1 Running head: maize sRNAs expression in rmr6 and abiotic stress 1 2 3 Corresponding author 4 ...

43 Pages·2016·2.43 MB·English

Plant Physiology Preview. Published on January 8, 2016, as DOI:10.1104/pp.15.01205
1 Running head: maize sRNAs expression in rmr6 and abiotic stress
2
3
4 Corresponding author
5 Prof. Serena Varotto
6 University of Padova
7 DAFNAE Agripolis Viale dell’Università, 16
8 35020 Legnaro (PD) Italy
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
Copyright 2016 by the American Society of Plant Biologists
27
28
29 Genome-wide characterization of maize small RNA loci and their regulation in the required
30 to maintain repression6-1 (rmr6-1) mutant and long-term abiotic stresses
31 Alice Lunardon1,2, Cristian Forestan1, Silvia Farinati1, Michael J. Axtell2, Serena Varotto1
32
33 1Department of Agronomy, Animals, Food, Natural Resources and Environment, University of
34 Padova, Agripolis Viale dell’Università 16, 35020 Legnaro PD Italy
35 2Department of Biology and Huck Institutes of the Life Sciences, Penn State University, University
36 Park, PA 16802 USA
37
38 Summary: Agronomically realistic, long-term drought stress mis-regulates some miRNAs and
39 induces the down-regulation of a set of small RNA loci in the maize leaf.
40
41 Authors' contributions
42 SV, AL, CF, and SF designed the experiments. AL, CF, SF performed the experiments and
43 collected the samples. AL, MJA and CF analyzed sRNA-seq and RNA-seq data. AL and SV was
44 the originator of the concept of this report. AL, SV, and MJA wrote the manuscript.
45
46 Founding informations
47 Research was supported by EU FP7 Project AENEAS and Italian EPIGEN Flagship CNR Project.
48 AL was granted by a CARIPARO PhD fellowship.
49
50 Corresponding author
51 Prof. Serena Varotto
52 [email protected]
2
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
53
54
55
56
57 Abstract
58 Endogenous small RNAs (sRNAs) contribute to gene regulation and genome homeostasis but
59 their activities and functions are incompletely known. The maize genome has a high number of
60 transposable elements (TEs; almost 85%), some of which spawn abundant sRNAs. We
61 performed sRNA and total RNA sequencing from control and abiotically stressed B73 wild-type
62 (wt) plants and rmr6-1 mutants. RMR6 encodes the largest subunit of the RNA polymerase IV
63 (Pol IV) complex, and is responsible for accumulation of most 24 nucleotide (nt) small interfering
64 RNA (siRNAs). We identified novel MIRNA loci and verified miR399 target conservation in maize.
65 RMR6-dependent 23-24 nt siRNA loci were specifically enriched in the upstream region of the
66 most highly expressed genes. Most genes mis-regulated in rmr6-1 did not show a significant
67 correlation with loss of flanking siRNAs, but we identified one gene supporting existing models of
68 direct gene regulation by TE-derived siRNAs. Long-term drought correlated with changes of
69 miRNA and sRNA accumulation, in particular inducing down-regulation of a set of sRNA loci in the
70 wt leaf.
71
72 Keywords: Zea mays, required to maintain repression6, sRNAome, miRNAs, siRNAs,
73 transcriptome, drought, salinity
74
75 Introduction
76 Plant endogenous sRNAs range in length from 20 to 24 nts and contribute to regulate gene
77 expression through RNA-mediated transcriptional gene silencing (TGS) and post-transcriptional
78 gene silencing (PTGS) mechanisms. Their activity is essential for the maintenance of genome
3
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
79 integrity, the intrinsic normal growth of cells, and proper plant development.
80 MiRNAs are mostly 21 nt long and are encoded by endogenous MIRNA genes, which are
81 transcribed by polymerase II (Pol II) to single-stranded precursors that fold into stem-loop
82 secondary structures. The stems are precisely cleaved by the DICER-LIKE 1 (DCL1) protein,
83 liberating a miRNA/miRNA* duplex (reviewed in Rogers and Chen, 2013). After the duplex is
84 loaded into ARGONAUTE 1 (AGO1), the miRNA guide strand is most frequently retained and
85 directs target repression, while the miRNA* is more often rapidly degraded. Most plant miRNAs
86 induce PTGS of their targets but some cases of TGS have been described (Bao et al., 2004; Wu
87 et al., 2010; Hu et al., 2014). Plant miRNAs share extensive sequence complementarity with their
88 target RNAs: there are only few examples with more than five mismatches between the miRNA
89 and the target (Axtell, 2013a). Many known miRNAs and their targets are conserved across
90 different plant species (Montes et al., 2014). In maize, many conserved miRNA targets have been
91 experimentally confirmed (Shen et al., 2013; Zhai et al., 2013; Zhao et al., 2013; Liu et al., 2014)
92 but a few exceptions exist, for example those of miR168 and miR399. MiRNAs are members of
93 multiple regulatory networks controlling plant development, and many miRNA families also play
94 roles in stress responses and tolerance (reviewed in Sunkar et al., 2012). For example miR156,
95 which targets SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL) genes, coordinates the
96 balance between development and abiotic stress tolerance and is important for heat stress
97 memory in Arabidopsis (Cui et al., 2014; Stief et al., 2014).
98 A specific class of endogenous siRNAs, mostly 24 nts long, participate in the
99 RNA-directed DNA methylation (RdDM) process. Pol IV transcribes target loci, typified by
100 high-levels of DNA methylation, into single-stranded RNAs that are copied into short
101 double-stranded RNAs (dsRNAs) by RNA-DEPENDENT RNA POLYMERASE 2 (RDR2; Blevins
102 et al., 2015; Zhai et al., 2015) . These short dsRNAs are then processed by DICER-LIKE 3 (DCL3)
103 into 24 nt siRNAs. Once loaded onto ARGONAUTE 4 (AGO4), these siRNAs are thought to target
104 nascent, chromatin-associated non-coding RNAs transcribed by RNA polymerase V (Pol V).
4
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
105 Successful targeting correlates with the deposition of de novo DNA methylation and other
106 repressive epigenetic marks at the target loci, inducing TGS (reviewed in Matzke and Mosher,
107 2014). Pol IV-dependent siRNAs are often produced from TEs, TE-like sequences, and other
108 repeats (Zhang et al., 2007; Mosher et al., 2008). They contribute to the reinforcement of TE
109 silencing (Slotkin et al., 2005; Marí-Ordóñez et al., 2013; Nuthikattu et al., 2013), and in some
110 cases are also essential to control the expression of protein-coding genes in cis or in trans (Liu et
111 al., 2004; Kinoshita et al., 2007; McCue et al., 2013).
112 In maize the majority of genes are located within one kilobase (kb) of an annotated TE
113 (Baucom et al., 2009; Schnable et al., 2009), and loci undergoing RdDM are primarily located in
114 gene flanking regions (Gent et al., 2013; Gent et al., 2014). The enrichment of RdDM-associated
115 siRNAs near maize genes (Wang et al., 2009; Gent et al., 2013; Gent et al., 2014; Xin et al., 2014)
116 has been suggested to ensure the continuous silencing of potentially deleterious TEs and TE-like
117 sequences in an active and accessible chromatin environment required for the Pol II transcription
118 of close genes (Gent et al., 2014). In maize, gene expression can be influenced by gene-proximal
119 TEs and repeats (Erhard et al., 2013), and both genes and TEs can be regulated by the direct
120 competition between Pol IV and Pol II occupancy at their promoters (Hale et al., 2009; Erhard et
121 al., 2015). At the genome-wide level, the expression of maize genes positively correlates with the
122 accumulation levels of upstream 24 nt siRNAs (Gent et al., 2013). In the absence of Pol
123 IV-dependent siRNAs, Pol II transcription globally decreases around the transcription start site
124 (TSS) and increases at 3’ end of genes (Erhard et al., 2015), and gene-proximal TEs lose DNA
125 methylation (Li et al., 2015). Thus the available data suggest that Pol IV-dependent 24 nt siRNAs
126 in maize primarily serve to mark and enforce boundaries between areas of transcriptionally active
127 euchromatin and transcriptionally repressed heterochromatin.
128 RdDM-associated siRNAs may also be important in adaptation to biotic and abiotic
129 stresses. Arabidopsis ago4 mutants are impaired in bacterial disease resistance (Agorio and
130 Vera, 2007) and genome-wide DNA methylation patterns are altered in Arabidopsis plants during
5
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
131 bacterial infection (Dowen et al., 2012). RdDM is also required for basal heat tolerance in
132 Arabidopsis (Popova et al., 2013). Environmental stresses trigger the expression of siRNAs that
133 modulate target genes involved in the stress response (Tricker et al., 2012; Wang et al., 2015).
134 RdDM-associated siRNAs can also defend the genome from heat-induced movements of TEs (Ito
135 et al., 2011). Here, we sought to extend our knowledge of maize stress-induced changes in
136 miRNA and siRNA accumulation to agronomically realistic drought and salinity stresses, which
137 are the major environmental stresses worldwide adversely affecting crop productivity, using
138 coupled small RNA and total RNA sequencing.
139
140
6
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
141 Results
142
143 De novo identification of sRNA loci by high-throughput sequencing
144 46 sRNA-seq libraries were sequenced from leaves of wt and rmr6-1 plants, grown under control
145 conditions (C), after ten days of three different abiotic stress treatments, and seven days
146 post-stress recovery (Data set S1). The abiotic stresses were drought (D), high salinity (S), and
147 D+S combined. An additional four libraries were obtained from C and D-treated wild-type shoot
148 apical meristematic areas. Each small RNA library had two or three biological replicates (Data set
149 S1). The majority of genome-aligned sRNAs from wild-type leaves were 24 nts long (Figure
150 1A-C). The trend toward 24 nt RNAs was even more pronounced in the meristematic libraries
151 (Figure 1A-C). Maize RMR6 encodes the largest subunit of Pol IV (Erhard et al., 2009). As
152 expected, 24 nt RNAs were mostly eliminated in the rmr6-1 mutant, and 22 nt RNAs had a slightly
153 higher accumulation level (Figure 1A-C). No major alterations in the sRNA size distribution were
154 observed in the stressed samples (Figure 1A-C).
155 We performed de novo annotation of maize sRNA loci using the merged set of all aligned
156 sRNA reads. 320,110 clusters were found (Figure 1D, Data set S2, Data set S3), of which
157 251,496 were dominated by RNAs 20 to 24 nts in length. The other 68,614 clusters were not
158 examined further because they were not likely generated by the catalytic activity of DCL proteins.
159 The remaining sRNA clusters were classified based upon their most abundant RNA size ('size
160 class'; Figure 1D). 48 MIRNA loci and 251,448 non-MIRNA loci were identified, with most frequent
161 size class of 21 nts and 24 nts, respectively (Figure 1D).
162
163 Analysis of miRNAs and their targets
164 Of the 48 MIRNA loci that we initially de novo identified (Figure 1D), 30 were previously known loci
165 annotated in miRBase (version 20; Data set S4). miRBase 20 contains a total of 159 annotated
166 maize MIRNA loci. Our de novo MIRNA discovery method is likely rather insensitive, both
7
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
167 because we wish to avoid false positive annotations, and because our automated cluster
168 discovery can sometimes incorrectly define the start and stop positions of MIRNA loci. Therefore
169 we tested the exact coordinates of each of the 159 miRBase (version 20) MIRNA loci against our
8
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
170 combined sRNA dataset (Data set S4). 67 of the 159 annotated MIRNA loci passed our stringent
171 analyses, but 64 of them failed (Data set S4). The remaining 28 loci had insufficient aligned small
172 RNAs in our data and as such couldn't be analyzed. The rather low rate of previous MIRNA
173 annotations supported by our data is consistent with observations that many miRBase MIRNA
174 annotations are questionable (Taylor et al., 2014).
175 The 18 putative novel MIRNA loci found by our initial analysis were tested for
176 reproducibility. Ten of them independently passed all criteria for MIRNA loci in at least two of our
177 libraries (Data set S5, Data set S6). As expected for true plant miRNAs, they were either 21 nts or
178 22 nts long and their accumulation was not affected in the rmr6-1 background. Three were new
179 members of the miR166 family and the other seven belong to six new miRNA families without
180 obvious homology with any other previously annotated miRNAs in miRBase.
181 Maize miR399 was previously predicted to target several messenger RNAs (mRNAs),
182 including one of unknown function (GRMZM2G165734), mRNAs encoding inorganic phosphate
183 transporters (Zhang et al., 2009), and an mRNA encoding a putative ubiquitin-like 1-activating
184 enzyme E1A (Wang et al., 2014). We found that the putative GRMZM2G165734 gene is actually
185 a MIR399 homolog. We identified more possible target sites in genomic DNA located immediately
186 upstream of GRMZM2G381709, an ortholog of Arabidopsis PHO2 (Calderón-Vázquez et al.,
187 2011). Arabidopsis PHO2 encodes an ubiquitin-conjugating E2 enzyme that plays a role in the
188 maintenance of Pi homeostasis (Bari et al., 2006). The Arabidopsis PHO2 mRNA is targeted by
189 miR399 in multiple sites in its 5’-untranslated region (5'-UTR; Allen et al., 2005). Complementary
190 DNA (cDNA) sequencing demonstrated that the GRMZM2G381709 5'-UTR encompassed the
191 putative miR399 target sites (Figure 2A). RNA-ligase mediated 5' rapid amplification of cDNA
192 ends (RLM-5'-RACE) found evidence for miR399-directed slicing at two of these sites (Figure
193 2A-B). This result indicates that miR399 targeting of PHO2 is conserved across angiosperms
194 (Bari et al., 2006), including maize.
195
9
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.
196 Long-term abiotic stresses affect the accumulation of a set of miRNAs
197 The applied abiotic stress treatments mimicked agronomically realistic long-term drought, salinity
198 and combined drought plus salinity stress conditions (Morari et al., 2015). To test their effects on
10
Downloaded from on November 18, 2018 - Published by www.plantphysiol.org
Copyright © 2016 American Society of Plant Biologists. All rights reserved.