Impact of blue, red, and far-red light treatments on gene expression and steviol glycoside accumulation in Stevia rebaudiana

Phytochemistry 137 (2017) 57e65

Contents lists available at ScienceDirect

Phytochemistry journal homepage: www.elsevier.com/locate/phytochem

Impact of blue, red, and far-red light treatments on gene expression and steviol glycoside accumulation in Stevia rebaudiana Yuki Yoneda, Hiroshi Nakashima, Juro Miyasaka, Katsuaki Ohdoi, Hiroshi Shimizu* Graduate School of Agriculture, Kyoto University, Sakyo-ku, Kyoto, 606-8502, Japan

a r t i c l e i n f o

a b s t r a c t

Article history: Received 10 August 2016 Received in revised form 1 February 2017 Accepted 3 February 2017 Available online 16 February 2017

Stevia rebaudiana (Bertoni) Bertoni is a plant that biosynthesizes a group of natural sweeteners that are up to approximately 400 times sweeter than sucrose. The sweetening components of S. rebaudiana are steviol glycosides (SGs) that partially share their biosynthesis pathway with gibberellins (GAs). However, the molecular mechanisms through which SGs levels can be improved have not been studied. Therefore, transcription levels of several SG biosynthesis-related genes were analyzed under several light treatments involved in GA biosynthesis. We detected higher transcription of UGT85C2, which is one of the UDP-glycosyltransferases (UGTs) involved in catalyzing the sugar-transfer reaction, under red/far-red (R/ FR) 1.22 light-emitting diodes (LEDs) and blue LEDs treatment. In this study, it was demonstrated that transcription levels of SG-related genes and the SGs content are affected by light treatments known to affect the GA contents. It is expected that this approach could serve as a practical way to increase SG contents using specific light treatments. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Stevia rebaudiana Asteraceae Light treatments Gibberellin UDP-glycosyltransferases Steviol glycosides Phytochrome Cryptochrome Shade-avoidance responses

1. Introduction Stevia rebaudiana (Bertoni) Bertoni is a perennial plant that belongs to the Asteraceae family and is native to Brazil and Paraguay. Its principal sweetening components are steviol glycosides (SGs), which can confer strong sweetness and are mainly extracted from leaves (Singh and Rao, 2005). S. rebaudiana is used by diabetic patients as a diet therapy, and its extracts exhibit pharmacological effects, such as anti-insulin resistance, insulin secretion promotion, anti-hypertensive and anti-obesity (Chen et al., 2005, 2006; Dyrskog et al., 2005) properties. SGs are diterpene secondary metabolites and share their biosynthesis pathway with gibberellins (GAs) through the formation of ent-kaurenoic acid (Fig. 1 i). Steviol (Fig. 1 ii) serves as a basic skeleton to which bglucose is added to advance SG biosynthesis (Geuns, 2003). According to Geuns (2003), the contents of SGs are 4e20% of the dry leaf weight, but depend on the cultivar and growing conditions. The largest percentage of SGs are composed of stevioside (Fig. 1 v), whose yield is 0.6e7.9% (w/w), followed by rebaudioside-A (reb-A)

* Corresponding author. E-mail address: [email protected] (H. Shimizu). http://dx.doi.org/10.1016/j.phytochem.2017.02.002 0031-9422/© 2017 Elsevier Ltd. All rights reserved.

(Fig. 1 vi), whose yield is 0.3e6.5% (w/w) (Vouillamoz et al., 2016). The sweetness intensity of stevioside (Fig. 1 v) is approximately 193 times greater than that of 2% sucrose, and the sweetness intensity of reb-A (Fig. 1 vi) is approximately 400 times higher than that (Schiffman et al., 1995). Schiffman et al. (1995) also investigated the relationship between the sweetness and taste of sweeteners. SGs have strong sweetness. However, they can also taste bitter at the same time. The SG reb-A (Fig. 1 vi) (Fig. 1 vi) is sweeter and less bitter than stevioside (Fig. 1 v). In addition, recent investigations have demonstrated that several minor SGs are stronger sweettasting compounds than stevioside (Fig. 1 v) and reb-A (Fig. 1 vi) (Hellfritsch et al., 2012; Espinoza et al., 2014). As an example, rebaudioside M, which has more b-glucose bound to steviol (Fig. 1 ii) than stevioside (Fig. 1 v) or reb-A (Fig. 1 vi), is 200e350 times sweeter than sucrose, and its bitter aftertaste is reduced compared to reb-A (Fig. 1 vi) (Prakash et al., 2014). In previous genetic research on S. rebaudiana, Ohlrogge and Benning (2000) and Brandle et al. (2002) analyzed expressed sequence tags (ESTs), and various genes from the SG biosynthetic pathway have been sequenced. In recent years, SGs have even been investigated at the molecular level. Kaurene oxidase (KO) is an enzyme that is common to the SG and GA biosynthetic pathways. It is an especially important enzyme

58

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65

Fig. 1. Metabolic pathway of steviol glycosides (SGs) and gibberellins (GAs). The enzymes are: ent-kaurene 19-oxidase (KO), and uridine diphosphate-dependent glycosyltransferase (UGT). This pathway is a simplified schematic pathway referred to by Richman et al. (2005); Mohamed et al. (2011).

in SG and GA biosynthesis (Helliwell et al., 1998, 1999) because it is one of the most highly expressed SG-related genes (Kumar et al., 2012). Many studies have investigated the enzymes in SG biosynthesis that are shared by the GA biosynthesis pathway, but are unique to these pathways. UGT85C2, UGT74G1, and UGT76G1 are UDPglycosyltransferases (UGT) that catalyze sugar-transferring reactions to add glucose to the steviol (Fig. 1 ii) basic skeleton (Richman et al., 2005). UGT85C2 catalyzes the synthesis of steviolmonoside (Fig. 1 iii) from steviol (Fig. 1 ii). UGT74G1 catalyzes the synthesis of several SGs, including stevioside (Fig. 1 v), from steviolbioside (Fig. 1 iv). UGT76G1 also catalyzes several SGs and can add glucose to stevioside (Fig. 1 v) to form reb-A (Fig. 1 vi).

The relationship between these enzymes and the SGs content has been previously researched. For example, Guleria et al. (201 1) reported that transcription levels of KO, UGT85C2, and UGT76G1 were upregulated after treatment with 5% sucrose compared to their transcription levels after treatment with 3% sucrose. Mohamed et al. (2011) suggested that the glycosylation involved in UGT85C2 is a rate-limiting step in SG biosynthesis. Based on these studies, transcription analysis of these enzymatic genes can serve as an indirect analysis of the levels of SGs. Some studies have demonstrated a relationship between SGs and GA. Because these compounds share the same biosynthesis pathway up to ent-kaurenoic acid (Fig. 1 i), an effect on the accumulation of SGs can be expected. Kumar et al. (2012) reported that SrUGT74G1 was upregulated by GA treatment. Similarly, Hajihashemi et al. (2013) reported that transcription of KO was significantly upregulated following GA treatment compared to transcription in the control. Recent investigations have further demonstrated that SG biosynthesis is closely related to GA biosynthesis. Numerous studies have investigated genes encoding GA biosynthesis enzymes. The transcription level of these genes is controlled by light (Ait-Ali et al., 1999; Zhao et al., 2007). It is well known that a red-light photoreceptor, called phytochrome, is involved in GA regulation (Toyomasu et al., 1998; Smith, 2000). There are two phytochrome conformations: Pfr (active form) and Pr (inactive form). Pfr is converted into Pr under 730 nm far-red light. Pr is converted into Pfr under 660 nm red light. In several species, such as cowpea, Arabidopsis thaliana, and cucumber, GA biosynthesis is promoted by far-red light (Kamiya and García-Martínez, 1999; García-Martinez and Gil, 2001). In studies examining the effects of light quality on S. rebaudiana, the levels of SGs were higher when S. rebaudiana was grown under short-day conditions with night-interruption by red light-emitting diodes (LEDs) compared to the control conditions (Ceunen et al., 2012). Another property of light has also been implicated in GA regulation. It has been reported that cryptochrome, which is a blue light photoreceptor, contributes to the inactivation of GA1, which is a bioactive GA under blue light treatment (Zhao et al., 2007). Given that the SGs content can be easily modulated by changing supplemental light, it can be expected to be able to increase the yield of SGs using particular light qualities. However, the relationship between SGs and the light qualities that affect GA biosynthesis has not been investigated thoroughly. The details about these light conditions are still unclear. Therefore, S. rebaudiana we investigated under various light conditions that are involved in GA biosynthesis to increase the volume of SGs. To confirm that SGs-rich

Fig. 2. Effect of various light quality treatments. Various red and far-red LEDs light conditions (A). Various blue and red LEDs light conditions (B). Various blue, red and far-red LEDs light conditions (C). Fluorescent lamp (FL), red LEDs (Red), red and far-red LEDs ratio (R/FR), blue and red LEDs ratio (B/R), blue LEDs (Blue), blue, red and far-red LEDs mix (BRFR mix). Scale bar indicates 1 cm. All of samples were collected 6 weeks after cutting.

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65

59

Fig. 3. Morphological analysis of S. rebaudiana under various light qualities. Various red and far-red LEDs light conditions (left). Various blue and red LEDs light conditions (middle). Various blue, red and far-red LEDs light conditions (right). White bars are 4 weeks after cutting. Black bars are 6 weeks after cutting. Each bar represents the mean. Error bars represent the standard deviation of the mean (n ¼ 4). Several treatments had a sample size of n ¼ 3 (see Materials and Methods). Different letters indicate significant differences (Tukey-Kramer method, P < 0.05). Small letters indicate a significant difference at 4 weeks after cutting. Capital letters indicate a significant difference at 6 weeks after cutting.

S. rebaudiana plants were produced, both morphological and genetic changes were evaluated. 2. Results 2.1. Morphological analysis 2.1.1. Effects of red and far-red LEDs on S. rebaudiana Phytochrome-photoequilibrium and GA biosynthesis are affected by the ratio of red (R) and far-red (FR) light. Therefore, S. rebaudiana was grown under different red/far-red (R/FR) ratios to examine the effect on its growth (Fig. 2A and Fig. 3). No significant differences were observed in the morphology of S. rebaudiana between fluorescent lamp (FL) and red light growth conditions. Compared to the FL condition, S. rebaudiana grown under higher far-red ratio conditions tended to have elongated stems, with the longest stem among them being 16.3 cm ± 0.9 under the R/FR 0.16 condition. By contrast, the total leaf area,

weights, numbers and stem diameters tended to decrease. There were no differences between S. rebaudiana grown under the R/FR 1.95 and FL conditions. The ratio of red and far-red in the FL was 7.54, which is nearly the same R/FR ratio as the R/FR 7.60 condition (Table 1). However, total leaf area and weight in the R/FR 7.60 condition were significantly decreased, compared to the FL condition. There was little difference between samples cut at 4 and 6 weeks, except that the results after 6 weeks of cutting had a greater change than that seen at 4 weeks after cutting. 2.1.2. Effects of blue and red LEDs on S. rebaudiana The increase in the percentage of blue light resulted in a decrease of all measured traits (Fig. 2B and Fig. 3). The shortest stems obtained from any of the treatments were under the blue (B) light condition (4 weeks: 4.5 cm ± 0.7; 6 weeks: 5.3 cm ± 0.3). Compared to the stem lengths of all blue/red (B/R) conditions, stems grown under the B/R 0.12 condition were significantly

60

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65

Table 1 Summary of light treatments: Fluorescent lamp (FL), red LEDs (Red), red and far-red LEDs ratio (R/FR), blue and red LEDs ratio (B/R), blue LEDs (Blue), blue, red and far-red LEDs mix (BRFR mix). Light treatment

PPFa (mmol m2 s1)

Peak wave length (nm)

Far-red (mmol m2 s1)

B/R ratio

R/FR ratio

Pfr/Ptotalb

FL

120

e

4.55

0.94

7.54

0.83

120

660

0.27

e

449

0.89

R/FR

7.60 3.90 1.95 1.22 0.16

120 120 120 120 120

660,730 660,730 660,730 660,730 660,730

15.9 30.6 61.7 98.9 751

e e e e e

7.60 3.90 1.95 1.22 0.16

0.86 0.83 0.78 0.73 0.35

B/R

0.12 0.25 0.42 0.67 1.00 8.57

120 120 120 120 120 120

470,660 470,660 470,660 470,660 470,660 470,660

e e e e e e

0.12 0.25 0.42 0.67 1.00 8.57

e e e e e e

e e e e e e

120

470

e

e

e

e

120 120

470,660,730 470,660,730

61.1 505

0.98 2.61

0.98 0.06

0.69 0.21

Red

Blue BRFR mix a b

B60R60 FR60 B88R32 FR505

PPF: photosynthetic photon flux (400e700 nm). Pfr/Ptotal: phytochrome-photoequilibrium was calculated using the photoconversion cross-sections from Sager et al. (1988).

elongated. The extent of stem elongation tended to be similar between plants sampled 4 and 6 weeks after cutting.

2.1.3. Effects of blue, red and far-red LEDs on S. rebaudiana Two light mix conditions using all three LEDs, blue, red and farred (BRFR), were set up to analyze the effects that the combination of three light qualities had on the plants (Fig. 2C and Fig. 3). Stem lengths of both BRFR conditions were significantly elongated, compared to the FL condition. In particular, the BRFR mix B88 R32 FR505, which included more FL than other BRFR conditions, resulted in the longest stem (21.7 cm ± 2.0). The length was approximately 1.8 times longer than stems of plants grown under FL (12.0 cm ± 1.2). However, there were no significant differences in the other measured items compared to the FL condition.

2.2. Genetic analysis 2.2.1. Effects of red and far-red LEDs on S. rebaudiana The transcription of KO under the R/FR 1.22 condition was higher than under the FL condition (Fig. 4A). S. rebaudiana grown under the R/FR 0.16 condition, which was the most far-red condition in this experiment, demonstrated a significantly enhanced transcription of UGT74G1 compared to that under the FL condition. The transcription pattern of UGT85C was similar to that of UGT76G1. The transcription of UGT85C2 in the R/FR 1.22 condition was increased by approximately 5.7 times over that of UGT85C2 in the FL condition, and transcription of UGT76G1 in the R/FR 1.22 condition was approximately 92.9 times higher than its transcription under the FL condition.

2.2.3. Effects of blue, red and far-red LEDs on S. rebaudiana The results were showed that R/FR 1.22 treatment and blue treatment were effective to enhance UGT85C2 and UGT76G1 transcription. However, the BRFR mixed conditions of blue, red and farred lights did not significantly enhance transcription of SG-related genes, compared to the FL condition (Fig. 4C). The BRFR B60 R60 FR60 treatment included the same light intensities of blue, red and far-red light. Although the R/FR ratio of the BRFR B60 R60 FR60 treatment was approximately the same as that of the R/FR 1.22 treatment, transcription levels of SG-related genes were lower compared to those the R/FR 1.22 treatment. 2.3. Analysis of SG contents According to genetic analysis, transcription of UGT85C2 was enhanced under the R/FR 1.22 and blue light treatments. Therefore, the relationship between SG-related gene and SGs contents were investigated using high performance thin layer chromatography (HPTLC) (Fig. 5). Consequently, the two main S. rebaudiana sweet components, stevioside (Fig. 1 v) and reb-A (Fig. 1 vi), were more strongly detected und the R/FR 1.22 and blue light conditions than under the FL condition. Genetic analyses were performed on extracts from the 7th leaves, whose averages of the areas under FL, R/FR 1.22 and blue conditions were 262, 82, and 75 mm2, respectively (data not shown). Averages of the levels of UGT85C2 under the FL, R/FR 1.22 and blue conditions were 1, 5.7, and 5.9, respectively. The leaf area  UGT85C2 following R/FR 1.22 and blue treatments were, respectively, approximately 1.8 and 1.7-fold larger than that of the FL control plants. 3. Discussion

2.2.2. Effects of blue and red LEDs on S. rebaudiana Compared to the FL condition, all of the SG-related genes were significantly enhanced under the blue light condition (Fig. 4B). Transcriptions of KO, UGT85C2, UGT74G1, and UGT76G1 under the blue condition were approximately 3, 5.9, 3.1 and 66 times greater than that under the FL condition. In addition, the pattern of gene transcription between UGT85C2 and UGT76G1 was similar. There were no significant differences in terms of gene transcription between all of the B/R ratio conditions and the FL condition.

3.1. Effects of red and far-red LEDs on S. rebaudiana 3.1.1. Consideration of morphological changes under red and far-red LEDs Results of anatomic analysis of S. rebaudiana under 16 light quality treatments showed that the stem tended to be longer and the leaf area became smaller with decrease in the R/FR ratio (Fig. 2A and Fig. 3). These morphological changes seemed to be consistent

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65

61

Fig. 4. Relative transcription of SG-related genes. Various red and far-red LEDs light conditions (A). Various blue and red LEDs light conditions (B). Various blue, red and far-red LEDs light conditions (C). Transcription levels were normalized against FL. Each bar represents the mean. Error bars represent the standard deviation of the mean (n ¼ 3). The different alphabets above each bar indicate significant differences (Tukey-Kramer method, P < 0.05). All of samples were collected 6 weeks after cutting.

Fig. 5. SGs content according to HPTLC. HPTLC of the fluorescent lamp (FL), blue LEDs (Blue), and red/far-red LEDs ratio (R/FR) 1.22 treatments (A). Analysis of the fold-change relative to the FL treatment based on HPTLC using ImageJ software (B). Each bar represents the mean. Error bars represent the standard deviation of the mean. Each data bar represents three independent experiments. The asterisks indicate a significant difference (t-test, P < 0.05). All of the samples were measured 6 weeks after cutting.

with shade-avoidance responses (Smith, 1982). Phytochrome, the red photoreceptor, detects the ratio of R/FR to initiate a shadeavoidance response. Phytochrome has two conformations, the active form (Pfr) and inactive form (Pr). The ratio of Pfr/Ptotal serves as an indicator if the amount of Pfr. If the value of the R/FR ratio is small in a far-red-enriched environment, the Pfr/Ptotal value is also small. In terms of light quality, far-red-rich environments tend to be found in the lower canopy rather than in the upper canopy because upper leaves preferentially absorb blue and red light, which are important for photosynthesis. Therefore, leaves in the lower canopy try to avoid shaded environments, and to do so, plants prioritize stem elongation over leaf growth. In the experiment here, R/FR ratios were set from 7.60 to 0.16 (Pfr/Ptotal: 0.86e0.35). According to Sager et al. (1986), an environment giving a Pfr/Ptotal ratio of 0.8 is considered to be a red light-rich environment and an environment giving a Pfr/Ptotal ratio of 0.35 is considered a far-red light-rich environment. Furthermore, phytochrome perceives the difference in light quality, so shadeavoidance responses are induced by changing the balance of the phytochrome-photoequilibrium to the Pr form. Altogether, it is believed that the R/FR 0.16 treatment has the maximum effects on stem elongation and reduction of leaf areas, compared to the FL treatment because of shade-avoidance responses. The leaf area and weight in the R/FR 7.60 condition were significantly smaller than in the FL condition, even though the R/FR ratio of FL was approximately 7.54 (Fig. 3). This result suggests that morphogenetic changes, such as an increase in leaf area, are affected by other light qualities, such as a blue light. A similar result was reported in lettuce, where the increase in leaf area was greater

62

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65

when using light treatment with wide-band wavelengths (Kim et al., 2004). Therefore, FL treatment was more effective than the R/FR 7.60 treatment because FL has wide-band wavelengths. 3.1.2. Effect of light treatments on SG-related genes under red and far-red LEDs KO is an important gene shared by the biosynthesis pathways of SGs and GA. Similarly, other SG-related genes, UGT85C2, UGT74G1, and UGT76G1, are important in SG biosynthesis. These four SGrelated genes were more strongly expressed under low R/FR ratio conditions than under the FL condition (Fig. 4A). According to morphological analysis, this change in transcription levels has the potential to be related to shade-avoidance responses caused by decreasing the R/FR ratio or caused by a far-red-rich environment. It is presumed that phytochrome and GA are involved in stem elongation from shade-avoidance responses (Vandenbussche et al., 2005). In cowpea, when the Pr form of phytochrome is increased by far-red irradiation, the amounts of GA1 increased. A likewise response of increased GA1 levels was observed in Arabidopsis thaliana and cucumber (Cucumis sativus L.) (Kamiya and GarcíaMartínez, 1999). As a result, stem elongation was induced. Recent investigations of the relationship between SGs and GA in S. rebaudiana demonstrated that external GA treatment changes transcription of SG-related genes (Kumar et al., 2012; Hajihashemi et al., 2013). In other words, there is the possibility that the four SGrelated genes examined in this study were more strongly expressed under low R/FR ratio conditions (far-red-rich environment) than under the FL condition, because the change of GA requirements was caused by shade-avoidance responses. SGs and GA are synthesized through a partially shared pathway. Therefore, it may be that levels of the substrates involved in SG-biosynthesis were increased following induction of bioactive GA levels. The transcriptions of KO under the R/FR 1.22 and R/FR 0.16 conditions were higher than under the FL condition. The R/FR ratio was set to 1.22, because the range of the daylight R/FR ratio is from 1.05 to 1.25 (Smith, 1982), whereas the R/FR 0.16 condition is a farred-rich environment that contains more of the Pr form of phytochrome. GA biosynthesis may be induced in an R/FR 1.22 environment, which results in strong KO transcription. UGT85C2 was maximally expressed under the R/FR 1.22 condition, and its transcription level was reduced under the R/FR 0.16 condition. These results suggest that there will be a shortage of precursor for GAs synthesis under the R/FR 0.16 condition. In other words, transcription of SG-related genes may be adjusted by the required amount of bioactive GA. In addition, transcription of UGT76G1 behaves similarly to that of UGT85C2 and may also be an important enzymatic gene that controls the amount of SGs. UGT74G1 showed a different behavior that UGT85C2 or UGT76G1 (Fig. 4C). The study by Guleria et al. (2011) indicates that transcription levels of KO, UGT85C2 and UGT76G1 under a 5% sucrose treatment were upregulated more than under a 3% sucrose treatment and that the amount of SGs was also increased. However, transcription of UGT74G1 was downregulated under the 5% sucrose treatment compared to the 3% sucrose treatment. For these reasons, there appears to be little relationship between the behavior of UGT74G1 and the amount of SGs. 3.2. Effects of blue and red LEDs on S. rebaudiana 3.2.1. Consideration of morphological changes under blue and red LEDs Stem lengths under the blue light condition were obtained the shortest among all of the conditions tested (Fig. 2B). They had a tendency to decrease as the B/R ratio increased. Generally, blue light is involved in the inhibition of stem elongation, except in some

species. In lettuce, which is also a member of the Asteraceae, stem elongation was inhibited by blue light treatment (Dougher and Bugbee, 2004). Therefore, inhibition of stem elongation was caused by a high B/R ratio, whereas stems from plants under the B/ R 0.12 condition were the longest among all of the conditions (Fig. 3). Hoenecke et al. (1992) reported that the most suitable growth conditions are blue and red mixed light conditions, which includes 10e20% blue light at a light intensity of 300 mmol m2 s1, which supports the results here that blue light and the B/R 0.12 condition (10% blue light) affect stem elongation in S. rebaudiana. 3.2.2. Consideration of genetic changes under blue and red LEDs Transcriptions of all SG-related genes under the blue light condition were significantly upregulated, as compared to the FL condition (Fig. 4b). Because blue light can inactivate GA, it is hypothesized that all SG-related genes were being influenced by changes in GA. Cryptochrome negatively regulates the GA-related genes AtGA20ox1 and AtGA3ox1, which are involved in activation of GA through positive regulation and positively regulate AtGA2ox1, which is involved in the inactivation of GA (Zhao et al., 2007). This previous study demonstrated that active GA becomes inactivated and GA biosynthesis is inhibited by blue light. Generally, a plant tissue can be affected by a small amount of bioactive GA. The concentration of GA20, which is the only GA detected in S. rebaudiana, was 1.2 mg per kg fresh weight (Alves and Ruddat, 1979). This amount was approximately 5,000,000 times lower than that of SGs. Therefore, an inactivating mechanism was developed to maintain a constant amount of bioactive GA. One of the GA conjugates is GA glucose conjugate, through which GAs are linked to glucose (Schneider and Schliemann, 1994). It is also believed that GA glucosyl esters have the potential to inactivate GA (Schneider et al., 1992). Because SGs are glycosides that share part of their biosynthesis with GA (Fig. 1), there is the possibility that the changes in SG-related gene transcriptions were caused by the blue light irradiation that inactivates GA. Judging from the R/FR experiment results, we consider that the shade-avoidance response is occurring due in part to the blue light response. Blue light detected via cryptochrome is also involved in shade-avoidance responses (Vandenbussche et al., 2005). However, Vandenbussche et al. (2005) also reported that cryptochromeinduced shade-avoidance responses occur under low intensities of blue light. The light intensity was set at 120 mmol m2 s1, and stem elongation was inhibited by blue light. That is, shadeavoidance responses did not occur. This result suggests that this light intensity (120 mmol m2 s1) is not considered to be low intensity for S. rebaudiana. Furthermore, it is possible that there are other light signaling pathways that are different from those in the R/FR experiments that affect SG-related genes. 3.3. Effects of blue and red and far-red LEDs on S. rebaudiana 3.3.1. Morphological changes under blue and red and far-red LEDs It was suggested that there are relationships between blue, red, and far-red light qualities in terms of transcription levels of SGrelated genes. In all of the experiments, R/FR 1.22 and blue light irradiations were the most effective for upregulating UGT85C2 transcription. Whether these effects remained under mixed conditions of three light qualities (blue, red and far-red) were thus examined. All of the mixed conditions, especially the BRFR mix B88 R32 FR505 condition, caused greater stem elongation compared to the FL condition. The Pfr/Ptotal was 0.21 under the BRFR mix B88 R32 FR505 condition, and there was much more Pr (phytochrome inactive form) than in the FL condition (Pfr/Ptotal ¼ 0.83), which is consistent with induction of shade-avoidance responses from increasing Pr.

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65

However, there was no difference in leaf development in the BRFR mixed conditions compared to the FL condition. This result indicated that a mix of the three light treatments did not cause shade-avoidance responses. From the viewpoint of the photosynthetic rate, the combination of red and far-red was more effective than the total number of single light treatments (Fork and Amesz, 1969). Under the BRFR mix B88 R32 FR505 condition, blue light was added, as it is involved in photomorphogenesis. Therefore, this mix of three light treatments had an effect on stem development because it was an ideal condition for photosynthesis. 3.3.2. Consideration of genetic changes under blue and red and farred LEDs There was no difference between the mixtures of the three light qualities and the FL treatment (Fig. 4C). Under the B60 R60 FR60 condition, the R/FR ratio was 0.98, which was close to the R/FR ratio under the 1.22 condition. However, the B60 R60 FR60 condition did not show any significant changes compared to the FL condition. The experiments conducted suggested the participation of two photoreceptors: phytochrome and cryptochrome. There has been much discussion about the overlapping functions or effects of the interactions between several photoreceptors on gene transcription (M
as et al., 2000; Chory and Wu, 2001). However, the results here do not show any effects of mixing the three light conditions on transcription of SG-related genes. This result suggests that there are multiple light signaling transduction pathways through phytochrome or cryptochrome that are involved in regulating transcription of SG-related genes and that these light signaling pathways can cancel the effects of each other. 3.4. Analysis of the amount of SGs It was expected that the R/FR 1.22 and blue light treatments would have greater effects on the upregulation of UGT85C2 than the FL treatment. The actual amount of SGs increased under the same treatment (Fig. 5). UGT76G1 is involved in the biosynthesis of reb-A (Fig. 1 vi). Via HPTLC analysis, the spots indicating the amount of reb-A (Fig. 1 vi) under the R/FR 1.22 and blue light conditions were larger than those indicating the FL condition. Transcription of UGT76G1 was similar to that of UGT85C2 via quantitative real-time (RT)-PCR (qPCR) analysis, so it may also be an important gene that regulates the amount of stevioside (Fig. 1 v) and reb-A (Fig. 1 vi). The 7th leaf area  UGT85C2 under the R/FR1.22 and blue conditions was larger than that of the FL condition. This result means that the R/FR1.22 and blue light treatments had an effect on the total yield of SGs per leaf. 4. Conclusion According to these results, there is a possibility that SG synthesis is affected by light quality, which can also play a role in upregulating GA biosynthesis. Moreover, there may be multiple light signaling pathways involved, including phytochrome or cryptochrome. In this study, which light qualities could be used to enrich SGs in S. rebaudiana was investigated and it was demonstrated that the R/FR 1.22 and blue light treatments had an effect on the amount of SGs per leaf. In particular, S. rebaudiana under blue light treatment had short stems and small leaf areas, but more concentrated SG contents than that of other light treatments. These smaller stems and leaves may be advantageous to reduce plants from being blown over by the wind, as well as being suitable for vertical farming. Therefore, the results may be commercially beneficial because blue light treatment may be able to save cultivation area space while also increasing the SGs content.

63

5. Experimental 5.1. Plant material To reduce genetic variation, experiments were conducted on cutting samples from a parent plant. Similar to a prior experiment, 10 different S. rebaudiana plants were purchased from several different garden centers (Kyoto, Japan) or were provided by a botanical garden affiliated with Kyoto University (Kyoto, Japan). Eight stems were cut from each plant and grown for 6 weeks under the same growth conditions. Each parent plant was selected and provided by Kyoto University because of the small variance among individuals after cutting. This parent plant is stored in our laboratory. All plants were cultivated hydroponically in growth chambers with urethane sponge culture medium. Stems with leaves from the third node (4 cm in height) were cut from the shoot apex, and the pair of 3rd node leaves was then cut in half. Before the stem was inserted in urethane, it was placed in H2O for 1 h. Subsequently, the shear part of the stem was coated with a rooting promoter (Rooton, Sumitomo Chemical Garden Products, Tokyo, Japan). The stem was grown for 2 weeks in order to take root (photoperiod: 16 h, temperature of light on: 24  C, temperature of light off: 20  C, humidity: 99%), and urethane was soaked in a nutrient solution (Menedael; Menedael Co., Ltd., Osaka, Japan) diluted 100 times with H2O. The light source for rooting was a fluorescent lamp (100MHF142DR, Monocoqtex, Inc., Tokyo, Japan), and the photosynthetic photon flux (PPF, 400e700 nm) was 120 mmol m2 s1. PPF was measured using a light quantum meter (I-250A, LI-COR, Inc., USA). One cultivation tank (30 cm  60 cm  7 cm) per test section was filled with nutrient solution (OAT house A No. 1 and No. 2, OAT Agrio Co., Ltd., Tokyo, Japan), which was then replaced by a new solution once a week (pH 6.4 and EC 1.3 dS m1). An air pump (Nonnoise S100, Japan Pet Design CO., LTD., Tokyo, Japan) was used for the aeration of each test section. 5.2. Treatments After rooting for 2 weeks, S. rebaudiana was transplanted to 16 differing light conditions (Table 1). Except for differences in light quality, S. rebaudiana was grown under the same conditions (photoperiod: 16 h, temperature of light on: 25  C, temperature of light off: 21  C, humidity of light on: 41%, humidity of light off: 67%, CO2 concentration of light on: 535 ppm, CO2 concentration of light off: 585 ppm). The temperatures, humidity levels, and CO2 concentrations were recorded using a data logger (TR-72Ii, T&D, Nagano, Japan). The light sources were LEDs (ISL-305X302-RFGB, ISL305X302-H4RFGB, ISL-150X150-FR, CCS, Inc., Kyoto, Japan) and a fluorescent lamp (100MHF142DR, Monocoqtex, Inc., Tokyo, Japan). All light intensities of the differing light conditions were 120 mmol m2 s1, not including far-red light, and the intensity of light exposure to the shoot apex was kept constant by adjusting the height using lab jacks. The R/FR 7.60 condition was set to approach an R/FR ratio of the FL 7.54. Because the R/FR ratio of daylight is 1.05e1.25 (Smith, 1982), an R/ FR condition was set at 1.22. The irradiance of far-red light to the shoot apex was adjusted with a spectroradiometer (LI-1800, LI-COR, Inc., USA). Pfr/Ptotal indicates the phytochrome-photoequilibrium of each R/FR condition according to the numbers of Sager et al. (1988) (Table 1). Ptotal indicates the total number of phytochrome molecules. 5.3. Morphological analysis Three or four S. rebaudiana leaves were sampled after 4 and 6 weeks of cutting, and the average values of the measurements were

64

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65

calculated along with the standard deviation. Total leaf areas and weights without leaves from the first of the pair leaves to the fourth pair of leaves from the bottom were measured because leaves were already developed to the fourth leaves from the bottom when S. rebaudiana was transplanted to the different light conditions. The stem lengths, stem diameters and leaf numbers were also measured including the first leaves through the fourth leaves from the bottom of the stem. Stem lengths were measured from the top of urethane to the shoot apex, and stem diameters were measured at the center of stem lengths. Leaf areas were measured using ImageJ software (http://rsbweb.nih.gov/ij/). 5.4. Quantitative RT-PCR analysis The seventh leaf from the bottom was selected for genetic analysis, because this leaf was newly developed after transplantation. After growing for 6 weeks after cutting, the seventh leaf was sampled 8 h after radiation treatment. Leaves were frozen in liquid N2 and ground with a mortar and a pestle. Total RNA was extracted with a RNA extraction kit (RNA suisui-P, Rizo, Inc., Tsukuba, Japan) and RNA extraction column (FARBC-C50, Favorgen, Taiwan). cDNA was synthesized from a total RNA (0.5 mg) template via reverse transcription-polymerase chain reaction (PCR) using a kit (Revere Tra Ace, TOYOBO CO., LTD. Life Science Department, Osaka, Japan). Next, quantitative real-time (RT)-PCR (qPCR) was performed for transcription analyses with the template cDNA. The total volume of the qPCR reaction solution was 20 mL, and qPCR was conducted using a qPCR kit (KAPA SYBR FAST qPCR kit, Kapa Biosystems, Inc., Woburn, MA, USA). The qPCR program was two steps: 95  C 30 s  1 cycle, 95  C 3 s and 60  C 30 s  40 cycles. Two internal standards (18S rRNA and b-Actin) were selected and the Vandesompele method was adopted to determine relative gene transcription levels (Vandesompele et al., 2002). Gene Expression Macro™ (Bio-Rad, Hercules, CA, USA) was used for calculations. Primers for SG-related genes, 18S rRNA, b-Actin, UGT85C2, UGT74G1 and UGT76G1, were based on Mohamed et al. (2011). Primers for KO were used according to Hajihashemi et al. (2013). Three biological replicates per light condition were analyzed. 5.5. Statistical analysis Statistical analysis was adopted from the Tukey-Kramer method (P < 0.05). S. rebaudiana was sampled 4 and 6 weeks after cutting (n ¼ 4), and nearly all of the conditions were measured for morphological analysis except for six conditions (n ¼ 3): R/FR 1.95, B/R 0.12, B/R 0.42, B/R 8.57 (4 weeks) and R/FR 0.16, BRFR mix B88 R32 FR505 (6 weeks). Data from Fig. 5B was analyzed by t-test to determine if FL conditions are significantly different compared to the blue or the R/FR 1.22 conditions (P < 0.05). 5.6. HPTLC The SGs content was analyzed using the HPTLC method from Jaitak et al. (2008), as HPTLC is a simple and cost effective means of measuring SGs. In general, SGs are measured using high performance liquid chromatography (HPLC); however, sample volumes in these experiments were too small for HPLC measurement. Therefore, HPTLC was adopted in these experiments. Six weeks after cutting, samples were spotted on a plate for HPTLC analyses (HPTLC 60 F254, E. Merck, Darmastadt, Germany). A sample 1 cm in diameter was cut from the sixth leaf from the bottom of each S. rebaudiana. Distilled H2O (50 mL) was added to the sample, and this was then triturated. Samples of 20 mL were applied to the plate.

Acknowledgements S. rebaudiana was provided by Michiho Ito, Associate Professor, Graduate School and Faculty of Pharmaceutical Sciences, Kyoto University. This work has been partly supported by the programs of the Grant-in-Aid for Scientific Research (B) from the Japan Society for the Promotion of Science (25292150). Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.phytochem.2017.02.002. References Ait-Ali, T., Frances, S., Weller, J.L., Reid, J.B., Kendrick, R.E., Kamiya, Y., 1999. Regulation of gibberellin 20-oxidase and gibberellin 3b-hydroxylase transcript accumulation during de-etiolation of pea seedlings. Plant Physiol. 121, 783e791. Alves, L.M., Ruddat, M., 1979. The presence of gibberellin A20 in Stevia rebaudiana and its significance for the biological activity of steviol. Plant Cell Physiol. 20, 123e130. Brandle, J.E., Richman, A., Swanson, A.K., Chapman, B.P., 2002. Leaf ESTs from Stevia rebaudiana: a resource for gene discovery in diterpene synthesis. Plant Mol. Biol. 50, 613e622. Ceunen, S., Werbrouck, S., Geuns, J.M.C., 2012. Stimulation of steviol glycoside accumulation in Stevia rebaudiana by red LED light. J. Plant Physiol. 169, 749e752. Chen, T.H., Chen, S.C., Chan, P., Chu, Y.L., Yang, H.Y., Cheng, J.T., 2005. Mechanism of the hypoglycemic effect of stevioside, a glycoside of Stevia rebaudiana. Planta Med. 71, 108e113. Chen, J., Jeppesen, P.B., Nordentoft, I., Hermansen, K., 2006. Stevioside counteracts the glyburide-induced desensitization of the pancreatic beta-cell function in mice: studies in vitro. Metabolism 55, 1674e1680. Chory, J., Wu, D., 2001. Weaving the complex web of signal transduction. Plant Physiol. 125, 77e80. Dougher, T., Bugbee, B., 2004. Long-term blue light effects on the histology of lettuce and soybean leaves and stems. J. Am. Soc. Hort. Sci. 129, 467e472. Dyrskog, S.E.U., Jeppesen, P.B., Colombo, M., Abudula, R., Hermansen, K., 2005. Preventive effects of a soy-based diet supplemented with stevioside on the development of the metabolic syndrome and type 2 diabetes in Zucker diabetic fatty rats. Metab. Exp. 54, 1181e1188. Espinoza, M.I., Vincken, J.P., Sanders, M., Castro, C., Stieger, M., Agosin, E., 2014. Identification, quantification, and sensory characterization of steviol glycosides from differently processed Stevia rebaudiana commercial extracts. J. Agric. Food Chem. 62, 11797e11804. Fork, D.C., Amesz, J., 1969. Action spectra and energy transfer in photosynthesis. Annu. Rev. Plant Physiol. 20, 305e328. García-Martinez, J.L., Gil, J., 2001. Light regulation of gibberellin biosynthesis and mode of action. J. Plant Growth Regul. 20, 354e368. Geuns, J.M.C., 2003. Stevioside. Phytochemistry 64, 913e921. Guleria, P., Kumar, V., Yadav, S.K., 2011. Effect of sucrose on steviol glycoside biosynthesis pathway in Stevia rebaudiana. Asian J. Plant Sci. 10, 401e407. Hajihashemi, S., Geuns, J.M.C., Ehsanpour, A.A., 2013. Gene transcription of steviol glycoside biosynthesis in Stevia rebaudiana Bertoni under polyethylene glycol, paclobutrazol and gibberellic acid treatments in vitro. Acta Physiol. Plant 35, 2009e2014. Hellfritsch, C., Brockhoff, A., St€ ahler, F., Meyerhof, W., Hofmann, T., 2012. Human psychometric and taste receptor responses to steviol glycosides. J. Agric. Food Chem. 60, 6782e6793. Helliwell, C., Sheldon, C., Olive, M., Walker, A., Zeevaart, J., Peacock, W., Dennis, E., 1998. Cloning of the Arabidopsis H-kaurene oxidase gene GA3. Proc. Natl. Acad. Sci. U. S. A. 95, 9019e9024. Helliwell, C.A., Poole, A., Peacock, W.J., Dennis, E.S., 1999. Arabidopsis ent-kaurene oxidase catalyzes three steps of gibberellin biosynthesis. Plant Physiol. 119, 507e510. Hoenecke, M.E., Bula, R.J., Tibbitts, T.W., 1992. Importance of ‘blue’ photon levels for lettuce seedlings grown under red-light-emitting diodes. HortScience 27, 427e430. Jaitak, V., Gupta, A., Kaul, V., Ahuja, P.S., 2008. Validated high-performance thinlayer chromatography method for Steviol glycosides in Stevia rebaudiana. J. Pharm. Biomed. Anal. 47, 790e794. Kamiya, Y., García-Martínez, J.L., 1999. Regulation of gibberellin biosynthesis by light. Curr. Opin. Plant Biol. 2, 398e403. Kim, H.H., Goins, G.D., Wheeler, R.M., Sager, J.C., 2004. Green-light supplementation for enhanced lettuce growth under red-and blue-light-emitting diodes. HortScience 39, 1617e1622. Kumar, H., Kaul, K., Bajpai-Gupta, S., Kaul, V.K., Kumar, S., 2012. A comprehensive analysis of fifteen genes of steviol glycosides biosynthesis pathway in Stevia rebaudiana (Bertoni). Gene 492, 276e284.

Y. Yoneda et al. / Phytochemistry 137 (2017) 57e65
s, P., Devlin, P.F., Panda, S., Kay, S. a, 2000. Functional interaction of phytochrome Ma B and cryptochrome 2. Nature 408, 207e211. Mohamed, A.A.A., Ceunen, S., Geuns, J.M.C., Van den Ende, W., De Ley, M., 2011. UDP-dependent glycosyltransferases involved in the biosynthesis of steviol glycosides. J. Plant Physiol. 168, 1136e1141. Ohlrogge, J., Benning, C., 2000. Unraveling plant metabolism by EST analysis. Curr. Opin. Plant Biol. 3, 224e228. Prakash, I., Markosyan, A., Bunders, C., 2014. Development of next generation Stevia sweetener: rebaudioside M. Foods 3, 162e175. Richman, A., Swanson, A., Humphrey, T., Chapman, R., McGarvey, B., Pocs, R., Brandle, J., 2005. Functional genomics uncovers three glucosyltransferases involved in the synthesis of the major sweet glucosides of Stevia rebaudiana. Plant J. 41, 56e67. Sager, J.C., Smith, W.O., Edwards, J.L., Cyr, K.L., 1986. Use of spectral data to determine photosynthetic efficiency and phytochrome photoequilibria. In: Proceedings, American Society of Agricultural Engineers. California Polytechnic Institute, San Luis Obispo, CA. Paper No. 86e3052. Sager, J.C., Smith, W.O., Edwards, J.L., Cyr, K.L., 1988. Photosynthetic efficiency and phytochrome photoequilibria determination using spectral data. Trans. ASAE 31, 1882e1889. Schiffman, S.S., Booth, B.J., Losee, M.L., Pecore, S.D., Warwick, Z.S., 1995. Bitterness of sweeteners as a function of concentration. Brain Res. Bull. 36, 505e513. Schneider, G., Jensen, E., Spray, C.R., Phinney, B.O., 1992. Hydrolysis and reconjugation of gibberellin A20 glucosyl ester by seedlings of Zea mays L. Proc. Natl. Acad. Sci. U. S. A. 89, 8045e8048.

65

Schneider, G., Schliemann, W., 1994. Gibberellin conjugates: an overview. Plant Growth Regul. 15, 247e260. Singh, S.D., Rao, G.P., 2005. Stevia: the herbal sugar of 21st century. Sugar Tech. 7, 17e24. Smith, H., 1982. Light quality, photoperception, and plant strategy. Annu. Rev. Plant Physiol. 33, 481e518. Smith, H., 2000. Phytochromes and light signal perception by plantsean emerging synthesis. Nature 407, 585e591. Toyomasu, T., Kawaide, H., Mitsuhashi, W., Inoue, Y., Kamiya, Y., 1998. Phytochrome regulates gibberellin biosynthesis during germination of photoblastic lettuce seeds. Plant Physiol. 118, 1517e1523. Vandenbussche, F., Pierik, R., Millenaar, F.F., Voesenek, L.A.C.J., Van Der Straeten, D., 2005. Reaching out of the shade. Curr. Opin. Plant Biol. 8, 462e468. Vandesompele, J., Preter, K.D., Pattyn, F., Poppe, B., Roy, N.V., Paepe, A.D., Speleman, F., 2002. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol. 3 research0034.1eresearch0034.11. Vouillamoz, J.F., Wolfram-Schilling, E., Carron, C.A., Baroffio, C.A., 2016. Agronomical and phytochemical evaluation of Stevia rebaudiana genotypes. Julius Kühn Archiv 453, 86e88. Zhao, X., Yu, X., Foo, E., Symons, G.M., Lopez, J., Bendehakkalu, K.T., Xiang, J., Weller, J.L., Liu, X., Reid, J.B., Lin, C., 2007. A study of gibberellin homeostasis and cryptochrome-mediated blue light inhibition of hypocotyl elongation. Plant Physiol. 145, 106e118.