UNIT 4.12 Construction of Normalized RNA-seq Libraries for Next-Generation Sequencing Using the Crab Duplex-Specific Nuclease Danos C. Christodoulou,1 Joshua M. Gorham,1 Daniel S. Herman,1 and J.G. Seidman1 1Harvard Medical School, Boston, Massachusetts ABSTRACT RNA-seq is a method for studying the transcriptome of cells or tissues by massively parallel sequencing of tens of millions of short DNA fragments. However, the broad dynamic range of gene expression levels, which span more than five orders of magni- tude, necessitates considerable over-sequencing to characterize low-abundance RNAs at sufficient depth. Here, we describe a method that enables efficient sequencing of low- abundance RNAs by normalizing or reducing the range spanned by the most abundant RNA species to the least abundant RNA species. This normalization is achieved using an approach that was developed for generating expressed sequence tag (EST) libraries that uses the crab duplex-specific nuclease and exploits the kinetics of DNA annealing. That is, double-stranded cDNA is denatured, then allowed to partially re-anneal, and the most abundant species, which re-anneal most rapidly, are digested with crab duplex-specific nuclease. This procedure substantially decreases the proportion of sequence reads from highly expressed RNAs, facilitating assessment of the full spectrum of the sequence and structure of transcriptomes. Curr. Protoc. Mol. Biol. 94:4.12.1-4.12.11. C 2011 by John Wiley & Sons, Inc. Keywords: RNA-seq library normalization crab duplex nuclease high-throughput DNA sequencing BASIC PROTOCOL This unit describes the generation of a normalized RNA-seq library for next-generation sequencing by utilizing the preference of the crab duplex nuclease (DSN) for digest- ing double-stranded, rather than single-stranded DNA (UNIT 5.12 also see Zhulidov et al., 2004). In this approach (Basic Protocol), polyadenylated RNA is used to generate a complex RNA-seq library. The library is denatured and incompletely renatured, and then digested with DSN. The kinetics of DNA annealing are such that at any given time abundant DNA molecules are more likely to have re-annealed and become double stranded, while rare molecules are more likely to remain single stranded. Thus, prefer- entially digesting double-stranded DNA with DSN yields a library markedly enriched for more rare DNA species. Massively parallel sequencing of these normalized libraries allows for efficient and comprehensive assessment of the sequence and structure of the polyadenylated transcriptome. The first step in the construction of a normalized RNA-seq library is to make a high- complexity RNA-seq library (i.e., a library that includes sufficient depth of sequence from all positions of all polyadenylated RNAs). This involves starting with ���40 to 400 ng polyadenylated RNA from ���2 to 20 ��g total RNA. This method should also be suitable for non-polyadenylated RNA, in which case the protocol would be started at step 5. The quality of the RNA is crucial and can be determined using a Bioanalyzer (Agilent) (Schroeder et al., 2006) or an agarose gel. All steps in the protocol are performed with an excess amount of reagents to ensure each reaction approaches completion. Current Protocols in Molecular Biology 4.12.1-4.12.11, April 2011 Published online April 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/0471142727.mb0412s94 Copyright C 2011 John Wiley & Sons, Inc. Preparation and Analysis of RNA 4.12.1 Supplement 94

Construction of Normalized RNA-Seq Libraries 4.12.2 Supplement 94 Current Protocols in Molecular Biology fragmentation (optional) cDNA synthesis, double-strand synthesis end repair, A-tailing, adapter ligation amplification, denaturation annealing at 68��C crab duplex-specific nuclease treatment amplification with final primers Figure 4.12.1 Normalized RNA-seq library construction. Illustration of the normalization process starting from a rare (gray) and abundant (black) RNA transcript. The resulting high-complexity RNA-seq library is amplified to generate 500 to 1200 ng DNA using 19- to 20-bp primers corresponding to the inner sequences of the Illumina Paired-End adapters. Normalization of the RNA-seq library is then achieved by denatur- ing and re-annealing at 68���C, followed by treatment with the duplex-specific nuclease (DSN). The annealing temperature of the resulting amplified RNA-seq library���s adapter sequence is ���57���C, which minimizes nonspecific targeting by DSN. A second ampli- fication and normalization is then performed. The normalized library is then amplified with the full-length Illumina paired-end primers. See Figure 4.12.1 for an overview of the library construction. Materials Dynabeads mRNA DIRECT kit (Invitrogen, cat. no. 610-12) containing: Lysis/binding buffer Washing buffer B DynaMag-2 magnet (Invitrogen, cat. no. 123-21D) RNA, purified using TRIzol (Invitrogen, cat. no 15596-018) or RNeasy kit (Qiagen, cat. no. 74104) Quant-iT RNA assay kit (Invitrogen, cat. no. Q32852) 5�� fragmentation buffer (see recipe)