Production of santalenes and bergamotene in Nicotiana tabacum plants

Terpenes play an important role in plant–insect relationships, and these relationships can potentially be modified by altering the profile of terpenes emitted from plants using metabolic engineering methods. Transgenic plants generated by employing such methods offer the prospect of low-cost sustainable pest management; in this regard, we used chloroplast targeting and cytosolic mevalonic acid pathway enhancement in this study to investigate the interaction of santalenes and bergamotene with insects. The santalene- and bergamotene-emitting transgenic tobacco plants thus generated were utilized to study host preference in the green peach aphid (Myzus persicae (Sulzer)). The results showed that co-expression of either 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR) or truncated HMGR with santalene synthase led to the production of higher amounts of santalenes and bergamotene in transgenic tobacco plants, and that these santalene- and bergamotene-emitting plants were attractive to green peach aphids. We accordingly propose that such transgenic plants may have potential application in pest management as a trap crop to prevent green peach aphid infestation of wild-type tobacco plants.


Introduction
Volatile terpenoids play important roles in plant defense, as they normally act as deterrents or repellents against herbivorous insect, or can function as attractants to attract carnivorous insect [1][2][3][4][5][6]. Although the application of insecticides is a common and effective method of pest management, these chemicals can have undesirable side effects on the environment, promote rapid evolution of resistance in pests, and can reduce predator populations [7]. Previous studies have demonstrated that transgenic plants in which emitted volatiles had been modified by metabolic engineering can be used to deter feeding by hornworms [8], decrease aphid colonization [9][10][11], or attract mite predators of infesting arthropods [12]. Modification of plant volatile terpenoids through metabolic engineering thus has considerable potential to lower the cost of pest management and facilitate more sustainable agricultural practices by reducing reliance on chemical pesticide usage [13]. PLOS  Volatile terpenes, including monoterpenes (C 10 ), sesquiterpenes (C 15 ), and diterpenes (C 20 ), play an important role in a wide range of plants. Plants interact with insects, surrounding plants, and pathogens by releasing volatile molecules. Monoterpenes (C 10 ) and diterpenes (C 20 ) are mainly biosynthesized via the plastidial methylerythritol 4-phosphate (MEP) pathway, while sesquiterpenes (C 15 ) are mainly biosynthesized from isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP) via the cytosolic mevalonic-acid (MVA) pathway. [14]. In recent years, Nicotiana tabacum has served as a platform for the production of several sesquiterpenes, as this plant can be sustainably cultivated at low cost [15]. Several specific metabolic engineering methods have been developed to enhance the production of sesquiterpenes in tobacco. For instance, the reaction catalyzed by 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR) is believed to be an important rate-limiting step in the tobacco MVA pathway; tobacco plants over-expressing HMGR have been shown to accumulate three to six times more sterol than wild-type plants [16]. In addition to enhancing HMGR expression, transgenic tobacco plants in which both sesquiterpene synthase and farnesyl diphosphate synthase (FPS) had been targeted into chloroplasts produced 1000-fold more patchoulol or amorpha-4,11-diene compared with transgenic lines in which the responsible sesquiterpene synthase gene was expressed only in the cytosol [8,17]. Co-expression of the mitochondrial amorpha-4,11-diene synthase (ADS) gene and a truncated HMGR (tHMGR) gene has been found to increase the production of amorpha-4,11-diene in transgenic tobacco [18,19], whereas transgenic tobacco co-expressing tHMGR and FPS with a valencene synthase gene showed increased production of (+)-valencene [20]. Furthermore, squalene biosynthesis was influenced by engineering genes of FPS and squalene synthase (SQS) via the N. tabacum chloroplast genomes [21]. Therefore, co-expression of HMGR, subcellular localization of sesquiterpene synthase, and chloroplast genome transformation are considered to be three important factors that determine the production of sesquiterpenes in transgenic tobacco plants.
Sandalwood oil is an essential oil that is extracted from the heartwood of sandalwood plants (Santalum album) and is used as a popular fragrance and insect repellent [22]. However, given that sandalwood is a rare resource, the utilization of sandalwood oil is somewhat limited. Santalols and bergamotol are the major constituents of this essential oil, and are generated from santalene/bergamotene by santalene/bergamotene oxidases. In S. album plants, santalenes and bergamotene can be generated from farnesyl diphosphate (FPP) in the MVA pathway via the activity of santalene synthase (SaSSy) [23,24]. Green peach aphid (Myzus persicae (Sulzer)), an important pest affecting tobacco production, has a broad spectrum of hosts worldwide [25,26]. To date, however, metabolic engineering of santalenes and bergamotene in tobacco plants, and their interaction with green peach aphid have not been intensively studied. In the present study, we employed chloroplast targeting methodology and co-expression of FPS, HMGR or tHMGR to enhance the production of santalenes and bergamotene in N. benthamiana and N. tabacum plants. The results showed that co-expression of either HMGR or tHMGR with SaSSy enhanced the production of santalenes and bergamotene in tobacco plants; transgenic tobacco plants emitting santalenes and bergamotene showed increased attraction to the green peach aphid M. persicae (Sulzer), indicating that these transgenic tobacco plants have considerable potential as a trap crop to protect wild-type tobacco from M. persicae feeding damage.

Plant materials and growth conditions
The seeds of tobacco plants (N. benthamiana and N. tabacum SR1) provided by Prof. Nam-Hai Chua (The Rockefeller University), were grown in a glasshouse under natural light conditions. Final binary vectors were introduced into Agrobacterium tumefaciens AGL1 by electro-transformation, and the transformed A. tumefaciens were then infiltrated into the leaves of tobacco plants using a standard protocol [16].

Vector construction
All the vectors used in this study are shown in S1 Fig Proof HF MasterMix (Bio-Rad, USA) and Rapid DNA ligation kit (ROCHE, Swiss) were used to perform PCR and ligation, respectively.

Infiltration of N. benthamlana leaves
A. tumefaciens cells were cultured in LB medium containing kanamycin (25 mg/L), spectinomycin (50 mg/L), or refampicillin (50 mg/L) and incubated at 28˚C at 250 rpm for 2 d. Thereafter, the cells were centrifuged for 10 min at 4000 g and 4˚C, resuspended with 10 mM MES buffer (containing 10 mM MgCl 2 and 100 μM acetosyringone) to a final OD 600 of 1.0, and then incubated at room temperature for 2 h. An A. tumefaciens strain encoding the TBSV P19 protein was added to suppress gene silencing [28]. Using a 1-mL syringe, A. tumefaciens cells were infiltrated into 3-week-old N. benthamiana plants grown in a glasshouse under natural light conditions. Three days after infiltration, the terpene products of the treated plants were analyzed by gas chromatography-mass spectrometry (GC-MS).
leaves. The A. tumefaciens strain encoding the TBSV P19 protein was added to suppress gene silencing. After 2-4 days, protein localization was observed by confocal laser scanning microscopy.
Plant transformation. Agrobacterium-mediated leaf disc co-cultivation method was used for tobacco transformation [16]. Tobacco seeds were firstly treated with 70% ethanol for 5 min, followed by sterilization with 20% bleach (2% NaOCl) for 10 min and washed three times with sterile distilled water. Sterilized seeds were germinated on MS medium. After one month, the mature leaves were cut into 0.5~1 cm 2 and transformed via A. tumefaciens AGL1 containing different vectors. The transformed plants were selected on MS medium containing 50 mg/L of hygromycin or 10 mg/L of Basta as plant selection antibiotics [29]. PCR using specific primers was performed to screen transgenic tobacco lines for transgenic integration.

Analyses of extracts and headspace volatiles
Two hundred milligrams of infiltrated leaves was ground in liquid nitrogen and extracted with 0.5 mL ethyl acetate at room temperature. One microliter of camphor-methanol solution (10 mg/mL), which was used as an internal standard, was added to the extracts. The extracts were prepared by shaking for 1 h at room temperature. Thereafter, the extracts were centrifuged for 5 min at 4000 g, and the upper organic layer was transferred to a new tube and dehydrated using anhydrous sodium sulfate (Na 2 SO 4 ). GC-MS was used to analyze the extracts. Volatiles were collected from 10 g of detached transgenic tobacco leaves using a Push-pull headspace collection system [30] under standard growth chamber conditions (25˚C, 16-h light:8-h dark photoperiod). Two microliters of camphor-methanol solution (10 mg/mL), which was used as an internal standard, was added to the leaf surfaces. A Porapak Q trap (80~100 mesh size; Sigma, USA) was used to trap volatiles during a 24-h period. Volatiles were eluted two times with 200 μL of dichloromethane and analyzed using an Agilent GC 7890A system incorporating a 5975C inert mass-selective detector and an HP-5 MS column (30 m × 0.25 mm; 0.25 μm film thickness). Samples (2 μL) were injected at an initial oven temperature of 50˚C, which was held for 1 min and then subsequently increased to 300˚C at a rate of 8˚C/min and then held for 5 min. Compounds were identified by comparison of their mass spectra with those in the NIST MS 2014 library.

RNA isolation and quantitative real-time PCR
Total RNA was isolated from transgenic tobacco using an RNeasy Plant Mini Kit (Qiagen, German), following the manufacturer's instructions, and cDNA was synthesized using iScript RT SuperMix (Bio-Rad, USA). Gene-specific primers (shown in S3 Table) were designed using Primer3Plus. Quantitative PCR was performed using Power SYBR Green PCR Master Mix (Life Technology, USA) in an ABI 7900 RT-PCR system (Applied Biosystems, USA). The transcript levels of all genes were analyzed relative to a reference gene, elongation factor-1 alpha (EF1a; GenBank: D63396.1), using the ΔΔCt method [31]. For each experiment, three replicates were performed with at least three independent biological samples.

Aphids
Green peach aphids (M. persicae (Sulzer)) were provided by Temasek Life Sciences Laboratory and reared in a 52 × 52 × 50 cm cage containing eight A. thaliana plants and maintained in a growth chamber (25˚C, 16-h light:8-h dark photoperiod).

Short-and long-term insect choice experiments using transgenic tobacco
A short-term choice experiment using green peach aphids was modified from that described previously [32]. Two plants of similar size and leaf number (one wild-type (WT) and one transgenic tobacco plant) were placed in a cage (30 × 30 × 30 cm), into which 30 adult green peach aphids were released. Twenty-four hours after release, the number of aphids on each plant was recorded.
To investigate the long-term choice preferences of these aphids, two plants of similar size and leaf number (one WT and one transgenic tobacco plant) were grown in a cage (30 × 30 × 30 cm) (25˚C, 16-h light:8-h dark photoperiod), into which 30 adult green peach aphids were released. Four weeks after release, the number of aphids on each plant was recorded.

Aphid two-choice experiments using hexane solutions of santalenes and bergamotene
Two leaves from 3-week-old N. tabacum plants were placed on the surface of agar medium in plates, one leaf was treated with 20 μL of hexane solutions of santalenes and bergamotene (1 g/ L), and the other leaf was treated with 20 μL of hexane as a control. After evaporation of the solvent, 10 green peach aphids were released in each plate, and 15 min after release, the number of aphids on or above the leaf was recorded.

Statistical analysis
Values are indicated as 'mean ± SD' for three to six biological replicates. Statistical comparison was analyzed using a two-tailed Student's t-test or unequal variance t 0 -test, and indicated by asterisks. � indicates P < 0.05; �� indicates P < 0.01.

Targeting of FPS and SaSSy into chloroplasts using chloroplast transit peptides
Most chloroplast proteins are encoded by the nuclear genome and synthesized as precursors containing N-terminal targeting transit peptides. A chloroplast target peptide is a short peptide chain that directly transports a protein to a chloroplast, after which most of the transit peptide is cleaved from the protein [33]. Targeting efficiency is affected by the structure and sequence motifs of the transit peptide [34]. Sesquiterpenes are naturally synthesized from IPP and DMAPP via the cytosolic MVA pathway [14]. In order to synthesize sesquiterpenes in chloroplasts, we used a S. lycopersicum linalool synthase chloroplast transit peptide (LTP) and an A. thaliana ribulose bisphosphate carboxylase small subunit chloroplast transit peptide (RTP) to target both A. thaliana FPS and SaSSy proteins into chloroplasts.
FPS, SaSSy, LTP, and RTP sequences were inserted into pCambia1300-GFP vectors to generate 1300-35SP-LTP-FPS-GFP, 1300-35SP-LTP-SaSSy-GFP, 1300-35SP-RTP-FPS-GFP, and 1300-35SP-RTP-SaSSy-GFP vectors, respectively (S1 Fig). Three days after N. benthamiana agro-infiltration, the subcellular locations of FPS and SaSSy proteins were examined using confocal microscopy (Fig 1). The presence of a green fluorescent protein signal (GFP) showed that native FPS and SaSSy proteins are both typically located in the cytoplasm of WT plants (Fig 1A and 1B), whereas after fusion with LTP, the GFP signal of LTP-SaSSy was observed in chloroplasts. In contrast, the LTP-FPS protein appeared in both cytoplasm and chloroplasts (Fig 1C and 1D). Compared with the localization of LTP-SaSSy and LTP-FPS proteins, the GFP signals of RTP-FPS and RTP-SaSSy were exclusively visible in the chloroplasts (Fig 1E  and 1F). These observations indicated that the targeting efficiency of RTP is higher than that

Effects of different metabolic engineering methods on the production of santalenes and bergamotene in N. benthamiana
To assess the effects of chloroplast targeting and MVA enhancement strategies using A. thaliana FPS, HMGR or tHMGR on the production of santalenes and bergamotene, we transiently co-expressed the two genes together in N. benthamiana (RTP-FPS + RTP-SaSSy, FPS + SaSSy, HMGR + SaSSy, or tHMGR + SaSSy). To avoid gene silencing mediated by promoter homology [35], the SaSSy or RTP-SaSSy genes were placed under the control of the RbcS promoter of A. thaliana, whereas each of the RTP-FPS, FPS, tHMGR, and HMGR sequences was expressed under the control of the Cauliflower mosaic virus (CaMV) 35S promoter.
Three days after N. benthamiana agro-infiltration, 200 mg of treated leaves from each plant was ground in liquid nitrogen and extracted using ethyl acetate, and the resulting extracts were analyzed by GC-MS. Although no santalenes or bergamotene were observed in the wildtype N. benthamiana plant and N. benthamiana leaves transiently expressing a single gene (SaSSy) (Fig 2A), trace amounts of santalenes and bergamotene were found in lines transiently co-expressing RTP-FPS + RTP-SaSSy or FPS + SaSSy (Fig 2A), which indicated that the chloroplast targeting and FPS enhancement strategies slightly increased the production of santalenes and bergamotene. When tHMGR was transiently co-expressed with SaSSy, the tobacco plants accumulated 1025.8 ng�g -1 fresh weight (FW) of α-santalene, 354.8 ng�g -1 FW of α-bergamotene, 68.8 ng�g -1 FW of Epi-β-santalene, and 885.1 ng�g -1 FW of β-santalene, the values of which were approximately 50 times higher than the amounts produced using the chloroplast targeting and FPS enhancement strategies (Table 1 and Fig 2B). In addition, co-expression with HMGR also yielded high amounts of santalenes and bergamotene: 984.8 ng�g -1 FW of αsantalene, 324.1 ng�g -1 FW of α-bergamotene, 40.5 ng�g -1 FW of Epi-β-santalene, and 830.0 ng�g -1 FW of β-santalene (Table 1 and Fig 2B). Therefore, the HMGR or tHMGR MVA enhancement strategy is suggested to be a more suitable method for enhancing the yields of santalenes and bergamotene in N. benthamiana. To further evaluate the effects of the different metabolic engineering methods on the production of santalenes and bergamotene, we transformed N. tabacum plants with RTP-FPS + RTP-SaSSy, FPS + SaSSy, tHMGR + SaSSy, or HMGR + SaSSy.

HMGR and tHMGR enhance the production of santalenes and bergamotene in transgenic N. tabacum
The GC-MS results of N. benthamiana agro-infiltration experiments indicated that chloroplast targeting and co-expression of FPS, HMGR or tHMGR improved the production of santalenes and bergamotene to different levels. To further investigate the effects of these metabolic engineering methods on the production of santalenes and bergamotene in stable transgenic N. tabacum plants, different gene combinations (SaSSy, RTP-FPS + RTP-SaSSy, FPS + SaSSy, tHMGR + SaSSy, or HMGR + SaSSy) were used to transform N. tabacum. For each transformation event, three independent lines with similar relative expression levels of SaSSy were selected from at least 10 lines of transgenic plants and GC-MS was performed to analyze the volatiles produced by each of the selected transgenic plants.

Production of santalols and bergamotol in transgenic tobacco
Santalols and bergamotol, which are major components of sandalwood essential oil, are generated from santalenes and bergamotene by S. album santalene/bergamotene CYP76F cytochrome P450 [23]. The results of the present study indicated that co-expression of HMGR in transgenic tobacco could enhance the production of santalenes and bergamotene. We subsequently employed HMGR, SaSSy, and S. album santalene/bergamotene oxidase SaCYP76F39v1 [23] to produce santalols and bergamotol in tobacco. To avoid gene silencing caused by promoter homology [35], three strong constitutive promoters, Arabidopsis ubiquitin 10 [27], RbcS, and 35S promoters were used to control these three genes respectively. Unfortunately, only santalenes and bergamotene were detected in both agro-infiltrated N. benthamiana plants and transgenic N. tabacum plants co-expressing HMGR, SaSSy, and SaCYP76F39v1) (Fig 3E  and S3 Fig) To detect the presence of glycosylated santalols and bergamotol in transgenic tobacco plants, homogenized HSP1 leaves were treated with either a saturated CaCl 2 solution or a glycosidase [36]. In the present study, we were unable to detect any santalols or bergamotol in either untreated or treated leaves of HSP1 line plants (S4 Fig). These results indicate that santalols and bergamotol were not stored as nonvolatile santalols/bergamotol-glycoside forms in leaf tissues or cells of transgenic tobacco plants.

Interaction of aphids with transgenic N. tabacum emitting santalenes and bergamotene
Green peach aphids are major pests of many commercial crops, not only causing feeding damage but also transmitting various viruses that can result in substantial losses [37]. To investigate the interaction of aphids with transgenic N. tabacum plants emitting santalenes and bergamotene, two-choice experiments were performed to examine green peach aphid preference between WT and transgenic tobacco plants [38]. Initially, we conducted short-term twochoice experiments using plants placed in a cage (Fig 4A), which revealed that 59.8% to 63.6% of aphids showed a preference for the transgenic tobacco plants (S3, 4, and 16), indicating that the transgenic plants emitting santalenes and bergamotene tended to attract aphids to a greater extent than the WT plants. Subsequently, to examine the long-term preference of insects for either WT or transgenic tobacco plants. Four weeks after releasing aphids, we observed that 67.1% to 67.3% of aphids showed preference for transgenic tobacco (S3, 4, and 16; Fig 4B and  4D, S5 Fig). These results indicated that the transgenic plants emitting santalenes and bergamotene were more attractive to aphids over the long-term. To further confirm the aphidattractant properties of santalenes and bergamotene, solutions of these terpenes were also utilized in two-choice tests using agar medium plates. Consistent with the transgenic tobacco choice tests, the leaves of wild-type N. tabacum plants treated with santalene and bergamotene solutions also proved to be more attractive to green peach aphids (Fig 4C and S6 Fig).

Discussion
Insects can locate their host plants by detecting specific olfactory signals. Plant volatiles play an important role in this host-location process, and herbivorous insects can distinguish certain host plants from the ratio-specific blends of specific volatiles produced by these plants [39]. In the short-term two-choice experiment conducted in the present study, we found that when given the choice between transgenic and WT tobacco plants, green peach aphids show a preference for the former. Similar results were obtained in long-term choice experiments, with most green peach aphids showing a preference for transgenic tobacco (Fig 4B and 4C).
trans-α-Bergamotene has been identified as an active compound in the sex attractant recognized by male Melittobia digitata [40], and may play a similar important role in the interaction with green peach aphids. However, given that purified standards of α-santalene, trans-α-bergamotene, epi-β-santalene, and β-santalene are currently commercially unavailable, this somewhat limits our scope for further investigating the interaction of each of these compounds with aphids.
The use of trap crops is one of the most effective pesticide-free pest management strategies, in which phytophagous pests are typically attracted away from the main crop [41]. Green peach aphids are very common and a source of considerable damage to agricultural plants. To combat these pests, employment of trap crops is one of the several major control methods [42]. Our short-and long-term choice experiments demonstrate that transgenic tobacco can potentially be deployed to serve as a trap crop for green peach aphid management.
In conclusion, we utilized an optimized HMGR or tHMGR MVA pathway strategy to generate higher amounts of santalenes and bergamotene in transgenic tobacco plants. However, the santalenes and bergamotene thus produced could not be further hydroxylated to santalols and bergamotol by S. album santalene/bergamotene oxidase. In both short-and long-term choice experiments, green peach aphids were strongly attracted to transgenic tobacco lines emitting santalenes/bergamotene, indicating that such lines could be employed as a trap crop to prevent green peach aphid infestation in wild-type tobacco. In the future, we plan to conduct field trials to assess the potential utility of these transgenic lines in pest management.
Supporting information S1