Pro-seq library preparation and sequencing

The PRO-seq protocol was performed as described by Mahat et al. (2016). Briefly, nuclei isolation washed away endogenous nucleotides, halting elongating RNA polymerases bound to chromatin. Precision run-on reactions were performed in the presence of equimolar amounts of unaltered ATP and GTP, as well as biotin-11-CTP and biotin-11-UTP (Perkin-Elmer). Notably, this two-biotin run-on will produce polymerase profiles with slightly reduced (∼ 2–4 bp) resolution compared to a more typical four-biotin run-on, given that RNA polymerases will primarily stall when incorporating the modified CTP and UTP (Kwak et al. 2013). RNA was extracted and base-hydrolyzed with NaOH. Hydrolyzed, biotin-labeled nascent RNAs were passed through a RNase-free P-30 spin column (Bio-Rad) and then enriched using M-280 streptavidin Dynabeads (Invitrogen). T4 RNA ligase 1 (NEB) was used to attach a 3′ RNA adaptor containing a six-nucleotide unique molecular index (UMI) for the removal of duplicate sequences produced by PCR (5′-/5Phos/NNNNNNGAUCGUCGGACUGUAGAACUCUGAAC/Inverted dT/-3′). After a second biotin-enrichment, RNAs were submitted to RNA 5′ Pyrophosphohydrolase (RppH, NEB) treatment for 5′ de-capping, and then 5′ phosphorylation with T4 polynucleotide kinase (T4 PNK, NEB), before ligation of the 5′ RNA adaptor (5′-CCUUGGCACCCGAGAAUUCCA-3′). Reverse transcription was performed with SSIII RT (Invitrogen) after a third biotin-enrichment. The cDNAs produced were PCR amplified for 13 cycles with Phusion polymerase (NEB) and size selected (120–400 bp) before sequencing. This protocol generated strand-specific libraries with every read starting from the 3′ end of the RNA. The RNA adapters used are TruSeq-compatible and libraries were reverse transcribed and amplified using primers from the Illumina TruSeq small RNA sequencing kit. Amplified libraries were assessed for quality on a bioanalyzer prior to sequencing on a HiSeq2500 with 100 bp single reads.