Drought-inducible expression of Hv-miR827 enhances drought tolerance in transgenic barley
Drought is one of the major abiotic stresses reducing crop yield. Since the discovery of plant microRNAs (miRNAs), considerable progress has been made in clarifying their role in plant responses to abiotic stresses, including drought. miR827 was previously reported to confer drought tolerance in transgenic Arabidopsis. We examined barley (Hordeum vulgare L. ‘Golden Promise’) plants over-expressing miR827 for plant performance under drought. Transgenic plants constitutively expressing CaMV-35S::Ath-miR827 and drought-inducible Zm-Rab17::Hv-miR827 were phenotyped by non-destructive imaging for growth and whole plant water use efficiency (WUEwp). We observed that the growth, WUEwp, time to anthesis and grain weight of transgenic barley plants expressing CaMV-35S::Ath-miR827 were negatively affected in both well-watered and drought-treated growing conditions compared with the wild-type plants. In contrast, transgenic plants over-expressing Zm-Rab17::Hv-miR827 showed improved WUEwp with no growth or reproductive timing change compared with the wild-type plants. The recovery of Zm-Rab17::Hv-miR827 over-expressing plants also improved following severe drought stress. Our results suggest that Hv-miR827 has the potential to improve the performance of barley under drought and that the choice of promoter to control the timing and specificity of miRNA expression is critical.
KeywordsAth-miR827 Hv-miR827 Non-destructive imaging Promoter
A number of abiotic stresses affect crop yields. Drought is one of the major stresses constraining plant growth and crop productivity in many parts of the world (Boyer and Westgate 2004). With the expected increase in world population and declining water resources for crop production, it is imperative that new strategies for enhancing food production are explored. The development of drought-tolerant varieties will be particularly important for increasing food production given the expected climate changes and understanding plant drought tolerance mechanisms will be crucial (Lawlor 2013). However, the genetic mechanisms of drought response and tolerance are highly complex involving multiple pathways and large suites of genes (Nevo and Chen 2010; Blum 2011). Molecular tools and transgenic approaches for the manipulation of genes have been used to elucidate these mechanisms (Zhang et al. 2004; Shinozaki and Yamaguchi-Shinozaki 2007; Nakashima et al. 2009; Morran et al. 2011). However, considerable additional work will be needed to identify master regulatory genes and their signalling networks required to improve plant performance under drought.
microRNAs (miRNAs) are an important class of regulatory molecules. Mature miRNAs are non-protein coding, approximately 18–21 nucleotide (nt), single-stranded, highly conserved RNA sequences (Axtell and Bartel 2005; Zhang et al. 2006; Jones-Rhoades et al. 2006; Sunkar et al. 2012). In association with Argonaute (AGO) proteins, mature miRNAs recognize target mRNAs based on sequence complementarity and, via target cleavage, function as negative transcriptional/translational regulators across multiple regulatory networks both in plants and animals (Bartel 2009).
Several studies report the involvement of plant miRNAs in various abiotic stress responses (Sunkar et al. 2007; Lu and Huang 2008; Li et al. 2008; Trindade et al. 2010; Wei et al. 2009; Hackenberg et al. 2015; Liu et al. 2015). Research groups have investigated transcription profiles of miRNAs upon exposure to drought stress for several plant species including Arabidopsis, maize, tobacco, poplar, soybean, wheat, sorghum, rice and barley (Sunkar and Zhu 2004; Liu et al. 2008; Wei et al. 2009; Frazier et al. 2011; Qin et al. 2011; Kulcheski et al. 2011; Kantar et al. 2011; Budak et al. 2015; Katiyar et al. 2015; Hackenberg et al. 2015; Cheah et al. 2015; Ferdous et al. 2016; Liu et al. 2016a, b). Constitutively expressed miRNA Ath-miR827 increased drought tolerance in Arabidopsis (Aukerman and Park 2009). In the present study, barley plants were genetically modified to express either the precursor miRNA (pre-miRNA) of Ath-miR827 or its cereal pre-miRNA orthologue Hv-miR827. Ath-miR827 was constitutively expressed using the promoter CaMV-35S while Hv-miR827 expression was drought inducible through the action of the Zm-Rab17 promoter. The aim of the present study was to evaluate miR827 function in transgenic barley plants expressing either Ath-miR827 or Hv-miR827 and to test plant performance upon exposure to drought stress.
Materials and methods
Generation of transgenic barley expressing Ath-miR827 and Hv-miR827
Barley (Hordeum vulgare cv. ‘Golden Promise’) was transformed with the precursor miRNA827s (pre-miR827, MI0005383) from Arabidopsis and barley (Schreiber et al. 2011) under the control of the constitutive promoter CaMV-35S or the drought-inducible promoter Zm-Rab17, respectively, with the Nos-terminator. The transformation method, selection and regeneration procedures were as described in Morran et al. (2011). Zm-Rab17::GUS barley ‘Golden Promise’ plants were obtained as described in Ismagul et al. (2014). In brief, the pre-miR827 was isolated from Arabidopsis, reverse transcribed and cloned into the pMDC32 vector (Curtis and Grossniklaus 2003). The isolated pre-miR827 from barley was reverse transcribed and cloned into another pMDC32 vector in which the CaMV-35S promoter was excised using HindIII-KpnII restriction sites and replaced with a Zm-Rab17 promoter (634-bp fragment) (Busk et al. 1997). Sequence alignment of pre-Hv-miR827 and pre-Ath-miR827 is shown in Fig. S1. The mature miRNA sequence of Hv-miR827 and Ath-miR827 are not identical, differing at two nucleotides (Fig. S1). Both of the constructs were transformed into the barley genotype ‘Golden Promise’ using Agrobacterium-mediated transformation. T0 seedlings/events were obtained for the transgenic plants containing constructs of CaMV-35S::Ath-miR827 and Zm-Rab17::Hv-miR827. Here the term ‘event’ denotes the group of plants from each respective (CaMV-35S::Ath-miR827 and Zm-Rab17::Hv-miR827) transformation episode that regenerated from tissue culture. Plantlets were then transferred to coco-peat medium in pots under greenhouse conditions. Genomic DNA was extracted from mature leaf tissue of wild-type (WT) transgenic lines and null plants following the protocol of Edwards et al. (1991). The presence or absence of the transgenic fragment DNA was confirmed in CaMV-35S::Ath-miR827 over-expressed (OX) plants and in Zm-Rab17::Hv-miR827 drought-inducible (DI) plants by PCR amplification using 1 μl of genomic DNA with CaMV-35S or Zm-Rab17 promoter specific forward primers (CaMV-35S forward: 5′-TTCATTTCATTTGGAGAGGACCTCGACT-3′; Zm-Rab17 forward: 5′-CGGGCTGGTATTTCAAAACTAT-3′) and Nos-terminator specific reverse primer (5′-AACCCATCTCATAAATAACGTCATGCA-3′) for the respective construct. PCR amplification was carried out in a DNA Engine Tetrad Peltier Thermal Cycler (Bio-Rad Lab, Hercules, CA, USA) using IMMOLASE™ DNA Polymerase according to the manufacturer’s instructions. Each 20 μl PCR reaction consisted of 0.2 mM dNTP mixture, 1.5 mM MgCl2, 1× buffer, 1.25 units DNA polymerase (Bioline, Sydney, Australia), 0.5 μM of each primer and 50 ng DNA template. The thermal cycler was programmed for an initial step at 95 °C for 10 min followed by 35 cycles with denaturation for 30 s at 95 °C, annealing for 30 s and extension for 30 s at 72 °C. The reaction was terminated with a final extension at 72 °C for 10 min. For annealing temperatures, 56 and 55 °C were used for the aforementioned primer pairs of OX and DI lines, respectively. Barley S-adenosylmethionine decarboxylase (Hv-SAMDC) gene was used as a control gene for PCR (forward 5′-CTCAAACTGCAACAATGGCCG-3′ and reverse 5′-ACAGACGGAACAGCGACAGC-3′) with annealing temperature of 58 °C. Here, the events regenerated from each respective transformation through tissue culture lacking the transgene are called ‘null’ events. Southern blotting was carried out in order to select single copy transgenic plants following the method described in Shi et al. (2010). Single copy homozygous lines were derived from self-pollination of low copy number plants from the T0 and the subsequent generations. Thus, we obtained three single copy, homozygous events for CaMV-35S::Ath-miR827 and four single copy, homozygous events for Zm-Rab17::Hv-miR827 constructs.
Determination of drought treatment using transgenic barley plants expressing a Zm-Rab17 controlled GUS gene
To determine the timing and strength of our drought-inducible promoter, we conducted an experiment using Zm-Rab17 controlled expression of GUS in transgenic barley (Cv. ‘Golden Promise’) plants, and observed the drought-inducible expression during the progression and recovery from water deficit stress. Seeds were sown in UC (University of California) mix/coco peat/clay loam = 1:1:1 in a growth room at 23 °C day and 18 °C night temperature, 12-h/12-h light/dark photoperiod, 450 μmol m−2 s−1 photosynthetically active radiation and 60 % relative humidity, and plants were well-watered for 4 weeks. Four weeks after germination, watering was stopped to provide the drought treatment for 14 days. At 0 (65 % gravimetric water content (GWC)), 3 (45 % GWC), 5 (28 % GWC), 7 (18 % GWC), 9 (16 % GWC), 12 (12 % GWC) and 14 (10 % GWC) days after water withdrawal, and 3 (59 % GWC), 6 (60 % GWC), 9 (61 % GWC) days after re-watering, leaves were collected for GUS staining. The leaf samples were stained for GUS activity as described by Jefferson et al. (1987). Samples were immersed in GUS staining solution containing 0.1 M Na2HPO4/NaH2PO4 pH 7.0, (pH 7), 10 mM Na-EDTA, 0.5 mM K4Fe[CN]6 and 0.5 mM K3Fe[CN]6, 0.1 % Triton X-100, 2 mM x-gluc and infiltrated under vacuum. Leaf samples were incubated for 24 h at 37 °C. The reaction was stopped and cleared in 70 % ethanol. The drought-treated pots were weighed at each time point of sample collection to determine the GWC at each sampling stage. GWC was calculated using the formula: GWC (%) = [(wet soil weight / dry soil weight) − 1] × 100 where wet soil weight is the soil weight during sample collection and dry soil weight is the oven dried soil weight. Six individual plants were used for this experiment. The following drought phenotyping experiments were carried out under mild (18 and 16 % GWC) and severe (12 and 10 % GWC) drought treatments.
Non-destructive imaging of shoot area under mild drought
Seeds of single copy T3 transgenic barley lines, respective null segregants and WT were surface sterilized (using 5 % sodium hypochloride for 30 min), then sown in cellular plug trays. After confirming the presence of transgene by PCR, 10-day-old seedlings at the same physiological stage were transplanted to sealed white pots (19.46 cm height × 14.94 cm diameter, Berry Plastics, item number T51386CP) filled with potting mix, as before. One seedling was transplanted per pot. White marble chips (National Terrazzo & Cement Works Pty Ltd in North Plympton, 8294 1233, 3 mm size) covered the soil surface to minimize evaporation and the residual soil evaporation was calculated gravimetrically in pots with no plants. Twenty days after transplanting, the pots were loaded onto a fully automated conveyor system in a temperature-controlled Smarthouse in The Plant Accelerator® facility (Honsdorf et al. 2014). The plants of constitutively OX lines, DI lines, WT and null plants were subjected to a constant water deficit treatment maintaining 18 % GWC, while control pots were maintained at 45 % GWC. Watering was at automatic watering stations programmed to water each pot to weight (Bizerba, Balingen, Germany). The temperature of the Smarthouse was 22 °C day and 15 °C night. Plants were grown in natural light.
With the onset of the drought treatment, control and drought-treated plants were imaged in The Plant Accelerator® facility using a LemnaTec 3D Scanalyzer (LemnaTec, GmbH, Germany). Each day, high-resolution visible light (RGB) digital images were taken, including two side and one top view for each plant, for 30 days. The projected shoot area of an individual plant was calculated as in Honsdorf et al. (2014). To separate the plant tissue area from the background, background-foreground separation was applied. Kilopixels per image per plant were estimated, and the sum of kilopixels from the three images was the projected shoot area. Eight independent plants were assessed per genotype per treatment.
Water use efficiency (WUEwp) was determined as the ratio of shoot biomass at the end of the imaging period against the amount of water (g/ml/plant) supplied during the imaging period. Anthesis time (day) and grain weight per plant (g/plant) were recorded during anthesis and after harvesting, respectively. Upon harvesting, the above-ground shoot material was collected and oven (Contherm Scientific Ltd, Wellington, New Zealand) dried for a week at 60 °C for dry biomass measurements. Eight independent replicates were used per genotype and treatment.
Experimental layout and statistical analysis
The experimental design for non-destructive imaging considered 10 genotypes (DI L-1, DI L-2, DI L-3, DI L-4; DI L-4 was excluded from analysis due to germination failure) of Zm-Rab17:Hv-miR827 transgenic and respective null plants; OX L-1, OX L-2, OX L-3 of CaMV-35S:Ath-miR827 transgenic and respective null plants and the WT under two watering conditions (drought or well-watered). The design employed was a split-plot design with main plots being two consecutive pots/carts. The experiment involved 160 pots arranged in eight lanes by 20 positions. Eight biological replicates were used for each treatment. The lines were assigned to pairs of pots using a nearly-trend-free, randomized complete-block design. The water conditions were randomly assigned to the two consecutive pots/carts within each main plot. The design was generated in the R statistical software environment (R Development Core Team 2014) using the Digger and dae packages (Coombes 2009; Brien 2011). The layout for the experiment is given in Fig. S2.
E[Y] = Genotypes × Treatments
var[Y] = Half/Lanes/Mainplots/Carts
It allows for (i) differences amongst the varieties and water conditions and (ii) for variability within the smarthouse across the main plots, between lanes, between pairs of carts that formed main plots and between individual carts. The overall significance tests used an F test statistic and tests for significance between means, which was conducted using a least significant difference (LSD) value at the 5 % significance level. The remaining responses were analysed using a two-way ANOVA for genotypes and treatments.
Drought tolerance following mild and severe periodic drought stress
To further understand the drought performance of the Zm-Rab17 controlled Hv-miR827, the T4 generation of DI lines and the controls (WT and null) plants were grown in a growth room with conditions as before. Eight independent plants from each event were grown as one plant per pot. Four-week-old (stem elongation stage) plants were first subjected to drought treatment avoiding inter-pot variation by maintaining watering to weight for 9 days until the pots reached 16 % GWC. At the end of day 9 (16 % GWC), re-watering started for all the plants. Four days after re-watering, the plants were harvested and fresh weight of above-ground biomass (n = 8) was measured.
Subsequently, the T4 generation of DI lines and the controls were grown for 4 weeks maintaining regular watering in the above-mentioned growth chamber conditions. Four-week-old (stem elongation stage) plants were subjected to drought treatment avoiding inter-pot variation by maintaining watering to weight for 14 days until the pots reached 10 % GWC. At day 12 after drought inception (12 % GWC), photosynthetic assimilation (A) and stomatal conductance (gs) were measured on a fully expanded, mature leaf (n = 3) at photosynthetically active radiation inside the chamber (PARi) 1500 μmol m−2 s−1 using the LI-6400XT Portable Photosynthesis System (Li-Cor, Lincoln, NE, USA). Fully expanded mature leaf blades (five independent plants, three leaves from each plant) were used for measuring the relative chlorophyll content (per 2 × 3 mm area of leaf) using SPAD-502 Chlorophyll Meter (Konica Minolta, Australia) at days 0, 6, 9 and 12 after drought inception. At day 12 after drought inception (12 % GWC), fully expanded mature leaf blades (n = 3) were also collected from well-watered and drought-treated plants of control and DI lines to isolate RNA for qRT-PCR.
After the severe drought episode, at day 14 after drought inception (10 % GWC), re-watering started to observe the recovery and survival rate of the plants. Plant survival was determined at 5 weeks after re-watering.
Expression of miR827 and its putative target
Total RNA was extracted from fully expanded mature leaves of WT, transgenic lines and null plants using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Additional genotypes were grown in water-limited and well-watered conditions and mature leaf samples collected exactly as described in Ferdous et al. (2016). To remove genomic DNA contamination, RNA samples were treated with DNA-free™ reagents twice (Ambion, Life Technologies, Grand Island, NY, USA) according to the manufacturer’s instructions. For semi-quantitative RT-PCR, 1 μg of total RNA from the WT, transgenic lines and nulls was reverse transcribed with random hexamer priming (100 ng) using SuperScript® III RT (Life Technologies, Carlsbad, CA, USA) following the manufacturer’s instructions. Expression of the pre-Ath-miR827 was confirmed in the T1 generation using precursor-Ath-miR827 specific forward (5′-TTAGATGACCATCAACAAACT-3′) and Nos-terminator specific reverse (5′-AACTAGTTAATTAAGGAATTATCGAA-3′) primers in the CaMV-35S::Ath-miR827 lines. Precursor-Hv-miR827 specific forward (5′-TTAGATGACCATCAGCAAACA-3′) and Nos-terminator specific reverse (5′-AACTAGTTAATTAAGGAATTATCGAA-3′) primers were used to assess the expression of pre-Hv-miR827 in the Zm-Rab17::Hv-miR827 lines.
The expression of barley glycolytic glyceraldehyde-3-phosphate dehydrogenase (Hv-GAPDH, accession number: X60343.1) gene provided an internal control. Primers (forward 5′-GCCAAGACCCAGTAGAGC-3′ and reverse 5′-CACATTTATTCCCATAGACAAAGG-3′) were designed to flank an intron in the endogenous gene Hv-GAPDH.
To quantify expression of Hv-miR827 and any inverse correlation with its predicted target, qRT-PCR was used. Specific stem-loop reverse transcription primer (5′-GTCGTATCCAGTGCAGGGTCCGAGGTATTCGCACT GGATACGACTGTTTG-3′) and appropriate forward (5′-TCTGACGACGTTAGATGACCATC-3′) and reverse (5′-GTGCAGGGTCCGAGGT-3′) primers were designed following the method described by Chen et al. (2005) and Varkonyi-Gasic et al. (2007).
Target prediction and validation through degradome analysis was conducted exactly as in Ferdous et al. (2016) except that the degradome library from only one genotype ‘Golden Promise’ was used in this experiment.
We designed the primers for the putative target, an SPX (SYG1/PHO81/XPR1) domain encoding transcript, spanning the miRNA binding site (forward: 5′-GCACCTGGAGCCATCGTC-3′ and reverse 5′-TTCACCATCTTGCTGTTTCTACC-3′) using AlleleID software (Premier Biosoft International, Palo Alto, CA, USA). All oligo primers used in this study were synthesized by Sigma-Aldrich (Sydney, Australia).
Total RNA extraction and removal of genomic DNA were conducted exactly as before. The concentration and integrity of the DNA-free RNA was determined by Agilent-2100 Bioanalyzer using RNA 6000 NanoChips (Agilent Technologies, Santa Clara, CA, USA). The RNA integrity number (RIN) for the experimental samples ranges from 5.7 to 7.1. Two micrograms of total RNA from the well-watered and drought-treated samples was reverse transcribed with Hv-miR827-specific stem-loop reverse transcription primer using SuperScript® III RT (Life Technologies, Carlsbad, CA, USA) by the pulsed RT method (Varkonyi-Gasic et al. 2007). For target qRT-PCR, 2 μg of total RNA from the same experimental samples was reverse transcribed by random hexamer (100 ng) using SuperScript® III RT (Life Technologies, Carlsbad, CA, USA) following the manufacturer’s instructions. For qRT-PCR, three biologically independent plants were assayed for each treatment per line with three technical replicates per biological replicate. qRT-PCR was carried out exactly as in Ferdous et al. (2015) using the RG6000 Rotor-Gene real-time thermal cycler (Qiagen, Valencia, CA, USA). Mature miRNA qRT-PCR product was sequenced using the M13 reverse primer following the method described by Ferdous et al. (2015). The sequence of a qRT-PCR amplicon of the un-cleaved target was also verified by Sanger sequencing. The melt curve plots for miRNA and mRNA qRT-PCR are shown in Fig. S3.
Generation of transgenic barley constitutively over-expressing (OX) Ath-miR827 or under the control of a drought-inducible promoter (DI) Hv-miR827
Transgenic barley expressing Ath-miR827 under the control of constitutive CaMV-35S promoter and Hv-miR827 under the control of Zm-Rab17 promoter were generated via Agrobacterium-mediated transformation. Three representative single copy lines for CaMV-35S::Ath-miR827 (OX L-1, OX L-2 and OX L-3) and Zm-Rab17::Hv-miR827 (DI L-1, DI L-2 and DI L-3) were selected based on southern blot analysis (Fig. S4). Transgene presence in these lines was confirmed using transgene specific primers designed to amplify either a 765-bp (OX plants) or 1111-bp (DI plants) fragment spanning the promoter through to the terminator (Fig. S5a, b). Transgene specific semi-quantitative RT-PCR confirmed pre-miRNA expression for both OX (282 bp amplicon) and DI (101 bp amplicon) transgenic lines. These amplicons were not detected in either the WT or null segregants (Fig. S5c, d). The expression of Hv-GAPDH as internal control generated a 78-bp amplicon in all genotypes. (Fig. S5c, d).
Drought inducibility of Zm-Rab17
Shoot area and WUEwp of the OX and DI transgenic plants
Anthesis, shoot biomass and grain weight of the transgenic plants
Based on phenotyping results from non-destructive imaging, we observed several unexpected phenotypes; that is, reduced shoot area, WUEwp, delayed anthesis, grain weight per plant and reduced shoot biomass of OX lines compared with non-transgenics upon drought and under well-watered conditions. DI lines did not exhibit growth penalties nor altered time to anthesis compared to non-transgenics, but showed higher WUEwp compared with non-transgenics exposed to drought. Therefore, DI lines were further evaluated.
DI lines recovered from mild drought stress
DI lines exhibit improved physiological response and survival rate during severe drought stress
Under severe drought stress, when GWC declined from 65 to 10 % (Fig. S7a) before re-watering, a higher percentage of DI L-1 and DI L-3 survived relative to non-transgenic plants; however, the survival rate for DI L-2 did not differ from non-transgenics (Fig. S7b, c).
Zm-Rab17 driven expression of Hv-miR827 in DI lines
Putative targets of Hv-miR827 and expression analysis of target genes in DI lines
From the in silico analysis, we obtained predicted targets of Hv-miR827 (Supplementary Table 1). The putative targets included SPX encoding transcripts, nucleotide-binding site–leucine-rich repeat (NBS-LRR) proteins, caseinolytic protease (Clp) amino terminal domain, aberrant pollen transmission 1 (APT1) and TCP (TB1, CYC and PCFs) family transcription factors. Experimental confirmation of these targets was undertaken by interrogation of the degradome library to identify cleavage products in leaf tissue derived from the pooled leaves and roots of well-watered and drought-treated plants. However, no cleavage products were identified for any of the putative targets. Previously, Osa-miR827 was reported to target two genes encoding SPX, with mRNA cleavage of both targets at the predicted canonical site being confirmed in rice (cv. Nipponbare) (Lin et al. 2010). To understand whether SPX was a likely target for Hv-miR827 in barley, we determined whether there was an inverse correlation in SPX abundance (JLOC1_37332/MLOC_57566) relative to Hv-miR827 at the transcript level. For this purpose, we used transgenic and non-transgenic plants differing in total Hv-miR827 abundance. We did not observe any inverse correlation between Hv-miR827 and SPX in either the DI lines or non-transgenic controls (Fig. 6 and Fig. S9). Interestingly, we observed increased SPX transcript abundance under drought for DI L-1 and DI L-3 when Hv-miR827 expression was induced. For DI L-2, the expression of the SPX transcript did not significantly differ between the two watering regimes.
Sequencing revealed that there are three and two nucleotides mismatches between the complementary pairing of Ath-miR827: barley SPX target site and Hv-miR827: barley SPX target site, respectively (Fig. S10).
The availability of barley deep-sequencing datasets and the application of bioinformatic tools have facilitated the identification of both conserved and novel miRNAs (Schreiber et al. 2011). Such analyses have revealed a number of conserved and novel miRNAs differentially expressed in barley under drought stress (Hackenberg et al. 2015; Ferdous et al. 2016). However, functional analysis of miRNAs and their cognate targets is challenging as methodologies for analysis typically require mutants or the generation of transgenics. Because mutant and other genomics resources in barley were not as well established as for the model crops Arabidopsis and rice, we took a transgenic approach towards functional validation of barley miR827. To date, stable genetic modification (GM) of a miRNA in barley has not been reported. We used the barley transgenic plants to determine the phenotypic effect of both transgenic and cisgenic miR827 on plant performance under drought, given that miR827 was reported to enhance wilting avoidance in the model species Arabidopsis (Aukerman and Park 2009). Phenotyping of Hv-miR827 in transgenic barley indicated that this miRNA could play a regulatory role leading to improved plant performance under drought.
One of the first reactions of plants to declining leaf water potential is reduced growth rate (Boyer 1970). Growth reduction allows plants to minimize water consumption. In our study, we observed reduced shoot area of the Ath-miR827 over-expressers (OX lines) under both watering conditions. Other observed phenotypic differences included WUEwp, time to anthesis, shoot biomass and grain weight per plant, which were negatively affected compared to the WT plants under both watering regimes (Figs. 2a, b and 3a–c). Delayed anthesis might have affected final grain weight in the OX lines. These findings were somewhat unexpected considering these abnormal growth and developmental phenotypes were not reported in the model species Oryza (Lin et al. 2010). This indicated that miR827 derived from the dicot Arabidopsis may not function normally in barley. One plausible explanation is that Ath-miR827 over-expression causes aberrant down-regulation of its mRNA targets, resulting in perturbed growth and development. Predicted targets for Ath-miR827 in barley include membrane protein SPX, NBS-LRR domain containing proteins, aberrant pollen transmission 1 (APT1), seven in absentia (SINA) family proteins, protein kinase domain containing proteins and F-box/LRR-repeat proteins (Table S1). SPX was negatively regulated by Osa-miR827 and has been reported to function in phosphate (Pi) sensing in rice (Lin et al. 2010). Maize APT1 was previously reported to be involved in pollen tube growth (Xu and Dooner 2006). APT1 is also a member of a conserved protein family vital for cell elongation in higher plants (Xu and Dooner 2006). Another putative target of Ath-miR827, SINA, was reported to promote drought tolerance in Arabidopsis in an ABA-dependent manner (Bao et al. 2014). However, when aligned to the degradome sequences of ‘Golden Promise’, no cleavage products were obtained for any of the putative target sequences. This observation suggests that Ath-miR827-mediated target gene regulation might not be through post-transcriptional cleavage or that Ath-miR827 might have other yet unknown target(s) in barley. It is possible that the target genes are involved in development or flowering. This might have caused disruption of genes related to phenology affecting the combined phenotypes such as shoot area, anthesis time, biomass and grain yield in the Ath-miR827 over-expresser plants compared to that of the non-transgenic plants.
We tested the expression of mature Hv-miR827 and a predicted target SPX encoding transcript in the DI lines and found no inverse correlation of Hv-miR827 and SPX transcripts under drought (Fig. 6 and Fig. S9). Interestingly, we observed drought-induced expression of this target transcript in DI L-1 and DI L-3 compared to the non-transgenic plants (Fig. S9). We also examined the miR827 target site sequence in ‘Golden Promise’ which revealed that there were mismatches between the miRNA:target duplex including at the 5′ 9th nt position of both Ath-miR827 and Hv-miR827 (Fig. S10). While, typically extensive complementarity (≤5 mismatches) is required to ensure functional targeting, base pairing between the 5′ 2 to 13 nt positions of a miRNA with its target is critical for miRNA-mediated target suppression (Liu et al. 2014). Additionally, base pairing at the central positions, the 9 to 11 nt of miRNA:target sites, is particularly important for miRNA and the target pairing in the vicinity of the AGO-catalyzed slicing site (Parizotto et al. 2004; Schwab et al. 2005; Liu et al. 2014). Liu et al. (2014) demonstrated that a single nucleotide mismatch between miRNA and the target at positions 9 and 10, as well as combinations of mismatches at positions 9-11, resulted in complete elimination of the responsiveness of a miR164-targeted sensor sequence in Nicotiana benthamiana.
In five examined barley genotypes (including the WT ‘Golden Promise’), endogenous Hv-miR827 was not induced by drought, although it was expressed above detectable limits in three of the five genotypes and its expression has been confirmed in combined small RNA libraries of the barley genotypes ‘Golden Promise’ and ‘Pallas’ through deep-sequencing (Schreiber et al. 2011). Consistent with our results, native Hv-miR827 was not differentially expressed under drought in ‘Golden Promise’ in the deep-sequencing libraries, although it was induced by phosphorus deficit in ‘Pallas’ (Schreiber et al. 2011; Hackenberg et al. 2013). This contrasts with the drought-responsive expression of orthologous, conserved miR827 in maize, rice, bread and durum wheat (reviewed in Liu et al. 2016b). Liu et al. (2016a) were also able to confirm differential expression of durum homologous Bdi-miR827-3p between more and less drought-sensitive genotypes and between tissues but, similarly to our results for barley, were unable to identify cleavage products of the predicted target gene (in this case Aberrant pollen transmission 1) or inverse correlation between the miRNA and target mRNA expression. We generated transgenic barley plants constitutively expressing Hv-miR827 (driven by the CaMV-35S promoter) through tissue culture, but these failed to germinate from seed in the T1 generation (data not shown). This implies that the target of endogenous Hv-miR827 is critical for early development and that early over-expression of the miRNA might have negative consequences.
Constitutive over-expression of a transgene can result in a different degrees of growth retardation and delayed flowering in plants (Oh et al. 2007). It has been suggested that the use of stress-inducible promoters can minimize the undesirable phenotypes observed during the constitutive expression of a transgene (Kasuga et al. 1999; Morran et al. 2011). The Zm-Rab17 promoter is reported to be responsive to abscisic acid (ABA) and water stress (Busk et al. 1997). In our experiment, plants expressing Hv-miR827 under the control of the DI promoter Zm-Rab17 showed several promising phenotypes under drought treatment. The higher WUEwp of the two DI lines, DI L-1 and DI L-3 (Fig. 2d), and higher grain weight per plant of DI L-1 and DI L-3 (Fig. 3f) indicated that these two lines required less water than other genotypes to provide the same shoot area and maintain of grain weight .
In this study, efforts were made to ensure that all the factors other than the desired treatments (mild and severe drought) were non-limiting during the experimental period. Our observations suggest that the extent of water deficit might result in variation in the phenotypic performance of the drought-induced lines. In a mild drought experiment, all plants of the tested DI lines and the non-transgenic plants recovered after the drought period. However, the transgenic plants showed faster recovery compared to the non-transgenic plants (Fig. 4). Additionally, the increased plant biomass of the transgenics upon re-watering (Fig. S6) showed that drought-induced expression of Hv-miR827 improved the ability of barley plants to recover from mild drought stress. Under severe drought stress, when the GWC declined from 65 to 10 %, differences in both plant mortality and in the speed of recovery among the DI lines were observed (Fig. S7a, b) with DI L-1 and DI L-3 plants’ higher survival indicating their better performance than the other genotypes (Fig. S7c). Under severe drought stress, the super-abundant accumulation of mature Hv-miR827 in DI L-2 (Fig. 6) led to high mortality of this line relative to the other transgenics. Under mild drought stress, this was not the case indicating that recovery after re-watering depends on drought intensity and duration (Xu et al. 2010). This observation suggests that the level of drought-induced accumulation of miR827 has an impact on the phenotypic performance of the transgenic plants and that high levels of expression of Ath/Hv-miR827 might not be effective for improving the performance of barley plants under drought.
In our study, as one of the early observations of DI L-1 and DI L-3 was improved WUEwp under mild drought, we expected that the main driver of increased water-use efficiency in these two DI lines would be decreased transpirational water loss through low stomatal conductance compared to the non-transgenic plants. However, we still observed higher stomatal conductance, photosynthetic assimilation and relative leaf chlorophyll content of the DI L-1 and DI L-3 compared to the non-transgenic plants under severe drought stress (Fig. 5a–c), whereas DI L-2 behaved similarly to the non-transgenic plants under severe drought stress (Fig. 5a–c). These observations suggest that higher photosynthetic assimilation in these two lines could potentially contribute to the increased WUEwp. The ability of DI L-1 and DI L-3 to maintain key physiological processes, such as photosynthetic assimilation during severe drought treatment, is indicative of the potential of these lines to support productivity under water deficit (Centritto et al. 2009). A preliminary field trial with the three DI lines, WT and null plants was conducted at a single site in 2015 with encouraging results (data not shown). However, extensive field trials under drought at multiple sites and in large plots will be necessary to confirm the yield benefits seen in the preliminary trial.
Drought-induced expression of Hv-miR827 improved whole plant water use efficiency of transgenic barley under drought. The drought-induced expression of Hv-miR827 influenced grain weight per plant and enabled plants to recover after drought treatment. Our findings suggested that, unlike the drought-inducible expression of Hv-miR827, constitutive over-expression of Ath-miR827 had a negative effect on growth, time to anthesis and grain weight of transgenic plants. It was clear that the phenotypic performance of transgenic barley plants expressing miR827 could depend on the nature of promoter. These findings suggest the possibility of creating water use efficient transgenic barley utilizing Hv-miR827 under the control of the Zm-Rab17 promoter.
This research was supported by a grant to the Australian Centre for Plant Functional Genomics, supported through research funding from DuPont/Pioneer (USA). Our grateful thanks to Patricia Warner and ACPFG Transformation Group for barley transformation; Margaret Pallotta and Suzanne Manning for their assistance with Southern blotting; and Dr. Sergiy Lopato, Dr. Ainur Ismagul and Dr. Nataliya Kovalchuk for generation and selection of the Zm-Rab17::GUS transgenic barley germplasm. We specially thank the team of The Plant Accelerator for technical support in running the experiment and conducting the image analysis. The Plant Accelerator, Australian Plant Phenomics Facility, is funded under the National Collaborative Infrastructure Strategy. We are thankful to Dr. Ursula Langridge, Alex Kovalchuk and Yuri Onyskiv for their assistance with growing plants at different phases of this experiment, and Yuan Li and Hui Zhou for performing the qRT-PCR assays. We also thank Dr. Bu-Jun Shi for his advice at the early stage of this experiment.
- Aukerman M, Park W (2009) Drought tolerant plants and related constructs and methods involving genes encoding mir827, Patent application number: US20090165168Google Scholar
- Bao Y, Wang C, Jiang C, Pan J, Zhang G, Liu H, Zhang H (2014) The tumor necrosis factor receptor-associated factor (TRAF)-like family protein SEVEN IN ABSENTIA 2 (SINA2) promotes drought tolerance in an ABA-dependent manner in Arabidopsis. New Phytol 202(1):174–187Google Scholar
- Brien CJ (2011) dae: functions useful in the design and ANOVA of experiments. Version 2.1-7. http://cran.r-project.org/
- Butler D, Cullis BR, Gilmour AR, Gogel BJ (2010) Analysis of mixed models for S language environments: ASReml-R reference manual. DPI Publications, BrisbaneGoogle Scholar
- Cheah BH, Nadarajah K, Divate MD, Wickneswari R (2015) Identification of four functionally important microRNA families with contrasting differential expression profiles between drought-tolerant and susceptible rice leaf at vegetative stage. BMC Genomics 16:692–709CrossRefPubMedPubMedCentralGoogle Scholar
- Coombes NE (2009) Digger design search tool in R. http://www.austatgen.org/les/software/downloads/
- Edwards K, Johnstone C, Thompson C (1991) A simple and rapid method for the preparation of plant genomic DNA for PCR analysis. Nucleic Acids Res 19(6):1349–1349Google Scholar
- Kantar M, Lucas SJ , Budak H, Kantar M, Lucas SJ, Budak H (2011) miRNA expression patterns of Triticum dicoccoides in response to shock drought stress. Planta 233(3):471–484Google Scholar
- Kulcheski FR, de Oliveira LFV, Molina LG, Almerão MP, Rodrigues FA, Marcolino J, Barbosa JF, Stolf-Moreira R, Nepomuceno AL, Marcelino-Guimarães FC, Abdelnoor RV, Nascimento LC, Carazzolle MF, Pereira GAG, Margis R (2011) Identification of novel soybean microRNAs involved in abiotic and biotic stresses. BMC Genomics 12:307–323CrossRefPubMedPubMedCentralGoogle Scholar
- Lin SI, Santi C, Jobet E, Lacut E, El Kholti N, Karlowski WM, Verdeil JL, Breitler JC, Périn C, Ko SS, Guiderdoni E, Chiou TJ, Echeverria M (2010) Complex regulation of two target genes encoding SPX-MFS proteins by rice miR827 in response to phosphate starvation. Plant Cell Physiol 51:2119–2131CrossRefPubMedGoogle Scholar