Plant materials, RNA isolation, and cDNA synthesis
Helianthus annuus L. cv. HA300 and Lactuca sativa var. capitata plants were grown under greenhouse conditions with an additional 16 h illumination (330 μmol s-1 m-2) and a night length of 8 hours. H. annuus capitate glandular trichomes were mechanically isolated from anther appendages as previously described . Studies on the organ development revealed that the formation of trichomes on anthers starts early and parallels the consecutive centripetal maturation of florets within the capitulum. Different trichome stages were determined microscopically by direct analysis of trichomes and pollen development; and light microscopic images were taken as previously described .
For total RNA extraction, trichomes were isolates from fresh plant material and immediately transferred to 200 μl ice-chilled RNA extraction buffer (Aurum total RNA isolation Kit, Biorad, Munich, Germany) in 2 ml vials. A mixer mill (MM20, Retsch, Haan, Germany) was used for cell disruption (16 Hz, 1 min) using 2 ceramic beads (2.8 mm diameter, Precellys, Peqlab, Erlangen, Germany). After cell disruption, an additional 500 μl lysis buffer was added. All subsequent steps were carried out as described in the manual. For total RNA isolation from pure glands, glandular trichomes from 200 florets (approx. 30,000 to 40,000 trichomes) were used. For the purification of RNA from different trichome stages, the glands from anther appendages of 50 florets were isolated. RNA quantity and integrity was verified by the Bioanalyzer 2100 using a RNA 6000 Pico Chip (Agilent, Böblingen, Germany). For routine PCR, cDNA was synthesised from total RNA using the RevertAid First Strand cDNA Synthesis Kits (Fermentas, St. Leon-Rot, Germany) with VNdT18 primer.
Identification of HaGAS1 and HaCS
For identification of sesquiterpene synthase genes by PCR, degenerate primers were used to obtain fragments of sunflower sesquiterpene synthases (forward primer, 5'-GAY GAR AAY GGI AAR TTY AAR GA-3', and reverse primer, 5'-CCR TAI GCR TCR AAI GTR TCR TC-3' ). PCR was performed in a total volume of 25 μl containing 1 μmol of the two primers, 0.25 μmol dNTPs, 1 unit of Taq DNA Polymerase (Fermentas, St. Leon-Roth, Germany) and 2 μl of cDNA. The PCR reaction was performed on a Mastercycler Gradient (Eppendorf, Hamburg, Germany) with 3 min of initial denaturation at 94°C, followed by 35 cycles of 1 min denaturation, 1 min annealing at 42°C, and 2 min of elongation at 72°C. Agarose gel electrophoresis showed a single band with a length of approximately 500 bp. Separation of the same PCR-reaction on 10% polyacrylamide gel revealed two bands with approximately 560 and 600 bp in length. Both bands were excised and transferred to 2 ml reaction tubes. After the addition of 150 μl ddH2O, the gel fragments were disrupted in a mixer mill for 1 min at 10 Hz using 2 ceramic beads. After centrifugation (5 min, 10,000 g), the supernatant was removed and used directly for reamplification of the PCR fragments with the same primer pair as before. The PCR products were gel-purified (Qiagen Gel Extraction Kit, Hilden, Germany) and used for direct sequencing using the same primers.
To obtain full length sequences, 3'-RACE was performed according to the protocol of Sambrook & Russell . The reaction was carried out in a thermocycler with 5 min initial denaturation (94°C), 5 min of annealing at 49°C and a first elongation for 40 s at 72°C, followed by 30 cycles with 40 s denaturation (94°C), 1 min annealing (49°C), and 3 min elongation at 72°C. The PCR amplifications and subsequent direct sequencing of the resulting fragments were carried out using the RACE-adapter primer (5'-GAC TCG AGT CGG ACA TCG A-3' ) and the gene specific primer 5'-TTG AGA TTG AAA GGG AAA AC-3' for HaGAS1 and the gene specific primer 5'-CCA ACT AAG AAT AAG AGG AGA ATC-3' for HaCS. For identification of the 5'-ends of the mRNA sequences of HaGAS1 and HaCS, trichome total RNA was reversely transcribed to cDNA using VNdT18 primers. The enzyme assay was purified with the Eppendorf PCR Purification Kit (Hamburg, Germany) and a single stranded DNA-oligonucleotide (5'-ACT AGG ATC CAA GCT TGG AAT TCG TAC GTC TAG AGA TAT C-3', blocked by fluoresceine at the 3'-end, phosphorylated at the 5'-end) was ligated to the 3' end of cDNA by T4 RNA-ligase (Fermentas) at 37°C overnight [modified protocol from Edwards et al. and Troutt et al. [66, 67]; T4-RNA ligase buffer (Fermentas), 20 μM ATP (Fermentas), 0.25 μg PEG 6000 (Roth GmbH, Karlsruhe, Germany) per μl ligation assay, 10 μg BSA (Fermentas) per μl ligation assay, 1 mM CoCl2 (Sigma-Aldrich, Taufkirchen, Germany), 25 nM DNA-oligonucleotide, 0.05 μl T4 RNA ligase per μl ligation assay]. The 5'-ends were amplified by PCR using gene-specific reverse primers (5'-GAC TTC AGA GTA ATA CGG CTC C-3' for HaGAS1 and 5'-GAC TTC AGA GTA ATA CGG CTC C-3' for HaCS) and a nested forward primer (5'-GAT ATC TCT AGA CGT ACG AAT C-3') for the ligated oligonucleotide at the 3' end of the cDNA. PCR was performed with 5 min initial denaturation, followed by 35 cycles with 40 s denaturation (94°C), 1 min annealing (52°C), 80 s elongation (72°C) and a final elongation step of 10 min using PCR reaction conditions as described above.
Sequencing of the genomic DNA for HaGAS1 and HaCS
The genomic DNA (gDNA) sequences of HaGAS1 and HaCS were identified from genomic DNA isolated from H. annuus cv. HA300 leaves using the GenElute Plant Genomic DNA Miniprep Kit (Sigma-Aldrich GmbH, München, Germany). For subsequent amplification of the gDNA of HaGAS1, the following primer pairs were used: forward primer, 5'-CCT TCC ATC AAA TAA TTT TGA AG-3' and reverse primer, 5'-GTC TCT TGA AAC CTC ATA TCC-3'; forward primer, 5'-TGG TGC TAG ATG ACA CAT ATG AC-3' and reverse primer, 5'-CAC GAT TGA GAT ATT GTC CTA G-3'; forward primer, 5'-TTG AGA TTG AAA GGG AAA AC-3' and reverse primer, 5'-AGC ATC TTC ACT CAC TAT CTC AC-3'. The gDNA for HaCS1 was amplified with the following primer pairs: forward primer, 5'-TTG CAC CAA CTC CCA TTC-3' and reverse primer, 5'-GAC TTC AGA GTA ATA CGG CTC C-3'; forward primer, 5'-GGA GCC GTA TTA CTC TGA AGT C-3' and reverse primer, 5'-gga gcc gta tta ctc tga agt c-3'; forward primer, 5'-GAC TTC AGA GTA ATA CGG CTC C-3' and reverse primer, 3'-CCA ACT AAG AAT AAG AGG AGA ATC-5'. While amplification of HaCS gDNA revealed a single band, the PCR amplification of HaGAS1 resulted in two products with different length. The PCR products were gel-purified using the illustra GFX PCR Gel Band Purification Kit (GE Healthcare GmbH, Munich, Germany) and cloned into the pSC-A plasmid by UA-cloning following the instructions of the StrataClone PCR Cloning Kit (Stratagene Inc., La Jolla, USA). The fragments were sequenced. Sequences were aligned using the Seqman module of the DNASTAR software package (Lasergene, Madison, WI, USA). Introns and exons were identified by comparison of the gDNA sequences with the previously established mRNA sequences for HaGAS1 and HaCS. Both PCR products for HaGAS1 with the same primer pairs showed highly similar sequences in the exon parts but differed within the intron sequences in length and nucleotide composition. This resulted in the identification of a second germacrene A synthase gene (HaGAS2). For all sequencing work the sunflower cultivar HA300 was used, but the presence of all three sesquiterpene synthase genes was verified in wild type H. annuus (data not shown).
Detection of the HaGAS1 and HaGAS2transcripts in trichomes and roots
The full length coding sequence for HaGAS1 and HaGAS2 was amplified from trichome and root cDNA using the forward primer 5'-ATG GCA GCA AGT TGG AGC CAG-3' and the reverse primer 5'-TTA CAT GGG TGA AGA ACC AAC AAA C-3'. The PCR conditions were 2 min initial denaturation followed by 30 cycles of 40 s denaturation (94°C), 40 s annealing (60°C), 2 min 20 s elongation (72°C). The PCR amplicons were gel-purified. To distinguish between HaGAS1 and HaGAS2 amplicons, a restriction digestion was performed using PauI and DraI (Fermentas GmbH). The HaGAS1 contains a PauI but no DraI recognition site while HaGAS2 contains a DraI site but no PauI recognition site.
Semi-quantitative RT-PCR and real-time quantitative RT-PCR
The cDNAs for RT-PCR were generated as described above. The constitutively expressed ubiquitin mRNA  served as the reference transcript. After reverse transcription, each cDNA sample was diluted several fold and used for PCR with the forward primer 5'-CAA AAC CCT AAC CGG AAA GA-3' and the reverse primer 5'-ACG AAG ACG GAG GAC GAG-3' to amplify ubiquitin cDNA. Equal initial cDNA concentrations within different samples were defined by equal amplification intensity for the ubiquitin transcript. PCR-Cycle number for PCR reactions was chosen to be in the linear range. Transcipts for HaGAS1 and HaGAS2 were traced by amplification with the forward primer 5'-TTG AGA TTG AAA GGG GAA AAC-3' and the reverse primer 5'-TGC CAA CAG AGT ATC TAG GTT CA-3'. To determine the expression level of HaCS, the forward primer 5'-CCA ACT AAG AAT AAG AGG AGA ATC-3' and the reverse primer 5'-GAC TTC AGA GTA ATA CGG CTC C-3' were used. The transcript for farnesyl diphosphate synthase [FDPS; Genbank: AF071887] was amplified with the following primer pair: forward primer 5'-ACT GCT TGT ACG GCT TTG CTT G-3' and reverse primer 5'-TTT CTT GCA TCT GCC CTT GGT TG-3'. For all semi-quantitative RT-PCR-experiments PCR products were separated on 1% agarose gels, stained for 30 min in a water bath containing 1.5 μg ml-1 ethidium bromide and exposed to UV light (312 nm) for documentation.
For quantitative real-time PCR 5× master mixes (LightCycler FastStart DNA MasterPLUS SYBR Green I, Roche Diagnostics GmbH, Mannheim, Germany) were used according to manufacturer's recommendations on a LightCylcer 1.5 instrument (Roche Diagnostics GmbH). Same primer combinations as for semi-quantitative RT-PCR were used. Initial denaturation time for all samples was 10 min (95°C), followed by 50 cycles with 10s annealing (57°C for FPPS and HaCS, 59°C for HaGAS, 60°C for Ubiquitin), 20s elongation (72°C). Ramp temperature was set to 20°C/s. The melting curve of all samples was analyzed. Additionally, all reactions were loaded on 1% agarose gels to ensure amplification of products with the expected length. Relative gene expression was calculated using the Pfaffl method  with Ubiquitin as the reference transcript and compared to the expression levels of the specific transcripts of cDNA obtained from RNA isolations of glandular trichomes in presecretory stage.
Heterologous expression of HaGAS1, HaGAS2, and HaCS in E. coli
The production of soluble proteins was made possible by expressing the proteins as N-terminal thioredoxin fusion proteins with the pET-32 EK/LIC plasmid (Novagen, Darmstadt, Germany). For generation of recombinant protein, the coding sequences for HaGAS1 and HaGAS2 were amplified with the forward primer 5'-GAC GAC GAC AAG ATG GCA GCA ATT GGA GC-3' and the reverse primer 5'-GAG GAG AAG CCC GGT TTA CAT GGG TGA AGA ACC AAC-3' from cDNA with KOD Polymerase (Novagen). The cDNA for HaCS was amplified using the forward primer 5'-GAC GAC GAC AAG ATG GCA ACA ACT GAA GC-3' and the reverse primer 5'-GAG GAG AAG CCC GGT TAC ATG GGG ACT GGA AC-3' and use of native Pfu-Polymerase (Fermentas). The PCR products were inserted in the pET-32 EK/LIC plasmid (Novagen, Darmstadt, Germany) according to the recommended protocol. The constructs were designated as pET32::HaGAS1, pET32::HaGAS2, and pET32::HaCS. These were used to transform NovaBlue cells (Novagen) which were grown overnight on Luria-Bertani (LB) plates supplemented with ampicillin (100 μg/ml). Plasmids were isolated from overnight cultures (LB medium supplemented with ampicillin) and their sequences were verified. Expression vectors for HaGAS1 and HaGAS2 were subcloned into the E. coli strain Rosetta-gami B (DE3)pLysS (Novagen), pET32::HaCS was subcloned into Rosetta 2 (DE3)pLysS cells (Novagen). Rosetta 2 cells were selected on LB-plates supplemented with carbenicillin (100 μg/ml), chloramphenicol (34 μg/ml). For Rosetta-gami B LB-plates kanamycin (15 μg/ml) and tetracycline (12.5 μg/ml) was also added.
Heterologous expression of sesquiterpene synthase genes in E. coliand recombinant protein purification
LB-medium supplemented with carbenicillin (100 μg/ml) and chloramphenicol (34 μg/ml) was inoculated with 250 μl of an overnight culture of E. coli containing either pET32::HaGAS1, pET32::HaGAS2, or pET32::HaCS and grown at 37°C to an OD of 0.5 – 0.7 (600 nm). Cultures expressing HaGAS1 or HaGAS2 were shifted to 30°C over 30 min, induced with 0.5 mM isopropyl-β-d-thiogalactopyranoside (IPTG) and incubated for 4 h at 30°C (220 rpm). Cultures expressing HaCS were treated in the same way, but were shifted to 10°C over 45 min, induced with 0.1 mM IPTG and cultivated for a further 24 h at 10°C (180 rpm).
E. coli cells were concentrated by centrifugation (9,000 g, 5 min, 4°C) and proteins were extracted using 5 ml BugBuster Protein Extraction Reagent (Novagen) per g bacteria cells. 1 μl Benzoase (Novagen) per ml was added and the assay was incubated at room temperature (RT) for 20 min in a shaker (100 rpm). Insoluble proteins and cell fragments were removed by centrifugation (9,000 g, 25 min, 4°C) and the supernatant was cleared using a 0.45 μm filter. To every 5.0 ml protein extract 1.0 ml pre-equilibrated Ni-NTA agarose (Novagen) was added and incubated at 4°C for 60 min in a shaker. The complete slurry was transferred to chromatography column and the supernatant with the unbound proteins was removed. The agarose was washed twice with ice-cold wash buffer (50 mM NaH2PO4, 300 mM NaCl, 20 mM imidazol, pH 8.0), followed by a single washing step with 100 mM imidazol (same buffer as before). For elution of the 6x-His Tag proteins ice-cold 250 mM imidazol was used (same buffer as before). Soluble proteins were analyzed by SDS-PAGE.
Proteins were concentrated by ultracentrifugation (21,000 g, 20 min, 4°C) using Vivaspin 500 columns (exclusion size 30 kDa, Sartorius AG, Göttingen, Germany). The concentrated samples were desalted and diluted twice with enzyme assay buffer (ESB; 15 mM MOPSO, 10% glycerol, 1 mM ascorbic acid, 10 mM MgCl2, 1 mM MnCl2, pH 7.0; modified from Bennett et al.  and Bertea et al. ). After each step, the concentrated samples were diluted with ESB.
Functional characterization was performed in 2 ml reaction tubes. Between 200 and 400 μg recombinant protein dissolved in ESB was used within a reaction volume of 750 μl. Farnesyl diphosphate (FDP, Sigma-Aldrich GmbH, München, Germany) was added in a final concentration of 50 μM. The reaction assay was carefully overlaid with 250 μl pentane and incubated for 60 min at 30°C in a shaker (100 rpm). Afterwards, the assay was extracted by vigorous shaking, the pentane layer was removed and the reaction assay was extracted with another 250 μl pentane followed by extraction with 500 μl pentane/diethyl ether (1:4, v/v). The pentane extracts were combined and dried over a short column of aluminium oxide overlaid with MgSO4 in a Pasteur pipet. The pentane/diethyl ether extract was also passed over the aluminium oxide column. The column was washed with 1.5 ml diethyl ether. All extracts were combined and carefully concentrated to 50 μl under a constant nitrogen flow. For GC-FID measurements, 1 μl of the concentrated sample was directly injected. For GC-MS measurements, 25 μl of the sample were diluted with 375 μl pentane and 1 μl was injected.
For determination of catalytic properties, the recombinant fusion-proteins were digested with enterokinase to obtain the native protein. Digest was performed for 16 h at 20°C. Subsequently enterokinase was removed following the recommendations of the Enterokinase Cleavage Capture Kit (Novagen). To remove the cleaved thioredoxin fusion-part, affinity-chromatography on Ni-NTA agarose was performed. The flow-through, containing the native enzyme, was concentrated and diluted in ESB as described above and the proteins were analysed by SDS-PAGE. Protein concentrations were determined by the Bradford method . Uncleaved thioredoxin-HaGAS1 fusion-protein was also used for determination of biochemical characteristics for comparison with the native HaGAS1 protein. Appropriate enzyme concentrations and incubation times were determined so that the reaction velocity was linear during the incubation time using 5 μM FPP (Sigma-Aldrich) spiked with [1-3H]FPP (Perkin Elmer, Rodgau-Jügesheim, Germany, 26.2 Ci/mmol).
A standard assay for determining biochemical properties was carried out in a final volume of 50 μl with 0.05 to 0.2 μg purified protein. The reactions were carefully overlay with 900 μl of dodecane and incubated for 15 min at 30°C in a thermoshaker-incubator (Thriller, Peqlab GmbH, Erlangen) at 300 rpm. Reactions were stopped by the addition of 50 μl of a solution containing 4 M NaOH and 1 M EDTA. To extract sesquiterpenes, the assays were vortexed for 1 min, centrifuged and 500 μl of the dodecane overlay was removed and mixed with 9.5 ml of liquid scintillation cocktail (Ultima Gold F, Perkin Elmer). Total radioactivity of the reaction products was determined using liquid scintillation counting (Wallac 1411 Liquid Scintillation Counter, Perkin-Elmer).
For pH optimum evaluation, assays were carried out using MES (pH 5.5, 6.0) MOPSO (pH 6.5, 7.0, 7.5) or Bis-tris propane (pH 8.0 to 9.5). These assays were done in duplicate. For determination of enzyme kinetics the concentration of FPP was varied from 0.125 to 30 μM with a fixed ratio of [1-3H]FPP. Ten different concentrations of FPP were used for each enzyme. Each concentration was done in triplicate. Calculation of the apparent Km value were obtained by Lineweaver-Burk plot analysis using Enzyme Kinetics!Pro software (ChemSW, Fairfield, USA).
Protein expression in vivo in S. cerevisiae
For in vivo expression of the sesquiterpene synthases, the genes were cloned into the pESC-Leu2d plasmid . HaGAS1 was amplified by PCR with the forward primer 5'-ACG TGC GGC CGC GAA CAT GGC AGC AGT TGG AGC CAG TG-3' and the reverse primer 5'-ACG TAG ATC TTT ACA TGG GTG AAG AAC CAA CAA ACA A-3'. HaGAS2 was amplified using the primer pair: forward primer 5'-ACG TCT CGA GAA TGG CAG CAG TTG GAG CCA GTG-3' and reverse primer 5'-ACG TGC TAG CTT ACA TGG GTG AAG AAC CAA CAA ACA A-3'. To generate the insert for HaCS, a PCR amplicon was generated using the forward primer 5'-ACG TCT CGA GAA TGG CAA CAA CTG AAG CTA ACA-3' and the reverse primer 5'-ACG TGC TAG CTT ACA TGG GGA CTG GAA CAC A-3'. The pET32::HaGAS1 and pET::HaCS plasmids served as template for the generation of the amplicons for HaGAS1 and HaCS. HaGAS2 was amplified from cDNA. To generate the thioredoxin fusion construct for HaCS in a yeast expression vector, the pET32 plasmid containing HaCS as a thioredoxin fusion-protein was amplified with the forward primer 5'-ACG TGG ATC CAA CAT GAG CGA TAA AAT TAT TCA C-3' and the reverse primer 5'-ACG TGC TAG CTT ACA TGG GGA CTG GAA CAC A-3'. As no germacrene A standard was available, the previously characterised germacrene A synthase (LsLTC2) from Lactuca sativa  was amplified from lettuce cDNA using the primer pair: forward primer 5'-ACG TGG ATC CAA CAT GGC AGC AGT TGA CAC TAA TG-3' and reverse primer 5'-ACG TGC TAG CTT ACA TGG ATA CAG AAC CAA C-3'. PCR reactions were performed using Phusion DNA Polymerase (New England Biolabs). HaGAS2, HaCS, and LsLTC2 were cloned into MCS2 of the pESC-Leu2d plasmid, HaGAS1 was cloned into MCS1 of the pESC-Leu2d plasmid. All amplicons were digested with the appropriate restriction enzymes overnight at 37°C, gel purified, and ligated into the pESC-Leu2d plasmid.
For all in vivo expression experiments, EPY300 S. cerevisiae cells were used. These cells were engineered for a high level of FDP production [51, 52]. EPY300 were transformed with purified plasmids following the protocol from Gietz & Woods . For protein expression, 5 ml SC medium (without Met, His, Leu and with 2% glucose supplement) were inoculated with single colony and grown overnight at 30°C (200 rpm). 250 ml culture flasks containing 50 ml YPAD medium (0.2% glucose, 1.8% galactose) supplemented with 1 mM methionine were inoculated with 1 ml overnight culture and overlaid with 5 ml of dodecane (Sigma-Aldrich GmbH, München, Germany). After 3–4 days incubation at 30°C (200 rpm), the cultures were transferred to 50 ml falcon tubes and centrifuged at 10,000 g (5 min). The dodecane overlay was carefully removed, diluted 1:100 with ethyl acetate and used directly for GC-FID and GC-MS analyses.
Identification of products of enzyme expression in vivo in S. cerevisiae and of in vitroassays
GC-MS analyses of terpenes produced by recombinant enzyme in S. cerevisiae were performed on an Agilent 6890N gas chromatography system coupled to an Agilent 5975B mass spectrometer. In vitro assays with recombinant HaGAS1 and HaGAS2 protein, produced by heterologous expression in E. coli, were analyzed on an Agilent 6890N gas chromatograph coupled to an MS5973 mass spectrometer (Agilent). 1 μl samples were injected at 250°C and analysed on a HP-5MS column (30 m × 250 μm i.d. × 0.25 μm film thickness, Agilent). Helium (constant flow rate of 1 ml min-1) was used as carrier gas. The temperature program was 40°C for 1 min followed by a linear gradient of 10°C min-1 to 250°C. Terpenes produced by in vitro assays with heterologously expressed HaCS were analyzed on a GC3400 gas chromatograph (Varian GmbH, Darmstadt, Germany) coupled to a Saturn 4D ion-trap mass spectrometer (Varian). 1 μl injection at a temperature of 250°C, He as carrier gas with 1.0 ml min-1. Temperature program: 50°C for 2 min, then 10°C min-1 to 300°C, which was held for 3 min. Spectra were interpreted with NIST02 Mass Spectra Library (Wiley & Sons, Mississauga, Canada). Reference compounds were obtained from Sigma-Aldrich (δ-cadinene) or Fluka (β-Caryophyllene, Buchs, Swiss). α-copaene, α-humulene and α-muurolene were identified by comparison with characterized compounds of Aloysia sellowii oil (generous gift from J. Degenhardt, Jena). Alkane standard (C8-C20, Fluka) was used to determine retention indices.
For analysis of sesquiterpenes present in glandular trichomes, approximately 10,000 glandular trichomes were extracted with pentane (1 ml) for 10 min. The pentane extract was carefully concentrated using nitrogen gas and analysed by GC-MS. Headspace trap experiments for detection of emitted terpenes were done using solid phase micro extraction (SPME). Two young flower heads were placed in an Erlenmeyer flask covered with aluminium foil. After one hour the SPME fibre was placed in the flask together with the flower heads for 60 minutes and trapped volatiles were analysed by GC-MS (data not shown).