Co-expression of N-terminal truncated 3-hydroxy-3-methylglutaryl CoA reductase and C24-sterol methyltransferase type 1 in transgenic tobacco enhances carbon flux towards end-product sterols


For correspondence (fax +44 1234 248010; e-mail

Present address: Alligator Biosciences, IDEON Delta 5, Scheele vägen 19 A, SE-223 70 LUND, Sweden.


The enzymes 3-hydroxy-3-methylglutaryl CoA reductase (HMGR) and C24-sterol methyltransferase type 1 (SMT1) have been proposed to be key steps regulating carbon flux through the sterol biosynthesis pathway. To further examine this hypothesis, we co-expressed the catalytic domain of Hevea brasiliensis HMGR (tHMGR) and Nicotiana tabacum SMT1 in tobacco, under control of both constitutive and seed-specific promoters, resulting in increased accumulation of total sterol in seed tissue by 2.5- and 2.1-fold, respectively. This enhancement is greater than when tHMGR and SMT1 were expressed singularly where, for example, seed-specific expression enhanced total sterols by 1.6-fold. Significantly, the relative level of 4-desmethyl sterols (end-product sterols) was higher in seed co-expressing tHMGR and SMT1 from seed-specific promoters (79% of total sterols) than when co-expressed from constitutive promoters (59% of total sterols) and similar to wild-type seed (80% of total sterols). These results demonstrate that HMGR and SMT1 work in concert to control carbon flux into end-product sterols and that the sterol composition can be controlled by the temporal activity of the promoters driving transgene expression. In addition, constitutive expression of the transgenes resulted in elevated accumulation of substrates for C4-demethylation reactions, which indicates that one or several enzymes catalysing such reactions limit carbon flow to end-product sterols, at least in a physiological situation when the carbon flow is upregulated.


Sterols are a class of natural compounds that possess a tetracyclic ring system. They are ubiquitous among plants, serving to modify the fluidity of membranes, also acting as precursors in the biosynthesis of plant brassinosteroid hormones (Parish and Nes, 1997) and as minor components of seed oils. Higher plants contain a mixture of sterols (phytosterols) that can be classified into three groups based on methylation state at C4: 4-desmethyl sterols, 4-monomethyl sterols and 4,4-dimethyl sterols. 4-desmethyl sterols are the commonly occurring plant end-product sterols, such as campesterol, a C24-methyl sterol and, sitosterol and stigmasterol, C24-ethyl sterols. In addition, certain plants, such as tobacco, have relatively high levels of cholesterol, a C24-desmethyl sterol synthesised from cycloartenol via a distinct branch of the sterol biosynthesis pathway (Figure 1). 4,4-Dimethyl sterols include cycloartenol and 24-methylene cycloartenol, and examples of 4-monomethyl sterols are 24-methylene lophenol and 24-ethylidene lophenol.

Figure 1.

A schematic representation of the sterol biosynthetic pathway in plants.

Solid lines indicate a single enzymatic step, and dashed lines indicate several enzymatic steps.

The sterol biosynthesis pathway in plants is divided in two parts and comprises some 30 enzymatic steps (Bach and Benveniste, 1997). In the first part, the mevalonate pathway, acetate is converted into squalene epoxide via 3-hydroxy-3-methylglutaryl CoA (HMG CoA), mevalonic acid (MVA), isopentenyl pyrophosphate (IPP)/dimethylallyl pyrophosphate (DMA-PP), geranyl pyrophosphate (GPP), farnesyl pyrophosphate (FPP) and squalene (Figure 1). In the second part, squalene epoxide is cyclized to give cycloartenol, which is the first committed step in the sterol biosynthesis pathway. Cycloartenol is further converted into end-product sterols by removal of the C4-methyl groups, opening of the cyclopropane ring (C9–C10), introduction of a double bond at C5–C6 and modification of the C24 side chain (Figure 1).

Two enzymes have been proposed to play pivotal roles in regulating the carbon flux into end-product sterols, namely 3-hydroxy-3-methylglutaryl CoA reductase (HMGR) and C24-sterol methyltransferase type 1 (SMT1). HMGR converts HMG CoA into MVA, and a large body of evidence supports the theory that this enzyme is key in controlling carbon flux into the sterol biosynthesis pathway. A sterol-overproducing tobacco mutant (LAB1-4) was shown to have increased HMGR activity (Gondet et al., 1992). Moreover, a high-sterol mutant cell line of Solanum xanthocarpum, selected by mevinolin (an HMGR inhibitor) treatment, displayed elevated HMGR activity that correlated with enhanced sterol biosynthesis (Josekutty, 1998). In addition, overexpression of either truncated hamster HMGR or rubber tree HMGR in transgenic tobacco resulted in sterol overaccumulation (Chappell et al., 1995; Harker et al., 2003; Schaller et al., 1995).

HMGR is an integral membrane protein, located on the cytoplasmic side of the endoplasmic reticulum (ER). Because of its central role in sterol biosynthesis, much research has been performed to elucidate the mechanisms of its regulation. In mammalian and yeast systems, HMGR is controlled at many levels, such as transcription, translation, regulated proteolytic degradation and kinase-mediated phosphorylation (Gardner and Hampton, 1999; Gil et al., 1985). The non-catalytic amino-terminal membrane-spanning domain has been shown to be essential for regulation via proteolytic degradation (Gardner and Hampton, 1999; Gil et al., 1985). In plant systems, the activity of HMGR is under developmental regulation at the transcriptional and post-transcriptional level. In addition, HMGR activity is regulated by external factors such as light, pathogens and wounding (Bach, 1995; Chappell, 1995; Korth et al., 2000). Analogous with mammalian HMGR, the activity of plant HMGR is also regulated via proteolytic degradation. For instance, Korth et al. (2000) demonstrated that darkness induces proteolytic degradation of HMGR in potato. Nonetheless, it still remains uncertain if the membrane-spanning domain is essential for degradative-mediated regulation of HMGR in plants.

SMT1 catalyses the conversion of cycloartenol into 24-methylene cycloartenol, and this has been proposed to be the critical slow step in the biosynthesis of Δ5-sterols (Nes and Venkatramesh, 1999). Evidence to support this view is the massive accumulation of cycloartenol in HMGR overexpressing transgenic tobacco (Chappell et al., 1995; Harker et al., 2003; Schaller et al., 1995). It has also been shown that SMT activity is downregulated by sitosterol in sunflower, which suggests that SMT has a regulatory role in sterol biosynthesis (Janssen and Nes, 1992). In addition, it was recently demonstrated that overexpression of SMT1 in tobacco seed resulted in elevated endogenous HMGR activity and higher total sterol production (Holmberg et al., 2002).

To further investigate the roles of HMGR and SMT1 in seed-sterol biosynthesis, we co-expressed the catalytic domain of Hevea brasiliensis HMGR and Nicotiana tabacum SMT1 in tobacco under the control of either constitutive or seed-specific promoters. The truncated HMGR used, lacking the amino-terminal membrane-spanning domain, has recently been shown to significantly enhance sterol accumulation in tobacco seed (Harker et al., 2003). Here, we investigate the consequences of co-expressing HMGR and SMT1 in seed tissue of tobacco in terms of: (i) accumulation of total sterols; (ii) alteration of sterol composition; and (iii) identification of any additional slow enzymatic steps, downstream of SMT1, that limit carbon flux into 4-desmethyl sterols.


Analysis of transgenic plants that express HMGR and SMT1 from constitutive promoters

The truncated H. brasiliensis hmgr1 (thmgr1) gene and the N. tabacum sterol methyl transferase type 1 (Ntsmt1-1) genes were placed under control of the constitutive enhanced cauliflower mosaic virus (CaMV) 35S promoter and carnation-etched ring virus (CERV) promoter, respectively, to generate pMH7 (Figure 2). Vector pMH7 was transformed into tobacco via Agrobacterium-mediated gene transfer alongside control vectors pMH5 (2× 35S-thmgr1-NOS; NOS, nopaline synthase), pNH7 (CERV-Ntsmt1-1-NOS) and empty vector control pSJ34 (Figure 2). Rooted kanamycin- or hygromycin-resistant plants were selected and screened by PCR for the presence of the transgene(s). Seed tissue of 30 primary transgenic tobacco plants from each transformation was analysed for altered sterol content and composition. Plants with high sterol content were selected and taken onto the subsequent generation (T1). Given below are average values of total sterol content and sterol intermediates of at least five T1 plants.

Figure 2.

Schematic representation of the binary plant expression vectors pMH5, pMH15, pMH7 (kanamycin selection) and pNH7, pNH19 (hygromycin selection).

LB and RB correspond to the left and right borders, respectively. The direction of the arrows indicates the direction of the genes.

The average wild-type control contained 3677 µg sterol g−1 DW seed whilst the transgenic line MH7 contained, on average, 9296 µg sterol g−1 DW (Table 1), representing a 2.5-fold increase in total sterol level. Expression of either tHMGR or SMT1 individually gave rise to 2.2- or 1.2-fold increase over wild-type tobacco, respectively, which is significantly less than the increase obtained from co-expressing both genes (Table 1).

Table 1.  Sterol composition in seed tissue of transgenic T1 tobacco lines expressing tHMGR and/or SMT1 under control of constitutive promoters (µg sterol g−1 DW)
(n = 10)
(n = 5)
(n = 5)
(n = 5)
  • Average values of five individual plants of lines MH5 (2×35S-tHMGR1-NOS), NH7 (CERV-NtSMT1-NOS) and MH7 (CERV-NtSMT1-NOS + 2×35S-tHMGR1-NOS) are shown.

  • a

    Standard deviation.

Total3677 ± 416a7963 ± 11984361 ± 2819296 ± 629
Cycloartenol278 ± 741737 ± 592161 ± 22942 ± 103
24-Methylene cycloartenol27 ± 5510 ± 16929 ± 101466 ± 197
24-Methylene lophenol46 ± 20247 ± 5291 ± 5501 ± 90
24-Ethylidene lophenol310 ± 63561 ± 55459 ± 61715 ± 52
Δ7-Avenasterol34 ± 6191 ± 4455 ± 6214 ± 55
Isofucosterol594 ± 1101212 ± 131785 ± 691263 ± 60
Sitosterol1338 ± 912105 ± 1011734 ± 1632373 ± 106
Stigmasterol332 ± 39442 ± 71318 ± 15532 ± 73
Campesterol432 ± 35667 ± 67537 ± 71917 ± 79
Cholesterol225 ± 40293 ± 51194 ± 18217 ± 32

Constitutive co-expression of tHMGR and SMT1 (MH7) leads to dramatic changes in sterol composition compared to the wild type and significant changes with respect to expression of tHMGR alone (MH5). The absolute level of 4-desmethyl sterols (Δ7-avenasterol, isofucosterol, stigmasterol, sitosterol, campesterol and cholesterol) is substantially elevated (1.9-fold) in MH7 tobacco compared to wild-type tobacco but the relative level, as a percentage of total sterols, is reduced from 80 to 59% (Figure 3). The decreased relative level of 4-desmethyl sterols in MH7 lines is primarily because of substantially enhanced accumulation of sterol intermediates, in particular, 24-methylene cycloartenol (54-fold), and also 24-methylene lophenol (10.9-fold) and cycloartenol (3.4-fold; Table 1). Interestingly, cholesterol, which is synthesised from cycloartenol via a different branch of sterol biosynthesis to that of the other 4-desmethyl sterols (Figure 1), is not elevated in content.

Figure 3.

Absolute (grey bars) and relative (black bars) levels of 4-desmethyl sterols (Δ7-avenasterol, isofucosterol, stigmasterol, sitosterol, campesterol and cholesterol) in seed tissue co-expressing tHMGR and SMT1.

The MH7 line expresses both transgenes from constitutive promoters (CERV-NtSMT1-NOS + 2× 35S-tHMGR1-NOS). Line NH19:MH5 expresses the transgenes from one constitutive and one seed-specific promoter (tACP-NtSMT1-NOS + 2× 35S-tHMGR1-NOS), whilst NH19:MH15 expresses both the transgenes under control of the tACP seed-specific promoter (tACP-NtSMT1-NOS + tACP-tHMGR1-NOS). The data are derived from five individual plants of each line. The error bars represent the standard deviation.

In comparison to tobacco expressing solely tHMGR (MH5), constitutive expression of SMT1 and tHMGR (MH7) leads to a halving of the amount of seed cycloartenol with a concomitant increase in 24-methylene cycloartenol, the next intermediate in the pathway. 24-Methylene lophenol levels are also double compared to MH5, whilst the rest of the intermediate and end-product sterols are slightly elevated in the MH7 line as compared to the MH5 line (Table 1). Cholesterol levels are lower in MH7 than in MH5. The SMT1 overexpressing line NH7 has reduced levels of cycloartenol compared to both the wild type and the MH7 line (Table 1). Remaining intermediate and end-product sterols are slightly elevated in the NH7 line as compared to wild-type tobacco but substantially less than those in MH7 and MH5 tobacco.

Primary transgenic plants, selected on the basis of sterol content, were analysed for SMT1 and HMGR activity in developing seed tissue (14 days after anthesis). The transgenic line MH7:31 exhibited HMGR activity of 670 pmol h−1 mg−1 protein, which is approximately 15-fold higher than the average of the wild-type controls (Table 2). The SMT1 activity was significantly increased in all the transgenic lines analysed with the transgenic line MH7:53, exhibiting SMT1 activity of 66 pmol h−1 mg−1 protein, which is nearly a ninefold increase over the average wild-type activity (Table 2).

Table 2.  HMGR and SMT1 activities (pmol h−1 mg−1 protein) of seed tissue of two individual primary MH7 lines
  • a

    The average and the standard deviations of the SR1 wild types are calculated based on three individual plants.

SR146.3 ± 96a7.7 ± 4.51

Analysis of transgenic seed co-expressing HMGR and SMT1 under control of seed-specific or a combination of constitutive and seed-specific promoters

The consequences for sterol biosynthesis when the seed-specific truncated acyl carrier protein (tACP) promoter controlled the expression of tHMGR and SMT1 was compared to when tHMGR and SMT1 expression was under control of the constitutive CaMV 35S and tACP promoters, respectively. The aim was to study the influence of temporal expression pattern and strength of the promoters that regulate the transgene expression.

The thmgr1 gene was placed under control of the tACP promoter, resulting in binary vector pMH15 (Figure 2). A stable SMT1 overexpressing tobacco (line NH19) had previously been generated by transforming tobacco with vector pNH19, consisting of the Ntsmt1-1 gene under control of the seed-specific tACP promoter (Holmberg et al., 2002). Vectors pMH15 and pMH5, described in the previous section, were re-transformed into the NH19 background. Thirty primary transgenic plants from each transformation were selected as described in the previous section. Vector pMH15 was also transformed into wild-type tobacco as a control.

Seed-specific co-expression of tHMGR and SMT1 (NH19:MH15) gave rise to a 2.1-fold increase in total sterol level compared to the wild type (Table 3) and a 1.3-fold increase over tHMGR (MH15) or SMT1 (NH19) expressed individually under control of the seed-specific tACP promoter (Table 3). Using a combination of constitutive and seed-specific promoters to regulate transgene expression (NH19:MH5) resulted in similar enhancement of sterol content as was achieved with seed-specific promoters (Table 3).

Table 3.  Sterol composition in seed tissue of primary transgenic tobacco lines expressing tHMGR and/or SMT1 under control of seed-specific promoters (µg g−1 DW)
(n = 10)
(n = 5)
(n = 5)
(n = 5)
(n = 5)
  • Average values of five individual plants of lines MH15 (tACP-tHMGR1-NOS), NH19 (tACP-NtSMT1-NOS), NH19:MH15 (NH19 background re-transformed with MH15), and NH19:MH5 (NH19 background re-transformed with MH5 (2×35S-tHMGR1-NOS)) are shown.

  • a

    Standard deviation.

Total3677 ± 416a5761 ± 4855884 ± 1687785 ± 4697355 ± 203
Cycloartenol278 ± 74515 ± 69269 ± 6402 ± 781293 ± 193
24-Methylene cycloartenol27 ± 541 ± 1245 ± 571 ± 12393 ± 51
24-Methylene lophenol46 ± 2090 ± 18156 ± 6183 ± 38242 ± 26
24-Ethylidene lophenol310 ± 63608 ± 85640 ± 24966 ± 99573 ± 46
Δ7-Avenasterol34 ± 662 ± 2875 ± 939 ± 11178 ± 29
Isofucosterol594 ± 1101042 ± 1061138 ± 301538 ± 1381318 ± 57
Sitosterol1338 ± 912218 ± 1652226 ± 803094 ± 1652038 ± 104
Stigmasterol332 ± 39347 ± 24370 ± 18381 ± 35401 ± 46
Campesterol432 ± 35557 ± 55813 ± 28941 ± 103706 ± 59
Cholesterol225 ± 40282 ± 16154 ± 4170 ± 18214 ± 14

However, despite NH19:MH15 and NH19:MH5 plants having similar seed-sterol contents, they exhibited striking differences in terms of seed-sterol composition. In general terms, there were substantially lower levels of sterol intermediates and concomitantly higher levels of end-product sterols when expressing the transgenes via seed-specific promoters. Thus 4-desmethyl sterols accumulated to a greater extent, both in absolute terms (1.3-fold), and as a percentage of total sterols (79.2% versus 66.0%) with seed-specific expression of the transgenes compared to a combination of seed-specific and constitutive expression (Figure 3). Of particular note, seed-specific expression of transgenes resulted in markedly lower levels of the intermediates cycloartenol (0.3-fold) and 24-methylene cycloartenol (0.2-fold) and substantially higher amounts of sitosterol (1.5-fold) than that obtained with the combination of promoters (Table 3).

Genetic analysis of the transgenic plants

Selected high-sterol-accumulating MH7, NH19:MH15, NH19:MH5 and MH15 transgenic lines were analysed for the number of transgene insertion loci using Southern blot analysis. This analysis did not produce a clear correlation between single insertion loci and high sterol accumulation (Tables 1, 3 and 4). High-sterol-accumulating plants NH19:MH5:41 and NH19:MH15:39 have two and three transgene insertion loci, respectively (Table 4). It has been suggested that high copy number can be associated with co-suppression and gene silencing (Hobbs et al., 1990), but this does not appear to be the case in these re-transformed lines. Nevertheless, high sterol accumulation in plants MH7:53 and MH15:7 correlates well with insertion of the transgene into a single locus as these plants were among the top sterol-accumulating plants (Table 4, data not shown).

Table 4.  Number of transgene insertion loci in selected transgenic tobacco lines as determined by Southern blot analysis
LineCopy numbera
  • a

    The copy number was determined using a probe directed against the NPTII kanamycin resistance marker.



An interplay between HMGR and SMT1 controls the amount of carbon that is directed into production of end-product sterols

Previous studies in tobacco have shown that overexpression of HMGR leads to not only increases in end-product sterols, but also substantial accumulation of the sterol intermediate cycloartenol (Chappell et al., 1995; Harker et al., 2003; Schaller et al., 1995). These findings suggested that the next step in the pathway, namely, the conversion of cycloartenol to 24-methylene cycloartenol, catalysed by SMT1 was a critical slow step in the biosynthesis of end-product sterols (Chappell et al., 1995; Schaller et al., 1995). We have shown here that, compared to tHMGR expression on its own, co-expression of tHMGR and SMT1 does reduce the amount of cycloartenol synthesised (Tables 1 and 3) and, furthermore, significantly elevates levels of total sterol (Tables 1 and 3). One can thus conclude that both HMGR and SMT1 contribute to regulate the flux of carbon to end-product sterols. HMGR controls the carbon flux through the mevalonate pathway leading to cycloartenol, whilst SMT1 converts cycloartenol into 24-methylene cycloartenol and hence into downstream sterols. It should be noted, however, that expression of SMT1 in combination of HMGR still results in accumulation of cycloartenol, especially in the MH7 lines (Table 1), where constitutive promoters regulate the two transgenes. This might suggest that, under these conditions, SMT1 activity is limiting in terms of being able to convert the additional carbon flux, generated from enhanced HMGR activity, into downstream sterols. Support for this idea is provided by the enzyme data on MH7-seed tissue, which showed that the measured increases in SMT1 activity were significantly lower than those determined for HMGR (ninefold and 15-fold, respectively). Pertinently, we have previously shown (Holmberg et al., 2002) that overexpression of SMT1 activity leads indirectly to an increase in HMGR activity, and it was hypothesised that reduced levels of cycloartenol, arising from increased SMT1 activity, feed back to upregulate HMGR activity to maintain carbon flux into the pathway. It is interesting to consider whether such a mechanism is in operation in this situation as well. Other studies, with sunflower (Janssen and Nes, 1992), have shown that sitosterol is able to downregulate the activity of SMT1 and such end-product inhibition could of course be occurring in tobacco seed here. Another factor to consider in a situation where flux into the sterol biosynthesis pathway has been enhanced by overexpression of HMGR is the ‘availability’ of the sterol intermediates, such as cycloartenol, for conversion into end-product sterols. We have previously shown that the ‘additional’ sterols, including intermediates synthesised from overexpression of tHMGR in tobacco (Harker et al., 2003) accumulate in the form of sterol acyl esters. This raises interesting questions as to the timing of this esterification process and whether the esterified intermediates, in particular cycloartenol, remain substrates for subsequent enzymatic conversion or whether they are ‘sequestered’ and are no longer available for conversion into end-product sterols. In summary, constitutive co-expression of tHMGR and SMT1 clearly reduces cycloartenol levels and increases synthesis of end-product sterols compared to expression of HMGR alone, but other intermediates such as 24-methylene cycloartenol, 24-methylene lophenol and 24-ethylidene lophenol are seen to accumulate (see below for C4-demethylation substrates).

Temporal regulation of promoters driving the transgenes has a major impact on sterol profile

Relatively minor differences in accumulation of total sterol were seen between lines co-expressing the transgenes using constitutive promoters (MH7) and lines co-expressing the transgenes using seed-specific promoters (NH19:MH15; 9296 ± 629 versus 7785 ± 469 µg sterol g−1 DW, respectively). However, substantial differences in sterol composition were observed, with line NH19:MH15 clearly accumulating 4-desmethyl sterols more than line MH7 (Figure 3), and the relative accumulation of 4-desmethyl sterols as a percentage of total sterols being 79 and 59% for the NH19:MH15 and MH7 lines, respectively (Figure 3). These differences can be attributed to higher accumulation of intermediates cycloartenol, 24-methylene cycloartenol and 24-methylene lophenol in the MH7 line.

These observed differences in accumulation of 4-desmethyl and intermediate sterols may be explained by differences in the temporal expression pattern or strength of the constitutive and seed-specific promoters. We have previously determined, using promoter–reporter gene fusions, that the tACP promoter is activated slightly earlier during seed development than the CERV and enhanced CaMV 35S promoters are (de Silva, 1990; Holmberg et al., unpublished results). It is crucial that the overexpressed transgenes are present at the time during seed development when sterol biosynthesis occurs. Thus, one can speculate that the temporal expression pattern of the tACP promoter is synchronised with the expression of native sterol biosynthesis enzymes, resulting in lower accumulation of sterol intermediates. Conversely, the temporal expression pattern of the stronger constitutive promoters is probably not completely synchronised with the expression of the native sterol biosynthesis enzymes, which leads to higher accumulation of sterol intermediates and lower production of 4-desmethyl sterols (Figure 3).

It was also interesting to observe the situation where SMT1 expression was under control of the tACP promoter, and tHMGR expression was under control of the enhanced CaMV 35S promoter (NH19:MH5). The SMT1 activity seemed to have less of an effect on the accumulated cycloartenol pool, and as a consequence, less carbon was directed towards 4-desmethyl sterols (67%) than when both genes were expressed via seed-specific promoters (79%; Figure 3). There are two possible explanations for this observation. First, there is an imbalance between the strongly enhanced CaMV 35S promoter, which drives tHMGR expression, and the relatively weaker tACP promoter, which drives the SMT1 expression. In this case, the SMT1 activity would not be sufficient to cope with the increased amount of carbon flux generated by the relatively higher HMGR activity. Second, it is also possible that there are differences in the temporal expression pattern of the constitutive and seed-specific promoters, so that the SMT1 activity is present before the HMGR activity. As a consequence, the cycloartenol accumulation would thereby occur after the onset of enhanced SMT1 activity.

Enzymes catalysing C4-demethylation reactions may limit the carbon flux to end-product sterols

A consequence of co-expressing HMGR and SMT1, especially when controlled by constitutive promoters (MH7), was that intermediates downstream of cycloartenol, in particular 24-methylene cycloartenol, and to a lesser extent, 24-methylene lophenol and 24-ethylidene lophenol accumulated (Figure 4). All of these intermediates are substrates for C4-demethylation reactions. When overexpression of the transgenes was driven by seed-specific promoters (NH19:MH15), substrates for C4-demethylation reactions (24-methylene cycloartenol, 24-methylene lophenol and 24-ethylidene lophenol) accumulated to a lesser extent than when constitutive promoters or a combination of constitutive and seed-specific promoters were used (Figure 4). In fact, with seed-specific promoters, the relative proportions of 4-desmethyl sterols and C24-demethylation substrates were similar to those of the wild type (Figure 4). Compared to seed-specific expression, expression of HMGR and SMT1 with a combination of promoters (NH19:MH5) resulted in a relatively larger accumulation of cycloartenol, 24-methylene cycloartenol and 24-methylene lophenol, and a reduced relative accumulation of 4-desmethyl sterols (Figure 4). A plausible explanation for these observations is that the seed-specific promoter is more compatible, in terms of temporal expression pattern, with the endogenous sterol biosynthesis enzymes than the constitutive promoter used.

Figure 4.

Relative levels of sterol intermediates cycloartenol (CA), 24-methylene cycloartenol (24-MCA), 24-methylene lophenol (24-ML) and 24-ethylidene lophenol (24-EL) in seeds from primary transgenic plants co-expressing HMGR and SMT1. The relative levels of 4-desmethyl sterols (4-Des) are also shown. The MH7 line expresses both transgenes from constitutive promoters (CERV-Ntsmt1-NOS + 2× 35S-thmgr1-NOS). Line NH19:MH5 expresses the transgenes from one constitutive and one seed-specific promoter (tACP-Ntsmt1-NOS + 2× 35S-thmgr1-NOS), whilst NH19:MH15 expresses both the transgenes under control of the tACP seed-specific promoter (tACP-Ntsmt1-NOS + tACP-thmgr1-NOS). The data is derived from five individual plants of each line. The error bars represent the standard deviation.

Studies of maize microsomes have shown that conversion of 24-methylene cycloartenol to cycloeucalenol, i.e. the removal of the α-methyl group at C4 position on the sterol backbone, is performed by a demethylation enzyme system comprising of three enzymatic steps (Pascal et al., 1993; Rondet et al., 1999). In addition to the maximal catalytic velocity (Vmax) of two C4-oxidising enzymes, the first steps of the C4-demethylation systems have been shown to have lower catalytic velocity than most other enzymes in plant sterol biosynthesis (Pascal et al., 1993). One might hence speculate that critical slow enzymes catalyse the demethylation reactions in tobacco seed, at least in the physiological situation where the carbon flux is upregulated via HMGR overexpression. However, one must bear in mind the possibility that the activities of these enzymes might be suppressed in the transgenic lines. It is therefore probably necessary to measure their activities to determine the true nature of these enzymes, if their catalytic efficiencies limit the carbon flow or if they are under some form of regulation. We are currently investigating the consequences of overexpressing certain enzymes involved in the C4-demethylation process on the accumulation of sterol intermediates in transgenic plants.

Nutritionally enhanced seed oils

The present study has provided new insights into the regulation of sterol biosynthesis that could be exploited to generate nutritionally enhanced oil seed crops. Dietary ingestion of phytosterols has been shown to lower blood cholesterol levels, and it is known that the 4-desmethyl sterol class of phytosterols is considerably more efficacious in this property than either the 4,4-dimethyl sterol or 4-monomethyl sterol classes (Westrate and Meijer, 1998). This study has shown that co-expression of tHMGR and SMT1 increases the accumulation of total seed sterols compared to that obtained with expression of either gene individually. Importantly, it has also demonstrated that seed-specific co-expression of tHMGR and SMT1 enhances the proportion of 4-desmethyl sterols synthesised compared to that obtained with constitutive co-expression of the transgenes. This work therefore represents a proof of principle demonstration for enhancing the sterol content of seeds whilst maintaining wild-type sterol composition. In summary, these findings provide potential technology to generate oil crops substantially enriched in cholesterol-lowering phytosterols.

Experimental procedures

Strains, plasmids, media and culture conditions

Escherichia coli strain DH5α (Invitrogen, Carlsbad, CA, USA) was used as the host strain in all cloning procedures. Bacteria were cultivated in Luria-Bertani (LB) medium (10 g l−1 tryptone, 5 g l−1 yeast extract and 5 g l−1 NaCl) supplemented with ampicillin (100 µg ml−1) or kanamycin (50 µg ml−1) on a rotary shaker (210 r.p.m.) at 37°C. The construction of binary vectors pMH5 (enhanced CaMV 35S-H. brasiliensis thmgr1-NOS), pNH7 (CERV-Ntsmt1-1-NOS) and pNH19 (truncated Brassica napus ACP-Ntsmt1-1-NOS) have been described elsewhere (Harker et al., 2003; Holmberg et al., 2002). Furthermore, the construction of cloning vectors pNH3, pNH9 and pNH12 has previously been described by Holmberg et al. (2002). Klenow fill in of the BamHI restriction sites of pGPTV-Kan yielded pSJ34 (Becker et al., 1992).

Plant material

Tobacco (N. tabacum) SR1 cv. Petite Havana and NH19 cv. Petite Havana (Holmberg et al., 2002) were grown in either MS-medium (Murashig and Skoog, 1962) or a compost:perlite mixture (2 : 1 (v/v)). The temperature in the growth rooms was kept at 22°C, and a day/night cycle of 16/8 h was used. The light intensity was 40 µmol m−2 sec−1.

Construction of binary expression vectors

A truncated form of hmgr was cloned by PCR, using the H. brasiliensis hmgr1 as the template as described by Harker et al. (2003). The thmgr1 gene was inserted into pNH4 between the NcoI and NheI sites of the polylinker, which lie between the enhanced CaMV 35S promoter and the NOS terminator, giving pMH3. This chimaeric gene was isolated by digestion with XmaI and SalI, purified and cloned into the corresponding polylinker sites of pNH9 giving pMH7. The binary vector pNH9 contains the Ntsmt1-1 gene cloned from N. tabacum, which is under transcriptional control of the CERV viral promoter. Vectors pMH3 and pMH7 were sequenced to ensure that the thmgr1 gene had been inserted correctly, and there were no mistakes in the promoter-initiation and terminator sequences.

The thmgr1 gene was also cloned into the polylinker region of pNH12 in the NcoI and NheI restriction sites, which lie between the B. napus tACP (0.3 kbp) promoter (de Silva et al., 1990) and the NOS terminator to give construct pMH11. The chimaeric gene was cloned into the binary vector pSJ34 after digestion and purification with XmaI and EcoRI and named pMH15. The binary vector pMH15 was sequenced to ensure that the thmgr1 gene had been inserted correctly, and there were no errors in the promoter-initiation and terminator sequences.

Plant transformation and growth conditions

Binary vectors pMH5, pMH7, pMH15 and pSJ34 were transformed into Agrobacterium tumefaciens LBA4404 using electroporation as described by Shen and Forde (1989). Agrobacterium-mediated transformation of N. tabacum SR1 was carried out using the leaf-disc method as described previously by Horsch et al. (1985). Vectors pMH5 and pMH15 were also re-transformed into transgenic line NH19:27, which has previously been described by Holmberg et al. (2002). For this, leaf discs from T2 NH19 plants were transformed as described above. Kanamycin (25 mg l−1) or hygromycin (50 mg l−1) resistant seedlings were screened by PCR to identify transformants (Edwards et al., 1991). PCR-positive transformants were transferred from tissue culture to growth chambers and potted into a compost:perlite mix (2 : 1 (v/v)). Mature seeds were collected after about 12 weeks.

Southern blot analysis

Genomic DNA was isolated from leaf tissue of 4-week-old soil-grown T0 tobacco plants using the cetyl-trimethyl-ammonium bromide (CTAB) method as described by Rogers and Bendich (1985). Approximately 10 mg genomic DNA was digested with EcoRI and separated on a 0.7% (w/v) agarose gel. The DNA fragments were transferred onto a Hybond™ (Amersham Pharmacia Biotech Ltd, UK) nylon membrane using capillary transfer as described by Sambrook et al. (1989). Southern blot analysis was performed using a DIG-labelled probe directed against the hygromycin/kanamycin resistance marker gene, according to the method described in the Roche Users Guide (Roche Molecular Biochemicals, Mannheim, Germany).

HMGR assay

The HMGR activity in homogenates of developing seeds of primary transformants (14 days after anthesis) was analysed as described by Harker et al. (2003).

SMT1 assay

Homogenates were prepared from developing seeds of primary transformants (14 days after anthesis) and analysed for SMT1 activity as earlier described by Holmberg et al. (2002).

Sterol analysis

Mature seeds of primary and T1 plants were collected and freeze-dried. The sterol composition and content was analysed as described before by Holmberg et al. (2002).


We thank Ann Scarborough, Rachel Payne and Bernadette Marsh for their skilful sequencing assistance.