• Open Access

Overexpression of Brassica juncea wild-type and mutant HMG-CoA synthase 1 in Arabidopsis up-regulates genes in sterol biosynthesis and enhances sterol production and stress tolerance


(Tel +852 22990319; fax +852 28583477; email mlchye@hkucc.hku.hk)


Brassica juncea 3-hydroxy-3-methylglutaryl-CoA synthase (HMGS) is encoded by four isogenes (BjHMGS1-BjHMGS4). In vitro enzyme assays had indicated that the recombinant BjHMGS1 H188N mutant lacked substrate inhibition by acetoacetyl-CoA (AcAc-CoA) and showed 8-fold decreased enzyme activity. The S359A mutant demonstrated 10-fold higher activity, while the H188N/S359A double mutant displayed a 10-fold increased enzyme activity and lacked inhibition by AcAc-CoA. Here, wild-type and mutant BjHMGS1 were overexpressed in Arabidopsis to examine their effects in planta. The expression of selected genes in isoprenoid biosynthesis, isoprenoid content, seed germination and stress tolerance was analysed in HMGS overexpressors (OEs). Those mRNAs encoding enzymes 3-hydroxy-3-methylglutaryl-CoA reductase (HMGR), sterol methyltransferase 2 (SMT2), delta-24 sterol reductase (DWF1), C-22 sterol desaturase (CYP710A1) and brassinosteroid-6-oxidase 2 (BR6OX2) were up-regulated in HMGS-OEs. The total sterol content in leaves and seedlings of OE-wtBjHMGS1, OE-S359A and OE-H188N/S359A was significantly higher than OE-H188N. HMGS-OE seeds germinated earlier than wild-type and vector-transformed controls. HMGS-OEs further displayed reduced hydrogen peroxide (H2O2)–induced cell death and constitutive expression of salicylic acid (SA)-dependent pathogenesis-related genes (PR1, PR2 and PR5), resulting in an increased resistance to Botrytis cinerea, with OE-S359A showing the highest and OE-H188N the lowest tolerance. These results suggest that overexpression of HMGS up-regulates HMGR, SMT2, DWF1, CYP710A1 and BR6OX2, leading to enhanced sterol content and stress tolerance in Arabidopsis.


Isoprenoids (terpenoids or terpenes) from eukaryotes and prokaryotes represent a large class of over 50 000 natural products, of which most are derived from plants (Bach, 1995; Roberts, 2007; Kirby and Keasling, 2008). In higher plants, isoprenoids originate from a common 5-carbon unit, isopentenyl diphosphate (IPP) and its isomer dimethylallyl diphosphate (DMAPP), which are biosynthesized via either the cytosolic mevalonate (MVA) pathway (cf. Bach, 1995) or the plastidial methylerythritol phosphate (MEP) pathway (cf. Rohmer, 1999) (Figure 1). The MVA pathway is responsible for supplying the precursors for sterols, ubiquinone and some sesquiterpenes (Edwards and Ericsson, 1999). 3-Hydroxy-3-methylglutaryl-CoA synthase (HMGS) is the second enzyme in the MVA pathway and catalyses the condensation of AcAc-CoA with acetyl-CoA (Ac-CoA) to produce 3-hydroxy-3-methylglutaryl-CoA (HMG-CoA) (Lynen et al., 1958; Ferguson and Rudney, 1959; Rudney and Ferguson, 1959; Stewart and Rudney, 1966; Lynen, 1967; Balasubramanlam et al., 1977). 3-Hydroxy-3-methylglutaryl-CoA reductase (HMGR) then converts HMG-CoA to MVA, which further undergoes two phosphorylation and a decarboxylation reactions to form IPP, the universal precursor of all isoprenoids (cf. Nagegowda, 2010). IPP units for the formation of carotenoids, the side chains of chlorophylls, abscisic acid (ABA), gibberellins (GA), plastoquinones, monoterpenes and tocopherols are synthesized via the MEP pathway (cf. Eisenreich et al., 2001). Although the MVA and MEP pathways can operate independently, metabolic flow has been reported to occur between these pathways, suggesting some cross-talk between them (Hemmerlin et al., 2003; Laule et al., 2003; Chow et al., 2007).

Figure 1.

 Simplified isoprenoid biosynthesis pathways in Arabidopsis. Subcellular compartments are indicated in uppercase italics. Solid arrows indicate single-step reactions; dashed arrows denote several steps. Enzymes are shown in bold. Arrows between cytosolic and plastid compartments represent metabolic flow between them (thicker arrow for greater flux). At5g23960, Arabidopsis thaliana sesquiterpene synthase. Abbreviations: ABA, abscisic acid; BR6OX2, brassinosteroid-6-oxidase 2; CYP710A1, C-22 sterol desaturase; DMAPP, dimethylallyl diphosphate; DWF1, delta-24 sterol reductase; DXR, 1-deoxy-D-xylulose 5-phosphate reductoisomerase; FPP, farnesyl diphosphate; GA-3P, glycerol-3 phosphate; GAs, gibberellins; GGPP, geranylgeranyl diphosphate; GPP, geranyl diphosphate; HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA; HMGS, 3-hydroxy-3-methylglutaryl-CoA synthase; HMGR, 3-hydroxy-3-methylglutaryl-CoA reductase; IPP, isopentenyl diphosphate. SMT1, sterol methyltransferase 1; SMT2, sterol methyltransferase 2; SMT3, sterol methyltransferase 3.

Isoprenoids have been proven to play significant roles in plant–environment interactions including plant–plant, plant–insect, plant–microbe, plant–stresses and plant–animal communications (Pichersky and Gershenzon, 2002; Violeta et al., 2005; Keeling and Bohlmann, 2006; Floßet al., 2008; Dudareva and Pichersky, 2008). For example, accumulated sesquiterpenoid phytoalexins (rishitin and lubimin) in potato enhanced disease resistance to P. infestans (Choi et al., 1994). Furthermore, isoprenoids have essential physiological functions in membrane biogenesis (sterols), photosynthesis (carotenoids and chlorophyll), respiration (quinones) and regulation of growth and development (ABA, GA and brassinosteroids) (Bush and Grunwald, 1972; Bach and Lichtenthaler, 1983; Bach, 1995; Hedden and Kamiya, 1997; Clouse and Sasse, 1998; Montoya et al., 2005). Sterols are also important in embryogenesis (Clouse, 2000) and in mediating stress tolerance (Poséet al., 2009).

Given the importance of isoprenoids, attempts have been made to overaccumulate them in transgenic plants. Overexpression of Hevea brasiliensis HMGR in transgenic tobacco resulted in 2- to 6-fold increase in sterol production, with most detected as steryl esters (Schaller et al., 1995). Transgenic tobacco overexpressing a hamster HMGR also accumulated sterol intermediates such as cycloartenol, while end-product sterols (campesterol, sitosterol and stigmasterol) merely doubled (Chappell et al., 1995). Zook et al. (1996) expressed a fungal-derived trichodiene synthase in tobacco, resulting in overproduction of sesquiterpene in elicited cells. Furthermore, traits to combat plant pests have been added through isoprenoid engineering (cf. Dudareva and Pichersky, 2008).

In B. juncea (Indian mustard), four isogenes (BjHMGS1–BjHMGS4) encode HMGS (Alex et al., 2000). BjHMGS was down-regulated by ABA, osmotic stress and dehydration and up-regulated by salicylic acid (SA), methyl jasmonate (MeJA) and wounding (Alex et al., 2000). BjHMGS1 and BjHMGS2 were confirmed to be expressed in young floral buds and leaves, while BjHMGS3 and BjHMGS4 were mainly expressed in leaves (Nagegowda et al., 2005). BjHMGS1 has been demonstrated to be a cytosolic enzyme using transient expression of N- and C-terminal green fluorescent protein fusions in onion epidermal cells and tobacco BY-2 cells (Nagegowda et al., 2005). His-tagged BjHMGS1 is inhibited in vitro by one of its substrates (AcAc-CoA) and by both products (HMG-CoA and CoA-HS) (Nagegowda et al., 2004). Replacement of conserved amino acid residues in BjHMGS1 by site-directed mutagenesis and the expression of such mutants in Escherichia coli resulted in alterations in BjHMGS1 activity in vitro. Enzyme kinetic analyses revealed that the H188N (His→Asn) mutant did not display substrate inhibition by AcAc-CoA and showed 8-fold lower enzyme activity; substitution S359A (Ser→Ala) resulted in a 10-fold increased specific activity, while double mutant H188N/S359A demonstrated 10-fold increased specific activity and lacked inhibition by AcAc-CoA (Nagegowda et al., 2004).

The aim of this study was to investigate the effects of overexpression of wild-type (wt) and mutant (H188N, S359A and H188N/S359A) BjHMGS1 in Arabidopsis thaliana Col-0. HMGS overexpression in Arabidopsis HMGS-OEs not only elevated the expression of HMGR, given that both are co-ordinately regulated in B. juncea (Alex et al., 2000), but SMT2, DWF1, CYP710A1 and BR6OX2, further down in the MVA pathway related to sterol biosynthesis, were also induced. Subsequently, changes in end products and phenotype of HMGS-OEs were investigated.


Generation and molecular analyses of transgenic Arabidopsis HMGS-OEs

The wild-type and mutant BjHMGS1 cDNAs were cloned into the plant binary vector pSa13 (Xiao et al., 2008) in which the cauliflower mosaic virus 35S promoter (CaMV 35S) directs the expression of the cloned cDNA. Plasmid constructs pBj132 (encoding BjHMGS1 H188N/S359A), pBj134 (wtBjHMGS1), pBj136 (BjHMGS1 S359A) and pBj137 (BjHMGS1 H188N) were generated, and resultant Agrobacterium transformants were used to transform Arabidopsis yielding 30–60 kanamycin-resistant transgenic lines per construct. Putative Arabidopsis HMGS-OEs were designated as OE-wtBjHMGS1 (wild-type BjHMGS1), OE-H188N (BjHMGS1 H188N), OE-S359A (BjHMGS1 S359A) and OE-H188N/S359A (BjHMGS1 H188N/S359A). The presence of each transgene in Arabidopsis was preliminarily verified by PCR (Figure S1a,b). Mutations on the cDNA were further validated by DNA sequence analysis of PCR products amplified from genomic DNA of Arabidopsis HMGS-OEs. PCR-positive lines were then tested by Western blot analysis for overproduction of HMGS (Figure 2a). Given that the peptide used to generate anti-BjHMGS antibodies in rabbit showed 100% homology to Arabidopsis HMGS (At4g11820), the faint band in protein of the vector (pSa13)-transformant is attributed to cross-reaction with endogenous HMGS (Figure 2a).

Figure 2.

 Molecular analyses on rosette leaves of 3-week-old Arabidopsis HMGS-OEs. (a) Western blot analysis using antibodies against BjHMGS indicates overproduction of HMGS (52.4-kDa) in representative Arabidopsis vector-transformed control (pSa13) and HMGS-OEs (OE-wtBjHMGS1, OE-H188N, OE-S359A and OE-H188N/S359A). (b) Northern blot analyses showing the mRNA levels of HMG1 (2.2-kb), BjHMGS1 (1.7-kb), DWF1 (2.0-kb), PR2 (1.2-kb), PR1 (0.7-kb), PR5 (0.5-kb) and PDF 1.2 (0.8-kb) in representative pSa13 and OE lines. (c) RT-PCR demonstrates the relative transcript abundance (as compared to ACTIN) of endogenous BR6OX2, CYP710A1, SMT2 and DXR mRNAs in representative pSa13 and OE lines. BR6OX2, CYP710A1 and SMT2 mRNAs in HMGS-OE lines were significantly higher than in the vector (pSa13)-transformed line. h, Higher than the vector (pSa13)-transformed line. Values are means ± SD (= 3); **< 0.01 and *< 0.05 by U test. Bars represent SD. (d) Northern blot analysis of mRNA encoding floral sesquiterpene synthase (At5g23960, 1.8-kb), which was not induced in OE-wtBjHMGS1 and OE-S359A, representative of OE lines.

Northern blot analysis indicated that all HMGS-OEs accumulated higher levels of BjHMGS1 mRNA (Figure 2b) than the pSa13-transformant, which lacks the AtHMGS hybridizing band following stringent washing. OEs were confirmed to be independent lines by Southern blot analysis that also revealed the number of copies of BjHMGS1 inserted in the Arabidopsis genome (Figure S1c). Two independent OE lines from each construct with single BjHMGS1 inserts and of similar HMGS expression levels were subsequently chosen for further analyses on isoprenoid content, germination rate, resistance to fungal infection and sensitivity to H2O2 treatment.

Arabidopsis HMGS-OEs show induced expression of SMT2, DWF1, CYP710A1 and BR6OX2 besides HMGR

The co-ordinated regulation of HMGS and HMGR encoding two adjacent enzymes in the MVA pathway has been previously reported (Alex et al., 2000). However, it is not known whether the expression of mRNAs encoding these enzymes in early MVA biosynthesis is co-regulated with those further downstream in the pathway. In Arabidopsis, HMGR, DWF1, CYP710A1, BR6OX2 (or CYP85A2) and sesquiterpene synthase (At5g23960) catalyse the production of MVA (Suzuki and Muranaka, 2007), campesterol/sitosterol (Choe et al., 1999), stigmasterol (Arnqvist et al., 2008), brassinolide (BL; Nomura et al., 2005) and floral sesquiterpenes [(–)-α-copaene, (–)-E-β-caryophyllene and α-humulene] (Tholl et al., 2005) biosyntheses, respectively. All five enzymes occur after HMGS in the MVA pathway of isoprenoid biosynthesis, in contrast to 1-deoxy-D-xylulose 5-phosphate reductoisomerase (DXR), a key enzyme of the MEP pathway (Figure 1). SMT1, SMT2 and SMT3 have been reported to control the ratio of sterol content in Arabidopsis (Schaller et al., 1998; Schaeffer et al., 2001; Carland et al., 2010). To address whether their mRNA expression is affected in Arabidopsis HMGS-OEs, Northern blot (Figure 2b,d) and RT-PCR (Figures 2c and S2a,d) analyses were carried out. HMG1, HMG2, SMT2, DWF1, CYP710A1 and BR6OX2 mRNAs were induced in all OEs (Figures 2b–c and S2a), whereas the mRNAs encoding sesquiterpene synthase (At5g23960; Figure 2d), DXR (Figure 2c), SMT1 (Figure S2d) and SMT3 (Figure S2d) remained unaffected.

Arabidopsis HMGS-OEs accumulate sterols

The overexpression of HMGR has been studied in eukaryotes and results in sterol accumulation in plants (Chappell et al., 1995; Schaller et al., 1995). However, the overexpression of HMGS in eukaryotes has not been investigated; hence, Arabidopsis HMGS-OEs were analysed for sterol contents using gas chromatography–mass spectrometry (GC-MS) analysis. The results indicated that the sterol contents of OEs generally increased in both 14-day-old seedlings cultured in petri dishes (Table 1) and 3-week-old soil-grown rosettes leaves (Table 2). In comparison with total sterol content of wild-type seedlings (Tables 1 and S1) and leaves (Tables 2 and S2), there was a significant increase in all OE lines except for OE-H188N. OE-S359A showed the highest increment in seedlings (24.5%–29.0%) (Table S1) and leaves (18.7%–25.9%) (Table S2), followed by OE-H188N/S359A, while OE-H188N had the least increment. Among individual sterols, the sitosterol content also displayed a similar trend as that of total sterol with an increase of 19.4%–24.0% in seedlings and 12.0%–22.7% in leaves of OE-S359A (Tables S1 and S2). Campesterol content was significantly increased in OE-S359A (20.3%–26.2%) and OE-H188N/S359A (8.7%–12.6%) seedlings (Table S1), whereas the leaf campesterol content was significantly increased in all OE-lines except OE-H188N (Table S2). Stigmasterol content of seedlings showed the highest percentage increment (112.5%–150%) among the individual sterols in all OE lines (Table S1), whereas its leaf content was significantly increased only in OE-S359A (23.8%–54.8%) (Table S2). Although OE-H188N indicated no significant difference in total sterol, campesterol and sitosterol, its seedlings accumulated stigmasterol as other OE lines. Overall, OE-S359A displayed the highest increase in total as well as individual sterol contents both in seedlings and in leaves. The ratio of campesterol to sitosterol was calculated in wild-type, vector-transformed control and HMGS-OEs, but no significant differences were observed (Figure S2c) in contrast to results from analysis on sterol methyltransferase (SMT)-OE plants (Schaller et al., 1998; Schaeffer et al., 2001). Trace amounts of cholesterol, isofucosterol, cycloartenol, delta-7-sitosterol, delta-7-avenasterol, 24-methylene cycloartanol and 24-ethylidene lophenol were also detected in all lines tested, but no significant differences between BjHMGS-OEs and vector (pSa13)-transformed controls (data not shown) were observed. Also, no significant differences in floral (–)-E-β-caryophyllene (Figure S4a), plastid chlorophyll (Figure S4b) and carotenoid contents (Figure S4c) were observed between HMGS-OEs and pSa13-transformants. In all experiments, wild-type Arabidopsis was also tested, but it did not demonstrate any significant differences from the vector-transformed control (Tables 1 and 2).

Table 1.   Sterol composition (μg/mg dry weight) in HMGS-OE, wild-type and vector (pSa13)-transformed Arabidopsis seedlings
WtpSa13Line 1Line 2Line 1Line 2Line 1Line 2Line 1Line 2
  1. Wt, wild-type.

  2. Values are means ± SD, = 5. *< 0.05; **< 0.01 by U test.

  3. Bold font indicates value that is significantly higher than the vector-transformed control (pSa13).

Campesterol1.02 ± 0.021.03 ± 0.051.04 ± 0.031.05 ± 0.030.99 ± 0.051.00 ± 0.041.24 ± 0.04**1.30 ± 0.13**1.16 ± 0.07**1.12 ± 0.05**
Stigmasterol0.15 ± 0.050.16 ± 0.010.34 ± 0.05**0.36 ± 0.05**0.34 ± 0.06**0.35 ± 0.04**0.38 ± 0.03**0.40 ± 0.04**0.37 ± 0.04**0.35 ± 0.03**
Sitosterol3.86 ± 0.063.87 ± 0.224.21 ± 0.17**4.21 ± 0.13**3.78 ± 0.213.82 ± 0.184.62 ± 0.09**4.85 ± 0.40**4.38 ± 0.22**4.28 ± 0.18**
Total sterol5.11 ± 0.195.19 ± 0.135.73 ± 0.15**5.81 ± 0.19**5.26 ± 0.265.30 ± 0.216.46 ± 0.17**6.70 ± 0.56**6.04 ± 0.27**5.94 ± 0.21**
Table 2.   Sterol composition (μg/mg dry weight) in HMGS-OE, wild-type and vector (pSa13)-transformed Arabidopsis rosette leaves
WtpSa13Line 1Line 2Line 1Line 2Line 1Line 2Line 1Line 2
  1. Wt, wild-type.

  2. Values are means ± SD, = 5. *< 0.05; **< 0.01 by U test.

  3. Bold font indicates value that is significantly higher than the vector-transformed control (pSa13).

Campesterol0.43 ± 0.040.43 ± 0.020.46 ± 0.03*0.47 ± 0.05*0.44 ± 0.030.46 ± 0.050.54 ± 0.07**0.54 ± 0.05**0.51 ± 0.06**0.50 ± 0.06**
Stigmasterol0.042 ± 0.0070.042 ± 0.0070.036 ± 0.0040.043 ± 0.0100.046 ± 0.0050.044 ± 0.0050.052 ± 0.009**0.065 ± 0.012**0.046 ± 0.0080.044 ± 0.005
Sitosterol3.11 ± 0.202.99 ± 0.213.37 ± 0.14**3.58 ± 0.19**3.06 ± 0.103.10 ± 0.183.35 ± 0.31**3.67 ± 0.19**3.62 ± 0.35**3.28 ± 0.19**
Total sterol3.84 ± 0.163.79 ± 0.234.22 ± 0.27**4.39 ± 0.32**3.76 ± 0.163.84 ± 0.274.50 ± 0.38**4.77 ± 0.47**4.75 ± 0.46*4.34 ± 0.31**

Arabidopsis HMGS-OE seeds germinate earlier

Arabidopsis HMGS-OE seedlings showed induced DWF1 and CYP710A1 mRNA expression and consequently overaccumulated stigmasterol. In tobacco, stigmasterol is known to significantly increase during seed germination (Bush and Grunwald, 1972). Hence, the germination rates of HMGS-OE seeds were examined next. The results from germination assays indicated that OE seeds germinated approximately 1 day earlier than wild-type and pSa13-transformed seeds (Figure 3). The germination rates of wild-type and pSa13-transformed vector control seeds were significantly lower (< 0.01) than OE seeds at 36, 48 and 60 h following transfer from 4 °C to normal growth conditions (Figure 3). Furthermore, 3-day-old OE seedlings show faster growth than wild-type and vector (pSa13)-transformed seedlings (data not shown). Similar results were obtained in this test using two generations of seeds.

Figure 3.

 Seed germination in Arabidopsis HMGS-OEs. Statistical data on germination rates recorded at 36, 48, 60 and 72 post-germination, showing differences at 36, 48 and 60, but not 72 h. Values are means ± SD (= 5); H, Higher than pSa13-transformants; **< 0.01 by U test. Bars represent SD.

Arabidopsis HMGS-OEs show induced PR gene expression and enhanced resistance to Botrytis cinerea

As both BjHMGS and BjHMGR mRNAs are SA-inducible (Alex et al., 2000), we were interested to test whether the elevation of HMGR mRNA in HMGS-OEs would confer SA-associated fungal resistance. It has also been reported very recently that brassinosteroids (BR) which appears elevated in HMGS-OEs up-regulates PR-1 (Divi et al., 2010). Hence, the expression of SA-dependent PR genes was investigated, and HMGS-OEs were subject to fungal infection by B. cinerea. Northern blot analysis using PCR-generated cDNA probes demonstrated that the expression of SA-dependent PR genes, PR1 (At2g14610, unknown protein), PR2 (At3g57260, β-1,3-glucanase) and PR5 (At1g75040, thaumatin-like protein), was induced in OE lines similar to the induction of HMGR and DWF1 (Figure 2b). The results from infection assays with B. cinerea 3 days after treatment revealed that detached leaves of OE lines were more resistant and showed reduction (17.9%–22.8%) in diameter size of lesions (Figure 4a,b).

Figure 4.

Botrytis cinerea infection assay on rosette leaves of Arabidopsis HMGS-OEs. (a) Phenotype of B. cinerea infection on detached rosette leaves, 3 days after treatment. Bar, 50 mm. (b) Statistical data of lesion size shown in (a), demonstrating all OE leaves showed reduction in lesion sized in comparison with pSa13-transformed leaves (100%); reductions being OE-wtBjHMGS1 by 21.7%, OE-H188N 17.9%, OE-S359A 22.8% and OE-H188N/S359A 20.7%. Values are means ± SD (= 30); L, lower than pSa13; **< 0.01 by U test. Bars represent SD.

Detached leaves of Arabidopsis HMGS-OEs are delayed in H2O2-induced cell death

Paniagua-Pérez et al. (2008) have verified that β-sitosterol functions as an antioxidant by removing free radicals from diphenylpicrylhydrazyl. As sitosterol content was elevated in OE-wtBjHMGS1, OE-S359A and OE-H188N/S359A lines, their 3-week-old rosette leaves were subjected to H2O2 (3 or 5 mm) treatment. All OE leaves exhibited higher tolerance to H2O2 at both concentrations than vector-transformed leaves, following 72-h treatment. Figure 5a indicates the phenotype of H2O2-induced cell death photographed 72 h after treatment. The results of trypan blue staining confirmed that more cells of the vector-transformed leaves had undergone cell death in comparison with OE leaves treated with either 3 or 5 mm H2O2 (Figure 5b). According to the electrolyte conductivity data, OE-S359A leaves showed highest tolerance to 3 and 5 mm H2O2, while OE-H188N leaves were the least tolerant ones among OE lines (Figure 5c). Interestingly, the OE-H188N leaves, which had not exhibited a significant increase in sitosterol content in comparison with the other OEs, demonstrated highest electrolyte leakage (15.2% or 18.1% in Figure 5c) when treated with H2O2 (3 or 5 mm, respectively).

Figure 5.

 Treatment with H2O2 on rosette leaves of Arabidopsis HMGS-OEs. (a) Phenotype of H2O2-induced cell death photographed 72 h after treatment. Representative leaves are shown here. (b) Trypan blue-stained leaves after treatment for 72 h. Dark blue indicates cell death. Representative leaves were shown here. (c) Ion leakage indicates cell death of leaves treated by H2O2. L, lower than vector (pSa13)-transformed control; DW, dry weight; values are means ± SD (= 3); **< 0.01 by U test. Bars represent SD.


The aims of this study were to phenotypically characterize Arabidopsis plants transformed with wild-type and mutant BjHMGS1, including effects on expression of downstream isoprenoid genes and products, seed germination, as well as consequent stress responses. Northern blot analyses revealed that the expression of ectopic BjHMGS1 mRNA and endogenous Arabidopsis HMG1 (Figure 2b) and HMG2 (Figure S2a) is induced in all HMGS-OE lines. Co-ordinated expression between endogenous HMGR and HMGS has been previously reported in B. juncea (Alex et al., 2000) and in animals (Gil et al., 1986; cf. Goldstein and Brown, 1990). In this study, we present new evidence that the mRNAs encoding DWF1, CYP710A1, SMT2 and BR6OX2, located further downstream of HMGS in the MVA-dependent steroid pathway, are also up-regulated in Arabidopsis HMGS-OEs. The coordination between HMGS, HMG2 and BR6OX2 was also observed in wild-type Arabidopsis inflorescence (Figure S2b). Thus, in addition to HMGR (Bach, 1986), it is evident that HMGS plays a crucial role in sterol biosynthesis, as it is required for MVA synthesis. The importance of HMGS in plant pollen development has been recently verified in Arabidopsis (Ishiguro et al., 2010), indicating that HMGS has significant functions in the life cycle of the plant.

No significant changes were evident in plastidial DXR mRNA expression in all OEs. Correspondingly, chlorophyll and carotenoid contents remained unaltered in the sterol-accumulating Arabidopsis OEs (Figure S4), indicating that HMGS overexpression has little bearing on the MEP pathway, consistent with reports (Hemmerlin et al., 2003; Laule et al., 2003) that metabolic flow is mainly from the plastid located MEP pathway to the cytosolic MVA pathway. The expression of sesquiterpene synthase At5g23960 and the formation of its major product floral (–)-E-β-caryophyllene did not show significant changes in OE-S359A, further confirming a specific link between HMGS and sterol isoprenoid biosynthesis. SMT2 has been reported to control the ratio of campesterol to sitosterol content in Arabidopsis SMT2-OEs (Schaeffer et al., 2001). However, the ratio of campesterol to sitosterol in HMGS-OEs did not differ from wild-type or vector-transformed control (Figure S2c), although SMT2 mRNA increased in HMGS-OEs. Perhaps the slightly induced SMT2 mRNA was not sufficient to influence sterol composition in comparison with SMT2-OE Arabidopsis which was driven by the 35S promoter. Furthermore, HMGS is located in the second step of MVA isoprenoid biosynthesis pathway and its overproduction in HMGS-OEs most probably had resulted in a greater effect on the metabolic flux of the whole pathway. Nevertheless, SMTs located further downstream would need to depend on the steroid intermediates to influence sterol content. Hence, when SMT2 is overexpressed, sitosterol is overproduced at the expense of campesterol which shares a common precursor (24-methylene lophenol) with sitosterol biosynthesis and then the ratio of campesterol to sitosterol is significantly changed (Schaeffer et al., 2001).

Gas chromatography–mass spectrometry analyses revealed that the OE-S359A lines accumulated the highest amounts of total sterol, campesterol and sitosterol in both seedlings and rosette leaves, consistent with the highest activity of recombinant S359A mutant of HMGS in in vitro assays (Nagegowda et al., 2004). Correspondingly, OE-H188N leaves and seedlings accumulated the lowest amounts of total and individual sterols (with the exception of stigmasterol) among all OE lines. Loss of inhibition by superfluous AcAc-CoA was not apparent in the OE-H188N lines from analysis of sterol content, perhaps attributed to relatively limited level of endogenous AcAc-CoA, which is not inhibitory. Although the bacterial-expressed BjMHGS1-H1188N/S359A enzyme displayed both characteristics of the two single mutants, H188N and S359A, its corresponding Arabidopsis OE lines (OE-H188N/S359A) did not accumulate the highest contents of sterols and was even somewhat lower than the OE-S359A lines. It is likely that the H188N mutation masked the S359A phenotype of increased activity (10-fold) when expressed in vivo because of the limited affinity to AcAc-CoA, which is only present in submicromolar concentrations in the cytosol. Moreover, the enzyme preceding HMGS, AcAc-CoA thiolase favours the thiolytic cleavage of AcAc-CoA to Ac-CoA in the presence of HS-CoA (see Ahumada et al., 2008).

Interestingly, in the OE-H188N seedlings, stigmasterol significantly increased, whereas campesterol and sitosterol levels remained relatively unchanged. It is likely that more acetyl-CoA and AcAc-CoA units are available for MVA biosynthesis in seedlings than in leaves. During seed germination, AcAc-CoA is released from lipid mobilization and catabolism, while acetyl-CoA is further derived from seed protein and carbohydrate catabolism. Therefore, H188N may function more efficiently in seedlings than in leaves. Given that the enzyme activity of H188N is 8-fold lower than wild-type BjHMGS1 and 80-fold lower than the S359A mutant (Nagegowda et al., 2004), H188N-OEs may not have as much HMG-CoA for subsequent sterol biosynthesis in comparison with other HMGS mutants. Although H188N-OEs could not overaccumulate total sterols, sitosterol and campesterol, the stigmasterol level in H188N-OE seedlings was significantly greater than the controls (Tables 1 and S1), suggesting that stigmasterol (rather than sitosterol or campesterol) plays a very important role in the rapid early development of HMGS-OEs. This was further supported by observations in higher CYP710A1 expression in wild-type developing seedlings than in leaves (Figure S3). It should be noted that stigmasterol is an end product in the pathway. In comparison, sitosterol and campesterol are intermediates in stigmasterol and BL biosynthesis (Grove et al., 1979). When either CYP710A1 or CYP710A4 was overexpressed in Arabidopsis, Arnqvist et al. (2008) observed that stigmasterol increased at the expense of sitosterol. Given the induction of CYP710A1 in HMGS-OEs, OE-H188N seedlings may accumulate stigmasterol at the cost of sitosterol (Table S1). Although stigmasterol increased >2-fold in HMGS-OEs, it did not drastically alter total sterol content in Arabidopsis seedlings because only 3% of total seedling sterols consist of stigmasterol, while in leaves stigmasterol levels are even lower (1%). A previous report (Schaeffer et al., 2001) showed that stigmasterol constitutes only 0.2% of total sterol in Arabidopsis ground materials.

Several studies have demonstrated that phytosterols particularly β-sitosterol possess high antioxidant activity (Weng and Wang, 2000; Wang et al., 2002; Vivancos and Moreno, 2005; Li et al., 2007; Poséet al., 2009). Wegener et al. (1997) have demonstrated that Pinus sylvestris L. HMGS mRNA is induced by ozone. Such evidence suggests that HMGS and sterols are essential in protection against oxidative damage. Sitosterol was observed to account for about 80% of total sterols in Arabidopsis OE lines (Table S1). The phenotypes of H2O2-treated Arabidopsis OE leaves (Figure 5a) are consistent with the sitosterol content of leaves determined in this study (Tables 2 and S2). Both OE-H188N and vector-transformed leaves accumulated less sitosterol, in comparison with the other OE leaves, and were correspondingly more sensitive to H2O2. Thus, having an ability to accumulate sitosterol would be applicable for crops to combat stresses caused by H2O2 and to protect against ultraviolet-B (UVB) light, as the ozone layer declines as a consequence of atmospheric pollution.

Besides contribution to growth and development (Suzuki et al., 2004, 2009; Okada et al., 2008), the MVA pathway also plays a major role in plant–pathogen interaction. Arabidopsis HMG2, one of two genes encoding HMGR, is pathogen inducible because its root mRNA was co-activated with PR3 after Orobanche ramosa infection (Santos et al., 2003). Alex et al. (2000) showed that both BjHMGS and BjHMGR were highly induced following SA treatment in B. juncea leaves. In this study, we detected constitutive activation of HMG2 together with SA-dependent pathogenesis-related genes PR1, PR2 and PR5 mRNA in Arabidopsis HMGS-OE lines. The function of PR1 is yet unclear, but PR2 encoding β-1,3-glucanase can degrade the fungal cell wall and PR5 encoding a thaumatin-like protein permeabilizes fungal membranes (van Loon et al., 2006). Fungal resistance of HMGS-OEs could be attributed to the SA pathway as apparent BL induction would have had up-regulated PR-1 (Divi et al., 2010). In comparison, this study revealed that expression of the jasmonate (JA)-dependent pathogenesis-related gene PDF1.2 mRNA (van Loon et al., 2006) was not significantly different between vector-transformed control and OE lines (Figure 2b), implying that JA does not seem to play a role in the HMGS-mediated defence response in Arabidopsis. This is consistent with our previous report that BjHMGS1 mRNA is only slightly induced by MeJA in B. juncea, in comparison with greater induction by SA (Alex et al., 2000).

In conclusion, the overexpression of wild-type and mutant BjHMGS1 in Arabidopsis up-regulated the expression of HMGR, SMT2, DWF1, CYP710A1 and BR6OX2 involved in MVA-dependent steroid biosynthesis and enhanced seed germination, B. cinerea resistance and H2O2 tolerance. Further, the up-regulation of HMGR, SMT2, DWF1, CYP710A1 and BR6OX2 in HMGS-OEs tempts us to suggest that all these genes could be possibly co-regulated, which would be the subject of future investigations.

Materials and methods

Construction of plasmids overexpressing wild-type and mutant BjHMGS1 and generation of transgenic Arabidopsis

A 1.4-kb wild-type or mutant BjHMGS1 cDNA BamHI fragment was excised from plasmids pBj110 (H188N/S359A), pBj75 (wtBjHMGS1), pBj89 (S359A) or pBj76 (H188N) (Nagegowda et al., 2004) and cloned into binary vector pSa13 (Xiao et al., 2008) to generate plasmids pBj132, pBj134, pBj136 and pBj137, respectively. The resultant plasmids and pSa13 vector control were mobilized into Agrobacterium tumefaciens LBA4404 by triparental mating, after which A. thaliana Col-0 was transformed using the floral dip transformation method (Clough and Bent, 1998). T1 seeds from each transformant were selected on MS medium (Murashige and Skoog, 1962) for resistance to 50 mg/L kanamycin, and the resistant seedlings were transferred to soil. To preliminarily confirm Arabidopsis transformants, the 35S promoter forward primer (5′-CAATCCCACTATCCTTCGCAAGACC-3′) and a BjHMGS1 3′-end cDNA reverse primer ML264 (5′-GGATCCATAACCAATGGACACTGAGGATCC-3′) were used to amplify inserts (Figure S1a,b). Mutations on the BjHMGS1 cDNA were validated by DNA sequencing of PCR products amplified from total DNA of transgenic Arabidopsis using primer ML915 (5′-CATTGCTATGTTGATAGGAC-3′). Wild-type and transgenic Arabidopsis were grown under cycles of 8 h dark at 22 °C and 16 h light at 23 °C. T2–T6 transgenic plants were used for subsequent analyses of isoprenoid content, germination rates and stress tolerance.

Protein extraction and Western blot analysis

Total protein was extracted from 3-week-old Arabidopsis rosette leaves following Xu et al. (2002). Protein (20 μg per well) was separated on 12% sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS–PAGE) gel and transferred to Hybond-ECL membrane (Amersham, Piscataway, NJ). Western blot analysis was carried out according to ECL Western blotting instructions (Amersham) and Chen et al. (2010). Antibodies against HMGS were used in Western blot analysis. A synthetic peptide (DESYQSRDLEKVSQQ) corresponding to BjHMGS1 amino acids 290–304 was used for the immunization of rabbits.

Northern blot analysis

Total RNA was extracted from 3-week-old Arabidopsis rosette leaves using TRIzol reagent (Invitrogen, Carlsbad, CA) for Northern blot analysis. RNA (20 μg per well) was separated on 1.3% agarose gels containing 6% formaldehyde and transferred to a Hybond-N membrane. Northern blot analysis was carried out according to Chen et al. (2008) using cDNA probes encoding Arabidopsis HMGR (HMG1, At1g76490), DWF1(At3g19820), PR1 (At2g14610), PR2 (At3g57260) and PR5 (At1g75040), PDF1.2 (At5g44420) and sesquiterpene synthase (At2g37620). The gene-specific primers used in labelling probes were ML276 (5′-GGATCCATGGCGAAGAACGTAGGGATATTG-3′) and ML860 (5′-GGAGACTGGTTCTCGCAGAGAC-3′) for BjHMGS1; ML1038 (5′-TCTCTCCTTCACACTTATCACGC-3′) and ML1039 (5′-GTTGTCGTTGTCGTTGCTCCAG-3′) for HMG1; ML1084 (5′-ATGTCGGATCTTCAGACACCGC-3′) and ML1085 (5′-CCCAGTACAAGCACCTTGTGTG-3′) for DWF1; ML739 (5′-CTCTTTGTAGCTCTTGTAGGTGC-3′) and ML740 (5′-GGCTTCTCGTTCACATAATTCCC-3′) for PR1; ML755 (5′-TCTGAATCAAGGAGCTTAGC-3′) and ML756 (5′-CATACTTAGGCTGTCGATCT-3′) for PR2; ML757 (5′-ATCTCCAGTATTCACATTCT-3′) and ML758 (5′-CTTCGGTTTTTAAGGGCAGA-3′) for PR5; ML741 (5′-TAAGTTTGCTTCCATCATCACCC-3′) and ML742 (5′-TTAACATGGGACGTAACAGATAC-3′) for PDF1.2; and ML1100 (5′-AAGGACACTTGGTGACAGATGC-3′) and ML1101 (5′-TGCGACATCTGTCATTCTCACG-3′) for sesquiterpene synthase.


Those genes with low mRNA levels were detected using RT-PCR. Total RNA was extracted using TRIzol reagent (Invitrogen) from 14-day-old seedlings (except for Figure S2b) for RT-PCR analyses. RNA (2.5 μg) was treated by DNase (Promega, Madison, WI) and used for first-strand synthesis following the instructions of the SuperScript TM First-Strand Synthesis System for RT-PCR (Invitrogen, Carlsbad, CA). The RT product (2 μL) was used as a PCR template to amplify the 1.4-kb DXR (At5g62790), 1.0-kb BR6OX2 (At3g30180), 0.5-kb CYP710A1 (At2g34500), 234-bp ACTIN1 (At2g37620), 0.45-kb SMT1 (At5g13170), 0.5-kb SMT2 (At1g20330), 0.4-kb SMT3 (At1g76090) and 0.5-kb ACTIN8 (At1g49240) cDNAs in Arabidopsis pSa13-transformed and HMGS-OE lines under a temperature programme of 94 °C, 30 s; 54 °C, 30 s and 72 °C, 90 s; running for 30 cycles. Equal volumes of PCR products were loaded on 0.8% agarose in gel electrophoresis. Gels were scanned using PharosFX Plus Molecular Imager, and the expected bands were quantified using Quantity One software (Bio-Rad, Hercules, CA). The adjusted volume of each band for each gene was divided by that of the corresponding ACTIN1 (Figures 2c and S2d) or ACTIN8 (Figures S2a–b and S3) band to obtain the relative quantity. Primer pairs used were ML1042 (5′-AACTCACTATCTCCAGCTGAATC-3′) and ML1043 (5′-GAACTGGCCTAGCACCAGAAG-3′) for DXR; ML1086 (5′-TCTTCTCCGATGGAACCAGATGC-3′) and ML1087 (5′-ACCGTTGAGTTCTAAGTCGTGAG-3′) for BR6OX2; primer pairs 5′-CACCGCTTAACCCGAA-3′ and 5′-GTGAGGTCACGACCAG-3′ for ACTIN1; ML1175 (5′-ATGGCCGATGCTGATGACATTC-3′) and ML1176 (5′-GTAGATAGGCACAGTGTGAGAC-3′) for ACTIN8; ML1181 (5′-TGCAAGAAGAGCTGCTCTTGTC-3′) and ML1182 (5′-AGATGTGAAGGTCGTCACCATC-3′) for HMG2 and ML1258 (5′-GACAAGCAATCCTCCACGGC-3′) and ML1259 (5′-CTATGGGGAGAGCCATGGATC-3′) for CYP710A1; ML1564 (5′-CGTCGAATCTTGGTGGGAAGATC-3′) and ML1565 (5′-CACCTGCAAGTCGGTTTAGTTCC-3′) for SMT1; ML1566 (5′-ACACTCTTCTTCACCGGTGCAC-3′) and ML1567 (5′-CCTCGCAAAGCGCGTCGAGAC-3′) for SMT2; ML1568 (5′-CTCTCTACTGCACCGCTGGTC-3′) and ML1569 (5′-GGCTCGTTGCACTTGGTACTC-3′) for SMT3.

Quantitative analysis of sterols

Sterols were extracted and analysed according to Babiychuk et al. (2008) with some modifications. Fourteen-day-old freeze-dried Arabidopsis seedlings grown on MS medium or 3-week-old soil-grown rosette leaves (50 mg) were homogenized in 15 mL methanol containing 6% potassium hydroxide (KOH) and incubated in a 65 °C water bath for 2 h. Water (5 mL) was added, and then 20 mL hexane (HPLC grade, Fluka) was used twice to extract the neutral lipids (including sterols). The top hexane phases were pooled and evaporated to dryness. Acetylation was performed by adding toluene (100 μL), acetic anhydride (100 μL) and pyridine (50 μL). The mixture was incubated overnight at room temperature. The acetylation reactants were evaporated under hot drought, and lupenyl-3,28-diacetate (20 μg) was added as internal standard to the acetylated sterols. The mixture was then dissolved in hexane (500 μL), and 1 μL was used for GC-MS analyses (GC: Hewlett Packard 6890 with an HP-5MS capillary column: 30 m long, 0.25 mm i.d., film thickness 0.25 μm; MS: Hewlett Packard 5973 mass selective detector, 70 eV) to determine the sterol content, with He as the carrier gas (1 mL/min). The column temperature programme used included a fast rise from 60 to 220 °C (30 °C/min) and a slow rise from 220 to 300 °C (5 °C/min), then kept at 300 °C for 10 min. The inlet temperature was 280 °C. Compounds were identified using mass spectral libraries NIST (Agilent, Santa Clara, CA). The sterol masses were determined by comparing the peak area of each compound with that of the internal standard. Each sample was injected twice in GC-MS analyses, and an average of its sterol mass was taken. Sitosterol, campesterol and stigmasterol contents in Arabidopsis OE were compared with those of vector (pSa13)-transformed plants (Tables 1 and 2).

Quantitative analysis of floral sesquiterpene (–)-E-β-caryophyllene in HMGS-OEs

Volatiles were collected using solid phase micro extraction (SPME) (Tholl et al., 2005; Abel et al., 2009) from flowers of 5-week-old Arabidopsis transformed with vector pSa13 and representative constructs OE-wtBjHMGS1 and OE-S359A. The fibre coating was PDMS-30 for the manual holder (30 μm polydimethylsiloxane) (57308; Supelco, Bellefonte, PA). Twenty detached flowers were placed in a 1.5-mL glass vial sealed with a septum cap containing 100 μL water mixed with 0.01 μL (8.67 μg) anhydrous toluene (internal standard). Water with toluene minus flowers was set as a blank control. Samples were kept at 25 °C, and the fibres were exposed for 10 min. GC-MS system and temperature programme used in this assay were similar to those described by Abel et al. (2009). (–)-E-β-caryophyllene was identified according to the Agilent mass library and commercial standard β-caryophyllene (Aldrich, Hercules, CA). The retention time and peak pattern were similar to those obtained by Abel et al. (2009). The (–)-E-β-caryophyllene level (Figure S4a) was determined by comparing the peak area of the volatile compound to that of the internal standard.

Spectrometry quantification of chlorophyll and carotenoid contents

Fresh rosette leaves of 3-week-old Arabidopsis plants were weighed and transferred to eppendorf tubes. The material was ground on ice using a Teflon grinding bar, and chlorophyll and carotenoids were extracted in dim light using 90% methanol, following Laule et al. (2003). After centrifugation at 4 °C, absorbance was measured at 470, 653 and 666 nm. The amount of chlorophyll a, chlorophyll b and total carotenoids was calculated according to Lichtenthaler (1987).

Germination assay

Seeds collected simultaneously from wild-type, vector (pSa13)-transformed and HMGS-OE Arabidopsis lines were dried at 37 °C for 2 days and subsequently sterilized in 20% bleach and 70% ethanol. After five rinses in sterile water, seeds were spread onto MS medium agar plates. Plates were incubated at 4 °C for 4 days (Uchida and Yamamoto, 2002) and transferred to a growth room (22 °C 8-h dark/23 °C 16-h light). The emergence of the radicle was defined as germination. The number of germinated seeds was counted every 12 h within 3 days.

Fungal infection assays

Botrytis cinerea was incubated on Potato Dextrose Agar medium (BD, Lakewood, CO) for 3 weeks at 23 °C. Spores were washed off from the medium using 1% glucose, and the number of spores was calculated using a haemocytometer. Spores (3.5 × 105 mL−1, 5 μL per drop) were used for the droplet infection (Dhawan et al., 2009). Fungal infection assays were carried out at 23 °C in 100% humidity. The diameters (width) of fungal lesions were measured in the droplet infection assay after 3 days (Ferrari et al., 2003; Dhawan et al., 2009).

Hydrogen peroxide treatment assay

The detached rosettes leaves from 3-week-old plants (pSa13-transformed control and HMGS-OE lines) were soaked in water, 3 or 5 mm H2O2 for 72 h at room temperature under dim light. Ion leakage (IL) was determined and expressed as increased electrolyte conductivity according to Gechev et al. (2005). To estimate cell death induced by H2O2, H2O2-treated leaves were soaked in Milli Q water and shaken for 1 h before electrolyte conductivity (C) was measured using a conductivity meter (YSI model 55) (Chen et al., 2008). Total conductivity (Tc) was obtained after boiling the treated leaves. Ion leakage was calculated by the equation: IL = (C/Tc)%. Treated leaves were also stained with trypan blue (Heese et al., 2007) to directly indicate the dead cells.


This work was supported by the University of Hong Kong (Grants 10208034 and 10400945, and a Postgraduate Studentship to W.H.) and the France-Hong Kong Joint Research Scheme F-HKA04/00 (to TBJ and MLC). We thank Dr. Hubert Schaller (CNRS-IBMP, Strasbourg) for instructions on sterol extraction and quantitative analysis by the use of the instrumentation installed at the IBMP’s metabolomics platform and Prof. S. F. Chen (The University of Hong Kong) for providing the conductivity meter.