Widespread occurrence of microRNA-mediated target cleavage on membrane-bound polysomes

Abstract

Background: Small RNAs (sRNAs) including microRNAs (miRNAs) and small interfering RNAs (siRNAs) serve as core players in gene silencing at transcriptional and post-transcriptional levels in plants, but their subcellular localization has not yet been well studied, thus limiting our mechanistic understanding of sRNA action.

Results: We investigate the cytoplasmic partitioning of sRNAs and their targets globally in maize (Zea mays, inbred line "B73") and rice (Oryza sativa, cv. "Nipponbare") by high-throughput sequencing of polysome-associated sRNAs and 3' cleavage fragments, and find that both miRNAs and a subset of 21-nucleotide (nt)/22-nt siRNAs are enriched on membrane-bound polysomes (MBPs) relative to total polysomes (TPs) across different tissues. Most of the siRNAs are generated from transposable elements (TEs), and retrotransposons positively contributed to MBP overaccumulation of 22-nt TE-derived siRNAs (TE-siRNAs) as opposed to DNA transposons. Widespread occurrence of miRNA-mediated target cleavage is observed on MBPs, and a large proportion of these cleavage events are MBP-unique. Reproductive 21PHAS (21-nt phasiRNA-generating) and 24PHAS (24-nt phasiRNA-generating) precursors, which were commonly considered as noncoding RNAs, are bound by polysomes, and high-frequency cleavage of 21PHAS precursors by miR2118 and 24PHAS precursors by miR2275 is further detected on MBPs. Reproductive 21-nt phasiRNAs are enriched on MBPs as opposed to TPs, whereas 24-nt phasiRNAs are nearly completely devoid of polysome occupancy.

Conclusions: MBP overaccumulation is a conserved pattern for cytoplasmic partitioning of sRNAs, and endoplasmic reticulum (ER)-bound ribosomes function as an independent regulatory layer for miRNA-induced gene silencing and reproductive phasiRNA biosynthesis in maize and rice.

Keywords: Membrane-bound polysome; Oryza sativa; PARE; Target cleavage; Zea mays; miRNA; phasiRNA; siRNA.

Fig. 1

Distinct subcellular partitioning of 21-nt/22-nt and 24-nt sRNAs in maize and rice. a Schematic illustration of experimental design for this study. b, c Size distribution of sRNAs (b) and relative abundance of 21-nt/22-nt and 24-nt sRNAs (c) in various fractions from maize immature tassels. d, e Size distribution of sRNAs (d) and relative abundance of 21-nt/22-nt and 24-nt sRNAs (e) in various fractions from rice immature panicles. In b and d, sRNA abundance is displayed as mean ± standard deviation (SD) of three biological repeats, and “RPMR” is short for “reads per million rRNA fragments”. In c and e, comparisons between Total, TP, and MBP were performed by two-tailed unpaired t-test, and P values are displayed above the corresponding comparisons

Fig. 2

Retrotransposons and DNA transposons contribute differentially to polysome association of TE-derived siRNAs in maize. a Identification of 22-nt TE-siRNAs that were differentially accumulated between TP and input (Total) (left panel), and between MBP and TP (right panel) in maize immature tassels. b Identification of 24-nt TE-siRNAs that were differentially accumulated between TP and Total (left panel), and between MBP and TP (right panel) in maize immature tassels. TE-siRNA abundance is displayed as the mean of three biological repeats. “RPMR” is short for “reads per million rRNA fragments”. c Overlap of TE loci that generated differentially accumulated siRNAs in the various comparisons in a and b. d Contributions of retrotransposons and DNA transposons to the differentially accumulated 22-nt and 24-nt TE-siRNAs in maize immature tassels. “Genome-wide” denotes the contributions of the transposon types to total TE-siRNAs. LINE: long interspersed nuclear element; LTR: long terminal repeat; SINE: short interspersed nuclear element; TIR: terminal inverted repeat. The cutoff parameters for differentially accumulated TE-siRNAs are fold change ≥ 2 and P value ≤ 0.05. “Group I,” “Group II,” “Group III,” and “Group IV” represent “TE loci with 22-nt or 24-nt siRNAs that were polysome-depleted,” “TE loci producing 22-nt or 24-nt siRNAs that were polysome-associated,” “TE loci with 22-nt or 24-nt siRNAs that were likely on free polysomes (FPs),” and “TE loci with 22-nt or 24-nt siRNAs enriched on MBPs,” respectively

Fig. 3

Enrichment of miRNAs on membrane-bound polysomes in maize and rice. a, c Size and abundance of all miRNAs in various samples from maize immature tassels (a) and rice immature panicles (c). b, d Identification of differentially accumulated miRNAs between TP and input (Total) (left panels), and between MBP and TP (right panels) in maize immature tassels (b) and rice immature panicles (d). The cutoff parameters for differentially accumulated miRNAs are fold change ≥ 2 and P value ≤ 0.05. “Group I,” “Group II,” “Group III,” and “Group IV” represent “miRNAs that were polysome-depleted,” “miRNAs enriched on polysomes,” “miRNAs that were associated with polysomes but MBP-depleted,” and “miRNAs that were MBP-enriched,” respectively. e Abundance of representative miRNAs (ZmmiR390a/b-5p, ZmmiR528a/b-5p, ZmmiR529-5p, ZmmiR529-3p, OsmiR390-5p, and OsmiR528-5p) in Total, TP, and MBP samples from maize immature tassels and rice immature panicles. miRNA abundance is displayed as mean ± standard deviation (SD) (a, c, and e) or the mean of three biological repeats (b and d). “RPMR” is short for “reads per million rRNA fragments.” f Northern blotting verification for miRNAs in (e) with Total, TP, and MBP RNA preparations that were used for sRNA library construction. 5S rRNA was used as an internal control, and U6 was used as a nuclear RNA marker

Fig. 4

Widespread occurrence of miRNA-mediated target cleavage on membrane-bound polysomes in maize and rice. a, b Number (left panels) and overlap of identified miRNA target genes (right panels) in input (Total), TP, and MBP samples from maize immature tassels (a) and rice immature panicles (b). Cleavage sites that are classified as category 0 with P value ≤ 0.05 in at least two biological repeats were filtered as miRNA target sites. Category 0 means that the 3′ cleavage fragments with the maximum read count (> 1) are mapped to only one indicated position on the transcript. In the left panels of (a) and (b), the number of miRNA target genes is displayed as mean ± standard deviation (SD) of three biological repeats, and comparisons between Total, TP, and MBP were performed by two-tailed unpaired t-test and P values are displayed above the corresponding comparisons. c PolyA RNA-seq read (top panel) and PARE 3′ cleavage fragment (bottom panel) coverage for Zm00001d018024, an identified MBP-unique target transcript cleaved by miR529-5p in Total, TP and MBP samples from maize immature tassels. d PolyA RNA-seq read (top panel) and PARE 3′ cleavage fragment (bottom panel) coverage for LOC_Os06g03920, an identified MBP-unique target transcript cleaved by miR396c-5p, in Total, TP, and MBP samples from rice immature panicles. The gene models are shown below the RNA-seq panels, with the thicker rectangles, lines, and thinner rectangles representing exons, introns, and UTR regions, respectively. In the PARE panels, the red dots indicate the cleavage sites on the transcripts targeted by miRNAs. The sequences of the target sites and the miRNAs are shown below the PARE panels. “PARE” and “RPM” are short for “parallel analysis of RNA ends” and “reads per million mapped reads,” respectively

Fig. 5

Reproductive 21PHAS precursors are widely cleaved by miR2118 on membrane-bound polysomes in maize and rice. a, c Abundance of different miR2118 members in input (Total), TP, and MBP samples from maize immature tassels (a) and rice immature panicles (c). b, d Number (left panels) and overlap of 21PHAS precursors cleaved by miR2118 (right panels) in Total, TP, and MBP samples from maize immature tassels (b) and rice immature panicles (d). Cleavage sites with category = 0, which means that the reads with the maximum count (> 1) are mapped to only one indicated position on the transcript, and P value ≤ 0.05 in at least two biological repeats were filtered as miR2118 target sites. e, f Abundance of reproductive 21-nt phasiRNAs in Total, TP, and MBP samples from maize immature tassels (e) and rice immature panicles (f). Abundance of miR2118 and reproductive 21-nt phasiRNAs (a, c, e, and f), and number of target 21PHAS precursors (left panels of b and d) are displayed as mean ± standard deviation (SD) of three biological repeats. “RPMR” is short for “reads per million rRNA fragments” (a, c, e, and f). Comparisons between MBP and TP (a and c), and between Total, TP, and MBP (left panels of b and d, e and f) were performed by two-tailed unpaired t-test. P values are displayed to the right of the corresponding MBP miR2118s (a and c) or above the corresponding comparisons (left panels of b and d, e and f)

Fig. 6

Reproductive 24PHAS precursors are widely cleaved by miR2275 on membrane-bound polysomes in maize and rice. a, c Abundance of different miR2275 members in input (Total), TP, and MBP samples from maize immature tassels (a) and rice immature panicles (c). b, d Number (left panels) and overlap of 24PHAS precursors cleaved by miR2275 (right panels) in Total, TP, and MBP samples from maize immature tassels (b) and rice immature panicles (d). Cleavage sites with category = 0, which means that the reads with the maximum count (> 1) are mapped to only one indicated position on the transcript, and P value ≤ 0.05 in at least two biological repeats were filtered as miR2275 target sites. e, f Abundance of reproductive 24-nt phasiRNAs in Total, TP, and MBP samples from maize immature tassels (e) and rice immature panicles (f). Abundance of miR2275 and reproductive 24-nt phasiRNAs (a, c, e, and f), and the number of target 24PHAS precursors (left panels of b and d) are displayed as mean ± standard deviation (SD) of three biological repeats. “RPMR” is short for “reads per million rRNA fragments” (a, c, e, and f). Comparisons between MBP and TP (a and c) and between Total, TP, and MBP (left panels of b and d, e and f) were performed by two-tailed unpaired t-test, and P values are displayed to the right of the corresponding MBP miR2275s (a and c) or above the corresponding comparisons (left panels of b and d, e and f)

Fig. 7

Widespread occurrence of reproductive 21-nt phasiRNA-mediated target cleavage on membrane-bound polysomes in maize and rice. a, b Number (left panels) and overlap of identified target genes of reproductive 21-nt phasiRNAs (right panels) in input (Total), TP, and MBP samples from maize immature tassels (a) and rice immature panicles (b). Cleavage sites with category = 0, which means that the reads with the maximum count (> 1) are mapped to only one indicated position on the transcript, and P value ≤ 0.05 in at least two biological repeats were filtered as 21-nt phasiRNA target sites. Number of 21-nt phasiRNA target genes is displayed as mean ± standard deviation (SD) of three biological repeats, and comparisons between Total, TP, and MBP were performed by two-tailed unpaired t-test and P values are displayed above the corresponding comparisons (left panels of a and b). c PolyA RNA-seq read (top panel) and PARE 3′ cleavage fragment (bottom panel) coverage for Zm00001d021201, an identified MBP-unique target transcript cleaved by the 21PHAS_NO31 21-nt phasiRNA, in Total, TP, and MBP samples from maize immature tassels. d PolyA RNA-seq read (top panel) and PARE 3′ cleavage fragment (bottom panel) coverage for LOC_Os01g60440, an identified MBP-unique target transcript cleaved by the 21PHAS_NO1824 21-nt phasiRNA, in Total, TP, and MBP samples from rice immature panicles. The gene models are shown below the RNA-seq panels, with the thicker rectangles, lines, and thinner rectangles representing exons, introns, and UTR regions, respectively. In the PARE panels, the red dots indicate the cleavage sites on the transcripts targeted by reproductive 21-nt phasiRNAs. The sequences of the target sites and the corresponding phasiRNAs are shown below the PARE panels. “PARE” and “RPM” are short for “parallel analysis of RNA ends” and “reads per million mapped reads,” respectively

Fig. 8

A proposed model of miRNA-mediated target cleavage for RNA degradation or reproductive phasiRNA biosynthesis in plants. miRNA-AGO1 complexes act on target transcripts at different locations in the cytoplasm with various outcomes: (a) miRNAs guide AGO1 proteins to independently cleave target transcripts for degradation on MBPs and FPs, and in the cytosol; (b) miRNAs such as miR2118 or miR2275 can guide AGO1 proteins to cleave reproductive 21PHAS or 24PHAS precursors on MBPs, then the phasiRNA-generating cleavage fragments dissociate from MBPs and undergo phasiRNA production through previously identified players such as SGS3, RDR6, HEN1, DCL4, and DCL5. The 21-nt phasiRNAs are loaded into AGO proteins such as rice MEL1 and can cleave target transcripts on MBPs and FPs, and in the cytosol; the 24-nt phasiRNAs are shown to be polysome-depleted in this study but their functions remain unknown. 21PHAS: 21-nt phasiRNA-generating locus; 24PHAS: 24-nt phasiRNA-generating locus; AGO: ARGONAUTE; DCL: DICER-LIKE; FP: free polysome; HEN1: HUA ENHANCER 1; MBP: membrane-bound polysome; MEL1: MEIOSIS ARRESTED AT LEPTOTENE 1; RDR6: RNA-directed RNA polymerase 6; SGS3: Suppressor of Gene Silencing 3