Running head:

The role of HMGRs in synthesizing ginseng saponins

Correspondence:

Deok-Chun Yang

Dept. OMMP, College of Life Science, Kyung Hee University

Suwon 449-701, Korea

E-mail: dcyang@khu.ac.kr

Tel: +82 (0) 31 201 2100

Fax: +82 (0) 31 202 2687

Title of article:

Functional Analysis of HMGRs in Biosynthesizing Triterpene Saponin in Panax ginseng

Authors’ names:

Yu-Jin Kim, Ok Ran Lee, Ji Yeon Oh, Moon-Gi Jang, Deok-Chun Yang

Addresses:

Department of Oriental Medicinal Materials and Processing, College of Life Science, Kyung Hee

University, Suwon 449-701, Korea

One-sentence Summary

PgHMGR, functional ortholog of AtHMGR1, contributes to the production of triterpene saponin

in Panax ginseng.

Footnotes:

The author responsible for the distribution of materials integral to the findings presented in this

article in accordance with the policy described in the Instructions for Authors

(www.plantphysiol.org) is: Deok-Chun Yang (dcyang@khu.ac.kr).

ABSTRACT

Ginsenosides are glycosylated triterpenes, which are considered to be the most pharmaceutically

active ingredients in ginseng (Panax ginseng), known as an adaptogenic herb. However, the

mechanism underlying the biosynthesis of triterpene saponin in ginseng remains unclear. In this

study, we characterized the role of 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR)

in ginseng saponin synthesis. Firstly we showed that P. ginseng has two HMGRs, PgHMGR1

and PgHMGR2 with high sequence identity by analyzing the full-length cDNAs. The expression

of PgHMGR1 and PgHMGR2 is strong in the root vesiculture, and increase in the older ginseng,

supporting its role in accumulation of saponin in roots. In addition, treatment with mevinolin, a

competitive inhibitor of HMGR, on ginseng adventitous roots caused significant reduction of

total ginsenoside. Moreover, we showed that continuous dark exposure for 2 days to 3-year-old

ginseng increased total ginsenoside contents, and PgHMGR1 showed dark-induced expression,

suggesting that PgHMGR1 may be associated with dark-dependent promotion of secondary

metabolite biosynthesis in the ginseng plant. Furthermore, overexpression of PgHMGR1

enhanced the production of sterol and triterpene in Arabidopsis and ginseng. The conserved

biochemical role of PgHMGR1 was shown by the genetic complementation of hmgr1-1 in

Arabidopsis. Taken together, our data suggest that ginseng HMGRs play a key role in the

triterpene pathway for ginseng.

INTRODUCTION

Ginseng (Panax ginseng Meyer), a member of the Araliaceae family, is a perennial

herbaceous plant and has been cultivated for its highly valued roots used for medicinal purposes.

The root of ginseng has been used as a drug by the people in the Eastern Asia for thousands of

years because it contains polyacetylenes, polysaccharides, peptidoglycans, phenolic compounds,

and saponin (Park, 1996; Kitagawa et al., 1987; Radad et al., 2006). The triterpene saponin in

ginseng, referred to as ginsenoside, has especially been noted as the active compound

contributing to the efficacy of ginseng. Triterpene saponins are secondary metabolites of

isoprenoid compounds and are common in a large number of plant species, particularly in dicot

plants. They exhibit great structural diversity and notable biological activity (Hostettmann and

Marston, 1995; Augustin et al., 2011). Ginsenosides are found exclusively in the plant genus

Panax, with a content of about 6 to 10%. More than 150 naturally occurring ginsenosides have

been isolated from Panax species (Shi et al., 2010). To date, more than 40 ginsenosides have

been isolated and identified from white and red ginseng, showing different biological activities

based on their structural differences (Gillis, 1997; Fuzzati, 2004; Lu et al., 2009; Xie et al., 2005;

Tung et al., 2009).

The main ginsenosides are glycosides that contain an aglycone with a dammarane skeleton

(Fig. 1A). They include protopanaxadiol-type (PPD) saponins (where sugar moieties are attached

to the β-OH at C-3 and/or C-20) such as ginsenosides Rb1, Rb2, Rc, and Rd and

protopanaxatriol-type (PPT) saponins (where sugar moieties are attached to the α-OH at C-6

and/or the β-OH at C-20) such as ginsenosides Re, Rg1, Rg2, and Rf, together constituting more

than 80% of the total ginsenosides (Kim, 1987). The oleanane group has a pentacyclic structure,

and only one ginsenoside, Ro, was identified, which is found in minor amounts in P. ginseng.

These ginsenoside compounds contribute to the various pharmacological effects of ginseng, such

as anti-aging (Cheng et al., 2005), anti-diabetes (Attele et al., 2002), anti-inflammatory (Wu et al.,

1992), and anti-cancer activities, such as the inhibition of tumor-induced angiogenesis (Liu et al.,

2000; Nakajima et al., 1998; Yue et al., 2007), anti-tumor activity, and the prevention of tumor

invasion and metastasis (Mochizuki et al., 1995; Sato et al., 1994).

Ginsenosides are synthesized from the 30-carbon intermediate 2,3-oxidosqualene, a

common precursor of sterols and which undergoes further cyclization, hydroxylation, and

glycosylation (Fig. 1B). Triterpene saponins, including ginsenosides, are derived from a

universal precursor, isopentenyl diphosphate (IPP), which can be synthesized via the cytoplasmic

mevalonate (MVA) pathway, conserved in some prokaryotes and in all eukaryotes. The MVA

pathway is catalyzed by the key regulatory enzyme 3-hydroxy-3-methylglutaryl coenzyme A

reductase (HMGR, EC1.1.1.34) (Stermer et al., 1994). HMGR is known as a rate-limiting

enzyme in the isoprenoid pathway in plants (Bach and Lichtenthaler, 1982) and in mammals

(Goldstein and Brown, 1990). Since HMGR is a rate-controlling enzyme for cholesterol

biosynthesis, its role has been studied in human health, such as how the inhibition of HMGR is a

major intervention strategy for the treatment of cardiovascular disease and blood pressure

reduction (Liao and Laufs, 2005). In contrast to the single HMGR gene in animals, plant HMGR

is encoded by a multigene family, where the different isoforms exhibit different tissue and

developmental patterns of expression. Two HMG1 and HMG2 proteins of Arabidopsis have the

same structural organization and intracellular localization, but their expression profiles are

different. The broad expression of HMG1 suggests that it may encode a housekeeping form of

HMGR; correspondingly, the loss of HMG1 caused senescence and sterility as well as a dwarf

phenotype, which are likely related to the reduced sterol content (Suzuki et al., 2004). In contrast,

the hmg2 mutant did not display a distinct phenotype. hmg1 hmg2 double mutants are not viable

because of the requirement of two genes for gametophyte development (Suzuki et al., 2009).

Both HMG1 and HMG2 genes were also shown to play a major role in the biosynthesis of

triterpene metabolites (Ohyama et al., 2007).

So far, even though much advance revealing the pharmacological effects of ginsenosides

and the mass production of ginsenosides by genetic engineering and biotechnology, little is

known about the regulatory function of ginsenoside biosynthesis at the molecular level in

ginseng. Because of the lack of genome sequence information, some attempts have been tried to

clone complete cDNA clones encoding several enzymes from ginseng involved in the post-

squalene step (indicated according to the accession number in Fig. 1B) (Kushiro et al., 1998; Lee

et al., 2004; Kim et al., 2010; Han et al., 2006; 2010; 2011; 2012). In this present study, the full-

length cDNA sequences, together with promoter sequences of two HMGR genes, were isolated

and characterized from P. ginseng for the first time. The phylogenetic analysis, the expression

profiles of two HMGR genes, and functional analysis by heterologous overexpression of HMGR1

in Arabidopsis and ginseng were also investigated. Our finding suggests that HMGR is essential

for the formation of triterpene ginsenoside.

RESULTS

Isolation and Sequence Analysis of PgHMGR1 and PgHMGR2

In order to identify the first committed enzyme of ginseng in the mevalonic acid (MVA)

pathway for the biosynthesis of isoprenoids, two expressed sequence tag (EST) clones coding 3-

hydroxy-3-methylglutaryl coenzyme A reductases, (HMGR from Panax ginseng, PgHMGR, EC

1.1.1.34) were selected from previously constructed EST libraries of 14-year-old ginseng and

hairy root, respectively (Kim et al., 2006b). By RACE PCR, full-length complementary DNA

(cDNA) sequences of PgHMGR1 and PgHMGR2 were obtained. PgHMGR1 has a length of

1,722 bp encoding 573 amino acids, whereas PgHMGR2 has 1,785 bp encoding 594 amino acids

(Supplementary Fig. S1A, B). Both genes contain four exons and three introns (Fig. 2A), a

typical feature of the HMGR genes from other plant species. PgHMGR2 is 63 bp longer than

PgHMGR1 in the first exon region, although the other three exons are the same length. Moreover,

a full genomic DNA sequence of each PgHMGR with the promoter sequence was obtained by

genomic DNA walking. The promoter region of PgHMGR1 contains G-box (CACGTG),

although PgHMGR2 does not (Supplementary Fig. S1 and S2).

HMGR proteins are comprised of three domains (Campos and Boronat, 1995); the membrane-

anchor domain, a flexible linker domain, and the catalytic domain. PgHMGR1 and PgHMGR2

also contain conserved substrate and cofactor binding sites that are proposed to be characteristic

of the HMGR protein (Fig. 2C and Supplementary Fig. S7): two putative HMG-CoA binding

sites (EMPI/VGYVQ, TTEGCLVA) and two NADP(H)-binding sites (DAMGMNM,

GTVGGGT) in the catalytic domain. Transmembrane prediction and the hydrophobicity profile

suggest that both the PgHMGR1 and PgHMGR2 proteins have two membrane-spanning domains,

indicated as H1 and H2 in Fig. 2C. Both proteins contained a motif rich in arginines (RRR)

(Supplementary Fig. S1 and S2), which is predicted to be specific for endoplasmic reticulum (ER)

retention (Schutze et al., 1994).

The three-dimensional (3D) structures of PgHMGR1 and PgHMGR2 were predicted by the

Swiss-Model using sequence homology-based structural modeling. The molecular modeling

results (Fig. 2B) showed an intricate spatial architecture, which is very similar to the human

HMGR structure (Istvan et al., 2000). The catalytic domain consists of three domains: the small

and helical N-terminal N-domain; the large, central L-domain harboring two HMG-CoA binding

sites; and the small helical S-domain which is inserted into the L-domain. The PgHMGR1 and

PgHMGR2 catalytic domains possess three catalytic active residues EGC, DKK, and GQD in the

three conserved motif sequences predicted by Multiple EM for Motif Elicitation (MEME) (Fig.

2B, 2C and Supplemental Fig. S8). These are similar to Cantharanthus roseus HMGR, which

Abdin et al. (2012) recently showed as responsible for the active site that can bind to HMG-COA

and NADPH2.

PgHMGR1 and PgHMGR2 Are Mainly Expressed in Roots

Ginseng is a perennial medicinal plant; an age-dependent increase of ginsenosides was

reported in ginseng roots (Shi et al., 2007). To investigate the expression patterns of PgHMGR1

and PgHMGR2, quantitative reverse transcription (qRT)-PCR was performed on RNA extracted

from ginseng seed, an early seedling and leaf, petiole and root of a 2-week-old seedling, as well

as on RNA extracted from flower, leaf, stem, main root, and lateral root of 3-year-old and 6-year-

old plants (Fig. 3A). A distinct anatomical feature of ginseng is the long petiole structure having

a relatively shortened stem in a 2-week-old seedling (Fig. 3A). The number of leaf petioles

corresponds to the number of years cultivated, and 2-year-old ginseng possesses five leaves.

PgHMGR1 was most highly expressed in the petiole in the 2-week-old seedling (Fig. 3B). In the

3-year-old ginseng, when most of the known ginseng organs are formed, and the 6-year-old

ginseng, transcripts of PgHMGR1 were highly observed in the main roots and lateral roots (Fig.

3C, D). Interestingly, its expression was similarly expressed to high level in 3-year and 6-year

old ginseng (Fig. 3C, D). PgHMGR2 was also comparatively highly expressed in root, but the

level of expression of PgHMGR2 in the root was at least 4-fold lower than that of PgHMGR1

(Fig. 3E, F, G). These results suggest that PgHMGR1 and PgHMGR2 are closely linked with the

age-dependent production of ginsenosides in roots.

When a promoter sequence of PgHMGR1 (-1317~1 bp)-fused GUS was transformed into a

ginseng adventitious root, it was expressed highly in the whole root, with more restriction in the

vasculature and root tip (Fig. 3J). When PgHMGR1::GUS and PgHMGR2::GUS were expressed

in heterologous Arabidopsis, expressions were also observed in the whole root vasculature (Fig.

3Hb, Ib) and in an 8-day-old main root tip (Fig. 3Hc, Ib), similar to the ginseng plant, which

suggests that Arabidopsis is an ideal plant for functional characterization of PgHMGR. Precise

expression of PgHMGR1::GUS and PgHMGR2::GUS in all tissue of a germinating embryo (Fig.

3Ha, Ia), cotyledon, true leaf, and root were also observed in Arabidopsis (Fig. 3Hd, Id).

PgHMGR1 and PgHMGR2 GUS expression was high in the filament, flower sepal and pistil

style (Fig. 3He, Ie) and in the junction between the silique and pedicel of a 50-day-old plant (Fig.

3Hf, If).

HMGR Activity Positively Correlates with Ginsenoside Production

To understand whether HMGR activity is involved in ginsenoside biosynthesis, inhibition of

HMGR activity by mevinolin was conducted. Mevinolin (6α-methylcompactin), also referred to

as lovastatin, competitively inhibits the binding of substrate HMG-CoA to the active site of the

enzyme HMGR and consequently blocks the synthesis of cytosolic IPP (Bach and Lichtenthaler,

1982) and phytosterol biosynthesis (Bach and Lichtenthaler, 1987; Bach et al., 1990). Mevinolin

treatment for 1 day in 4-week-old adventitious roots significantly decreased the total ginsenoside

content to about 34% lower than the control (Fig. 4A), with a decrease in the expression of

PgHMGR1 and PgHMGR2 (Fig. 4B). On the other hand, methyl jasmonate (MJ), known as an

elicitor for triterpene biosynthesis, increased the ginsenoside contents (Fig. 4D) and an increase

in PgHMGR1 expression were observed in ginseng adventitious root and pHMGR1::GUS (Fig.

4E and 4F), confirming the essential role of HMGRs in ginsenoside biosynthesis (Supplementary

Fig. S3).

Expression of PgHMGR1 Is Light-Inhibited and Possibly Associated with Ginseng

Shadowing Growth

The evidence for a dark-dependent hypocotyl expression pattern of PgHMGR1::GUS (Fig. 5G)

led us to hypothesize that ginsenoside synthesis may be inhibited by light. As shown in Fig. 5,

placing the ginseng plants under completely dark conditions for 48 h caused a dramatic increase

in the ginsenoside contents compared to the control grown under a 16 h light and 8 h dark

condition. Dark conditions for 72 h also resulted in an increase in ginsenosides. The total

ginsenoside contents were decreased in both leaves and roots, in spite of different patterns

demonstrated by the individual ginsenosides (Supplementary Fig. S4). Transcriptions of

PgHMGR1 and PgHMGR2 were significantly decreased in the ginseng leaves and roots exposed

to darkness for three days (Fig. 5G, E). However, the HMGR activity was increased in both

leaves and roots (Fig. 5C).

PgHMGR1::GUS promoter activity was differently modulated by light exposure in the leaf

and hypocotyl. In a young seedling that had not yet undergone true leaf emergence, dark-induced

hypocotyl expression of PgHMGR1::GUS was decreased by light exposure within 2 h (Fig. 5G).

However, GUS expression in a 7-day-old seedling was decreased by dark treatment in the true

leaves (Fig. 5H).

PgHMGR1 Localizes to the ER and Peroxisome in Arabidopsis

Fractionation analysis detected plant HMGR in three subcellular sites – the plastids,

mitochondria, and ER (Bach et al., 1999). Despite enzyme activity detection in several organelles,

immunofluorescence confocal microscopy and immunogold electron microscopy analyses

showed that AtHMGR1 is localized in the ER and unidentified spherical vesicles and is not co-

localized with peroxisomal catalase (Leivar et al., 2005). Cyan fluorescent protein (CFP) or

monomeric red fluorescent protein (mRFP) was tagged to the C-terminal ends of PgHMGR1 and

PgHMGR2 to identify subcellular localization patterns by confocal laser scanning microscopy.

Both PgHMGR1 and PgHMGR2 were localized in ER-like vesicles (Fig. 6A, B, C). But the

spherical vesicles and other unidentified vesicles of PgHMGR1 were co-merged with

peroxisome and plastids when PgHMGR1-mRFP was crossed with endosomal markers (Fig. 6E).

Only the catalytic domain of PgHMGR1 (H1CD) was localized to the cytosol and nucleus (Fig.

6D), as in the case of Arabidopsis HMGR1 (Leivar et al., 2005).

Overexpression of PgHMGR1 Enhances Production of Triterpenes in Arabidopsis and

Ginseng

To investigate whether increased HMGR activity contributed to the metabolite profiles, sterol

contents in Arabidopsis HMGR1ox lines were analyzed using gas chromatography-mass

spectrometry (GC-MS) analysis. Higher plants synthesize a mixture of phytosterols, including

campesterol, stigmasterol, and β-sitosterol (Hartmann and Benveniste, 1987); thus, three major

phytosterols were analyzed. Pentacyclic β-amyrin, one of the most common triterpenes in plants,

and α-amyrin are present in Arabidopsis; therefore, these two triterpenes were also analyzed

together with squalene, a common precursor of triterpene and sterol. Total sterols were extracted

from the rosette leaves and inflorescence of 5-week-old plants. Compared to wild type,

HMGR1ox No. 15-8 rosette leaves were 2-times higher in β-sitosterol, 1.8-times higher in

campesterol and cycloartenol, 2.5-times higher in β-amyrin, and 2-times higher in α-amyrin (Fig.

7A). The inflorescence was 1.6-times higher in campesterol, β-sitosterol, and β-amyrin, 2.6-

times higher in cycloarternol, and 3.7-times higher in α-amyrin (Fig. 7B) than the WT. The

squalene and stigmasterol contents were not significantly changed compared with the control.

Volatile (-)-E-b-caryophyllene, which is a sesquiterpene, was also increased in the HMGR1ox

line analyzed by SPME GC-MS (data not shown).

To test whether PgHMGR1 can constitute a functional triterpene pathway in planta, PgHMGR1

was constitutively expressed in Arabidopsis as well as ginseng under the control of the 35S

promoter.

To understand the contribution of HMGR to ginsenoside biosynthesis, adventitious roots were

induced from transgenic ginseng callus after transformation of the PgHMGR1 gene to ginseng.

The ginsenoside contents of different transgenic lines were 1.5-2 times higher than the control,

with no alteration of the ratio of individual major ginsenosides (Fig. 7C).

Conserved Role of PgHMGR1

To understand the evolutionary role of PgHMGRs, we performed amino-acid sequence

alignments of HMGRs from various plants searched by GenBank BlastX from NCBI database

(Supplementary Fig. S5). Alignment analysis revealed that two PgHMGRs had high similarity

with other plant HMGRs. PgHMGR1 shares 93.7% identity with HMGR from Eleutherococcus

senticosus (EsHMGR, AFM77981), which belongs to the same Araliaceae family. PgHMGR2

shared 88.4% identity with HMGR from Panax quinquefolius (PqHMGR, ACV65036). Both

PgHMGR1 and PgHMGR2 showed 74.8% and 67.8% amino acid sequence identity to

AtHMGR1 (At1g76490) and AtHMGR2 (At2g17370) (Enjuto et al., 1994), respectively, and

shared a highly conserved domain with other plant homologs containing a membrane domain

(Campos et al., 1995) and a catalytic domain harboring two HMG-CoA binding motifs and two

NADP(H)-binding motifs, which is comparable to other HMGRs reported previously (Shen et al.,

2006), except for the highly divergent N-terminal region and the linker domain (Supplementary

Fig. S5).

PgHMGR was clustered into plant HMGRs and was structurally distinct from fungus and

animal enzymes (Fig. 8A). PgHMGR1 has less than 30% identity with HMGR from animals,

such as Homo sapiens (NP_000850), Drosophila melanogaster (P14773), and Mus musculus

(NP_032281) and with the fungi, Pichia jadinii (O74164). Three conserved motif sequences

were identified in plant as well as animal HMGRs (Fig. 8B, Supplementary Fig. S5).

The conserved role of ginseng HMGRs was revealed by the genetic complementation of

Arabidopsis null mutant hmgr1-1 (Fig. 8C and 8D). Overexpression of the full length of the

cDNA sequence of PgHMGR1 and the catalytic domain of PgHMGR1 under the 35S promoter

complemented the phenotypic defect of hmgr1-1, whereas PgHMGR2 did not, indicating that

PgHMGR1 has a conserved role with AtHMGR1.

DISCUSSION

Ginseng root, one of the most famous well-known medicinal plants, particularly in traditional

oriental medicine, contains more than 4% ginsenoside by dry weight (Shibata, 2001), which

contribute to its many pharmaceutical activities. Ginsenoside, dammarane-type tetracyclic

triterpene, is unique to ginseng (Fig. 1), although other oleanane-type triterpenes are observed in

other plants. Manipulation of ginsenoside biosynthesis is of particular importance to the increase

of ginsenoside in ginseng or other species. Currently, much of the research about ginsenoside has

been conducted for pharmaceutical efficacy, and there is less information about ginsenoside

biosynthesis. In common, all triterpenes and sterols are biosynthesized via the MVA pathway.

The enzyme HMGR catalyzes the formation of MVA, a rate-limiting step in isoprenoid

metabolism. Here, we report the characterization of the HMGR gene, an essential gene for

ginsenoside biosynthesis in ginseng. Although one HMGR sequence of Panax quinquefolius

(ACV65036) was reported in NCBI, having 76% identity to the amino acid sequence of

PgHMGR1, a functional study of HMGR has not yet been studied in ginseng.

The roots of P. ginseng plants are usually used as important components of traditional oriental

medicine, as the ginsenoside content increases with plant age (Shi et al., 2007). Specifically, the

transcript of PgHMGR1 was enhanced in the roots of 3-year-old and 6-year-old ginseng,

corresponding to the identification of EST homologous from the ginseng root EST library. This

suggests that PgHMGR1 has a major role in ginseng roots with regard to providing sterol or

triterpene. Sterol biosynthesis-related genes were supposed to be expressed constitutively in all

plant tissues in order to synthesize sterol, which is necessary for plant development (Shen et al.,

2006, He et al., 2003). A detailed functional understanding of sterol and triterpene in plant

development needs to be clarified further. Corresponding with the expression of PgHMGR1 and

PgHMGR2 in the vasculature tissue of the roots (Fig. 3Hb, 3Ib), ginsenoside biosynthesis-related

genes PgSSs and PgSEs were also characterized as being expressed in vascular bundle tissues

including phloem cells, parenchymal cells near xylem, and resin ducts in the petioles of ginseng

(Han et al., 2010; Kim et al., 2011). It can be postulated that the phloem and resin ducts might

serve as the metabolically active sites for sterol and saponin biosynthesis and play a role in

conducting squalene oil. In fact, ginsenoside is preferably accumulated more in the phloem than

in the xylem (Fukuda et al., 2006) and is observed in parenchymal and phloem cells (Yokota et

al., 2011).

PgHMGR1 is expressed in a range of ginseng tissues in addition to the roots. Our qRT-PCR

results showed that PgHMGR1 was expressed not only in ginseng roots, but also in the leaves

and flowers, followed by weaker expression in the stem, and PgHMGR2 was expressed at a very

low level (Fig. 3). The results indicate that the expression of PgHMGR1 is in agreement with the

ginsenoside biosynthesis in different tissues, where there is lower ginsenoside in the stems

compared to the roots, leaves, and fruit (Shi et al., 2007). PgHMGR1::GUS and

PgHMGR2::GUS were detected in ginseng seedlings at early post-germination stage, and GUS

expression was also triggered during the germination stage in Arabidopsis. This suggests that

triterpene is required in the seedling stage during germination, corresponding to active β-amyrin

production in pea seedlings (Baisted, 1971). Interestingly, PgHMGR1 was highly expressed in

petiole tissue (Fig. 3B), which is similar to the high expression of other ginsenoside biosynthesis-

related SS (Kim et al., 2011) and SE genes (Han et al., 2010). A long petiole is a distinct anatomic

structure that can be found in the ginseng plant. Triterpene production by petiole-specific gene

expression might provide metabolites required for special petiole development in ginseng,

although there has been no study on the relationship between petioles and triterpene. The

promoter activity of PgHMGR1 and PgHMGR2 was high in the early flowering stage, with no

detection in matured siliques or seed. High HMGR expression in the early developmental stage

has also been shown in other plants: potato HMGR1 is expressed in early flower development

(Korth et al., 1997), and tomato HMGR1 is expressed in young tomato fruit (Jelesko et al., 1999).

The flower phenotype in tobacco is altered by overexpression of AtHMGR1 (Hey et al., 2006),

which suggests that HMGR-derived phytosterols and metabolites play roles in flower

development. Interestingly, overexpression of PgHMGR1 resulted in an increase of volatile oil,

caryophyllene (a sesquiterpene), triterpene, and sterol in the inflorescence of Arabidopsis (Fig.

7B), which supports the idea that higher activity of PgHMGR1 provides more isoprenoid

compounds required for flower development. It is noticeable that the constitutive inhibition of

HMGR by mevinolin ultimately blocked fruit development (Narita and Gruissem, 1989).

All of these results imply that PgHMGR1 plays a key role in ginsenoside biosynthesis. Hence,

it is interesting to determine whether or not PgHMGR1 is positively correlated with the

ginsenoside content in the P. ginseng plant. Unfortunately, ginseng is not practically amenable to

loss- or gain-of-gene-function studies. To gain insight into the effects of the loss of HMGR

activity on ginsenoside biosynthesis, an inhibitor was used to deplete the metabolic flux through

the MVA pathway. Treatment with mevinolin, a highly potent competitive inhibitor of HMGR,

reduced the total ginsenoside content in adventitious roots compared to the control after 24 h due

to reduced gene expression, in contrast to the effect of the elicitor MJ, which increased the

ginsenoside level expression of PgHMGR1 (Fig. 4). This is the first report that the activity of

HMGR is possibly playing an important role in ginsenoside biosynthesis. It was confirmed again

with increased ginsenoside levels in transgenic ginseng adventitious roots overexpressing

PgHMGR1 (Fig. 7C).

Light is known to be one factor that has an effect on plant HMGR. Light-regulated

suppression of HMGR1 promoter activity in Arabidopsis immature leaves and seedlings

(Learned and Connolly, 1997) as well as increased HMGR activity in etiolated pea seedlings

(Brooker and Russsell, 1979) are in direct contrast to the suppression of HMGR activity in dark-

treated potato plants (Korth et al., 2000). In the case of ginseng HMGR, in contrast to the

enhanced gene expression in the etiolated hypocotyl region (Fig. 5G), dark treatment caused a

significant decrease in GUS activity in true leaves of PgHMGR1::GUS in Arabidopsis (Fig. 5H)

and a decrease in the expression level of PgHMGR1 and PgHMGR2 in ginseng (Fig. 5D and E),

consistent with reduced HMGR2 in potato leaves due to dark treatment (Korth et al., 2000).

Differential regulation of PgHMGR1::GUS expression in leaf and hypocotyl tissue can be

explained by the idea that light-mediated control of HMGR1 is an organ-autonomous response

(Learned and Connolly, 1997). Downregulation of HMGR in Arabidopsis due to light was shown

to occur via photoreceptor phytochrome B (PHYB) and transcription factor HY5 (Rodríguez-

Concepción et al., 2004). Analysis of the promoter sequence implied that PgHMGR1 was

regulated by binding PIF3 to the G-box in the promoter of the PgHMGR1 gene, although further

experimental analysis is needed for confirmation. PIFs, basic helix-loop-helix (bHLH)

transcription factors, play a central role in the light-mediated response (Castrillon et al., 2007).

Shin et al. (2007) showed that PIF3 and HY5 regulate anthocyanin biosynthesis by binding to the

promoters of anthocyanin biosynthesis-related genes. PIF3 is also known as a negative

component in the PHYB-mediated inhibition of hypocotyl elongation (Kim et al., 2003). The

PgHMGR1 gene contains a PIF3 binding motif corresponding with dark-dependent hypocotyl

expression, suggesting that ginsenoside biosynthesis might be tightly regulated by light, possibly

through PIF3-mediated suppression of PgHMGR1. When the whole ginseng plant was in dark

conditions for up to 48 h, the ginsenoside content was increased. This could be affected by dark

stress via the light-to-dark switch (Souret et al., 2003). HMGR activity in ginseng was also

increased under dark treatment for 2-3 days, possibly at the post-transcriptional level (Korth et

al., 2000), and it was positively correlated with total ginsenoside content. Our research may

provide new insight into the relationship between light and ginsenosides.

Although HMGR activity was reportedly detected in the ER, plastid, and mitochondria in

plants (Bach et al., 1999), there has been no observation of subcellular localization of plant

HMGR in multiple organelles, except for the localization of AtHMGR1 in the ER and

unidentified spherical vesicles (Leivar et al., 2005). Both PgHMGR1 and PgHMGR2 contain an

RRR motif at their N terminus, which may act as an ER retention signal (Schutze et al., 1994).

When PgHMGR1 and PgHMGR2 were expressed individually in Arabidopsis as C-terminal CFP

fusions, both proteins were also targeted to the ER and spherical vesicles. We further confirmed

that the spherical vesicles and other unidentified vesicles of PgHMGR1 were co-merged with the

peroxisome and plastids (Fig. 6E). Discovery of the subcellular localization of ginseng HMGR

suggests that ginseng HMGR could be targeted to different subcellular organelles through post-

translational modifications. In addition, alternative mRNA processing could contribute to the

different localization. For example, the IPP isomerase long isoform is targeted to the chloroplasts

and mitochondria, and its short isoform is localized to the peroxisome (Sapir-Mir et al., 2008).

The catalytic domain of ginseng HMGR (H1CD) was localized in the nucleus and cytosol,

confirming that the N-terminal domain of PgHMGR is necessary for binding to the membrane,

which confirms the previous report (Leivar et al., 2005). In mammalian research, peroxisomal

localization of HMGR in rat (Keller et al., 1985) was reported in addition to ER localization

(Goldfarb, 1972); thus, dual localization of HMGR in the peroxisome and ER is now commonly

accepted (Kovacs et al., 2007). Localization of HMGR in the peroxisome could have related

roles in lipid metabolism and cholesterol degradation in animals (Keller et al., 1985). Recently,

in plants, acetoacetyl-CoA thiolase (Reumann et al., 2007) and IPP isomerase (Sapir-Mir et al.,

2008) were reported to be localized in peroxisomes (Reumann et al., 2007), and other MVA

pathway genes except HMGR, HMGS, MVK, PMK, DMD, and FPS were predicted to be

partially localized in the peroxisome (Sapir-Mir et al., 2008). The presence of PgHMGR1 in

multiple subcellular localizations supports that MVA biosynthesis can occur in different cell

compartments, in parallel with the localization of ginsenoside Rb1 in the cytosol, plastid, and

peroxisome in the ginseng plant (Yokota et al., 2011). Peroxisome-localized HMGR in animals is

functionally and structurally different from ER-localized HMGR (Aboushadi et al., 2000).

Different compartments could contain separate biosynthetic pathway with separate enzymes for

producing organelle-specific isoprenoids, such as sterol in the ER, pigments in plastids,

ubiquinone in mitochondria, or rubber particles in specialized vacuoles (McGarvey and Croteau,

1995).

Furthermore, to test whether PgHMGR1 can contribute to the triterpene pathway in other

plants, PgHMGR1 was expressed in Arabidopsis. Although HMGR is known as the key

regulatory enzyme in sterol biosynthesis in plants (Babiychuk et al., 2008), overexpression of

HMGR in Arabidopsis did not show any altered morphology or isoprenoid content (Re et al.,

1995). However, overexpression of the HMGR gene in tobacco plants increased the amount of

sterol and intermediates of the sterol pathway in the form of fatty acyl esters in accordance with

the increased activity of HMGR (Chappell et al., 1995; Schaller et al., 1995). In the case of

overexpression of HMGR in tomato, phytosterols were elevated without alteration of the HMGR

activity (Enfissi et al., 2005). The constitutive overexpression of PgHMGR1 driven by the 35S

promoter resulted in the accumulation of not only sterol, but also sesquiterpene, volatile oil, and

triterpenes in Arabidopsis (Fig. 5A, B), with modest increases in the levels of HMGR activity.

This result is comparable with the previous study where overexpression of AtHMGR1 resulted in

co-activation of the HMGR1 gene without altering the accumulation of the isoprenoids end-

products (Re et al., 1995).

Although PgHMGR1 was expressed in different levels among tested organs, the PgHMGR1

gene appears to be expressed constitutively throughout ginseng development, which was similar

to characterized homologs from Salvia miltiorrhiza (Liao et al., 2009) and Eucommia ulmoides

(Jiang et al., 2006). In Arabidopsis, AtHMGR1 can be detected in all tissues (Enjuto et al., 1994),

but AtHMGR2 is expressed exclusively in meristematic and floral tissues (Enjuto et al., 1994).

As reported for the AtHMGR1 gene, PgHMGR1 may play a housekeeping role in ginseng,

whereas PgHMGR2 may play a more specialized role. Indeed, heterologous expression of two

full-length ginseng HMGRs in an Arabidopsis hmgr1-1 mutant background exhibited that only

PgHMGR1 could complement the dwarf and sterile phenotype of hmgr1-1 (in this study; Suzuki

et al., 2004), which suggests that PgHMGR1 is a functional ortholog of AtHMGR1. Phylogenetic

analysis also showed that both PgHMGR1 and PgHMGR2 have three conserved motifs with

conserved amino acid sequences in the homologues (Fig. 9A, B and Supplemental Fig. S6).

Among all of the identified plant HMGR genes, corresponding with the topology of HMGR as a

membrane-binding protein, the N-terminal region exposing to the cytosol and linker domain

differs greatly both in length and amino acid sequence, while the membrane-spanning domain

that serves as a membrane anchor (Campos & Boronat, 1995) and C-terminal catalytic domain

are highly conserved. These two domains contain the binding motifs of two HMG-CoA substrate

and NADP (H) cofactor (Ruiz-albert et al., 2001; Shen et al., 2006). A defect of AtHMGR1 could

be complemented by only the catalytic domain of PgHMGR1, suggesting that the product of

PgHMGR1 is enough to complement without proper targeting to an organelle. However,

overexpression of PgHMGR2 driven by the 35S promoter and not the native promoter could not

complement. This implies that PgHMGR1 and PgHMGR2 could possibly be involved in

different metabolite pathways, ruling out the difference of the promoter or membrane domain

region. Although PgHMGR1 shares more identity with AtHMGR1 than AtHMGR2, it has a

closet relationship with EsHMGR in Acanthopanax, which is a small and woody shrub in the

same family, that is known to have similar herbal properties to those of P. ginseng (Huang et al.,

2011), and that clusters separately from the AtHMGR1 in Arabidopsis (Fig. 9A). PgHMGR2

shares high identity with HMGR2 (CaHMGR, AAB69727) from Camptotheca acuminata

(Maldonado-Mendoza et al., 1997), which also showed lower expression than HMGR1. This

suggests that PgHMGR1 and PgHMGR2 originated from a gene evolution event after the split

from Arabidopsis and before the separation of PgHMGR1 and its homologues. Considering these

findings, it is possible that each of the isozymes of HMGR could have evolved separately

according to the production of specific products rather than duplication.

In conclusion, our data suggest that PgHMGR1 is very likely to contribute to the ginsenoside

synthesis pathway, playing a housekeeping role as a functional ortholog of AtHMGR1. However,

despite the PgHMGR2 expression in ginseng roots, the specific role of PgHMGR2 is unclear. To

understand the contribution of PgHMGR2 to the ginsenoside pathway, we are now generating

PgHMGR2-overexpressing ginseng. Examining this transgenic ginseng should further clarify the

function of PgHMGR2. Further work will be required to clarify the specific role of the HMGR

isozymes in plants.

MATERIAS AND METHODS

Plant Materials and Growth Conditions

The Columbia ecotype of Arabidopsis thaliana was used as a model plant in this study. ER -

YFP (ER-yk, CS16251; Nelson et al., 2007)-, PX-YFP (PX-yk, CS16261; Nelson et al., 2007)-,

MT-YFP (MT-yk, CS16264; Nelson et al., 2007)-, and PT-YFP (PT-yk, CS16267; Nelson et al.,

2007)-expressing lines and hmgr1-1 (SALK_125435) were purchased from the Arabidopsis

stock center (http://www.Arabidopsis.org/). Seeds were surface sterilized and then sown on 1/2

MS medium (Duchefa Biocheme, The Netherlands) containing 1% sucrose, 0.5 g/L MES (2-[N-

morpholino]ethanesulphonic acid), pH 5.7 with KOH, and 0.8% agar. Three-day cold-treated

seeds were germinated under a long-day condition of 16 h light/8 h dark at 23°C. Transformants

were selected on hygromycin-containing plates (50 µg/mL). Ten-day-old seedlings were

transplanted into soil and allowed to grow for up to 5 weeks under the same light/dark conditions.

For RNA extraction, seedlings were grown on plates for 15 days. For metabolite analysis, leaves

and the inflorescence from 5-week-old plants were collected.

Ginseng Materials and Treatment

Mevinolin (Mev) and methyl jasmonate (MJ) were purchased from Sigma (St. Louis, MO,

USA). Stock solutions of Mev (10 mM) and MJ (100 mM) in ethanol were stored at -20°C. For

inhibitor treatment of ginseng, after 4 weeks of pre-cultivation of ginseng adventitious roots, 10

µM Mev or 10 µM MJ was added. After 1, 3, and 7 days, harvested adventitious roots were

frozen and used for RNA extraction, an HMGR activity assay, and ginsenoside analysis.

Three-year-old ginseng plants hydroponically cultured in perlite and peatmoss grown at

23°C±2 under white fluorescent light (60-100 µmol m-2

s-1

) in a controlled greenhouse (kindly

provided by i-farm in Yeo-Ju, Korea) were used for dark treatment. Control plants were grown in

16 h light/8 h dark and sampled in light conditions, while dark treatments lasted for 2 and 3 days

using a black-cover with an aerial part. Leaf and root parts were separately used for RNA

isolation, an HMGR activity assay, and ginsenoside analysis.

Identification of PgHMGR Genes and Sequence Analysis

In order to obtain a full-length coding sequence of the PgHMGR gene, rapid amplification of

cDNA ends (RACE) PCR was performed with a CapfishingTM

full-length cDNA premix kit

(Seegene, Korea) according to the product manual, with HotStarTaq (Qiagen) as DNA

polymerase. The first-strand cDNA synthetic reaction from total RNA was catalyzed by

superscript III RNase H reverse transcriptase (Invitrogen Life Technologies), according to the

product instruction manual, and the cDNA was prepared using a commercial cDNA synthesis kit

according to the manufacture's instruction manual (Clontech, US). Specific primers were

designed according to the 3’-end and 5’-end sequences of partial PgHMGR ESTs, which were

derived from our ginseng EST library. The primer sequences used were as follows: (HMGR1;

HMG_F2: 5’- ATG GAG GCC ATT AAC GAT GGA AAA G-3’, HMG_R2: 5’- ACC AAC CTC

AAT TGA TGG CAT TGT G-3’, HMG_R3: 5’- CCT TTT TGC CGT AGA AAA CCT AAC CA-

3’; HMGR2: HMG2_F2: 5’- TAC TCA CTC GAG TCC AAA CTG GGA GAC -3’, HMG2_R1:

5’-GGC ATG CTA ATT GGG TCC CAC CTC CT-3’, HMG2_R3: 5’- ATT GCC TCA CAA ACA

ACC GAT TTA CC -3’). RACE PCR was performed by the hot start method with the following

conditions: 30 cycles of 94ºC for 40 s, 60 - 66ºC for 40 s, 72ºC for 1 min 20 s, and a final

extension of 72ºC for 5 min. The PCR product was purified and ligated into a pGEM-T vector

(Promega, USA) followed by sequencing. By assembling the sequence of the 3’-RACE and 5’-

RACE products, the full-length cDNA sequence of PgHMGRs was deduced. A 60 ng genomic

DNA of ‘Yunpoong’ and a pair of PCR primers (Start codon and Stop codon) were used for

amplifying the genomic sequence of PgHMGR.

Deduced amino acid sequences were searched for homologous proteins in the databases using

BLASTX network services at the NCBI. ClustalX with default gap penalties was used to perform

multiple alignment of HMGRs isolated in ginseng and previously registered in other species. A

phylogenetic tree was constructed by the neighbor-joining method, and the reliability of each

node was established by bootstrap methods using MEGA4 software. Identification of conserved

motifs of HMGR was accomplished with MEME. A three-dimensional model was prepared by

using PgHMGR as a template on a SWISS-MODEL WORKSPACE in automated mode. The

generated 3-D structure was visualized by the UCSF Chimera package.

Isolation of Promoter Sequences

The Universal Genome Walker™ Kit (Clontech Laboratories, Inc., Palo Alto, CA, USA) was

used to isolate fragments of the PgHMGR1 and PgHMGR2 promoter. Ten 6 bp-recognizing and

blunt end-forming restriction enzymes (DraI, EcoRV, PvuII, StuI, SspI, SmaI, MscI, ScaI,

Eco105I, and HpaI) were used to digest the isolated ‘Yungpoong’ genomic DNA. DNA

fragments containing the adaptors at both ends were used as templates for amplifying the

PgHMGR promoter regions. The first PCR, using AP1 and H1–GSP1 and H2-GSP1, respectively,

(H1-GSP1, 5’-ATA ACA GTC CCC GAG TTT TGA TTC CAG-3’; H2-GSP1, 5’-GCT TGG

GAG GAA GAG AAC AGT CGA TAG C-3’) and the second nested PCR, using AP2 and

H1_GSP2 and H2_GSP2, respectively, (H1_GSP2, 5’-ACG TGA GAC AAA TGA TTG GAC

GAA ATC-3’; H2_GSP2, 5’-ATG GCT CTC GAC GAC TAT CTT CCT CGT T-3’) were

performed using i-MAXII (Intron, Korea). The amplified fragment was cloned in the pGEM-T

vector and sequenced. The promoters were analyzed using PlantPan (Plant Promoter Analysis

Navigator).

Vector Construction and Arabidopsis Transformation

To visualize the subcellular localization patterns of ginseng HMGRs, cDNA sequences of

PgHMGR1 and PgHMGR2 were cloned into a pCAMBIA1390 vector containing the cauliflower

mosaic virus (CaMV) 35S promoter and eCFP. The PgHMGR1 cDNA and its catalytic domain

(PgHMGR1CD; amino acid residues 154 to 573) were amplified using primers with XhoI and

EcoRI sites (PgHMGR1: 5’-TA CTC GAG ATG GAC GTC CGC CGG AGA-3’ and 5’-TG GAA

TTC AGA TCC AAT TTT GGA CAT-3’, PgHMGR1CD: 5’-TA CTC GAG ATG CCC GTA GTG

ATG TCA-3’ and 5’-TG GAA TTC AGA TCC AAT TTT GGA CAT-3’). The PgHMGR2 cDNA

was amplified using primers with SalI and AvrII sites: 5’-CA GTC GAC ATG GAC GTT CGC

CGG CGA CCT-3’ and 5’- GA CCT AGG CTA GGA GGA GAG TTT TGT-3’. Enzyme-digested

PCR products were cloned into the PLA2a gene site of Pro35S:PLA2a-

eCFP/pCAMBIA1390 (provide by Dr. Lee) as eCFP fusion proteins (Pro35S:PgHMGR1-eCFP,

Pro35S:PgHMGR1CD-eCFP or Pro35S:PgHMGR2-eCFP). To express the PgHMGR1 fusion

protein with mRFP at the C terminus, the mRFP was amplified from 326-mRFP using primers

with EcoRI and SpeI sites (5’-TG GAA TTC ATG GCC TCC TCC GAG GAC-3’ and 5’-GG

ACT AGT TTA GGC GCC GGT GGA GTG-3’) and was replaced with eCFP of

Pro35S:PgHMGR1-eCFP.

The 1317 and 477 nucleotides of genomic PgHMGR1 and PgHMGR2 DNA, respectively,

were amplified with HindIII and BamHI sites (ProPgHMGR1; 5’-TG AAG CTT AGC TAA

AGA AAG TTA GGC-3’ and 5’-TA GGA TCC GGA AGA GTA TAT TCC GGC-3’,

ProPgHMGR2; 5’-CG AAG CTT GTA ATG TGT TCC CAC TTC CAT-3’ and 5’-TA GGA TCC

CTT TAT GGT GGG GGA ACT CCG-3’) and cloned to a pCambia 1390 vector containing the

GUS gene. All transgene constructs were confirmed by nucleotide sequencing before plant

transformation. The constructs were transformed into Arabidopsis using Agrobacterium

tumefaciens C58C1 (pMP90) (Bechtold and Pelletier, 1998). The insertion of transgenes into the

transformants was confirmed by confocal microscopy or PCR analysis of the genomic DNA

from the transformants. Homozygous plants with a 3:1 segregation ratio were selected on

antibiotic plates for further analyses. For each construct, 20-50 T1-independent lines were

obtained, and the phenotypic significance of the transgenes was analyzed in the chosen lines.

Transformation of Ginseng

Zygotic embryos of P. ginseng cv. “Yunpoong” seeds (provided by Ginseng Genetic Resource

Bank) were stratified in humidified sand to mature for three months at 5°C. After stratification,

the seeds, in which zygotic embryos were in a mature state (4 mm in length), were immersed in

70% ethanol for 1 min, surface sterilized in 2% NaOCl for 15 min, and rinsed three times with

sterilized distilled water. After carefully dissecting the zygotic embryos, they were placed on MS

basal medium containing 3% sucrose and 0.8% agar. Five-day-old zygotic embryos were excised

transversely, and cotyledon explants (4 mm) were used for transformation. The explants were

dipped in the bacterial solution for 15 min, subsequently blotted with sterile filter paper, and co-

cultivated with A. tumefaciens C58C1 on MS medium containing 3% sucrose and 1 µg/mL 2,4-

dichlorophenoxyacetic acid (2,4-D) for 2 days. Thereafter, the explants were cultured on

selection MS medium with 3% sucrose, 1 µg/mL 2,4-D, 0.5 µg/mL 6-benzylaminopurine (BA),

250 µg/mL cefotaxime, and 50 µg/mL hygromycin. After subculturing six times every two weeks

on selection medium, the surviving cotyledons producing calli were cultured on the same

medium without antibiotics, and adventitious roots were induced from calli on B5 medium with

3% sucrose and 3 mg/L indole-3-butyric acid (IBA). More than 10 transgenic lines were

generated, and adventitious roots were excised from the maternal explants and sub-cultured in

liquid B5 medium with 3% sucrose and 2 mg/L IBA every 5 weeks.

Expression Analysis

Total RNA was extracted from frozen samples with the RNeasy plant mini kit (Qiagen, USA),

including the DNase I digestion step. Next, 2 µg of the total RNA was reverse transcribed with

RevertAid™ H Minus M-MuLV Reverse Transcriptase (Fermentas, USA). RT-PCR was

performed in a 25 µl reaction volume consisting of 0.2 µl of the cDNA product and 5 pmol of

each primer using Super-Taq DNA polymerase (Super Bio, Korea) by a Bio-Rad PCR machine

(3 min at 95ºC, followed by 28 cycles at 94ºC for 20 s, 60ºC for 20 s, and 72ºC for 30, with a

final extension at 72ºC for 5 min). The products were analyzed on 1.2% agarose gels.

Real-time quantitative PCR was performed using 100 ng of cDNA in a 10-µl reaction volume

using SYBR®

Green Sensimix Plus Master Mix (Quantace, Watford, England). The following

thermal cycler conditions recommended by the manufacturer were used: 10 min at 95ºC,

followed by 40 cycles at 95ºC for 10 s, 60ºC for 10 s, and 72ºC for 20 s. The fluorescent product

was detected during the final step of each cycle. Amplification, detection, and data analysis were

carried out on a Rotor-Gene 6000 real-time rotary analyzer (Corbett Life Science, Sydney,

Australia). To determine the relative fold-differences in template abundance for each sample, the

Ct value for each of the gene-specific genes was normalized to the Ct value for ß-actin and

calculated relative to a calibrator using the formula 2ㅡΔCt

or 2ㅡΔΔCt

. Three independent

experiments were performed, and the primer efficiencies were determined according to the

method of Livak and Schmittgen (2001) to validate the ΔΔCt method used in our experiment.

The observed slopes were close to zero, indicating that the efficiencies of the gene and the

internal control ß-actin were equal.

Confocal Microscopy Analysis

The fluorescence from reporter proteins and organelle markers was observed by confocal laser

scanning microscopy (LSM 510 META, Carl Zeiss, Jena, Germany). GFP, CFP, YFP, and mRFP

were detected using 488/505-530, 458/475-525, 514/>530, and 543/560-615 nm

excitation/emission filter sets, respectively. Fluorescence images were digitized with the Zeiss

LSM image browser.

HMGR Activity Assay

Extraction of crude proteins was performed as described (Rodríguez-Concepción et al., 2004),

with minor modifications. Two-week-old seedlings of Arabidopsis, leaves or roots of 3-year-old

ginseng, or adventitious roots of ginseng (~200 mg) were homogenized in liquid nitrogen and

mixed with 1 ml of prechilled extraction buffer containing 100 mM sucrose, 40 mM sodium

phosphate pH 7.5, 30 mM EDTA, 50 mM NaCl, 10 mM DTT, 1 mM AEBSF, 1 µM bestatin, 15

µM E64, 20 µM leupeptin, 15 µM pepstatin, 0.5 mM pheanthroline, 0.5 mM

phenylmethylsulfonyl fluoride, and 0.25% (w/v) Triton X-100. The slurry was centrifuged at 200

g and 4°C for 10 min to remove cell debris, and the supernatant was used for HMGR activity as

described (Dale et al., 1995) with the following modifications, and the protein concentration was

determined with the BCA Protein Assay kit (Intron, Korea). The supernatant was analyzed using

a spectrophotometric assay at 37°C (total volume 200 µl) containing 0.3 mM of HMG-CoA, 0.2

mM of NADPH, and 4 mM of dithiothreitol in 50 mM Tris-HCl, pH 7.0. The decrease in

absorbance at 340 nm was monitored for 20 min after the addition of HMG-CoA. One unit of

HMGR activity is defined as the amount of enzyme which oxidizes 1 µmol of NADPH per

minute at 37°C.

GUS Histochemical Analysis

Four-day-old seedlings were treated with the indicated chemical for 3 h before visualizing the

GUS activity. GUS staining was performed by incubating whole seedlings in the staining buffer

containing 1 mM 5-bromo-4-chloro-3-indoyl-β-D-glucuronic acid cyclohexylammonium salt (X-

Gluc, Duchefa Biocheme, The Netherlands), 0.1 M NaH2PO4, 0.1% Triton-X, and 0.5 mM

potassium ferri- and ferrocyanide at 37°C until a blue color appeared (1-3 h). Stained seedlings

were cleared in 70% ethanol for 2 h and 100% ethanol for 2 h. In the final step of dehydration,

samples were sequentially exposed to 10% (v/v) glycerol/50% (v/v) ethanol and 30% (v/v)

glycerol/30% (v/v) ethanol. Seedlings were photographed under a microscope (ZEISS Axio

Observer D1, Germany).

HPLC Analysis of Ginsenosides Extracted from Ginseng

For the analysis of ginsenosides from different ginseng samples, 0.3 - 1 g of milled powder of

freeze-dried adventitious roots, leaves, and roots were soaked in 80% MeOH at 70°C. After the

liquid was evaporated, the residue was dissolved in H2O, followed by extraction with H2O-

saturated n-butanol. The butanol layer was then evaporated to produce the saponin fraction. Each

sample was dissolved in MeOH (1 g/5 ml) and then filtered through a 0.45 µm filter and used for

HPLC analysis. The HPLC separation was carried out on an Agilent 1260 series HPLC system

(Palo Alto, CA, USA). This experiment employed a C18 (250 x4.6 mm, ID 5 µm) column using

acetonitrile (solvent A) and distilled water (solvent B) mobile phases, with a flow rate of 1.6

ml/min and the following gradient: A/B ratios of 80.5:19.5 for 0-29 min, 70:30 for 29-36 min,

68:32 for 36-45 min, 66:34 for 45-47 min, 64.5:35.5 for 47-49 min, 0:100 for 49-61 min, and

80.5:19.5 for 61-66 min. The sample was detected by UV 203 nm. Quantitative analysis was

performed with a one-point curve method using external standards of authentic ginsenosides.

GC-MS Analysis of Sterols and Triterpenes Extracted from Arabidopsis

Using the method modified from Suzuki et al. (2004), freeze-dried plant materials (rosette

leaves, 200 mg; inflorescence, 25 mg) were powdered and then extracted twice with 1 ml CHCl3-

MeOH (7:3) at room temperature. 5-α-Cholestane (20 µg) was used as an internal standard. The

extract was dried in a rotary evaporator and saponified with 1.5 ml each of MeOH and 20%

KOH aq. for 1 h at 80°C to hydrolyze the sterol esters. After saponification, 1.5 ml each of

MeOH and 4 N HCL were added for 1 h at 80°C, and these reaction mixtures were extracted

three times with 4 ml of hexane. When the phases separated, the sterols partitioned to the hexane

layer, and the combined hexane layer was evaporated to dryness. The residue was

trimethylsilylated with pyridine and BSTFA+1% TMCS (1:1) at 37°C for 90 min and analyzed

by GC-MS.

Capillary gas chromatography-mass spectrometry (GC-MS) analysis was performed using a

mass spectrometer (HP 5973 MSD) connected to a gas chromatograph (6890A, Agilent

Technologies) with a DB-5 (MS) capillary column (30 m×0.25 mm, 0.25 μm film thickness). The

analytical conditions were as follows: EI (70 eV), source temperature 250°C, injection

temperature 250°C, column temperature program: 80°C for 1 min, then increased to 280°C at a

rate of 10°C/min and held at this temperature for 17 min; post temperature of 300°C, carrier gas

He, flow rate 1 ml/min, run time of 38 min; splitless injection. The endogenous sterol levels were

determined as the peak area ratios of molecular ions of the endogenous sterol and internal

standard. Standards including squalene, phytosterols (campesterol, β-sitosterol, stigmasterol),

and triterpene (β-amyrin, α-amyrin) were purchased from Sigma (USA).

For sesquiterpene analysis, the solid-phase microextraction (SPME)-trapped volatiles

(Aharoni et al., 2008) were analyzed by GC-MS as described by Verhoeven et al. (1997). Exactly

0.3 g of inflorescence of 5-week-old Arabidopsis was introduced into a 20 ml vial (Gerstel,

Mu l̈heim, Germany). The fused silica fiber of the SPME device coated with 100 µm of

polydimethylsiloxane was inserted into the vial through an aluminum cap, and volatiles were

trapped by exposing the fiber to the headspace for 30 min. A 65 µm PDMS-DVB

(polydimethylsiloxane – divinylbenzene)-coated fiber was used. The SPME fiber was exposed

for 20 min in the head-space at 45°C, and then fiber was withdrawn into the needle and

transferred to the injector of the GC-MS. An HP-5 column (50 m x 0.32 mm, film thickness 1.05

pro) was used with He (37 kPa) as the carrier gas. The GC oven temperature was programmed as

follows: 80°C for 2 min, increased to 250°C at a rate of 8°C /min, and held at 250°C for 5 min.

Mass spectra in the electron impact mode were generated at 70 eV. The compounds were

identified by comparison of GC retention indices and mass spectra with those of authentic

reference compounds.

LITERATURE CITED

Abdin MZ, Kiran U, Aquil S (2012) Molecular cloning and structural characterization of HMG-CoA

reductase gene from Catharanthus roseus (L.) G. Donn. cv. Albus. Indian J Biotechnol 11:16-22

Aboushadi N, Shackelford JE, Jessani N, Gentile A and Krisans SK (2000) Characterization of

peroxisomal 3-hydroxy-3-methylglutaryl coenzyme A reductase in UT2 cells: sterol biosynthesis,

phosphorylation, degradation, and statin inhibition. Biochem 39:237–247.

Aharoni A, Giri AP, Deurelein S, Griepink F, de Kogel WJ, Verstappen FWA, Verhoeven HA,

Jongsma MA, Schwab W and Bouwmeester HJ (2003) Terpenoid metabolism in wild-type and

transgenic Arabidopsis plants. Plant Cell 15:2866-2884.

Attele AS, Zhou YP, Xie JT and Wu JA (2002) Antidiabetic effects of Panax ginseng berry extract and

the identification of an effective component. Diabetes 51:1851–1858.

Augustin JM, Kuzina V, Andersen SB and Bak S (2011) Molecular activities, biosynthesis and

evolution of triterpenoid saponins. Phytochem 72:435-457.

Babiychuk E, Bouvier-Nave P, Compagnon V, Suzuki M, Muranaka T, Van Montagu M, Kushnir S

and Schaller H (2008) Allelic mutant series reveal distinct functions for Arabidopsis cycloartenol

synthase 1 in cell viability and plastid biogenesis. Proc Natl Acad Sci USA 105:3163-3168.

Bach TJ and Lichtenthaler HK (1982) Mevinolin a highly specific inhibitor of microsomal 3-hydroxyl-

3-methylglutaryl-coenzyme A reductase of radish plants. Z Naturforsch 37:46-50.

Bach TJ and Lichtenthaler HK (1987) Plant growth regulation by mevinolin and other sterol

biosynthesis inhibitors. In G Fuller, WD Nes, eds, Ecology and Metabolism of Plant Lipids. American

Chemical Society Symposium Series, Vol 325. American Chemical Society, Washington, DC, pp 109–

139

Bach TJ, Boronat A, Campos N, Ferrer A and Vollack KU (1999) Mevalonate biosynthesis in plants.

Critic Rev Biochem Mol Biol 34:107-122.

Bach TJ, Weber T and Motel A (1990) Some properties of enzymes involved in the biosynthesis and

metabolism of 3-hydroxy-3-methylglutaryl-CoA in plants. Recent Advances in Phytochem 24:1-82.

Baisted DJ (1971) Sterol and triterpene synthesis in the developing and germinating pea seed. Biochem J

124(2):375–383.

Bechtold N and Pelletier G (1998) In planta Agrobacterium-mediated transformation of adult

Arabidopsis thaliana plants by vacuum infiltration. In Arabidopsis Protocols, J.M. Martinez-Zapater

and J. Salinas, eds (Totowa, NJ: Humana), pp 259-266

Brooker JD, Russsell DW (1979) Regulation of microsomal 3-hydroxy-3-methylglutaryl coenzyme A

reductase from pea seedlings: rapid posttranslational phytochrome-mediated decrease in activity and

in vivo regulation by isoprenoid products. Arch Biochem Biophys 198:323-334

Campos N and Boronat A (1995) Targeting and topology in the membrane of plant 3-hydroxy-3-

methylglutaryl coenzyme a reductase. Plant Cell 7:2163-2174.

Castrillon A, Shen H and Huq E (2007) Phytochrome interacting factors: central players in

phytochrome-mediated light signaling networks. Trends Plant Sci 12:514-521.

Chappell J (1995) Biochemistry and molecular biology of the isoprenoid biosynthetic pathway in plants.

Annu Rev Plant Physiol Plant Mol Biol 46:521–547.

Cheng Y, Shen LH and Zhang JT (2005) Anti-amnestic and anti-aging effects of ginsenoside Rg1 and

Rb1 and its mechanism of action. Acta Pharmacol Sin 26(2):143-149.

Dale S, Arro ́ M, Becerra B, Morrice NG, Boronat A, Hardie DG and Ferrer A (1995) Bacterial

expression of the catalytic domain of 3-hydroxy-3-methylglutaryl-CoA reductase (isoform HMGR1)

from Arabidopsis thaliana, and its inactivation by phosphorylation at Ser577 by Brassica oleracea 3-

hydroxy-3-methylglutaryl-CoA reductase kinase. Eur J Biochem 233: 506–513.

Enfissi EM, Fraser PD, Lois LM, Boronat A, Schuch W and Bramley PM (2005) Metabolic

engineering of the mevalonate and non-mevalonateisopentenyldiphosphate-forming pathways for the

production of health-promoting isoprenoids in tomato. Plant Biotechnol J 3(1):17-27.

Enjuto M, Balcells L, Campos N, Caelles C, Arro ́ M and Boronat A (1994) Arabidopsis thaliana

contains two differentially expressed 3-hydroxy-3-methylglutaryl-CoA reductase genes, which encode

microsomal forms of the enzyme. Proc Natl Acad Sci USA 91:927–931.

Fukuda N, Shan S, Tanaka H and Shoyama Y (2006) New staining methodology: Eastern blotting for

glycosides in the field of Kampo medicines. J Nat Med 60:21-27.

Fuzzati N (2004) Analysis methods of ginsenosides. J Chromatogr B Analyt Technol Biomed Life Sci

812(1-2):119-133.

Gillis CN (1997) Panax ginseng pharmacology: a nitric oxide link? Biochem Pharmacol 54(1):1-8.

Goldfarb S (1972) Submicrosomal localization of hepatic 3-hydroxy-3-methylglutaryl coenzyme a

(HMG-CoA) reductase. FEBS Lett 24(2):153–155.

Goldstein JL and Brown MS (1990) Regulation of the mevalonate pathway. Nature 343:425-430.

Han JY, Hwang HS, Choi SW, Kim HJ and Choi YE (2012) Cytochrome P450 CYP716A53v2

catalyzes the formation of protopanaxatriol from protopanaxadiol during ginsenoside biosynthesis in

Panax ginseng. Plant Cell Physiol 53(9):1535-1545.

Han JY, In JG, Kwon YS and Choi YE (2010) Regulation of ginsenoside and phytosterol biosynthesis

by RNA interferences of squalene epoxidase gene in Panax ginseng. Phytochem 71(1):36-46.

Han JY, Kim HJ, Kwon YS and Choi YE (2011) The cyt P450 enzyme CYP716A47 catalyzes the

formation of protopanaxadiol from dammarenediol-II during ginsenoside biosynthesis in Panax

ginseng. Plant Cell Physiol 52:2062–2073.

Han JY, Kwon YS, Yang DC, Jung YR and Choi YE (2006) Expression and RNA interference-induced

silencing of the dammarenediol synthase gene in Panax ginseng. Plant Cell Physiol 47:1653–1662.

Hartmann MA and Benveniste P (1987) Plant membrane sterols: isolation, identification, and

biosynthesis. Methods Enzymol 148:632-650.

He JX, Fujioka SH, Li TCH, Kang SG, Seto H, Takatsuto S, Yoshida S and Jang JC (2003) Sterols

regulate development and gene expression in Arabidopsis. Plant Physiol 131(3):1258-1269.

Hey SJ, Powers SJ, Beale MH, Hawkins ND, Ward JL and Halford NG (2006) Enhanced seed

phytosterol accumulation through expression of a modified HMG-CoA reductase. Plant Biotechnol J

4(2):219-229.

Hostettmann K and Marston A (1995) Saponins. Cambridge University Press, Cambridge, UK.

Huang L, Zhao H, Huang B, Zheng C, Peng W, Qin L (2011) Acanthopanax senticosus: review of

botany, chemistry and pharmacology. Pharmazie 66(2):83-97

Istvan ES, Deisenhofer J (2000) The structure of the catalytic portion of human HMG-CoA reductase.

BBA-Mol Cell Biol Lipids 1529:9-18

Jelesko J, Jenkins SM, Rodríguez-Concepción M and Gruissem W (1999) Regulation of tomato

HMG1 during cell proliferation and growth. Planta 208:310-318.

Jiang J, Kai G, Cao X, Chen F, He D, Liu Q (2006) Molecular cloning of a HMG-CoA reductase gene

from Eucommia ulmoides Oliver. Biosci Rep 26:171-181

Keller GA, Barton MC, Shapiro DJ and Singer SJ (1985) 3-Hydroxy-3-methylglutaryl-coenzyme A

reductase is present in peroxisomes in normal rat liver cells. Proc Natl Acad Sci USA 82: 770-774.

Kim J, Yi H, Choi G, Shin B, Song PS and Choi G (2003) Functional characterization of phytochrome

interacting factor 3 in phytochrome-mediated light signal transduction. Plant Cell 15:2399–2407.

Kim MK, Lee BS, In JG, Sun H, Yoon JH and Yang DC (2006b) Comparative analysis of expressed

sequence tags (ESTs) of ginseng leaf. Plant Cell Rep 25:599-606.

Kim MW, Ko SR, Choi KJ and Kim SC (1987) Distribution of saponin in various sections of Panax

ginseng root and changes of its contents according to root age. Kor J Ginseng Res 11:10-16.

Kim OT, Bang KH, Jung SJ, Kim YC, Hyun DY, Kim SH and Cha SW (2010) Molecular

characterization of ginseng farnesyldiphosphate synthase gene and its up-regulation by methyl

jasmonate. Biol Planta 54(1):47-53.

Kim, TD, Han JY, Huh GH and Choi YE (2011) Expression and functional characterization of three

squalene synthase genes associated with saponin biosynthesis in Panax ginseng. Plant Cell Physiol

52(1):125-137.

Kitagawa O, Taniyama T, Shibuya H, Nota T, Yoshikawa M (1987) Chemical studies on crude drug

processing. V. on the constituents of ginseng radix rubra (2): Comparison of the constituents of white

ginseng and red ginseng prepared from the same Panax ginseng root. Yakugaku Zasshi 107:495-505

Korth KL, Jaggard DAW and Dixon RA (2000) Developmental and light-regulated post-translational

control of 3-hydroxy-3-methylglutaryl-CoA reductase levels in potato. Plant J 23:507–516.

Korth KL, Stermer BA, Bhattacharyya MK and Dixon RA (1997) HMG-CoA reductase gene families

that differentially accumulate transcripts in potato tubers are developmentally expressed in floral

tissues. Plant Mol Biol 33(3):545-551.

Kovacs WJ, Tape KN, Shackelford JE, Duan X, Kasumov T, Kelleher JK, Brunengraber H and

Krisans SK (2007) Localization of the pre-squalene segment of the isoprenoid biosynthetic pathway

in mammalian peroxisomes. Histochem Cell Biol 127(3):273-290.

Kushiro T, Shibuya Y and Ebizuka Y (1998) β-amyrin synthase: cloning of oxidosqualenecyclase that

catalyze the formation of the most popular triterpene among higher plants. Eur J Biochem 256:238-

244.

Learned RM and Connolly EL (1997) Light modulates the spatial patterns of 3-hydroxy-3-

methylglutaryl coenzyme A reductase gene expression in Arabidopsis thaliana. Plant J 11(3):499-511.

Lee MH, Jeong JH, Seo JW, Shim CG, Kim YS, In JG, Yang DC, Yi JS and Choi YE (2004)

Enhanced triterpene and phytosterol biosynthesis in Panax ginseng overexpressing squalene synthase

gene. Plant Cell Physiol 45(8):976-984.

Leivar P, Gonzalez VM, Castel S, Trelease RN, Lopez-lglesias C, Arró M, Boronat A, Campos N,

Ferrer A and Fernandez-Busquets X (2005) Subcellular localization of Arabidopsis 3-hydroxy-3-

methylgluraryl-coenzyme A reductase. Plant Physiol 137:57-69.

Liao JK and Laufs U (2005) Pleiotropic effects of statins. Annu Rev Pharmacol Toxicol 45:89-118.

Liao P, Zhou W, Zhang L, Wang J, Yan X, Zhang Y, Zhang R, Li L, Zhou G, Kai G (2009)

Molecular cloning, characterization and expression analysis of a new gene encoding 3-hydroxy-3-

methylglutaryl coenzyme A reductase from Salvia miltiorrhiza. Acta Physiol Plant 31:565-572

Liu WK, Xu SX and Che CT (2000) Anti-proliferative effect of ginseng saponins on human prostate

cancer cell line. Life Sci 67:1297–1306.

Livak KJ and Schmittgen TD (2001) Analysis of relative gene expression data using real-time

quantitative PCR and the 2(-delta delta C(T)) method. Methods 25:402-408.

Lu JM, Yao Q and Chen C (2009) Ginseng compounds: an update on their molecular mechanisms and

medical applications. Curr Vasc Pharmacol 7(3):293-302.

Maldonado-Mendoza IE, Vincent RM, Nessler CL (1997) Molecular characterization of three

differentially expressed members of the Camptotheca acuminate 3-hydroxy-3-methylglutaryl CoA

reductase (HMGR) gene family. Plant Mol Biol 34:781-790

McGarvey DJ and Croteau R (1995) Terpenoid metabolism. Plant Cell 7:1015-1026.

Mochizuki M, Yoo YC, Matsuzawa K, Sato K, Saiki I, Tonooka S, Samukawa K and Azuma I (1995)

Inhibitory effect of tumor metastasis in mice by saponins, ginsenoside-Rb2, 20(R)- and 20(S)-

ginsenoside-Rg3, of red ginseng. Biol Pharm Bull 18:1197–1202.

Nakajima S, Uchiyama Y, Yoshida K, Mizukawa H and Haruki E (1998) The effect of ginseng radius

rubra on human vascular endothelial cells. Am J chin Med 26:365-373.

Narita JO and Gruissem W (1989) Tomato hydroxyl-methylglutaryl-CoA reductase is required early in

fruit development but not during ripening. Plant Cell 1:181-190.

Nelson BK, Cai X and Nebenführ A (2007) A multicolored set of in vivo organelle markers for co-

localization studies in Arabidopsis and other plants. Plant J 51:1126–1136.

Ohyama K, Suzuki M, Masuda K, Yoshida S and Muranaka T (2007) Chemical phenotypes of the

hmg1 and hmg2 mutants of Arabidopsis demonstrate the in-planta role of HMG CoA reductase in

triterpene biosynthesis. Chem Pharm Bull 55:1518:1521.

Park JD (1996) Recent studies on the chemical constitutes of Korean ginseng. Korean J Ginseng Sci 20:

389-415

Radad K, Gille G, Liu L, Rausch WD (2006) Use of ginseng in medicine with emphasis on

neurodegenerative disorders. J Pharmacol Sci 100:175-186

Re EB, Jones D and Learned RM (1995) Co-expression of native and introduced genes reveals cryptic

regulation of HMG CoA reductase expression in Arabidopsis. Plant J 75(5):771-784.

Reumann S, Babujee L, Ma C, Wienkoop S, Siemsen T, Antonicelli GE, Rasche N, Luder F,

Weckwerth W and Jahn O (2007) Proteome analysis of Arabidopsis leaf peroxisomes reveals novel

targeting peptides, metabolic pathways, and defense mechanisms. Plant Cell 19:3170–3193.

Rodríguez-Concepción M, Fores O, Martinez-Garcia JF, Gonzalez V, Phillips MA, Ferrer A and

Boronat A (2004) Distinct light-mediated pathways regulate the biosynthesis and exchange of

isoprenoid precursors during Arabidopsis seedling development. Plant Cell 6:144–156.

Ruiz-albert J, Cerda-Olmedo E and Corrochano LM (2001) Genes for mevalonate biosynthesis in

Phycomyces. Mol Gen Genet 266:768-777.

Sapir-Mir M, Mett A, Belausov E, Tal-Meshulam S, Frydman A, Gidoni G and Eyal Y (2008)

Peroxisomal localization of Arabidopsis isopentenyldiphosphateisomerases suggests that part of the

plant isoprenoid mevalonic acid pathway is compartmentalized to peroxisomes. Plant Physiol

148:1219–1228.

Sato K, Mochizuki M, Saiki I, Yoo YC, Samukawa K and Azuma I (1994) Inhibition of tumor

angiogenesis and metastasis by a saponin of Panax ginseng, ginsenoside-Rb2. Biol Pharm Bull

17:635–639.

Schaller H, Grausem B, Benveniste P, Chye ML, Tan CT, Song YH and Chua NH (1995) Expression

of the Hevea brasiliensis (H.B.K.) Mull. Arg. 3-hydroxy-3-methyl-Coenzyme A reductase 1 in

tobacco results in sterol overproduction. Plant Physiol 109:761–770.

Schutze MP, Peterson, PA, Jackson MR (1994) An N terminal double arginine motif maintains type-II

membrane proteins in the endoplasmic reticulum. EMBO J 13:1696–1705

Shen G, Pang Y, Wu W, Liao Z, Zhao L, Sun X and Tang X (2006) Cloning and characterization of a

root-specific expressing gene encoding 3-hydroxy-3-methylglutaryl coenzyme A reductase from

Ginkgo biloba. Mol Biol Rep 33:117-127.

Shi W, Wang Y, Li J, Zhang H and Ding L (2007) Investigation of ginsenosides in different parts and

ages of Panax ginseng. Food Chem 102:664–668.

Shi YSC, Zheng B, Li Y and Wang Y (2010) Simultaneous determination of nine ginsenosides in

functional foods by high performance liquid chromatography with diode array detector detection.

Food Chem 123:1322-1327.

Shibata S (2001) Chemistry and cancer preventing activities of ginseng saponins and some related

triterpenoid compounds. J Korean Med Sci 16:S28-S37

Shin J, Park E and Choi G (2007) PIF3 regulates anthocyanin biosynthesis in an HY5-dependent

manner with both factors directly binding anthocyanin biosynthetic gene promoters in Arabidopsis.

Plant J 49:981-994.

Souret FF, Kim YJ, Wyslouzil BE, Wobbe KK and Weathers PJ (2003) Scale-up of Artemisia annua

L. hairy root cultures produces complex patterns of terpenoid gene expression. Biotechnol Bioeng

83(6):653-667.

Stermer BA, Bianchini GM and Korth KL (1994) Regulation of HMG-CoA reductases activity in

plants. J Lipid Res 35:1133-1140.

Suzuki M, Kamide Y, Nagat N, Seki H, Ohyama K, Kato H, Masuda K, Sato S, Kato T, Tabata S,

Yosida S and Muranaka T (2004) Loss of function of 3-hydroxy-3-methylglutaryl coenzyme A

reductase 1 (HMG1) in Arabidopsis leads to dwarfing, early senescence and male sterility, and

reduced sterol levels. Plant J 3:750-761.

Suzuki M, Nakagawa S, Kamide Y, Kobayashi K, Ohyama K, Hashinokuchi H, Kiuchi R, Saito K,

Muranaka T and Nagata N (2009) Complete blockage of the mevalonate pathway results in male

gametophyte lethality. J Exp Biol 60:2055-2064.

Tung N H, Song GY, Park YJ and Kim YH (2009) Two new dammarane-type saponins from the leaves

of Panax ginseng. Chem Pharm Bull 57:1412–1414.

Verhoeven H, Beuerle T and Schwab W (1997) Solid-phase microextraction: artifact formation and its

avoidance. Chromatographia 46:63-66.

Wu JY, Gardner BH, Murphy CI, Seals JR, Kensil CR, Recchia J, Beltz GA, Newman GW and

Newman MJ (1992) Saponin adjuvant enhancement of antigen-specific immune responses to an

experimental HIV-1 vaccine. J Immunol 148:1519–1525.

Xie JT, Wang CZ, Wang AB, Wu J, Basila D and Yuan CS (2005b) Anti hyperglycemic effects of total

ginsenosides from leaves and stem of Panax ginseng. Acta Pharmacol Sin 26(9):1104-1110.

Yokota S, Onohara Y and Shoyama Y (2011) Immunufluorescence and immunoelectron microscopic

localization of medicinal substance, Rb1, in several plant parts of Panax ginseng. Curr Drug

Discovery Tech 8:51-59.

Yue PY, Mak NK, Cheng YK, Leung KW, Ng TB, Fan DT, Yeung HW and Wong RN (2007)

Pharmacogenomics and the Yin/Yang actions of ginseng: antitumor, angiomodulating and steroid-like

activities of ginsenosides. Chin Med 2:6.

FIGURE LEGENDS

Figure 1. Biochemical pathway for biosynthesis of Panax ginseng saponins. Classification of

main ginsenosides based on attached glycosides and the dammarendiol-type structure (A). Glc,

β-D-glucopyranosyl; Ara (pyr), α-L-glucopyranosyl; Ara (fur), α-L-arabinofuranosyl; Rha, α-L-

rhamncpyranosyl. Ginsenoside biosynthesis pathway (B). FPS, farnesyl diphosphate synthase;

SS, squalene synthase; SE, squalene epoxidase; DDS, dammarenediol synthase; β-AS, β-amyrin

synthase; CAS, cycloartenol synthase; P450, cytochrome P450; GT, glucosyltransferase;

Mevinolin, a competitive inhibitor of HMGR. Dotted line displays putative pathway. Reported

enzymes in ginseng are shown with NCBI accession number in bracelet.

Figure 2. Analysis of sequence and domain characteristics of PgHMGR1 and PgHMGR2. The

DNA structures of PgHMGR1 and PgHMGR2 genes. Exons are represented by dark-filled

square boxes. Lines between boxes correspond to introns. Numbers above exons and below

introns indicate their length in bp. Indicated triangle represents the starting site used for H1CD

cloning in this study (A). The predicted 3-D structures of PgHMGR1 and PgHMGR2, which

consist of three domains: the small and helical N-terminal N-domain; the large, central L-domain;

and the small helical S-domain which is inserted into the L-domain. The catalytically active

residues (Abdin et al., 2012) are highlighted in yellow, and the sequences are indicated in

Supplemental Figure S5. (B). Schematic representation of the domain structures of PgHMGR1

and PgHMGR2. HMGR proteins are comprised of three domains: the membrane-anchor domain,

a flexible linker domain, and the catalytic domain. Within the catalytic domain, three subdomains

have been defined. H1 and H2 indicate two probable trans-membrane hydrophobic regions. LS,

lumenal sequence. The catalytic domain contains HMG-CoA- and NADH(H)-binding regions

and three conserved motifs. The active sites are indicated as *. The numbers indicate the position

in the amino acid sequence of the start and end of each domain and motif (C).

Figure 3. Ubiquitous expression of PgHMGR1 and PgHMGR2 in vascular tissue. Different aged

ginseng materials used for RT-PCR (A). Expression patterns of PgHMGR1 in seed, 1-week-old

seedlings, and leaves, petioles, and roots of 2-week-old of seedlings (B). Expression patterns of

PgHMGR1 in flower buds, leaves, stem, main roots, and lateral roots of 3-year-old (C) and 6-

year-old ginseng (D). Expression patterns of PgHMGR2 (E, F, G) were analyzed using the same

tissue as with PgHMGR1. Vertical bars indicate the mean value ± SE from three independent

experiments. Scale bars indicate 1 cm in A. Histochemical analysis of GUS expression in

transgenic Arabidopsis plants harboring the pHMGR1::GUS (H) and pHMGR2::GUS (I)

constructs at different developmental stages. (a) GUS expression in 2-day-old germinated

seedlings. Main roots (b), lateral roots (c), true leaves, and cotyledons (d) of 8-day-old seedlings.

Mature flowers (e) and siliques (f) of 50-day-old plants. All scale bars indicate 100 µm in H, I,

and J. PgHMGR1::GUS expression is shown in the root vasculature and root tip of ginseng (J).

Figure 4. Expression of PgHMGR1 and PgHMGR2 is associated with the production of

ginsenoside. Mevinolin (10 µM), an inhibitor of HMGR, was administered to 4-week-old

adventitious roots for 1, 3, and 7 days. We analyzed total ginsenoside contents (mg/g dry weight)

(A), gene expression of PgHMGR1 and PgHMGR2 (B), and relative HMGR activity (C). Methyl

jasmonate (MJ, 10 µM), an elicitor of saponin biosynthesis, was treated for 3 days, and the total

ginsenoside contents (mg/g dry weight) (D) and gene expression (E) were analyzed. Vertical bars

indicate the mean value ± SE from five independent experiments. *: Significantly different from

the control at P <0.05. (F) PgHMGR1::GUS- or PgHMGR2::GUS-expressing seedlings grown

on ½ MS media for 4 days were treated with ethanol control (Cont), 1 or 10 µM of mevinolin,

and 10 µM of MJ for 3 hours. Bar= 100 µM.

Figure 5. Expression of PgHMGR1 is light-inhibited and possibly associated with shadowing

growth of ginseng. Three-year-old ginseng plants cultured hydroponically in perlite and

peatmoss were kept in darkness for 2 or 3 days. Control plants (Cont) were grown in a 16-h

light/8-h dark cycle, and samples were taken under light conditions. Parts of the leaf and root

were separately sampled and used for RT-PCR, HMGR activity assay, and ginsenoside analysis.

Total ginsenoside contents in the leaves (A) and roots (B) and individual major ginsenosides

were analyzed (Supplemental Figure 4). Vertical bars indicate the mean value ± SE from three

independent experiments. * and **, Significantly different from the control at P <0.05 and **

P<0.01, respectively. Relative HMGR activity in the leaves and roots was analyzed (C). Data

represent mean value ± SE (n=10). Gene expression of PgHMGR1 and PgHMGR2 in leaves (D)

and roots (E) was analyzed by real-time PCR. Data represent mean value ± SE (n=5). Putative

PIF3 binding motifs are found on the G-box or ACE element of the analyzed promoter of the

PgHMGR1 gene, which are not found in PgHMGR2, by Plantpan (F). Hypocotyl expression in

4-day-old etiolated seedlings of PgHMGR1::GUS was decreased by 2 h of light exposure (G).

Expression in true leaves of 10-day-light-grown PgHMGR1::GUS seedling was decreased by 9 h

of darkness (H).

Figure 6. Subcellular localization of PgHMGR1, PgHMGR2, and the catalytic domain of

PgHMGR1. Fluorescent images of PgHMGR1-CFP (A) and PgHMGR1-mRFP (B) were

visualized by a confocal laser scanning microscope. A fluorescent image of PgHMGR2-CFP was

visualized with green color in an ER-like vesicle (C). The catalytic domain of PgHMGR1 tagged

with CFP at the C-terminus was localized in the cytosol and nucleus (D). Fluorescent signals of

PgHMGR1-mRFP [Pro35S:PgHMGR1–mRFP in the ER-YK (ER marker), PX-YK (peroxisome

marker), PT-YK (plastid marker) transgenic background] overlap with ER-YK, PX-YK, and PT-

YK foci (E). All scale bars indicate 5 µm.

Figure 7. Overexpression of PgHMGR1 or the catalytic domain of PgHMGR1 results in higher

production of sterol and triterpene in plants. Quantification of endogenous sterols and triterpenes

in Arabidopsis HMGR1ox and H1CDox. Rosette leaves (200 mg) and inflorescences (25 mg) of

5-week-old plants were harvested and freeze-dried. The samples were then analyzed using GC-

MS. Values represent the mean ±SE of content (µg per g dry weight) calculated from three

independent sterol isolations and quantification. * and **: Significantly different from control at

P <0.05 and P<0.01, respectively. Quantitative analysis of triterpene ginsenoside contents in the

control and transgenic ginseng adventitious roots of Panax ginseng (cv. Yunpoong)

overexpressing PgHMGR1 were analyzed by HPLC (C). Vertical bars indicate the mean ± SE

from three independent experiments.

Figure 8. Conserved role of PgHMGR1. A phylogenetic tree of HMGRs from various organisms

including PgHMGR1 and PgHMGR2 (A). Eleutherococcus senticosus (EsHMGR, AFM77981),

Eucommia ulmoides (EuHMGR, AAV54051), Salvia miltiorrhiza (SmHMGR, ACD37361),

Panax quinquefolius (PqHMGR, ACV65036), Camptotheca acuminata (CaHMGR, AAB69727),

Solanum chacoense (ScHMGR, AEX26934), Arabidopsis thaliana (AtHMGR1S, At1g76490;

AtHMGR2, At2g17370), fungi Pichia jadinii (O74164), and animals such as Homo sapiens

(NP_000850), Drosophila melanogaster (P14773), and Mus musculus (NP_032281). The

neighbor-joining method was used, and the branch lengths are proportional to the divergence,

with the scale of 0.1 representing 10% changes. The conserved motifs among the members are

highlighted in colored boxes with an arranged number, and the sequences of the motifs are listed

in Supplemental Figure S7 (B). PgHMGR1 complemented hmgr1-1. Structure of T-DNA

insertion position in the Arabidopsis genome of SALK line (SALK_125435) (C). Closed boxes

are exons. The T-DNA insertion site is indicated by a flag. T-DNA sequences were inserted in the

first exon in hmg1-1. Expression analysis of the AtHMGR1, AtHMG2, AtSE1, and SAG12, a

senescence marker gene, in 3-week-old seedlings (F) of wild-type (WT), heterozygous (HZ), and

homozygous (HM) lines of hmgr1-1 (D). The expression level of an actin gene was used as an

internal control. Real-time quantitative RT-PCR analysis of AtHMGR1 confirmed the knock-out

of hmgr1-1. Data are normalized with β-actin and are relative to the levels in WT seedlings.

Fifty-day-old Arabidopsis plants (E, G-I). Heterozygous (HZ) hmgr1-1 shows a similar

phenotype to the wild-type (Col), whereas the homozygous (HM) mutant shows a severe dwarf

and sterile phenotype (E, F). Exogenous squalene treatment rescued the dwarf phenotype (E).

PgHMGR1ox (G) and PgHMGR1CDox (H) complemented the hmgr1-1 dwarf phenotype,

whereas PgHMGR2ox (I) did not.

Figure 9. Presumptive model for the HMGR-triggered ginsenoside biosynthesis. Sterols and

triterpenes produced via mevalonate (MVA) are catalyzed by 3-hydroxy-3-methylglutaryl-CoA

reductase (HMGR) in the endoplasmic reticulum (ER) or peroxisome of the vasculature of

ginseng. HMGR can be up-regulated by the light signaling factor, phytochrome-interacting factor

3 (PIF3), regulated by phyB (phytochrome B).

Supplemental Table 1. Primers used in confirmation of gDNA insertion and RT-PCR

No. Organism Gene Accession

No. Annotation Primers used (5'-3')

Annealing

temp. (°C)

1 Ginseng PgHMGR1

HMG-F TGCTGCCAATATCGTCTCTG 60

2

HMG-R CCAAGAAGGTTCAAGCAAGC

3 Ginseng PgHMGR1

H1_RT_F2 ATGTAGATCGCTCGCCGT 60

4

H1_RT_R2 GTGGAGGAGATAGTATGCGAC

5 Ginseng PgHMGR2

H2_RT_F AGGCCTTCGCCTGATGCCT 60

6

H2_RT_R GCAAATGCTGACTTGGGATT

7 Ginseng Pgactin DC03005B05 actin-left AGAGATTCCGCTGTCCAGAA 60

8

actin-right ATCAGCGATACCAGGGAACA

9 Arabidopsis AtHMGR1 At1G76490 Arabi_H1_F ATTGTCACCGAATCGCTTCC 59

10

Arabi_H1_R GATCTCCCGGTGACTCTCTG

11 Arabidopsis AtHMGR2 At2G17370 Arabi_H2_1300F TGAAGCTGAAGGCAACGACCT 59

12

hmgr2-2RP CTGTTGACTTGAGACGAAGGG

13 Arabidopsis AtSE1 At1g58440 AtSE_F AGCTGGTGTTGCTGGTTCT 57

14

AtSE_R CTTCCACACAATCTTCAATTCC

15 Arabidopsis SAG12 At5g45890 AtSAG12_F ATTCCTGCCGGAAGAACTTT 54

16

AtSAG12_R TATCCATTAAACCGCCTTCG

17 Arabidopsis Atactin At5g09810 At-actin2F GTGTGTCTTGTCTTATCTGGTTCG 58

18

At-actin2R AATAGCTGCATTGTCACCCGATACT

19 Vector CFP, GFP

GFP342rv CTCGACCAGGATGGGCAC

20

35S

promoter 35S GCACAATCCCACTATCCTTCG

21

NOS terminator p33NOS ACCGGCAACAGGATTCAATCT

37

Supplemental Data

Supplemental Figure S1. Genomic DNA gene sequence of PgHMGR1 with its promoter (-1317

to -1). The full-length genomic DNA sequence with its deduced amino acid sequence. A one-

letter code below the corresponding codons and intron sequence is shown. Thin-line boxes in the

promoter region=putative ‘TATA’ boxes; thin-line underlines=putative ‘CAAT’ boxes; shaded

bold box=putative ‘G-box’; bold underline=putative ‘ACE element’; shaded box in the N-

terminal=putative motif for ER retention; square boxes in the intron region=putative intron-exon

boundaries with the GT-AG rule.

Supplemental Figure S2. Genomic DNA gene sequence of PgHMGR2 with its promoter (-477

to -1). The full-length genomic DNA sequence with its deduced amino acid sequence. A one-

letter code below the corresponding codons and intron sequence is shown. Thin-line boxes in the

promoter region=putative ‘TATA’ boxes; thin-line underline=putative ‘CAAT’ boxes; shaded box

in the N-terminal=putative motif for ER retention; square boxes in the intron region=putative

intron-exon boundaries with the GT-AG rule.

Supplemental Figure S3. Ginsenoside contents by inhibitor or elicitor treatment. Individual

ginsenosides (A) were analyzed in 4-week-old adventitious roots (B) following 3 day treatment

with the chemicals; Mevinolin (10 µM), an inhibitor of HMGR and methyl jasmonate (MeJA, 10

µM), an elicitor of saponin biosynthesis. Vertical bars indicate the mean ± SE from three

independent experiments. *: Significantly different from control at P <0.05.

Supplemental Figure S4. Ginsenoside contents by dark treatment. Three-year-old ginseng

plants cultured hydroponically in perlite and peatmoss were kept in darkness for 2 or 3 days.

Control plants (Cont) were grown in a 16-h light/8-h dark cycle, and samples were taken under

light conditions. Individual major ginsenosides in leaves (A) and roots (B) were analyzed.

Vertical bars indicate the mean ± SE from three independent experiments. * and **, Significantly

different from control at P <0.05 and ** P<0.01, respectively.

Supplemental Figure S5. Multiple alignment of the deduced amino acid sequences of

38

PgHMGR1 and PgHMGR2 with homologous HMGRs from other plants (Fig. 8):

Eleutherococcus senticosus (EsHMGR, AFM77981), Eucommia ulmoides (EuHMGR,

AAV54051), Salvia miltiorrhiza (SmHMGR, ACD37361), Panax quinquefolius (PqHMGR,

ACV65036), Camptotheca acuminata (CaHMGR, AAB69727), Solanum chacoense (ScHMGR,

AEX26934), and Arabidopsis thaliana (AtHMGR1S, At1g76490; AtHMGR2, At2g17370).

Black boxes indicate identical residues; gray boxes indicate identical residues for at least two of

the sequences. Conserved domains are indicated with arrowed lines and three conserved motifs

are highlighted in colored boxes. Two putative HMGR-CoA-binding sites and two NADP(H)-

binding sites are indicated with a thin square box (Shen et al., 2006), and the motif responsible

for ER retention is highlighted with bold square box (Merret et al., 2007). The indicated triangle

represents the starting site for the H1CD cloning in this study.

Supplemental Figure S6. Conserved motifs among plant and animal HMGRs. Conserved

residues analyzed by MEME are represented by three motifs. The catalytically active residues

marked by a box (Abdin et al., 2012) were conserved in these three motifs.