Bioengineering of the Plant Culture of Capsicum frutescens with Vanillin Synthase Gene for the Production of Vanillin

Production of vanillin by bioengineering has gained popularity due to consumer demand toward vanillin produced by biological systems. Natural vanillin from vanilla beans is very expensive to produce compared to its synthetic counterpart. Current bioengineering works mainly involve microbial biotechnology. Therefore, alternative means to the current approaches are constantly being explored. This work describes the use of vanillin synthase (VpVAN), to bioconvert ferulic acid to vanillin in a plant system. The VpVAN enzyme had been shown to directly convert ferulic acid and its glucoside into vanillin and its glucoside, respectively. As the ferulic acid precursor and vanillin were found to be the intermediates in the phenylpropanoid biosynthetic pathway of Capsicum species, this work serves as a proof-of-concept for vanillin production using Capsicum frutescens (C. frutescens or hot chili pepper). The cells of C. frutescens were genetically transformed with a codon optimized VpVAN gene via biolistics. Transformed explants were selected and regenerated into callus. Successful integration of the gene cassette into the plant genome was confirmed by polymerase chain reaction. High-performance liquid chromatography was used to quantify the phenolic compounds detected in the callus tissues. The vanillin content of transformed calli was 0.057% compared to 0.0003% in untransformed calli.


Introduction
Vanilla is one of the most important flavors in the food and beverage industries, and it is also used in perfumery and pharmaceutical products. Natural vanilla extract from reported that vanillin is able to suppress the proliferation of cancer cells and prevent chemically and physically induced mutagenesis [2]. It was also reported that vanillin exhibits antimicrobial properties [3].
Despite the usefulness of vanilla or more specifically, vanillin, natural vanillin is very expensive to produce. This is largely attributed to the laborious and timeconsuming process to extract vanillin from vanilla beans. In the market, only a small portion of vanilla flavoring is derived from natural vanilla beans due to the low supply of vanilla beans, which is often subjected to extensive price fluctuations. The market price of natural vanilla has recently soared from over USD 200 per kg to USD 400-500 per kg in 2016, which is more than a ten times increase from its lowest price at just USD 20 per kg ten years ago [4]. Only 2000 tons of the global demand, which is more than 15,000 tons, is provided by vanilla beans [5]. The rest is supplied by synthetic vanillin produced from lignin and eugenol. Nevertheless, the market share of natural vanilla is believed to be not affected by their artificial counterparts due to the shift in demand towards food regarded as natural and organic. US and EU labeling regulations allow only goods produced using natural vanilla to be labeled "vanilla" [6,7]. In addition, bioengineered vanillin from plant tissues and microorganisms is still of low success because of the high cost incurred in culture fermentation and the requirement to optimize various culture conditions [8,9].
and Garsson [12]. Similar enzymes in fungi have also been reported by Hansen and co-workers [13]. The enzymes were generally referred to as aldehyde oxidase, CoA ligase, dehydrogenase, hydratase or reductase.
This research explores a plant-based alternative to the current vanillin production systems by the heterologous expression of a VpVAN gene in callus cultures of Capsicum frutescens (C. frutescens) L. var. Hot Lava (chili). Ferulic acid (4-hydroxy-3-methoxycinnamic acid) and vanillin were found to be the precursors for the biosynthesis of capsaicin in chili ( Fig. 1) [14]. Thus, the constitutive expression of VpVAN could potentially enable bioconversion of endogenous ferulic acid to vanillin in the callus cultures of C. frutescens at a higher level compared to the untransformed callus. This would potentially lead to the production of natural, pure vanillin using an alternative bioengineered plant-based system in another food crop.

Bacterial strain and growth media
The expression vector was propagated using One Shot ccdB Survival 2 T1 R chemically competent Escherichia coli (Invitrogen) following transformation by heat shock. The bacteria were cultured on Luria Bertani (LB) agar, then in LB broth containing 100 µg/mL ampicillin and 15 µg/mL chloramphenicol as the selective antibiotics. Purification of the expression vector was carried out using Hybrid-Q Plasmid Rapidprep kit (GeneAll) according to manufacturer's protocol.

Plant material
Seeds of Capsicum frutescens L. cv. Hot Lava were surface sterilized using 70% (v/v) ethanol, followed by washing in sterile distilled water. They were then soaked in 20% (v/v) commercial Clorox (1.05% sodium hypochlorite), and were subsequently washed twice in sterile distilled water. The sterilized seeds were germinated on Murashige and Skoog (MS) agar (4.42 g/L MS basal salt, 20 g/L sucrose, and 3.5 g/L agar (Phytagel)) for two weeks. Hypocotyls of germinated seedlings were excised and incubated in the dark overnight prior to particle bombardment.

Particle bombardment
The expression vector was coated onto 1.6 µm gold particles by mixing of the DNA and the gold particles with spermidine and calcium chloride with constant vortex.
Coated gold particles were pelleted and the resulting supernatant was removed, followed by washing with 70% (v/v) ethanol. Subsequently, the particles were washed with 100% (v/v) ethanol and were resuspended in 100% (v/v) ethanol prior to loading on macrocarriers. Particle bombardment was performed on the explants using a PDS-1000/He system at 1350 psi helium pressure in a vacuum chamber at 28 mm mercury pressure. Three replicates of twenty explants per replicate were subjected to the bombardment. The target distance was set at 6 cm. Bombarded explants were then incubated in the dark overnight for recovery.

Selection and regeneration of putative transformants
After an overnight incubation, the bombarded explants were transferred to selective (v/v) ethanol. Finally, the DNA was resuspended in sterile nuclease free water.

Confirmation of transformants by polymerase chain reaction (PCR)
Amplification of VpVAN by PCR from the extracted genomic DNA of calli was

Extraction of Phenolic Compounds
Phenolic compounds from callus cultures were extracted using maceration and sonication. Two grams of callus tissue was weighed and ground with 80% (v/v) ethanol, followed by sonication. Suspension with ethanol and sonication was repeated twice. All liquid extracts were collected. Extraction solvent was removed by rotary evaporation. Target compounds were then redissolved in 80% (v/v) methanol.

High Performance Liquid Chromatography (HPLC)
All callus extracts were filtered through a 0.45 µm syringe filter prior to injection into the HPLC system. A gradient HPLC was performed with a mobile phase ratio changing from 1:3 to 1:1 methanol-1% acetic acid over 15 min. Flow rate was set at 1 mL/min. The HPLC column used was Hypersil GOLD (Thermo) C18 analytical column (250 × 4.6 mm ID, 5 µm particle size). Target phenolic compounds were detected using ultraviolet (UV) photodiode array at 260 -325 nm wavelengths.

VpVAN-V5)ΔccdB was successfully transformed into chemically competent
Escherichia coli, polymerase chain reaction (PCR) was performed on the bacterial colonies grown overnight on Luria Bertani agar containing ampicillin and chloramphenicol. Colonies that showed the amplification of the 1133 bp VpVAN gene (in Fig. S1) were selected for subculture to propagate the expression vector.
Subsequently, the expression vector that was extracted was subjected to another PCR of the VpVAN gene and to cleavage by PstI and BamHI restriction endonucleases for further verification (in Fig. S2). The double restriction digest gave the expected DNA bands of 6028 bp and 1460 bp. Additionally, the expression vector was sequenced and the resulting reads showed up to 100% sequence identity to that of the known cDNA sequence of VpVAN gene (data not shown).

Selection, regeneration and screening of plant transformants
The particle bombardment procedure with pcDNA6.2::(35Sp-VpVAN-V5)ΔccdB achieved 5% transformation efficiency, whereby one out of twenty explants in each of the three replicates survived and proliferated into callus on the blasticidin selective media over two months (Fig. 3A). Explants that did not survive showed signs of shrinking and extensive browning with very little or no callus proliferation at all (Fig.   3B). Separately, transformation with pcDNA6.2::(35Sp-sGFP)ΔccdB achieved 20% transformation efficiency and the surviving explants (Fig. 3C) showed the expression of sGFP (Fig. 3D). This demonstrated effective expression of the gene cassette transferred with the expression vector. In addition, the sGFP gene was a modified GFP gene with a chromophore mutation at position 65, where serine was replaced with threonine, to give 100-fold higher fluorescence signal compared to the original jellyfish GFP [15].

Integration of the gene cassette into the plant genome was verified by PCR of
VpVAN after the extraction of genomic DNA from the callus tissues. The gene was detected in the genomic DNA of all callus samples, hence indicating successful integration of the gene cassette into the plant genome (Fig. 4). PCR amplification of VpVAN was not achieved from the genomic DNA of non-transformed C. frutescens callus (Fig. 5). On the other hand, genomic DNA from the leaf of V. planifolia
Maximum UV absorbance for the four compounds was measured at 280 nm, 260 nm, 270 nm, and 280 nm wavelengths, respectively. The retention times in chromatograms acquired for the target compounds were compared to those of the external standards ( Fig. 6). Transformed calli produced vanillin at an average of 573.39 (±120.70) µg per gram tissue. This was equivalent to 0.057% of vanillin in the fresh callus. The amount of vanillin produced was significantly higher than that from the untransformed calli, which produced detectable levels of vanillin at an average of only 3.32 (±0.83) µg per gram tissue (0.0003%) (Fig. 7). The increase in vanillin level was 190 times. The VpVAN is able to catalyze the synthesis of vanillin and its glucoside from ferulic acid and its glucoside, respectively, as described by Gallage

Conclusion
The   via the phenylpropanoid biosynthesis, followed by steps from ferulate to vanillylamine and subsequently capsaicin, which is unique to Capsicum [14]. Pathway diagram was generated by MetaCyc (http://metacyc.org/cytoscapejs/ovsubset.html?orgid=META&pwys=PWY-5710).        a negative selectable marker, a lethal cytotoxic ccdB gene, that was originally present downstream of the chloramphenicol resistance gene (Cm R ). Non-resistant E. coli that was transformed with recircularized (non-recombinant) destination vector would be killed with the expression of intact ccdB gene. The presence of the ampicillin resistance gene (Amp R ) and the chloramphenicol resistance gene (Cm R ) allows selection of the bacterial transformants in Luria Bertani agar containing ampicillin and chloramphenicol to maintain the integrity of the vector. Plant transformants would confer resistance to blasticidin in the Murashige and Skoog (MS) medium with the expression of blasticidin S deaminase gene (BSD). Image above was modified from the vector representation diagram generated using SnapGene.