INCREASING DIOSGENIN PRODUCTION THROUGH METABOLIC ENGINEERING OF STEROIDAL SAPONIN BIOSYNTHESIS

(1)

1

INCREASING DIOSGENIN PRODUCTION THROUGH

METABOLIC ENGINEERING OF STEROIDAL SAPONIN

BIOSYNTHESIS

Bachelor Thesis Biology Student: L. (Lars) Tierolf

Swammerdam Institute for Life Sciences

Date: 01-07-2018

Supervisor: dr. L. (Lemeng) Dong Plant Hormone Biology

(2)

2

Abstract

This study aims to explore a strategy for increasing diosgenin production through metabolic engineering of the plant steroid metabolic pathway. In order to shift the metabolic flux toward steroidal saponin biosynthesis, the GAME4 gene located at the branching point between steroidal alkaloid and steroidal saponin biosynthesis from the same substrate was identified to be silenced in Solanum lycopersicum and Nicotiana benthamiana. Part of the steroidal pathway was to be transiently over- expressed in N. benthamiana. GC-MS was to be used for metabolite quantification. GAME7, GAME8, GAME11, GAME6 and GAME4 were selected for this research, sequences were retrieved from public databases. Primers were designed in order to create cDNA using RNA isolated from S. lycopersicum and Solanum tuberosum as template. This cDNA was used to construct transient expression and gene silencing plasmids with pEAQ and pYL156 vectors, respectively. Transformation of Agrobacterium tumefaciens GV3101 strain with the constructs failed or could not be confirmed with colony PCR. Transformation of A. tumefaciens EHA105 was performed and confirmed through colony PCR. Agro-infiltration and VIGS treatment of S. lycopersicum was performed, however not enough time for phenotype development was available. Agro-infiltration and metabolite analysis in N. benthamiana could not be performed.

(3)

3

1. Introduction

1.1 Diosgenin

Steroidal saponins are secondary plant metabolites and are considered the active constituents of many plant derived folk medicine (Sparg, Light & van Staden, 2004). Diosgenin ((25R)-5-spirosten-3H-ol, Pubchem ID 99474) is a steroidal saponin found in several plant species, including Dioscorea nipponoca (yam), Chlorophytum borivilianum (safed musli), Helicteres isora L, Trillium

govanianum (nagchhatri) and Trigonella foenum graecum (fenugreek) (Sharma, Malhotra, and Sood,

2016; Kumar, Desai, and Shriram, 2014; Patel et al, 2012). Pharmaceutical industry has a great interest in the production of diosgenin, since it is explored as a precursor molecule for various hormonal drugs. The pharmaceutical company Syntex started large scale production of progesterone derived from diosgenin extracted from Dioscorea mexicana in the late 1940s, and manufactured the first oral contraceptives with processed diosgenin in 1951 (Renneberg, Berkling & Loroch, 2017). The diosgenin aglycon is also known to have anticancer, anti-inflammatory, and anti-infectious activities (Jesus et al., 2016). Patel e al. (2012) suggest diosgenin could be used to treat numerous types of disorders in the future, including cardiovascular diseases and various types of diabetes.

Large scale production of diosgenin is at present realised through cultivation of and extraction from Dioscorea zingiberensis and Dioscorea mexicana tubers (Li et al., 2018). However, diosgenin production from Dioscorea species is time and cost consuming, due to the years required for the tubers to acquire the sufficient diosgenin content needed for pharmaceutical processing. This makes the rate of diosgenin accumulation the price determining factor for diosgenin derived medicine (Vaidya, 2013). Developing alternative approaches for diosgenin production may provide opportunities for decreasing the price of diosgenin derived medicine through lowering the costs of plant cultivation. Metabolic engineering of the biosynthetic pathways in plants can be used as a tool for enhancing secondary metabolite production. Jiang et al. (2016) realised a twofold to threefold increase of the antimalarial drug artemisinin in Artemisia

annua through overexpression of a gene in the terpenoid biosynthetic pathway. Mahmoud and

Croteau (2001) increased essential oil yield in peppermint through cutting one branch of the mono-terpenoid metabolic network through gene silencing, changing the metabolic flux towards the menthol pathway. Thus, metabolic engineering of the steroidal saponin biosynthetic pathway in plants may provide new strategies for increasing diosgenin production.

(5)

5

Itkin et al. (2013) proposed a metabolic pathway for steroidal saponin biosynthesis in L.

lycopersicum (figure 2). This pathway shows resemblance to the pathway of diosgenin biosynthesis

in fenugreek proposed by Ciura et al. (2017) (figure 1). In L. lycopersicum, GLYCOALKALOID METABOLISM (GAME)-7, GAME8, GAME11, and GAME6 synthesize a furostanol type saponin aglycone. GAME4 converts the furostanol type saponin aglycone to furostanol-26-aldehyde, which is further converted to various steroidal alkaloids. Silencing of GAME4 led to an accumulation of steroidal saponins, since both the steroidal saponin and steroidal alkaloid biosynthesis pathway compete for the same furostanol type saponin aglycone substrate (Itkin et al., 2013). The same steroidal pathway is involved in diosgenin biosynthesis (Ciura et al., 2017).

GAME7, GAME8, GAME11. GAME6 and GAME4 are known to be present in L. lycopersicum

and L. tuberosum (Itkin et al., 2013).

This study aims to test an accumulation of steroidal saponins after gene silencing of the enzymatic step at the branching point (GAME4) between steroidal saponin and steroidal alkaloid biosynthesis in S. lycopersicum and in N. benthamiana. Another approach to increasing steroidal saponin production consists of overexpressing the steroidal saponin biosynthetic pathway in S.

lycopersicum and S. tuberosum into N. benthamiana. It is hypothesised that silencing of the gene leads

to a change of metabolic flux from steroidal alkaloid to steroidal saponin production, and that transient overexpression of the biosynthetic genes involved in the steroidal saponin pathway in

(6)

6 1.2 Theoretical Background

1.2.1 Metabolic pathways involved in Diosgenin biosynthesis

Three metabolic pathways are involved in diosgenin biosynthesis, namely glycolysis, the mevalonate, and steroid pathway, and cholesterol was found to be a precursor of diosgenin (Mehrafarin et al., 2010). However, not much is known on the biosynthetic steps occurring between cholesterol and diosgenin. First squalene is transformed into cholesterol, a precursor for most steroidal saponins (Mehrafarin et al., 2010). Ciura et al. (2017) discovered genes regulating the pathway from cholesterol to diosgenin in Trigonella foenum graecum through elicitation of secondary metabolite production with methyl jasmonate and precursor feeding with cholesterol and squalene for representational difference analysis of cDNA. A detailed pathway of diosgenin biosynthesis from cholesterol was hypothesised and is shown in figure 1.

Li et al. (2018) performed an extensive transcriptome study in leaf and rhizome tissue of

Dioscorea zingiberensis in order to identify genes regulating the biosynthetic steps in the pathway

from cholesterol to diosgenin. In this study the transcriptome of tissue containing a relatively high diosgenin content, the rhizome, was constructed and compared with the transcriptome of tissue containing a relatively low diosgenin content, the leaf. The transcriptome was then annotated using public databases. Li et al. (2018) proposed that site specific oxidation of cholesterol by cytochromes P450 (CYPs) yields diosgenin.

Chaudhary (2015) found, through elicitation of fenugreek plant defence mechanisms with methyl jasmonite, increased transcription levels of hydroxymethylglutaryl-CoA reductase (HMGR), a pivotal enzyme in the mevalonate pathway, and increased transcription levels of sterol-3-ß-glucasyl transferase (STRL), a key enzyme in the pathway of diosgenin biosynthesis not via cholesterol but via sistosterol from cycloartenol to sterol-3-ß-D-glucoside. The latter is converted to diosgenin by sterol-3-ß-glucasyl transferase (STRL).

Identification and annotation of genes regulating diosgenin biosynthesis contributes to the scientific knowledge required for the metabolic engineering of plants for higher diosgenin production. In the next section, strategies for increasing diosgenin production found in literature are reviewed, and a strategy for this experimental study is discussed.

(7)

7

1.2.2 Strategies for Increasing Diosgenin production

Ye et al. (2014) cloned and expressed a putative squalene synthase gene, named DzSQS, from

Dioscora zingiberensis into Escherichia coli and found that after incubation with farnesyl diphosphate

squalene was present in the in vitro reaction mixture. Enhancing the diosgenin yield in D.

zingiberensis cultivars through overexpression of DzSQS is discussed. Another research group

discusses the possibility of cloning the diosgenin biosynthesis pathway from D. zingiberensis into yeast or other microbes in order to produce diosgenin without the need for actually cultivating the plant (Hua et al., 2017). Their paper aims to aid the discovery of genes related to secondary metabolite pathways and the engineering of plants with enhanced active ingredients. Other strategies for increasing diosgenin production include hairy root induction in Helicteres isora L., inducing secondary metabolite production in micro-tubers of Chlorophytum borivilianum with jasmonic acid and salicylic acid, and enhancing diosgenin production in Dioscorea zingiberensis cell cultures through Palmarumycin C13 from an endophytic fungus (Kumar, Desai & Shriram, 2014; Chauhan, Keshavkant & Quraishi, 2018; Mou et al., 2015).

Since genomic resources for fenugreek and wild yam species have not yet been fully developed, and the time available for this project is limited, the chose was made to make use of knowledge on pathways leading to steroidal saponin biosynthesis in tomato and potato. Steroidal saponins were found to increase after gene silencing of a cytochrome P450 encoding gene, named GLYCOALKALOID METABOLISM 4 in Solanaceous crops (Itkin et al., 2013). The proposed pathway of steroid biosynthesis from cholesterol by Itkin et al (2013) shown in figure 2 shows resemblance to the proposed pathway of diosgenin biosynthesis of Ciura et al (2017). The intermediate compound (3β, 25R)-furost-5-en-3,22,26-triol is present in both pathways. It is assumed that the pathway leading to steroidal saponin biosynthesis may be homologous to steroidal saponin biosynthesis in diosgenin producing plant species.

It can be inferred that gene silencing of functional homologs in other plant species leads to an accumulation of various steroidal saponins, which in the case of fenugreek and wild yam species may including diosgenin. In order to further investigate the role of the various GAME genes, (1) N. benthamiana is transformed through agro-infiltration, and (2) GAME4 gene in S.

(8)

8 Figure 2. This pathway involved in steroidal saponin biosynthesis proposed by Itkin et al. (2013) shows resemblance to the pathway proposed by Ciura et al. (2017). Both include 22,- and 26C hydroxylation and E-ring closure. Gene GAME7, GAME8, GAME11, GAME6, and GAME4 are selected for further analysis. Figure reprinted from “Biosynthesis of antinutritional alkaloids in solanaceous crops is mediated by clustered genes”, by Itkin et al., 2013, Sciencexpress, 1240230.

Figure 1.Biosynthetic steps leading to diogenin from cholesterol hypothesized by Ciura et al. (2017). The figure was created through differential cDNA analysis after elicitation of steroidal saponin biosynthesis with Figure reprinted from “Next-generation sequencing of representational difference analysis products for identification of genes involved in diosgenin biosynthesis in fenugreek (Trigonella foenum-graecum)”, Ciura et al., 2017, Planta, 245(5), pp 977- 991. Creative Commons Attribution 4.0 International License.

(9)

9

1.2.3 Virus Induced Gene Silencing (VIGS)

Virus induced gene silencing (VIGS) methods make use of the RNA-mediated defence mechanisms in plants against viruses (Baulcombe, 1999). When a plant is infected with a genetically modified virus containing a RNA fragment similar to an endogenous gene, its RNA mediated defence mechanisms will mount a defensive response. The method makes use of the RNAi pathway present in plants to knock down genes selected by the researcher.

The VIGS process, as described by Baulcombe (1999) and Gould & Kramer (2007), starts with Agrobacterium mediated transformation of the plant with a modified virus vector containing a cDNA fragment homologous to the endogenous gene selected for gene silencing. Double stranded (ds) RNA is transcribed from the viral vector as the vector spreads through the plant. DICER- like enzymes degrade the viral dsRNAs into small interfering siRNAs, which in turn function as template for RNA-induced silencing complex (RISC). Since this template was designed with a nucleotide sequence homologue to the gene targeted for gene silencing, endogenous mRNA transcribed from this gene is degraded through RISC as well, hereby eliminating gene expression.

Various virus vectors are available for application in VIGS. In this study we make use of the protocol provided by A. Abd-El-Haliem (2017), describing the application of the Tobacco rattle virus (TRV) based pYL156 vector. A helper TRV-RNA1 plasmid encodes for proteins required for virulence, while the VIGS fragment of the gene targeted for silencing is cloned into the multiple cloning site of pYL156 (Liu et al., 2002). A combination of both vectors is needed for VIGS treatment. For positive control, a pYL156 - phytoene desaturase (PDS) construct is mostly used, inhibiting carotenoid biosynthesis, resulting in a phenotype of photo-bleached leafs. Another positive control construct is pYL156 - enhanced green fluorescent protein (eGFP). The resulting phenotype is green fluorescence of affected plant tissue. A negative control is usually performed with a pYL156 -GUS construct, a gene without any homologs in plans and generally of similar length as the VIGS fragment used in the experimental condition. Infection with an empty pYL156 vector usually results in stem lesions and foliar necrosis, hereby potentially biasing the results of the VIGS experiment (Wu, Jia & Goggin, 2011). The pTRV-RNA1 helper plasmid is from here on referred to as the TRV1 plasmid, and the pYL156 vector carrying the insert is from her on referred to as the TRV2 vector.

(10)

10

1.2.4 Transient gene expression through agro-infiltration

Agroinfiltration is a method used for inducing transient gene expressing in plants, through utilizing the properties of T-DNA present in gall inducing Agrobacterium tumefaciens. A. tumefaciens are soil born tumour inducing bacteria, and are pathogenic to plants. Upon infection with a host plant, the bacteria transfers part of its DNA, a specific region on the tumour inducing (Ti) plasmid called transfer DNA (T-DNA), into the host plant cells. Proteins encoded on the

virulence-region of the plasmid are essential for this process (Hoekema et al., 1983). There are two

main strategies for manipulating the T-DNA of the Ti-plasmid, either by cloning the gene of interest onto the same plasmid as the virulence-genes, or by cloning the gene on a separate replicon from the virulence genes (Gelvin, 2003). This last method yields a so called binary vector, such as the pEAQ vector used in this study.

The pEAQ (Easy And Quick) vector series facilitates the production of recombinant proteins in plants through exploiting the Cowpea Mosaic Virus hypertranslational ‘‘CPMV-HT’’ expression system (Peyret & Lomonossoff, 2013). This vector increases protein production through increased translation rather that increased replication. The pEAQ-HT--DEST1- USER vector used in this experiment has USER-cassette incorporated in the vector sequence (figure 3). This cassette facilitates assembly of the gene insert into the pEAQ vector through uracil excision- based cloning. Uracil excision- based cloning is a PCR- based DNA engineering technique capable of seamless assembly of different fragments of DNA in one construct. This method applies PCR amplifications of inserts with primers containing uracil usually eight nucleotides from the 5’end. Treatment with the USER®_{enzyme excises the uracil and generates a} single nucleotide overhang in the insert. This overhang is designed compatible with the overhang generated through restriction of specific endonuclease sites in the vector, and when mixed the insert hybridises with the vector without the use of a DNA ligase enzyme (Bitinaite et al., 2007; Salomonsen, Mortenses & Halkier, 2014).

(11)

11

2. Methods and Materials

Two approaches for enhancing steroidal saponin production are explored. (1) agroinfiltrating part of the steroid pathway starting from cholesterol from S. lycopersicum 'Moneymaker' and S.

tuberosum into N. benthamiana, and (2) gene silencing of the branch point gene GAME4 for

steroidal alkaloid formation in S. lycopersicum and N. benthamiana.

Nucleotide sequences of genes involved in steroidal saponins biosynthesis in Solanaceae plants identified by Itkin et al. (2013) were retrieved from the SPUD DB and Sol Genomics Network database (Hirsch et al., 2014; Muelller et al., 2005). A BLAST search was performed aiming to identify homologs in N. benthamania, results are shown in table 1. Open reading frames were identified and translated into amino acid sequences in order to check for protein functionality. Supplementary data, including gene sequences, NanoDrop results, sequence data obtained from Macrogen sequencing services in CLC format and annotated gel electrophoresis images, is available at https://drive.google.com/open?id=13aVQWW6wI4iGEr3A6-UM2Kz_CzNz-TAE. A zipped version of the folders located on Google Drive can be found at:

https://figshare.com/s/56509faf60bd7ba8b786

2.1 Construction of the transient expression construct with pEAQ vector and the various GAME genes

RNA was isolated from a mixture of young and old leaves from a 6-week old potato plant, and from leaves from a 3.5 week old tomato plant. Leaves were grinded in liquid nitrogen, 1 ml of trizol was added to 100mg of leave material. The QIAGEN RNeasy Mini Kit was used following the protocol provided by the manufacturer. TURBO RNAse was added to the spin column. Isolated RNA was loaded on a 1% agarose gel stained with EtBr to check the RNA content. cDNA was synthesized with 1 µg of RNA template with the ThermoFisher RevertAid RT Kit following the protocol provided by the manufacturer. 20 µl of reaction mixture was incubated in 42°C for 60’ and 70°C for 10’. Primers for GAME4, GAME6, GAME7, GAME11 (tomato), and GAME8a, GAME8b (potato) including USER cassettes were used to amplify the various

GAME genes from the acquired cDNA, sequences are shown in table 2. ThermoFisher Phusion

U Hot Start DNA Polymerase was used in the reaction, PCR program is shown in table 5. A 50 µl reaction mixture was prepared following the protocol provided by the manufacturer, PCR products were checked and extracted from 1% agarose gel stained with EtBr with the use of the

(12)

12

ThermoFisher Gel Extraction kit following the protocol included by the manufacturer. Purified

GAME cDNA was ligated into the USER cassette of the pEAQ-HT-DEST1-USER vector with

the New England BioLabs inc. USER enzyme, yielding pEAQ+GAME4, pEAQ+GAME6, pEAQ+GAME7, pEAQ+GAME8a, pEAQ+GAME8b, and pEAQ+GAME11 constructs. A general overview of the pEAQ-HT-DEST1-USER vector is shown in figure 3.

2.2 Transformation of E. coli cells with the pEAQ constructs

Competent E. coli cells were transformed through heat shock with the pEAQ vector plus insert and cultured overnight in solid LB (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, 15 g/L Daishin agar) containing kanamycin (50 µg/mL) in 37ºC on a Petri. Dish. Four colonies were selected for colony PCR, primers amplifying the insert region of the pEAQ vector were used, sequences are shown in table 4. Colonies positively transformed with the pEAQ vector plus insert were inoculated in 5 ml of liquid LB (10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl) medium containing kanamycin (50µg/mL). The E. coli was cultured overnight in 37ºC at 300 rpm. Plasmid DNA from the overnight cultures was isolated with the GENEJET Plasmid Miniprep kit. Nanodrop was used to estimate the DNA concentration, and plasmid DNA combined with pEAQ forward and reverse primer amplifying the region of the insert was sent to Macrogen for sequencing.

2.3 Transformation of A. tumefaciens GV3101 with the pEAQ construct

Electro competent Agrobacterium tumefaciens of the GV3101 strain were transformed with an empty pEAQ vector for control, and pEAQ vectors containing GAME4, GAME6, GAME7, GAME8a, GAME8b and GAME11 inserts with the GENE PULSER Electroporation System in Gene Pulser®/MicroPulser™ Electroporation Cuvettes with a 0.1 cm electrode gap. 50ng of vector DNA was added to 50μl of electro competent A. tumefaciens per construct. An electro shock of 2.5 kV was used for transformation, and electrode arcing was observed in all transformations. Transformed cultures were plated on solid LB medium (50µg/ ml kanamycin, 50µg/ ml rifampicin, 10µg/ ml gentamycin). Petri dishes were incubated at 28ºC for two days, and left on the bench for room temperature incubation for an extra two days. Colonies were picked up from the plate and boiled in 10 µl of milli-Q water at 95ºC for 12 minutes to lyse the

(13)

13

cells. 3 µl of boiled product was used as template in a PCR mixture with Thermo Scientific DreamTaq DNA Polymerase following the protocol provided by the manufacturer. Colonies were inoculated in 3 ml of liquid LB (50µg/ ml kanamycin, 50µg/ ml rifampicine, 10µg/ ml gentamycin). After overnight culture growth, 20 µl of liquid culture was boiled for 12 minutes at 95ºC. 1 µl of boiled product was used in another colony PCR with the ThermoFisher Phusion U Hot Start DNA Polymerase enzyme. A 20µl reaction mixture was prepared following the protocol provided by the manufacturer. PCR product was checked on a 1% agarose gel stained with EtBr.

Another electroporation with newly isolated purified plasmids was performed one week later on A. tumefaciens EHA105 strain. A 1.8kV 400Ω shock was applied and the same electroporation cuvettes with a 0.1 cm electrode gap washed with ethanol were used. All samples exhibited arcing, except for isolated plasmid pEAQ+GAME11(1) and pEAQ+GAME11(2). Samples were grown on agar plates and checked through colony PCR with BioLine MangoMix™ in a 10μl PCR reaction mixture.

2.4 Construction of the TRV2 vector

RNA for cDNA synthesis was isolated from N. benthamiana and S. lycopersicum with use of the QIAGEN RNeasy Mini Kit. cDNA was created with the ThermoFisher RevertAid cDNA Synthesis Kit, and 300bp long VIGS sequences were amplified with two different primer pairs for both N. benthamiana and S. lycopersicum. The first primer pair did not include a restriction site, the second primer pair included a restriction site sequence for the SmaI endonuclease protected with a three nucleotide addition on both sites of the restriction site to improve restriction efficiency.

After gel extraction of the VIGS-fragment, both the TRV2 pYL156 vector provided by Dinesh-Kumar lab and PCR product of amplified cDNA with primer pair 2 were cut with SmaI restriction enzyme. 2 µl of vector DNA, 1µl of SmaI and 1 µl of Thermo Scientific 10X Tango Buffer were mixed in 6µl of Milli-Q water. The mixture was incubated at room temperature for 2 hours. PCR product of both primer pair one and primer pair two was ligated into the cut TRV2 vector with Thermo Scientific T4 DNA Ligase. The reaction mixture was incubated overnight in room temperature, after which the ligation mixture was incubated at 60°C for 10’. Only the

(14)

14

VIGS fragment for gene silencing in S. lycopersicum was ligated into the TRV2 vector, amplification of the GAME4 homolog in N. benthamiana showed no positive result on the gel.

2.5 Transformation of E. coli and A. tumefaciens cells with the TRV2 vectors

Competent E.coli cells were transformed through heat shock with the TRV2 vector containing

GAME4 insert treated with, and without SmaI restriction, and with the TRV1 vector containing

the DNA required to facilitate viral RNA transcription. E. coli was also transformed with a TRV2+GUS, TRV2+PDS, TRV2+eGFP construct as negative controls. After plating and overnight incubation in solid LB-medium containing kanamycin (50 µg/mL), colonies containing TRV2 with insert were selected through colony PCR with TRV2 primers amplifying the multiple cloning site. Selected colonies were inoculated in 5 ml liquid LB-medium containing kanamycin (50µg/ ml) for overnight culture growth in 37°C at 200 rpm. Plasmids were isolated with the GENEJET Plasmid Miniprep kit, autoclaved milliQ water was used in the final dilution step. DNA concentrations were estimated through NanoDrop, and plasmids were send to the Macrogen sequencing service.

After sequence analysis in CLC work-bench, plasmids containing the correct insert were used for transformation of A. tumefaciens from the GV3101 strain. Electro competent cells were transformed with the TRV2 vector containing the GAME4 VIGS sequence, a TRV2 vector plus eGFP, a TRV2 vector plus PDS, a TRV2 vector plus GUS, an empty TRV2 vector for control and a TRV1 vector needed for virulence. Agrobacteria were transformed using the GENE PULSER II Electroporation System at 2.2 kilo Volt at 400 Ohm. Transformed colonies were plated on LB plus agar (50µg/ ml kanamycin, 50µg/ ml rifampicin, and 10µg/ ml gentamycin). Colonies were grown for two days in 30°C, and for two more days at room temperature. A colony PCR was performed with ThermoScientific DreamTaq polymerase. Colonies were inoculated in 3 ml of liquid LB medium ( 50µg/ ml kanamycin, 50µg/ ml rifampicin, and 10µg/ ml gentamycin). The inoculate was grown overnight at 30°C at 200 rpm. A colony PCR was performed again on the overnight liquid culture, of which 30 µl was heated to 95°C for 12 minutes. 1 µl of heated product was used as template following the ThermoFisher Phusion Hot Start DNA Polymerase enzyme protocol for a 20 µl reaction mixture.

One of each four-day old cultures of Agrobacterium containing the TRV2 vector plus PDS, GUS, eGFP and GAME4 inserts were selected for agro-infiltration in S. lycopersicum. One day prior to the procedure, 200 µL of bacterial culture was inoculated in 10 ml of fresh LB medium (50µg/ ml kanamycin, 50µg/ ml rifampicin) for colony refreshment. After overnight growth,

(15)

15

cells were spun down at 4000 rpm for 10 minutes, and cell pellet was resuspended in MMA buffer (10mM MgCl2, 10mM MES, 400µM acetosyringone). MMA was added until an OD of 2.4 was reached. Tubes were incubated for 3 hours at room temperature in dark. After incubation, the culture transformed with the TRV1 vector was further diluted until an OD of 1.2 was reached. 1 ml of each resuspended Agrobacterium culture transformed with TRV2+PDS, eGFP, GUS, and GAME4 was mixed with 1ml of resuspended Agrobacterium transformed with TRV1 vector. A needleless syringe was used to agro-infiltrate three replicate plants per gene construct.

2.6 Gene analysis and sequence alignment

Gene ontologies were obtained from the Sol Genomics Network Database, the Spud DB Potato Genomics Resource. NCBI Blast was used in search for homologs.

Sequence data obtained through the SILS Sequencing service was analysed and aligned with sequences obtained through the Sol Genomics Network database with the use of QIAGEN Bioinformatics CLC Main Workbench software. Sequences of short length (<1000bp) and sequences with trace data indicating an error in colony selection or sequencing error were discarded. Alignments were constructed, and sequences were checked for point mutations, deletions, and insertions. Plasmids without mutation were selected for Agrobacterium transformation. However, due to errors in sequencing, no alignment of GAME7 could be constructed.

(16)

16 2.7 Tables and Figures

Table 1. Genes and homologs in tomato, potato, and in tobacco (found in the SolGenomics database). As defined by Itkin et al., 2013.

Table 2. Primers designed with Primer BLAST from NCBI.

Table 3. VIGS sequences designed with the VIGS tool provided by Sol Genomics Network.

Gene Species VIGS Sequence Fw Rev Tm (fw-rev) BP

product GAME4 Niben101Scf 02437g0001 1.1 N. benthamiana ATGGAGTACTACAATTTAGCTATCTTCTACACAG TTTTGGCAGTAGGGGTTTTAGCTCTTTATAGTAT ACTAAAGAGAGCAAATGGATGGTTTTATTCAATC AAATTTGGTTCAAAGAAATATCTCATACCTCCAG GTGAAATGGGCTGGCCATTTATTGGAAATACTC TTTTTTTCTTCAGTTTTGCTGGCGATCATGGCTC ATTTCTATCCTACTTTTCTACTAGGTTTGGGCCA GGAGGGATGTACAAGGCACACATATTTGGAAAG CCAACCATTATCATTACAAAGCCAGAAACA (1)ATGGAGTA CTACAATTTAG CTATCT (2)TCCCCCGG GGAATGGAGT ACTACAATTTA GCTATCT (1)TGTTT CTGGCTT TGTAATG ATAAT (2)TCCC CCGGGG GTGTTTC TGGCTTT GTAATGA TAAT 55.13 - 55.08 300 GAME4 Solyc12g006 460 S. lycopersicum ATGGATTTCTACAATTTAGCCTTGTTCTTCATAG CTTTAATACTTGGAATTTTCACATTTTATGCCATA TTAATGAGAATAAATGGTTGGTATTATGCAATCA AATTTTGTTCAAACAAATATAACATCCCAAATGG TTATATGGGTTTGCCATATTTTGGTAACACACTT TCTTACTTCAAAGCTTCAATGTGTGGTGATCCAA AATCATTCATTGATTTCTTTGCTACTAGGTTTGG AGAAGGAGGAATGTATAGGGCATACATATTTGG GAAGCCAACAATTATGGTGACAAAGCCA (1)ATGGATTTC TACAATTTAGC CTTG (2)TCCCCCGG GGGAATGGAT TTCTACAATTT AGCCTTG (1)TGGC TTTGTCA CCATAAT TGT (2)TCCC CCGGGG ATGGCTT TGTCACC ATAATTG T 55.47 – 56.04 300

1_{Annotated as 1-aminocyclopropane-1-carboxylate oxidase, so probably not a true homolog.}

Gene S. lycopersicum S. tuberosum N. benthamiana Annotation

GAME4 Solyc12g006460 PGSC0003DMG400024274 Niben101Scf02437g00011.1 Cytochrome P450 superfamily protein

GAME6 Solyc07g043460 PGSC0003DMG400011750 Niben101Scf00748g01005.1 Cytochrome P450 superfamily protein

GAME7 Solyc07g062520 PGSC0003DMG402012386 Niben101Scf00031g01003.1 Niben101Scf05615g03001.1 Cytochrome P450 superfamily protein GAME8a NA PGSC0003DMG400026594 NA Cytochrome P450 superfamily protein GAME8b NA PGSC0003DMG400026586 NA Cytochrome P450 superfamily protein

GAME11 Solyc07g043420 PGSC0003DMG400011751 Niben101Scf00748g00004.11

2- oxoglutarate-dependent dioxygenase

Gene Species Fw Rev Tm (fw-rev) BP product

GAME4 S. lycopersicum ATGGATTTCTACAATTTAG

CCTT TTAATTACCATTACTAGAGATCGATAG 54.02-54.36 1464

GAME4 N. benthamiana ATGGAGTACTACAATTTAG

CTATCTT CTAGGCTAAAAGCTTCTTGAATCT 55.80 - 56.76 1446

GAME6 S. lycopersicum ATGGCTATTGCAATTTTCA

TAGCT TCATTCTAGCTTTCGAACAATCATAG 57.58 - 57.64 1551

GAME11 S. lycopersicum ATGGCGGATCTTCTCTCGA

A TTAATTAGCTTCTGTTTTGAAGGGC 58.6 - 58.31 1029

GAME8a S. tuberosum ATGGCAATTGGAACAGTAA

TTGG TCAAAGCTTGTTGAGGATTATCTG 57.91 - 57.25 1497

GAME8b S. tuberosum ATGGCAATTGCAACAGTAA

TTGG TCAAAGCTTGTTGAGGATTATCTG 58.74 - 57.25 1497

GAME7 S. lycopersicum ATGGACGAAATTCAAATAT

(17)

17 Figure 3 Overview of the pEAQ-HT-DEST1-USER vector. The various GAME genes were ligated into the USER cassette. Primers amplifying the insert region are shown in this figure as F1seq_pEAQ and R1seq_pEAQ.

Figure 4 General overview of the pYL156 vector used for virus induced gene silencing. GAME4 VIGS fragment was ligated into the Multiple cloning site. Primers amplifying the insert region are shown in this figure as forward and TRV2-reverse.

(18)

18

Table 4 Primers used for insert region amplification

Vector Fw Rev Tm

pEAQ-HT-DEST1 GCTTCTGTATATTCTGCCCAAATTCG CCGCTCACCAAACATAGAAATGC 55

TRV2 pYL156 GTGTGTCAACAAAGATGGAC CTTCAGACACGGATCTAC 52

Table 5 PCR program used with Phusion U HS and gene specific primers

Step Temperature Time

1 98°C 1’ 2 98°C 20’’ 3 55°C 30’’ 4 72°C 1’30’’ 5 > Step 2 35 cycles 6 72°C 7’ 7 4°C ∞

Table 6 PCR program used for colony PCR with DreamTaq and pEAQ primers

Table 7 PCR program used for colony PCR with DreamTaq and TRV2 primers

(19)

19

3. Results

3.1 Gene analysis results

GLYCOALKALOID METABOLISM 7 (GAME7)

GAME7 encodes a member of the cytochrome P450 proteins. GAME7 is located on

chromosome 7 in S. lycopersicum. The total gene length is 3,916 base pairs, without introns the mRNA open reading frame is 1,661 nucleotides long. The protein encoded by GAME7 is 520 amino acids long and catalyzes hydroxylation of C22 of cholesterol, the first step in the steroidal alkaloid and saponin biosynthetic pathway in S. lycopersicum (figure 1). Homologs identified in S.

tuberosum and N. benthamiana are shown in table 1.

GLYCOALKALOID METABOLISM 8a and 8b (GAME8a and 8b)

GAME8a and GAME8b both encode a member of the cytochrome P450 proteins, and

catalyzes further hydroxylation at C26, yielding 22,26-dihydrocholesterol (figure 1). GAME8a is located on the reverse strand chromosome 6 in S. tuberosum and has a genomic sequence of 2,095 nucleotides. The open reading frame in the mRNA transcript has a length of 1,497 nucleotides. The protein encoded is 498 amino acids long. GAME8b is located on the forward strand on chromosome 6 and has a genomic sequence of 2,145 nucleotides, with the mRNA open reading frame having a length of 1,497 nucleotides. The protein has a length of 498 amino acids. No homologs in nucleotide sequence were identified in S. lycopersicum, although the protein sequence is highly conserved within the Solanaceous plant species.

GLYCOALKALOID METABOLISM 6 and 11 (GAME6 and GAME11)

GAME6 and GAME11 facilitate C16 oxidation and E-ring closure of the previously formed

22,26-dihydrocholesterol, yielding a furostanol-type saponin aglycone (figure 1) in Solanaceous plant species. GAME6 is located on chromosome 7 in S. lycopersicum and has a genomic sequence of 2,120 nucleotides. The open reading frame in the mRNA transcript has a length of 1,551 nucleotides. The protein encoded has a length of 517 amino acids. GAME11 is located on chromosome 7 in S. lycopersicum and has a genomic sequence of 1,726 nucleotides. The open reading frame in the mRNA transcript has a length of 1,029 nucleotides, with an encoded protein of 343 amino acids. Homologs identified in S. tuberosum and N. benthamiana are shown in table 1.

(20)

20 GLYCOALKALOID METABOLISM 4 (GAME4)

GAME4 encodes a member of the 88D subfamily of cytochrome P450 proteins and is located

on chromosome 12 in S. lycopersicum (Itkin et al., 2013). The genomic sequence is 4,693 nucleotides long, with a mRNA open reading frame of 1,464 nucleotides. The protein encoded has a length of 488 amino acids. GAME4 functions at the branching point between steroidal saponin and steroidal alkaloid biosynthesis. Silencing of GAME4 in S. lycopersicum showed an reduction of a factor 74 of steroidal alkaloid concentrations in the leafs and tubers, while levels of steroidal spanonins in leaf tissue increased. Plant morphology remained unaffected in this study (Itkin et al., 2013). Homologs identified in S. tuberosum and N. benthamiana are shown in table 1.

(21)

21 3.2 E. coli transformation results

GAME4, GAME6, GAME7, and GAME11 were amplified from S. lycopersicum RNA. GAME8a and GAME8b were isolated from S. tuberosum RNA. Extracted GAME cDNA was ligated into the pEAQ expression vector, and E. coli was transformed for plasmid amplification. TRV2 constructs functioning as control vectors for the VIGS experiment were cloned to E. coli for plasmid amplification as well. A colony PCR screening for E. coli colonies containing the pEAQ vector plus GAME inserts is shown in figure 5 and 6, the colony PCR used for screening of colonies transformed with the TRV2 vectors is shown in figure 7.

Colonies showing a band for at the length corresponding with the expected length of the vector containing the gene insert were selected for overnight culture and plasmid isolation. Colonies used for plasmid isolation were: pEAQ+GAME4 colony 1,2 and 3; pEAQ+GAME6 colony 1, 2 and 3; pEAQ+GAME7 colony 1, 2 and 3; pEAQ+GAME8b colony 1, 2, and 3; pEAQ+GAME8b colony 1 and 4; pEAQ+GAME11; colony 2 and 3.

Figure 5 and 6. Agarose gels showing a colony PCR performed on E. coli transformed with the pEAQ constructs using primers amplifying the insert region of the pEAQ vector. All colonies of E.coli transformed with pEAQ+GAME4 show a band on the expected length of the insert, colony 2 of E.coli transformed with pEAQ+GAME8a shows a band on the length of the insert, and all colonies of E.coli transformed with pEAQ+GAME11 show a band on the length of the insert. All colonies of pEAQ+GAME6, GAME7, and GAME8b show a band corresponding the expected gene length.

Figure 7. Agarose gel showing colony performed on E. coli transformed with the various TRV2 control vectors. TRV2 primers amplifying the region of the insert were used for amplification of the TRV2+PVL and TRV2+PDS constructs. The TRV2+eGFP construct was amplified with primers amplifying the region of the eGFP insert, since it is not located in the multiple cloning site of the TRV2 vector.

(22)

22 3.3 Sequence alignment results

Alignments of vector insert DNA sequences obtained from Macrogen with sequences retrieved from Solgenomics.org are discussed in this section. CLC Workbench files and nucleotide sequences constructed with the sequence data obtained from Macrogen are available in the the supplementary data folder.

pEAQ+GAME4

The sequencing results obtained with the forward and reverse primer of the pEAQ insert region of the purified pEAQ+GAME4 plasmid could be completely aligned from position 1 to 1464 of the GAME4 sequence obtained from solgenomics.org, no point mutations are observed in the sequence data.

pEAQ+GAME6

Sequencing results obtained with the forward and reverse pEAQ primer amplifying the insert region of the pEAQ vector could be aligned without error from position 1 till position 1551 of the GAME6 sequence obtained from solgenomics.org. No mutations in nucleotide sequence were observed.

pEAQ+GAME8b

Sequence results obtained with both the forward and reverse primer were aligned with the GAME8b sequence obtained from Solgenomics.org. Point mutations occurred at position 384, 510, 580, 620 and 1218 in reference to the sequence obtained from Solgenomics. Both sequence construct and sequence obtained from solgenomics were translated to amino-acid sequences. An alignment was performed in which two changes in amino-acid sequence were observed. At position 194 alanine was replaced with threonine in the vector construct. At position 207 lysine was replaced with arginine.

pEAQ+GAME11

Sequence data of both forward an reverse primer could be aligned without error. No point mutations had occurred and the sequence corresponds with the sequence obtained from SolGenomics Database.

(23)

23 3.4 A. tumefaciens transformation results

Isolated plasmid pEAQ+GAME4 colony 3, pEAQ+GAME6 colony 2, pEAQ+GAME7 colony 2, pEAQ+GAME8b colony (2), pEAQ+GAME11 colony 3 were selected for A. tumefaciens transformation. Four day old cultures were selected from the plate and a colony PCR was performed with pEAQ primers amplifying the region of the insert. Thermo Scientific DreamTaq was used in the PCR reaction, the gel on which PCR product was loaded is shown in figure 8. No bands were visible on the gel. The colonies were inoculated in 3 ml of LB with kanamycin, rifampicin and gentamycin, regardless the result of the colony PCR due to time considerations.

Since no bands were visible on the gel, another colony PCR was performed on the liquid culture the next day with the Thermo Scientific Phusion HS enzyme and pEAQ primers. The gel loaded with PCR product is shown in figure 9. Bands are visible with a length of about 1700 bp for all of the pEAQ insert. However, this fragment length does not correspond with the expected length of the inserts. Another PCR with gene specific primers and Phusion HS was attempted on the liquid culture, as shown in figure 10 no bands are visible. A final colony PCR with gene specific and pEAQ primers on the A. tumefaciens cultures transformed with the pEAQ and TRV2 constructs was performed, the gel is shown in figure 11. This colony gel shows uniform band in the PCR product amplified with pEAQ primers, while no bands are visible in PCR product in which the same template was amplified using gene specific primers.

Figure 9 Gel showing results from a colony PCR performed on overnight liquid cultures. Thermo Scientific Phusion Hot Start enzyme was used in the reaction. Almost all pEAQ vectors plus GAME insert show a band around 1700 base pair length, while TRV constructs show no band, or a uniform band around 500 bp (+eGFP, +VIGS).

Figure 8 Gel showing results from colony PCR performed on plated A. tumefaciens transformed with pEAQ and TRV2 constructs. Thermo Scientific DreamTaq was used in the reaction. No bands are visible in this gel.

(24)

24

Electro competent A. tumefaciens of the EHA105 strain were transformed with pEAQ+GAME4, pEAQ+GAME6, pEAQ+GAME8b, pEAQ+GAME11 and TRV2+PDS following the same protocol as transformation of agrobacterium of the GV3101 strain. An electroporation was performed with 1.8 kV, at 400 Ohm. While some of the cuvettes arced, transformation was successful with pEAQ+GAME11, pEAQ+GAME8b and pEAQ+GAME4 plasmids. The

Agrobacteria transformed with pEAQ+GAME6 and TRV2+PDS plasmids did not exhibit colony

growth. An agarose gel showing the colony PCR performed with Mango mix and pEAQ primers is shown in figure 12.

Figure 10 Gel showing colony PCR results on the liquid culture of A. tumefaciens with gene specific primers. An empty pEAQ vector was used with pEAQ primers for control, and an empty TRV2 vector was used with TRV2 primers for control. No bands are present in the PCR product with gene specific primers for the pEAQ constructs. In the TRV2 constructs a band around 500 bp is present in almost all PCR products. However, a band at 500 bp is also present in the empty vector. TRV2+GAME4 shows a band around 300 bp, which is the expected length of the VIGS fragment.

Figure 11 Gel showing results from a colony PCR performed on two colonies of plated of A. tumefaciens per construct with both gene specific (a) and pEAQ (b) primers. Numbers above the wells indicate the GAME insert. Again no bands are visible in the PCR product with gene specific primers, while the pEAQ primer all show the same bands around the 1600-1700 and 200 base pair length.

Figure 12 Electrophorese gel results for A. tumefaciens EHA105 transformed with pEAQ+GAME11(1), pEAQ+GAME11(2), pEAQ+GAME4 and pEAQ+GAME8b plasmids. All colonies selected for colony PCR show an insert length corresponding with the expected length of each gene insert. No uniform band at 1700bp is observed such as in the colony PCR results on GV3101. Bioline MangoMixtm_{was used in}

(25)

25

4. Discussion

Amplification of the GAME genes and E. coli transformation with the various vector constructs was successful, however transformation of A. tumefaciens GV3101 could not be confirmed. PCR product of all transformed Agrobacteria colonies show a band at the same fragment length, which does not correspond with the different fragments obtained from the colony PCRs performed on

E. coli. This meant that agro-infiltration of N. benthamiana with the various pEAQ constructs

could not be carried out. It is hypothesized that the failure in Agrobacterium GV3101 transformation may be due to several factors.

Electroporation was performed with an electro pulse of 2.5kV, 2.2kV, and 1.8kV. The Gene Pulser®/MicroPulser™ Electroporation Cuvettes with a 0.1 cm gap exhibited arcing in all of these voltages. This may be due to the fact that the protocol used in the electroporation was written for electroporation with cuvettes containing a 0.2 cm electrode gap. In order to maintain the same field strength with the 0.1 cm gap cuvettes, the voltage should have been lowered by a factor 2. A voltage of 1.25kV would have been sufficient in this experiment, and could have possibly avoided arcing due to a lower field strength. High salt levels in the purified vector DNA mixture may have caused arcing, despite Milli-Q water was used in each purification step. When electroporation with A. tumefaciens EHA105 strain was performed using a 1.8kV pulse, arcing was observed only in the reused cuvettes that were suspected to have exhibited arcing before. Transformed A. tumefaciens EHA105 colonies furthermore showed positive results on the agarose gel, implicating that transformation was successful despite of the arcing.

However, failed electroporation does not explain colony growth on kanamycin and rifampicin added culture media. This may be due to contamination of the medium, despite all bacterial inoculation steps were performed under the laminar flow hood. The band visible at 1700 bp after colony PCR with pEAQ primers is also unaccounted for. E.coli were transformed with the purified pEAQ-GAME plasmids a second time to study plasmid separation though colony PCR, however no separation was observed. The plasmids were purified from these E. coli cells, and A. tumefaciens EHA105 transformed with the purified plasmid. A colony PCR on EHA105 yielded no similar uniform band of 1700 bp, but instead showed bands corresponding with the expected fragment length for each gene insert. The cells were transformed with a voltage of 1.8kV. The experiment could continue with this strain, however, since time is of limited availability in this study, the experiment was discontinued.

(26)

26

Knowledge on the metabolic pathway involved in steroidal saponin biosynthesis in plants capable of diosgenin production may open new possibilities for to higher diosgenin production observed in wild type Dioscorea species. A change in metabolic flux toward steroidal saponin through silencing of an enzymatic step in steroidal alkaloid formation in tomato was demonstrated in the literature. Homologs of GAME7, GAME6, GAME11, and GAME4 genes researched in this study were identified in N. benthamiana (table 1), an experiment altering metabolic flux toward steroidal saponins and incorporating part of the diosgenin biosynthetic pathway in N. benthamiana may provide new strategies for increased diosgenin production. However, genomic resources in public databases devoted to these plants capable of diosgenin production have yet to be developed. More research into the genome and genetics of important diosgenin rich crops, such as Dioscorea mexicana and Trigonella foenum-graecum, is needed for the study of these diosgenin biosynthetic pathways.

Engineering of the plant steroidal pathway may have applications for the production of other plant derived medicine as well. Plants produce thousands of different phytochemicals, each with different biological activities, making plants a valuable resource for the production of pharmaceutical compounds. Studies into the biological activities of the various secondary plant metabolites in combination with studies aiming to elucidate the metabolic pathways involved in the biosynthesis of these compounds contributes to the development of a plant based green production system for pharmaceuticals.

(27)

27

5. References

Baulcombe, D. C. (1999). Fast forward genetics based on virus-induced gene silencing. Current opinion in plant biology, 2(2), 109-113.

Bitinaite, J., Rubino, M., Varma, K. H., Schildkraut, I., Vaisvila, R., & Vaiskunaite, R. (2007). USER™ friendly DNA engineering and cloning method by uracil excision. Nucleic acids research, 35(6), 1992-2002.

Chaudhary, S., Chikara, S. K., Sharma, M. C., Chaudhary, A., Syed, B. A., Chaudhary, P. S., et al. (2015). Elicitation of diosgenin production in trigonella foenum-graecum (fenugreek) seedlings by methyl jasmonate. International Journal of Molecular Sciences, 16(12), 29889-29899. Chauhan, R., Keshavkant, S., & Quraishi, A. (2018). Enhanced production of diosgenin through elicitation in micro-tubers of Chlorophytum borivilianum Sant et Fernand. Industrial Crops and Products, 113, 234-239.

Chandra, S., & Chandra, R. (2011). Engineering secondary metabolite production in hairy roots. Phytochemistry reviews, 10(3), 371.

Cheng, P., Zhao, H., Zhao, B., & Ni, J. (2009). Pilot treatment of wastewater from Dioscorea zingiberensis CH Wright production by anaerobic digestion combined with a biological aerated filter. Bioresource technology, 100(12), 2918-2925.

Ciura, J., Szeliga, M., Grzesik, M., & Tyrka, M. (2017). Next-generation sequencing of representational difference analysis products for identification of genes involved in diosgenin biosynthesis in fenugreek (Trigonella foenum-graecum). Planta, 245(5), 977-991.

Fernandez-Pozo, N., Rosli, H. G., Martin, G. B., & Mueller, L. A. (2015). The SGN VIGS tool: user-friendly software to design virus-induced gene silencing (VIGS) constructs for functional genomics. Molecular plant, 8(3), 486-488.

Gelvin, S. B. (2003). Agrobacterium-mediated plant transformation: the biology behind the “gene-jockeying” tool. Microbiology and molecular biology reviews, 67(1), 16-37.

Gould, B., & Kramer, E. M. (2007). Virus-induced gene silencing as a tool for functional analyses in the emerging model plant Aquilegia (columbine, Ranunculaceae). Plant Methods, 3(1), 6.

(28)

28

Hirsch, C. D., Hamilton, J. P., Childs, K. L., Cepela, J., Crisovan, E., Vaillancourt, B., ... & Buell, C. R. (2014). Spud DB: A resource for mining sequences, genotypes, and phenotypes to accelerate potato breeding. The Plant Genome, 7(1).

Hoekema, A., Hirsch, P. R., Hooykaas, P. J., & Schilperoort, R. A. (1983). A binary plant vector strategy based on separation of vir-and T-region of the Agrobacterium tumefaciens Ti-plasmid. Nature, 303(5913), 179.

Hua, W., Kong, W., Cao, X., Chen, C., Liu, Q., Li, X., & Wang, Z. (2017). Transcriptome analysis of Dioscorea zingiberensis identifies genes involved in diosgenin biosynthesis. Genes & Genomics, 39(5), 509-520.

Itkin, M., Heinig, U., Tzfadia, O., Bhide, A. J., Shinde, B., Cardenas, P. D., et al. (2013). Biosynthesis of antinutritional alkaloids in solanaceous crops is mediated by clustered genes. Science, 341(6142), 175-179.

Jesus, M., Martins, A. P., Gallardo, E., & Silvestre, S. (2016). Diosgenin: Recent highlights on pharmacology and analytical methodology. Journal of analytical methods in chemistry, 2016.

Jiang, W., Fu, X., Pan, Q., Tang, Y., Shen, Q., Lv, Z., et al. (2016). Overexpression of AaWRKY1 leads to an enhanced content of artemisinin in artemisia annua. Biomed Research International, , 7314971.

Kumar, V., Desai, D., & Shriram, V. (2014). Hairy root induction in Helicteres isora L. and production of diosgenin in hairy roots. Natural products and bioprospecting, 4(2), 107-112.

Li, J., Liang, Q., Li, C., Liu, M., & Zhang, Y. (2018). Comparative Transcriptome Analysis Identifies Putative Genes Involved in Dioscin Biosynthesis in Dioscorea zingiberensis. Molecules, 23(2), 454.

Liu, Y. , Schiff, M. , Marathe, R. and Dinesh‐Kumar, S. P. (2002), Tobacco Rar1, EDS1 and NPR1/NIM1 like genes are required for N‐mediated resistance to tobacco mosaic virus. The Plant Journal, 30: 415-429.

Mehrafarin, A., Qaderi, A., Rezazadeh, S. H., Naghdi Badi, H., Noormohammadi, G. H., & Zand, E. S. K. A. N. D. A. R. (2010). Bioengineering of important secondary metabolites and metabolic pathways in fenugreek (Trigonella foenum-graecum L.). Journal of Medicinal Plants, 3(35), 1-18.

(29)

29

Mou, Y., Zhou, K., Xu, D., Yu, R., Li, J., Yin, C., & Zhou, L. (2015). Enhancement of Diosgenin Production in Plantlet and Cell Cultures of Dioscorea zingiberensis by Palmarumycin C 13 from the Endophytic fungus, Berkleasmium sp. Dzf12. Tropical Journal of Pharmaceutical Research, 14(2), 241-248.

Mueller, L. A., Solow, T. H., Taylor, N., Skwarecki, B., Buels, R., Binns, J., ... & Herbst, E. V. (2005). The SOL Genomics Network. A comparative resource for Solanaceae biology and beyond. Plant physiology, 138(3), 1310-1317.

Patel, K., Gadewar, M., Tahilyani, V., & Patel, D. K. (2012). A review on pharmacological and analytical aspects of diosgenin: a concise report. Natural Products and Bioprospecting, 2(2), 46–52.

Peyret, H., & Lomonossoff, G. P. (2013). The pEAQ vector series: the easy and quick way to produce recombinant proteins in plants. Plant molecular biology, 83(1-2), 51-58.

Ratcliff, F., Martin‐Hernandez, A. M., & Baulcombe, D. C. (2001). Technical advance: tobacco rattle virus as a vector for analysis of gene function by silencing. The Plant Journal, 25(2), 237-245.

Raju, J., & Rao, C. V. (2012). Diosgenin, a steroid saponin constituent of yams and fenugreek: emerging evidence for applications in medicine. In Bioactive Compounds in Phytomedicine. InTech.

Renneberg, R., Berkling, V., & Loroch, V. (2017). In Renneberg R., Berkling V. and Loroch V.(Eds.), Chapter 4 - white biotechnology: Cells as synthetic factories. Boston: Academic Press.

Salomonsen, B., Mortensen, U. H., & Halkier, B. A. (2014). USER-derived cloning methods and their primer design. In DNA Cloning and Assembly Methods (pp. 59-72). Humana Press, Totowa, NJ.

Sharma, S., Malhotra, N., & Sood, H. (2016). Expression analysis of steroid pathway genes revealed positive correlation with diosgenin biosynthesis in Trillium govanianum. Acta physiologiae plantarum, 38(11), 272.

Sparg, S., Light, M. E., & Van Staden, J. (2004). Biological activities and distribution of plant saponins. Journal of ethnopharmacology, 94(2-3), 219-243.

(30)

30

Vaidya, K., Ghosh, A., Kumar, V., Chaudhary, S., Srivastava, N., Katudia, K., ... & Chikara, S. K. (2013). De novo transcriptome sequencing in Trigonella foenum-graecum L. to identify genes involved in the biosynthesis of diosgenin. The Plant Genome, 6(2).

Wu, C. , Jia, L. and Goggin, F. (2011), The reliability of virus‐induced gene silencing experiments using tobacco rattle virus in tomato is influenced by the size of the vector control. Molecular Plant Pathology, 12: 299-305.

Ye, Y., Wang, R., Jin, L., Shen, J., Li, X., Yang, T., Zhou, M., Yang, Z., Chen, Y. (2014). Molecular cloning and differential expression analysis of a squalene synthase gene from Dioscorea zingiberensis, an important pharmaceutical plant. Molecular biology reports, 41(9), 6097-6104.

INCREASING DIOSGENIN PRODUCTION THROUGH METABOLIC ENGINEERING OF STEROIDAL SAPONIN BIOSYNTHESIS