Plants produce short-chain aldehydes and hydroxy fatty acids, which are important industrial materials, through the lipoxygenase pathway. Based on the information that lipoxygenase activity is up-regulated in tobacco leaves upon infection with tobacco mosaic virus (TMV), we introduced a melon hydroperoxide lyase (CmHPL) gene, a tomato peroxygenase (SlPXG) gene and a potato epoxide hydrolase (StEH) into tobacco leaves using a TMV-based viral vector system to afford aldehyde and hydroxy fatty acid production. Ten days after infiltration, tobacco leaves infiltrated with CmHPL displayed high enzyme activities of 9-LOX and 9-HPL, which could efficiently transform linoleic acid into C9 aldehydes. Protein extracts prepared from 1 g of CmHPL-infiltrated tobacco leaves (fresh weight) in combination with protein extracts prepared from 1 g of control vector-infiltrated tobacco leaves (as an additional 9-LOX source) produced 758 ± 75 μg total C9 aldehydes in 30 min. The yield of C9 aldehydes from linoleic acid was 60%. Besides, leaves infiltrated with SlPXG and StEH showed considerable enzyme activities of 9-LOX/PXG and 9-LOX/EH, respectively, enabling the production of 9,12,13-trihydroxy-10(E)-octadecenoic acid from linoleic acid. Protein extracts prepared from 1 g of SlPXG-infiltrated tobacco leaves (fresh weight) in combination with protein extracts prepared from 1 g of StEH-infiltrated tobacco leaves produced 1738 ± 27 μg total 9,12,13-trihydroxy-10(E)-octadecenoic acid isomers in 30 min. The yield of trihydroxyoctadecenoic acids from linoleic acid was 58%. C9 aldehydes and trihydroxy fatty acids could likely be produced on a larger scale using this expression system with many advantages including easy handling, time-saving and low production cost.
The lipoxygenase (LOX) enzymatic pathway has been intensively studied in the plant kingdom. LOX catalyses the oxidation of unsaturated C18 fatty acids into either 9- or 13-hydroperoxy fatty acids. These hydroperoxides are highly reactive molecules and are further metabolized by enzymes like hydroperoxide lyase (HPL), allene oxide synthase, divinyl ether synthase, epoxy alcohol synthase and peroxygenase (PXG) into a series of compounds endowed with various biological activities (Gardner, 1991; Grechkin, 1998). For example, HPL cleaves the C–C bond adjacent to the hydroperoxy group, resulting in the formation of ω-oxo acids and volatile aldehydes (C6 or C9). PXG catalyses an intramolecular transfer of oxygen from hydroperoxides to cis double bonds yielding epoxy alcohols. Subsequently, the epoxy alcohols are converted into trihydroxyoctadecenoates by epoxide hydrolase (EH). Trihydroxy derivatives of linoleic and linolenic acids have been reported as plant self-defence substances, called phytoalexins (Blée, 1998; Hamberg, 1999; Weber, 2002).
Volatile C6 and C9 aldehydes formed by LOX and HPL are important components of the aroma and flavour of fruits, vegetables and green leaves. These short-chain aldehydes and alcohols are produced by higher plants in response to wounding and pathogen attack (Bate and Rothstein, 1998; Croft et al., 1993). The sequence starts with the oxygenation of linoleic acid and linolenic acid by 9- or 13-LOX to form 9- or 13-hydroperoxy-octadecadienoic/octadecatrienoic acids (HPOD/T), respectively. The 13-hydroperoxy fatty acids are subsequently cleaved by 13-HPL into 12-oxo-9(Z)-dodecenoic acid and hexanal or 3(Z)-hexenal, whereas the 9-hydroperoxy fatty acids are broken down by 9-HPL into 9-oxononanoic acid and 3(Z)-nonenal or 3(Z),6(Z)-nonadienal (Matsui, 2006; Matsui et al., 2001). The 3(Z)-aldehydes isomerize either spontaneously or enzymatically catalysed to their 2(E)-enal isomers and can be reduced to their corresponding alcohols by alcohol dehydrogenase. Because of their pleasant odour, C6 and C9 aldehydes are commonly used as flavours to confer a fresh green odour of grass and vegetable to food products, respectively (Gigot et al., 2010; Kuroda et al., 2003; Lam and Proctor, 2002; Palma-Harris et al., 2002; Schieberle et al., 1990; Whitehead et al., 1995). These compounds can be extracted from plants. However, extraction is very expensive because of the low abundance of the short-chain aldehydes and alcohols in plants and cannot meet the increasing market demand. The natural green notes market is estimated at 5–10 ton/year and US$ 3000/kg (Muller et al., 1995a). Although most of them can be synthesized in a chemical way, this is not favoured because consumers have a strong preference for natural food additives (Dubal et al., 2008; Schrader et al., 2004). Thus, development of a biocatalytic process is required to produce these compounds on a large scale.
Hydroxy fatty acids are important industrial materials because the hydroxyl group gives fatty acids special properties such as higher viscosity and reactivity compared with their nonhydroxylated derivatives (Bagby and Carlson, 1989). Hydroxy fatty acids are used in a wide range of industrial products including resins, waxes, nylons, plastics, lubricants, cosmetics and additives in coatings and paintings (Naughton, 1974; Odian, 2004). Hydroxy fatty acids are also encountered in nature as cyclic esters known as lactones, which are used in perfumes and as flavour components in food (Hayes, 1996). The most important lactone for flavour application with a market volume of several hundred tons per year is γ-decalactone, which is transformed from ricinoleic acid (12-hydroxyoctadec-9-enoic acid; Gopinath et al., 2008; Schrader et al., 2004). Besides, trihydroxy fatty acids, 9(S),12(S),13(S)-trihydroxy-10-octadecenoic acid and 11,12,13-trihydroxy-9(Z),15(Z)-octadecenoic acid, with unknown stereochemistry have been reported to exhibit antifungal activity (Kato et al., 1985, 1986; Masui et al., 1989). Although the biological significance of these hydroxy fatty acids was reported, their production in nature is restricted mainly to plant systems, and in trace amounts. Therefore, much effort has been focused on the microbial production of hydroxy fatty acids from various substrates (Hou, 1994; Hou and Bagby, 1991; Kim et al., 2000; Kuo et al., 2001).
During the past several years, the use of enzymes or whole cells to produce aroma compounds from natural substrates on a large scale has become popular. Bringing a naturally rich source of substrate in contact with highly active enzymes can result in the production of flavour compounds in mass fractions of the order of several g/kg, instead of mg/kg encountered in raw materials (Aguedo et al., 2004; Schrader et al., 2004). Moreover, lipids are an important source of aroma compounds. Here, we report on the production of flavours and polymer precursors using LOX/HPL and LOX/PXG/EH routes for the generation of C9 aldehydes and trihydroxy fatty acids from linoleic acid, respectively (Figure 1). To this end, a 9-HPL gene from melon (CmHPL), a PXG gene from tomato (SlPXG) and a EH gene from potato (StEH) were overexpressed in Nicotiana benthamiana using a viral vector system (Marillonnet et al., 2004, 2005). High expression levels of CmHPL, SlPXG and StEH were achieved in tobacco leaves yielding considerable HPL, PXG and EH enzymatic activities. Owing to the concomitant induction of 9-LOX gene expression in tobacco by the viral vectors, CmHPL-infiltrated tobacco leaf extracts contained significant activities of 9-LOX and 9-HPL that can transform linoleic acid to C9 aldehydes, while the mixture of SlPXG-infiltrated and StEH-infiltrated tobacco leaf extracts contained high activities of 9-LOX, PXG and EH that can produce trihydroxy fatty acids from linoleic acid.
Overexpression of CmHPL, SlPXG and StEH in Nicotiana benthamiana
Previously, we reported that 9-LOX enzyme activity was highly induced in N. benthamiana by agroinfiltration using a tobacco mosaic virus (TMV)-based vector system (Huang and Schwab, 2011; Huang et al., 2010). Taking advantage of this effect, we introduced the genes encoding enzymes downstream in the LOX pathway, namely a HPL gene from melon (CmHPL, accession number AF081955), a PXG gene from tomato (SlPXG, accession number SGN-U567002) and a EH gene from potato (StEH, accession number AAA81892) into N. benthamiana leaves using a viral vector system (Marillonnet et al., 2004, 2005).
To determine the optimal harvest time, the tobacco leaves infiltrated with CmHPL, SlPXG or StEH were analysed at 4, 7 and 10 days after infiltration. The leaf infiltrated with GFP was used as a control. The SDS-PAGE gel analysis (Figure 2) showed that the GFP protein (27 kDa) has already been detected 3 days after infiltration, and the highest amount was observed 10 days after infiltration. Because of the similarity of the protein molecular weight of Rubisco large subunit (about 55 kDa) and CmHPL (54.4 kDa), it was not clear whether the CmHPL protein was expressed. However, HPL enzyme activity assays revealed that CmHPL was highly expressed. SlPXG (27.8 kDa) and StEH proteins (37.3 kDa) could be clearly detected on SDS-PAGE gels in the leaf preparations harvested seven and 10 days after infiltration (Figure 2b,c).
Production of C9 aldehydes by CmHPL-infiltrated tobacco preparations
The expression levels of Nb-9-LOX and CmHPL were examined in N. benthamiana leaves infiltrated with CmHPL by real-time PCR using the 18S-26S interspacer gene as an internal control for normalization (Figure 3a,b). As described in a previous study (Huang and Schwab, 2011), Nb-9-LOX gene expression was highly induced in tobacco leaves after treating with viral vectors. The leaf harvested 10 days after infiltration displayed the highest level of Nb-9-LOX transcripts (Figure 3a). CmHPL transcripts were detected in all leaves infiltrated with the CmHPL gene, and the highest level was found in the leaf harvested 7 days after infiltration (Figure 3b). HPL activity was also analysed in untreated leaf and in leaves infiltrated with GFP and CmHPL (Figure 3c). A low level of HPL enzyme activity was detected in leaf harvested 10 days after infiltration with GFP and in leaf harvested 4 days after infiltration with CmHPL. The highest level of HPL activity was determined in the leaf harvested 7 days after infiltration with CmHPL. In general, the results of the enzyme assays coincided with that of real-time PCR analysis.
To evaluate the short-chain aldehyde-forming activity of different leaf extracts, linoleic acid was incubated with the crude enzyme solutions prepared from untreated leaf (UT), GFP-infiltrated (GFP), as well as CmHPL-infiltrated leaves and harvested at 4, 7 and 10 days after infiltration (Figure 3d). Large amounts of 3(Z)-nonenal (Figure 4a, peak 1), 2(Z)-nonenal (Figure 4a, peak 2) and 2(E)-nonenal (Figure 4a, peak 3) were detected by SPME-GC–MS when linoleic acid was incubated with the crude enzyme solutions prepared from CmHPL-infiltrated leaves. Low levels of 3(Z)-nonenal, 2(Z)-nonenal and 2(E)-nonenal were also detected in the mixture containing GFP-infiltrated leaf extract (Figure 4a, GFP). These could be formed by the internal tobacco HPL, which was up-regulated through agroinfiltration. The amount of total C9 aldehydes produced by GFP sample was about 1/53 of that produced by Cm-10 d. In contrast, C9 aldehydes were absent in the reaction containing untreated leaf extract (Figure 4a, UT). The highest amount of nonenal was produced in the reaction containing the leaf extract harvested 10 days after infiltration with CmHPL (Figure 3d, row 10 d). Although sample 10 d did not display the highest HPL activity, it showed the highest expression level of Nb-9-LOX (Figure 3a). Besides, LC-MS analysis of the different reaction mixtures prepared from leaves infiltrated with CmHPL confirmed that 9-HPOD was completely consumed by HPL (data not shown). It appears that LOX is the rate-limiting enzyme in the leaf preparations. Protein extracts prepared from 1 g of CmHPL-infiltrated tobacco leaf (Cm-10 d, fresh weight) produced 303 ± 30 μg total C9 aldehydes in 30 min.
To increase the production of C9 aldehyde, more 9-LOX enzyme should be added to the reaction. Because GFP-infiltrated leaf extract contained high activity of 9-LOX (2.6 ± 0.1 U/g fresh weight of leaf tissue), it was used as additional 9-LOX source. When GFP- and CmHPL-infiltrated leaf extracts were simultaneously incubated with linoleic acid, the production of C9 aldehydes was increased two times compared with the reaction containing CmHPL-infiltrated leaf extract only. Higher levels (2.5 times) of 9-aldehydes were also produced when linoleic acid was preincubated with GFP leaf extract (LOX source) for 30 min, and then, CmHPL leaf extract was added to the reaction after 30 min (Figure 4b). Protein extracts prepared from 1 g of CmHPL-infiltrated tobacco leaves (fresh weight) in combination with protein extracts prepared from 1 g of GFP-infiltrated tobacco leaves produced 758 ± 75 μg total C9 aldehydes in 30 min. The yield of C9 aldehydes from linoleic acid was 60%.
Analysis of SlPXG-infiltrated and StEH-infiltrated tobacco leaves
The expression levels of Nb-9-LOX and SlPXG were examined in N. benthamiana leaves infiltrated with SlPXG by real-time PCR using the 18S-26S interspacer gene as an internal control for normalization (Figure 5a,b). As expected, internal 9-LOX (Nb-9-LOX) gene was up-regulated by treatment with viral vectors. The leaf harvested at 10 days after infiltration displayed the highest Nb-9-LOX gene expression level (Figure 5a). SlPXG transcript was detected in all leaves infiltrated with the SlPXG gene construct. Its level was low in leaves harvested 4 days after infiltration, but high in leaves harvested 7 days and 10 days after infiltration (Figure 5b). PXG activity was also analysed in tobacco leaves using oleic acid as a substrate. PXG activity was undetectable in untreated tobacco (Figure 5c), and low PXG activity was detected in the leaf extract infiltrated with GFP. Coinciding with the result of real-time PCR, the highest level of PXG activity was detected in the leaves harvested 7 days after infiltration (Figure 5c). EH activity was analysed in tobacco leaves infiltrated with StEH using trans-stilbene oxide as a substrate. The highest EH activity was measured in the leaves harvested 7 days after infiltration in accordance with the results obtained for HPL and PXG. Leaves harvested 10 days after infiltration still displayed a high level of EH activity (Figure 5d).
Metabolism of linoleic acid by SlPXG- and StEH-infiltrated tobacco leaf preparations
The proteins extracted from SlPXG-infiltrated tobacco leaf were incubated with linoleic acid for 30 min. Reaction products were extracted and analysed by LC-MS and UPLC–TOF-MS. Five main products (peaks 1–5) were detected by LC-MS (Figure 6). Peaks 4 and 5 could also be detected by ultraviolet monitoring at 234 nm (data not shown). Peak 4, identified as 9(S)-hydroperoxy-10(E),12(Z)-octadecadienoic acid (9-HPOD), is known to be the major product formed by tobacco 9-LOX (Huang and Schwab, 2011). It was detected in all leaf preparations incubated with linoleic acid (Figure 6). Peak 5, putatively identified as 9-(nona-1′,3′-dienoxy)non-8-enoic acid (colneleic acid; Figure S1E), is a divinyl ether fatty acid formed by the action of divinyl ether synthase on 9-HPOD (Galliard and Phillips, 1972; Weber et al., 1999). It was found in all assays containing infiltrated leaf extracts (Figure 6). Peak 3, putatively identified as 12,13-epoxy-9(S)-hydroxy-10(E)-octadecenoic acid (Figure S1D), is an epoxy product formed from 9-HPOD through SlPXG action. It was only detected in reactions containing SlPXG-infiltrated tobacco leaf extracts (Figure 6). Peak 1, identified as a mixture of 9,12,13-trihydroxy-10(E)-octadecenoic acid (peak 1a, Figure S1A) and 9,10,13-trihydroxy-11(E)-octadecenoic acid (peak 1b, Figure S1B), is chemically formed from 12,13-epoxy-9(S)-hydroxy-10(E)-octadecenoic acid during the chromatographic separation because of the acidic LC solvent (pH 3) (Hamberg, 1991; Hamberg and Hamberg, 1996). These compounds were also only detected in reactions bearing SlPXG-infiltrated tobacco leaf extracts (Figure 6). The compounds representing peak 2 were putatively identified as 12-O- and 13-O-methyl ether of 9,12,13-trihydroxy-10(E)-octadecenoic acid (Figure S1C). They are also chemically formed from 12,13-epoxy-9(S)-hydroxy-10(E)-octadecenoic acid during the chromatographic separation in the methanol/formic acid solution. Thus, the epoxy derivative was solely detected in preparations containing SlPXG-infiltrated tobacco leaf extracts (Figure 6). When LC was performed under neutral pH condition (without formic acid), peaks 1 and peak 2 were reduced, whereas peak 3 was largely increased (Figure S2). Peaks 1, 2, 3 and 5 were further analysed by UPLC–TOF-MS, and the results confirmed those of LC-MS analysis (Figures S3–S6).
When linoleic acid was preincubated with SlPXG leaf extract for 30 min, and then, StEH leaf extract was added to the reaction after 30 min, two isoforms of trihydroxyoctadecenoic acid were observed (peaks 1a and 1c; Figure S7). However, peak 1b, peak 2 and peak 3 were disappeared. Consequently, epoxy-hydroxy fatty acid (peak 3) was completely converted into trihydroxyoctadecenoic acids (peak 1a and peak 1c) by StEH (Figure S7, panel 4 and 5). Therefore, chemically formed trihydroxyoctadecenoic acid (peak 1b) and the methyl ethers of 9,12,13-trihydroxy-10(E)-octadecenoic acid (peak 2) were not detected. Peak 1a showed a retention time (29.2 min) and a MS/MS spectrum (m/z 329→229, 211, negative mode) identical to those of the standard 9(S),12(S),13(S)-trihydroxy-10(E)-octadecenoic acid. Peak 1c displayed a retention time of 30 min and the same MS/MS spectrum (m/z 329→229, 211, negative mode) as peak 1a, indicating the presence of an additional isomer of 9,12,13-trihydroxy-10(E)-octadecenoic acid. Furthermore, enzyme assays were performed in the presence of H218O (Figure 7). The result showed that both peaks 1a and 1c exhibited 18O incorporation in one of the hydroxyl groups (Figure 7b). Peak 1a showed a MS/MS spectrum of m/z 331→231 (negative mode), indicating the incorporation of 18O at C12 position, whereas peak 1c displayed a MS/MS spectrum of m/z 331→229 (negative mode), demonstrating that the incorporation of 18O occurred at position C13. Moreover, products formed by incubation of linoleic acid with SlPXG-infiltrated and StEH-infiltrated leaf extracts were methylated, trimethylsilylated, and then subjected to GC–MS. Peak 1a was unambiguously identified as 9(S),12(S),13(S)-trihydroxy-10(E)-octadecenoic acid, and peak 1c is 9(S),12(R),13(R)-trihydroxy-10(E)-octadecenoic acid (Figure S8). Taken together, StEH converts 12,13-epoxy-9(S)-hydroxy-10(E)-octadecenoate into two isomers of trihydroxyoctadecenoic acid, namely 9(S),12(S),13(S)- and 9(S),12(R),13(R)-trihydroxy-10(E)-octadecenoic acid. Protein extracts prepared from 1 g of SlPXG-infiltrated tobacco leaves (fresh weight) in combination with protein extracts prepared from 1 g of StEH-infiltrated tobacco leaves produced 1738 ± 27 μg total 9,12,13-trihydroxy-10(E)-octadecenoic acid isomers in 30 min. The yield of trihydroxyoctadecenoic acids from linoleic acid was 58%.
The present study aimed to produce high levels of LOX/HPL and LOX/PXG/EH enzymatic activities for the production of short-chain aldehydes and hydroxy fatty acids, respectively. Heterologous gene expression in plants is a central component of the many experimental and industrial processes in plant biotechnology (Biesgen et al., 2002; Streatfield, 2007). The main advantage foreseen for the plant system is reduced production cost (Biesgen et al., 2002; Streatfield, 2007; Twyman et al., 2003). The costs for growing plants on large acreage are low, and the production can be scaled up easily. Being able to produce high levels of industrial enzymes at low costs is a prerequisite for successful production of flavours and polymer precursors. Bacterial protein expression systems are also popular because bacteria are easy to culture, grow fast and produce high yields of recombinant protein. However, multi-domain eukaryotic proteins expressed in bacteria often are nonfunctional because the cells are not equipped to accomplish the required post-translational modifications or molecular folding. Also, many proteins become insoluble as inclusion bodies that are very difficult to recover without harsh denaturants and subsequent cumbersome protein refolding procedures (Demain and Vaishnav, 2009; Goldstein and Thomas, 2004). For example, we also expressed the SlPXG gene in bacteria, but did not obtain sufficient and stable PXG activity from the crude bacteria extracts. In contrast, the expression of SlPXG in tobacco showed high and stable PXG enzyme activity.
The viral vector system described by Marillonnet et al. (2004, 2005) not only enabled us to overexpress CmHPL, SlPXG and StEH genes in N. benthamiana within a short time (7–10 days), but also highly induces the enzyme activity of internal tobacco 9-LOX, which allowed us to successfully produce high levels of 9-LOX/9-HPL, 9-LOX/PXG and 9-LOX/EH enzyme activities. Vacuum infiltration guaranteed an even and homogenous distribution of the Agrobacterium strains within the entire plant (Huang et al., 2010). The highest levels of CmHPL activity under optimized infiltration conditions were 29 times the HPL activity of N. benthamiana infiltrated with a control vector (GFP), and 163 times that of untreated leaf. PXG activity in tobacco infiltrated with SlPXG was 83 times that in tobacco infiltrated with GFP control vector (Figure 5c). EH activity in tobacco infiltrated with StEH was 285 times that in tobacco infiltrated with GFP control vector (Figure 5d). However, no PXG or EH enzyme activities were detected in untreated leaves. The total aldehyde-forming activity of extracts prepared from CmHPL-infiltrated tobacco leaves is 50 times that of extracts prepared from control leaves (infiltrated with GFP; Figure 3d). The total trihydroxy fatty acids formed by SlPXG- and StEH-infiltrated tobacco leaf extracts from linoleic acid are 60 times that formed by leaf extracts prepared from control leaves. Expression of internal 9-LOX gene was highly induced after infiltration of TMV vectors. The highest level of 9-LOX gene transcript in infiltrated leaf was over 200 times that detected in untreated leaf (Figures 3a and 5a). The system also contains other advantages, for example, a very rapid process; short time from cloning to expression in plants; easy protein purification; no risk for outcrossing in most cases because the system is transient and not transmitted via germ-line.
Consumers have a strong preference for naturally synthesized additives and aromas with regard to food applications. Owing to the high demand for such natural flavours, many groups have attempted to develop biocatalytic processes to produce these compounds on an industrial scale. Processes that contain at least one enzymatic reaction catalysed by crude plant material have been patented (Belin et al., 1998; Brunerie and Koziert, 1997; Goers et al., 1989; Holtz et al., 2001; Kanisawa and Itoh, 1988; Muller et al., 1995b). Among them, the highest yields were reported by Muller et al. (1995b), for example, 4.2 g 3(Z)-hexen-1-ol kg−1 and 1.5 g 2(E)-hexenal kg−1 were produced. Owing to its low stability and the difficulty to purify it, HPL is the limiting component for the conversion of fatty acids into food flavour. Thereby, considerable efforts have been made to clone and produce this enzyme with enhanced stability and activity in Escherichia coli, yeast and transgenic tobacco (Brash et al., 2001; Fukushige and Hildebrand, 2005; Häusler et al., 2001; Huang and Schwab, 2011; Huang et al., 2010; Noordermeer et al., 2002). Also, different sources of HPLs have been explored for industrial production of C6 aldehydes in combination with soybean flour as a source of stable LOX (Fukushige and Hildebrand, 2005; Németh et al., 2004; Noordermeer et al., 2002; Rabetafika et al., 2008). Such LOX/HPL combinations yielded about 26%–60% of hexanal or hexenal from vegetable oils, linolenic acid or linseed oil. In this study, we present a viral vector system for heterologous expression of a 9-HPL in tobacco. Extracts prepared from 1 g of fresh leaves infiltrated with CmHPL and one gram of fresh leaves infiltrated with a control vector (as an additional source of LOX) could convert linoleic acid to 758 ± 75 μg of C9 aldehydes in 30 min. The yield was 60%. The system present here is competitive compared with the yeast expression system described before (Huang and Schwab, 2011). Therefore, this system has potential for producing C9 compounds on a large scale.
The irreversible inhibition of hydroperoxide lyase by its substrate (fatty acid hydroperoxide) has been reported (Santiago-Gómetz et al., 2007; Suurmeijer et al., 2000) when concentrations of substrate were used in excess. Therefore, the optimum LOX/HPL ratio used in the LOX/HPL system has to be considered. Moreover, we observed a phenomenon that the Nb-9-LOX activity was inhibited by CmHPL products (data not shown). That is why the highest amount of nonenal (Figure 3d) was produced in the reaction containing the leaf extract Cm-10 d, which showed the highest expression level of 9-LOX (Figure 3a) but contained less HPL activity (Figure 3c). Also, the production amount of 9-aldehydes in the reaction where GFP- and CmHPL-treated leaf extracts were simultaneously incubated with linoleic acid was lower than that in the reaction where linoleic acid was preincubated with GFP leaf extract for 30 min, and then, CmHPL leaf extract was added to the reaction (Figure 4b). It has also been reported that 4-hydroxy-2(E)-nonenal inhibited soybean LOX-1 activity (Gardner and Deighton, 2001). To overcome this problem, linoleic acid should be incubated with 9-LOX enzyme at first for at least 30 min, and 9-HPL can then be added to the reaction.
Molecular modifications of fatty acids can often lead to value-added products for a variety of new uses. Hydroxy and keto fatty acids are used as monomers for polymerization in plasticizers, lubricants, coatings and in the detergent industries (Hagemann and Rothfus, 1991; Kenney et al., 1974). Much effort has been focused on the microbial production of hydroxy fatty acids from various fatty acid substrates (Hou and Bagby, 1991; Hou et al., 1997; Kim et al., 2000; Kuo et al., 2001). Bacterium strains ALA2 and Pseudomonas aeruginosa PR3 were used to yield about 25%–45% of trihydroxyoctadecenoates from linoleic acid. In this study, we produced sufficient quantities of LOX, PXG and EH enzymes in transgenic tobacco. The procedure elaborated for the production of 9,12,13-trihydroxy-10(E)-octadecenoic acid is composed of three biocatalytic steps. The first step is the hydroperoxidation of linoleic acid by tobacco internal 9-LOX, which was highly induced by treating with the viral vector. The co-substrate for the fatty acid hydroperoxidation is the oxygen. For the large scale production of fatty acid hydroperoxides (9-HPOD), sufficient oxygen should be applied in the reaction system (Elshof et al., 1996; Fauconnier and Marlier, 1996). The second step is the epoxidation of 9-HPOD by heterologously expressed SlPXG. PXG requires hydroperoxide for its activity. As 9-HPOD can fulfil this purpose, no additional hydroperoxide was required to be added to the reaction. PXG is an integral membrane protein (Blée, 1998); therefore, detergent is required for its solubilization. The third step is the conversion of epoxy alcohols into trihydroxyoctadecenoates by EH. We observed a phenomenon that the PXG activity was inhibited when it was simultaneously incubated with EH. To overcome this problem, linoleic acid should be incubated with SlPXG-infiltrated leaf extracts at first for at least 30 min. Then, StEH-infiltrated leaf extracts can be added to the reaction. In this way, a large amount of trihydroxy fatty acids is produced in one experimental step. The isolation of trihydroxy fatty acids from the reaction mixture was simply realized by extraction with ethyl acetate.
We introduced the genes encoding enzymes downstream in the LOX pathway into tobacco leaves using the viral vector system described by Marillonnet et al. (2004, 2005). This expression system not only provides high protein yields of foreign genes, but also highly induces the gene expression of 9-LOX, which allowed us to successfully produce high levels of 9-LOX/9-HPL, 9-LOX/PXG and 9-LOX/EH enzyme activities in N. benthamiana within a short time (7–10 days), by a single infiltration of CmHPL, SlPXG or StEH gene constructs into tobacco, respectively. The biocatalytic processes described here allow effective generation of C9 aldehydes and trihydroxy fatty acids in one experimental step with many advantages including easy handling, time-saving and low production cost.
Nicotiana benthamiana was grown in a growth room maintained at 23 ± 1 °C with a 16-h light/8-h dark photoperiod and a light intensity of 70 ± 10 μmol/m2/s. Leaves used for transformation were taken from plants about 8–10 weeks old.
The viral vector system based on cr-TMV (crucifer-infecting TMV) is an expression system that relies on in planta assembly of functional viral vectors from separated pro-vector modules (Marillonnet et al., 2004, 2005). The 5′ module (pICH17388) contains the 5′ part of the viral vector including the RNA-dependent RNA polymerase, movement protein genes, the coat protein subgenomic promoter and a loxP site. The 3′ module (pICH11599) contains a loxP site, cloning sites for cloning of the gene of interest and the 3′ end of the viral vector. Both modules are assembled inside a plant cell with the help of a site-specific recombinase (pICH14011) to form a fully functional RNA replicon. pICH7410 containing GFP gene was used as a control construct. pICH17388, pICH11599, pICH14011 and pICH7410 were kindly provided by Icon Genetics AG (Halle, Germany).
Vector construction and agroinfiltration
The full-length open reading frames of SlPXG (accession number SGN-U567002), CmHPL (accession number AF081955) and StEH (accession number AAA81892) were amplified by PCR from plasmid DNAs pYES2-SlPXG (Aghofack-Nguemezi et al., 2011), pYES2-Cm-9/13-HPL (Huang and Schwab, 2011) and potato first-strand cDNA, respectively. The PCR primers used for each gene are listed in Table 1. The PCR products were double digested with restriction enzymes, the recognition sequences (underlined) of which are contained in the primers, and then ligated with pICH11599 vectors which were digested with the same restriction enzymes as for the inserts to yield pICH11599-SlPXG, pICH11599-CmHPL and pICH11599-StEH. The recombinant genes were subjected to sequencing to confirm the sequence of the inserts.
Table 1. Primer sequences for cloning of SlPXG, CmHPL and StEH into vector pICH 11599
Agrobacterium was used to deliver various modules into plant cells. pICH17388, pICH14011, pICH7410 and pICH11599 containing SlPXG, CmHPL or StEH were separately transformed into the Agrobacterium tumefaciens strain AGL0 using the freeze–thaw technique as described by Höfgen and Willmitzer (1988), and integrity was confirmed by PCR. Agrobacterium strains carrying each pro-vector module were mixed and infiltrated into N. benthamiana using a syringe without a needle as described (Huang et al., 2010).
Real-time RT-PCR analysis
Total RNA was extracted from leaves of transfected N. benthamiana and untreated leaves using the CTAB extraction procedure (Liao et al., 2004). RNA samples were treated with RNase-free DNase I (Fermentas, St. Leon-Rot, Germany) for 1 h at 37 °C. First-strand cDNA synthesis was performed in duplicate in a 20-μL reaction volume, with 1 μg of total RNA as the template, random primer (random hexamer, 100 pmol) and M-MLV reverse transcriptase (200 U; Invitrogen, Karlsruhe, Germany) according to the manufacturer's instructions. Real-time PCR was performed as described by Huang et al. (2010). Relative quantification of gene expression was performed using an 18S-26S interspacer gene as a reference (Pfaffl, 2001). Primers used are listed in Table 2.
Table 2. Primer sequences for real-time PCR analysis
IS, interspacer gene.
Enzyme extraction and assay
One hundred milligram (fresh weight) samples of N. benthamiana leaves infiltrated with Agrobacterium were ground into a fine powder in liquid nitrogen using a mixer mill MM 400 (Retsch, Haan, Germany), followed by being resuspended in 300 μL of protein extraction buffer (for CmHPL: 50 mm sodium phosphate buffer, pH 7.5, 10 mm EDTA, 0.1% Triton X-100 (v/v), 5 mm β-mercaptoethanol; for SlPXG and StEH: 50 mm sodium phosphate buffer, pH 7.5, 10 mm EDTA, 10% glycerol (v/v), 0.2% Tween 20 (v/v), 5 mm β-mercaptoethanol). The homogenate was centrifuged at 4 °C, 16 000 g for 10 min to remove the cell debris. Total protein content was determined by Bradford assay (Bradford, 1976).
HPL activity was determined in 120 μL of 50 mm sodium phosphate buffer (pH 7.0) containing 2 μL of tobacco leaf extract and 50 μm of 9(S)-HPOD at room temperature. The decrease in fatty acid hydroperoxide was measured spectrophotometrically by following the decrease in A234 because of cleavage of the substrate by HPL. The concentration of remaining substrate was calculated using an extinction coefficient of 23 000/m/cm. One unit of activity (U) corresponds to the amount of enzyme that converts 1 μmol of substrate per minute.
PXG activity was determined by GC-MS using oleic acid as a substrate. SlPXG-infiltrated tobacco extract (50 μL) was incubated with 2.5 mm hydrogen peroxide and 200 μm oleic acid in 500 μL of buffer C (10 mm sodium acetate buffer, pH 6.0, 2% glycerol, 0.01% Tween 20). After 30 min at 30 °C, the reaction mixture was extracted once with the same volume of dichloromethane and evaporated to dryness under a nitrogen stream. After formation of the methyl esters with (trimethylsilyl) diazomethane (Aghofack-Nguemezi et al., 2011), the reaction products were analysed by GC-MS. Relative PXG activity was determined by comparing the production amount of methyl 9,10-epoxystearic acid (mass peak area, m/z 155). The PXG activity measured in sample GFP was defined as 1.
EH activity was determined by LC-MS using trans-stilbene oxide (TSO) as a substrate. StEH-infiltrated tobacco extract (50 μL) was incubated with 300 μm TSO in 500 μL of 1× PBS buffer (140 mm NaCl, 8 mm Na2HPO4, 2.7 mm KCl, 1.76 mm KH2PO4, pH 7.4) for 30 min at 30 °C. The reaction products were extracted with the same volume of ethyl acetate, evaporated to dryness, resuspended in methanol and analysed by LC-MS. Relative EH activity was determined by comparing the production amount of stilbene diol (mass peak area, m/z 237 [M + Na]+). The EH activity measured in sample GFP was defined as 1.
For analysis of PXG and EH products, 50 μL of SlPXG-infiltrated or StEH-infiltrated N. benthamiana leaf extracts was incubated at 30 °C in 500 μL of buffer C containing 300 μm of linoleic acid with constant shaking for 30 min. The reaction products were extracted with the same volume of ethyl acetate, evaporated to dryness, resuspended in 100 μL of 100% methanol and analysed by LC-MS and UPLC–TOF-MS. The amount of trihydroxyoctadecenoic acid was determined using standard curves calculated from various known concentrations of 9(S),12(S),13(S)-trihydroxy-10(E)-octadecenoic acid (generous gift from Mats Hamberg Karolinska Institutet, Stockholm, Sweden) against the mass peak areas (m/z 329 [M-H]−, negative mode) that were recorded by LC-MS. For labelling experiments with H218O, freeze-dried leaf extracts with reaction buffer were resuspended in the initial volume of H218O. Assays were carried out as described above.
For analysis of aldehyde formation, 50 μL of CmHPL-infiltrated N. benthamiana leaf extracts was diluted to 2 mL of 50 mm sodium phosphate buffer (pH 7) containing 75 μm linoleic acid. The mixture was incubated for 30 min at 25 °C with constant shaking in a 20 mL reaction vial closed with a septum. Headspace compounds were trapped by SPME (65-μm polydimethylsiloxane–divinylbenzene-coated fibre; Supelco, Steinheim, Germany) at 45 °C for 30 min and analysed by GC-MS. The amount of nonenal isomers was determined using standard curves calculated from various known concentrations of (2E)-nonenal against the mass peak areas (m/z 69, 83) that were recorded by SPME-GC-MS.
High-performance liquid chromatography–electrospray ionization mass spectrometry (LC-MS)
The HPLC system consisted of a quaternary pump and a variable wavelength detector, all from Agilent 1100 (Bruker Daltonics, Bremen, Germany). The column was a LUNA C18 100A 150 × 2 mm (Phenomenex, Aschaffenburg, Germany). HPLC was performed with the following binary gradient system: solvent A, water with 0.1% formic acid and solvent B, 100% methanol with 0.1% formic acid. The gradient programme was as follows: 0–10 min, 70% A/30% B to 50% A/50% B; 10–40 min, 50% A/50% B to 100% B, hold for 5 min; 100% B to 70% A/30% B, in 1 min, then hold for 6 min. The flow rate was 0.2 mL min−1. Absorbance was recorded at 234 nm for the detection of hydroperoxy fatty acids (9- and 13-HPOD) and colneleic acid. Attached to the HPLC was a Bruker esquire 3000 plus mass spectrometer with an ESI interface that was used to record the mass spectra. The ionization voltage of the capillary was 4000 V, and the end plate was set to −500 V.
Gas chromatography-mass spectrometry (GC-MS)
The volatile compounds collected from the headspace were analysed on a Thermo Finnigan Trace DSQ mass spectrometer coupled to a 0.25 μm BPX5 20 m fused silica capillary column with a 30 m × 0.25 mm inner diameter. He (1.1 mL min−1) was used as carrier gas. The injector temperature was 250 °C, set for splitless injection. The temperature programme was 40 °C for 1 min, 40–60 °C at a rate of 2 °C min−1, and 60–325 °C at 10 °C min−1. The ion source temperature was 250 °C. Mass range was recorded from m/z 50–300, and spectra were evaluated with the Xcalibur software version 1.4 (Thermo, Dreieich, Germany) supplied with the device. Diagnostic ions for nonenal were m/z 69 and 83.
The temperature programme for the analysis of products formed by the incubation of PXG and oleic acid was 100 °C for 5 min, 100–190 °C at a rate of 30 °C min−1, 190–250 °C at a rate of 8 °C min−1, and 250–280 °C at a rate of 20 °C min−1. Finally, it was held at 280 °C for 10 min. The injector temperature was 220 °C, set for split injection, and the ion source temperature was 250 °C. Mass range was recorded from 50 to 600 m/z and spectra were evaluated with the Xcalibur software version 1.4. The diagnostic ion for methyl epoxystearic acid was m/z 155.
Ultra-performance liquid chromatography–time-of-flight mass spectrometry (UPLC–TOF-MS)
Mass spectra of the compounds were measured on a Waters Synapt G2 HDMS mass spectrometer (Waters, Manchester, UK) coupled to an Acquity UPLC core system (Waters, Bedford, MA) consisting of a binary solvent manager, sample manager and column oven. Aliquots (3 μL) of SlPXG reaction products were injected into the UPLC–TOF-MS system equipped with a BEH C18, 2 × 150 mm, 1.7 μm, column (Waters, Manchester, UK). Operated with a flow rate of 0.4 mL min−1 at a temperature of 45 °C, the following gradient was used for chromatography: starting with a mixture (20/80, v/v) of acetonitrile (0.1% HCOOH) and aqueous formic acid (0.1% HCOOH), the acetonitrile content was increased to 95% within 3 min and, then, kept constant for 1 min. Scan time for the MSE method (centroid) was set to 0.1 s. Measurements were performed using negative ESI and the resolution mode consisting of the following ion source parameters: capillary voltage -2.0 kV, sampling cone 20 V, extraction cone 4.0 V, source temperature 150 °C, desolvation temperature 450 °C, cone gas 30 L/h and desolvation gas 850 L/h. Data processing was performed using the elemental composition tool for determining the exact mass. All data were lock mass corrected on the pentapeptide leucine enkephaline (Tyr-Gly-Gly-Phe-Leu, m/z 554.2615, [M-H]−) in a solution (2 ng/μL) of acetonitrile/0.1% formic acid (1/1, v/v). Scan time for the lock mass was set to 0.3 s, an interval of 15 s and 3 scans to average with a mass window of ± 0.3 Da. Calibration of the Synapt G2 in the range from m/z 50–600 was performed using a solution of sodium formate (5 mm) in 2-propanol/water (9/1, v/v).
The authors thank Professor Hans-Ulrich Koop and Dr Yuri Gleba (Icon Genetics, Halle, Germany) for the viral expression system and N. benthamiana seeds. 9(S),12(S),13(S)- trihydroxy-10(E)-octadecenoic acid was kindly provided by Prof. Mats Hamberg (Karolinska Institute, Stockholm, Sweden). We are grateful for the UPLC-MS analyses performed by Dr Timo Stark (Chair of Food Chemistry and Molecular Sensory Science, Technische Universität München). Financial supports from AIF 15088N/1 and 2 as well as SynRg® (BMELV-FNR) are acknowledged.