The electron transfer molecules plastoquinone and ubiquinone are formed by the condensation of aromatic head groups with long-chain prenyl diphosphates. In the present paper we report the cloning and characterization of two genes from tomato (Solanum lycopersicum) responsible for the production of solanesyl and decaprenyl diphosphates. SlSPS (S. lycopersicum solanesyl diphosphate synthase) is targeted to the plastid and both solanesol and plastoquinone are associated with thylakoid membranes. A second gene [SlDPS (S. lycopersicum solanesyl decaprenyl diphosphate synthase)], encodes a long-chain prenyl diphosphate synthase with a different subcellular localization from SlSPS and can utilize geranyl, farnesyl or geranylgeranyl diphosphates in the synthesis of C45 and C50 prenyl diphosphates. When expressed in Escherichia coli, SlSPS and SlDPS extend the prenyl chain length of the endogenous ubiquinone to nine and ten isoprene units respectively. In planta, constitutive overexpression of SlSPS elevated the plastoquinone content of immature tobacco leaves. Virus-induced gene silencing showed that SlSPS is necessary for normal chloroplast structure and function. Plants silenced for SlSPS were photobleached and accumulated phytoene, whereas silencing SlDPS did not affect leaf appearance, but impacted on primary metabolism. The two genes were not able to complement silencing of each other. These findings indicate a requirement for two long-chain prenyl diphosphate synthases in the tomato.
- virus-induced gene silencing (VIGS)
The prenylquinones PQ (plastoquinone) and UQ (ubiquinone) are electron carriers involved in numerous biochemical processes. In the chloroplast PQ is a component of the electron-transport chain, mediating the flow of electrons from photosystem II to the cytochrome b6f complex . PQ is a component of the redox chain associated with the desaturation of phytoene , whereas the redox state of the PQ pool has been shown to regulate the expression of nuclear- and plastid-encoded genes involved in photosynthesis . UQ acts in the mitochondria where it has a central role in the mitochondrial respiratory chain, mediating electron transfer from the NADH dehydrogenase complex through to the cytochrome bc1 complex . Its reduced form, ubiquinol, is an antioxidant preventing the initiation of lipid peroxidation and photo-oxidation of mitochondrial lipids.
Structurally PQ and UQ comprise an aromatic head group with an isoprenoid (prenyl) side chain to confer lipid solubility. Isoprenoids are derived from a five carbon precursor, IPP (isopentenyl diphosphate), and its isomer DMAPP (dimethylallyl diphosphate). In plants these are produced by the cytosolic MVA (mevalonic acid) pathway and in plastids via the MEP (methylerythritol 4-phosphate) pathway . Labelling experiments have demonstrated that the IPP utilized in the formation of the UQ side chain is synthesized from MVA, whereas PQ originates from IPP formed via the MEP pathway [6,7]. Thus these two prenylquinones not only function in separate organelles, but also use IPP formed in different subcellular compartments. Prenyl chains are assembled from the consecutive condensation of IPP and its allylic isomer DMAPP, involving the formation of short chains, such as FPP (farnesyl diphosphate, C15) and GGPP (geranylgeranyl diphosphate, C20), which in turn are elongated further by a class of enzymes known as long-chain prenyl diphosphate synthases. The length of the chain is determined by the activity of these enzymes and is known to vary among different organisms and is determined by the specificity of the long-chain prenyl diphosphate synthases present . SPSs (solanesyl diphosphate synthases) and DPSs (decaprenyl diphosphate synthases) are responsible for the production of the C45 and C50 prenyl chains respectively (Figure 1A).
To produce PQ, solanesyl diphosphate is attached to the aromatic ring of homogentisate by the prenyl transferase , known as HST (homogentisate solanesyl transferase). This reaction marks the first committed step in PQ formation and results in the formation of 2-methyl-6-solanesyl-1,4-benzoquinol . This is then methylated by MSBQ MT (2-methyl-6-solanesyl-1,4-benzoquinone methyltransferase) to produce PQ. These reactions are believed to occur exclusively in the plastid . In the mitochondrion, the condensation of the aromatic intermediate PHB (p-hydroxybenzoic acid) with the polyprenyl diphosphate is catalysed by PPT (PHB polyprenyltransferase), leading to UQ formation (Figure 1B and ). However, little is known about their functional equivalents in crop species, including tomato.
The hp (high-pigment) tomato mutants are characterized by simultaneous elevation of multiple classes of phytochemicals and exaggerated responsiveness to light. The hp phenotype results from defects to genes encoding negative regulators of phytochrome signal transduction [13,14] and manipulation of light signalling pathways has been seen as an effective strategy to improve the nutritional value of the tomato . Transgenic suppression of the gene responsible for the hp2 mutant, DET1 (DE-ETIOLATED1) resulted in substantial elevations in the carotenoid and flavonoid content in fruit . Multi-level ‘omic’ characterization of these varieties has revealed that co-ordinated up-regulation of core metabolic processes, such as the Calvin cycle and photorespiration, as well as plastid size are the probable progenitors of the hp chemotype, rather than transcriptional activation of isoprenoid biosynthetic pathway genes . In the present study, we have identified a putative trans-long-chain prenyl diphosphate synthase gene that is up-regulated in response to DET1 silencing in tomato fruit, concomitant with increased accumulation of solanesol and PQ. This and a further putative tomato long-chain prenyl lipid synthase has been characterized in vitro and in vivo, their biochemical role in prenyl lipid formation elucidated and the effect of modulating PQ and UQ levels on the metabolome determined.
Tobacco (Nicotiana tabacum cv Wisconsin 38) was used for the generation of transgenic plants. Plants were grown under a 16 h photoperiod at 25°C. The tomato (Solanum lycopersicum) T56 genotype was used for the isolation of long-chain prenyl diphosphate synthases and VIGS (virus-induced gene silencing) treatments. Homozygous TFM7 was used as the DET1 down-regulated genotype and grown as described in . Nicotiana benthamiana was used in the transient expression system.
The tomato SPS full-length coding sequence (1194 bp, GenBank® accession number DQ889204) was amplified from tomato leaf cDNA using KOD Hot Start DNA Polymerase (Novagen) with oligonucleotide primers containing restriction sites (underlined) SPSNdeI_For (5′-CGGACATATGTCTGTGACTTGCCATAATC-3′) and SPSBamHI_Rev (5′-CTGAGGATCCCTATTCAATTCTCTCCAGAT-3′). The amplified products were cloned into the NdeI-BamHI sites of the Escherichia coli expression vector pET14b (Novagen).
For construction of the plant transformation vector pEXP35SSPS, two oligonucleotide primers (SPSTOPO_For, 5′-CACCATGTCTGTGACTTGCCATAATC-3′ and SPSTOPO_Rev, 5′-CTATTCAATTCTCTCCAGATTAT-3′) were used to amplify the complete SPS coding sequence. SPSTOPO_For includes the CACC sequence upstream of the ATG start codon (in bold) to allow the amplified product to be introduced into the Gateway® transformation vector pENTR/D-TOPO (Invitrogen) by the TOPO reaction, following the manufacturer's instructions. This produced pENTR-SPS, which was recombined by the LR reaction (Invitrogen) into the plant transformation vector pK7WG2, which allows strong expression from the CaMV35S (where CaMV is cauliflower mosaic virus) promoter . This yielded vector pEXP35S-SPS and transformants were selected in LB (Luria–Bertani) broth supplemented with 50 mg/l spectinomycin. The SPS sequences in E. coli and plant transformation vectors were verified by sequencing.
The vector pEXP35S-SPS was transformed into electrocompetent Agrobacterium tumefaciens strain LBA4404 and selected on YEB (yeast extract broth) plates (50 mg/l spectinomycin, 100 mg/l streptomycin and 50 mg/l rifampicin). N. tobaccum was transformed as described by Horsch et al. . A total of 30 independent transformants were screened by PCR using the primers 35SFor (5′-CAATCCCACTATCCTTCGC-3′) and SPSTOPO_Rev (5′-CTATTCAATTCTCTCCAGATTAT-3′) to detect the presence of the inserted T-DNA.
The vectors pEXP35S-SPS:GFP, pEXP35SDPS:GFP and the helper plasmid P19K  were separately transformed into AGL1 Agrobacterium containing the virG expression enhancer . Cultures of Agrobacterium containing P19K or the GFP (green fluorescent protein) vectors were pelleted by centrifugation (5000 g for 20 min at 4 °C) and resuspended to a D600 of 0.1 in infiltration medium (as in ). The GFP and P19K cultures were mixed 1:1 and used to inoculate the abaxial surface of mature leaves of N. benthamiana using a needleless syringe. Leaves were collected 4 days after infiltration and imaged immediately. Confocal microscopy was performed with a FluoView FV1000 (Olympus). GFP and chlorophyll were excited using the 488 nm line from the argon laser, and emission signals were collected in separate channels, at wavelengths between 495 and 526 nm, and between 631 and 729 nm respectively. Transmitted light was collected in a separate channel. The GFP signal was false coloured green and chlorophyll autofluorescence was false coloured red.
The TRV (tobacco rattle virus)-based silencing vectors pTRV1, pTRV2-MCS and pTRV2-LePDS were obtained from the Arabidopsis Biological Resources Centre (Ohio State University, Colombus, OH, U.S.A.). For the construction of pTRV2-SlSPS, a 409 bp fragment of the SlSPS (S. lycopersicum SPS) gene was PCR amplified from tomato leaf cDNA using the primers SPS VIGS forward (5′-CGGTCTAGACAAGAACTTGCATAATATTG-3′) and SPS VIGS reverse (5′-CGGGGATCCCACCACTTGCAAAGTCTTTAA-3′). For the pTRV2-SlDPS construction a 409 bp fragment of the SlDPS (S. lycopersicum DPS) gene was PCR amplified using DPS VIGS forward (5′-CGGTCTAGAGTTGCAGAGTAATTCATTC-3′) and DPS VIGS reverse (5-CGGGGATCCCAGTACATCATCATGAAGTA-3′). The resulting PCR products were cloned into the XbaI and BamHI sites of pTRV2-MCS to form pTRV2-SPS and pTRV2-DPS respectively. For pTRV2-SPS:DPS construction, the SlSPS and SlDPS fragments above were fused together by PCR using primers SPS VIGS forward, DPS VIGS reverse and the SPS:DPS fusion (5′-TAAAGACTTTGCAAGTGGTGGTTGCAGAGTAATTCATTCA-3′) and subsequently cloned into pTRV2-MCS. As a control pTRV2-PDS was constructed using a 409 bp fragment of the tomato PDS (phytoene desaturase) gene (GenBank® accession number X71023), using primers PDS VIGS forward (5′-CGGTCTAGAGGCACTCAACTTTATAAACC-3′) and PDS VIGS reverse (5′-GCTGGATCCCTTCAGTTTTCTGTCAAACC-3′) cloned into the XbaI and BamHI sites of pTRV2 MCS. pTRV1 and the pTRV2-derived vectors were separately transformed into A. tumefaciens strain GV3101.
Plant infiltration was performed as described previously , using 2-week-old tomato seedlings. Plants were incubated in darkness overnight after infiltration. A visible phenotype was observed 10 days after inoculation and leaf material was harvested in liquid nitrogen 12 days later.
Protein purification and digestion
pET14b-SPS was transformed into BL21 Star (DE3) pLysS E. coli cells (Sigma–Aldrich). For high level expression the recombinant proteins were expressed by adding isopropyl β-D-thiogalactopyranoside to a final concentration of 0.4 mM overnight at 24°C. Cells were pelleted by centrifugation at 5000 g for 20 min at 4°C and pellets were resuspended in 10 ml of protein-binding solution [20 mM Tris/HCl (pH 8.0), 0.5 M NaCl and 5 mM imidazole) and stored at −20°C. Harvested cells were defrosted and lysed by sonication using a Sonic Vibracell Ultrasonic processor (Sonics). The lysate was centrifuged at 5000 g for 15 min at 4°C and supernatants transferred at room temperature (20 °C) to an Amersham Biosciences Chelating SepharoseTM Fast Flow column to purify His-tagged protein from the soluble fraction.
Purified proteins were visualized by SDS/PAGE and for identification of purified proteins by MS, bands corresponding to SlSPS or SlDPS (by size estimation) were excised from polyacrylamide gels prior to destaining and trypsin digestion as described in . Peptides were dried down under vacuum and resuspended in 40 μl of 0.1% TFA (trifluoroacetic acid) and 2% (v/v) ACN (acetonitrile) and filtered through a 0.45 μm nylon membrane.
MALDI (matrix-assisted laser-desorption ionization)–TOF (time-of-flight)-MS
Digested peptides, desalted and concentrated using Zip Tips and eluted in 1:1 H2O/ACN with 0.1% TFA were spotted (0.5 μl) on to a 600 μm Anchor chip 384 format MALDI target plate (Bruker Daltonics) and overlaid with 0.5 μl of DHB (2,5-dihydroxybenzoic acid) matrix (ACN+0.1% TFA) and air dried. Mass spectra were obtained on a Bruker Autoflex MALDI—TOF/TOF-MS (Bruker Daltonics) equipped with a nitrogen-pulsed laser (377 nm). Operating conditions were as follows: positive reflectron mode, ion source 1=19.1 kV, ion source 2=16.74 kV, lens voltage=8.85 kV, reflectron voltage 1=21.11 kV, reflectron voltage 2=9.7 kV, laser power=14%, laser frequency=200 Hz and matrix suppression=400 Da. Mass spectra were collected in the range 500–4500 m/z. At total of ten random points per sample were each obtained with 30 laser shots per point. Peptides from mass spectra were analysed in Flex Analysis 3.3 and Biotools v.3.2 (Bruker Daltonics) and matched against the databases NCBInr Viridiplantae and SwissProt Viridiplantae using an in-house Mascot search engine (Matrix Science). Search parameters were: mass tolerance of 0.5 Da; enzyme, trypsin; modifications, deamidation; and missed cleavages, 1.
nESI (nano electrospray ionization)-LC (liquid chromatography)-MS/MS (tandem MS) identification of purified SlSPS and SlDPS
Identification of trypsin-digested SlSPS and SlDPS peptides was carried out using the Amazon ETD (Bruker Daltonics) on-line with a UHPLC UltiMate 3000 (Dionex Softron). Chromatographic separations were performed at 35°C using an Acclaim PepMap RSLC nano C18 2 μm 100 Å (1 Å=0.1 nm) column (75 μm×15 cm, LC Packings/Dionex), coupled to an Acclaim PepMap C18 3 μm 100 Å (75 mm×2 cm) nano-trap column (LC Packings/Dionex). The mobile phases consisted of (A) 0.1% formic acid in water and (B) 80% (v/v) ACN and 0.1% formic acid. The gradient used was 96% (A) isocratically for 5 min at a flow rate of 0.25 μl/min, followed by a linear gradient over 20 min to 50% (A), and another linear gradient for 5 min to 90%. This last condition was held in isocratic mode for 5 min. The ionization mode used was nESI operating in positive mode. Vaporization temperatures were set at 200°C and a full MS scan was performed from 100 to 2000 m/z.
MS/MS spectra were exported to Biotools v. 3.2 (Bruker Daltonics) and submitted for Mascot search of Viridiplantae entries with the following parameters: one missed cleavage, mass error MS 0.4 Da, search for single, double and triple charged ions, and searching the decoy database. The MS/MS spectra were also compared with the deduced amino acid sequences of SlSPS and SlDPS following theoretical digestion (performed using trypsin, with two missed cleavages and the following modifications: oxidation to methionine, carbamidomethylation and deamidations) and the percentage coverage of observed fragments was determined.
The extraction, HPLC separation, PDA (photodiode array) detection and quantification of carotenoids were performed as described previously . Three extractions were performed from each biological replicate. For quantification of solanesol, aliquots of 20 mg of ground freeze dried plant tissue were saponified in 200 μl of methanol, 500 μl of dH2O (distilled water) and 6% (v/v) KOH and incubated for 1 h at 40°C. Solanesol was extracted in 500 μl of chloroform and the organic hypophase was removed and re-extracted in 500 μl of chloroform. Pooled extracts were dried under nitrogen and resuspended in 100 μl of ethyl acetate. Samples were analysed by HPLC according to , with three extractions performed from each biological replicate. Solanesol was identified by co-chromatography with an authentic standard (Sigma–Aldrich) and quantified at 210 nm by comparison with dose–response curves (0.2–1.0 mg).
Ratios of UQ-8/UQ-9/UQ-10 were calculated from their corresponding peak areas obtained from 275 nm chromatograms. Under these chromatographic conditions  retention times for UQ-8, UQ-9 and UQ-10 were 20.6 min, 23.3 min and 25.5 min respectively. Quantification of ratios was performed using this system as it provided better separation of UQ-10 (25.5 min) and Mk-8 (24.6 min) than the LC-PDA-MS system (Figure 4) described in the section below.
LC-PDA-MS analysis of isoprenoids and prenylquinones
Identification of solanesol, decaprenol, α-tocopherol, γ-tocopherol and UQ-10 was carried out using the high-resolution Q-TOF mass spectrometer UHR-MAXIS (Bruker Daltonics), on-line with a UHPLC UltiMate 3000 equipped with a PDA detector (Dionex Softron). Chromatographic separations were performed in a similar manner to , with the exception that a reverse-phase C30 3 μm column [150 mm×2.1 mm i.d. (internal diameter)] coupled to a 20 mm×4.6 mm C30 guard column was used (YMC). The mobile phase was altered to facilitate ionization and was composed of (i) methanol, containing 0.1% formic acid and (ii) tert-butyl methyl ether, containing 0.1% formic acid. These solvents were used in a gradient mode starting at 100% (i) for 2 min, then stepped to 80% (i) for 1 min, held for 3 min and followed by a linear gradient over 4 min to 30% (i). This last condition was kept for 10 min in isocratic mode and after that initial conditions (100% i) were restored for 2 min. The column was then re-equilibrated for 5 min. The flow rate used was 0.2 ml/min and the injection volume was 10 μl. The positive ionisation mode was APCI (atmospheric pressure chemical ionization). The capillary and APCI vaporization temperatures were 250°C and 450°C respectively and the dry gas (nitrogen) and nebulizer were set at 4 l/min and 2 bar (1 bar=100 kPa) respectively. The APCI source settings were: corona discharge voltage at 4000 nA and a capillary voltage of 4 kV. A full MS scan was performed from 200–1000 m/z and MS/MS spectra were recorded at an isolation width of 0.5 m/z. A collision energy ramp from 40 to 80 eV was applied for target masses between 650 and 900 m/z. Instrument calibration was performed externally prior to each sequence with APPI (atmospheric pressure photoionization)/APCI calibrant solution (Fluka). Automated post-run internal calibration was performed by injecting the same APPI/APCI calibrant solution at the end of each sample run via a six port divert valve equipped with a 20 μl loop.
GLC-MS analyses of metabolites
Extraction and analysis of polar metabolites was performed as described by , with slight modifications. Freeze dried powder (10 mg) was extracted in 1 ml of methanol/0.21 M HCl (80:20 by volume), containing the internal standard ribitol (final concentration 0.04 mg/ml). A 20 μl aliquot was removed from pelleted samples and dried under nitrogen gas. Four extractions were performed on each biological replicate. Derivatization was performed by the addition of 30 μl of methoxyamine-HCl (Sigma–Aldrich) at 20 mg/ml in pyridine. Samples were incubated at 40°C for 1 h, after which 80 μl of MSTFA (N-methyl-N-trimethylsilyltrifluoroacetamide; Sigma–Aldrich) was added and the samples incubated for 2 h at 40°C before analysis. GLC-MS was performed as described previously , using a 20:1 split injector. Metabolites were quantified using ChemStation software (Agilent), facilitating integrated peak areas for specific compound targets (qualifier ions) relative to the internal standard (ribitol) peak. Heat maps were generated using Qlucore software, using mean metabolite levels for each treatment.
In vitro solanesol diphosphate synthase assays
Solanesol diphosphate synthase activity was determined in vitro as described , with slight modifications. The incubation mixture contained, in a total volume of 200 μl, 50 mM Tris/HCl buffer (pH 7.5), 5 mM MgCl2, 1 mM DTT (dithiothreitol), 46 μM [1-14C]IPP (1.85 GBq/mmol), 0.5% Triton X-100 and 38.6 μg of purified protein, and GPP (geranyl diphosphate; 13.7 μM), GGPP (11.1 μM) or FPP (11.5 μM). Samples were incubated for 1 h at 30°C and reactions stopped by cooling on ice and the addition of 200 μl of 1 M NaCl. Samples were extracted in 2 ml of water-saturated butanol overnight at −20°C. Butanol extractable products were dephosphorylated according to . Potato acid phosphatase (1 unit; Sigma–Aldrich) was added to 1 ml of extract and incubated in 700 μM 0.1% Triton X-100 and 4 ml of methanol in a total volume of 10 ml at 37°C overnight on a shaking platform water bath. Samples were extracted in 10 ml of 10:90 (v/v) diethyl ether/petroleum ether (boiling point 40–60°C) and the organic phase dried under nitrogen. Solanesol was identified by TLC using reverse-phase C18 F254S plates (Merck), with a mobile phase of 19:1 (v/v) acetone/water and visualized by exposure to iodine vapour. Comparison of the effect of different divalent cations was performed as described above with slight modifications. The incubation mixture (100 μl) contained 50 mM Tris/HCl buffer (pH 7.5), 177.6 nM GGPP (for the SlSPS assay) or 184.6 nM FPP (for the SlDPS assay), 1 mM DTT, 36 μM [1-14C]IPP, 0.5% Triton X-100 and 2.5 μg of purified protein and 10 mM CaCl2, 10 mM MgCl2, 10 mM MnCl2 or 10 mM ZnCl2. Samples were incubated for 15 min at 30°C and then the reactions stopped by cooling on ice and the addition of 100 μl of 1 M NaCl. Samples were extracted in 1 ml of water-saturated butanol and after centrifugation the amount of [1-14C]IPP incorporated into butanol-extractable polyprenyl diphosphate was measured in triplicate by scintillation counting. For kinetic studies, the concentration of the allylic substrate (FPP/GPP) was varied, while the co-substrate ([1-14C]IPP) was at 36 μM. Kinetic parameters and their standard errors were estimated by non-linear regression analysis using GraphPad Prism software v.5 (GraphPad Software).
Solanesol isomers were identified using LC-MS. The solanesol standard (1 mg, Sigma, tobacco extract, 90% pure) was resuspended in chloroform and separated by reverse-phase TLC [solvent 19:1 (v/v) acetone/water]. Two solanesol bands were visualized by exposure to iodine vapour, heated to 50°C and subsequently scraped from the TLC plate. The compounds were extracted with 1 ml of chloroform, dried under nitrogen, redissolved in 100 μl of chloroform and analysed by LC-MS, as described above.
TEM (transmission electron microscopy)
Leaf tissues from representative VIGS-treated plants were fixed in 2.5% (w/v) glutaraldehyde in 50 mM sodium cacodylate buffer overnight at 4°C. Samples were rinsed twice in the same buffer prior to fixing in 1% (w/v) osmium tetroxide in 50 mM sodium cacodylate buffer for 1 h at room temperature. Samples were dehydrated in a graded ethanol series, then embedded in LV (low viscosity) resin (Agar Scientific) and polymerized at 60°C for 24 h. Sections (75 nm) were prepared with a RMC-MTXL ultramicrotome (Diatome) and counterstained with 4.5% (w/v) uranyl acetate in 1% (v/v) acetic acid for 45 min and Reynolds lead citrate stain for 7 min. Microscopy was performed on a Jeol 1230 transmission electron microscope and images captured with a Gatan digital camera.
Lipid peroxidation in tomato leaf tissue was measured as TBARS (2-thiobarbituric acid-reacting substances) formation, according to the method described in , using approximately 100 mg of intact leaflets from five plants per genotype. TBARS values are expressed as MDA (malonaldehyde) equivalents, using the molar absorption coefficient 155 mM−1·cm−1.
TEAC (Trolox equivalent antioxidant capacity) total antioxidant activity assays
A non-polar extract was generated using the carotenoid/isoprenoid extraction procedure described above, with 10 mg of homogenized freeze-dried tobacco leaf tissue. TEAC assays were performed as described in  by generating two ABTS [2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid); Sigma–Aldrich] free radical cations, ABTS+. Results are expressed as a TEAC in mmol of Trolox/g of DW (dry mass).
Measurement of gene expression by real-time qRT–PCR (quantitative reverse-transcription PCR)
RNA was extracted from fresh-frozen plant material extracted using RNeasy reagents (Qiagen). The QuantiFast SYBR Green one-step real time qRT–PCR kit (Qiagen) was used to determine gene expression levels using gene specific primers. Melt-curve analyses verified product specificity and the CT value calculations were performed by Rotor-Gene software (Qiagen), calibrated against a dilution series of cloned gene products, run simultaneously with experimental samples. The tomato and tobacco actin genes served as references for normalization.
RESULTS AND DISCUSSION
The effect of DET1 down-regulation on the transcription of SlSPS and SlDPS in the tomato
Previous transcriptomic studies using microarrays have demonstrated that the increases in isoprenoids in tomato plants silenced for DET-1 are generally not achieved through increases in expression of isoprenoid biosynthetic genes, but by post-transcriptional regulation . However, transcripts of a SlSPS (GenBank® accession number DQ889204) were found to be elevated at the mature green stage of fruit development in the DET-1 silenced variety TFM7, relative to the wild-type control (T56) . In the present study, we have used qRT–PCR to accurately analyse transcript levels of SlSPS. The SlSPS transcript was detected in immature green fruit of T56 and TFM7 (10- and 15-mm diameter fruit, Figure 2A) and remained virtually constant through development into ripe fruit in T56, but in TFM7 fruit levels were elevated at the mature green stage onwards compared with those in T56 until a significant 3-fold increase in the transcript levels was found within ripe fruit tissue (Figure 2A). These elevated transcript levels in TFM7 mirrored the increases in solanesol and PQ (Figures 2B and 2C). TFM7 fruit accumulated 5-fold greater levels at the B+7 stage than did T56 fruit (Figure 2C). Similarly, solanesol, the alcohol derivative of solanesyl diphosphate, was more abundant in TFM7 fruit than T56 fruit at all developmental stages and both varieties displayed a similar profile through to the mature green stage (Figure 2B). In TFM7, however, the levels became dramatically increased at the breaker (B) and 7 days post-breaker (B+7) stages, to produce a 12-fold difference between the two varieties. These increases in solanesol were in same order of magnitude as the elevations in carotenoids, xanthophylls, chlorophylls and tocopherols reported previously . Since the DET1 phenotype is characterized by the overproduction of both the photosynthetic apparatus, and phenolic and ascorbate antioxidants, it is plausible that up-regulation of SlSPS in DET1 mutants may serve to allow for greater production of PQ to function in photosynthesis and as an antioxidant. The reason for this difference in the mechanism for regulating levels of these compounds in not yet understood.
Recombinant expression, purification and in vitro properties of SlSPS and SlDPS
The tomato Unigene library (http://www.solgenomics.net) contains a candidate DPS (SGN-U573523), exhibiting significant similarity to the putative tomato solanesol diphosphate synthase. The ORF (open reading frame) encoded by SGN-U573523 shares 71% amino acid identity with the rice mitochondrial SPS [OsSPS (Oryza sativa SPS) 1] and 40% identity with the plastidial SPS (OsSPS2)  and has been designated SlDPS. Conversely, SlSPS shares 67% amino acid identity with OsSPS2 and 40% identity with OsSPS1. Accordingly, SlSPS possesses a putative transit peptide sequence for plastid localization, whereas SlDPS contains a putative N-terminal mitochondrial signal peptide sequence, as predicted by the protein localization programs TargetP and ChloroP . The deduced amino acid sequences of SlSPS and SlDPS each contain two conserved aspartate-rich motifs [DD(X)nD; Supplementary Figure S1 at http://www.BiochemJ.org/bj/449/bj4490729add.htm] and phylogenetic analysis positioned these two tomato genes among known long-chain prenyl diphosphate synthases from other species and distinct from the tomato GGPPSs (geranylgeranyl diphosphate synthases; Figure 3).
In order to functionally characterize these prenyl diphosphate synthase gene products, the full-length coding regions were amplified from tomato leaf cDNA and separately cloned into the E. coli expression vector pET-14b. Expression in DPS BL21star (DE3) pLysS E. coli cells and subsequent purification yielded His-tagged SlSPS and SlDPS proteins of 45446 and 45443 Da respectively. The identity of the purified proteins was verified by MS/MS of proteolytic digests, with observed amino acid coverage of 58.5 and 41.6% for the SlSPS and SlDPS sequences respectively (Supplementary Table S1 at http://www.BiochemJ.org/bj/449/bj4490729add.htm).
When purified SlSPS was assayed with either FPP or GGPP, two prominent compounds were observed on TLC (Rf values 0.52 and 0.46, Figure 4A). Authentic solanesol also produced two bands on TLC with these Rf values. Analysis by LC-MS indicated that the two products, which had different retention times (18.17 and 18.38 min), both have the appropriate molecular mass for solanesol (m/z 613.57 [M+H]±H2O). Purified SlDPS was able to use FPP, GPP or GGPP as its allylic substrate and the resulting products included a major component with an Rf of 0.37 (Figure 4B). In the presence of FPP, bands co-migrating with the products obtained with SlSPS were also visible. The products of both enzyme activities were confirmed by LC-MS analyses as solanesol, whereas SlDPS was shown to produce solanesol and decaprenol (Supplementary Figure S1).
Direct involvement of SlSPS and SlDPS in polyprenyl diphosphate synthase activity in vivo was determined by LC-MS analysis of UQ species in E. coli DH5a BL21 cells (Supplementary Table S2 at http://www.BiochemJ.org/bj/449/bj4490729add.htm). Cells transformed with an empty pET-14b vector were shown by LC-MS to produce UQ-8 (m/z 727.57 [M+H]+) which is synthesized by the endogenous E. coli octaprenyl diphosphate synthase and eluted with a retention time (Rt) of 10.3 min (Supplementary Figure S2 at http://www.BiochemJ.org/bj/449/bj4490729add.htm). Menaquinone-8 (m/z 717.56 [M+H]+) was also detected. In contrast, transformation with pET14-b SPS produced a distinct quinone profile resulting from accumulation of the C45 compound UQ-9 (m/z 795.63 [M+H]+), which eluted at an Rt of 11.2 min (Figure 4C). LC-MS analysis of extracts of E. coli harbouring pET14-b DPS demonstrated that recombinant SlDPS expression had resulted in production of UQ-9 as well as UQ-10 (m/z 863.69 [M+H]+), which eluted after menaquinone-8 at an Rt of 11.7 min. These findings demonstrate that SlDPS encodes a functioning DPS. Therefore expression of SlSPS or SlDPS alone in E. coli was sufficient to modify the chain length of the endogenous UQ species, demonstrating that like the long-chain prenyl diphosphates characterized in other plant species [29,31–33] those in tomato function as single gene products. SlSPS expression provided for UQ prenyl side chains consisting strictly of nine isoprene units, whereas SlDPS provided side chains with nine or ten isoprene units. Thus the tomato genome, unlike that of Arabidopsis and rice which produce only single species of side chain (C45), contains genes allowing the production of C45 and C50 side chains.
The divalent cation requirements of purified tomato SlSPS and SlDPS (Supplementary Table S3 at http://www.BiochemJ.org/bj/449/bj4490729add.htm) were shown to be similar to those from the homodimer type enzymes found in rice [OsPPT (O. sativa PPT) 1] and S. cerevisiae (COQ2, ). Purified SlSPS was most active in the presence of MgCl2, whereas MnCl2 and MgCl2 were both effective at stimulating SlDPS activity. In addition the apparent Km values for SlSPS and SlDPS indicate a preference for GGPP as a substrate rather than FPP for both enzymes (0.46 and 1.51 μM respectively, Table 1), in agreement with Arabidopsis enzymes with SPS activity, AtSPS (Arabidopsis thaliana SPS) OsSPS1 and AtSPS2 [32,33]. It is unclear why those enzymes not targeted to the plastid (SlDPS and AtSDS1) do not show a preference for the cytosolic substrate FPP, although here the Vmax value of SlDPS for FPP was higher than that for GGPP (Table 1). Long-chain prenyl diphosphate synthases exhibit a requirement for divalent metal cations, particularly Mg2+, to allow binding of the substrates to the active site , as do the aromatic prenyl transferases [36,37].
Subcellular localization of tomato long-chain prenyl transferases and solanesol
To experimentally confirm the subcellular localization of SlSPS and SlDPS, their full length coding regions were transcriptionally fused to the GFP gene, placed under the control of the CaMV35S promoter and the vectors expressed in N. benthamiana. SlSPS was shown to be localized in chloroplasts as judged by chlorophyll autofluorescence in the confocal microscope (Figure 5). The fluorescence signal of SlDPS–GFP was more diffuse and it was not possible to determine precisely its subcellular location, although it may resemble the mitochondrial localization previously reported for the rice OsSPS1 . The first Arabidopsis long-chain diphosphate synthase to be isolated, AtSPS1, was shown to be targeted to the ER [32,33] and presumed to be responsible for the production of solanesol moieties for UQ biosynthesis. Recently, however, a further enzyme, AtPPPS (A. thaliana trans-type polyprenyl pyrophosphate synthase), which possesses SPS activity in vitro, was shown to be targeted to the mitochondrion and plastid and identified as the main contributor to SPS activity in UQ biosynthesis . Since the Arabidopsis p-hydroxybenzoate prenyl transferase AtPPT1 (A. thaliana PPT1) is also localized to the mitochondrion, this is likely to be the main site of UQ synthesis . Silencing AtPPPS did not affect PQ accumulation and hence, despite plastid targeting, does not appear to participate in production of PQ . A further Arabidopsis SPS, AtSPS2, is believed to provide the prenyl diphosphate for this purpose and is also targeted to the plastid [31,32], where all of the enzymes involved in the final steps of PQ synthesis are localized [40,41].
Functional characterization of SlSPS and SlDPS in planta
VIGS was used to gain further insight into the function of SlSPS and SlDPS in tomato vegetative tissues. A mixture of Agrobacterium cultures containing the pTRV2-derived vectors and pTRV1, a plant binary transformation plasmid containing a cDNA clone of tobacco mosaic virus RNA 1, was infiltrated into tomato seedlings. pTRV1, mixed with an empty vector (pTRV2-MCS), was infiltrated as a negative control (TRV) and a vector (pTRV2-PDS), containing a 409 bp fragment of the tomato phytoene desaturase gene was used as a positive control (TRV2-PDS, ). Tomato plants infected with pTRV2-PDS developed a bleaching phenotype in the upper leaves 4 weeks after Agrobacterium infiltration. Plants infected with either pTRV2-SPS or pTRV2-SPS-DPS also became photobleached, although less dramatically than pTRV2-PDS. The leaflets contained pale green sectors at the base, giving a variegated appearance and were abnormal (Figure 6A). There was no visible phenotype in plants infected with TRV2-DPS. The mottled phenotype is similar to that observed by phytoene desaturase silencing, although less pronounced, and presumably results from inhibited production of PQ since plants silenced for SlSPS were found to contain significantly decreased levels of PQ (reduced 42% in TRV2-SPS and 38% in TRV2-SPS-DPS, although not statistically significant in the latter case compared with the control; Supplementary Table S4 at http://www.BiochemJ.org/bj/449/bj4490729add.htm). PQ acts as a cofactor in carotenoid biosynthesis by accepting the hydrogen atoms released during phytoene desaturation, re-oxidizing the reduced flavin (FADH2) and it is subsequently reoxidized by the plastid terminal oxidase . Leaf bleaching is observed in mutants defective in PQ reoxidation  or biosynthesis [2,40]. We propose that as a result of SlSPS silencing reduced biosynthesis of PQ inhibits carotenoid biosynthesis, making plastids vulnerable to photooxidative damage and preventing their normal development. Plastids in bleached tissue (Figure 6B, ii) resembled those resulting from treatment with the PDS inhibitor norflurazon , having a modified morphology and stroma largely void of thylakoid membranes. It is noteworthy that in our experiments SlDPS was not able compensate for losses of SlSPS activity in TRV2-SPS-treated plants. The TRV2-SPS plastids also contained small aggregates of electron dense globuli as well as large less dense membrane bound bodies.
Transcripts levels of SlSPS and SlDPS genes were measured in the treated leaf tissue. In pTRV2-SlSPS-infected plants the transcripts of SlSPS were reduced by 69% and, similarly, SlDPS expression was reduced to 55% in pTRV2-SlDPS plants (Supplementary Figure S3 at http://www.BiochemJ.org/bj/449/bj4490729add.htm). In the TRV2 SPS-DPS-treated plants the SlSPS and SlDPS transcripts were shown to be reduced by 62 and 52% respectively (Supplementary Figure S4 at http://www.BiochemJ.org/bj/449/bj4490729add.htm). These decreases in transcript levels are relatively modest, but this is thought to reflect the uneven tissue distribution of the VIGS phenotype penetration observed previously in the tomato .
Chlorophyll and carotenoid analyses of VIGS-treated tissues showed that the predominant effect was the accumulation of phytoene in plants silenced for SlSPS (TRV2-SlSPS, 49-fold and TRV2-SPS-DPS, 40-fold over TRV controls; Supplementary Table S4), whereas phytoene levels in plants treated with TRV2-DPS alone were not significantly altered. TRV2-SPS plants also had lower levels of chlorophylls a and b, as well as lutein and the xanthophylls neoxanthin and violaxanthin.
TRV2-SPS plants were further analysed for accumulation of other prenyl lipid species, since the tocopherols and PQ share the precursor homogentisate. Compared with the TRV controls, TRV2-SPS-treated plants were found to accumulate elevated levels of α-tocopherol (2-fold), but there was no change in the levels of γ-tocopherol. Interestingly, levels of UQ-10 were also increased 3-fold in plants silenced for SlSPS (Supplementary Figure S4A; see Supplementary Table S5 at http://www.BiochemJ.org/bj/449/bj4490729add.htm for compound identification). Since quinones and carotenoids have a role in the protection of membranes against oxidative stress, leaves from glasshouse-grown TRV2 and TRV2-SPS were analysed for their levels of lipid peroxidation. TRV2-SPS plants contained 50% greater levels of lipid peroxides (MDA/g of fresh mass, Supplementary Figure S4), indicating increased susceptibility to reactive oxygen species generated as products of photosynthesis and that the chlorotic phenotype observed is a consequence of oxidative damage to lipids. In agreement with the photobleached appearance and perturbed plastid ultrastructure of the TRV2-SPS plants, the maximum quantum yield of photosystem II was also reduced. This is estimated from the Fv/Fm ratio (Fm−Fo/Fm, where Fo and Fm are minimum and maximum chlorophyll a fluorescence respectively of dark-adapted leaves) and is a reliable indicator of photosynthetic performance. TRV2 leaves had a mean Fv/Fm of 0.824±0.001 and TRV2-SPS bleached leaves had a mean Fv/Fm of 0.462±0.109.
To provide insight into the sectors of metabolism affected by the VIGS treatments, metabolite profiling was undertaken. A heatmap, incorporating metabolites extracted in polar and non-polar fractions, is shown in Figure 7 and the data presented in full in Supplementary Table S6 (at http://www.BiochemJ.org/bj/449/bj4490729add.htm). For comparative purposes, each metabolite is presented as variance from its mean abundance across all genotypes. Noteworthy are the similarities in clusters between those plants treated with TRV2-PDS and TRV2-SPS-DPS, marked by relative reductions in multiple classes of metabolites, including sugars, organic acids and isoprenoids, and increases in a number of amino acids. TRV2-SlDPS plants are differentiated from TRV controls by a cluster containing elevated levels of sugars glucose, fructose, mannose and arabinose, whereas a cluster containing reduced levels of chlorophylls, lutein and PQ is present in the TRV2-SlSPS plants. Silencing of either SlSPS or SlDPS alone thus produced metabolite profiles that were distinct from one another and the TRV2-PDS, TRV2-SlSPS-SlDPS cluster. The metabolic profile of TRV2-SlSPS-SlDPS plants does not appear to represent a combination of the profiles resulting from silencing the two long-chain prenyl transferases individually, but clustering of TRV2-SlSPS-SlDPS and TRV2-PDS away from the control does reflect the severity of the phenotype resulting from VIGS treatment. The difference in profiles between the PDS- and SlSPS-silenced phenotype may be explained by the differences in the effectiveness of VIGS treatment. Alternatively, TRV2-SPS plants may compensate for lack of solanesyl diphosphate by incorporating shorter prenyl diphosphate chains into PQ. These were not detected in the present study, but have been reported previously . Furthermore, increased accumulation in SlSPS-silenced plants of UQ-10 and α-tocopherol may be a compensatory antioxidant mechanism or, in the case of α-tocopherol, owing to reduced competition with PQ biosynthesis for homogentisate.
In order to assess the potential of SlSPS as a tool to elevate the synthesis of solanesol and its role in prenyl quinone formation, transgenic tobacco plants overexpressing tomato SlSPS were produced. From a total of 30 plants displaying resistance to kanamycin, 90% were found to be PCR positive for SlSPS. Each of these plants was regenerated in the glasshouse and analysed for their total solanesol content. The levels of solanesol in the vegetative tissues were highly variable and a number of plants displayed unusual phenotypic characteristics, such as reduced stature and mottled leaf appearance (Supplementary Table S7 at http://www.BiochemJ.org/bj/449/bj4490729add.htm). A total of four lines (S-9, S-11, S-19 and S-29), selected on the basis of solanesol content, were self-pollinated to create the T1 generation plants. PCR screening of T1 generation plants from these selected lines identified that the majority (89%) of the T1 progeny contained the SlSPS transgene with azygous plants only identified for lines 9 and 19, indicating that S-11 and S-29 contained multiple transgene inserts. Of the 48 T1 plants screened, 31 individuals, including three azygous individuals from line 9, showed solanesol content above the mean wild-type level (412.5 μg/g of DW) in fully expanded mature leaves. The solanesol content was greatest in the line S-29-6, at 1383.6 μg/g of DW, 3.4-fold greater than the wild-type mean content (Table 2). The four transgenic plants with the highest solanesol content and a single azygous plant were chosen to investigate the effects of high solanesol content on the biosynthesis of other isoprenoids. Only line 9-6 showed altered pigment contents, with an increased abundance of β-carotene, lutein and chlorophyll a of 33%, 39% and 34% respectively. The PQ levels in the transgenic lines were no different from those determined for the wild-type plants (Table 2). The same compounds were then screened in immature leaves to see whether overexpression of SlSPS had any effects in developing tissue. The solanesol content in immature leaves was one order of magnitude lower than that found in mature leaves and in only one transgenic line measured (S9-6) was the solanesol content elevated relative to the wild-type levels (Table 2). However, immature leaves of all transgenic lines showed a consistent elevation between 36 and 55% in PQ levels above that in the wild-type. Line 9-16 also showed small, but significant, elevations in lutein (12%), β-carotene (13%) and chlorophyll (18%) levels compared with the wild-type plants, whereas line 29-6 contained amounts of lutein, violaxanthin, β-carotene and chlorophylls reduced by 26%, 39%, 41% and 30% respectively.
Since plastoquinol (PQ-H2) is known to be an effective antioxidant, TEAC analysis was performed on non-polar extracts from immature leaf tissue to assess whether overexpression of SlSPS had resulted in elevated total antioxidant capacity in tobacco. Statistically significant increases in antioxidant capacity of 15 and 24% were observed in lines S-9 (P<0.01) and S-29 (P<0.001) respectively, over the wild-type control (Supplementary Figure S5 at http://www.BiochemJ.org/bj/449/bj4490729add.htm).
These data suggest that expression of SlSPS and supply of solanesol is limiting for the production of PQ in immature leaves. The increase in PQ content may also explain the enhanced antioxidant activity in non-polar extracts of immature transgenic tobacco leaves. Like tocopherols, PQ levels are elevated in plant tissues in response to a range of abiotic and biotic stresses and there is strong evidence that in its reduced form it has an important function as an antioxidant . Furthermore, solanesol levels are also induced in tobacco leaves following viral infection and are implicated as part of the natural response to pathogens [47,48]. Whether plants that accumulate greater solanesol levels are more resistant to pathogens has yet to be determined.
In summary, we have shown that two enzymes from tomato (S. lycopersicum) with different subcellular distributions, are responsible for the production of solanesyl and decaprenyl diphosphates, namely SlSPS and SlDPS. The two enzymes cannot complement each other and have different substrate specificities. In silenced DET-1 lines their expression is up-regulated, unlike other genes encoding isoprenoid biosynthetic enzymes. When expressed in E. coli, SlSPS and SlDPS extend the prenyl chain length of the endogenous UQ to nine and ten isoprene units respectively.
Matthew Jones performed the experimental work with the assistance of Laura Perez-Fons and Francesca Robertson, especially related to the MS analysis of metabolites and proteins respectively. Matthew Jones, Paul Fraser and Peter Bramley contributed to the design of the experimental approach. All authors contributed to data interpretation, preparation and writing of the paper. Peter Bramley and Paul Fraser obtained the funding and Peter Bramley edited the paper prior to submission.
This work was supported by the U.K. Biotechnology and Biological Sciences Research Committee [grant number BB/F005644/1].
We thank Chris Gerrish for excellent technical support.
Abbreviations: ABTS, 2,2′-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid); ACN, acetonitrile; APCI, atmospheric pressure chemical ionization; APPI, atmospheric pressure photoionization; AtPPPS, Arabidopsis thaliana trans-type polyprenyl pyrophosphate synthase; CaMV, cauliflower mosaic virus; DET1, DE-ETIOLATED1; DMAPP, dimethylallyl diphosphate; DPS, decaprenyl diphosphate synthase; DTT, dithiothreitol; DW, dry mass; FPP, farnesyl diphosphate; GFP, green fluorescent protein; GGPP, geranylgeranyl diphosphate; GGPPS, GGPP synthase; GPP, geranyl diphosphate; hp, high-pigment; IPP, isopentenyl diphosphate; LC, liquid chromatography; MALDI, matrix-assisted laser-desorption ionization; MDA, malonaldehyde; MEP, methylerythritol 4-phosphate; MS/MS, tandem MS; MVA, mevalonic acid; nESI, nano electrospray ionization; PDA, photodiode array; PDS, phytoene desaturase; PHB, p-hydroxybenzoic acid; PPT, PHB prenyltransferase; PQ, plastoquinone; qRT–PCR, quantitative reverse-transcription PCR; SlDPS, Solanum lycopersicum DPS; SPS, solanesyl diphosphate synthase; AtSPS, A. thaliana SPS; OsSPS, Oryza sativa SPS; SlSPS, S. lycopersicum SPS; TBARS, 2-thiobarbituric acid-reacting substance; TEAC, Trolox equivalent antioxidant capacity; TFA, trifluoroacetic acid; TOF, time-of-flight; UQ, ubiquinone; VIGS, virus-induced gene silencing
- © The Authors Journal compilation © 2013 Biochemical Society