was added for 25 min at 30°C, and RNA was purified and analyzed by northern blotting with a U6atac snRNA probe (5′-CCG TAT GCG TGT TGT CAG GCC CGA GGG CCT-3′).