INVESTIGATION The Role of UV-B light on Small RNA Activity During Grapevine Berry Development Sukumaran Sunitha,* Rodrigo Loyola,†,‡ José Antonio Alcalde,‡ Patricio Arce-Johnson,† José Tomás Matus,§,1 and Christopher D. Rock*,2 *Department of Biological Sciences, Texas Tech University, Lubbock TX 79409-3131, †Departamento de Genética Molecular y Microbiología, Pontificia Universidad Católica de Chile, Santiago, Chile, ‡Departamento de Fruticultura y Enología, Pontificia Universidad Católica de Chile, Santiago, Chile, and §Center for Research in Agricultural Genomics (CRAG) CSIC-IRTA-UAB-UB, Barcelona, Spain ORCID IDs: 0000-0002-1547-9902 (S.S.); 0000-0001-5638-0610 (R.L.); 0000-0002-9196-1813 (J.T.M.); 0000-0002-2090-0650 (C.D.R.) ABSTRACT We explored the effects of ultraviolet B radiation (UV-B) on the developmental dynamics of KEYWORDS microRNAs and phased small-interfering-RNA (phasi-RNAs)-producing loci by sequencing small RNAs in veraison vegetative and reproductive organs of grapevine (Vitis vinifera L.). In particular, we tested different UV-B miRNAs conditions in in vitro-grown plantlets (high-fluence exposition) and in berries from field-grown (radiation siRNAs filtering) and greenhouse-grown (low- and high-fluence expositions) adult plants throughout fruit develop- oxidative stress ment and ripening. The functional significance of the observed UV-coordinated miRNA responses was anthocyanin supported by degradome evidences of ARGONAUTE (AGO)-programmed slicing of mRNAs. polyphenols Co-expression patterns of the up-regulated miRNAs miR156, miR482, miR530, and miR828 with cognate MYBA target gene expressions in response to high-fluence UV-B was tested by q-RT-PCR. The observed light signaling UV-response relationships were also interrogated against two published UV-stress and developmental transcriptome datasets. Together, the dynamics observed between miRNAs and targets suggest that changes in target abundance are mediated transcriptionally and, in some cases, modulated post- transcriptionally by miRNAs. Despite the major changes in target abundance are being controlled pri- marily by those developmental effects that are similar between treatments, we show evidence for novel miRNA-regulatory networks in grape. A model is proposed where high-fluence UV-B increases miR168 and miR530 that target ARGONAUTE 1 (AGO1) and a Plus-3 domain mRNA, respectively, while decreas- ing miR403 that targets AGO2, thereby coordinating post-transcriptional gene silencing activities by different AGOs. Up-regulation of miR3627/4376 could facilitate anthocyanin accumulation by antagonizing a calcium effector, whereas miR395 and miR399, induced by micronutrient deficiencies known to trigger anthocyanin accumulation, respond positively to UV-B radiation. Finally, increases in the abundance of an anthocyanin-regulatory MYB-bHLH-WD40 complex elucidated in Arabidopsis, mediated by UV-B-induced changes in miR156/miR535, could contribute to the observed up-regulation of miR828. In turn, miR828 would regulate the AtMYB113-ortologues MYBA5, A6 and A7 (and thereby anthocyanins) via a widely conserved and previously validated auto-regulatory loop involving miR828 and phasi TAS4abc RNAs. The sessile nature of plants makes them vulnerable to myriad abiotic and two percent passes through the atmosphere and causes detrimental biotic stress factors to which they must adapt to survive. Ultraviolet effects on all life forms. A highly-conserved UV-B perception and (UV) irradiation, drought, heat, cold, salt, and oxygen deficiency are signaling system is evolved in plants (Heijde and Ulm 2012; Tilbrook the major abiotic factors restricting plant growth, development, and et al. 2016). Exposure to low-fluence doses of UV-B irradiation acts as productivity whereas microbial pathogens and herbivores are major an environmental trigger facilitating the expression of a wide array of biotic stress factors. Genome-wide association studies have identified genes involved in development or hormone signaling, including auxin, UV-B responses with autoimmune-like reactions due to overexpres- ethylene, abscisic acid (ABA), and cell wall modifying related genes sion of defense related genes (Piofczyk et al. 2015). UV-B radiation (Frohnmeyer and Staiger 2003; Jenkins 2017; Mackerness 2000; (280-315 nm) is mostly absorbed by the ozone layer although about Pontin et al. 2010; Ries et al. 2000). In addition to the up-regulation of Volume 9 | March 2019 | 769 antioxidant/free radical scavenging enzymes and proteins involved in high-fluence radiation expositions) (Loyola et al. 2016). HYH is in- DNA repair and cell cycle regulation (Brosché et al. 2002), a conserved duced by high-fluence UV-B three weeks after the onset of ripening, and highly specific signaling pathway involving photomorphogenic (veraison, manifested with the red/black pigmentation of berry skins receptors is also activated (Jenkins 2017). Furthermore, a few studies and a transition from organic acid to sugar accumulation). UV-B irra- have shown that photomorphogenic responses to UV influence and/or diation also promoted flavonol accumulation in leaves and berries. are influenced by several miRNA activities. For instance, Crypto- Thus, both high and low intensity UV-B exposures in grapevine may chrome1 and Cryptochrome2 blue light receptors mediate the expres- facilitate flavonol accumulation in vegetative and reproductive organs sion of miR172 (Jung et al. 2007), whereas miR408 is coordinately through the activation of HY5 and HYH at different developmental regulated by SQUAMOSA PROMOTER BINDING PROTEIN-LIKE7 stages (reviewed by Matus 2016). (SPL7) and ELONGATED HYPOCOTYL5 (HY5) transcription factors Anthocyanin accumulation is highly responsive to UV although is (TFs) in response to light (Zhang et al. 2014). developmentally regulated by subspecialized R2R3-type MYB tran- Plants fend off harmful effects of UV-B with phenylpropanoids scription factors. For example, in Arabidopsis SPL9, a target of (a group of plant-specific secondary metabolites, also known as poly- miR156, negatively regulates anthocyanin accumulation during phenols) derived from phenylalanine and deposited in the vacuoles of phase transition to flowering by directly preventing expression of epidermal cells (Dotto and Casati 2017; Li, Ou-Lee et al. 1993). Flavo- anthocyanin biosynthetic genes through destabilization of a MYB/ nols, chalcones, and anthocyanins constitute a branch of metabolites PAP1-bHLH/TT8-WD40 transcriptional activation complex (Gou from this pathway (known as flavonoids) that are synthesized by co- et al. 2011). The grapevine berry color locus in chromosome 2 en- ordinated transcriptional control of a suite of biosynthetic enzymes, codes the anthocyanin regulators MYBA1 and MYBA2 (Walker including CHALCONE SYNTHASE (CHS). The UV-B induction of et al. 2007). Conversely, the red pigmentation observed in vegetative CHS (Jenkins 2017) and the downstream flavonoid biosynthetic tissues is attributed to the activity of MYBA6, MYBA7, and potentially genes has been well characterized in several species. UV RESISTANCE MYBA5 genes, all clustered on chromosome 14 (Matus et al. 2017). LOCUS 8 (UVR8), a UV-B receptor, upon absorbing UV-B radiation MYBA5/6/7 accumulate at high levels in early berry development and monomerizes into its active form (Rizzini et al. 2011; Wu et al. 2012) decrease after the onset of veraison. UV-B radiation induces MYBA1 and interacts with CONSTITUTIVELY PHOTOMORPHOGENIC1 genes and significantly delays the down-regulation of MYBA6 and (COP1) (Cloix et al. 2012). COP1 is a WD40/RING protein that facil- MYBA7 at the latter stages of berry development (Matus et al. 2017). itates proteasome-mediated degradation in the dark of HY5, a basic Interestingly, MYBA6 and MYBA7 follow the exact expression pattern of leucine zipper TF. COP1 interaction with UVR8 releases HY5 from grapevine HY5 (Loyola et al. 2016), suggesting that MYBA6 and MYBA7 proteasome-mediated degradation and promotes photomorphogenic may also contribute to fruit pigmentation upon increased radiation stress, signaling (Huang et al. 2013). Subsequently, HY5 binds to bZIP-binding in addition to MYBA1. promoter elements of CHS, MYB12 (a regulator of flavonol biosynthesis; Grapevine VviMYBA6 and VviMYBA7 are orthologs of AtPAP1/ Lee et al. 2007, Stracke et al. 2010), and MYB75/PRODUCTION OF PAP2 and AtMYB113, the Arabidopsis MYBs known to be under ANTHOCYANIN PIGMENT1/PAP1 (a TF inducer of anthocyanin negative regulatory control by the tasiRNA species TAS4-39D4(-). biosynthesis genes; Shin et al. 2013), increasing their transcription These regulatory small RNAs are generated by miR828-mediated cleavage and facilitating the accumulation of flavonoids. of the Trans Acting Small-interfering RNA locus4 (TAS4) (Rajagopalan The grapevine (Vitis vinifera L.) is an important fruit crop abundant et al. 2006). AtMYB113 is targeted by miR828 directly in addition to TAS4 in health-promoting- and flavor-enhancing polyphenolics, with resil- 39D4(-), suggesting a close evolutionary relationship between AtPAP1/ ient metabolic and physiological responses to UV radiation. Despite the PAP2/MYB113, miR828, and TAS4 (Rajagopalan et al. 2006). TAS4 is availability of its genome for over a decade, UV-B signaling compo- functionally conserved among various eudicots (Luo et al. 2012; Rock nents and their functions in integrated control of secondary metabo- 2013) and in fact we have shown that TAS4-39D4(-) targets VviMYBA6 lism have not been identified in this species until recently. Loyola et al. and VviMYBA7, triggering 21 nt phased small interfering RNA (phasi- (2016) characterized the grape’s UV-B RECEPTOR (UVR1), HY5 and RNAs) (Rock 2013; Zhang et al. 2012). The biological significance of 21 nt HY5 HOMOLOG (HYH) that function in UV-B perception and sig- phasi-RNAs is largely unknown, but is hypothesized to be important for naling. The expression profile of the above-mentioned photomorpho- defining a concentration gradient for silencing activity across cell layers genic factors was studied in vegetative and reproductive tissues at and for widespread trans suppression of large families, including Leucine different developmental stages and in response to UV stress. While Rich Repeat (LRR) resistance genes (Fei et al. 2013; Shivaprasad et al. UVR1 is strongly expressed in early berry developmental stages, tran- 2012) or Pentatrico Peptide Repeat (PPR) genes (Howell et al. 2007). script abundance is mostly influenced by light and temperature but not Studies to date have focused on the influence of UV-B radiation in UV-B radiation. HY5, however, is induced by low-fluence UV-B in regulating anthocyanin biosynthesis at different developmental stages of plantlets and at early fruit developmental stages (both with low- and berry development. The effect of UV-B on miRNA profiles is well- documented in Arabidopsis, poplar, maize and wheat (Zhou et al. 2007; Copyright © 2019 Sunitha et al. Jia et al. 2009; Casati 2013; Wang et al. 2013), however the role of UV-B doi: https://doi.org/10.1534/g3.118.200805 in post-transcriptional regulation of MYB TFs by miR828/TAS4 path- Manuscript received December 16, 2018; accepted for publication January 9, 2019; way is unexplored. Further, there is a large annotation gap between published Early Online January 15, 2019. the empirical knowledge of siRNA expression, especially the role of This is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/ miRNAs and tasi-RNAs on production and functions of phasi-RNAs licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction (Coruh et al. 2014; Luo et al. 2009). This work focuses on a data- in any medium, provided the original work is properly cited. driven systems approach to explore the effects of UV-B radiation Supplemental material available at Figshare: https://doi.org/10.25387/g3.7471553. on miRNA/phasi-RNA expression and elucidate functional con- 1 Present Address: Institute for Integrative Systems Biology (I2SysBio) CSIC-UV, Parc Científic de la Universitat de València, 46980 Valencia, Spain. sequences for stress adaptation by controlling the biosynthesis of 2 Correspondence: 2901 Main St., Lubbock, TX 79409-3131. E-mail: chris.rock@ UV-B absorbing anti-oxidant secondary metabolites. We performed ttu.edu a detailed computational re-analysis of empirically characterized 770 | S. Sunitha et al. grapevine miRNAs for slicing activities with a degradome library using NanoDrop 1000 spectrometer (Thermo Fisher Scientific). The (Pantaleo et al., 2010). We then performed quantitative real-time integrity of RNA was assessed by an Agilent 2100 Bioanalyzer using PCR (qRT-PCR) assays on miRNA target mRNAs in same-source a sRNA chip (Agilent Technologies) according to the manufacturer’s samples used to generate our small RNA (sRNA) libraries and obtained instructions. supporting evidence for a regulatory network model of UV responses sRNA libraries were prepared using total RNA as input (1 mg) mediated by sRNAs. Meta-analysis of available mRNA transcriptome according to the instructions provided by TruSeq sRNA Sample Prep- results for similar experiments across UV treatments and developmen- aration Kit (Illumina). Twenty-eight bar-coded sRNA libraries were tal time points (Carbonell-Bejerano et al. 2014; Massonnet et al., 2017; constructed. The quality of each library was assessed using an Agilent Suzuki et al. 2015) further corroborate the model for sRNAs fine-tuning High Sensitivity DNA chip on an Agilent 2100 Bioanalyzer. Equi-molar abiotic and biotic stress responses. concentrations of libraries were pooled and sequenced on an Illumina NextSeq500v2 instrument at the Institute for Integrative Genome Bi- MATERIALS AND METHODS ology, UC Riverside. One library (minus UV-B, -3 WAV, high-fluence UV greenhouse experiment) was re-sequenced to increase reads depth UV-B treated in vitro plantlets and fruits on a MiSeqv3 platform, and another library was solely sequenced on In order to gain perspective in defining berry-specific relationships MiSeqv3 (plus UV-B, +6 WAV, field grown; Supplementary Table 1a). (Loyola et al. 2016; Czemmel et al. 2017) relative to leaf UV-B This latter library had consequently fewer reads and the 21 and 24 nt (Cavallini et al. 2015) sRNA responses, in vitro grown plantlets (four species of small RNAs were not as abundant due to overrepresentation negative control and four UV-B treated biological replicates; n = 4) of a specific 48 nt non-coding RNA homologous mapping to chromosome from the cultivar (cv.) ‘Cabernet Sauvignon’ were exposed to 6 hr of 11 intergenic region. low-fluence UV-B radiation (0.15 W m-2 irradiance) exactly as con- ducted in Cavallini et al. (2015). Leaves were harvested after +UV-B Bioinformatics and Statistical Analyses of treatment and frozen in liquid nitrogen for RNA extraction. Control Sequencing Data plants were grown under the same conditions described above but The quality assessment of the raw reads was done by FastQCv0.11.5 covered with a polyester filter (100-mm clear safety polyester plastic (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and film) that absorbed total UV-B from the spectrum without affecting the adaptor sequences were trimmed using fastx_clipper (http://hannonlab. photosynthetic active radiation. cshl.edu/fastx_toolkit/index.html). Reads longer than 18 nt were retained Potted nine-year-old greenhouse-grown cv. ‘Cabernet Sauvignon’ and were sequentially mapped to Arabidopsis thaliana rRNAs, tRNAs, fruit clusters were exposed to high (0.3 W m-2 daily for 5 hr) UV-B or snRNAs, and transposable elements (TEs; downloaded from https://plants. low (0.1 W m-2 daily for 10 hr) UV-B irradiance during 2011-2012 ensembl.org/info/website/ftp/index.html and www.arabidopsis.org) using and 2012-2013 growing season, respectively (both contrasted to control Bowtie (Langmead et al. 2009) to remove reads matching the same. Bowtie ‘filtered UV-B’ conditions; more details in Loyola et al. 2016). Sample was used to map RNAseq datasets of Massonnet et al. (2017; NCBI Bio- collection was carried out at four stages of grapevine development: -3, 0, Projects PRJNA265039 and PRJNA265040) to pre-MIR828 and down- 3 and 6 weeks before (-) and after (+) veraison (WAV, defined as the stream effector targets quantified by qRT-PCR (see below) in our field time at which clusters are 30–50% colored and sugar concentration samples. There remained a few abundant 32-53 nt long reads in the libraries reaches 5 Brix). Berries were immediately peeled, deseeded and skins after pre-filtering that mapped to XR_002032162.1 “Vitis uncharacterized were frozen in liquid nitrogen for RNA extraction. A total of 18 berries ncRNA”, XM_002283070.4 “DELETION OF SUV3 SUPPRESSOR 1”, a were sampled every 3 weeks from 9 grape clusters (n = 3, three bunches snoRNA, and a 5.8S rRNA. The pre-filtered reads were mapped to per plant and two berries per cluster). One and three biological repli- the reference Vitis vinifera 12X genome sequence, version NCBI cates were considered for deep sequencing and qRT-PCR analyses, RefSeq GCF_000003745.3 (https://www.ncbi.nlm.nih.gov/assembly/ respectively. GCF_000003745.3) (Jaillon et al. 2007) using ShortStack (version A solar UV-B filtering treatment was applied in a commercial cv. 3.8.2) (Johnson et al. 2016). In order to increase the power of Short- ‘Cabernet Sauvignon’ field, located in the Maipo Valley (Chile; 33◦ 439 Stack to call MIRNAs by sequencing at least one rare miRNA star 2899 S, 70◦ 459 9.7199 W) during 2011-2012 growing season exactly as species, we included publicly available sRNA datasets for grapevine from conducted in Czemmel et al. (2017). Berry skin samples were collected the following sources available from NCBI: Belli Kullan et al. (2015; at -3, 0, +3 and +6 WAV from a row of UV-B filtered (UV-B radiation SRR1528350- SRR1528419), Paim Pinto et al. (2016; SRR4031591- was blocked by installing a 100-mm clear polyester film at the position SRR4031638), Pantaleo et al. (2010; 2016; GSM458927- GSM458930; of grape clusters) and +UV-B irradiated plants (natural solar radiation, GSM1544381- GSM1544384), Sun et al. (2015; SRR2029783, no filter). The experiment was designed in four blocks (rows) with five SRR2029784), Wang et al. (2012, 2014; SRR390296, GSM1907875- plants each (biological replicates n = 4). Three berries per cluster (ran- GSM1907880), and Zhai et al. (2011; GSM803800- GSM803802). The domly sampled) and four clusters per plant were used for each sample. ShortStack counts output for these libraries were not used for subse- One biological replicate was considered for deep sequencing. quent differential expression analysis. ShortStack counts output gener- In total, 28 samples were used for sRNA library construction; four ated from our UV treatment libraries alone were subjected to differential for in vitro plantlets (two biological replicates, two conditions), 16 for expression analysis. greenhouse-grown berries (one biological replicate, four-time points, A few novel candidate MIRNAs were called de novo by ShortStack as four conditions [with and without high- or low-fluence UV-B), and having met all annotation requirements (viz., . 50% of reads map to eight for field-grown berries (one biological replicate, four time points, hairpin duplex, .80% of reads map to one strand of the hairpin, and at two conditions [with or without solar UV-B]). least one miRNA star species sequenced). The raw counts output of ShortStack for 40,678 sRNA-producing loci with greater than 261 total RNA extraction and sRNA libraries construction reads per cluster summed across all libraries (equivalent to 0.5 rpm Total RNA was isolated following the procedure of Reid et al. (2006) raw reads, the default for ShortStack cluster calling) was used as input using a CTAB- spermidine extraction buffer. The RNA was quantified to DESeq2 “R” Package (version 1.16.1) (Love et al. 2014) to analyze the Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 771 differential expression of sRNA cluster loci under different conditions. approach. For example, Pantaleo et al. (2010) identified targets for 13 con- The cutoff was chosen to capture the low-expressed (1.5 reads per served miRNA families out of 24, but not miR168, miR395, miR396 and million, rpm) MIR828 and MYBA6 and MYBA7 clusters, which were of miR403 despite high effector abundances. We validated 126 targets from particular interest but at the cost of more independent statistical tests. 27 of 31 deeply conserved miRNA families, including for miR168, -395, For comprehensive genome-wide assessment, an additional 25 clusters -396, and -403 (Supplementary Table 2a and Supplementary File S4). mapping to miRBase21-annotated grapevine MIRNAs (Kozomara et al. When a known canonical miRNA target was not validated due to ab- 2018) were included which were below the cutoff of 261 total reads. sence of slice signature reads in degradome datasets (miR397, -399, -408, This preprocessing step of limiting candidate clusters increases the and -827), we relied on Generic Small RNA-Transcriptome Aligner power of DESeq by removing poorly expressed loci that would other- (GSTAr) output with low Allen scores to identify the missing canonical wise contribute to ‘shot noise’ variation and excessive multiple test cor- predicted targets (all found except for miR827 targets; Supplementary rections for statistical significance, while adequately encompassing nearly Table 2a). It is noted that vvi-miR397 target LACCASE11 (Supplementary all annotated MIRNA loci (e.g., MIR858, 0.4 rpm; MIR395f, 0.05 rpm (see Table 2a) has been previously validated by 59 RNA Ligase-Mediated Supplementary Table 3). Rapid Amplification of cDNA Ends (RLM-RACE) in grapevine (Sun The experimental comparisons made for differential expression et al. 2015). We further obtained evidence for novel targets of some analyses of biologically coordinated replicates are outlined in Table 1. conserved miRNAs (miR169, -395, -396, -477, -530, -535, and -828) The differential expression of MIRNA loci and phasi-RNA-producing and for grapevine-specific miRNAs miR3623, -3624, -3626, -3629, loci as a time-series spanning berry development was tested using the -3631, -3632, 3633, and miR3635 (Supplementary Table 2a, Supple- likelihood ratio test (LRT) for inference, which assumes that the mentary File S4). The JAZ4 homolog of miR169 novel target JAZ3_1 expression data follows negative binomial general linear model. To transcription corepressor (Supplementary File S4, p. 14) has recently detect differentially expressed genes under +UV/-UV and high/low been demonstrated along with canonical miR169 targets NF-YAs to be fluence UV radiation, Wald-Log test was performed. Publicly avail- an important effector of temperature adaptation and phenology in Arab- able grape degradome data (Pantaleo et al. 2010; GSM458931) and idopsis by interaction with APETELA2-like TARGET OF EAT1/2 (targets the sRNA library datasets were subjected to PhaseTank analysis (Guo, of miR172; Supplementary File S4, p. 17; Supplementary Table 2a) that Qu, & Jin 2015) to identify the genome-wide phased siRNA-producing negatively regulate FLOWERING TIME (Gyula et al. 2018). The miRNA loci and regulatory cascades triggered by miRNAs. PhaseTank is built on targets were also predicted in parallel using the psRNATarget: A Plant CleaveLandv4.4.3 (Addo-Quaye et al. 2009). CleaveLand4 is more sen- sRNA Target Analysis Server (Dai et al. 2018) and Plant Non-coding sitive than earlier versions employed in prior analyses of GSM458931 RNA Database (PNRD) (Yi et al. 2015). The expectation threshold of (Pantaleo et al. 2010) by implementation of Generic Small RNA- psRNATarget was set to an Allen score = 4 (Fahlgren and Carrington Transcriptome Aligner (GSTAr) which calculates duplex parameters 2010) for target predictions (Supplementary Table 2b). on RNA-RNA thermodynamics instead of sequence-based alignment. ShortStack output (Supplementary Table 3 columns Q and S) lists the For False-Discovery Rate correction, the genome-wide numbers of “Major RNA” (i.e., most abundant) species and “Complexity”, calcu- MIRNA loci and phasi-producing loci identified by ShortStack and lated as (n distinct alignments)/(abundance of alignments). Low com- PhaseTank, respectively, were used as the basis for independent tests. plexity values indicate loci dominated by just a few sRNAs. Although the ShortStack metrics as reported were the average across all libraries, Identification of conserved miRNAs and targets the Major Species reported is invariably the mature miRNA species (or To identify the conserved miRNAs, all miRNA candidates and sRNA star, when it is claimed as functional, as with miR3624-3p). Further, the clusters called by ShortStack were annotated to genome coordinates of Complexity reported is quite low such that inference can be made that 163 vvi-miRNAs deposited in miRBase (version 21; Kozomara et al. differentially expressed sRNAs are predictable. The average Complexity 2018; http://www.mirbase.org/). To identify the mRNA targets that are was 0.03 for claimed MIRNA clusters (data not shown), with several sliced by miRNAs, CleaveLand 4.4 (Addo-Quaye et al. 2009) analysis loci having zero Complexity (see Supplementary File S6 for documen- was performed. Four degradome datasets were used for this analysis. tation of complexity vis-à-vis reads abundances for novel MIRNAs). Publicly available grape degradome data (Pantaleo et al. 2010; Complexity averaged 0.07 for PHASI loci across all claimed loci. For GSM458931) was pre-filtered through Vitis vinifera ncRNAs in Rfam TAS4a, where D1(+) was the major species, the Complexity was 0.002 (Kalvari et al. 2017) (Suppl. Table. 1b) and was used as the bona fide and our 28 libraries accounted for 11% of all mapped TAS4a reads degradome input. We also created three independent “pseudo-degradome” (data not shown); for TAS4b the functional 39-D4(-) siRNA was the libraries by concatenating small RNA libraries. The rationale for using major species and Complexity was 0.001 wherein our libraries small RNA libraries as pseudo-degradome inputs for CleaveLand analyses accounted for 1% of all TAS4b reads vs. the public data used for is grounded in our interest to study those 59-monophosphate sliced annotation power. For TAS4c, 39-D4(-) was the major species and polyadenylated transcripts, which are templates for production of the measured Complexity was 0.01; for this key hypothesized effector phasiRNAs triggered by miRNA activities. Such species of 21 nt of MYBA6 and MYBA7 expression our libraries accounted for 91% of diced dsRNA products are well-represented in sRNA libraries all TAS4c cluster reads. Taken together, these data support our view (Luo et al., 2009; Rock 2013). Further, independent corroboration that statistical inference of miRNA and siRNA dynamics using clus- of slicing activities from different sources improves the statistically ters as proxy for bona fide effectors is reasonable and useful. defensible rigor of CleaveLand (Addo-Quaye et al. 2009). We there- fore concatenated all 28 UV-B test sRNA libraries and employed RNA blotting and Densitometry them as a biological replicate, as well as the four sRNA libraries from Total RNA (15 mg) was loaded onto a 17% polyacrylamide gel with 7M Pantaleo et al. (2010) and independently the 14 sRNA datasets used urea and was electrophoresed in 0.5X TBE at 200 V. RNA was blotted for MIRNA empirical characterizations with ShortStack (Belli Kullan onto the positively charged Amersham Hybond-N+ nylon membrane et al. 2015; Paim Pinto et al. 2016; Sun et al. 2015; Wang et al. 2012, (GE Healthcare Life Sciences, USA) using the Owl semidry trans- 2014; and Zhai et al. 2011). The identification of most of the canonical, fer apparatus (Thermoscientific, USA) and the membrane was deeply conserved miRNA targets validated the robustness of the UV-crosslinked (SpectroLinker XL-1500, Spectroline, Westbury NY). 772 | S. Sunitha et al. n Table 1 Schematic of various differential expression comparisons described in Results Experiment Comparisons DESeq2 Statistical test Supplementary Table S3a. Effect of +UV-B plantlets vs -UV-B plantlets. Two Wald-Log test low-fluence UV-B on in vitro plantlets biological replicates Supplementary Table S3b. Effect +UV-B (low-fluence, high-fluence and field Wald-Log test of UV-B on berries samples) vs coordinate -UV-B (low-fluence, high-fluence and field samples). Three replicates Supplementary Table S3c. Effect +UV-B and -UV-B high-fluence vs co-ordinate +UV-B Wald-Log test of UV-B fluence on berries and -UV-B low-fluence. Four replicates Supplementary Table S3d. sRNAs 23 WAV, 0, and +3 vs +6 WAV coordinated to three Likelihood Ratio Test differentially expressed during berry conditions (low-fluence, high-fluence and field samples) development and two treatments (+UV, -UV). Three replicates A 22-nt anti-vvi-miR828 Locked-Nucleic Acid oligonucleotide Data availability (Válóczi et al. 2004) (Exiqon Inc., Woburn, MA; www.exiqon.com) The raw data files of UV sample libraries can be accessed from the was end labeled with [g-32P]ATP, 6,000 Ci/mmol, (Perkin Elmer, National Center for Biotechnology Information (NCBI) under Sequence www.perkinelmer.com) and used as probe. Hybridization was per- Read Archive accession SRP131396. Supplemental material available at formed at 37 for 16-20 h with PerfectHyb Plus hybridization buffer Figshare: https://doi.org/10.25387/g3.7471553. (Sigma). Post-hybridization washes were performed as follows: The hybridization solution was discarded, and the blots were washed se- RESULTS quentially with 2X SSC/0.2% SDS, 0.5X SSC/0.2% SDS and 0.1X SSC/ As UV-B irradiation increases the expression of VviHY5 and of its 0.2% SDS. Each wash was done for 30 min at 37 in the hybridization regulatory network (i.e., MYBF1, FLS4, MYBA6 and MYBA7) in both oven. As an equal loading control, the membrane blot was re-probed leaves and berries of UV-treated plants, favoring flavonol and antho- with end labeled [g-32P]ATP 5S rRNA oligonucleotide. RNA blots cyanin accumulation (Loyola et al. 2016; Matus et al. 2017; Czemmel were scanned using a Storm 860 PhosphorImager and sRNA signals et al., 2017), we choose to explore sRNA abundances in all these pre- were quantified using ImageLab software (v6.0, Bio-Rad. www.bio- vious conditions in order to provide a perspective in defining organ- rad.com). We segmented the miR828, 5S rRNA band areas into seven specific UV-B responses. Small RNA (sRNA) libraries were constructed vertical subsections of equal area per lane. Values of five subsections for four in vitro-grown leaf samples (control and UV-B treated, two were averaged after discarding the leftmost and rightmost subsections, biological replicates) and 24 berry skin samples (control and UV-B which in some lanes showed gel migration artifacts. An area of the same irradiated/filtered samples collected from greenhouse and fields at -3, subsection dimensions adjacent to bands was subtracted from the sig- 0, +3 and +6 Weeks After Veraison, WAV) and sequenced by Illumina nal to remove background noise. A ratio of miR828 average signal NextSeq500v2 (1 · 75 · 6 cycles for indexes). We obtained a total of divided by the control 5S RNA average signal from each lane was 356,090,480 raw reads (Supplementary Table 1a). After adaptor trim- calculated to normalize across samples for lane loading variations. ming and removal of rRNA, tRNA, snRNA, and TE-like sequences The test comparison (UV treatment -fold change) for expression longer than 18 nt found in the 28 libraries (Supplementary Table 1a), strength was calculated as the ratio of normalized plus UV sample there remained 231,853,790 clean reads (65.7% of raw reads), of which signal divided by the cognate minus-UV sample normalized signal 8% on average were unique within libraries and 5.2% were unique (set to unity) for a given tissue (in vitro plantlet replicates), develop- across all libraries. Reads longer than 27 nt were retained in the dataset mental time point, greenhouse fluence condition, and field condition. for quality control assessment of libraries, which contained discreet fragments of grapevine mRNAs and a few ncRNAs, snoRNA and Quantitative Real-Time PCR rRNA species of 32, 48, 51, and 53 nt sizes not filtered with Arabidopsis cDNA amplification was conducted as in Matus et al. (2009). Quanti- sequences (Supplementary Fig. S1). ShortStack mapped 31% of reads fication of relative gene expression was carried out as in Dauelsberg uniquely to the reference genome, while 32.8% of reads mapped to et al. (2011), following a thermocycling program of 95 for 10 min multiple loci and were binned proportionally based on flanking se- followed by 40 cycles at 95 for 20 s, 55-60 for 20 s, and 72 for 20s, quence uniqueness (Johnson et al. 2016). Only 0.3% of the reads followed by 71 cycles increasing from 50 to 96 at increments of 0.5 per mapped to more than 50 loci and were discarded. Principal component cycle for 30s to obtain a melting curve. Standard quantification curves analysis of berry samples showed a near majority of variation in abun- with serial dilutions of PCR products were constructed for each gene dances of sRNAs mapping to 40,678 loci between libraries that was to calculate amplification efficiency. Melting temperatures, primer strongly associated with berry development (48%, PC1) (Figure 1). The sequences and amplification efficiency coefficients are shown in chronological sequence from green berry (-3 WAV), veraison, +3 Supplementary File S7. VviUBIQUITIN1 was used as reference gene WAV, and +6 WAV clustered coordinately and continuously irrespec- for calibration. To differentiate the transcripts of Tas4a, Tas4b and Tas4c tive of samples being collected from greenhouse or field (Figure 1). The and to amplify specific transcripts, primers were designed flanking the +6 WAV sRNA genome-wide variations were more dynamic (clustered endonucleolytic slicing sites for miR828 and TAS4 39D4(-) in highly loosely) than earlier well-staged/constrained time points, consistent variable regions (Supplementary Fig. S2A). qRT-PCR amplicons were with greater transcriptome variation at harvest reported by run in 1% agarose gels by electrophoresis and purified and sequenced Massonnet et al. (2017) among red berry cultivars than between to validate appropriate designs. All experiments were performed with red and white cultivars. three (greenhouse experiment) and four (field experiment) biological The size distributions of trimmed reads were analyzed showing replicates and two or three technical replicates. that all libraries had distinct peak abundances at 21- and 24-nt Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 773 Figure 1 Principal component (PC) analysis of all 24 grape berry library sRNA-generating loci subjected to differential expression anal- ysis. The percentage of variation is depicted in the PC1 and PC2 axes. Based on temporal clustering of colors, PC1 is inferred to capture the developmental stage (WAV, weeks after veraison) of samples. HF: high-fluence UV-B greenhouse berries; LF: low-fluence UV-B greenhouse berries. (Supplementary Fig. S1). Consistent with previous reports in grapevine sample size and high read number), PhaseTank analysis revealed a non- (Pantaleo et al. 2010; Paim Pinto et al. 2016), the 21-nt species was coding phasi-RNA candidate TAS locus (chr13:15496204-15497757) more abundant when compared to the 24-nt sRNA species. To iden- (ShortStack phase score 1,737) generating a tasiRNA [39D3(-)] by vvi- tify the conserved miRNAs, all miRNA candidates and sRNA clusters miR482 cleavage (Allen score = 5.0; P , 0.07) (Supplementary Table called by ShortStack were annotated to genome coordinates deposited 4 and File S5). This novel candidate TAS locus significantly up-regulated in miRBase. The following MIRNAs were not found in the ShortStack by UV treatment (Supplementary Table 3a) was named vvi-TAS11 dataset: vvi-MIR169akw, -171dgj, -395k, -399f, -535a, -845ade, and following the nomenclature of Zhang et al. (2012). vvi-TAS11 maps to vvi-MIR3631c. Other MIRNAs identified in the dataset and reported the 59 end of an unannotated 87aa peptide-encoding locus VIT_ in the literature but not in miRBase were vvi-miR529, vvi-miR530 13s0047g00100 with ESTs (GenBank XR_002031633.1, XR_787251.2) (Pantaleo et al. 2010), vvi-miR827 (Alabi et al. 2012; Han et al. 2016; annotated as ’uncharacterized non-coding RNA’. Additionally, degra- Pantaleo et al. 2010), vvi-miR858 (Pantaleo et al. 2010), vvi-miR4376, dome evidence (Allen score 4.0; Supplementary Table 4) supports vvi-miR5225 (Wang et al. 2011; Xia et al. 2013), and vvi-miR7122 cleavage of a phasi-RNA-producing (phase score =93.5) Leucine- (Pantaleo et al. 2016; Xia et al. 2013). Also present in the dataset Rich-Repeat (LRR) transcript by the tasiRNA39-D3(-), which is pre- as differentially regulated we identified 13 novel candidate MIRNAs, dicted to target another phasi-RNA-producing LRR (phase score 29.2), including several with weak hairpin homology to other plant species 12 NB-ARC-LRRs (Supplementary Tables 2b, 3a, 4), and the putative in the miRBase, e.g., miR5741/8685, miR7099, and miR8761. The cation/hydrogen exchanger Vvi-CHX15. novel MIRNA hairpin sequences, their refseq genome coordinates, Induction of MYBA6 and MYBA7 in vegetative tissues upon UV-B and read counts are documented in Supplementary File S6. The radiation has been observed previously (Matus et al. 2017). Similarly, miRNA expression profiles under different experimental set up and miR828/TAS4-mediated post-transcriptional regulation of MYBA6 and their experimentally confirmed/putative mRNA targets are discussed MYBA7 by production of TAS"4-39D4(-)-triggered phasi-RNAs is also in detail below. well documented in leaves (Rock 2013). The extremely low levels of mature miR828 (1 per 20 million reads) in the sRNA libraries, in- UV-B responsive miRNAs in in vitro-grown plantlets cluding leaves (where it is expected to be maximal), made it difficult to Two biological replicates from low-fluence UV-B irradiated plants and quantify by deep sequencing. Hence, a more sensitive RNA blot analysis two samples from the control treatment were analyzed for differential with Locked Nucleic Acid-anti-vvi-miR828 as probe was performed to expression of miRNAs (Supplementary Table 3a). Two UV-B responsive verify the expression profile of MIR828 cluster siRNAs (miR828 was miRNAs, vvi-MIR5225 and vvi-MIR160e, which target the phasi-RNA the major species found at 2.6 reads per million; Supplementary Table spawning locus Ca2+-ATPase10 (Supplementary Table 2a) and ARF10/ 3c,d) quantified by ShortStack. As predicted, mature miR828 was max- 17 TFs, respectively (Supplementary Table 2a; 2b; Supplementary File imal in leaves, but there was no obvious difference in miR828 levels in S4, p5), were identified as significantly up-regulated (Supplementary response to UV-B in plantlets (Figure 2; Panel A; but see below). Table 3a). miR5225 is a 22 nt species evolutionarily related to miR3627 and miR4376 (Xia et al. 2013; Wang et al. 2011). In grape miR5225, UV-B responsive small RNAs in berry skins from miR3627 and miR4376 target the 59UTR intron2 region of the calcium- greenhouse and field experiments ATPase10, 14-40 nt upstream of the translation start site as independently The high- and low-fluence UV irradiated greenhouse and field samples demonstrated by the CleaveLand analyses of miR4376 and iso-miR3627 were classed as +UV-B test samples and the corresponding minus (-) in public and Pantaleo et al. (2010) sRNA pseudo-degradome datasets UV-B as the control samples for differential expression analysis (Table 1). (Supplementary Table 2a; Supplementary File. S4, pp. 56, 57). The orthol- Eleven UV-B responsive miRNAs were identified with statistically ogous genes in cassava, hop, and mangrove have expressed sequence tags significant differential expression (independent of developmental (NCBI JG970650.1, ES653410.1, DB992577.1) mapping to the miR5225/ time points), of which nine were up-regulated (vvi-MIR395n, vvi- 3627/4376 target site and downstream through the translation start site, MIR3627, vvi-MIR535c, vvi-MIR3624, vvi-MIR171b, vvi-MIR156f, suggesting calcium-ATPase10 intron2 is mis-annotated. vvi-miR4376, vvi-MIR319c) and two were overall down-regulated Although vvi-MIR482 expression was not significantly changed by (vvi-MIR477b, vvi-MIR530) (Supplementary Table 3b). Conversely, UV treatment in in vitro plantlets (P = 0.21, possibly due to small further analysis (described below) showed those two down-regulated 774 | S. Sunitha et al. Figure 2 sRNA blot of miR828 abundance in UV response samples subjected to deep sequencing. The quantitation of mature miR828 abundances by densitometry was used to validate ShortStack quantitation of relative vvi-MIR828 sRNA cluster abundances, which are used in statistical analyses as a proxy for mature miR828 species abundance since mostly miR828 species were sequenced from sRNA libraries. A) UV-induced expression of miR828 in in vitro plantlets. –UV: No UV-B treatment; +UV-B: 6 hr of low UV-B radiation (0.15 W m-2 irradiance). B) UV-induced expression of miR828 at different stages of berry development in greenhouse conditions subjected to high-fluence UV-B. –UV: No UV-B treatment; +UV-B: (0.3 W m-2 daily for 5 hr). C) UV-induced expression of miR828 at different stages of berry development in green house conditions when subjected to low-fluence UV-B. –UV: No UV-B treatment; +UV-B: (0.1 W m-2 daily for 10 hr). D) UV-induced expression of miR828 at different stages of berry development in field conditions UV-B. –UV: Solar UV-B blocked by 100 mm clear polyester film; +UV-B: No filters. Panels B-D: t1, t2, t3 and t4 corresponds to Weeks -3, 0, 3, and 6 after veraison (WAV), respectively. As loading control, 5S rRNA probe was hybridized to the same membrane. The relative abundance of miR828 in test samples (+UV-B) is presented as the ratio compared to normalized abundance of -UV-B controls (set to unity). MIRNA loci are up-regulated in response to high-fluence UV-B. The triggered by miR390 (Zhang et al. 2012; Supplementary File S4, up-regulated UV-B-responsive miRNA targets were validated by p. 19; Supplementary File S5, see below). In addition, a phasi-RNA- CleaveLand analysis and corroborated by psRNATarget, Plant producing locus (unannotated, but homologous to rice Oryza|ChrUn. Non-coding RNA Database (PNRD) (Yi et al. 2015), and evidences fgenesh.mRNA.47_GX23P; https://phytozome.jgi.doe.gov) was pre- from available literature (see Supplementary Table 3b; Supplementary dicted to be targeted by miR2111 (Supplementary Table 4, prediction Tables 2a, 2b; Supplementary File S4). CleaveLand analysis validated sheet), as well as some evidence for miR2111 slicing of NHD1, Sodium Metal Ion Binding Protein mRNA, previously reported by Sun et al. hydrogen antiporter (Supplementary File S4, p. 58), but miR2111 was (2015) and Pagliarani et al. (2017) as the target of MIR3624-3p (Sup- not significantly regulated by UV-B and not considered further. plementary Table 2a; Supplementary File S4, p. 62), despite the Matus et al. (2017) reported that berries irradiated with UV-B MIR3624 locus produces abundant phasi-RNAs from the antisense delayed the decline in MYBA5/6/7 expression after the onset of veraison, strand (Supplementary Table 3b). supporting the role of MYBA6 and MYBA7 in favoring pigmentation CleaveLand and PhaseTank analysis identified vvi-MIR482 targeting during fruit development under increased abiotic stress conditions. Luo a second TAS locus (chr18: 20665102-20666018), annotated as overlapping et al. (2012) and Hsieh et al. (2009) characterized in Arabidopsis an the 39 UTR of an 83aa transmembrane peptide transcript VIT_ auto-regulatory feedback loop involving the up-regulation of miR828/ 18s0072g01090 (Supplementary Table 2a, 2b; Supplementary File S4, TAS4 by AtPAP1/MYB75 (an ortholog of VviMYBA6/7) and being p. 37). An abundant tasiRNA 39D12(-) significantly up-regulated by themselves targets of the negative regulators miR828/TAS4. Piya et al. UV is analogous by concordance of phase (but not by homology) to sly- (2017) recently described a similar auto-regulatory loop for miR858 and TAS5 (Li et al. 2012; data not shown) and likewise is predicted to target MYB83. If such feedback loops are conserved in grapevine, an increase eight NB-ARC/TIR-LRR transcripts (Supplementary Table 2b, File S5), in MYBA6/7 transcript levels would increase the levels of miR828- and/ thus we named it vvi-slyTAS5-like. This candidate TAS locus produces or miR828-triggered TAS4-39-D4(-) from VviTAS4abc loci, which a tasiRNA39-D6(+) predicted to target NB-LRRs (Supplementary Table cleave MYBA6/7 to generate phasi-RNAs that in turn amplify the 4 PhaseTank prediction, Supplementary File S5). Degradome evidence post-transcriptional silencing of these MYBA genes (Rock 2013). Infer- demonstrated the tasi-RNA D6(+) from vvi-slyTAS5-like also directs ence of miR828 expression by deep sequencing read abundances of cleavage (Allen score 3.5) of an Armadillo-type fold domain found miR828 (being much more abundant than the rarely sequenced in nuclear cap-binding protein CBP80 and BB-LRR variants an miR828), suggested overall down-regulation of MIR828 locus by S-adenosyl-methionine-dependent methyltransferase in the family of UV-B treatment (Supplementary Table 3b). Yet we observed a signifi- catechol/caffeoyl-CoA phenylpropanoid biosynthesis enzymes. An- cant UV-B induced accumulation of TAS4b-39D4(-) (Supplementary other PHASI locus with an unknown trigger produced an abundant Table 3b ^ footnote) that is functionally contingent upon miR828 ac- siRNA up-regulated by UV-B with homology to miR6173 (Liu et al. tivity, consistent with the results of Matus et al. (2017) and the hypoth- 2015; data not shown). We identified from degradome analysis a phasi- esized auto-regulation of miR828/VvTAS4abc loci. This hypothesis was RNA-producing and cis-cleavage active locus triggered by miR390 substantiated by CleaveLand analysis of sRNA pseudo-degradome in possibly involving the 221 two-hit mechanism (Axtell et al. 2006; Sup- which we observed miR828 slicing of VviTAS4a/b/c and the phased plementary Table 2a; Supplementary File S4, p. 18; Supplementary File S5) TAS4ab-39D4(-) species targeting MYBA7, MYBA6, and MYBA5/ associated with AGO7 complex (Montgomery et al. 2008; Rajeswaran MYB113-like/VIT_14s0006g01340 (Supplementary Table 2a; File S4, and Pooggin 2012). These phasi-RNAs are generated from VIT_ pp. 38-43). Because vvi-MIR828 was not significantly correlated to 12s0059g01410 encoding a 66aa signal peptide with limited homology UV-B by DESeq2 analysis across all fluence and field treatments, we to Arabidopsis TASIR-ARF/At5g57735 and distinct from VviTAS3 characterized mature miR828 expression alone by RNA blot (Figure 2). Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 775 Figure 3 Functional validation of differentially expressed miRNA activities under high-fluence UV-B by qRT-PCR of target genes. Upper Panel A-C, left panel D: miRNA/siRNA deep-sequencing profiles at different berry developmental stages (weeks after veraison, WAV) under high- fluence UV-B in greenhouse. Lower Panels A-C, right panel D: miRNA target gene expression profile at different berry developmental stages under high-fluence UV-B. A) miR156f and its target Squamosa promoter binding protein-like 2 (SPL2) expression profile. B) miR3632-3p and its target Leucine-rich repeat protein (LRR) expression profile. C) miR530a and its target Plus3 domain protein expression profile. D) miR482 expression profile (left panel). Middle Panel: TAS11 39 D3(-) trans acting siRNA expression profile triggered by miR482. Right Panel: Expression profile of NB-ARC LRR VIT_12s0034g01460 targeted by TAS11 39 D3(-) tasiRNA. E) TAS4c 39 D4(-) trans acting siRNA expression profile (left panel) and the expression profile of TAS4c undiced primary transcript (right panel; “NA”: not available. Test samples were normalized to 0 WAV minus UV-B [set to unity]). Error bars are s.d. Asterisks () denote significant differences based on analysis of variance (n = 3 biological replicates) and comparisons using the Tukey-Kramer honestly significance test (HSD; P , 0.05). Insets, A-E: Target-miRNA/siRNA binding positions (mRNA nucleotide coordinates) and primer binding sites (arrows) are depicted to scale. Target:miRNA binding depicted in Red; Target:tasiRNA binding depicted in Green. Interestingly, semi-quantitative densitometry analysis of band intensi- demonstrating a robust UV-B response of mature miR828 despite ties showed reproducible up-regulation of miR828 by high-fluence the very low abundance found in our sRNA libraries. These results UV-B treatments in both greenhouse and field fruits, and down-regulation suggest a complex regulation (up-regulation by high-fluence despite of miR828 by low-fluence UV-B greenhouse treatments (Figure 2B-D). being down-regulated by low-fluence UV-B) of vvi-MIR828 locus, The binomial distribution probability for the concordant UV-B up- and representing a behavior possibly shared by other miRNAs. miR828 down-regulations of mature miR828 across all 11quantified greenhouse also appeared to be down-regulated over developmental time in and field sample signals of Figure 2 to occur by chance was P , 0.0005, high-fluence UV-B experiments (Figure 2B, D). 776 | S. Sunitha et al. The functional significance of differentially expressed miRNAs upon n Table 2 Normalized RNAseq quantification of pre-MIR828, high UV-B influence in the greenhouse and field was tested by real-time TAS4, and target MYBA6 and -MYBA7 mRNAs in berry skins of quantitative PCR (qRT-PCR). The expression levels of validated target five field-grown red and white cultivars during development genes were assayed from the same total RNA samples used to construct (datasets of Massonnet et al. 2017) the sRNA libraries. The results for greenhouse treatments are shown in Berry Developmental Stage Figure 3. An inverse correlation was indicated between the observed Datasets/gene Pea-sized Just prior veraison Soft Harvest miRNA dynamic and some target gene expression profiles during berry Red cultivars (reads per 100 Million) development, although the relationship was positively correlated in pre-MIR828^ 1.1 0.0 0.2 0.0 discrete time points. Differentially expressed miR156f, miR3632-3p TAS4a^ 31.4 16.5 6.0 0.4 (homologous to miR482) and miR530a displayed at certain times of TAS4b^ 0.3 0.2 0.0 0.0 berry development an inverse relationship with their cognate target TAS4c& 0.0 0.0 13.6 8.7 genes Squamosa promoter binding protein-like2_3 (Supplementary Ta- MYBA6$ 5.1 1.6 10.3 0.9 ble 2a; Supplementary File S4, p. 1), numerous homologs of VIT_ MYBA7^ 19.9 13.3 0.0 0.0 13s0067g00790 encoding Leucine-rich repeat protein (LRR) (Figure pre-MIR166e 26.5 20.2 13.2 7.8 3B; Supplementary Table S2b; Supplementary File S4, p. 66), and Plus3 reference^ Domain protein-coding genes (Supplementary Table S2a; Supplemen- White cultivars pre-MIR828^ 1.3 0.2 0.0 0.0 tary File S4, p. 11), respectively. However, qRT-PCR results for (-) TAS4a^ 53.8 41.7 19.1 3.9 UV-B control samples indicated miR156f likely did not change during TAS4b^ 4.6 4.3 0.0 0.0 development whereas assayed paralogue SPL2_1 decreased sharply TAS4c 0.0 0.0 0.0 0.0 from -3 to 0 (Figure 3A). A similar developmental dynamic was ob- MYBA6$ 6.0 3.7 5.2 3.8 served for miR530 target Plus3 Domain suggesting that UV-B inverse MYBA7^ 8.2 0.0 0.2 0.0 relationships may not be observable before +6 WAV (Figure 3C). A pre-MIR166e 23.0 14.4 5.8 3.9 direct relationship was observed between miR482 and derivative tasi- reference^ RNA TAS11 39-D3(-) abundance changes during berry development  96 nt reads mapping completely to amplicons quantified by qRT-PCR in Fig. and in response to high UV-B (Figure 3D, left and middle panels). Likewise S2, ,= 4 mismatches. See Supplementary Table 7 for amplicon target details. to miRNA:target dynamics, the TAS11 regulatory cascade showed evidence Pre-vvi-MIR166e target sequence from miRBase22; TAS4b reads are for 105 bp amplicon plus 100 nts upstream and downstream because 96 nt read lengths of activity against downstream target NB-ARC LRR protein gene (Supple- precluded results. mentary Table 2b; Figure 3D, compare middle and right panels). Taken ^ The decreasing pre-MIR828, TAS4ab, MYBA7, and pre-MIR166e expressions together and in general, the dynamics observed between miRNAs and as development progresses correlate with the derivative miR828, TAS4 siRNA, target expressions suggest that changes in target abundance are mediated MYBA7 mRNA, and mature miR166 abundances documented in field samples, Supplementary Table 3d and Supplementary Fig. S2. transcriptionally and modulated in some cases post-transcriptionally by & The increasing TAS4c expression as development progresses (bold) correlates miRNAs, with the major changes in target abundance controlled primarily with mRNA and derivative TAS4c tasiRNA 39D4(-) abundances documented in by developmental effects that are similar between treatments. miRNAs field samples, Supplementary Table 3d and Fig. S2E. $ may act not as dominant negative effectors to drive down target expression The fluctuations of MYBA6 expressions at veraison and soft berry stages (italics) (i.e., presenting inverse correlations), but more like a “rheostat” to limit or correlate with the derivative siRNA abundance documented in field samples, Supplementary Table 3d and Fig. S2F, left panel. fine-tune the accumulation of target mRNAs. Interestingly, a direct concordance was observed in greenhouse high-UV treatments between the tasi-RNA TAS4c 39D4(-) and its pri- that miR828 (Figure 2B, D) and targets TAS4ab (Supplementary Fig. S2) mary (undiced) TAS4c transcript (Figure 3E), supporting the notion decrease during berry development, as does MYBA7 (TAS4-39D4(-)’s that miR828 induction by high-fluence UV-B throughout berry devel- target), whereas miR828 target TAS4c (Figure 3E, Supplementary Fig. opment (as shown in Figure 2B, D) results in increased TAS4 expres- S2E) paradoxically increases during berry development, but remarkably sion, marked by the accumulation of TAS4 phasi-RNAs triggered by only in red-skinned cultivars (Table 2). This, along with the concordant miR828 (Figure 3E, left panel; Supplementary Table 2a; File S4, pp. observation that MYBA6-derived tasiRNA expression dynamics (Sup- 38-40). Despite their radiation-responsiveness, the UV-induced and plemental Fig. S2F) are associated with higher MYBA6 expression at developmental trends of TAS4c-39D4(-) and miR828 are opposite to post-veraison stages (Table 2), may indicate some unique function of each other (see below). Similar results correlating the apparent de- TAS4c in MYBA6 regulation for anthocyanin accumulation. creases in mature miR828 expression after veraison (Figure 2B,D) were Our findings raised the question of why TAS4 primary transcript observed for pre-MIR828, derived miR828 abundance, TAS4a relative abundances are not reduced by high UV-B treatments; i.e., an 39D1(+) and TAS4b 39D4(-) tasi-RNAs and their respective primary inverse relationship with miR828 at veraison (Figure 2, Supplementary TAS4 transcripts from high-fluence UV-B field samples (Supplemen- Fig. S2) as seen for other miRNA targets and tasi-RNA cascades tary Fig. S2B-E). Despite the very low mature miR828 expressions (above). We observed high-fluence UV-B treatments in greenhouse observed at veraison (Figure 2B, D), its high induction (.threefold and field significantly delayed the down-regulation of MYBA6 and change) by UV treatment measured by blot densitometry was also MYBA7 at the latter stages of berry development (Supplementary reflected in the normalized abundance of UV-induced miR828 reads Fig. S2F, G right panels; Matus et al. 2017). We propose, based on these and in the UV-induced pre-MIR828 levels assayed from the same total results, that the observed concordant expressions of miR828 (Figure 2, RNA field samples (Supplementary Fig. S2B). In order to further sub- Supplementary Fig. S2B), TAS4abc primary transcripts and diced tasi- stantiate these results over developmental time points for a very low- RNAs (Figure 3E, Supplementary Fig. S2C-E) and the derived MYBA6 expressed miR828 and derivative effectors, we analyzed the RNAseq and MYBA7 phasi-RNAs spawned by the activity of TAS4ab-39D4(-) datasets of Massonnet et al. (2017) for reads mapping to our amplicon (Supplementary Fig. S2F, G left panels; Supplementary File S4, pp. 41, 42), pre-MIR828, TAS4, and MYBA6/7 targets quantified by qRT-PCR. The are strong evidence in support of a functional orthologous MYBA6/7 results are shown in Table 2 and further corroborate our observations auto-regulatory loop as described in Arabidopsis (Hsieh et al. 2009; Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 777 Luo et al. 2012) acting either directly or indirectly to activate expression sRNAs and is not stranded (it produces abundant antisense tran- of both MIR828 and TAS4abc. scripts; Supplementary Table 3b). vvi-MIR399i, which is predicted The observed complex regulation of miR828 by UV-B detected by to target phosphate transporters (Supplementary Table 2a) (Mica RNA blot and corroborated by sRNA reads and qRT-PCR prompted us et al. 2010), and vvi-MIR167c, targeting ARF TFs (Supplementary to examine other significantly differentially expressed MIRNAs (all Table 2a; Supplementary File S4, p. 9), were some of the miRNAs analyzed by DESeq2) for a pattern of up-regulation by high UV fluence up-regulated by high-fluence UV-B vs. low-fluence reference samples. and down-regulation by low UV fluence. Table 3 lists those signifi- Two interesting observations were the coordinated downregulation of cant UV-B differentially regulated miRNAs and siRNAs mimicking the all six miR403 family members (five of them significantly) by high-fluence pattern (up in high fluence, down in low fluence) observed by RNA UV-B, that target ARGONAUTE 2 (AGO2) in dicots (Supplementary blot for miR828 (Figure 2), miR828 and pre-MIR828 at veraison Tables 2a, 2b; Supplementary File S4, p. 34). AGO2 is correspondingly (Supplementary Fig. S2B). We further endeavored to impute support- up-regulated by UV-C and UV-B (Table 3) and associated with stress ing evidence for miRNA function by meta-analysis with respect to the responses including DNA damage (Fátyol et al. 2016; Wei et al. 2012). observed UV-B changes of miRNA abundances using co-expression The up-regulation of MIR403f primary transcript during berry matura- analyses of validated mRNA targets. The subject datasets employed tion in field samples (Supplementary Fig. S3) validated our observation of were three genome-wide transcriptomics datasets of grapevine berry significant up-regulation across developmental time points for miR403 skin tissues in response to 1 hr of UV-C exposure (Suzuki et al. 2015), species (Supplementary Table 3d). TAS4a targeted by vvi-miR828 (phase five weeks exposure to UV-B after veraison (Carbonell-Bejerano et al. Score: 18,631; Supplementary Table 2a; File S4, p. 38), a miR828-targeted 2014), and from five red-skinned varieties sampled across berry de- MYB (phase Score: 1,260, Supplementary File S4, p. 47), and a miR482/ velopmental stages (Massonnet et al. 2017). It is noted that UV-C lamps Vv-sly-TAS5-like-39D6 (+)-triggered LRR PHASI locus (phase score have about 5% of their energy emitted in the UV-B band (UV-C light 398.3; Supplementary Table 2b) were significantly up-regulated by source GL-15 lamps have power emission in the UV-B range; http:// high-fluence UV-B, consistent with observed trigger miRNA abundances www.nelt.co.jp/english/products/safl/) supporting our view that these (Figure 2, Supplementary Table 3c). However, TAS4c (which has a comparisons are worthwhile. Table 3 shows the normalized abun- 39D4(-) variant in our cv. ‘Cabernet Sauvignon’ sRNA libraries that dances (ratio of +UV/Control) of selected miRNAs for four time points targets VviMYBA6/7 less well due to a mismatch at seed position 10; and three environments in response to UV-B (minus UV-B set to Supplementary Table 2b), despite evidence for being up-regulated unity), and the reported fold-change effects on validated miRNA target by high-fluence UV-B per se (like TAS4ab; Table 3), was expressed mRNAs (Supplementary File S4; Supplementary Table 2b) after 1 hr highly in both control and low-fluence UV treatments compared to UV-C and UV-B treatments. Binomial distribution tests of significance the high-fluence conditions (not shown), resulting in an overall negative for the observed inverse relationships to occur by chance (a function of coefficient for high-fluence effect like seen for miR403 family members number of targets) revealed significant results collectively, and individ- (Table 3; Supplementary Table 3b,c). These results suggest an additional ually for all 10 predicted SPL targets of miR156f, 12 of 14 LRR targets of regulatory mechanism affecting TAS4c expression, which we speculate miR482-triggered TAS11 39D3(-), and in four out of the five validated might involve the TAS4c-39-D4(-) mismatches to target MYBA6/A7 impact- MYB targets of miR828 (Pantaleo et al. 2010) (Supplementary Table 2b; ing a functional interaction with the hypothesized auto-regulatory loop. Supplementary File S4, pp. 44-48). Taken together these results are consistent with the hypothesis that high-fluence UV-B can induce Small RNAs differentially expressed during miRNA accumulations that consequently have an effect in decreasing berry development berry skin target gene expressions. To characterize the overall accumulation of miRNAs and siRNAs affected by berry ripening, sRNA libraries constructed from berries UV-B regulated miRNAs with different fluence- (-3, 0, +3 and +6 WAV) harvested from greenhouse and fields were rate responses tested for differential expression across developmental time using the Based on the apparent coordination for miRNA expressions by high- vs. likelihood ratio test. By assuming (null model) there is no difference in low-fluence UV-B treatments, the high- and low-fluence irradiated miRNA expression profiles between different berry development stages, greenhouse berry samples were tested against each other (high/low we identified 48 miRNAs to be significantly differentially expressed at ratio across developmental time, including across (-) UV controls). one or more developmental time points across environmental treat- The field UV-B irradiated berries were not included in the test since ments (Supplementary Table 3d). As expected for phenotypic changes the environmental variables were likely confounding vs. within-greenhouse associated with veraison, we observed profound dynamics in miRNA comparisons. Supplementary Table 3c lists 18 miRNAs with significant abundances. 21 miRNAs were up-regulated while 27 displayed a sig- differential expression by high-fluence UV-B treatment per se, of which nificant negative log fold-change during development. Linear regression 11 were up-regulated and seven were down-regulated. Three novel (Pearson coefficient; Supplementary Table 3d) as a metric to correlate miRNAs (called by ShortStack as valid MIRNAs or missing only evi- time (three equally spaced time points before berry maturity at harvest) dence for the star species) were identified to be differentially expressed with normalized miRNA/siRNA abundances demonstrated a strong in response to high-fluence UV-B. We found multiple candidate family linear correlation across the developmental time points with replicates members of vvi-MIR477 of which miRBase has only two annotated: (Supplementary Table 3d). This strong linear relationship held for nearly miR477a and miR477b. The five others that respond differentially to every MIRNA and PHASI locus even when fold changes were small and UV-B high vs. low fluence are listed in Supplementary Table 3c and are non-significant. This evidence demonstrates that the experimental pa- presumed to be encompassed by the nine vvi-miRC477c,i-p candidate rameters were highly reproducible and the results warrant inference of family members previously described by Paim Pinto et al. (2016). Sim- miRNA/siRNA functions toward validated targets as shown above. We ilar to the UV-B response results across greenhouse and field berry performed co-expression analysis of our differentially expressed (DE) samples (Supplementary Table 3b), vvi-MIR3627 was highly miRNAs (Supplementary Table 3d) with fruit developmental RNA-Seq up-regulated under high-fluence UV-B (Supplementary Table 3c). datasets from Massonnet et al. (2017) that reported significantly DE vvi-MIR3627 is an atypical MIRNA locus in that it produces phased mRNAs at four comparable time points in five red-skinned varieties. 778 | S. Sunitha et al. n Table 3 Co-expression analysis of high-fluence UV-B induction of miRNAs/phasi-RNAs abundances during berry development which are inversely correlated with predicted target mRNA expressions in the same skin tissues in response to UV-C pre-veraison and UV-B five weeks post-veraison from independent experiments (Suzuki et al. 2015; Carbonell-Bejerano et al. 2014) UV-B Fold Change& (bold= up; italics = down) Inverse correlation of mRNA target expression vs. High-Fluence Greenhouse Low-Fluence Greenhouse High-Fluence Field miRNA/siRNA? + + + + + + LFC, -3 Verai- 3 6 -3 Verai- 3 6 -3 Verai- 3 6 Validated Targets (n; FC, FC, FC P- miRNA/siRNA UV-B WAV son WAV WAV WAV son WAV WAV WAV son WAV WAV annotation) UV-C$ UV-B† val^ miR156f 0.68 1.332 2.755 3.266 5.520 0.644 10.25 1.863 0.567 1.440 2.149 1.144 1.430 11; SPB 0.75 0.89 0.02 miR530b 20.77+ n.d. 2.556 0.488 1.183 0.258 0.853 0.200 0.065 n.d. 5.908 5.853 6.316 2; Plus3, DYW 0.80 0.86 0.09 miR477c 20.32+ 0.010 3.253 3.129 2.772 0.017 0.104 0.116 0.003 3.082 1.177 3.039 0.374 2; GRAS-domain n.d. 0.84 0.51 TAS4a, miR828 target 0.32 0.561 3.540 5.119 1.249 0.102 2.155 0.554 0.357 2.772 2.602 1.961 1.192 2; MYBA6/A7 0.63 0.97 0.30 TAS4b, miR828 target 1.10 0.089 1.704 2.115 60.16 1.006 2.713 2.562 0.572 4.956 7.022 6.213 2.953 TAS4c, miR828 target 0.14 5.381 4.825 1.712 0.955 0.180 1.348 1.342 0.513 0.318 1.777 1.953 0.750 miR828 target MYB, phasi- 20.29+ 0.218 1.425 1.023 1.725 0.130 2.948 1.260 0.377 1.151 1.367 0.569 1.316 5; MYBs 0.70 0.87 0.04 RNA producing miR482 0.07 0.340 0.965 1.188 2.936 0.393 2.583 2.840 0.624 1.450 1.678 0.774 0.855 2;TAS11,slyTAS5-Like 0.80 0.96 0.09 TAS11-D3(-) miR482 target 0.17 0.287 1.467 1.290 5.191 0.515 5.015 1.697 0.419 1.673 1.167 0.733 1.279 14; LRRs 0.84 0.86 0.02 miR403a-e average 20.19 0.504 1.206 0.205 1.212 2.764 2.238 0.347 0.382 1.566 0.987 0.678 2.044 1; AGO2 1.1 1.07 0.17 Overall 38 targets, two experiments 5.3E-6 Volume 9 March 2019 | & Normalized to 20M reads (+1 if zero reads). Denominator is corresponding normalized -UV-B sample (unity). n.d.: not detected. + sign negative overall because low-fluence UV-B effect dominant, or low read counts independently filtered out by PhaseTank.  See Supplementary Table 2a. $ Data for 1 hr FC by UV-C treatment of pre-veraison berries from Suzuki et al. (2015). If called significantly different expression in response to UV-C, in bold. ^ Binomial distribution probability for FCs of al target mRNA responses to UV-C and UV-B inverse to miRNA/siRNA (up) under high-fluence UV-B. † Data for FC by UV-B treatment on 26 Brix (ripe) berries irradiated for five weeks post-veraison from Carbonell-Bejerano et al. (2014). For miR403 target AGO2, includes 23Brix sample (3 independent tests for significance). UV-B Responsive sRNAs in Grapevine | 779 We identified 15 cases of miRNA/target mRNAs with statistically significant Initial prediction of UV-B-responsive miRNAs in Arabidopsis was (binomial test of chance) inverse correlations, supporting functional effects founded on computational analyses of UV-regulated co-expressed of miRNA changes during development (Supplementary Tables 2b, 3d). genes, which shared promoter motifs with candidate MIRNAs, coupled miR530ab may target 39 UTR intronic regions of AGO1 orthologs, with predicted target expression that inversely correlated with inferred supported by GSTAr alignment data (Supplementary Table 2a). miRNA profiles (Zhou et al. 2007). UV-B responsive miRNAs were miR393ab, which degradome evidence shows triggers phasi-RNA later identified in poplar (Jia et al. 2009), wheat (Wang et al. 2013), production and phasi-RNA auto- and trans cleavages (Si-Ammour maize (Casati 2013), and Brassica rapa in relation to UV-A (Zhou et al. et al. 2011) in AUXIN TRANSPORT INHIBITOR RESPONSE1/TIR1 2016). In maize, miR156 and miR529 are decreased while their several co-receptor mRNAs (Supplementary Tables 2a, 4; Supplementary targets Squamosa Promoter-Binding-like (SPL transcripts) are increased Files S5, S4 p. 20), were also coordinately up-regulated during berry after 8 hr of UV-B exposure of maize leaves (Casati 2013). Wang et al. development (Supplementary Table 3d). (2013) also reported down-regulation of miR156 by UV-B in wheat. As shown previously, TAS4a, -b and -c siRNAs (triggered by We also observed a few inverse correlations between vvi-miR156 abun- miR828), that in turn trigger phasing of MYBA6 and MYBA7 siRNA dances and SPL target expressions in our experiments which were production via TAS4ab-39-D4(-), showed differential expressions consistent with published datasets for UV-C and UV-B responses during berry development. TAS4a siRNAs expression was maximal (Table 3, Supplementary Table 2b), but in contrast to maize and at -3 WAV, and decreased at veraison, +3 WAV and +6 WAV wheat we found vvi-miR156f and vvi-miR535c to be up-regulated in (Supplementary Table 3d). TAS4b siRNAs displayed a lower expression grapevine in response to UV-B. It is possible that the low-fluence than TAS4a, and also decreased along developmental stages. Interest- down-regulation effect on numerous miRNAs (including miR156f; ingly, TAS4c reads increased at every developmental stage when com- Table 3) could account for the observed differences between grape pared to the previous stage (Supplementary Table 3d), consistent with and maize or wheat, where a coordinated antagonist response pathway the observation that TAS4c and miR403 were regulated differently than for low-fluence UV-B may be involved. Previous work on grapevine TAS4ab and most miRNAs (Supplementary Table 3c), which decrease leaves showed high-fluence rate UV-B specifically modulates pathways during berry development (Supplementary Table 3d). and processes associated with oxidative and biotic stress, cell cycle progression, and protein degradation, whereas low-fluence rate UV-B DISCUSSION promotes photomorphogenic responses and regulates the manifesta- We have endeavored to characterize the miRNA and phasi-RNA space tion of auxin, ABA, and cell wall modification processes (Pontin et al. in response to UV-B light by target degradome validations, statistical 2010). Consistent with this notion is the differential UV-B fluence inference of sRNA dynamics, and direct qRT-PCR measurement and effects on VviTAS3 phasi-RNAs and miR828 abundances (up in meta-analysis of miRNA and target mRNA abundances in published high-, down in low-fluence), and with prior work showing TAS4 is datasets from the same organs and tissues. Shortstack ver3, by virtue of negatively regulated by ABA/sugar crosstalk signaling (Luo et al. its multimapper functionality (Johnson et al. 2016), was our method of 2012). The higher expression level and lack of UV-B induction of choice for quantifying locus-specific miRNA precursor (i.e., cluster) miR828 or pre-MIR828 in in vitro-grown plantlets (Figure 2A; data abundance as a proxy for mature miRNA and phasi-RNA levels. This not shown) in contra-distinction to berries is consistent with a possible approach presumes that when a sRNA cluster is called as differentially negative effect of increased ABA in leaves (Loveys 1984). expressed, the mature miRNA (or most abundant phasi-RNA) is the species changing in response to a treatment or condition. Even if the A bimodal pathway operates under low-fluence UV-B subject sRNA ostensibly undergoes dynamic fluctuations in abundance, regulating photomorphogenesis it cannot function as a negative effector of target gene activity without UV-B signaling mechanisms affecting grape berry polyphenolics (flavor being loaded in a functional AGO-containing RISC, which along with enhancers and fruit astringency components) have been previously base pairing to target mRNAs, finally controls miRNA/tasiRNA stabil- characterized (Loyola et al. 2016; Martínez-Lüscher et al. 2013), but ity, activity, and subcellular retention (Li et al. 2013; Pitchiaya et al. there is no description of grapevine UV-B responsive miRNAs and 2017; Speth et al. 2013). It is remarkable that miR828, a 22 nt species their significance for regulation of polyphenolic biosynthesis, stress with a very low estimated abundance (the sequenced star species = responses, or berry development. Our analysis reveals a dynamic reg- 2.6 rpm from 411 raw reads, meaning miR828 was sequenced only a ulation of certain miRNAs following UV-B-induction and pervasive few times across all libraries 0.01 rpm), manifested the highest in vivo changes in miRNA profiles during fruit development. We have formu- slicing activities based on phasi-RNA abundances (15,000 rpm) and lated a model (Figure 4) to frame high-fluence miRNA dynamic re- phasing scores from TAS4abc and most of the miR828 target MYBs. sponses to UV-B in the context of post-transcriptional gene silencing Single nucleotide polymorphisms of sRNAs across mapping datasets processes impacting oxidative and nutrient (sulfur and phosphate, as- could also be a confounding factor in interpreting our differentially sociated with anthocyanin accumulation) stress adaptation during expressed candidate miRNAs. Caution is warranted when interpreting grape berry development. We suggest that these high-fluence UV-B phasi-RNA abundances (e.g., TAS4a,b,c) where there are many species response processes may be different between dicots vs. monocots, like generated, some of which can accumulate to high levels yet only one the case for MIR828/TAS4 regulation of anthocyanin biosynthesis (e.g., TAS4 39-D4[-]; Rock 2013) is claimed as functional. The func- (Rock 2013) based on three grounds: i) observed differences between tional siRNA and miRNA may not be the major species in clusters or our results for deeply conserved and functionally significant miR156 the one that changes significantly in response to stimulus, confounding changes (Figure 3, Table 3) in contradistinction of results in maize and interpretations. For TAS4a, the major species called by ShortStack is wheat (Casati 2013; Wang et al. 2013). ii) The proposed coordinate documented as 39D1(+) (Supplementary Table 3d, column “I”). How- antagonist response pathway for UV-regulated miRNAs revealed by ever, Supplementary Table 3d and Supplementary File S4 (p. 38) pro- the low-fluence UV-B experiment (Figure 2, Table 3) affecting predom- vides direct evidence supporting the assumption that the claimed inantly dicot-specific miRNAs (Supplementary Table 3a-c). iii) Differ- differentially expressed mature miRNAs and phasi-RNAs (e.g., TAS4b, ences in overrepresented gene ontology categories between eudicots and c 39D4[-]) are the major species changing their abundances per se. monocots suggest the evolution of oxygen and radical detoxification 780 | S. Sunitha et al. Figure 4 A) Model of UV-B effects on grape berry stress responses and development regulated by miRNAs. Defense, DNA repair and accumula- tion of anthocyanin are the major responses regulated by UV-B respon- sive miRNAs. Increase in miR168/ miR530 facilitate post-transcriptional gene silencing activity. The increase in PTGS activity would in turn facil- itate targeting of LRR genes by miR482 and regulate the defense genes. miR477 family members tar- get DNA Pol III involved in base excision repair (BER) and facilitates photo-reactivation and nucleotide excision repair (NER) to remove UV-B induced DNA lesions. Upre- gulation of miR3627/4376/5225 by UV-B targets ER-specific Calcium ATPase10 and facilitate increase in cytosolic Ca2+ levels and thereby increased anthocyanin accumula- tion. miRNAs shown to be regulated by sulfur (miR395) and phosphorous starvation (miR399; miR827) are also responsive to UV-B radiation and facilitate anthocyanin accumulation. miR156/miR535 target SPL genes and thereby increase the accumula- tion of MYB-bHLH-WD40 TFs and anthocyanin as elucidated in Arabi- dopsis. Increase in MYB-bHLH-WD40 TFs would trigger the conserved auto- regulatory loop involving miR828/ TAS4 to regulate MYBA6/A7 levels and thereby anthocyanin levels. B) Schematic representation of the auto-regulatory feedback loop of miR828/TAS4 exerted by MYBA genes from chromosome 14. (i.e., UV associated) and RNA silencing processes mark key divergence Calcium-transporting ATPase10 pre-mRNA (Supplementary Table 2a; points between dicots and monocots (Lee et al. 2011). Our results suggest Supplementary File S4, pp. 56-57), we propose these miRNA family no overlap between the miRNAs regulating photomorphogenic and de- members as modulators of free cytosolic Ca2+ levels and CHS expres- fense pathways in response to low-fluence UV-B, in accordance with those sion (Figure 4A). of Yoon et al. (2016) and Pontin et al. (2010), who showed low levels of Considering that nutrient availability impacts on miRNA networks UV-B irradiation result in photomorphogenic and auxin responses rather regulating anthocyanin metabolism, vvi-MIR399i, which targets PHO2 than induction of defense- and abiotic oxidative stress-related genes. (a ubiquitin-conjugating E2 enzyme regulating the stability of phos- phate transporters) is up-regulated by high-fluence UV-B. PHO2 re- Polyphenolic pathway and cellular transcription are pression by miR399 in response to UV would affect negatively Pi uptake targets of UV-B responsive miRNAs (Lin et al. 2008) (Figure 4A). Consistent with this model, phosphate Frohnmeyer et al. (1999) demonstrated that UV-B triggers an increase deficiency increases anthocyanin content in grapevine cell cultures in free cytosolic Ca2+ levels associated with UV-B induction of CHS, (Yamakawa et al. 1983). and Loyola et al. (2016) demonstrated UV-B induction of CHS3 in vvi-MIR395 is a 59 RLM-RACE-validated slicer of ATP Sulfurylase irradiated grapevine plantlets. Calcium transporting ATPase10 is 1supported by GSTAr alignment prediction, which showed a strong known to be involved in transporting free cytosolic Ca2+ to the endo- and significant inverse correlation between miRNA and target mRNA plasmic reticulum (Geisler et al. 2000). The evolutionarily related abundances in response to cold stress in grapevine (Sun et al. 2015). miR5225, miR3627, and miR4376 (offset by 4 nt) were up-regulated ATP Sulfurylase, along with validated miR395 target AST68, a low upon UV-B irradiation. As they all target the 59 UTR intron2 region of affinity sulfate transporter (Supplementary Table 2a, File S4 p. 22), Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 781 functions to increase sulfur transfer to aerial parts under sulfur starva- Novel TAS11 and vvi-slyTAS5-like loci are triggered by miR482 tion (Kawashima et al. 2011). MIR395 up-regulation in response to up-regulation under high-fluence UV, generating several phasi-RNAs drought stress in rice (Zhou et al. 2010) and salinity stress in Zea mays from multiple registers. These in turn target numerous LRR genes, (Ding et al. 2009) is consistent with our evidence for a role in abiotic which themselves undergo UV-B dependent siRNA amplification stress responses including UV-B. Sulfur starvation increases anthocy- (Supplementary Table 3a-c) and subsequent functional silencing (Table anin levels in rice (Lunde et al. 2008) and in grapevine sulfur- depen- 3, Figure 3D; Supplementary Table 2b; Supplementary File S4, p. 37). dent oxidative response effectors GLUTATHIONE S TRANSFERASE The repression of defense genes by miRNAs until the plant is exposed (GST) and glutaredoxin are induced during veraison (Terrier et al. to severe stress is a hypothesized adaptive mechanism to keep the de- 2005). Different GSTs have been recently related to flavonoid transport fense response under post-transcriptional control to conserve energy by (Pérez-Díaz et al. 2016), with VviGST1/4 having major roles in antho- preventing LRR protein synthesis (Shivaprasad et al. 2012). A domain cyanin transport (Conn et al. 2008; Gomez et al. 2011). of the nuclear Armadillo-type found in Cap Binding Protein Large UV-B induced the accumulation of vvi-MIR535c and vvi-MIR156f, Subunit/CBP80 and LRR variants is targeted by D6(+) phase of vvi- the latter of which we show and impute (Table 3) to coordinately down slyTAS5-like, based on degradome evidence. Mutations in cap binding regulate SPL transcripts. Up-regulation of miR535 also may contribute protein20 and cbp80 are ABA hypersensitive, drought tolerant, and to observed down-regulation of SPLs, although a bulge at nt 6 of have abnormal miRNA biogenesis (Gregory et al. 2008; Papp et al. the seed region and lower degradome p-values qualify this assertion 2004). It is plausible that targeting of a cap binding complex by elevated (Supplementary Table 2a; 2b). Several groups have reported SPL miR482/TAS phasi-RNA expression may be an adaptive mechanism transcripts in grapevine targeted by both miR156 and miR535 for UV-B response/tolerance and a crosstalk mechanism linking UV-B (Carra et al. 2009; Han et al. 2016; Pantaleo et al. 2010). SPL transcript fluence dynamics of miRNAs (Table 3) to ABA antagonism of MIR828/ reduction increases anthocyanin levels in Arabidopsis (Gou et al. 2011) TAS4 feedback homeostasis (Luo et al. 2012). and our functional co-expression analysis supports that UV-B regula- tion of miR156 is an important component of a regulatory network for Oxidative stress by high-fluence UV-B combated anti-oxidant and UV-protective polyphenolic biosynthesis (Figure 4A). by miRNAs? The preponderance of transcriptional regulators and TAS/PHASI tar- The 18 differentially expressed miRNAs identified as changing under gets of UV-B regulated miRNAs in grapevine and across species sup- high vs. low-fluence per se were more than found when UV-B radiation ports a model integrating nodes of conserved regulatory networks across all environments (including field trials) was tested as a single important for integrating abiotic stress and developmental programs. parameter (Supplementary Table 3b). This is possibly because of the noise associated with different (i.e., field) environments. However, Phenylpropanoid pathway and defense genes are Pontin et al. (2010) observed the same degree of increased effects of targets of UV-B responsive tasiRNAs high- vs. low-fluence UV-B changes in a grapevine leaf transcriptome Recent papers report a role of miR858 in Arabidopsis related to comparison. One upshot from the alternative perspectives afforded by immunity upon nematode and fungal pathogen infections by negative the controlled greenhouse experiments was the finding that certain regulation of flavonoid-specific MYB regulators (Camargo-Ramírez miRNAs/siRNAs (e.g., miR403, miR530, TAS4c) are regulated differ- et al. 2018; Piya et al. 2017). Other researchers have invoked the ently and possibly coordinately by UV and development. Six miRNAs ’two-hit’ model to explain observed phasi-RNA production of (five of them novel; Supplementary File S6) belonging to the MIR477 miR828- and miR858-dual targeted MYBs involved in proanthocyani- family were up-regulated by UV-B and had star species that were much din and anthocyanin biosynthesis, based on weak predicted binding more abundant. vvi-miR477b is predicted to uniquely target RepC/ affinity of miR858 in a conserved domain 12 nt upstream of miR828 DNA Pol III epsilon subunit t (Supplementary Table 2b) associated binding sites (Guan et al. 2014; Xia et al. 2012; Zhu et al. 2012). Several with nucleotide excision/mismatch repair pathways. Interestingly, grapevine MYBs with a conserved miR858-miR828 dual binding do- MIR477, MIR482, and MIR3627 are conserved across woody fruit trees main as described in peach, apple, cotton, and Arabidopsis (Guan et al. (Solofoharivelo et al. 2014) including grapevine. We speculate that the 2014; Xia et al. 2012; Zhu et al. 2012) showed compelling evidence of prevalence of vvi-MIR477 family members induced by high-fluence slicing by miR828 that spawns phasi-RNAs in register. However, these UV-B may indicate a regulatory role in countering oxidative stress. targets only showed marginal evidence of miR858 slicing that correlates UV-B radiation is known to induce cyclobutane pyrimidine dimers with high (4.5- 6.5) Allen scores (Supplementary File S4, pp. 44, 47, 49, that kill cells by blocking DNA replication (Ries et al. 2000). Photo- 50). Although it is plausible that dual auto-regulatory loops for miR828: repair by photolyase and nucleotide excision repair (NER) are the MYBA6/7 (Rock 2013) and miR858:MYB82 (Piya et al. 2017) operate major pathways in plants for repairing UV-B-induced DNA damage coordinately in grapevine to regulate anthocyanin and flavonol biosyn- (Tuteja et al. 2001), while base excision repair (BER) involving RepC/ thesis, respectively like in Arabidopsis, our functional analysis supports DNAPol III epsilon is involved in repairing oxidized or hydrated bases miR828 being the dominant effector module. It is noteworthy that for and single-stranded breaks caused by ionizing radiation (Gill et al. MYBA6/7 slicing by 21 nt TAS4ab-39D4(-) and production of abun- 2015). We speculate that the observed up-regulation of vvi-MIR477 dant phased registers downstream, there is no obvious trigger mecha- family members may represent an adaptation to reduce BER mecha- nism like described for the ’two-hit’ and 221 hit miRNA models for nisms and facilitate photoreactivation and NER, thus removing UV-B phasi-RNA production (Axtell et al. 2006). induced DNA lesions (Figure 4A). Primary TAS4 transcripts, miR828 and TAS4abc 39D4(-) tasiR- Liang et al. (2013) demonstrated that in addition to miR168, AGO1 NAs triggered by miR828 are up-regulated by high-fluence UV-B in is a target of miR530 in papaya and a predicted target of miR530 in greenhouse and field, which is in line with the function of a con- Salvia miltiorrhiza (Shao and Lu 2013). We predict that two grape served auto-regulatory feedback loop proposed by Hsieh et al. AGO1 orthologs could also be targets of vvi-miR530-isomiR. Harvey (2009) and Luo et al. (2012). Increases in VvMYBA6/7 induce et al. (2011) demonstrated AGO2 acting as a second layer of control in TAS4 (and MIR828) transcription and thereby trigger a feedback antiviral defense that is activated when AGO1 is suppressed. When repression of MYB genes targeted by miR828 and TAS4 (Figure 4B). AGO1 is active it represses AGO2 via miR403 slicing. Interestingly, 782 | S. Sunitha et al. AGO2 induction by Pseudomonas syringae pv. tomato bacterial infec- MYBA7 phasi-RNAs triggered by TAS4ab-39D4(-) corresponding to tion without suppression by its negative effector miR403 was observed the sixth phase are the most abundant (Rock 2013). Similar to earlier in the ago1-27 mutant, suggesting a post-transcriptional regulation of work we find here that the sixth phase of MYBA6/7 phasi-RNAs are the AGO2 by bacteria, which is mechanistically different from the induc- major species where MYBA7 phasi-RNA D6(+) is significantly down- tion triggered by viruses (Zhang et al. 2011). Our assay of AGO2 tran- regulated during berry development, in stark contrast to TAS4c39- script abundances in samples from low- and high-fluence UV-B D4(-) significant up-regulation. Further evidence consistent with the greenhouse-treated berries showed inverse trends to documented functioning of this feedback loop comes from RNA-Seq transcriptome miR403 dynamics, but results were not statistically significant. Based data for several red-skinned berries (Table 2; Massonnet et al. 2017). on the observed coordinate down-regulation of all vvi-miR403 family The significant up-regulation of MYBA6 mRNA in the pre-veraison-to- members and imputed up-regulation (albeit not statistically significant; veraison transition is inversely correlated with the observed down reg- P = 0.17) of validated target AGO2 in response to UV-B (Table 3; ulation of TAS4ab and downstream MYBA6/7 phasi-RNAs, suggesting Supplementary Table 3c; Supplementary File S4, p. 34), we speculate a positive feed-forward activity for the TAS4ab 39D4(-) tasi-RNAs on a third component to the model that might operate in UV-B-radiated slicing MYBA6/7 and MYBA5/MYB113-like. The subsequent wholesale grapes in which AGO2 could play a role in combating UV oxidative inversion of the MYBA6 target mRNA:TAS4a and TAS4b primary stress without inactivation of AGO1 (Figure 4). transcripts and tasi-RNA relationships (i.e., both significantly down) at berry maturity is parsimonious with the predicted silencing functions Auto-regulatory loop involving miR828-TAS4-MYBA6/ of TAS4ab-39D4(-) on MYBA6/7 late in berry development. Matus A7/A5-MYB113-Like is conserved in berry development et al. (2017) showed high level expression and significant UV-B in- Our specific rationale for studying miRNAs/phasi-RNAs is their in- duction of MYBA6/7 before veraison and Table 2 shows complete volvement in phenylpropanoid accumulation in relation to pigmenta- down-regulation of MYBA7 (but not MYBA6, which was delayed per tion, one of the most important traits for grape berry ripening and Matus et al. 2017) during and after veraison. subject to nutrient and phytohormone effects (Vanhaelewyn et al. 2016). Our results of differential UV-B fluence effects on miR828 (Figure UV-B radiation reduces berry weight and increases anthocyanin con- 2B-D) and pre-MIR828 (Supplementary Fig. S2B) and developmental tent (Berli et al. 2010; Berli et al. 2011). In addition to phosphate de- stage-specific TAS4a-c expressions (Supplementary Table 3b, d) suggest ficiency (Yamakawa et al. 1983), rises in ABA concentration and that the auto-regulatory circuit may function to fine-tune expression of biosynthesis/response effectors such as MYBPA, ATHB-12, ABI5, MYBA6/7 in response to UV-B and during berry skin maturation. Our Nine-cis-epoxycarotenoid dioxygenase4, and PAP1/PAP2 increase an- results reveal complex activities of MIR828/TAS4 in response to envi- thocyanin levels (Hiratsuka et al. 2001; Koyama et al. 2010; Luo et al. ronmental and developmental queues to post-transcriptionally silence 2012). Interestingly, UV-B radiation does not affect berry ABA content effectors and underscore the importance of post-transcriptional con- whereas leaves exposed to UV-B increase ABA levels (Berli et al. 2010; trols in the observed delays by UV-B exposure on repression of MYBA6 Berli and Ruben 2013). The vegetative anthocyanin loci MYBA6, late in berry development. Understanding the significance of miRNA/ MYBA7, and presumably MYBA5/MYB113-like and their regulator phasi-RNA silencing of target mRNAs awaits additional data such as HY5 are up-regulated during early berry development when exposed rates of PolII transcription, splicing, polyadenylation, RNA-templated to UV-B and despite expression declines with berry ripening, this be- antisense transcription, dicing, slicing, and exosome degradation of havior is delayed in response to UV (Supplementary Fig. S2F,G; Loyola target mRNAs. Our exploratory results and meta-analysis provide a et al. 2016; Matus et al. 2017). Pilati et al. (2017) reported up-regulation basis for systems approaches to better understand non-coding RNA of MYBA7 when berries are treated with ABA for 20 h but not when the functions in response to UV light and strategies for improving wine treatment was prolonged to 44 h. This observation is consistent with the grape production of anti-oxidant flavor enhancers. notion that MYBA7 is induced directly by ABA but does not have a role in berry ripening per se. vvi-MIR828 and TAS4 are validated in this work ACKNOWLEDGMENTS to regulators of expression of MYBA6, MYBA7, and MYB5/MYB113- The authors thank Glenn Hicks and John Weger at the Institute like. The auto-regulatory feedback loop is shown to function in grapevine for Integrative Genome Biology UC Riverside for Illumina in a similar fashion to Arabidopsis (Figure 4B; Hsieh et al. 2009; Luo et al. Nextseq500 sequencing, Rao Kottapalli at the Texas Tech Center 2012), whereby developmental induction (early in berry development) of for Biotechnology and Genomics for MiSeq sequencing, and the MYBA6 and MYBA7 correlates with observed elevated miR828 and TTU High Performance Computer Center for support in use of high-fluence UV-B inductions at the green berry and veraison stages. the Quanah supercluster. In addition, MYBA gene expression is concordant with temporal accu- mulations of TAS4abc primary transcripts and derivative siRNAs. LITERATURE CITED Interestingly TAS4c 39D4(-), which in our analysis of cv. ‘Cabernet Addo-Quaye, C., W. Miller, and M. J. Axtell, 2009 CleaveLand: a pipeline Sauvignon’ had two different single nucleotide polymorphisms com- for using degradome data to find cleaved small RNA targets. Bioinfor- pared to TAS4ab and the reference genome sequence for TAS4c, dis- matics 25: 130–131. https://doi.org/10.1093/bioinformatics/btn604 played a trend of increasing reads abundance with advancement of Alabi, O. J., Y. Zheng, G. Jagadeeswaran, R. Sunkar, and R. Naidu, berry ripening. This result is intriguing because earlier studies on Vitis 2012 High-throughput sequence analysis of small RNAs in grapevine (Vitis vinifera L.) affected by grapevine leafroll disease. Mol. Plant Pathol. amurensis and Vitis vinifera (cv. ‘Pinot Noir’) grape datasets could not 13: 1060–1076. https://doi.org/10.1111/j.1364-3703.2012.00815.x detect appreciable TAS4c phasi-RNA expression (Rock 2013), and here Axtell, M. J., C. Jan, R. Rajagopalan, and D. P. Bartel, 2006 A two-hit trigger we show (Table 2) that green-skinned cultivars do not express TAS4c. It for siRNA biogenesis in plants. Cell 127: 565–577. https://doi.org/ will be critical to ascertain if the trait of TAS4c-39D4(-) functional 10.1016/j.cell.2006.09.032 variant expression late in berry development is unique to cv. ’Cabernet Belli Kullan, J., D. Lopes Paim Pinto, E. Bertolini, M. Fasoli, S. Zenoni et al., Sauvignon’, or if it constitutes a key regulator of a branch pathway in 2015 miRVine: a microRNA expression atlas of grapevine based on red cultivars for polyphenolic biosynthesis, or a specificity determinant small RNA sequencing. BMC Genomics 16: 393. https://doi.org/10.1186/ for the MYBA6/A7/A5-MYB113-like autoregulatory loop. MYBA6 and s12864-015-1610-5 Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 783 Berli, F. J., M. Fanzone, P. Piccoli, and R. Bottini, 2011 Solar UV-B and Fátyol, K., M. Ludman, and J. Burgyán, 2016 Functional dissection of a ABA are involved in phenol metabolism of Vitis vinifera L. increasing plant Argonaute. Nucleic Acids Res. 44: 1384–1397. https://doi.org/ biosynthesis of berry skin polyphenols. J. Agric. Food Chem. 59: 10.1093/nar/gkv1371 4874–4884. https://doi.org/10.1021/jf200040z Fei, Q., R. Xia, and B. C. Meyers, 2013 Phased, secondary, small interfering Berli, F. J., D. Moreno, P. Piccoli, L. Hespanhol-Viana, M. F. Silva et al., RNAs in posttranscriptional regulatory networks. Plant Cell 25: 2010 Abscisic acid is involved in the response of grape (Vitis vinifera L.) cv. 2400–2415. https://doi.org/10.1105/tpc.113.114652 Malbec leaf tissues to ultraviolet-B radiation by enhancing ultraviolet- Frohnmeyer, H., L. Loyall, M. R. Blatt, and A. Grabov, 1999 Millisecond absorbing compounds, antioxidant enzymes and membrane sterols. UV-B irradiation evokes prolonged elevation of cytosolic-free Ca2+ and Plant Cell Environ. 33: 1–10. stimulates gene expression in transgenic parsley cell cultures. Plant J. 20: Berli, F. J., and B. Ruben, 2013 UV-B and abscisic acid effects on grape 109–117. https://doi.org/10.1046/j.1365-313X.1999.00584.x berry maturation and quality. J. Berry Res. 3: 1–14. Frohnmeyer, H., and D. Staiger, 2003 Ultraviolet-B radiation-mediated Brosché, M., M. A. Schuler, I. Kalbina, L. Connor, and A. Strid, 2002 Gene responses in plants. Balancing damage and protection. Plant Physiol. 133: regulation by low level UV-B radiation: identification by DNA array 1420–1428. https://doi.org/10.1104/pp.103.030049 analysis. Photochem. Photobiol. Sci. 1: 656–664. https://doi.org/10.1039/ Geisler, M., N. Frangne, E. Gomes, E. Martinoia, and M. G. Palmgren, B202659G 2000 The ACA4 gene of Arabidopsis encodes a vacuolar membrane Camargo-Ramírez, R., B. Val-Torregrosa, and B. San Segundo, calcium pump that improves salt tolerance in yeast. Plant Physiol. 124: 2018 miR858-mediated regulation of flavonoid-specific MYB 1814–1827. https://doi.org/10.1104/pp.124.4.1814 transcription factor genes controls resistance to pathogen infection Gill, S. S., N. A. Anjum, R. Gill, M. Jha, and N. Tuteja, 2015 DNA damage in Arabidopsis. Plant Cell Physiol. 59: 190–204. https://doi.org/ and repair in plants under ultraviolet and ionizing radiations. Scientif. 10.1093/pcp/pcx175 World J. 2015: 250158. https://doi.org/10.1155/2015/250158 Carbonell-Bejerano, P., M. P. Diago, J. Martínez-Abaigar, J. M. Martínez- Gomez, C., G. Conejero, L. Torregrosa, V. Cheynier, N. Terrier et al., Zapater, J. Tardáguila et al., 2014 Solar ultraviolet radiation is necessary 2011 In vivo grapevine anthocyanin transport involves vesicle-medi- to enhance grapevine fruit ripening transcriptional and phenolic re- ated trafficking and the contribution of anthoMATE transporters sponses. BMC Plant Biol. 14: 183. https://doi.org/10.1186/1471-2229- and GST. Plant J. 67: 960–970. https://doi.org/10.1111/j.1365- 14-183 313X.2011.04648.x Carra, A., E. Mica, G. Gambino, M. Pindo, C. Moser et al., 2009 Cloning Gou, J. Y., F. F. Felippes, C. J. Liu, D. Weigel, and J. W. Wang, and characterization of small non-coding RNAs from grape. Plant J. 59: 2011 Negative regulation of anthocyanin biosynthesis in Arabidopsis by 750–763. https://doi.org/10.1111/j.1365-313X.2009.03906.x a miR156-targeted SPL transcription factor. Plant Cell 23: 1512–1522. Casati, P., 2013 Analysis of UV-B regulated miRNAs and their targets in https://doi.org/10.1105/tpc.111.084525 maize leaves. Plant Sig. Behav. 10: e26758. Gregory, B. D., R. C. O’Malley, R. Lister, M. A. Urich, J. Tonti-Filippini et al., Cavallini, E., J. T. Matus, L. Finezzo, S. Zenoni, R. Loyola et al., 2015 The 2008 A link between RNA metabolism and silencing affecting Arabi- phenylpropanoid pathway is controlled at different branches by a set of dopsis development. Dev. Cell 14: 854–866. https://doi.org/10.1016/j. R2R3-MYB C2 repressors in grapevine. Plant Physiol. 167: 1448–1470. devcel.2008.04.005 https://doi.org/10.1104/pp.114.256172 Guan, X., M. Pang, G. Nah, X. Shi, W. Ye et al., 2014 miR828 and miR858 Cloix, C., E. Kaiserli, M. Heilmann, K. J. Baxter, B. A. Brown et al., regulate homoeologous MYB2 gene functions in Arabidopsis trichome 2012 C-terminal region of the UV-B photoreceptor UVR8 initiates and cotton fibre development. Nat. Commun. 5: 3050. https://doi.org/ signaling through interaction with the COP1 protein. Proc. Natl. Acad. 10.1038/ncomms4050 Sci. USA 109: 16366–16370. https://doi.org/10.1073/pnas.1210898109 Guo, Q., X. Qu, and W. Jin, 2015 PhaseTank: genome-wide computational Conn, S., C. Curtin, A. Bézier, C. Franco, and W. Zhang, 2008 Purification, identification of phasiRNAs and their regulatory cascades. Bioinformatics molecular cloning, and characterization of glutathione S-transferases 31: 284–286. https://doi.org/10.1093/bioinformatics/btu628 (GSTs) from pigmented Vitis vinifera L. cell suspension cultures as pu- Gyula, P., I. Baksa, T. Tóth, I. Mohorianu, T. Dalmay et al., 2018 Ambient tative anthocyanin transport proteins. J. Exp. Bot. 59: 3621–3634. https:// temperature regulates the expression of a small set of sRNAs influencing doi.org/10.1093/jxb/ern217 plant development through NF-YA2 and YUC2. Plant Cell Environ. 41: Coruh, C., S. Shahid, and M. J. Axtell, 2014 Seeing the forest for the trees: 2404–2417. https://doi.org/10.1111/pce.13355 Annotating small RNA producing genes in plants. Curr. Opin. Plant Biol. Han, L., K. Weng, H. Ma, G. Xiang, Z. Li et al., 2016 Identification and 18: 87–95. https://doi.org/10.1016/j.pbi.2014.02.008 characterization of Erysiphe necator-responsive microRNAs in Chinese Czemmel, S., J. Höll, R. Loyola, P. Arce-Johnson, J. A. Alcalde et al., wild Vitis pseudoreticulata by high-throughput sequencing. Front. Plant 2017 Transcriptome-wide identification of novel UV-B- and light Sci. 7: 621. https://doi.org/10.3389/fpls.2016.00621 modulated flavonol pathway genes controlled by VviMYBF1. Front. Plant Harvey, J. J., M. G. Lewsey, K. Patel, J. Westwood, S. Heimstadt et al., Sci. 8: 1084. https://doi.org/10.3389/fpls.2017.01084 2011 An antiviral defense role of AGO2 in plants. PLoS One 6: e14639. Dai, X., Z. Zhuang, and P. X. Zhao, 2018 psRNATarget: a plant small RNA https://doi.org/10.1371/journal.pone.0014639 target analysis server (2017 release). Nucleic Acids Res. 46: W49–W54. Heijde, M., and R. Ulm, 2012 UV-B photoreceptor-mediated signalling in https://doi.org/10.1093/nar/gky316 plants. Trends Plant Sci. 17: 230–237. https://doi.org/10.1016/j. Dauelsberg, P., J. T. Matus, M. J. Poupin, A. Leiva-Ampuero, F. Godoy et al., tplants.2012.01.007 2011 Effect of pollination and fertilization on the expression of genes Hiratsuka, S., H. Onodera, Y. Kawai, T. Kubo, H. Itoh et al., 2001 ABA and related to floral transition, hormone synthesis and berry development in sugar effects on anthocyanin formation in grape berry cultured in vitro. grapevine. J. Plant Physiol. 168: 1667–1674. https://doi.org/10.1016/j. Sci. Hortic. (Amsterdam) 90: 121–130. https://doi.org/10.1016/S0304- jplph.2011.03.006 4238(00)00264-8 Ding, D., L. Zhang, H. Wang, Z. Liu, Z. Zhang et al., 2009 Differential Howell, M. D., N. Fahlgren, E. J. Chapman, J. S. Cumbie, C. M. Sullivan et al., expression of miRNAs in response to salt stress in maize roots. Ann. Bot. 2007 Genome-wide analysis of the RNA-DEPENDENT RNA POLY- 103: 29–38. https://doi.org/10.1093/aob/mcn205 MERASE6/DICER-LIKE4 pathway in Arabidopsis reveals dependency on Dotto, M., and P. Casati, 2017 Developmental reprogramming by UV-B miRNA- and tasiRNA-directed targeting. Plant Cell 19: 926–942. https:// radiation in plants. Plant Sci. 264: 96–101. https://doi.org/10.1016/j. doi.org/10.1105/tpc.107.050062 plantsci.2017.09.006 Hsieh, L. C., S. I. Lin, A. C. Shih, J. W. Chen, W. Y. Lin et al., Fahlgren, N., and J. C. Carrington, 2010 miRNA target prediction in 2009 Uncovering small RNA-mediated responses to phosphate defi- plants. Methods Mol. Biol. 592: 51–57. https://doi.org/10.1007/978- ciency in Arabidopsis by deep sequencing. Plant Physiol. 151: 2120–2132. 1-60327-005-2_4 https://doi.org/10.1104/pp.109.147280 784 | S. Sunitha et al. Huang, X., X. Ouyang, P. Yang, O. S. Lau, L. Chen et al., 2013 Conversion Love, M. I., W. Huber, and S. Anders, 2014 Moderated estimation of fold from CUL4-based COP1-SPA E3 apparatus to UVR8–COP1-SPA com- change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15: plexes underlies a distinct biochemical function of COP1 under UV-B. 550. https://doi.org/10.1186/s13059-014-0550-8 Proc. Natl. Acad. Sci. USA 110: 16669–16674. https://doi.org/10.1073/ Loveys, B. R., 1984 Abscisic acid transport and metabolism in grapevine pnas.1316622110 (Vitis vinifera L.). New Phytol. 98: 575–582. https://doi.org/10.1111/ Jaillon, O., J. M. Aury, B. Noel, A. Policriti, C. Clepet et al., 2007 The j.1469-8137.1984.tb04150.x grapevine genome sequence suggests ancestral hexaploidization in major Loyola, R., D. Herrera, A. Mas, D. C. Wong, J. Höll et al., 2016 The pho- angiosperm phyla. Nature 449: 463–467. https://doi.org/10.1038/ tomorphogenic factors UV-B RECEPTOR 1, ELONGATED HYPO- nature06148 COTYL 5, and HY5 HOMOLOGUE are part of the UV-B signalling Jenkins, G. I., 2017 Photomorphogenic responses to ultraviolet-B light. pathway in grapevine and mediate flavonol accumulation in response to Plant Cell Environ. 40: 2544–2557. https://doi.org/10.1111/pce.12934 the environment. J. Exp. Bot. 67: 5429–5445. https://doi.org/10.1093/jxb/ Jia, X., L. Ren, Q. J. Chen, R. Li, and G. Tang, 2009 UV-B-responsive erw307 microRNAs in Populus tremula. J. Plant Physiol. 166: 2046–2057. https:// Lunde, C., A. Zygadlo, H. T. Simonsen, P. L. Nielsen, A. Blennow et al., doi.org/10.1016/j.jplph.2009.06.011 2008 Sulfur starvation in rice: the effect on photosynthesis, carbohy- Johnson, N. R., J. M. Yeoh, C. Coruh, and M. J. Axtell, 2016 Improved drate metabolism, and oxidative stress protective pathways. Physiol. placement of multi-mapping small RNAs. G3: Genes|Genomes|Genetics Plant. 134: 508–521. https://doi.org/10.1111/j.1399-3054.2008.01159.x (Bethesda) 6: 2103–2111. Luo, Q.-J., A. Mittal, F. Jia, and C. D. Rock, 2012 An autoregulatory feed- Jung, J. H., Y. H. Seo, P. J. Seo, J. L. Reyes, J. Yun et al., 2007 The back loop involving PAP1 and TAS4 in response to sugars in Arabidopsis. GIGANTEA-regulated microRNA172 mediates photoperiodic flower- Plant Mol. Biol. 80: 117–129. https://doi.org/10.1007/s11103-011-9778-9 ing independent of CONSTANS in Arabidopsis. Plant Cell 19: Luo, Q.-J., M. P. Samanta, F. Köksal, J. Janda, D. W. Galbraith et al., 2736–2748. https://doi.org/10.1105/tpc.107.054528 2009 Evidence for antisense transcription associated with microRNA Kalvari, I., J. Argasinska, N. Quinones-Olvera, E. P. Nawrocki, E. Rivas et al., target mRNAs in Arabidopsis. PLoS Genet. 5: e1000457. https://doi.org/ 2017 Rfam 13.0: shifting to a genome-centric resource for non-coding 10.1371/journal.pgen.1000457 RNA families. Nucleic Acids Res. 46: D335–D342. https://doi.org/ Mackerness, S. A. H., 2000 Plant responses to Ultraviolet-B (UV-B: 10.1093/nar/gkx1038 280–320 nm) stress: What are the key regulators? Plant Growth Regul. 37: Kawashima, C. G., C. A. Matthewman, S. Huang, B. R. Lee, N. Yoshimoto 27–39. https://doi.org/10.1023/A:1006314001430 et al., 2011 Interplay of SLIM1 and miR395 in the regulation of sulfate Martínez-Lüscher, J., F. Morales, S. Delrot, M. Sanchez-Diaz, E. Gomes et al., assimilation in Arabidopsis. Plant J. 66: 863–876. https://doi.org/10.1111/ 2013 Short- and long-term physiological responses of grapevine leaves j.1365-313X.2011.04547.x to UV-B radiation. Plant Sci. 213: 114–122. https://doi.org/10.1016/j. Koyama, K., K. Sadamatsu, and N. Goto-Yamamoto, 2010 Abscisic acid plantsci.2013.08.010 stimulated ripening and gene expression in berry skins of the Cabernet Massonnet, M., M. Fasoli, G. Battista Tornielli, M. Altieri, M. Sandri et al., Sauvignon grape. Funct. Integr. Genomics 10: 367–381. https://doi.org/ 2017 Ripening transcriptomic program in red and white grapevine 10.1007/s10142-009-0145-8 varieties correlates with berry skin anthocyanin accumulation. Plant Kozomara, A., M. Biraoanu, and S. Griffiths-Jones, 2018 miRBase: from Physiol. 174: 2376–2396. https://doi.org/10.1104/pp.17.00311 microRNA sequences to function. Nucleic Acids Res. 47: D155–D162. Matus, J. T., R. Loyola, A. Vega, A. Peña-Neira, E. Bordeu et al., 2009 Post- https://doi.org/10.1093/nar/gky1141 veraison sunlight exposure induces MYB-mediated transcriptional regu- Langmead, B., C. Trapnell, M. Pop, and S. L. Salzberg, 2009 Ultrafast lation of anthocyanin and flavonol synthesis in berry skins of Vitis and memory-efficient alignment of short DNA sequences to the human vinifera. J. Exp. Bot. 60: 853–867. https://doi.org/10.1093/jxb/ern336 genome. Genome Biol. 10: R25. https://doi.org/10.1186/gb-2009-10-3-r25 Matus, J. T., E. Cavallini, R. Loyola, J. Höll, L. Finezzo et al., 2017 A group Lee, E. K., A. Cibrian-Jaramillo, S.-O. Kolokotronis, M. S. Katari, of grapevine MYBA transcription factors located in chromosome A. Stamatakis et al., 2011 A functional phylogenomic view of the seed 14 control anthocyanin synthesis in vegetative organs with different plants. PLoS Genet. 7: e1002411. https://doi.org/10.1371/journal. specificities compared with the berry color locus. Plant J. 91: 220–236. pgen.1002411 https://doi.org/10.1111/tpj.13558 Lee, J., K. He, V. Stolc, H. Lee, P. Figueroa et al., 2007 Analysis of tran- Matus, J. T., 2016 Transcriptomic and metabolomic networks in the grape scription factor HY5 genomic binding sites revealed its hierarchical role berry illustrate that it takes more than flavonoids to fight against ultra- in light regulation of development. Plant Cell 19: 731–749. https://doi. violet radiation. Front. Plant Sci. 7: 1337. https://doi.org/10.3389/ org/10.1105/tpc.106.047688 fpls.2016.01337 Li, F., R. Orban, and B. Baker, 2012 SoMART: a web server for plant Mica, E., V. Piccolo, M. Delledonne, A. Ferrarini, M. Pezzotti et al., miRNA, tasiRNA and target gene analysis. Plant J. 70: 891–901. https:// 2010 High throughput approaches reveal splicing of primary micro- doi.org/10.1111/j.1365-313X.2012.04922.x RNA transcripts and tissue specific expression of mature microRNAs in Li, J., T. M. Ou-Lee, R. Raba, R. G. Amundson, and R. L. Last, Vitis vinifera. BMC Genomics 11: 109. https://doi.org/10.1186/1471- 1993 Arabidopsis flavonoid mutants are hypersensitive to UV-B irra- 2164-11-109 diation. Plant Cell 5: 171–179. https://doi.org/10.1105/tpc.5.2.171 Montgomery, T. A., M. D. Howell, J. T. Cuperus, D. Li, J. E. Hansen et al., Li, S., L. Liu, X. Zhuang, Y. Yu, X. Liu et al., 2013 microRNAs inhibit the 2008 Specificity of ARGONAUTE7-miR390 interaction and dual translation of target mRNAs on the endoplasmic reticulum in Arabi- functionality in TAS3 trans-acting siRNA formation. Cell 133: 128–141. dopsis. Cell 153: 562–574. https://doi.org/10.1016/j.cell.2013.04.005 https://doi.org/10.1016/j.cell.2008.02.033 Liang, G., Y. Li, H. He, F. Wang, and D. Yu, 2013 Identification of miRNAs Pagliarani, C., M. Vitali, M. Ferrero, N. Vitulo, M. Incarbone et al., and miRNA-mediated regulatory pathways in Carica papaya. Planta 238: 2017 The accumulation of miRNAs differentially modulated by drought 739–752. https://doi.org/10.1007/s00425-013-1929-6 stress is affected by grafting in grapevine. Plant Physiol. 173: 2180–2195. Lin, S. I., S. F. Chiang, W. Y. Lin, J. W. Chen, C. Y. Tseng et al., https://doi.org/10.1104/pp.16.01119 2008 Regulatory network of microRNA399 and PHO2 by systemic Papp, I., L. A. Mur, A. Dalmadi, S. Dulai, and C. Koncz, 2004 A mutation in signaling. Plant Physiol. 147: 732–746. https://doi.org/10.1104/ the Cap Binding Protein 20 gene confers drought tolerance to Arabidopsis. pp.108.116269 Plant Mol. Biol. 55: 679–686. https://doi.org/10.1007/s11103-004-1680-2 Liu, F., W. Wang, X. Sun, Z. Liang, and F. Wang, 2015 Conserved and Pantaleo, V., G. Szittya, S. Moxon, L. Miozzi, V. Moulton et al., novel heat stress-responsive microRNAs were identified by deep se- 2010 Identification of grapevine microRNAs and their targets using quencing in Saccharina japonica (Laminariales, Phaeophyta). Plant Cell high-throughput sequencing and degradome analysis. Plant J. 62: Environ. 38: 1357–1367. https://doi.org/10.1111/pce.12484 960–976. Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 785 Pantaleo, V., M. Vitali, P. Boccacci, L. Miozzi, D. Cuozzo et al., 2016 Novel leaves. Plant Physiol. 157: 683–691. https://doi.org/10.1104/ functional microRNAs from virus-free and infected Vitis vinifera plants pp.111.180083 under water stress. Sci. Rep. 6: 20167. https://doi.org/10.1038/srep20167 Solofoharivelo, M. C., A. P. van der Walt, D. Stephan, J. T. Burger, and Paim Pinto, D. L., L. Brancadoro, S. Dal Santo, G. De Lorenzis, M. Pezzotti S. L. Murray, 2014 MicroRNAs in fruit trees: discovery, diversity and et al., 2016 The influence of genotype and environment on small RNA future research directions. Plant Biol. 16: 856–865. https://doi.org/ profiles in grapevine berry. Front. Plant Sci. 7: 1459. https://doi.org/ 10.1111/plb.12153 10.3389/fpls.2016.01459 Speth, C., E. M. Willing, S. Rausch, K. Schneeberger, and S. Laubinger, Pérez-Díaz, R., J. Madrid-Espinoza, J. Salinas-Cornejo, E. González- 2013 RACK1 scaffold proteins influence miRNA abundance in Arabi- Villanueva, and S. Ruiz-Lara, 2016 Differential roles for VviGST1, dopsis. Plant J. 76: 433–445. https://doi.org/10.1111/tpj.12308 VviGST3, and VviGST4 in proanthocyanidin and anthocyanin transport Stracke, R., J. J. Favory, H. Gruber, L. Bartelniewoehner, S. Bartels et al., in Vitis vinifera. Front. Plant Sci. 7: 1166. https://doi.org/10.3389/ 2010 The Arabidopsis bZIP transcription factor HY5 regulates expres- fpls.2016.01166 sion of the PFG1/MYB12 gene in response to light and ultraviolet-B Pilati, S., G. Bagagli, P. Sonego, M. Moretto, D. Brazzale et al., 2017 Abscisic radiation. Plant Cell Environ. 33: 88–103. acid is a major regulator of grape berry ripening onset: new insights into Sun, X., G. Fan, L. Su, W. Wang, Z. Liang et al., 2015 Identification of cold- ABA signaling network. Front. Plant Sci. 8: 1093. https://doi.org/10.3389/ inducible microRNAs in grapevine. Front. Plant Sci. 6: 595. https://doi. fpls.2017.01093 org/10.3389/fpls.2015.00595 Piofczyk, T., G. Jeena, and A. Pecinka, 2015 Arabidopsis thaliana natural Suzuki, M., R. Nakabayashi, Y. Ogata, N. Sakurai, T. Tokimatsu et al., variation reveals connections between UV radiation stress and plant 2015 Multiomics in grape berry skin revealed specific induction of the pathogen-like defense responses. Plant Physiol. Biochem. 93: 34–43. stilbene synthetic pathway by ultraviolet-C irradiation. Plant Physiol. 168: https://doi.org/10.1016/j.plaphy.2015.01.011 47–59. https://doi.org/10.1104/pp.114.254375 Pitchiaya, S., L. A. Heinicke, J. I. Park, E. L. Cameron, and N. G. Walter, Terrier, N., D. Glissant, J. Grimplet, F. Barrieu, P. Abbal et al., 2005 Isogene 2017 Resolving subcellular miRNA trafficking and turnover at single- specific oligo arrays reveal multifaceted changes in gene expression dur- molecule resolution. Cell Reports 19: 630–642. https://doi.org/10.1016/j. ing grape berry (Vitis vinifera L.) development. Planta 222: 832–847. celrep.2017.03.075 https://doi.org/10.1007/s00425-005-0017-y Piya, S., C. Kihm, J. H. Rice, T. J. Baum, and T. Hewezi, 2017 Cooperative Tilbrook, K., M. Dubois, C. D. Crocco, R. Yin, R. Chappuis et al., regulatory functions of miR858 and MYB83 during cyst nematode par- 2016 UV-B perception and acclimation in Chlamydomonas reinhardtii. asitism. Plant Physiol. 174: 1897–1912. https://doi.org/10.1104/ Plant Cell 28: 966–983. pp.17.00273 Tuteja, N., M. B. Singh, M. K. Misra, P. L. Bhalla, and R. Tuteja, Pontin, M. A., P. N. Piccoli, R. Francisco, R. Bottini, J. M. Martinez-Zapater 2001 Molecular mechanisms of DNA damage and repair: progress in et al., 2010 Transcriptome changes in grapevine (Vitis vinifera L.) cv. plants. Crit. Rev. Biochem. Mol. Biol. 36: 337–397. https://doi.org/ Malbec leaves induced by ultraviolet-B radiation. BMC Plant Biol. 10: 10.1080/20014091074219 224. https://doi.org/10.1186/1471-2229-10-224 Válóczi, A., C. Hornyik, N. Varga, J. Burgyan, S. Kauppinen et al., Rajagopalan, R., H. Vaucheret, J. Trejo, and D. P. Bartel, 2006 A diverse 2004 Sensitive and specific detection of microRNAs by northern blot and evolutionarily fluid set of microRNAs in Arabidopsis thaliana. Genes analysis using LNA-modified oligonucleotide probes. Nucleic Acids Res. Dev. 20: 3407–3425. https://doi.org/10.1101/gad.1476406 32: e175. https://doi.org/10.1093/nar/gnh171 Rajeswaran, R., and M. M. Pooggin, 2012 RDR6-mediated synthesis of Vanhaelewyn, L., E. Prinsen, D. Van Der Straeten, and F. Vandenbussche, complementary RNA is terminated by miRNA stably bound to template 2016 Hormone-controlled UV-B responses in plants. J. Exp. Bot. 67: RNA. Nucleic Acids Res. 40: 594–599. https://doi.org/10.1093/nar/gkr760 4469–4482. https://doi.org/10.1093/jxb/erw261 Reid, K. E., N. Olsson, J. Schlosser, F. Peng, and S. T. Lund, 2006 An Walker, A. R., E. Lee, J. Bogs, D. A. McDavid, M. R. Thomas et al., optimized grapevine RNA isolation procedure and statistical determina- 2007 White grapes arose through the mutation of two similar and tion of reference genes for real-time RT-PCR during berry development. adjacent regulatory genes. Plant J. 49: 772–785. https://doi.org/10.1111/ BMC Plant Biol. 6: 27. https://doi.org/10.1186/1471-2229-6-27 j.1365-313X.2006.02997.x Ries, G., W. Heller, H. Puchta, H. Sandermann, H. K. Seidlitz et al., Wang, B., Y. F. Sun, N. Song, X. J. Wang, H. Feng et al., 2013 Identification 2000 Elevated UV-B radiation reduces genome stability in plants. Na- of UV-B-induced microRNAs in wheat. Genet. Mol. Res. 12: 4213–4221. ture 406: 98–101. https://doi.org/10.1038/35017595 https://doi.org/10.4238/2013.October.7.7 Rizzini, L., J. J. Favory, C. Cloix, D. Faggionato, A. O’Hara et al., Wang, C., J. Han, C. Liu, K. N. Kibet, E. Kayesh et al., 2012 Identification of 2011 Perception of UV-B by the Arabidopsis UVR8 protein. Science microRNAs from Amur grape (Vitis amurensis Rupr.) by deep se- 332: 103–106. https://doi.org/10.1126/science.1200660 quencing and analysis of microRNA variations with bioinformatics. BMC Rock, C. D., 2013 Trans-acting small interfering RNA4: key to nutraceutical Genomics 13: 122. https://doi.org/10.1186/1471-2164-13-122 synthesis in grape development? Trends Plant Sci. 18: 601–610. https:// Wang, C., X. Leng, Y. Zhang, E. Kayesh, Y. Zhang et al., doi.org/10.1016/j.tplants.2013.07.006 2014 Transcriptome-wide analysis of dynamic variations in regulation Shin, D. H., M. Choi, K. Kim, G. Bang, M. Cho et al., 2013 HY5 regulates modes of grapevine microRNAs on their target genes during grapevine anthocyanin biosynthesis by inducing the transcriptional activation of the development. Plant Mol. Biol. 84: 269–285. https://doi.org/10.1007/ MYB75/PAP1 transcription factor in Arabidopsis. FEBS Lett. 587: s11103-013-0132-2 1543–1547. https://doi.org/10.1016/j.febslet.2013.03.037 Wang, Y., A. Itaya, X. Zhong, Y. Wu, J. Zhang et al., 2011 Function and Shivaprasad, P. V., H. M. Chen, K. Patel, D. M. Bond, B. Santos et al., evolution of a microRNA that regulates a Ca2+-ATPase and triggers the 2012 A microRNA superfamily regulates nucleotide binding site-leu- formation of phased small interfering RNAs in tomato reproductive cine-rich repeats and other mRNAs. Plant Cell 24: 859–874. https://doi. growth. Plant Cell 23: 3185–3203. https://doi.org/10.1105/tpc.111.088013 org/10.1105/tpc.111.095380 Wei, W., Z. Ba, M. Gao, Y. Wu, Y. Ma et al., 2012 A role for small RNAs in Shao, F., and S. Lu, 2013 Genome-wide identification, molecular cloning, DNA double-strand break repair. Cell 149: 101–112. https://doi.org/ expression profiling and posttranscriptional regulation analysis of the 10.1016/j.cell.2012.03.002 Argonaute gene family in Salvia miltiorrhiza, an emerging model me- Wu, D., Q. Hu, Z. Yan, W. Chen, C. Yan et al., 2012 Structural basis of dicinal plant. BMC Genomics 14: 512. https://doi.org/10.1186/1471-2164- ultraviolet-B perception by UVR8. Nature 484: 214–219. https://doi.org/ 14-512 10.1038/nature10931 Si-Ammour, A., D. Windels, E. Arn-Bouldoires, C. Kutter, J. Ailhas et al., Xia, R., H. Zhu, Y. Q. An, E. P. Beers, and Z. Liu, 2012 Apple miRNAs and 2011 miR393 and secondary siRNAs regulate expression of the TIR1/ tasiRNAs with novel regulatory networks. Genome Biol. 13: R47. https:// AFB2 auxin receptor clade and auxin-related development of Arabidopsis doi.org/10.1186/gb-2012-13-6-r47 786 | S. Sunitha et al. Xia, R., B. C. Meyers, Z. Liu, E. P. Beers, and S. Ye, 2013 MicroRNA response to light and copper. Plant Cell 26: 4933–4953. https://doi.org/ superfamilies descended from miR390 and their roles in secondary small 10.1105/tpc.114.127340 interfering RNA biogenesis in eudicots. Plant Cell 25: 1555–1572. https:// Zhang, Z. L., M. Ogawa, C. M. Fleet, R. Zentella, J. Hu et al., doi.org/10.1105/tpc.113.110957 2011 Scarecrow-like 3 promotes gibberellin signaling by antagonizing Yamakawa, T., S. Kato, K. Ishida, T. Kodama, and Y. Minoda, master growth repressor DELLA in Arabidopsis. Proc. Natl. Acad. Sci. 1983 Production of anthocyanins by Vitis cells is suspension culture. USA 108: 2160–2165. https://doi.org/10.1073/pnas.1012232108 Agric. Biol. Chem. 47: 2185–2191. Zhou, X., G. Wang, and W. Zhang, 2007 UV-B responsive microRNA Yi, X., Z. Zhang, Y. Ling, W. Xu, and Z. Su, 2015 PNRD: a plant non- genes in Arabidopsis thaliana. Mol. Syst. Biol. 3: 103. https://doi.org/ coding RNA database. Nucleic Acids Res. 43: D982–D989. https://doi. 10.1038/msb4100143 org/10.1093/nar/gku1162 Zhou, L., Y. Liu, Z. Liu, D. Kong, M. Duan et al., 2010 Genome-wide Yoon, M. Y., M. Y. Kim, S. Shim, K. D. Kim, J. Ha et al., identification and analysis of drought-responsive microRNAs in 2016 Transcriptomic profiling of soybean in response to high-intensity Oryza sativa. J. Exp. Bot. 61: 4157–4168. https://doi.org/10.1093/jxb/ UV-B irradiation reveals stress defense signaling. Front. Plant Sci. 7: 1917. erq237 https://doi.org/10.3389/fpls.2016.01917 Zhou, B., P. Fan, Y. Li, H. Yan, and Q. Xu, 2016 Exploring miRNAs in- Zhai, J., D. H. Jeong, E. De Paoli, S. Park, B. D. Rosen et al., volved in blue/UV-A light response in Brassica rapa reveals special reg- 2011 MicroRNAs as master regulators of the plant NB-LRR defense ulatory mode during seedling development. BMC Plant Biol. 16: 111. gene family via the production of phased, trans-acting siRNAs. Genes https://doi.org/10.1186/s12870-016-0799-z Dev. 25: 2540–2553. https://doi.org/10.1101/gad.177527.111 Zhu, H., R. Xia, B. Zhao, Y. Q. An, C. D. Dardick et al., 2012 Unique Zhang, C., G. Li, J. Wang, and J. Fang, 2012 Identification of trans-acting expression, processing regulation, and regulatory network of peach siRNAs and their regulatory cascades in grapevine. Bioinformatics 28: (Prunus persica) miRNAs. BMC Plant Biol. 12: 149. https://doi.org/ 2561–2568. https://doi.org/10.1093/bioinformatics/bts500 10.1186/1471-2229-12-149 Zhang, H., X. Zhao, J. Li, H. Cai, X. W. Deng et al., 2014 MicroRNA408 is critical for the HY5–SPL7 gene network that mediates the coordinated Communicating editor: B. Gregory Volume 9 March 2019 | UV-B Responsive sRNAs in Grapevine | 787