Virus-induced down-regulation of GmERA1A and GmERA1B genes enhances the stomatal response to abscisic acid and drought resistance in soybean

Drought is a major threat to global soybean production. The limited transformation potential and polyploid nature of soybean have hindered functional analysis of soybean genes. Previous research has implicated farnesylation in the plant’s response to abscisic acid (ABA) and drought tolerance. We therefore used virus-induced gene silencing (VIGS) to evaluate farnesyltransferase genes, GmERA1A and GmERA1B (Glycine max Enhanced Response to ABA1-A and -B), as potential targets for increasing drought resistance in soybean. Apple latent spherical virus (ALSV)-mediated GmERA1-down-regulated soybean leaves displayed an enhanced stomatal response to ABA and reduced water loss and wilting under dehydration conditions, suggesting that GmERA1A and GmERA1B negatively regulate ABA signaling in soybean guard cells. The findings provide evidence that the ALSV-VIGS system, which bypasses the need to generate transgenic plants, is a useful tool for analyzing gene function using only a single down-regulated leaf. Thus, the ALSV-VIGS system could constitute part of a next-generation molecular breeding pipeline to accelerate drought resistance breeding in soybean.


Introduction
Climate change has increased the frequency and severity of diverse abiotic stresses worldwide [1,2]. Drought is by far the most devastating abiotic stress affecting plant growth and productivity [3]. In soybean (Glycine max L.), drought reduces seed quality and quantity, and induces changes in plant morphology, by reducing CO 2 assimilation, leaf area development, and symbiotic N 2 fixation [4,5]. Functional validation of candidate genes for improved drought resistance identified in model plants such as Arabidopsis thaliana could provide gene resources for breeding varieties of non-model staple crops, such as soybean, that can withstand drought conditions. Based on our knowledge of the mechanism underlying abscisic acid (ABA)/stress- 25˚C with 50% humidity and 400 ppm CO 2 under a 12 h light/12 h dark photoperiod and a light intensity of 150 μmol photons m -2 s -1 . The soybean growth stages such as V4 and V5 were determined as described previously [39]. For virus propagation, Chenopodium quinoa plants were grown in a greenhouse or growth chamber at 25˚C, as described previously [40]. C. quinoa was germinated on Sakata Super Mix A soil (Sakata Seed, Yokohama, Kanagawa, Japan) and planted in culture soil (Tsuchitaro: Sumitomo Forestry or Gacchirikun N-120: Tokita Seed, Saitama, Japan).

RNA isolation and RT-PCR analysis
Total RNA was isolated from C. quinoa and soybean using RNAiso Plus (Takara Bio, Otsu, Shiga, Japan) according to the manufacturer's instructions. Total RNA was pre-treated with RQ1 RNase-free DNase (Promega, Madison, WI, USA) and complementary DNA (cDNA) was synthesized using PrimeScript RT Master Mix (Takara Bio). Quantitative reverse-transcription (qRT) polymerase chain reaction (PCR) was performed using an ABI7500 Real-Time PCR system (Life Technologies, Grand Island, NY, USA), and SYBR Premix Ex Taq (Takara Bio). Virus RNA was detected by RT-PCR analysis using GoTaq Green Master Mix (Promega). The specific oligonucleotides used to amplify each gene are described in S1 Table. Statistical analysis was performed using StatPlus (v6 for Mac, AnalystSoft Inc.).

Virus inoculation
C. quinoa and soybean plants were inoculated with ALSV as described [19,23,41]. The plasmids for ALSV-RNA1 and RNA2 were mixed in equal amounts, and the DNA solution was mechanically inoculated onto the true leaves of C. quinoa using carborundum. Reverse osmosis (RO) water was used for mock inoculation. The inoculated C. quinoa plants were grown for two to three weeks. The upper leaves were sampled and ground in three volumes of extraction buffer (0.1 M Tris-HCl, pH 8.0, 0.1 M NaCl, 5 mM MgCl 2 [23]). Debris was precipitated by centrifugation, and the supernatants were used for secondary inoculation of C. quinoa. After two to three weeks, the upper leaves were sampled and stored at -80˚C. Total RNA, extracted from the upper leaves of C. quinoa plants subjected to a second inoculation, was used as an inoculum for soybean. Biolistic inoculation of germinated soybean seeds was performed using a PDS-1000/He Particle Delivery System (Bio-Rad, Hercules, CA, USA) and 1.0 Micron Gold Microcarrier (Bio-Rad) particles that had been coated with total RNA from C. quinoa as described [23]. Seven to ten germinated soybean seeds were placed onto a Petri dish and bombarded twice at 1,100 psi. Approximately 7 μg of total RNA was used per shot.

Measurement of physiological changes in VIGS plants
Water loss rates were measured as described [42], with minor modifications. A leaflet from the fourth or fifth trifoliate leaves of a virus-infected soybean plant was excised and fresh weight was measured over time. Thermal images of leaves were taken using an FLIR E60 thermal imaging camera according to the manufacturer's protocol (FLIR Systems, Wilsonville, OR, USA). The H 2 O conductance rates and leaf temperature were measured with a LI-6400XT portable photosynthesis system (Li-Cor Bioscience, Lincoln, NE, USA). The CO 2 concentration of the input flow, chamber block temperature, and level of the LED light source were set at 400 μmol mol -1 , 25˚C, and 500 μmol m -2 s -1 , respectively. A fourth or fifth trifoliate leaflet was placed in the leaf chamber of the LI-6400XT system, and time 0 was set after the measured values stabilized. After a measurement was taken at time 0, the leaflet was detached and the data were collected every 30 s. Stomatal apertures were measured as described [31,42], with minor modifications. Leaf disks of 1 cm 2 were excised from the fourth trifoliate leaves of ALSVinfected soybean plants, incubated for 4 h in 20 mM KCl, 1 mM CaCl 2 , 5 mM MES-KOH, pH 6.15, and 0.003% Silwet-77, and treated with or without 10 μM ABA for 4 h under the light condition. Photographs of guard cells were taken through a color laser three-dimensional profile microscope (Keyence, Osaka, Japan). Stomatal apertures were measured using ImageJ software (http://imagej.nih.gov/ij/). Drought tolerance assays were conducted as described [42] with minor modifications. Water was withheld from whole plants at growth stage V6 [39] for 3 days. Six independent trials were carried out, including three blind tests.

Sequence analysis of GmERA1 genes
Arabidopsis ERA1 plays an important role in ABA signaling [29][30][31] and is a promising candidate gene for genetic manipulation of drought stress tolerance in Arabidopsis and canola plants [35,36]. Although drought is considered to be a major problem in soybean production worldwide [43][44][45], ERA1 homologs in soybean have not been reported to date. Therefore, we searched the Soybean Knowledge Base (SoyKB, http://soykb.org/) using BlastP program and the amino acid sequence of Arabidopsis ERA1 as query. Two soybean genes, Glyma06g19740 and Glyma13g23780, were identified as homologs of Arabidopsis ERA1 and named GmERA1A and GmERA1B, respectively (S1 and S2 Figs, S2 Table). S2 Fig shows the sequence alignment of ERA1 proteins from soybean and Arabidopsis. GmERA1A and GmERA1B share 56.0% and 56.3% amino acid sequence identity with Arabidopsis ERA1, respectively. GmERA1A shares 92.9% amino acid sequence identity with GmERA1B. These data are consistent with the previous finding that soybean is a partially diploidized tetraploid [12]. Although these proteins have high levels of amino acid sequence identity, the N-terminus of Arabidopsis ERA1 is longer than that of GmERA1A or GmERA1B (

VIGS of GmERA1 genes in soybean
We next sought to analyze the function of GmERA1 genes by manipulating their expression in soybean plants. Since the transformation efficiency of soybean is much lower than that of Arabidopsis and rice [13,46,47], we used VIGS to down-regulate the GmERA1 genes in wild-type soybean plants. To date, BPMV, CMV, and ALSV vectors have been used for VIGS in soybean [17,18,19]. Because soybean is not a natural host of ALSV, in contrast to BPMV and CMV, we used an ALSV vector in the current study to minimize potential interference effects imposed by naturally compatible virus-host interactions on abiotic stress responses. To examine the effect of loss-of-function of GmERA1 genes in soybean plants, we performed ALSV-mediated VIGS in the sequenced model soybean cultivar Williams 82. We obtained ALSV-GmERA1N, ALSV-G-mERA1C, ALSV-GmPDS, and ALSV-VC through propagating the virus in C. quinoa plants inoculated with the ALSV-RNA1 cDNA plasmid combined with ALSV-RNA2 VIGS vectors either harboring or not a specific 300-bp trigger sequence required for VIGS (S1 and S4 Figs). ALSV-GmERA1N and ALSV-GmERA1C were designed to target both GmERA1A and GmERA1B due to their high level of nucleotide sequence similarity (95.3%) (S1A Fig [19,48], indicating that ALSV-mediated VIGS can be observed throughout the leaves after the fourth trifoliate stage of development. We performed qRT-PCR analysis to analyze the silencing efficiency of GmERA1 genes in the fourth trifoliate leaves (Fig 1) using a primer set that could amplify both GmERA1s (S1 Down-regulation of GmERA1s reduces water loss and gas exchange under water limitation ERA1, which is expressed in guard cells, regulates the transpiration rate in Arabidopsis by modulating ABA-mediated stomatal responses [29,31]. We therefore measured water loss in the leaves of GmERA1-down-regulated plants to evaluate their response to drought stress. The water loss rate was lower in the detached fourth or fifth trifoliate leaves of soybean plants inoculated with ALSV-GmERA1s than in those of plants inoculated with ALSV-VC (Fig 2A), suggesting that transpiration from the leaves of GmERA1-down-regulated plants was reduced under limited water conditions. This result is consistent with our infrared thermography observation that the leaf temperature of detached fourth trifoliate leaflets of plants inoculated with ALSV-GmERA1s was higher than that of detached fourth trifoliate leaflets of plants inoculated with ALSV-VC (Fig 2B). To confirm these results, we analyzed the leaf surface temperature over time using a LI-6400XT portable photosynthesis system. At 3 or more min after detachment, the surface temperature of detached fourth or fifth trifoliate leaflets of GmERA1-down-regulated plants was~1˚C higher than that of control plants (Fig 2C). Furthermore, using the LI-6400XT system, we measured the rate of gas exchange in the detached GmERA1down-regulated leaflets over time. At 4 to 16 min after detachment, the water-stressed fourth or fifth trifoliate leaves of GmERA1-down-regulated plants exhibited statistically significant reductions in stomatal conductance compared to that of control plants inoculated with ALSV-VC empty viral vector, suggesting that down-regulation of GmERA1s reduces gas exchange under water-limiting conditions (Fig 2D). These results support the notion that down-regulation of GmERA1s reduces water loss from water-stressed soybean leaves through its influence on stomatal regulation.
Down-regulation of GmERA1s enhances the stomatal response to ABA in soybean leaves ERA1 is a well-known negative regulator of ABA responses in Arabidopsis [30]. To examine whether GmERA1s-down-regulation affects stomatal responses to ABA, we analyzed stomatal apertures in the leaves of GmERA1-down-regulated plants treated with or without ABA. In the absence of ABA, no marked difference in stomatal aperture in the fourth trifoliate leaves was detected between GmERA1-down-regulated and control plants. However, in the presence of 10 μM ABA, the stomatal aperture in these leaves was significantly narrower in plants inoculated with ALSV-GmERA1s than in ALSV-VC or mock-inoculated plants (Fig 3), indicating that down-regulation of GmERA1s enhances the stomatal closure response to ABA. This result also suggests that GmERA1s function as negative regulators of ABA responses, as does Arabidopsis ERA1. Taken together, these data support the notion that altered ABA-mediated

GmERA1-down-regulated plants exhibit enhanced resistance to waterdeficit stress
We demonstrated that the GmERA1-down-regulated leaves displayed reduced water loss and gas exchange through ABA-mediated stomatal regulation compared to the vector control

Fig 2. GmERA1-down-regulated leaves show reduced water loss rates and stomatal conductance. (A)
Rates of water loss in the fourth or fifth trifoliate leaves of soybean plants inoculated with ALSV vectors. Each data point represents the mean of duplicate measurements (n = 3 for ALSV-VC and n = 6 for ALSV-GmERA1). Bars indicate SD. Asterisks at 0.3 and 0.5 h indicate a significant difference between ALSV-VC and ALSV-GmERA1 by t-test (*P < 0.05, **P < 0.01). (B) The surface temperature of detached fourth or fifth trifoliate leaflets of soybean plants inoculated with ALSV vectors was measured using thermography during the waterloss experiment shown in (A). Thermal images and photographs were taken 11 min after detachment. (C and D) Evaluation of leaf temperature (C) and stomatal conductance (D) in detached fourth or fifth trifoliate leaflets of soybean plants inoculated with ALSV vectors using a LI-6400XT portable photosynthesis system. Values are presented as mean ± SD (n = 3 for ALSV-VC and n = 4 for ALSV-GmERA1), and asterisks at 4, 8, and 12 min indicate a significant difference between ALSV-VC and ALSV-GmERA1 by t-test (**P < 0.01).
https://doi.org/10.1371/journal.pone.0175650.g002 plants. Next, we examined whether whole GmERA1-down-regulated soybean plants exhibit enhanced drought tolerance. When watering was withheld from soybean plants at the V6 growth stage [39] for 24 h, the leaf temperature was higher in plants inoculated with ALSV-G-mERA1N and ALSV-GmERA1C than in mock-or ALSV-VC-inoculated plants ( S8 Fig). However, during the first 24 h of dehydration, no wilting was observed among the plants (S8 Fig). We then scored leaf rolling in the fourth to sixth trifoliate leaves after approximately 60 h of dehydration stress treatment in six independent experiments (Fig 4A). At 60 h post-treatment, the leaf rolling scores of the plants were similar (S9 Fig). However, after 72-75 h of dehydration stress treatment, although ambient humidity could not be controlled under our experimental conditions, in five out of six independent experiments, the ratio of severely wilted leaves to total leaves (score 3) was lower in GmERA1-down-regulated soybean plants than in the control, indicating that in many cases, the GmERA1-down-regulated soybean plants exhibited less wilting than the control (Fig 4B). At one day after rehydration, the survival rates of the plants in each experiment were quite variable (Fig 4C, S3 Table). Nevertheless, taking all six independent experiments into account, approximately 60% of the GmERA1down-regulated soybean plants survived, whereas only approximately 30% of the control plants survived (Fig 4C, S3 Table). Overall, the GmERA1-down-regulated soybean plants survived dehydration stress better than did the vector control plants in terms of wilting, supporting the view that GmERA1-down-regulated soybean plants exhibit enhanced resistance to water-deficit conditions.

Discussion
Our data show that GmERA1-down-regulated leaves displayed reduced water loss rates under water-limiting conditions through ABA-mediated stomatal regulation and, consequently, that these plants exhibited enhanced resistance to water deficit conditions, consistent with previous studies of plants with similar levels of down-regulation of ERA1 homologs [35][36][37]. These findings demonstrate that GmERA1s function as negative regulators of ABA signaling in guard cells in soybean. Therefore, GmERA1s appear to play an important role in the dehydration stress response in soybean, as does ERA1 in Arabidopsis [29][30][31]35]. Our data support the hypothesis that ERA1 genes can be downregulated to increase sensitivity to ABA, thereby leading to more rapid ABA-induced stomatal closure under water-limiting conditions. Thus, our study suggests a useful strategy for enhancing drought resistance in soybean. Plants harboring mutations in GmERA1s could be readily identified through high-throughput methods, such as targeting-induced local lesions in genomes (TILLING), high-resolution melting (HRM), and indexed amplicon sequencing by next generation sequencing (NGS), taking advantage of high-density mutant resources that have already been developed in soybean [49][50][51].
This is the first report describing the successful use of VIGS in soybean to functionally characterize genes involved in drought stress resistance and ABA signaling. As shown in this and previous work [15,52], the greatest merit of the VIGS system is that it allows gene functions to be characterized easily and rapidly without the need to generate transgenic plants or to perform large-scale forward mutagenesis screening. However, this and previous studies also indicate that the leaf positions and growth stage in which gene functions can be analyzed are limited. Indeed, we mainly analyzed the fourth or fifth trifoliate leaves based on the spread of silencing symptoms caused by ALSV-GmPDS in soybean plants (Figs 1-3, S5 Fig). More   Table), which is consistent with previous reports showing that statistically significant results could not be obtained in VIGS experiments yielding plants with enhanced drought resistance at the whole-plant level [37,53]. Thus, the weakness of the VIGS system is the instability in the extent of temporal and spatial gene silencing due to variations in virus infection caused by environmental factors and plant growth, as described previously [47,54]. Improving virus infection rates and stable spread within an infected plant would increase the utility of the VIGS system.
Nonetheless, the weaknesses of the VIGS system described above can also be perceived as strengths. Although the growth of ERA1-repressed Arabidopsis plants is retarded [55][56][57][58], under our experimental conditions, no significant negative effect on plant growth was observed in the GmERA1-down-regulated plants (S7 Fig). As specific leaf positions and growth stages (e.g., the fourth trifoliate leaves at the V4 growth stage) are more readily recognized in soybean than in other staple crops such as wheat and Solanum lycopersicum (tomato) [39], the ALSV-VIGS system is ideal for validating gene function in soybean. Thus, our findings demonstrate that the ALSV-VIGS system is a useful tool for validating candidate genes without the need for generating transgenic plants; however, the selected candidate genes will need to be evaluated using appropriate mutant or transgenic plants to assess the impact on physiological parameters, such as soil water contents, before applying this technology in the field. Therefore, the ALSV-VIGS system, combined with recent methods for screening for desired mutations from high-density mutant resources and the CRISPR/CAS9-mediated introduction of desired mutations, could contribute to next-generation molecular breeding strategies to improve drought resistance in soybean. During the water-withholding test, the drought status of the fourth to sixth trifoliate leaves was scored as described in Fig 4. Leaf drought scores at 55-62 h after water withholding are shown for six independent trials. (PDF) S1