Gene transcription and steviol glycoside accumulation in Stevia rebaudiana under polyethylene glycol‐induced drought stress in greenhouse cultivation

Stevia rebaudiana is a sweet herb of the Astraceae family, which is cultivated for the natural sweeteners it contains. The aim of this study was to assess the effect of drought, simulated by the application of polyethylene glycol (5%, 10%, and 15% w/v), on the content of steviol glycosides (SVglys) and transcription levels of six genes involved in the biosynthesis of these natural sweeteners. The transcription levels of ent‐kaurene synthase, ent‐kaurene oxidase, ent‐kaurenoic acid hydroxylase, and three UDP‐dependent glycosyltransferases, UGT85C2,UGT74G1 and UGT76G1 were downregulated under polyethylene glycol treatment. Polyethylene glycol treatment significantly decreased the amount of stevioside, rebaudioside A, B, C and F, steviolbioside, dulcoside A, rubusoside, and total SVglys. These results strongly suggest a close relationship of SVglys content with the transcription of genes involved in the SVglys biosynthesis pathway. Comparing the observations of the present study with other reports provided the knowledge that the Stevia response to drought stress can be influenced by different environmental and experimental factors, in addition to intensity of drought stress. In conclusion, these results strongly suggest that polyethylene glycol‐induced drought stress has a negative effect on the content of SVglys and transcription of SVglys biosynthetic genes and that this should be investigated further. We recommend that sufficient irrigation of Stevia is required to obtain a high content of SVglys.

Stevia rebaudiana is a sweet herb of the Astraceae family, which is cultivated for the natural sweeteners it contains. The aim of this study was to assess the effect of drought, simulated by the application of polyethylene glycol (5%, 10%, and 15% w/v), on the content of steviol glycosides (SVglys) and transcription levels of six genes involved in the biosynthesis of these natural sweeteners. The transcription levels of ent-kaurene synthase, ent-kaurene oxidase, ent-kaurenoic acid hydroxylase, and three UDPdependent glycosyltransferases, UGT85C2, UGT74G1 and UGT76G1 were downregulated under polyethylene glycol treatment. Polyethylene glycol treatment significantly decreased the amount of stevioside, rebaudioside A, B, C and F, steviolbioside, dulcoside A, rubusoside, and total SVglys. These results strongly suggest a close relationship of SVglys content with the transcription of genes involved in the SVglys biosynthesis pathway. Comparing the observations of the present study with other reports provided the knowledge that the Stevia response to drought stress can be influenced by different environmental and experimental factors, in addition to intensity of drought stress. In conclusion, these results strongly suggest that polyethylene glycol-induced drought stress has a negative effect on the content of SVglys and transcription of SVglys biosynthetic genes and that this should be investigated further. We recommend that sufficient irrigation of Stevia is required to obtain a high content of SVglys.
Stevia rebaudiana Bertoni is a perennial herb belonging to the Asteraceae family. It is a natural sweetener plant known as "Sweet Weed", "Sweet Leaf", "Sweet Herb", and "Honey Leaf", which is estimated to be 30 times sweeter than cane sugar. Leaves of S. rebaudiana accumulate up to 30% of steviol glycosides (SVglys), the compounds responsible for its sweet taste [1]. SVglys are used as a calorie-free sweetener in many countries of the world including China, Japan, Korea, Australia, New Zealand, and many countries of the European Union. Stevia is a medicinal plant with a huge demand in pharmaceutical, food and beverage industries as a source of low-calorie and high-potency natural sweeteners. Many studies showed that SVglys have different beneficial pharmacological effects, such as lowering of blood pressure, regulation of blood glucose in type 2 diabetes while enhancing insulin sensitivity, prevention of atherosclerosis, prevention of certain forms of cancers and as a free radical scavenger [2][3][4].
Plants frequently encounter unfavorable growth conditions. They are challenged by stressful environmental conditions like water deficit, which leads to changes in molecular, biochemical, and physiological processes. Consequently, plant growth and development are adversely affected [12,13]. The responses are regulated by multiple signaling pathways which result in changes in the patterns of gene expression. Stevia rebaudiana originates from Paraguay, a semihumid area with a yearly average rainfall of 1500 mm. Now cultivation of Stevia has spread to other regions of the world, including Canada and some regions of Asia and Europe. One critical point in Stevia cultivation is sufficient irrigation to obtain greater profits. According to the finding of Hajihashemi and Ehsanpour [14], S. rebaudiana showed little resistance to in vitro polyethylene glycol-induced drought stress under controlled conditions. Polyethylene glycol negatively affected plant growth, photosynthetic pigments, proteins, enzymes [14,15]. Also polyethylene glycol-induced drought stress under in vitro condition negatively affected the transcription of ent-KO, UGT85C2 and UGT76G1 genes [16]. Our previous study was carried out under in vitro condition which is an artificial medium supplemented with sucrose. Therefore, it was imperative to know how Stevia reacts to polyethylene glycol treatment under controlled greenhouse conditions which are close to field culture. Understanding how plants respond to drought stress can play an important role in improving Stevia management and performance, especially as the climate-change scenarios suggest an increase in aridity in many areas of the globe.
To make the first step toward the production of drought tolerant Stevia plants, the expression of some related genes should be analyzed under drought stress. Polyethylene glycol treatment under greenhouse culture provide us compelling evidence to heighten awareness of the mechanism by which the vital step of SVglys biosynthesis is regulated by drought stress and its correlation with gene transcription. The objective of this study was to (a) measure SVglys content (e.g., ST, Reb A, B, C, F, SB, Dul A, Rub, and total SVglys) and (b) distinguish the transcription of ent-KS, ent-KO, ent-KAH, UGT85C2, UGT74G1, and UGT76G1 genes to understand the correlation between genes transcription and SVglys accumulation in polyethylene glycol-treated S. rebaudiana.

Materials and methods
The S. rebaudiana plants were provided by the Laboratory of Functional Biology, KULeuven University, Leuven, Belgium. Stevia propagation was carried out by tissue culture. The in vitro propagated plants were grown to the seedling stage under normal greenhouse conditions with supplementary light to extend the photoperiod to 16 h per day. Air temperature ranged from 22°C to 26°C during the day and 15°C to 18°C during the night. Humidity range was between 40% and 60%. Two-month old plants of similar size were transferred to 1.5 L pots filled with perlite and vermiculite. The plants were irrigated every other day and fertilized once a week by nutrient solution (pH 6.0). The 6month old plants were irrigated with 0% (distilled water), 5%, 10%, and 15% w/v polyethylene glycol (molecular weight 6000; À0.08, À0.49, À1.4, À2.9 MPa, respectively) solutions. Polyethylene glycol treatment continued for 1 month. All plants were harvested 2 days after the last day of polyethylene glycol treatment to be used for further analysis. All plant samples were immediately frozen in liquid nitrogen and stored at À80°C until further use. Some of frozen leaves were freeze-dried for SVglys analysis. The experiment was designed as a completely randomized block. Each experiment was repeated three times at different periods. Each experiment consisted of six replicates per treatment.

Water extraction of leaves
Twenty mg of freeze-dried and pulverized leaves were extracted three times consecutively by boiling, each sample three times for 15 min in 300 lL HPLC quality water. The extracts were mixed, pooled, centrifuged, and filtered to remove particles. To avoid possible losses, no purification step was used. The extracts were stored at 4°C for further use.

Gene transcription analysis using RT-qPCR
The frozen leaves were used to extract total RNA using the Plant RNA Mini Kit (Thermo Fisher Scientific, UK), according to the manufacturer's instructions. The amount of 2.3 lg of total RNA was reversely transcribed with 50 U AMV-reverse transcriptase in a total volume of 25 lL of Master Mix, for cDNA synthesis. Master mix was a mixture of 2.5 lM oligo (dT) 16 primer, 20 U RNase inhibitor, 19 PCR buffer, 5 mM MgCl 2 , and 1 mM of each dNTP (GeneAmp RNA PCR kit; Roche Molecular Systems, Branchburg, NJ, USA). The house-keeping genes of b-Actin and 18S rRNA were used in this experiment. Oligo-Calc software was used to design primers for both target and house-keeping genes [18] (Table 1).
SYBR green binds to double-stranded DNA and the amplification products emit fluorescence. Real-time quantitative PCR based on the emitted fluorescence, to quantify the target transcript accumulation relative to the housekeeping gene transcript. A 96-well plate was used to carry out the SYBR green one-step RT-qPCR. The PCR reaction was containing 2.5 lL RNase-free water, 0.2 lL of each primer (10 lmolÁL À1 ), 5 lL SYBR green mix, 0.1 lL CXR and 2 lL of cDNA (5 ngÁlL À1 ), in a total volume of 10 lL. The reactions were run in triplicate in an ABI prism 7000 sequence detection system (ABI Prism 7000 SDS; Thermo Fisher Scientific). The thermal cycles of 95°C for 20 s, followed by 40 cycles of 95°C for 3 s and 62°C for 30 s. Melting curve (dissociation) analysis (60-95°C) was done to verify amplicon specificity after 40 cycles at 95°C for 15 s, 62°C for 1 min and 95°C for 15 s. b-Actin in addition to 18S rRNA were used to normalize the data individually. According to the 2 ÀDDCT method, the abundance of targeted gene transcripts was set relative to control plants (with no treatment) [19]. STEP ONE software (version 2.1; Thermo Fisher Scientific) was used to analyze the data of RT-qPCR.

Statistical analysis
A Randomized Complete Block Design was used for all experiments with three biological and technical replications. The Duncan test's SPSS (version 16; SPSS, New York, NY, USA) statistical package was applied to calculate the significant differences (at the 5% level) between means. The results of statistical analysis are shown by superscripted letters after the numbers in tables to reveal significant differences. Figure 1 shows the Stevia plants under different treatments of polyethylene glycol. The plants were treated with 0%, 5%, 10%, and 15% of polyethylene glycol which resulted in smaller plants. The SVglys content of leaves was analyzed in the treated plants which Table 1. List of primers used in RT-qPCR and house-keeping genes. Kaurene synthase (ent-KS), kaurene oxidase (ent-KO), kaurenoic acid hydroxylase (ent-KAH), UDP-dependent glycosyltransferases of UGT85C2, UGT74G1, and UGT76G1.

Gene
Primer sequence 5 0 ?3 0 (forward/reverse) Amplicon length (bp) Accession number According to results from SVglys analysis and some other preliminary experiments (data not shown), we decided to do the RT-qPCR analysis on plants treated with 10% of polyethylene glycol. The house-keeping genes of b-Actin or 18S rRNA were used to normalize the results of RT-qPCR. The transcription pattern of both house-keeping genes was similar. Samples had a good purity because no signal amplification was observed in the negative control sample. The values were reported relative to b-Actin because the expression patterns of samples relative to b-Actin or 18S rRNA were almost similar. Also, the melting curves of all samples were the same.  Figure 2 shows the transcription patterns of investigated genes in plants subjected to polyethylene glycol treatment. A significant reduction was observed in the transcription of ent-KS1, ent-KO, and ent-KAH genes under polyethylene glycol treatment by about 31%, 67%, and 30%, respectively (Fig. 2). The UGT74G1, UGT76G1, and UGT85C transcription significantly decreased in polyethylene glycol-treated plants by about 19%, 41%, and 50%, respectively (Fig. 2). Ent-KO showed the greatest decrease in gene transcription among the six investigated genes. In addition, the adverse effect of polyethylene glycol on the transcription of three genes (ent-KO, UGT85C2, and UGT76G1) was more than that of the other three genes (ent-KS1, ent-KAH, and UGT74G1).

Discussion
Stevia rebaudiana is a source of new sweeteners with a huge economic potential because of its SVglys accumulation (Fig. 3). Therefore, it is a critical point to study the elements which may negatively affect SVglys accumulation. Plant growth and secondary metabolism are commonly regulated by external cues such as light, temperature, and water availability [21,22]. Drought stress is one of the critical limiting factors, adversely affecting Stevia plant growth and performance [14][15][16]23]. Evidently, the negative effect of polyethylene glycol on plant growth was observed at high level of polyethylene glycol in the present study (Fig. 3). In this study, the changes of SVglys contents and the transcript levels of some corresponding biosynthetic genes were investigated under different polyethylene glycol treatment. The discovery of reduction in SVglys contents in polyethylene glycol-treated plants can be correlated with the downregulation of the transcription of ent-KS1, ent-KO, ent-KAH, UGT74G1, UGT76G1, and UGT85C2 genes, which are involved in SVglys biosynthesis pathway. The inhibition of ent-KS1 and ent-KO transcription can reduce ent-Kauronic acid which is the precursor of SVglys. According to Kumar et al. [24], a positive correlation can be detected between the transcription of ent-KO transcription and SVglys content, which is supported by the results from the present study. The reduction in plant height under polyethylene glycol treatment can be contributed to the downregulated transcription of ent-KS and ent-KO genes because ent-KS and ent-KO enzymes catalyze gibberellin biosynthesis (Fig. 1). Paclobutrazol treatment decreases plant growth via blocking GA biosynthesis, more specifically at the step of ent-kaurene oxidation [25,26]. It is known that under in vitro condition [16], ent-KO transcription significantly decreased in the polyethylene glycol-treated plants while no changes were observed in the transcription of ent-KS1. Yang et al. [27] reported that the transcription of ent-KS1 and ent-KO did not change under drought stress. Another conflict was observed about the transcription of ent-KAH. Polyethylene glycol treatment under in vitro culture had no significant effect on the transcription of ent-KAH [16], but in the present study, a significant reduction was observed in it.
Glycosyltransferases catalyze the transfer of a glycosyl moiety from an activated donor to an acceptor molecule, forming a glycosidic bond [28]. In S. rebaudiana, UGTs trigger the downstream steps in the synthesis of steviol glycosides [1]. The transcription of UGT85C2 and UGT76G1 genes showed a significant reduction in polyethylene glycol treatment which is supported by other reports [16,27]. UGT85C2 enzyme forms steviolmonoside by adding a C-13 glucose to steviol [18]. Thus, the significant correlation between total SVgly accumulation and transcription of UGT85C2 suggests the UGT85C2 enzyme as a ratelimiting step of SVgly biosynthesis. UGT76G1 modulates the synthesis of stevioside and rebaudioside A [9,10]. In the present study, the transcription of UGT74G1 was significantly reduced in polyethylene glycol-treated plants, which is supported by the results of Yang et al. [27], while it is opposed to Hajihashemi et al. [16].
It is known that the plant response to drought stress is related to the time and intensity of stress, plant species, genotypes, and environmental conditions [29]. The observed variations in the gene transcription of Stevia in different experiments can be correlated with different methods to induce drought stress, severity of drought stress, plant growing stages, plant cultivar, harvesting time, etc. Hajihashemi et al. [16] induced drought stress in Stevia plant by adding 5% polyethylene glycol to MS culture medium while in the present study the gene transcription was studied in plants treated with 10% of polyethylene glycol. Yang et al. [27] induced drought stress by stopping watering the pots until the plants wilted. The study of growing stages revealed changes of SVglys contents and the transcription of genes at different growing periods [27]. It was observed that moderate water-deficit stress (8-days irrigation period) did not significantly affect the SVglys content [30]. Accordingly, Karimi et al. [31] reported that SVglys content increased when soil moisture was depleted to 60% FC (9-day irrigation interval), but it significantly decreased under severe drought stress (45% FC). The harvesting time of Stevia affected the yield and quality of SVglys content in leaves [32]. In conclusion, it can be strongly suggested that the Stevia response to drought stress was influenced not only by the intensity of drought stress but also by environmental cues which have a major influence on the transcript levels of genes involved in SVglys biosynthesis and SVglys accumulation in S. rebaudiana. However, it should be noticed that Stevia cultivation need sufficient irrigation to obtain greater SVglys and profits.

Conclusion
The response of S. rebaudiana Bertoni to different polyethylene glycol treatments outlined in this study confirms the negative effects of polyethylene glycolinduced drought stress in Stevia plant. SVglys content as a critical issue in S. rebaudiana was adversely affected by polyethylene glycol treatment which is in a close relationship with the transcription of SVglys biosynthesis genes. These results provide insight into the mechanism by which SVglys accumulation is regulated by gene transcription. In present investigation, the downregulation of ent-KS1, ent-KO, ent-KAH, UGT74G1, UGT76G1, and UGT85C2 transcription led to a reduction in SVglys content. These findings provided the awareness that severity of drought stress, environmental cues, and methods of treatment resulted in a variation in gene transcription. It indicates that the genes included in the same biosynthesis pathway do not display the same reaction to stress and that this complicates the understanding of the mechanisms. Polyethylene glycol treatment inhibited the ent-KO, UGT85C2, and UGT76G1 in both the present study and our previous study [16]. We suggest that the genes involved in SVglys biosynthesis pathway might be classified into two groups of early-and late-induced genes. The three genes, ent-KO, UGT85C2, and UGT76G1, can be classified as early-induced genes in response to polyethylene glycol treatment, with the other genes (ent-KS1, ent-KAH, and UGT74G1) only engaged by increasing the intensity of polyethylene glycol-induced drought stress and thus classified as late-induced genes. The three genes of ent-KS1, ent-KAH, and UGT74G1 might be classified as late-induced genes. Overall, these findings provide new insights into the droughtresponse mechanisms in Stevia plants. This study on gene transcription in S. rebaudiana opens the way for further research to confirm these initial conclusions.