Prediction and Characterization of miRNA/Target Pairs in Non-Model Plants Using RNA-seq

Abstract

Plant microRNAs (miRNAs) are ∼20- to 24-nucleotide small RNAs that post-transcriptionally regulate gene expression of mRNA targets. Here, we present a workflow to characterize the miRNA transcriptome of a non-model plant, focusing on miRNAs and targets that are differentially expressed under one experimental treatment. We cover RNA-seq experimental design to create paired small RNA and mRNA libraries and perform quality control of raw data, de novo mRNA transcriptome assembly and annotation, miRNA prediction, differential expression, target identification, and functional enrichment analysis. Additionally, we include validation of differential expression and miRNA-induced target cleavage using qRT-PCR and modified RNA ligase-mediated 5' rapid amplification of cDNA ends, respectively. Our procedure relies on freely available software and web resources. It is intended for users that lack programming skills but can navigate a command-line interface. To enable an understanding of formatting requirements and anticipated results, we provide sample RNA-seq data and key input/output files for each stage. © 2019 The Authors. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.

Keywords: RNA-seq; de novo transcriptome assembly; differential expression; miRNA prediction; miRNA target; non-model plant; small RNA.

Figure 1

Overview of the article. Colored arrows indicate points of integration between miRNA and target analysis. Differential expression analysis is applied as a filtration step to identify the most biologically relevant miRNA/target pairs.

Figure 2

Generation of paired mRNA and small RNA (sRNA) samples for RNA‐seq. Sample data provided in this article are from Neller et al. (2018). Our study included two conditions: an ethanol control treatment and jasmonic acid (JA) experimental treatment. For each condition, three pooled biological replicates were prepared, with each replicate consisting of equal amounts of RNA from three independent plants. Both the sRNA and mRNA fractions were extracted for each plant to produce a paired sRNA:mRNA design. In total, 18 plants were used (3 plants per pooled replicate × 3 pooled replicates × 2 conditions).

Figure 3

Layout of the Galaxy workspace. The tool panel is shown, with arrows indicating the location of tools used in this article. The history panel is also shown, containing user‐uploaded files as well as output files.

Figure 4

Example of psRNAtarget results. ‘miRNA Acc.’ and ‘Target Acc.’ indicate IDs of miRNA and target, respectively; ‘Expect’ is the score for the miRNA/target pair, with zero indicating perfect sequence complementarity; ‘UPE’ is the energy required for unpairing secondary structure around the miRNA binding site (not calculated under default parameters); ‘Alignment’ depicts binding of the miRNA/target pair, with numbers indicating sequence position; ‘Target Description’ provides additional information in the target ID; ‘Inhibition’ is a prediction of whether the miRNA/target interaction results in transcript cleavage or translational inhibition (use with caution); ‘multiplicity’ is the number of miRNA binding sites in the target.

Figure 5

Overview of miRNA cleavage detection by modified RLM‐RACE. miRNAs that induce cleavage leave exposed 5′ phosphates on their target mRNAs. This allows ligation of an RNA adapter, followed by nonspecific PCR amplification with 5′ and 3′ adapter primers (5′AP, 3′AP) and gene‐specific amplification with 5′ nested adapter primers (5′ NAP) and gene‐specific primers (GSP1, GSP2). PCR products of the expected size are cloned, sequenced, and aligned to the target mRNA sequence. Sequences that align between the 10th and 11th nucleotide of the miRNA binding site indicate successful validation.

Figure 6

Thermal cycling profiles for RLM‐RACE. (A) Touchdown PCR is performed for nonspecific amplification of the cDNA pool. (B) For all other PCR amplifications, a standard profile is used, with or without a gradient annealing step. The number of cycles is indicated, along with respective temperature (°C) and duration (seconds).

Figure 7

Validation of SPL mRNA cleavage by miR156. The RLM‐RACE protocol was performed for the positive control miRNA/target pair, miR156/SPL. The miR156 sequence was reverse‐complemented by the MAAFT aligner for easy visualization of miRNA and target alignment. Sequences of five individual clones of cleaved SPL mRNA are shown, with 4/5 indicating the expected cleavage site.

Figure 8

Primer design for qRT‐PCR validation of mRNA and miRNA expression. (A) The ideal product size for a qRT‐PCR reaction is between 150 and 200 bp. Wherever possible, forward primers (FPs) and reverse primers (RPs) should be designed to span an intron, which will reduce the level of genomic DNA amplification from contamination and allow its detection on an agarose gel for troubleshooting. (B) miRNAs require reverse transcription with a reverse primer bearing a stem‐loop, which serves to increase the length and melting point of the PCR product, to be compatible with standard PCR cycling.

Figure 9

miRNA biogenesis and function. A MIR gene is transcribed by Pol II to yield a non‐coding, single‐stranded transcript that folds back on itself, forming a bulged hairpin flanked by unstructured arms. This primary miRNA (pri‐miRNA) is processed mainly by DCL1 to a miRNA precursor (pre‐miRNA) and miRNA/miRNA duplex (∼21 nt) in sequential steps. The duplex is 3′‐end methylated by HEN1 to protect from degradation, then exported to the cytosol. The miRNA guide strand is selected and incorporated into the RNA‐induced silencing complex (RISC), which contains an AGO protein, usually AGO1. The RISC binds to a target mRNA on the basis of sequence complementarity and either slices it or inhibits its translation.