Methods to get 5' UTR regions (5′-Anchored Reads)

Transcription Start Site Evolution in Drosophila

Bradley J. Main*,1, Andrew D. Smith1, Hyosik Jang1 and Sergey V. Nuzhdin1
+ Author Affiliations

1Section of Molecular and Computational Biology, Department of Biological Sciences, University of Southern California
↵*Corresponding author: E-mail: bmain@usc.edu.

New Approaches

Several molecular techniques can be used to locate TSS including cap analysis for gene expression (CAGE) (Shiraki et al. 2003) and several updated versions (Ni et al. 2010; Plessy et al. 2010; Kanamori-Katayama et al. 2011), 5′-rapid amplification of cDNA ends (RACE) (Harvey and Darlison 1991), robust analysis of 5′-transcript ends (Gowda et al. 2007), and FLcDNA assays (Suzuki et al. 1997). The original CAGE protocol involves the concatenation of short 5′-sequence tags (14–20 bp), followed by traditional Sanger sequencing (Shiraki et al. 2003). More recently, CAGE and similar 5′-targeting methods have been adapted to high-throughput sequencing (Sandelin et al. 2007; Ni et al. 2010; Plessy et al. 2010; Kanamori-Katayama et al. 2011). One major difference between the available methods is the approach used to target 5′-ends of full length transcripts. For example, some methods rely on the removal of the 5′-cap structure with tobacco acid pyrophosphatase (TAP), others use the 5′-cap structure to perform template switching, and 5′-caps can also be biotinyled and isolated with streptavidin beads. We wanted a simple and straightforward approach without specific limitations, such as short reads (tags) (Harbers and Carninci 2005; Kodzius et al. 2006; Ni et al. 2010) and single-end reads (Kodzius et al. 2006), which hinder mapping, and added sampling bias from a required semisuppressive polymerase chain reaction (PCR) step (Plessy et al. 2010; Salimullah et al. 2011). Thus, we generated 5′-anchored reads using a concise TAP-based protocol that employs standard Illumina adapters and barcode indexes and is free of the aforementioned drawbacks. We extracted total RNA from whole body, adult female flies from each Drosophila species. We purified mRNA using oligo-dT Dynabeads (Invitrogen) and ligated an RNA adapter oligo to the 5′-end of each mRNA molecule. We chemically fragmented the ligated mRNA using RNA fragmentation reagent (Ambion) and generated single-stranded cDNA with reverse transcriptase and random hexamers, followed by RNAse H treatment. We added a primer complementary to the 5′-ligated adapter sequence and performed one primer extension step at 72 °C with Taq polymerase to yield double-stranded fragments of all 5′-ends (fig. 2). This primer has a 5′-amine group to prevent concatenation and subsequent ligation. Taq adds an A-overhang in a template-independent fashion (Clark 1988), thus we can bypass the typical blunt-end repair and cleanup step and immediately ligate standard Illumina adapters in a strand-specific orientation. Standard Illumina indexing barcodes were then added during PCR enrichment of each sample. We sequenced the 5′-enriched fragments on an Illumina Genome Analyzer II (see supplementary file, Supplementary Material online, for a detailed protocol).