Cloning and Characterization of δ-Guaiene Synthase Genes Encoding a Sesquiterpene Cyclase from Aquilaria microcarpa Cell Cultures

Three cDNA clones encoding δ-guaiene synthase, a sesquiterpene cyclase, were isolated from tissue cultures of Aquilaria microcarpa, and data mining analysis of the orthologous genes suggested that 10 and 9 amino acid residues of Nand C-terminal ends of the translated products of these clones remained undefined. The recombinant enzyme proteins, to which the putative missing Nand C-terminal amino acid sequences (MSSAKLGSAS and ALLRHAIEI, respectively) were ligated, exhibited the catalytic activities of sesquiterpene biosynthesis. Among these three δ-guaiene synthases, two isoforms were capable of liberating α-guaiene, δ-guaiene, β-elemene plus α-humulene as a minor product, while remaining one isoenzyme generated α-, δ-guaiene and β-elemene but not α-humulene. Although the enzyme protein solely lacking in the N-terminal 10 amino acid residues was capable of synthesizing the sesquiterpenoids, the protein without 9 amino acids at Cterminal did not exhibit the catalytic activity. These results suggest that two types of δ-guaiene synthase; α-, δ-guaiene, β-elemene-producing type, and α-, δ-guaiene, β-elemene plus α-humulene-producing type; concomitantly occur in A. microcarpa cell cultures, and several amino acid residues at C-terminal of the synthase protein are essential to exhibit the catalytic activities as the sesquiterpene cyclase.


Introduction
It has been widely known [1] that Aquilaria, and Gyrinops species sometimes form dark resinous heartwood, called agarwood, in the trunk of the plants in response to mechanical wounding or microbial infection.Agarwood produces a variety of sesquiterpene compounds with aroma, and, therefore, it has been used as the scent, perfume and traditional medicines [1].These highly valuable tissues are not formed under healthy conditions, and artificial transformation of these trees to agarwood has been very difficult.Okudera and Ito, however, reported [2] that biosynthetic activities of several sesquiterpenoids, such as δ-guaiene, α-guaiene and α-humulene, are induced in cell cultures of A. crassna by the treatment with methyl jasmonate or yeast extract.Guaianolide sesquiterpene compounds are thought to be synthesized via two cyclization reactions [3]: i) C1-to-C10 cyclization generating germacrene-like macrocyclic intermediate; and ii) C2-to-C6 cyclization forming guaiane-skeleton (Figure 1).A part of the macrocyclic intermediate is converted to α-humulene or germacrene A without the second cyclization.It has been also demonstrated [4] [5] that β-elemene is an artificially formed compound by the pyrolysis of germecrene A. Recently, several cDNA clones encoding δ-guaiene synthase, a sesquiterpene cyclase, have been isolated from A. crassna and A. sinensis [3] [6].Catalytic specificities of the translated proteins of these genes have been extensively studied, and it has been shown that the enzymes liberate multiple sesquiterpene compounds, such as δ-guaiene, α-guaiene, β-elemene and α-humulene, from farnesyl diphosphate as the substrate.However, only very little is known about the molecular basis of the enzymatic reaction of the synthase generating multiple products.
We have recently shown [5] [7] that treatment of cell cultures of A. microcarpa with methyl jasmonate or yeast extract results in the marked transcriptional activation of δ-guaiene synthase gene, and a cDNA clone (GS-1, GenBank accession No. KF800046) encoding the enzyme protein has been isolated [5].It has been also demonstrated that GS-1 translate catalyzes the formation of δ-guaiene, α-guaiene and β-elemene, but not α-humulene [5].In the present experiments, we have isolated three additional clones of the synthase gene although several amino acid residues at both N-and C-terminal ends of the translated proteins remain undefined.In order to understand the inducible biosynthetic activities of A. microcarpa liberating multiple sesquiterpene compounds, we have ligated putative missing amino acid sequences to N-and C-terminals of the recombinant proteins, and examined the catalytic properties of these 'deficiencies supplied' δ-guaiene synthases.

Cloning of δ-Guaiene Synthase Genes from A. microcarpa Cell Culture
Seeds of A. microcarpa were generous gift from Professor Shigetoshi Kadota, Institute of Natural Medicine, University of Toyama, and they were germinated and grown in a greenhouse at Experimental Station for Medi- cinal Plant Research of our university.Sterilized young leaf tissues of the plant were placed on Murashige and Skoog's agar medium [8] supplemented with 3% (w/v) sucrose and 3 μM of 2,4-dichlorophenoxyacetic acid, and callus formation was observed in the leaf veins after 4 weeks.The dedifferentiated cells were then incubated on the medium in the presence of 3 μM of 2,4-dichlorophenoxyacetic acid and 3 μM of N 6 -benzyladenine at 26˚C under darkness.After establishment of the callus culture, the cells were transferred onto a fresh medium every 3 weeks as described previously in detail [9] [10].Total RNA was isolated from yeast extract-treated cultured cells (approximately 100 mg) with RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) [5] [9], and cDNA templates were generated by reverse transcription reaction using Transcriptor First Strand cDNA Synthesis Kit (Roche, Basel, Switzerland).The cDNA fragments presumably encoding the translatable region of δ-guaiene synthase were isolated by PCR amplification employing the primer pair designated from the nucleotide sequence of the homologous gene previously isolated [5] (5'-ATG TCT TCG GCA AAA CTA GGT TCT GCC TCC-3' as the forward and 5'-TCA GAT TTC AAT AGC ATG ACG CAA CAA GGC-3' as the reverse primer).The amplified cDNAs were subcloned into the pGEM-T Easy vector (Promega, Fitchburg, WI, USA), and the nucleotide sequences were determined on both strands using the dye-terminator method on a PRISM 3130 Genetic Analyzer (Applied Biosystems, Waltham, MA, USA).

Preparation and Purification of Recombinant δ-Guaiene Synthase Proteins
The modified δ-guaiene synthase genes, GS-2, 3 and 4, in which undefined sequences of the translatable regions were ligated, were amplified using a common primer pair (5'-ATG TCT TCG GCA AAA CTA GGT TCT GCC TCC-3' as the forward and 5'-TCA GAT TTC AAT AGC ATG ACG CAA CAA GGC-3' as the reverse primer, 1644 mer as the product).Amplification of GS-2ΔNΔC, the undefined regions-omitted gene, was performed with 5'-GAA GAT GTT AGC CGC CGA GAT GCC-3' as the common forward primer and 5'-TCA GGC AAT TCT ATC CTT GGT CAC-3' as the gene specific reverse primer employing pGEM-T Easy vector harboring GS-2 as the template.GS-3ΔNΔC and GS-4ΔNΔC were similarly amplified by the combination of the common forward primer and the gene specific reverse primer (5'-TCA GGC AAT TCT ACC CTT GGT CAC-3' for GS-3ΔNΔC, and 5'-TCA GGC AAT TCT ATC TTT GGT CAC-3' for GS-4ΔNΔC, respectively).GS-4ΔN and GS-4ΔC were also amplified by the appropriate combinations of the common primer and the specific primers (5'-GAA GAT GTT AGC CGC CGA GAT GCC-3' and 5'-TCA GAT TTC AAT AGC ATG ACG CAA CAA GGC-3' for GS-4ΔN while 5'-ATG TCT TCG GCA AAA CTA GGT TCT GCC TCC-3' and 5'-TCA GGC AAT TCT ATC TTT GGT CAC-3' for GS-4ΔC) using GS-4 subcloned into pGEM-T Easy vector as the template.
Over expression of δ-guaiene synthase genes in Escherichia coli was performed using the E. coli Expression System with Gateway Technology (Invitrogen, Waltham, MA, USA).Desired regions of the synthase genes amplified by PCR were successively subcloned into pENTR and pDEST15 (Invitrogen, Waltham, MA, USA), and then, E. coli BL21 cells were transformed with the constructed expression vectors.The cell cultures were grown in LB medium overnight at 25˚C, and, after transfer into the fresh medium, isopropyl β-D-1-thiogalactopyranoside (IPTG, final concentration 0.1 mM) was added to the cultures at an optical density of 0.4 -0.5 at 600 nm.The cultures were further incubated at 25˚C for 5 h.δ-Guaiene synthase proteins were recovered as the glutathione-S-transferase (GST) tagged form, and were purified by an affinity chromatography on a Glutathione-Sepharose 4B column (GE Healthcare Life Sciences, Little Chalfont, UK) according to the instruction manual.Protein concentrations were determined by the method of Bradford [11], and, if necessary, the purified enzyme solutions were appropriately concentrated by Amicon Ultra-15 (Merck Millipore, Billerica, MA, USA).The samples were, then, subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE, 10% gel) according to the method of Laemmli [12], and the separated proteins were stained with Coomassie brilliant blue.

Catalytic Activities of Recombinant δ-Guaiene Synthase Proteins
In order to examine sesquiterpene biosynthetic activities of the recombinant δ-guaiene synthases, the purified enzyme proteins were incubated with farnesyl diphosphate as the substrate, and the reaction products were analyzed by GC-MS.The assay mixtures (200 μl) contained 50 mM Tris-HCl (pH 7.0), 20 mM MgCl 2 , 10 mM dithiothreitol, 46 μM farnesyl diphosphate (Sigma-Aldrich, St. Louis, MO, USA), 10% (v/v) glycerol, purified GST-δ-guaiene synthase proteins (approximately 10 μg) and limonene (0.2 μg, nacalai tesque, Kyoto, Japan) as an internal standard.In a control experiment, the enzyme proteins were heat denatured (95˚C, 10 min) prior to the assay.The mixtures were incubated at 25˚C for 4 h, and the reaction products were extracted using headspace solid phase microextraction assembly with a 100 µm polydimethylsiloxane fiber (Supelco, Bellefonte, PA, USA).The fiber was exposed to the sample headspace for 30 min, and, after adsorption of the reaction products, sesquiterpene compounds were analyzed by GC-MS (Shimadzu, GCMS-QP5000; Agilent, J&W DB-1ms 0.32 mm × 30 m column) in the splitless mode.The injection port and the interface temperature were adjusted at 250˚C, and the mass range was set from m/z 40 to 400.The carrier flow (He) was at 1.4 ml/min.The oven temperature was started at 50˚C for 2 min and was increased to 170˚C at the rate of 4˚C/min, then raised to 300˚C at 30˚C/min.The reaction products were separated under the above conditions to calculate retention indices, and the identification of the compounds was based on the comparison of reported retention indices [13] [14] and mass spectra with those in the NIST02 and Wiley 7 MS databases.

δ-Guaiene Synthase Genes of A. microcarpa Cell Culture
It has been demonstrated [3] [6] that sesuqiterpene biosynthetic genes encoding δ-guaiene synthase in Aquilaria plants are usually organized as the multigene family.To our knowledge, seven orthologues of the synthase genes have been isolated from Aquilaria plants (A.crassna, A. sinensis and A. microcarpa), and the translated products of these genes are found to be highly conserved.All of the synthase proteins consist of 547 amino acid residues and the sequences also show very high homology (98% -99% identity) [3] [5] [6].In order to isolate δ-guaiene synthase genes other than GS-1 from A. microcarpa cell cultures, we employed the primer pair corresponding to the N-and C-terminal ends of the translatable region of GS-1 (30 nucleotides each, including stop codon for C-terminal).Three new clones presumably encoding δ-guaiene synthase were isolated (GS-2ΔNΔC, GS-3ΔNΔC and GS-4ΔNΔC; GenBank accession No. KT283579, KT283580 and KT283581, respectively), and the putative amino acid sequences of the translated products showed high homology with GS-1 (Figure 2).Two specific motifs, the double arginine residue (RRX8W) at the N-terminal end and the aspartate rich motif (DDXXD) in the reaction center of the protein [15] [16], were observed in these three clones, and replacement of amino acid residues toward GS-1 was found at only three positions in GS-2, four in GS-3 and eight in GS-4, respectively (Figure 2).
Since nucleotide sequences corresponding to N-and C-terminal ends of GS-1 were employed as primers for PCR amplification, 10 and 9 amino acid residues of the terminals remained to be defined.A number of attempts were made to elucidate the undefined translatable and untranslatable regions of these three δ-guaiene synthase genes, and the rapid amplification of cDNA ends method have been systematically performed to obtain the DNA fragments.However, these experiments have not succeeded possibly because only very few nucleotides are replaced in the newly isolated three clones.

Preparation and Purification of Recombinant δ-Guaiene Synthase Proteins
Among seven δ-guaiene synthases reported from Aquilaria [3] [5] [6], six enzyme proteins share the common N-terminal sequence MSSAKLGSAS while remaining one is MSSAKLGSTS (Table 1).On the other hand, C-terminal sequence of all of the seven proteins is identical (ALLRHAIEI).Therefore, we ligated the most probable "missing" amino acid sequences to N-and C-terminal ends, and catalytic properties of these deficiencies-supplied enzymes (GS-2, GS-3 and GS-4) were examined together with the terminals-lacking translates of GS-2ΔNΔC, GS-3ΔNΔC and GS-4ΔNΔC.The recombinant enzymes were obtained as GST-tagged proteins, and were purified by an affinity chromatography on a Glutathione-Sepharose 4B column.In SDS-PAGE analysis (Figure 3), the protein bands corresponding to the position of GST-fused GS-2, 3 and 4 in IPTG-treated cell extracts (approximately 91.5 kDa) were shown to be sufficiently purified.The recombinant enzymes lacking in the terminal structures were also prepared as GST-tagged forms and purified in a similar manner.

Catalytic Activities of Recombinant δ-Guaiene Synthase Proteins
The recombinant δ-guaiene synthases of A. microcarpa were incubated with farnesyl diphosphate, and the reaction products were analyzed by GC-MS.We have previously reported [5] that translate of GS-1 liberates δ-guaiene, α-guaiene and germacrene A as the reaction products.We have also demonstrated [5] that germa-

GS-1 MSSAKLGSASEDVSRRDANYHPTVWGDFFLTHSSNFLENNHSILEKHEELKQEVRNLLVVETSDLPSKIQLTDKIIRLGV 80 GS-2ΔNΔC ----------EDVSRRDADYHPTVWGDFFLTHSSNFLENNHSILEKHEELKQEVRNLLVVETSDLPSKIQLTDKIIRLGV 70 GS-3ΔNΔC ----------EDVSRRDANYHPTVWGDFFLTHSSNFLENNDNILEKHEELKQEVTNLLVVETSDLPSKIQLTDEIIRLGV 70 GS-4ΔNΔC
----------EDVSRRDANYHPTVWGDFFLTHSSNFLENNDNILEKHEELKQEVRNLLVVETSDLPSKIQLTDKIIRLGV 70 crene A is easily converted to β-elemene by pyrolysis, and, therefore, under the usual GC-MS conditions, together with α-and δ-guaiene, β-elemene was observed as the apparent product of the enzymatic reaction instead of germacrene A. As shown in Figure 4(a), GS-2 produced at least three sesquiterpene compounds, and peak numbers 1, 2 and 4 were assigned to be β-elemene, α-guaiene and δ-guaiene by the retention indices and mass spectra (Figure 5) in the data base, respectively [13] [14].In sharp contrast, detectable amount of the reaction products was not observed when the assay was carried out with the terminals sequences-omitted form, GS-2ΔNΔC (Figure 4(a)).Heat denatured enzyme proteins also did not show the catalytic activity (data not shown).GS-3 was also capable of producing β-elemene, α-guaiene and δ-guaiene as the major products, however, unlike in GS-1 [5] and GS-2, this protein was found to liberate α-humulene (peak 3) as an additional minor product (Figure 4(b) and Figure 5).As well as GS-2ΔNΔC, GS-3ΔNΔC did not exhibit the sesuqiterpene biosynthetic activity.GS-4 showed the catalytic specificity essentially similar to that of GS-3, and this enzyme produced β-elemene, α-guaiene, δ-guaiene and low amount of α-humulene (Figure 4(c) and Figure 5).As was in other δ-guaiene synthases, GS-4ΔNΔC was shown to be an inactive protein.In Figure 4(a), it appeared that a minor peak could be observed at the retention time similar to α-humulene (peak 3), and, therefore, we carefully examine the possible formation of this sesquiterpene compound in GS-2-catalyzed reaction.However, based on the mass spectra analysis, it was revealed that the apparent minor peak was due to the formation of unknown compound but not α-humulene.
In order to elucidate whether or not both N-and C-terminal structures of δ-guaiene synthase are essential to exhibit the catalytic activity, 10 amino acid residues at N-terminal or 9 amino acids at C-terminal of the enzyme were selectively omitted, and the catalytic activities of these proteins were examined.In this set of the experiments, GS-4 which showed the highest recovery in the protein purification processes was selected, and N-terminal-lacking and C-terminal-lacking enzyme proteins (GS-4ΔN and GS-4ΔC) was prepared as the GST-fused form.As well as GS-4, N-terminal lacking enzyme GS-4ΔN exhibited the catalytic activity, if somewhat reduced, and the peaks of the sesquiterpene compounds were clearly detected (Figure 6).In sharp contrast, GS-4ΔC apparently did not show the enzymatic activity, and no peak of the reaction product was observed as far as tested.These results strongly suggest that 9 amino acid residues at C-terminal of δ-guaiene synthase proteins should be an essential structure to function as the sesquiterpene cyclase.

Discussion
As shown in Figure 1, α-and δ-guaiene are thought to be synthesized from germacrene A as an intermediate.On the other hand, α-humulene biosynthesis shares the common pathway with guaienes-forming processes only in the early stage.Kumeta and Itoh [3] isolated three δ-guaiene synthase genes from A. crassna, and showed that the translated proteins of all of these genes generate δ-guaiene, α-guaiene plus α-humulene as the minor product.In contrast, Xu et al. have recently reported [6] that three homologous genes of δ-guaiene synthase were isolated from A. sinensis, and the translated proteins liberate δ-guaiene, α-guaiene and β-elemene but not α-humulene.Therefore, it is assumed that two classes of δ-guaiene synthase would occur in Aquilaria plants; δ-guaiene, αguaiene and β-elemene producing group (A.sinensis), and δ-guaiene, α-guaiene and α-humulene producing group (A.crassna).We have previously reported [5] that GS-1, the translated product of a δ-guaiene synthase gene isolated from A. microcarpa cell culture, is capable of producing δ-guaiene, α-guaiene and β-elemene, however, the formation of α-humulene by GS-1 was not observed.In the present experiments, we have isolated three additional cDNA clones encoding the partial structures of δ-guaiene synthase (GS-2ΔNΔC, GS-3ΔNΔC and GS-4ΔNΔC) from A. microcarpa (Figure 2).We have also shown that the modified proteins, in which probable missing amino acid residues at N-and C-terminals were supplied (GS-2, GS-3 and GS-4), showed unique catalytic properties (Figure 4 and Figure 5).The reaction products of GS-2 were α-guaiene, δ-guaiene and β-elemene, and liberation of α-humulene was not observed.Therefore, catalytic specificity of GS-2 was found to be similar to that of GS-1 isolated from A. microcarpa [5] and three δ-guaiene synthases from A. sinensis [6].In contrast, GS-3 and GS-4 produced α-and δ-guaiene, β-elemene plus α-humulene, and, therefore, characteristics of the enzyme reaction of GS-3 and GS-4 should resemble the synthases from A. crassna [3].These observations imply that two classes of δ-guaiene synthase, α-humulene producing and non-producing types, would concomitantly occur in A. microcarpa cells.In order to clarify the molecular bases of these unique specificities of δ-guaiene synthases in Aquilaria plants, preparation and characterization of several mutant proteins of the synthase are in progress in our laboratory in which the non-conserved amino acids presented in Figure 2 are appropriately replaced.
Although recombinant GS-2, 3 and 4 proteins showed the biosynthetic activity of the sesquiterpene compounds, N-and C-terminals-deleted proteins, GS-2ΔNΔC, GS-3ΔNΔC and GS-4ΔNΔC, did not catalyze the terpenoids production (Figure 4).However, GS-4ΔN, deleted solely N-terminal end, clearly showed the enzymatic activity (Figure 6) while GS-4ΔC without C-terminal was found to lose the catalytic function.These observations strongly suggest that several amino acid residues at C-terminal end of GS-4 protein are essential to exhibit the terpene cyclase activity.We are attempting to construct the 3D model of the enzyme protein employing the suitable related protein(s) as the template for the elucidation of the biochemical functions of C-terminal amino acid sequence of δ-guaiene synthase.

Figure 3 .
Figure 3. SDS-PAGE analysis of production and purification of δ-guaiene synthase proteins.GS-2, GS-3 and GS-4 were overexpressed in E. coli cells, and the translated products were obtained as GST-fusion proteins.The crude extracts prepared from non-treated control, IPTG-treated cells and the purified proteins were analyzed by SDS-PAGE, and the expected position of GST-fused δ-guaiene synthase (approximately 91.5 kDa) was indicated by an arrow.

Figure 5 .
Figure 5. Mass spectra of the compounds corresponding to the peaks shown in Figure 4.

Table 1 .
Amino acid sequences of N-and C-terminal ends of δ-guaiene synthases isolated from Aquilaria plants.