cDNA Amplification, Library Preparation and Sequencing

To generate cDNA, we used 65 ng or 50 ng of total RNA from each sample pool. The ERCC RNA spike-in mix (Invitrogen, Carlsbad, California) was added in quantities recommended by the manufacturer. RNA was added to lysis mixture (0.9X PCR buffer, 3 mM MgCl2, 0.45% NP40, 4.5 mM DTT, 0.18U/ul SUPERase-In, 0.36U/ul RNase Inhibitor, 0.125uM dNTP) and heated to 70 °C for 90 seconds. First strand cDNA synthesis was performed using SuperScript III reverse transcriptase (Invitrogen) along with a modified poly-dT primer in reactions containing 13.2U/ul SuperScript III, 0.5uM poly-dT hairpin primer, 0.4U/ul RNase Inhibitor, and 0.07ug/ul T4 gene 32 protein (Invitrogen). The primer (5′-T₁₈TGGAATTCTCGGGTGCCAACCCTTGGCACCCGAGAATTCCAT₆V-3′) contains a single non-T nucleotide at the 3′ end, which serves to anchor it to the transcript immediately upstream of the polyA tail, and an internal sequence that forms a hairpin loop. Prior to the RT reaction, the primer was heated to 98 °C for 10 minutes and allowed to slowly cool to room temperature in order to ensure proper formation of the hairpin. Following first strand cDNA synthesis, ExoSAP (Affymetrix, Santa Clara, California) was added to remove all remaining free primer. The cDNA was polyadenylated using a terminal transferase and treated with RNaseH to remove the RNA strand (New England Biolabs, Ipswitch, Massachusetts). Purification of the single-stranded cDNA was done using Ampure beads (Agencourt, Lynn, Massachusetts). The second strand of cDNA was synthesized using a UP2-polydT primer (5′-ATATCTCGAGGGCGCGCCGGATCCT₂₂-3′) and Takara Taq polymerase (Nojihigashi, Japan). Eighteen rounds of cDNA amplification were performed using a UP2 primer (5′-ATATCTCGAGGGCGCGCCGGATCC-3′) and a biotinylated primer containing only a part of the hairpin sequence (5′-Bio-CAACCCTTGGCACCCGAGAATTCCAT₆-3′) in order to limit residual polyA tail sequences to 6 nt in the final product. Half of the cDNA was re-amplified by another six cycles of PCR to generate sufficient cDNA for library preparation, repurified with Ampure beads, and examined on the Bioanalyzer to assess quality and quantity of the PCR products.

For library preparation, approximately 1ug of amplified cDNA was fragmented in a S220 Focused Ultrasonicator (Covaris, Woburn, Massachusetts). Fragments between 300–400 bp were selected with a gel-free size-selection protocol using Ampure beads. Streptavidin beads (Invitrogen) were added to capture biotinylated cDNA ends corresponding to the original 3′ ends of the transcript. 10ul of streptavidin beads were washed and resuspended in 20ul of 2x binding buffer (10 mM Tris pH7.5, 1 mM EDTA, 2 M NaCl, 0.02% Tween20). While still attached to beads, the captured cDNA was end-repaired and an adenosine was added to the 3′ end (NEBNext Ultra End Repair/dA-Tailing module). Illumina P5 adapters carrying a six base barcode were T/A-ligated to the free end of the DNA fragments (5′-AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT-XXXXXX-T-3′). The library was amplified from the beads for twelve cycles with KAPA HiFi HotStart ReadyMix (Kapa Wilmington, Massachusetts) using a P5 primer (5′-AATGATACGGCGACCACCGAGATCT-3′) and a P7-half-hairpin primer (5′- CAAGCAGAAGACGGCATACGAGATCCAACCCTTGGCACCCGAGAATTCCA -3′). During PCR amplification, the beads easily settle to the bottom. To ensure proper amplification of the library, the tubes were quickly vortexed after each denaturing step. Libraries were cleaned with Ampure beads and checked on a Bioanalyzer for quality and quantity before storage in −20 °C. Prior to sequencing libraries were quantified using the KAPA Library Quantification Kit, diluted to a concentration of 2nM, and pooled. A PhiX library was added at approximately 20% to increase the base diversity. A paired-end 100 bp rapid run was performed on the Illumina HiSeq 2500 platform (San Diego, California) using the standard Read 1 primer and a custom sequencing primer for Read 2 (5′-AGCAGAAGACGGCATACGAGATCCAACCCTTGGCACCCGAGAATTCCAT₆ -3′), which is designed to provide sequences starting from the last base prior to the poly(A) tail.