Transformation and Characterization of Δ12-Fatty Acid Acetylenase and Δ12-Oleate Desaturase Potentially Involved in the Polyacetylene Biosynthetic Pathway from Bidens pilosa

Bidens pilosa is commonly used as an herbal tea component or traditional medicine for treating several diseases, including diabetes. Polyacetylenes have two or more carbon–carbon triple bonds or alkynyl functional groups and are mainly derived from fatty acid and polyketide precursors. Here, we report the cloning of full-length cDNAs that encode Δ12-fatty acid acetylenase (designated BPFAA) and Δ12-oleate desaturase (designated BPOD) from B. pilosa, which we predicted to play a role in the polyacetylene biosynthetic pathway. Subsequently, expression vectors carrying BPFAA or BPOD were constructed and transformed into B. pilosa via the Agrobacterium-mediated method. Genomic PCR analysis confirmed the presence of transgenes and selection marker genes in the obtained transgenic lines. The copy numbers of transgenes in transgenic lines were determined by Southern blot analysis. Furthermore, 4–5 FAA genes and 2–3 OD genes were detected in wild-type (WT) plants. Quantitative real time-PCR revealed that some transgenic lines had higher expression levels than WT. Western blot analysis revealed OD protein expression in the selected transformants. High-performance liquid chromatography profiling was used to analyze the seven index polyacetylenic compounds, and fluctuation patterns were found.


Introduction
Bidens pilosa L., which belongs to the Asteraceae family, is an erect annual plant. It is native to South America and is now widely distributed in subtropical and tropical regions of the world. B. pilosa is commonly used as an herbal tea component or as traditional medicine in Latin America, Africa, and Asian countries for treating various disorders, such as inflammation, stomach illnesses, malaria, liver disorders, enteritis, dysentery, diabetes, and hypertension [1][2][3]. In Taiwan, three variants (namely, radiata, pilosa, and minor) of B. pilosa are often used as a folk medicine for curing diabetes. However, better anti-diabetic properties were observed in the plant extract from the variant radiata as compared with the other two variants [4]. To date, around 200 secondary metabolites have been identified from B. pilosa, including polyacetylenes, flavonoids, phenylpropanoids, and terpenes; 383 amino acids (aa) with a predicted pI of 8.55 and predicted molecular weight (MW) of 44122 Da. Our clone is the same length as the published ∆12 oleate desaturase (FAD2-2) mRNA from sunflower (Helianthus annuus) (GenBank accession no. AF251843) [19]. Nucleotide sequence alignment and protein sequence alignment between sunflower FAD2-2 and our BPOD clone revealed 87% and 93% sequence identity, respectively. In addition, the deduced protein sequence from our BPOD cDNA clone also showed homology to other desaturases. Furthermore, three regions of conserved histidine cluster motifs that contain eight histidine residues-HXXXH, HXX(X)H, and HXXHH [27]-were also found in our BPOD clone (Figure 1b). These histidine residues are catalytic sites and proposed to be the ligands for the iron atoms in stearoyl CoA desaturase, a membrane-associated enzyme [27]. The predicted function of BPOD is the conversion of oleic acid to linoleic acid. The other clone, BPFAA (Figure 2), is 1134 bp in length and contains a reading frame of 377 aa with a predicted pI of 8. 19 and predicted MW of 43,935 Da. This clone is the same length as the published ∆12 fatty acid acetylenase (FAA) from sunflower (H. annuus) (GenBank accession no. AY166773; [28]). Nucleotide sequence alignment and protein sequence alignment of sunflower FAA and our BPFAA clone revealed 85% and 92% sequence identity, respectively. FAA is recognized as a triple-bond-forming enzyme, catalyzing the conversion of linoleate into crepenynate through acetylenation at the ∆12 position. It has been reported that overexpression of this enzyme in transgenic soybean seeds leads to the accumulation of crepenynic and dehydrocrepenynic acids, two ∆12-acetylenic fatty acids [28]. Protein sequence alignment of our BPOD and BPFAA clones revealed 61% sequence identity; moreover, the eight conserved histidine residues in the three regions were also found in our BPFAA clone (Figure 2b). These two full-length cDNAs were cloned separately into the Gateway expression vector pK2GWIWG2, resulting in pBPFAA (Supplementary Figure S1a) and pBPOD (Supplementary Figure S1b). These vectors were transformed separately into A. tumefaciens strain LBA4404.

Phylogenetic
Analysis of ∆12-Oleate Desaturase and ∆12-Fatty Acid Acetylenase Genes from Bidens pilosa var. radiata The phylogenetic tree of OD ( Figure 3a) with 99% bootstrap replicates indicates that our BPOD is closest to Helianthus annuus (GenBank accession no. AF251843) and is grouped with other Asteraceae species. Sequences from three families, Asteraceae, Lamiaceae, and Brassicaceae, clearly form three different clades. Although there are not many complete plant FAA genes available in the GenBank database at the moment, we observed that the phylogenetic trees of FAA from all of the Asteraceae species are clustered in one clade, and BPFAA is closest to H. annuus (GenBank accession no. AY166773) ( Figure 3b). Phylogenetic relationships of (a) oleate desaturase (OD) and (b) fatty acid acetylenase (FAA) genes based on their complete coding DNA sequences (CDSs). The numbers in brackets are the accession numbers of the OD and FAA sequences. Molecular phylogenetic analyses were conducted using the Maximum Likelihood method based on the Tamura-Nei model [29]. Bootstrap values at the nodes are the percentages of 1000 replicates.

Confirmation of Putative Transformants
After Agrobacterium-mediated transformation and selection, over 10 putative transformants from each transformation vector were obtained and grown in a greenhouse. Typical images of selection and plant regeneration are shown in Supplementary Figure S2. No morphological differences between the WT and all transformants were observed. To verify the transformants, genomic DNA from the leaves of plantlets was isolated, and PCR analysis was carried out. Among the 14 putative OD transgenic plants, a PCR amplicon of 1366 bp, comprising a partial sequence (234 bp) of the cauliflower mosaic virus (CaMV) 35S promoter followed by a partial sequence (1132 bp) of BPOD, was detected in 13 out of the 14 samples that we examined (Figure 4a). No PCR band was detected in the wild-type (WT) sample or the transgenic OD24 line (Figure 4a). However, a unique PCR amplicon of 0.8 kb (nptII for the kanamycin selection marker) was detected in all 14 transgenic OD lines but not in the WT (Figure 4a). Similarly, a unique amplicon of 764 bp, comprising a partial sequence (234 bp) of the CaMV 35S promoter and a partial sequence (530 bp) of BPFAA, was detected in 13 out of 14 transgenic FAA lines (Figure 4b). Again, a unique PCR amplicon of 0.8 kb (nptII) was detected in all 14 transgenic FAA lines but not in the WT (Figure 4b). Furthermore, genomic DNA from WT and transgenic plants was digested with EcoRI and then probed with the selection marker nptII in order to estimate the copy number of the transgene ( Figure 5). For the transformation vector pBPOD (Figure 5a) or pBPFAA (Figure 5b), no internal EcoRI restriction site in nptII was found. As expected, no hybridization band was detected in the WT sample in either Southern blot. For OD transformants, the hybridization bands in all samples were different, indicating that all 14 OD transgenic plants that we obtained in this study were independent integration events ( Figure 5a). A single copy of the nptII selection marker gene was integrated into the genome of transgenic plants OD5 and OD14; two or more copies of the nptII transgene were found in other transgenic plants (Figure 5a). For the FAA transformants, the hybridization bands in all the samples were different, indicating that all 15 FAA transgenic plants were independent integration events ( Figure 5b). A single copy of the nptII transgene was integrated into the genome of transgenic plants FAA7, FAA10, and FAA14; two or more copies of the nptII transgene were found in other transgenic plants ( Figure 5b). Meanwhile, to determine the endogenous copies of FAA and OD, DNA from WT B. pilosa was digested with various restriction enzymes and probed with OD ( Figure 6a) or FAA (Figure 6b). No internal restriction sites for AflII, EcoRI, NcoI, SacI, or PacI were found in the OD or FAA cDNA sequence. Two to three hybridization bands were detected when the digested DNA sample was probed with OD (Figure 6a), suggesting the presence of two to three copies of the OD gene in the genome of B. pilosa. Four to five hybridization bands were detected when the digested DNA sample was probed with FAA (Figure 6b), suggesting the presence of four to five copies of the FAA gene in the genome of B. pilosa. To examine whether introns were present in our FAA and OD genes, genomic PCR analysis was performed using genomic DNA from the WT as well as plasmids from pBPFAA and pBPOD as templates. As shown in Supplementary Figure S3, the predicted size of the PCR product (1134 bp) from the pBPFAA plasmid (lane 3) is equal or very similar in size to the PCR product from the WT (lane 2), suggesting that our FAA gene may not contain an intron. In addition, the predicted size of the PCR product (1152 bp) from the pBPOD plasmid (lane 6) is equal or very similar in size to the PCR product from the WT (lane 5), suggesting that our OD gene may not contain an intron.

mRNA and Protein Expression in Transgenic Plants
The expression of foreign genes in transgenic plants was first analyzed by qRT-PCR analysis (  To evaluate the presence and accumulation of OD protein in transgenic plants, total protein was extracted from the leaf tissue of selected OD transgenic plants (OD1, OD5, and OD23), which had higher OD mRNA expression levels, as revealed by qRT-PCR analysis (Figure 7a), and subjected to Western blot analysis ( Figure 8). The protein expression of the OD gene in the WT and selected transgenic plants was confirmed by the presence of a 44-kDa band that was specific for the OD protein of B. pilosa (Figure 8). Among the transgenic plants, the highest protein level was observed in OD1, and this is consistent with the highest mRNA expression, as revealed by qRT-PCR analysis (Figure 7a). We also tried to examine the FAA protein level in FAA transgenic plants by using the anti-FAA antibody in the western blot analysis; however, the attempt was unsuccessful (data not shown).

Polyacetylenic Compound Profiling in Transgenic Plants
Transgenic and WT plants were grown to maturity, and one-month-old leaves were excised for HPLC analysis. As shown in Figure 9, the seven polyacetylenic (PA) compounds were present at higher levels in transgenic plants OD2, OD7, OD11, and OD23 than in the WT. Using the same HPLC approach, the amounts of the seven PA compounds in transgenic plants FAA10, FAA14, and FAA18 were found to be higher than those in the WT ( Figure 10). Representative HPLC profiles of the WT and a few transformants are shown in Supplementary Figure S4.

Discussion
Previously, we established optimal conditions for tissue culture, plant regeneration, and Agrobacterium-mediated transformation in B. pilosa var. radiata [30]. In this study, expression vectors (Supplementary Figure S1) containing two full-length cDNAs that encode ∆12-oleate desaturase (pBPOD) and ∆12-fatty acid acetylenase (pBPFAA) were transformed independently into B. pilosa var. radiata according to the protocol detailed in our previous paper [30]. Over 10 transgenic plants were obtained from each transformation vector. Genomic PCR analysis (Figure 4) revealed that all examined FAA and OD transformants, without exception, contained the foreign selection marker gene nptII in the transgenic plant genome. Moreover, Southern blot analysis ( Figure 5) revealed that all of the examined transformants, without exception, contained at least one copy of nptII in the transgenic plant genome, and the T-DNA integration site in each transformant was different. The transformation vector pBPFAA (Supplementary Figure S1a) contains the expression cassettes of FAA cDNA and the kanamycin resistance gene nptII in the T-DNA region, and the transformation vector pBPOD (Supplementary Figure S1b) contains the expression cassettes of OD cDNA and nptII in the T-DNA region. Thus, the copy number of nptII, which is detected in Figure 5, should be the same as that of the OD gene in transgenic OD plants (Figure 5a) or the same as that of the FAA gene in transgenic FAA plants (Figure 5b). However, in transgenic plant FAA11, the presence of nptII was detected, but the chimeric promoter/FAA expression cassette was not (Figure 4b). Similarly, in transgenic plant OD24, the presence of nptII was detected, but the chimeric promoter/OD expression cassette was not (Figure 4a). In brief, incomplete integration of T-DNA into the plant chromosome during the transformation process was found in transgenic plants OD24 and FAA11. Previously, using the transformation vector pCHS, which carries the Petunia chalcone synthase (chs) and nptII genes, we also observed incomplete integration of T-DNA in several plant species, including B. pilosa [30], the floricultural plant Cleome spinosa [31], and the medicinal plant Echinacea pallida [32]. The loss of one of the two transgenes within the same T-DNA has been clearly demonstrated in transgenic wheat [33] and rice [34]. A single T-DNA insertion, with one copy of nptII and the foreign expression cassette (35S promoter/OD or 35S promoter/FAA), was found in transgenic plants OD5, OD14, FAA7, FAA10, and FAA14 ( Figure 5). In addition, no introns are found in our BPFAA and BPOD genes (Supplementary Figure S3). In Brassica napus, no or two introns had been reported in FAD2 genes (equivalent to our BPOD gene in this study) [35]. In Medicago truncatula, three FAD2 genes have been reported; among them, MtFAD2.1 had no intron, while MtFAD2.2 had one intron and MtFAD2.3 had two introns [20]. The exon-intron organization in FAA genes in other plant species has not been reported yet.
In this study, overexpression of FAA or OD did not significantly increase the mRNA expression of those genes, as revealed by qRT-PCR analysis (Figure 7). One possible explanation is the presence of multiple copies of the endogenous FAA or OD genes in the genome of WT B. pilosa ( Figure 6); thus, overexpression of one homologous gene may not affect the total gene expression among those homologous genes. Nevertheless, preliminary HPLC profiling suggested that the seven polyacetylene (PA) compounds identified in transformants OD2, OD7, OD11, OD23, FAA10, FAA14, and FAA18 were slightly higher than in WT (Figures 9 and 10).
The fatty acid desaturase 2 (FAD2) gene encodes an enzyme that catalyzes the desaturation of oleic acid (C18:1) to linoleic acid (C18:2). Oleic acid and linoleic acid are two of the most abundant polyunsaturated fatty acids in plants [17]. For example, in safflower seed oil, oleic acid and linoleic acid together account for about 90% of the total fatty acids [14]. However, higher oleic acid content in oilseed crops is maintained by breeding programs because of the thermal stability of the resulting oil and its suitability as an edible oil [36]. Furthermore, oils containing higher oleic acid are beneficial for lowering cholesterol and reducing blood pressure [37]. Thus, by using the RNA interference (RNAi) technique to suppress the FAD2 gene, higher expression of oleic acid and lower expression of linoleic acid have been reported in several plants, such as Arabidopsis, rice, cotton, and flax [38]. More recently, clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated protein (Cas9) (CRISPR/Cas9) strategies have been developed for mutating the FAD2 genes in oilseed crops. For example, in rapeseed (Brassica napus), two guide RNAs were designed for BnaA.FAD2.a (FAD2_Aa), which is one of four FAD2 genes, followed by CRISPR/Cas9-mediated genome editing. Finally, two mature plants that contained the mutant alleles were obtained. Fatty acid composition analysis of seeds from homozygous lines revealed a statistically significant increase in oleic acid as compared with WT seeds [36]. Interestingly, the CRISPR/Cas9 system was successfully used to mutate all four copies of BnaFAD2 in tetraploid rapeseed (B. napus), and the oleic acid content in the seeds of mutants was significantly increased compared with WT seeds [39]. In parallel, the production of high oleic acid and low linoleic acid due to the disruption of specific FAD2 genes using CRISPR/Cas9-mediated mutagenesis was recently reported in rice [38], peanut [37], soybean [40], and tobacco [41]. It will be interesting to investigate whether the contents of oleic acid and the seven polyacetylenic compounds are affected if our BPOD gene (87% nucleotide sequence and 93% protein sequence identities to the sunflower FAD2-2) is disrupted by RNAi or CRISPR/Cas9 techniques.
In conclusion, herein, two full-length cDNAs that encode ∆12-fatty acid acetylenase (BPFAA) and ∆12-oleate desaturase (BPOD), which have been reported to be potentially involved in the polyacetylene biosynthesis pathway in other plants, were cloned from the medicinal herb B. pilosa var. radiata. These two cDNAs were constructed into plant expression vectors. Then, cotyledons were excised from B. pilosa plantlets that were grown in vitro, and A. tumefaciens infection, subculture, callus and shoot induction, and plant regeneration were conducted. Selection was conducted in an induction medium supplemented with 200 mg L −1 timentin to inhibit Agrobacterium growth and 200 mg L −1 kanamycin for selection. Finally, over 10 putative transgenic lines from each construct were obtained and grown in a greenhouse. A unique band of 795 bp (nptII, which confers kanamycin resistance) was detected in all the transgenic lines that we examined, but not in the WT, when nptII-specific primers were employed in genomic PCR analysis. This suggests that our transformation protocol for B. pilosa var. radiata was successful. The molecular characterization of these transformants was carried out.

Plant Material and Culture Conditions
Seeds of Bidens pilosa var. radiata were sterilized as previously described [30] and then germinated on MS basal medium (MS salts; 2% sucrose; 0.8% Bacto-agar; pH 5.7) [42] in a 25 • C growth chamber under a cycle of 16 h illumination (100 µmol m −2 s −1 ) and 8 h darkness. Explants of in vitro-grown plantlets were excised for Agrobacterium-mediated transformation.

Cloning of Full-Length cDNAs
To clone ∆12-fatty acid acetylenase and ∆12-oleate desaturase cDNAs in B. pilosa var. radiata, degenerate primers were designed based on homologous sequences in the National Center for Biotechnology Information (NCBI). Total RNA of B. pilosa var. radiata was used as a template to synthesize first-strand cDNA by 5 and 3 RACE using a SMARTer RACE 5 /3 Kit (Clontech Laboratories is now Takara Bio USA, Mountain View, CA, USA) according to the manufacturer's instructions. Specific primers for our FAA clone were designed according to ∆12 fatty acid acetylenase from sunflower (Helianthus annuus) (GenBank accession no. AY166773) [28], whereas specific primers for our OD clone were designed according to ∆12 oleate desaturase (FAD2-2) from sunflower (H. annuus) (GenBank accession no. AF251843) [19]. Oligonucleotide sequences of specific FAA primers (FAA-F, FAA-R, cFAA-f1, cFAA-f2, cFAA-r1, cFAA-r2) and specific OD primers (OD-F, OD-R, cOD-f1, cOD-f2, cOD-r1, cOD-r2) are listed in Supplementary Table S1. Protocols for gel purification, cloning into pGEM-T Easy Vector (Promega, Madison, WI, USA), and transformation into Escherichia coli JM109 competent cells were previously described [32]. Plasmid DNA was isolated and then sequenced. The nucleotide sequence that encodes ∆12-oleate desaturase from B. pilosa var. radiata was designated BPOD (1,152 bp) and submitted to NCBI with accession number MF318524, and the nucleotide sequence that encodes ∆12-fatty acid acetylenase from B. pilosa var. radiata was designated BPFAA (1134 bp) and submitted to NCBI with accession number MF318525.

Phylogenetic Analysis
Complete coding DNA sequences (CDSs) of the BPOD and BPFAA genes were studied with other oleate desaturase and fatty acid acetylenase complete CDSs obtained from the GenBank database. Desaturase and acetylenase CDSs were analyzed, respectively, and aligned using Clustal W [43]. Phylogenetic studies were carried out using the maximum likelihood method based on the Tamura-Nei model [29] in the MEGA7 software [44]. Phylogenetic analysis of these CDSs is represented by a bootstrap consensus tree (1000 replicates) in a traditional branch style.

Construction of Expression Vectors and Agrobacterium-mediated Transformation
Two full-length cDNAs that encode BPFAA (1134 bp) and BPOD (1152 bp) were separately cloned into the Gateway expression vector pK2GW7 [carrying neomycin phosphotransferase II (nptII) for kanamycin resistance], resulting in pBPFAA and pBPOD, respectively. These two vectors (i.e., pBPFAA and pBPOD) were transformed separately into Agrobacterium tumefaciens strain LBA4404 by electroporation (Bio-Rad Gene Pulser II, Hercules, CA, USA), and the individual Agrobacterium culture carrying a single transformation vector was used for plant transformation as previously described [30]. Briefly, cotyledon explants were excised from B. pilosa plantlets grown in vitro. Agrobacterium-mediated transformation, subculture, callus induction, and plant regeneration were performed. Selection was conducted in a medium containing MS salts, 2% sucrose, 1 mg L −1 BA, 0.5 mg L −1 IAA, and 0.8% Bacto-agar supplemented with 200 mg L −1 timentin and 200 mg L −1 kanamycin.

Transgenic Plant Verification by Genomic PCR Analysis
Total genomic DNA was extracted from the green leaves of wild-type (WT) and putative transgenic plants using the CTAB method [45]. PCR was carried out with the following primer sets (Supplementary  Table S1): 35S Pro-F1 and OD-R for amplification of a 1366-bp-long DNA fragment corresponding to the partial CaMV 35S promoter and partial OD sequences; 35S Pro-F1 and FAA-r2 for amplification of the 764-bp-long DNA fragment corresponding to the partial CaMV 35S promoter and partial FAA sequences; Kan-F and Kan-R for amplification of the 795-bp-long corresponding to the full-length nptII gene for kanamycin resistance [46]. Conditions for PCR and gel electrophoresis were as previously described [30].

Southern Blot Analysis
For Southern blot analysis, 20 µg of genomic DNA from putative transgenic and WT plants was independently digested with EcoRI. Probe preparation and detection of the non-isotope Digoxigenin (DIG)-labeled PCR product for the kanamycin selection marker nptII were carried out as previously described [30].
To determine the endogenous copy numbers of the OD and FAA genes, 20 µg of genomic DNA from WT plants was digested by various restriction enzymes for AflII, EcoRI, NcoI, SacI, and PacI. No restriction cutting sites were found using these restriction enzymes in the full-length cDNAs of either FAA or OD. For gene-specific probes, primers FAA-F and FAA-R (Supplementary Table S1) were used to amplify the 1134-bp-long DNA fragment corresponding to the full-length cDNA of FAA, and primers OD-F and OD-R (Supplementary Table S1) were used to amplify the 1152-bp-long DNA fragment corresponding to the full-length cDNA of the OD gene. The protocols for PCR amplification, gel purification, blotting onto a nylon membrane, preparing the non-isotope DIG-labeled PCR probe, hybridization, and detection were as previously described [30].

Quantitative Real-Time PCR Analysis
Total RNA was isolated from the green leaves of WT and transgenic plants by CTAB [47]. Quantitative real-time PCR (qRT-PCR) analysis was carried out as previously described [32]. Specific primer sets for the amplification of DNA fragments for FAA (187 bp), OD (234 bp), and ribosomal protein L2 (100 bp; internal control) from B. pilosa are listed in Supplementary Table S1. Three biological replicates were used for quantification.

Western Blot Analysis
One hundred milligrams of leaf tissue from WT and transgenic plants was harvested and frozen immediately in liquid nitrogen; the tissue was ground into a fine powder, and 400 µL of protein extraction buffer (50 mm Tris-HCL, pH 7.0; 0.5 mm EDTA; 1× protease inhibitor) was added in a 2 mL micro-centrifuge tube. The tube was vortexed for 10 s, put on ice for 20 min, and then centrifuged at 16,000 rpm for 30 min at 4 • C. The supernatant was transferred into a fresh micro-centrifuge tube, and protein concentration was determined by DC Protein Assay Reagent (Bio-Rad) with bovine serum albumin as the standard. To produce the antibody against BPOD, oligopeptide (N'-SHR RHH SNT GSI EHD EVF-C') conjugated to the carrier protein OVA was synthesized, HPLC purified, and then used as an antigen to immunize rabbits. Immunization and serum collection were performed by LTK BioLaboratories (New Taipei City, Taiwan). For immunoblot analysis, 30 µg of total protein was electrophoretically separated on 10% SDS-polyacrylamide gel and then transferred onto a PVDF membrane. The blot was incubated with anti-OD antibody or anti-actin antibody (control) and then visualized by incubation with horseradish phosphatase (HRP)-conjugated goat anti-rabbit IgG, followed by chemiluminescent HRP substrate detection. To ensure equal loading of protein, the SDS-polyacrylamide gel was stained with Coomassie Brilliant Blue.

High-Performance Liquid Chromatography Analysis
HPLC analysis was carried out as described previously [4,6] with modifications. Briefly, one-month-old leaves (100 mg) from WT and selected transgenic plants were harvested and frozen immediately in liquid nitrogen; the tissues were ground into a fine powder and transferred into a micro-centrifuge tube. Then, 1 mL of cold 70% ethanol was added into each tube and vortexed for 10 s. After sonication for 20 min (samples were kept in an ice-water bath), tubes were centrifuged for 10 min at 12,000× g. Supernatant from each sample was filtered out by using a 1-mL syringe and 0.2 µm PTFE membrane into a fresh micro-centrifuge tube. HPLC analysis was conducted using an Agilent 1200 Chemstation HPLC system and a C18 reverse-phase column (Phenomenex Luna 5 µ C18, 250 mm × 4.6 mm, Torrance, CA, USA). Polyacetylenic compounds were monitored at 245 nm. The mobile phase consisted of H 2 O (A) and methanol (B), and separations were performed using the following gradients: 60% B from 0 to 10 min, 70% B from 10 to 15 min, 80% B from 15 to 30 min, 90% B 30 to 40 min, 100% B 40 to 45 min, 100% B 45 to 65 min, and 60% B from 65 to 85 min. The injection volume was 30 µL.

Supplementary Materials:
The following is available online at http://www.mdpi.com/2223-7747/9/11/1483/s1: Figure S1: Construction of expression vectors for plant transformation; Figure S2: Typical photos demonstrating shoot regeneration under selection medium and then plant regeneration of transgenic Bidens pilosa plants by transforming two expression vectors and Agrobacterium-mediated method; Figure S3: Examination of PCR products between genomic DNA and cDNA for FAA and OD genes in Bidens pilosa; Figure S4: Representative HPLC profiles of WT and a few randomly selected transformants; Table S1: Primers used for this study. Acknowledgments: The authors thank Miranda Loney for critical reading and editing of the manuscript. We also thank Shih-Hon Lin for technical assistance in the Western blot analysis and Chin-Yu Lin at the Metabolomics Core Facility, Agricultural Biotechnology Research Center, Academia Sinica, for HPLC assistance.

Conflicts of Interest:
The authors declare that they have no conflict of interest.