Increased Drought Tolerance through the Suppression of ESKMO1 Gene and Overexpression of CBF-Related Genes in Arabidopsis

Improved drought tolerance is always a highly desired trait for agricultural plants. Significantly increased drought tolerance in Arabidopsis thaliana (Columbia-0) has been achieved in our work through the suppression of ESKMO1 (ESK1) gene expression with small-interfering RNA (siRNA) and overexpression of CBF genes with constitutive gene expression. ESK1 has been identified as a gene linked to normal development of the plant vascular system, which is assumed directly related to plant drought response. By using siRNA that specifically targets ESK1, the gene expression has been reduced and drought tolerance of the plant has been enhanced dramatically in the work. However, the plant response to external abscisic acid application has not been changed. ICE1, CBF1, and CBF3 are genes involved in a well-characterized plant stress response pathway, overexpression of them in the plant has demonstrated capable to increase drought tolerance. By overexpression of these genes combining together with suppression of ESK1 gene, the significant increase of plant drought tolerance has been achieved in comparison to single gene manipulation, although the effect is not in an additive way. Accompanying the increase of drought tolerance via suppression of ESK1 gene expression, the negative effect has been observed in seeds yield of transgenic plants in normal watering conditions comparing with wide type plant.


Introduction
Drought stress is a major limiting factor for crop production worldwide [1]. In 2012, a severe drought in the United States caused heavy losses in crop production, especially in corn, and farmers produced less than three-fourths of the corn that the U.S. Department of Agriculture anticipated [2]. In China, around 20 million hectares of land are at risk of drought each year [3]. Globally, estimated crop losses due to water limitation exceed $10 billion annually [4]. Improving yield under drought, therefore, is a continuous challenge for agriculture, especially for modern breeding.
The development of modern plant biotechnology provides new hope for generating crops with increased drought tolerance. Understanding the response of plants to drought stress is the first step for development of stress tolerance plant through plant biotechnology. Gene expression experiments have identified several hundred genes that are either induced or repressed during drought. Some of those genes encode proteins that play important role in protecting cells from dehydration, such as the enzymes required for biosynthesis of various osmoprotectants, late-embryo-genesis-abundant (LEA) proteins, antifreeze proteins, chaperones, and detoxification enzymes [5]. Some others are responsible for gene products including transcription factors, protein kinases, and enzymes involved in phosphoinositide metabolism. C-repeat/ dehydration-responsive element binding factors (CBFs) are AP2/ ERF-type transcription factors, which make up a critical gene cluster of the second group. During in the stress condition, CBF genes are rapidly induced in response to abiotic stress, such as dehydration and cold [6,7]. The CBF proteins in turn activate expression of a set of target effector genes by binding to a core sequence in their promoter, C-repeat (CRT) / dehydration response element (DRE) [8,9,10,11]. CBF genes appear to be ubiquitous in plant species and almost always present as a gene family [12,13]. In Arabidopsis, the three characterized CBF genes are CBF1, CBF2, and CBF3, which are organized in tandem on chromosome 4 [14]. CBF1 and CBF3 are positive regulators whereas CBF2 has a negative regulatory role [15]. CBF transcription factor genes are induced by the constitutively expressed inducer of CBF expression (ICE1) by binding to the CBF promoter [16,17]. The ICE1-CBF cold response pathway is conserved in diverse plant species [14,17,18]. Constitutive overexpression of CBF transcription factors in transgenic plants has increased the plant tolerance to freezing, salt, and drought stresses [8,19,20,21,22,23]. This functional conservation has suggested the ICE1-CBF genes are important targets for crop improvement for drought tolerance through genetic engineering [24].
ESK1 is a newly discovered member of the second group. It was initially identified as conferring freezing tolerance; a significantly high proline content accumulates in esk1 mutants [25]. The gene product of ESK1 belongs to an uncharacterized plant-specific protein family containing 48 members [26]. Bioinformatics analysis of genes whose expression modified by the eskimo1 mutation showed that a large number of genes were previously reported linking to plant response to salt, osmotic stress, and the stress hormone abscisic acid (ABA) [25,26]. Later work showed that the mutant has a clear advantage in response to drought and salt stress: In standard and drought conditions, transpiration rate of mutant is lower than in wild type (WT) [27]. A biologically relevant parameter is the water required per biomass unit, and with this measure, the esk1 mutants clearly have shown a higher water use efficiency and photosynthetic rate compared to WT [27]. This higher water use efficiency was independent of stomata closure through ABA biosynthesis. Measurement of root hydraulic conductivity suggests that the esk1 vegetative apparatus suffers water deficit because of a defect in water transport system [27,28]. ESK1 promoter-driven reporter gene expression has been observed in xylem and fibers, the vascular tissue which is responsible for the transport of water and mineral nutrients from the soil to the shoots, via the roots. Moreover, in cross sections of hypocotyls, roots, and stems, collapsed xylem vessel has been observed in esk1 mutant [28]. The ESK1 gene, therefore, was inferred to play a major role in whole plant water economy. ESK1 has homologues in numerous species, and it is reasonable to hypothesize that manipulation of ESK1 in crops could improve water use efficiency.
With an understanding of the molecular mechanism, several gene manipulation approaches have been employed to increase plant drought tolerance. The manipulated genes include those encoding enzymes required for the biosynthesis of various osmoprotectants or enzymes for modifying membrane lipids, LEA protein, and detoxification enzyme [29,30,31]. To date, the CBF genes are the most explored genes for improving crop stress tolerance because of the ability of this transcription factor to regulate an entire set of genes in a stress-response pathway [8,32,33]. When rice OsDREB1A [34] or corn ZmDREB1A is constitutively overexpressed in Arabidopsis, the downstream target genes regulated by the Arabidopsis DREB1 (e.g., RD29A) are induced, resulting in desiccation tolerance under 15% humidity [35]. In another study, 35S: CBF1 transgenic tomato plants are more resistant to water-deficit stress by showing less plant wilting and leaf curling than WT controls after 21-d water withdrawal in the same pot [36]. Constitutive overexpression of two wheat CBF factors in barley substantially improves survival under severe drought or cold. In addition, expression of DREB factors in wheat and barley under the control of drought-inducible promoters allows for normal development, together with significantly improved survival under severe drought [37].
In spite of the extensive evaluation of CBF factors, only a few studies have shown a clear improvement in drought tolerance in crops under field conditions [38]. Considering the complexity of the plant stress response, it has been assumed that better drought tolerance might be obtained if multiple genes involved in different stress response pathways could be manipulated together through molecular stacking. In this work, we selected ICE1 as well as the CBF1 and CBF3 genes involved in one stress-activated pathway and the ESK1 gene from another stress-regulating pathway as targets of gene manipulation. Here, ICE1 or CBF1 and CBF3 genes were overexpressed under the control of a constitutive promoter whereas the ESK1 gene was suppressed with siRNA technique. Using gene stacking, we combined the two expression cassettes into one transformation vector for Agrobacteriummediated plant transformation. The results obtained by testing the concept in the model plant Arabidopsis showed a significant increase in drought tolerance compared to non-transgenics, indicating that multiple gene manipulation might be a promising strategy for improving stress response and especially drought tolerance in plants.

Manipulation of target gene expression through knock-down and overexpression
To knock down gene expression, a siRNA targeting specifically the ESK1 gene was designed, and to overexpress the desired ICE1 or CBF gene, the 35S promoter was placed in front of the gene for constitutive expression. The gene suppression and overexpression cassettes were integrated into the Arabidopsis genome, either alone or combined, via the flora dipping method. In analysis of gene expression by qRT-PCR, the plant transformed with siRNA cassette that suppresses ESK1 showed clear knock-down of ESK1 expression. Although the level of suppression varied from line to line, most plants showed about or more than 50% reduction in gene expression (Fig. 1A). In comparison, the transgenic lines derived via using only the pGPTV-ESKi vector showed a better suppression of ESK1 gene expression than those obtained through use of the vector with the stacked overexpression and suppression cassettes (Figs. 1A, C, D).
Analysis of either ICE1 or CBF overexpression showed the transgenic lines with the relevant transgene cassette had significantly higher ICE1 or CBF expression compared to WT (Figs. 1B-E). For further evaluation of the concept, the bestperforming plants based on qRT-PCR results were selected and their T2 seeds produced following the protocol described in the Methods part.
The response of transgenic Arabidopsis to osmotic stress in vitro A distinct difference emerged in in vitro stress response between transgenic and WT plants. When growing on medium without any applied external stress, the majority of transgenic lines showed an almost identical phenotype to WT plants. However, when growing on medium with 30% polyethylene glycol (PEG), the transgenic plants with suppressed ESK1 had much better root system growth than WT plants: The main root was longer, and the number of lateral roots was greater (Fig. 2). Leaf growth also differed between the transgenic and WT plants although not as much as the root system: All transgenic lines had eight leaves after 18 days of growth on medium with PEG whereas WT had only six. 15 ESKi transgenic lines were tested and the result of response to PEG was similar (data not shown).
Compared to the transgenic lines with the suppressed ESK1 alone, the plants obtained through using stacking vector that suppress ESK1 and overexpress either ICE1 (14 lines) or CBF (3 lines) presented a similar performance on medium with PEG. After 14 days of growth with PEG, more and larger leaves were observed in transgenic lines compared to WT (Fig. 2). The growth  The responses to in vitro osmotic stress observed from transgenic plants with only the overexpression for CBF (5 lines), or ICE1 (4 lines) were also similar to those with ESK1 suppressed (Fig. 2). On normal 1/2 MS medium without PEG, no phenotype difference was observed between transgenics and WT control. On PEG medium, however, most transgenic plants clearly showed better growth than control plants after 14 days: The development of the root system from plants with CBF or ICE1 overexpression was more robust, with a longer main root and more lateral roots. Nevertheless, several lines of ICE1 transgenic plants showed a stress phenotype like that of WT, although the ICE1 expression in those plants was much higher than in WT.

The response of transgenic plants to ABA stress in vitro
Suppression of ESK1 gene expression did not significantly influence the sensitivity of plants to external ABA application. On medium without additional ABA, transgenic plants containing ESK1 suppression cassette appeared almost identical to WT plants in their growth (Fig. 3A). On medium with 20 mM ABA for 14 days, the two still did not differ, and both appeared to suffer the effects of growth on ABA, showing leaves with yellow-brown stress symptoms (Fig. 3B).
The only difference observed between transgenic lines and WT plants was in seed germination time. On medium containing 0.5 mM ABA (Fig. 4B), the transgenic seeds showed delayed germination compared to WT control whereas on medium without ABA (Fig. 4A), the transgenic seeds germinated at the same time as WT.
Similar results were obtained for plant overexpression of CBF or ICE1 in response to ABA stress in vitro (Figs. 3, 4). On medium without ABA, both transgenic and control plants showed similar growth with no abnormal phenotype observed in the majority of plants during the first 18 days of growth.
On medium with 20 mM ABA, both transgenic and WT plants were negatively affected. The elongation of the main root was inhibited, the number of lateral roots was reduced, and the leaves showed yellow-brown stress symptoms. No clear differences between transgenic lines and non-transgenic controls were observed (Fig. 3).

The performance of greenhouse transgenic plants in response to drought
To mimic most completely the interaction of plants and their natural environment, the transgenic plants showing the potential to increase drought tolerance via the in vitro osmotic test were transferred to greenhouse for evaluation of drought tolerance. The testing protocol was established based on the water-loss rate in greenhouse and the survival rate of plants in the protocol of water withdrawal and re-watering. Plants growing 2 weeks after germination in the described conditions were withdrawn from water for 14 days, so that the soil water content reached around 20%, which is considered a serious drought condition [27]. Following this threshold point, the plants were then fully watered and grown under a normal watering program for two more weeks before survival rates were evaluated.
Non-transgenic plants usually died completely under this protocol; however, transgenic plants obtained by the vector pGPTV-ESKi alone showed dramatic improvement in drought tolerance in contrast to the WT control plant (Fig. 5A). After 14 days without water, the transgenic plants still appeared green although with some level of dryness (Figs. 5A, B). Two weeks after re-watering, many of the transgenic plants had re-gained growth, and the survival rate of the best-performing line esk1i-3 reached 80% (Fig. 5C). In comparison, almost all WT plants became yellow and dry after 14 days without water, seldom regaining growth after water restoration (survival rate 1.25%, Fig. 5C).
Transgenic plants with ICE1 or CBF gene overexpression vector alone also showed increased drought tolerance under greenhouse conditions (Figs. 5B-E). After 14 days without watering, the transgenic plants still remained green and recovered soon after the re-watering; the non-transgenic controls did not recover. Among the transgenic lines with various vectors, the line transformed with the ICE1 overexpression vector showed relatively better drought tolerance than those transformed with CBF overexpression vector (Figs. 5C, D, E).
The improved drought tolerance test response under greenhouse conditions was also achieved in transgenic plants containing the combined ESKi cassette and overexpression cassette, especially those with the combination of the ESKi and ICE1 cassettes; the highest survival rate was obtained from the line transformed with ESKi-ICE1 stacking vector at 85.25% (Fig. 5C).
To compare the effects of different vectors on improvement of drought tolerance, the transgenic plants were divided into groups according to the vectors used, and the survival rate data was analyzed using t-tests. The results showed that the plants performing best were those transformed with ESKi-ICE1 stacking vector; the lines transformed with the ESKi vector showed the second-best improvement while improvement was weakest with overexpression of the CBF genes (Figs. 5D, E).

The effect of transgenes on seed biomass
To evaluate potential applications to agriculture, we investigated the effect of transgenes on plant biomass, especially seed mass, under greenhouse conditions, either normal growth conditions or drought test conditions.
Under normal growth conditions in the greenhouse, some ESK1 siRNA transgenic plants showed a dwarf phenotype with dark leaves (Fig. 6Aa); a delay in flowering was also observed, but most of the plants, otherwise, looked normal. Measuring the mass of seed harvested under normal greenhouse conditions revealed that the ESK1 siRNA transgenic plants produced less seed than non-transgenic controls. However, upon harvest of the seed from plants that survived the drought test procedure, a clear contrast was evident between the seed yield of transgenic plants and controls, given that most of the non-transgenic plants died of drought. Comparing the seed yield of transgenic plants under drought and normal conditions, a significant reduction was found under drought, with less than 10% of the yield under normal conditions (Fig. 6B).
The impact of transgenes on seed yield showed a similar trend in plants transformed with ICE1 or CBF vector alone and in those transformed with stacking vectors (Fig. 6B). Some dwarf phenotype development was observed during growth under normal conditions whereas in the drought condition, a higher stress tolerance compared to WT was observed, although accompanied by severe seed yield reduction.

Discussion
Plant genetic engineering has presented great potential for improving the drought tolerance of crops, especially since the discovery of CBF genes and their functions in stress tolerance [8,19,20,21,22,23]. ESK1 was recently found to be involved in a stress tolerance pathway that is separate from the CBF pathway [25,26]. The ESK1 pathway originally was identified as being related to plant freezing tolerance, but Bouchabke-Coussa et al. [27] later observed reduced respiration capacity in the eskl mutant of Arabidopsis, indicating a potential role of ESK1 in drought tolerance. In our experiment, siRNA specifically targeting ESK1 was employed to suppress its expression in Arabidopsis and the role of ESK1 was confirmed in drought tolerance: Transgenic plants with efficient ESK1 expression knock-down showed an obvious increase in osmotic stress tolerance in vitro and little effect on root system development by PEG treatment in vitro (Fig. 2). Further evaluation confirmed that suppression of the ESK1 gene significantly enhanced plant drought tolerance (Fig. 5). In ABA vitro assay, it has been observed that ESK1 knock-down transgenic plants has the same sensitivity to ABA change as control (Figs. 3B and 4B), which indicates that acquired stress tolerance of the transgenic plant may not be closely linked to the ABA-regulated network.
The ESK1 knock-down plants appeared phenotypically normal comparing to non-transgenic plants when grown for 18 days: Plant size, leaf color, and shape were basically same like WT under normal growth conditions (Fig. 3A). Morphological observation also revealed that the number of stomata in the leaves remained unchanged relative to controls (data not shown). The question that arises is how ESK1 knock-down confers higher tolerance to drought or other stress. One explanation is that the disruption in ESK1 function may block the normal development of the vascular system. Lefebvre et al. [28] observed a kind of collapsed xylem structure in cross sections of hypocotyls, roots, and stems of an Arabidopsis esk1 mutant. In the current study, we also observed that the vascular system appeared abnormal in both leaf and stem (data not shown). These observations may explain why the respiration rate of a transgenic plant is lower compared to WT and why drought tolerance is higher than in controls.
Measuring the weight of total seeds showed that knocking down the ESK1 gene can negatively affect seed production under normal growth conditions but that under the drought condition, the yield is much less affected, in contrast to complete loss in the non-transgenic control (Fig. 6B). Linking the yield penalty and the vascular distortion of the ESK1 knock-down plant, we can infer that manipulation of ESK1 gene expression can indeed increase the drought tolerance of plants, but expecting no reduction in biomass is not realistic considering the physical distortion of the vascular structure. Adoption of an inducible promoter or tissuespecific promoter may alleviate some negative consequences, but the enhancement of drought tolerance may be compromised at the same time. Therefore, use of ESK1 for stress tolerance should be carefully balanced. One potential application may be in improvement of ornamental plants, for which biomass is not important.
Many studies have illustrated the potential of manipulating CBF/DREB genes to confer improved drought tolerance [20,21,37,39]. For example, overexpression of CBF1/DREB1B from Arabidopsis improves tolerance to water-deficit stress in tomato, but few plants clearly show enhanced drought tolerance under natural conditions [36]. We speculate that the complexity of the stress response pathway could be the reason and that modification of a single gene in a complicated pathway might not be sufficient to alter plant drought tolerance dramatically. Gene discovery and functional genomics projects have revealed many mechanisms and gene families that confer improved adaptation to abiotic stresses. These gene families can be manipulated into novel combinations, expressed ectopically, or transferred to species in which they do not naturally occur or vary [40,41,42]. Therefore, we have designed a gene stacking strategy by combining manipulation of the CBF and ESK1 genes, hoping to obtain at least an additive effect on drought tolerance.
The result showed that gene stacking can indeed further improve drought tolerance in Arabidopsis, but an additive effect was not observed. One possibility for the lack of additive effect is that genes stacked in the same transformation vector may not work as efficiently as those in completely independent transformation vectors. qRT-PCR analysis in our work showed that the suppression effect of siRNA targeting ESK1 was less significant when the same vector was used for the stacked RNAi cassette and the overexpression cassette. In the future, the effect of combined gene manipulation will be evaluated through stacking the gene cassette by crossing the transgenic lines with individual vectors. For plate-grown plants, Arabidopsis thaliana seeds were surface sterilized with 20% (v/v) bleach and sown on medium containing half-strength Murashige and Skoog (MS) salts [43], 1% sucrose, and 0.8% agar. After stratification for 3 d at 4uC, the plates were kept in a growth incubator under a long-day photoperiod (16 h light, 8 h darkness) at 24uC for 10 d. For transgenic seeds, medium was supplemented with 50 mg/mL kanamycin sulfate.

Gene isolation and binary vector construct
The leaf of 2-week-old Arabidopsis Col-0 was used for DNA and RNA extraction. The Col-0 cDNA sequences were obtained from GenBank (http://www.ncbi.nlm.nih.gov/). Total DNA was extracted from the samples using the DNeasy Plant Mini Kit (Qiagen, Germany), and total RNA was extracted using the RNeasy Plant Kit (Qiagen, Germany). cDNA was synthesized as described in the Thermo Scientific protocol (#K1631). The AtCBF1 (AT4G25490) and AtCBF3 (AT4G25480) were amplified from Arabidopsis Col-0 DNA by PCR. AtCBF1 was amplified with a forward primer containing a BamHI restriction site at the 59 end (59-CGGGATCCCTCTGATCAATGAACT-CATT-39) and a reverse primer containing a SacI site at the 39 end (59-GCGAGCTCTTAGTAACTCCAAAGCGACA-39). AtCBF3 was amplified with a forward primer containing an ApaI site at the 59 end (59-GGGCCCGATCAATGAACT-CATTTTCTGC-39) and a reverse primer containing a XbaI site at the 39 end (59-GCTCTAGATTAATAACTCCATAACGA-TACGTCG-39).

Agrobacterium-mediated gene transformation
The floral dip method [44] was applied to stably transform the Arabidopsis plant by using the Agrobacterium strain ATHV containing the designed binary vector. Seeds from treated plants were germinated on media containing 1/2 MS salts, 1% sucrose, 0.8% agar and 50 mg/mL kanamycin. Resistant seedlings were transferred to soil and grown under 16 h light /8 h dark at 24uC in a growth chamber. After PCR confirmation, positive seedlings were used to produce T2 and T3 seeds, which were always subjected to kanamycin selection and PCR confirmation; only the positive plants were used for seed production of the next generation. The kanamycin-tolerant and PCR-positive T3 or T4 plants were used in all experiments. Ecotype Col-0 served as control.

Quantitative real-time PCR (qRT-PCR)
Total RNA from different Arabidopsis plants was isolated using the RNeasy plant mini kit (Qiagen, Germany) according to the manufacturer's instructions. To eliminate any residual genomic DNA, total RNA was treated with ribonuclease-free DNase I (Thermo Scientific). Two micrograms of the total RNA was used as template to synthesize cDNA employing the RevertAid H Minus First Strand cDNA Synthesis Kit (Thermo Scientific). Realtime PCR was performed according to Kant et al. [45]. The PCR reaction was performed with three replicates and repeated with three biological samples. Relative quantification values for each target gene were calculated by the 2 (-Delta Delta C(T)) method [46]. For normalizing the amount of total RNA in all Arabidopsis samples, Actin2 (AT3G18780) and beta tubulin (AT5G23860) were used as internal reference genes to compare data from different PCR runs or cDNA samples. The GenScript online tool was used to design the primers (https://www.genscript.com/), and primers used in q-PCR are listed in Supplemental Table S1. In vitro assay for plant response to ABA and PEG For germination, seeds were planted on the plate containing 1/ 2 MS salts, 1% sucrose, and 0.8% agar with 0 or 0.5 mM ABA as indicated. Plates were chilled at 4uC in the dark for 3 d and moved to 24uC with a 16 h light/8 h dark cycle. The percentage of seed germination was scored at the indicated times. Germination was defined as an obvious emergence of the radicle through the seed coat [47].
To study the inhibition effect of ABA in post-germinative growth, seeds were sown on 1/2 MS medium with 50 mg/mL kanamycin for 4 d after 3 d stratification, then transferred to 1/2 MS medium containing 0 or 20 mM ABA.
For response to in vitro dehydration, 3-day-old seedlings were transferred to 1/2 MS medium previously infused with 30% PEG8000 for 14 d [48]. After growing for 14 d on the treatment medium vertically, seedlings were photographed with a digital camera.
15 ESKi transgenic lines, 14 eski-ICE1 transgenic lines, 3 eski-CBF transgenic lines, 5 CBF transgenic lines and 5 ICE1 transgenic lines were used in both PEG and ABA in vitro assay. For each treatment, 30 plants of each line have been subjected to the treatment. Each test has been repeated three times. The influence of the tested element was concluded based the result pooled together from three repeats.
Greenhouse test for drought tolerance 15 ESKi transgenic lines, 14 eski-ICE1 transgenic lines, 3 eski-CBF transgenic lines, 5 CBF transgenic lines and 5 ICE1 transgenic lines, which showed desired gene manipulation result in qRT-PCR, were used for drought tolerance test. All seeds from transgenic lines and Col-0 together were sown in 40 5 cm65 cm pots. Plant growth conditions were as described [49]. All seeds were cold-treated (4uC) for 3 d immediately after planting to ensure uniform germination. Plants were grown in controlled environment chambers at 24uC under sunshine illumination or supplementary illumination from cool-white fluorescent lights (100-150 mmol m 22 s 21 ) and irrigated with water three times a week, ensuring soil saturation; the last watering was on day 14 after germination. Two-week-old seedlings were used for drought treatment, in which watering was withdrawn for two weeks. Afterwards, all plants were re-watered according to the predrought treatment. Tolerance to drought was determined by the capacity of plants to resume growth after 7 d of re-watering. The test was repeated at least three times. The data obtained was subjected to statistical analysis described as below.

Measurement of drought effect on seed biomass
14 ESKi transgenic lines, 12 eski-ICE1 transgenic lines, 3 eski-CBF transgenic lines, 4 CBF transgenic lines and 3 ICE1 transgenic lines were used for the test. At least 30 plants from each transgenic line and wild type were exposed to both control and drought treatment described as above. After the treatment, the survived plants were grown in normal water condition for seed setting. The seeds were harvested by bag that had been covered to plant in reproductive period. The harvested seeds were cleaned and then dried at room temperature. The weight of seed was measured after 2 weeks drying. The average weight (total harvested seed divided by the number of tested plant) was used as result to compare the drought effect on each line of plant. The data was analyzed based the statistical methods described below.

Statistics
The results shown are representative of three independent experiments, and within each experiment, treatments were replicated three times. Data were statistically analyzed using SigmaPlot 10.0 and GraphPad Prism 5.

Supporting Information
Table S1 Primers used in molecular analyses. (DOCX)