RNA sequencing by direct tagmentation of RNA/DNA hybrids

Abstract

Transcriptome profiling by RNA sequencing (RNA-seq) has been widely used to characterize cellular status, but it relies on second-strand complementary DNA (cDNA) synthesis to generate initial material for library preparation. Here we use bacterial transposase Tn5, which has been increasingly used in various high-throughput DNA analyses, to construct RNA-seq libraries without second-strand synthesis. We show that Tn5 transposome can randomly bind RNA/DNA heteroduplexes and add sequencing adapters onto RNA directly after reverse transcription. This method, Sequencing HEteRo RNA-DNA-hYbrid (SHERRY), is versatile and scalable. SHERRY accepts a wide range of starting materials, from bulk RNA to single cells. SHERRY offers a greatly simplified protocol and produces results with higher reproducibility and GC uniformity compared with prevailing RNA-seq methods.

Keywords: RNA-seq; Tn5 transposase; single cell.

Fig. 1.

Tn5 tagmentation activity on double-stranded hybrids and the experimental process of SHERRY. (A) RNase H-like (RNHL) domain alignment of Tn5 (TN5P_ECOLX), RNase H (RNH_ECOLI), and MMLV reverse transcriptase (POL_MLVMS). Active residues in the RNHL domains are labeled in bright yellow. Orange boxes represent other domains. (B) Superposition of the RNHL active sites in these three enzymes. Protein Data Bank IDs are 1G15 (RNase H), 2HB5 (MMLV), and 1MUS (Tn5). (C) Putative mechanism of Tn5 tagmentation of a RNA/DNA hybrid. Crooked arrows represent nucleophilic attacks. (D) Size distribution of mRNA/DNA hybrids with and without Tn5 tagmentation and after amplification with index primers. (E) Workflow of sequencing library preparation by SHERRY. The input can be a lysed single cell or extracted bulk RNA. After reverse transcription with oligo-dT primer, the hybrid is directly tagmented by Tn5, followed by gap-repair and enrichment PCR. Wavy and straight gray lines represent RNA and DNA, respectively. Dotted lines represent the track of extension step.

Fig. 2.

Verification of Tn5 tagmentation of RNA/DNA heteroduplexes. (A) Procedures of two ligation tests. Gray dotted box indicates negative results. The table below lists key experimental parameters that are different from standard SHERRY. (B) Comparison of two ligation tests and standard SHERRY with respect to mapping rate, exonic rate, and number of genes detected. Each test consisted of two replicates of 200 ng HEK293T total RNA. (C) Strand test procedures. (D) Comparison among three strand tests and dU-SHERRY with respect to mapping rate, exonic rate, and number of genes detected. Each test consisted of two replicates of 200 ng HEK293T total RNA.

Fig. 3.

Performance of SHERRY with large RNA input. (A) Coefficient of variation (CV) across three replicates was plotted against the mean value of each gene’s FPKM (Fragments Per Kilobase of transcript per Million mapped reads). All experiments used HEK293T total RNA as input. (B) Genes detected by SHERRY in three replicates of 200 ng HEK293T or HeLa total RNA are plotted in Venn Diagrams. Numbers of common genes are indicated. (C) Common genes detected by SHERRY and NEBNext in the three replicates of 200 ng HEK293T or HeLa total RNA. (D) Distance heatmap of samples prepared by SHERRY or NEBNext for three replicates using 200 ng HEK293T or HeLa total RNA. The color bar indicates the Euclidian distance. (E) Correlation of gene expression fold-change identified by SHERRY and NEBNext. Involved genes are differentially expressed genes between HEK293T and HeLa detected by both methods.

Fig. 4.

Performance of SHERRY with microinput samples. (A) Genes detected by SHERRY in three replicates with 100 pg HEK293T total RNA. (B) Correlations of normalized gene read counts between replicates with 100 pg HEK293T total RNA. (C) Gene number detected by scSHERRY under various experimental conditions in single HEK293T cells. Each condition involved three to four replicates. (D) The heatmap of R2 calculated from scSHERRY and Smart-seq2 replicates, and slope deviation in a linear fitting equation for the two methods. (E) Normalized gene numbers with different GC content.