To identify and localize GES activity, crude protein extracts were prepared from whole leaves and from isolated glands (prepared according to Gang et al., 2001) and assayed for monoterpene synthase activity with geranyl diphosphate as the substrate. Because small young leaves contained the highest levels of geraniol and related compounds, we used such leaves as the source of the glands. Crude protein extracts from glands produced >99% geraniol upon incubation with geranyl diphosphate (Fig. 1C). In particular, no nerol product was observed. GES-specific activity in the gland extracts was more than 50 times higher than in whole-leaf extracts.

BLAST searches with the complete sequence of GES indicate that compared with proteins of known function, GES shares the highest sequence similarity (i.e. percentage identity of amino acid residues) with 1,8-cineole synthase from sage (Wise and Croteau, 1999) and 4S-limonene synthase from spearmint (Colby et al., 1993), both monoterpene synthases from Lamiaceae species (Fig. 5A). However, the overall sequence identity of GES to either one is only approximately 30%, whereas the latter two enzymes are approximately 48% identical to each other. A phylogenetic analysis employing the nearest neighbor-joining method suggests that GES occupies a highly divergent branch of the terpene synthase family, but it most likely shares a most recent common origin with the snapdragon ocimene and myrcene synthases and with C. breweri linalool synthase (Dudareva et al., 1996, 2003), all terpene synthases catalyzing the formation of acyclic monoterpenes (Fig. 5B).

Basil leaves were categorized into three sizes: small (0.5-1.5 cm), medium (1.5-3 cm), and large (3-4 cm; Gang et al., 2001). Two hundred milligrams of each leaf was added to liquid N₂ and ground by mortar and pestle. The powder was soaked in 2 mL of methyl t-butyl ether containing 0.1 mg of linalool as an internal standard (linalool was used because basil cv Sweet Dani does not contain linalool) and extracted for 2 h at room temperature in 5-mL glass vials with tightly sealed rubber septa caps. The methyl t-butyl ether upper layer, which included the volatile oil, was removed and placed into another vial and concentrated to 200 μL under gentle N₂ gas flow for GC-MS analysis. Data points were obtained in triplicate.

Volatile oils were extracted from the glands of young leaves with a stretched glass pipette as previously described (Gang et al., 2001).

A Shimadzu QP-5000 system (Shimadzu, Columbia, MD) equipped with Shimadzu GC-17 gas chromatograph was used for GC-MS analysis of volatile compounds. Separation was performed on DB-WAX (30-m × 0.32-mm-i.d. × 0.25-μm film thickness, J&W Scientific, Folsom, CA) capillary column with electron impact mode (1.4 kV). However, some nonpolar compounds eluted with the solvent peak in the DB-WAX column, and their separation was achieved on a CP-5 column (30-m × 0.32-mm-i.d. × 1-μm film thickness, Alltech Associates Inc., Deerfield, IL). The oven temperature for DB-WAX methods was held at 60°C for 2 min and raised to 220°C at 4°C min^-1 with the injector set at 220°C and the interface set at 240°C. The GC condition for the CP-5 method was the same as the previous report (Gang et al., 2001). Ultrapure helium was used as the carrier gas at a rate of 1.3 mL min^-1. Samples (2 μL) were injected by the Shimadzu AOC-17 Autoinjector. Eluted compounds were identified by comparing their retention time and mass fragmentation patterns with standard compounds.

GES activity was assayed by incubating 5 μL of the enzyme sample in a final volume of 50 μL of buffer containing 50 mm HEPES-KOH (pH 8.0), 1 mm dithioerythritol, 0.5 mm MnCl₂, 20 mm MgCl₂, 10% (w/v) glycerol, and 0.025 μm [1-³H]-geranyl diphosphate (specific activity 20 Ci mol^-1, American Radiolabeled Chemicals, St. Louis). After incubation for 30 min at 32°C, 160 μL of hexane was added to the tube, vortexed briefly, and centrifuged to separated the phases. The hexane layer was directly placed into a scintillation vial containing 2 mL of nonaqueous scintillation fluid (Econo-Safe, Research Products International, Mount Prospect, IL). This extraction procedure was repeated twice, and the total hexane phase was counted by a liquid scintillation counter (LS-6500 model, Beckman Coulter, Fullerton, CA). Boiled enzyme extracts were used as controls.

GES enzyme assays were also performed by adding 100 μL of enzyme solution with 900 μL of assay solution containing 54 μm nonradioactive geranyl diphosphate (Echelon Research Laboratories, Salt Lake City) and the same buffer described above. The reactions were carried out in an 8-mL DuPont autosampler vial (DuPont-Dow Elastomers L.L.C., Wilmington, DE) with a white solid-top polypropylene cap (Alltech). After letting the reaction proceed for 2 to 4 h at 32°C, the liberated compounds were collected with an SPME device PDMS-100 with a polydimethylsiloxane fiber (Supelco, Bellefonte, PA) by inserting the fiber into the tube and leaving it in for 20 min at 42°C. After this incubation step, the SPME fiber was directly injected into the GC-MS.

Assays in buffer containing ¹⁸O-labeled water were carried out in 2-mL glass vials with screw cap of PTFE/Silicone Septa (Supelco) by the addition 20 μL of purified enzyme (approximately 1 μg of protein) to a 180-μL assay solution that contained 20 μL of 10× assay buffer with 150 μL of H₂¹⁸O (95% atom, Icon Service, Summit, NJ) and 10 μL of 5.4 mm geranyl diphosphate. The final concentration of H₂¹⁸O in this assay solution was 71.3% (w/v). This solution was incubated for 2 h at 32°C, cooled down on ice, and then extracted with 200 μL of pentane. After concentration to 50 μL, 2 μL of this solution was injected to GC-MS system. To compare the mass spectra pattern, a pentane extract of the product from a reaction in which normal water (H₂¹⁶O) was used was also analyzed.

Phosphatase activity was measured as described by Hernández and Whitton (1996) with the following modification: Assay samples were prepared by incubating 50 μL of enzyme solution in a final volume of 400 μL of assay buffer containing 2 mm p-nitrophenyl phosphate as substrate. The buffer composition was the same as with the GES assay (but without geranyl diphosphate). After incubation for 1 h at room temperature, the reaction was stopped by adding of 700 μL of 0.2 m Na₂CO₃. The yellow color generated from the hydrolysis of p-nitrophenyl phosphate was measured at 420 nm in a spectrophotometer (Beckman DU530). Phosphatase activity was calculated using a standard curve for p-nitrophenol. For the purified enzyme, this assay was scaled down to 10-fold.

All purification steps were carried out at 4°C unless stated otherwise. Glands were isolated from approximately 300 g of basil cv Sweet Dani, essentially following the procedures previously described by Gang et al. (2001) with a total yield of 4 mL of resuspended glands. The gland preparation was diluted 10:1 (v/v) in ice-cold enzyme extraction buffer (100 mm BisTris-HCl [pH 7.5], 5 mm dithioerythritol, 5 mm Na₂S₂O₄, 2% [w/v] polyvinylpolypyrrolidone, and 10% [w/v] glycerol), and sonicated on ice, with rest intervals for cooling down, until gland cells were completely lysed. After centrifugation for 20 min at 10,000g, the supernatant (39 mL) was loaded onto a DEAE-cellulose column (10 mL of DE53, Whatman, Fairfield, NJ) installed in a Pharmacia Biotech FPLC apparatus and pre-equilibrated with a solution containing 50 mm Tris-HCl (pH 7.5), 10% (w/v) glycerol, and 10 mm β-mercaptoethanol (buffer A). After elution of unbound material from the column with 25 mL of buffer A, GES activity was eluted with 200 mL of a linear gradient from 0 to 1 m KCl in buffer A. The flow rate was 1.0 mL min^-1, and 3-mL fractions were collected and then assayed for GES activity. The fractions with the highest GES activity were pooled (KCl concentration of 255-435 mm, a total of 38 mL) and dialyzed in buffer A for 4 h to remove KCl. This dialysis step did not result in any decrease in GES activity. The enzyme solution was subsequently loaded onto strong anion-exchange column (Mono Q, 0.5 × 6.0 cm, Pharmacia Biotech) pre-equilibrated with buffer A. After washing off the unbound material with 2 mL of buffer A, GES was eluted with 50 mL of a linear gradient of 0 to 700 mm KCl in buffer A at 0.5 mL min^-1, and 1-mL fractions were collected. The highest GES activity was detected in the 2-mL fraction containing 294 mm KCl. Octyl glucoside was added to this fraction (final concentration of 5 mm), and the enzyme was concentrated in an Ultrafree-4 centrifugal device (Millipore, Bedford, MA) to a total volume of 200 μL. The concentrated enzyme solution was loaded onto a size exclusion column (10 × 300 mm) packed with Superose 12 (Pharmacia Biotech), and active fraction was isocratically eluted with 100 mm KCl in buffer A at 0.2 mL min^-1. Fractions (0.5 mL each) were collected, and protein purity was examined by SDS-PAGE gel electrophoresis followed by Coomassie Brilliant Blue or silver staining of the gel. The protein concentrations were measured by the Bradford method or by staining intensity on SDS-PAGE compared with bovine serum albumin concentration standards.

Partially purified GES was run on a size exclusion column under the same conditions used during the purification procedure, except that 0.25-mL fractionations were collected instead, and fractions were assayed for GES activity. A standard curve was obtained by plotting the elution volume/void volume of the standard proteins against the log of the molecular mass. The protein standards used included cytochrome C (12.4 kD), carbonic anhydrase (29 kD), ovalbumin (45 kD), bovine serum albumin (66 kD), alcohol dehydrogenase (from equine liver, 80 kD), and alcohol dehydrogenase (from yeast [Saccharomyces cerevisiae], 141 kD). The subunit molecular mass was estimated by SDS-PAGE performed on 10% (w/v) polyacrylamide gel and calibrated with molecular mass standard in the range of 14 to 212 kD (New England Biolabs, Beverly, MA).

Mass spectrometric analysis of the purified GES was carried out in the Proteomics Core Facility (Southwest Environmental Health Sciences Center and Arizona Cancer Center, University of Arizona, Tucson). Proteins were first separated by SDS-PAGE, as described above, stained lightly with Coomassie Brilliant Blue R250, excised from the gel, and digested with trypsin (Shevchenko et al., 1996). Extracted peptides were analyzed by liquid chromatography (LC)-MS/MS using a ThermoFinnigan LCQ Classic quadrupole ion trap mass spectrometer (ThermoFinnigan, San Jose, CA) equipped with a Michrom MAGIC2002 HPLC (Michrom, Auburn, CA) and a nanospray ion source (University of Washington, Seattle). Peptides were loaded onto 10-cm capillaries (100-μm i.d., packed with 5-6 cm of Vydac C18 material) that were pulled to 3- to 5-μm tips using a Sutter Instruments P2000 capillary puller (Sutter Instruments, Novato, CA). Peptides were eluted at a flow rate of 200 to 300 nL min^-1 into the mass spectrometer using reversed phase solvent conditions (Shevchenko et al., 1996). Tandem MS spectra of peptides were analyzed with the TurboSequest program to assign peptide sequences to the spectra (Eng et al., 1994). TurboSequest analyses were performed against the sweet basil EST databases housed at the Arizona Genomics Institute (University of Arizona, Tucson). The nonidentified spectra were further analyzed by the ExPASy peptide mass program (http://us.expasy.org/tools/peptide-mass.html) for the calculated tryptic peptide masses from full length GES cDNA.

A basil cv Sweet Dani peltate gland EST database containing 3,200 unique sequences was developed at the Arizona Genomics Institute and the Arizona Genomics Computational Laboratory (University of Arizona, Tucson), using a cDNA library constructed from gland mRNAs as previously described (Gang et al., 2001). BLAST searches revealed numerous ESTs with sequence similarity to terpene synthases. Potential cDNAs encoding GES were examined by reverse transcriptase-PCR cloning of full-length cDNAs (complete sequence were obtained by 5′-RACE when necessary) into the pCRT7/CT-TOPO TA vector (Invitrogen, Carlsbad, CA), expressing these constructs in the E. coli expression system, and testing the resulting proteins for activity with geranyl diphosphate, as previously described (Chen et al., 2003). Constructs encoding truncated GES proteins missing the first 34 or 43 amino acids were also constructed, using the method described by Chen et al. (2003) in either the pCRT7/CT-TOPO TA vector or a pET-11a vector. After harvesting recombinant GES proteins from E. coli cells (Chen et al., 2003), the proteins were purified using the same method employed in the purification of GES from basil glands.

Alignment of multiple protein sequences was performed using the ClustalX program (Thompson et al., 1997). Sequence relatedness by the neighbor-joining method was determined using the protocol included in the ClustalX package. The phylogenic tree was drawn using the TREEVIEW program (http://taxonomy.zoology.gla.ac.uk/rod/treeview.html; Page, 1996).