Biosynthesis of t-Anethole in Anise: Characterization of t-Anol/Isoeugenol Synthase and an O-Methyltransferase Specific for a C7-C8 Propenyl Side Chain^,,

Abstract

The phenylpropene t-anethole imparts the characteristic sweet aroma of anise (Pimpinella anisum, family Apiaceae) seeds and leaves. Here we report that the aerial parts of the anise plant accumulate t-anethole as the plant matures, with the highest levels of t-anethole found in fruits. Although the anise plant is covered with trichomes, t-anethole accumulates inside the leaves and not in the trichomes or the epidermal cell layer. We have obtained anise cDNA encoding t-anol/isoeugenol synthase 1 (AIS1), an NADPH-dependent enzyme that can biosynthesize t-anol and isoeugenol (the latter not found in anise) from coumaryl acetate and coniferyl acetate, respectively. In addition, we have obtained a cDNA encoding S-[methyl-¹⁴C]adenosyl-l-methionine:t-anol/isoeugenol O-methyltransferase 1 (AIMT1), an enzyme that can convert t-anol or isoeugenol to t-anethole or methylisoeugenol, respectively, via methylation of the para-OH group. The genes encoding AIS1 and AIMT1 were expressed throughout the plant and their transcript levels were highest in developing fruits. The AIS1 protein is 59% identical to petunia (Petunia hybrida) isoeugenol synthase 1 and displays apparent K_m values of 145 μm for coumaryl acetate and 230 μm for coniferyl acetate. AIMT1 prefers isoeugenol to t-anol by a factor of 2, with K_m values of 19.3 μm for isoeugenol and 54.5 μm for S-[methyl-¹⁴C]adenosyl-l-methionine. The AIMT1 protein sequence is approximately 40% identical to basil (Ocimum basilicum) and Clarkia breweri phenylpropene O-methyltransferases, but unlike these enzymes, which do not show large discrimination between substrates with isomeric propenyl side chains, AIMT1 shows a 10-fold preference for t-anol over chavicol and for isoeugenol over eugenol.

The phenylpropenes are a class of volatile compounds found throughout the gymnosperms and angiosperms. When emitted from flowers, they serve as attractants for pollinators, which detect them through their olfactory systems. In addition, at high concentrations their general toxicity to cells renders them useful as defense compounds, and consequently they are found in vegetative tissues of many plant species, although perhaps due to their toxicity they are typically sequestered in specialized structures or quickly emitted from the plant. For example, eugenol and methylchavicol are synthesized and stored in glandular trichomes on the surface of leaves of sweet basil (Ocimum basilicum, Lamiaceae; Gang et al., 2001; Iijima et al., 2004), and eugenol, isoeugenol, and their methylated derivatives are synthesized and emitted from flowers of Clarkia breweri (Onagraceae; Wang et al., 1997). Because of the antimicrobial and aromatic properties of the phenypropenes, humans have historically used spices and herbs containing them as pesticides and food preservatives. t-Anethole, a main component of the essential oils of anise (Pimpinella anisum, Apiaceae) seeds, star anise (Illicium anisatum), and fennel (Foeniculum vulgare), has been used as a fumigant against mosquitoes (Chang and Ahn, 2002; Gross et al., 2002). Clove (Syzygium aromaticum) buds, which are rich in eugenol, have been used as a source of spice for millennia (Jirovetz et al., 2006).

The many compounds of the phenylpropene class differ from one another in two aspects: the moieties attached to the phenyl ring and the position of the double bond in the propenyl side chain (Fig. 1). For example, t-anol and chavicol both have a para-hydroxy functionality on the phenyl ring, but the double bond in the propenyl side chain in t-anol is between C7 and C8, whereas in chavicol it is between C8 and C9. On the other hand, the position of the double bond in the propenyl side chain is the same (C7 and C8) in t-anethole and methylisoeugenol, but the latter has an additional methoxy functionality at the meta-position of the phenyl ring as compared to t-anethole.

Figure 1.

Biochemical reactions leading to phenylpropenes in plants. The reactions catalyzed by known enzymes are indicated. The carbon numbering system used in the text is shown.

We have recently isolated two distinct NADPH-dependent enzymes, eugenol synthase (EGS) and isoeugenol synthase (IGS), which both utilize the same substrate, coniferyl acetate, and convert it to their respective product via a quinone methide intermediate (Koeduka et al., 2006, 2008; Louie et al., 2007). EGS and IGS proteins are approximately 50% to 95% identical to each other and are also closely related to several other NADPH-dependent reductases involved in phenylpropanoid biosynthesis in plants, including pinoresinol-lariciresinol reductase (PLR), isoflavone reductase (IFR), and phenylcoumaran benzyl ether reductase (PCBER). Of particular interest is the observation that there are two distinct lineages of EGS enzymes within this large family of reductases, whereas the two IGS enzymes previously characterized, from petunia (Petunia hybrida, Solanaceae) and C. breweri, are more closely related to each other as well as to one of the two lineages of EGS sequences (Koeduka et al., 2008). So far, all natural eugenol and IGSs produce only one product—either eugenol or isoeugenol—but in vitro mutagenesis of a few residues at the active site which crystallographic structures demonstrated to influence the binding position of the substrate resulted in an enzyme that produces a mixture of both phenylpropene regio-isomers (Koeduka et al., 2008).

The synthesis of some of the phenylpropenes found in plants, such as t-anethole, methylchavicol, methyleugenol, and methylisoeugenol, requires methylation of the para-hydroxyl group of their phenyl rings (Fig. 1). It has been previously reported that in basil glands, two closely related (90% identical) enzymes chavicol O-methyltransferase (CVOMT) and eugenol O-methyltransferase (EOMT) catalyze the formation of methylchavicol and methyleugenol from chavicol and eugenol, respectively (Gang et al., 2002). CVOMT has very little activity with t-anol whereas EOMT is 4-fold more active with eugenol than with chavicol. Basil EOMT is also 4-fold more active with eugenol than with isoeugenol, whereas isoeugenol/eugenol O-methyltransferase (IEMT) from C. breweri flowers, an enzyme that is only 45% identical to EOMT or CVOMT, converts both isoeugenol and eugenol to their methylated derivatives at roughly equal rates (Wang and Pichersky, 1998).

Previous studies showed that the essential oil of the seeds and leaves of anise consists of 90% t-anethole as well as some methylchavicol, anisaldehyde, and β-caryophyllene (Orav et al., 2008). The close similarity in structure between t-anol and eugenol and isoeugenol suggests that an anise enzyme similar to IGS or EGS might be involved in its synthesis, utilizing the analogous substrate coumaryl acetate. Likewise, a methyltransferase (MT) similar to CVOMT/EOMT or IEMT might be present in anise and responsible for the conversion of t-anol to t-anethole. Here we report the identification and characterization of anise cDNAs encoding t-anol/isoeugenol synthase 1 (AIS1) and t-anol/isoeugenol O-methyltransferase 1 (AIMT1). We show that AIS1 can indeed use coumaryl acetate to catalyze the formation of t-anol. We further show that AIMT1, unlike previously characterized phenylpropene O-methyltransferases (OMTs), is highly specific for substrates in which the double bond in the propenyl side chain is located between C7 and C8. Finally, we show that unlike the situation in basil, which is in the Lamiaceae, the phenylpropenes synthesized in the anise plant, which belongs to the Apiaceae family, are not stored in glands.

RESULTS

Distribution of Phenylpropenes in Different Parts of the Anise Plant

It was previously reported that t-anethole is present in anise seeds (Rodrigues et al., 2003; Orav et al., 2008). Here we examined the levels of this compound, along with other phenylpropenes, in different organs of anise plants and during different stages of development. Since visual inspection of the plants indicated that the leaves, stems, sepals, and the abaxial side of the petals are covered with hairy trichomes (Fig. 2), we first examined whether these trichomes contain phenylpropenes. Hexane extracts of approximately 100 pooled trichomes that were physically detached from the leaves with forceps were analyzed by gas chromatography (GC)-mass spectrometry (MS) but no phenylpropenes nor any other volatiles were detected (data not shown). A comparison of the amount of phenylpropenes extracted by dipping the leaves in organic solvent for 10 min versus the amount extracted when the tissue was ground and then extracted for 10 min showed that the latter extraction procedure yielded 3- to 4-fold more phenylpropenes (Table I). This result indicated that the phenylpropenes were not significantly present in the epidermal cells or the trichomes. Furthermore, peeled leaf epidermis sections were tested to measure the amount of t-anethole and only 0.08 ± 0.05 μg of t-anethole was extracted from a square centimeter of leaf epidermis, while 2.43 ± 0.41 of t-anethole was extracted from a square centimeter section of ground whole leaves, providing additional evidence that t-anethole is present mostly inside this organ. Therefore, to measure the concentration of phenylpropenes, we ground different parts of the anise plant and extracted the phenylpropenes with hexane for 15 h, followed by GC-MS analysis. The highest levels of t-anethole, 4.6 μg/mg fresh weight (FW), were found in developing fruits (seeds and pods), followed closely by flowers (Fig. 3). t-Anethole levels were much lower in young leaves and undetected in roots. In addition to t-anethole, low levels of methylchavicol were observed in flowers and developing fruits. In leaves, low levels were observed of a modified eugenol/isoeugenol compound with a molecular mass of 248 that could not be precisely identified. No other phenylpropenes were found in the plant.

Figure 2.

Aerial organs of anise are covered with trichomes. A, Adaxial surface of mature leaves. B, Side view of mature buds. C, Single flower of anise. The inset shows a closer view of the hairy trichomes on the abaxial side of petals. D, Surface of stems.

Table I.

Comparison of t-anethole between different extraction methods

t-Anethole was extracted with hexane for 10 min by dipping plant tissues or using ground tissues.

Plant Tissues	t-Anethole
	Dipped	Ground
	ng/mg FW
Mature leaves	27.5 ± 7.1a	124.4 ± 21.5
Stem	87.6 ± 18.1	353.8 ± 53.4
Buds	160.7 ± 3.8	478.7 ± 63.4

^aValues are averages of three independent experiments ±se.

Figure 3.

Accumulation of phenylpropenes in anise during development. Tissues were ground in liquid nitrogen and the compounds were extracted with hexane and analyzed by GC-MS. Each value in the figure is an average of three replicates from three independent experiments.

Isolation and Characterization of Anise AIS1

To isolate cDNAs encoding t-anol synthase from anise, we designed redundant oligonucleotides based on conserved amino acid sequences in phenylpropene synthases (see Supplemental Table S1) and used them in a reverse transcription (RT)-PCR experiment with leaf tissue mRNA. The RT-PCR reaction yielded one fragment encoding a protein with high similarity to petunia isoeugenol synthase 1 (PhIGS1). A full-length cDNA of the gene encoding this fragment was obtained by 3′- and 5′-RACE. This cDNA encodes a protein, designated PaAIS1, with 323 amino acid residues. A phylogenetic analysis indicated that PaAIS1 is 58.8% identical to PhIGS1 and also closely related to basil EGS1 and Clarkia EGS1 and IGS1 with 52% to 54% identities (Fig. 4). PaAIS1 is less similar to petunia EGS1 and Clarkia EGS2. The later two enzymes reside in a separate clade together with the previously characterized PCBER enzymes.

Figure 4.

Phylogenetic analysis showing the relatedness of PaAIS1 to other representative NADPH-dependent PIP reductases. An unrooted tree was constructed by using the neighbor-joining method with EGS/IGS proteins and the PIP family of NADPH-dependent reductases. Bootstrap values from a minimum of 1,000 trials are shown. Cb, C. breweri; Du, Desmodium uncinatum; Fi, Forsythia intermedia; Lj, Lotus japonicus; Md, Malus domes; Ms, M. sativa; Ob, O. basilicum; Pa, P. anisum; Ph, P. hybrida; Ps, Pisum sativum; Pt, Populus trichocarpa; Ptd, Pinus taeda; Th, Tsuga heterophylla; Tp, Thuja plicata; Vv, Vitis vinifera; At, A. thaliana; Ca, Cicer arietinum; Lt, L. tridentata. In addition to IGS/EGS-like proteins including AIS1 and chavicol/EGS, the PIP family includes PLRs, pinoresinol reductases, IFRs, PCBER, leucocyanidin reductase, and pterocarpan reductase. Clones biochemically characterized in C. breweri, basil, and petunia are underlined. Anise AIS1 cloned in this study is indicated with an oval. The accession numbers of the sequences analyzed in this figure are given in “Materials and Methods.”

Purified PaAIS1, obtained by expression in Escherichia coli, was able to catalyze the formation of t-anol from coumaryl acetate and of isoeugenol from coniferyl acetate (Fig. 5). An apparent K_m value of 229.9 ± 46.2 μm for coniferyl acetate was measured with an apparent k_cat value of 1.02 ± 0.17 per s, whereas the apparent K_m value for coumaryl acetate was 134.9 ± 36.4 μm with an apparent k_cat value of 0.07 ± 0.01 per second (Table II). We note however that in the assays containing high concentration of the coumaryl acetate substrate (>100 μm), some precipitation was observed, perhaps due to the low solubility of the product or the substrate.

Figure 5.

Product analysis by GC-MS of the reaction catalyzed by PaAIS1. A, The reaction product of anise AIS1 using coumaryl acetate as the substrate (and NADPH). B, The reaction product of anise AIS1 using coniferyl acetate as the substrate (and NADPH).

Table II.

Kinetic parameters of AIS1 and AIMT1

Enzymes	Substrates or Cofactor	Km	kcat	kcat/Km
		μm	s−1	s−1 mm−1
AIS1	Coumaryl acetate	134.9 ± 36.4a	0.07 ± 0.01	0.52
	Coniferyl acetate	229.9 ± 46.2	1.02 ± 0.17	4.44
AIMT1	Isoeugenol	19.3 ± 2.9	0.015 ± 0.001	0.78
	SAM	54.5 ± 6.5	–	–

^aValues are averages of three independent experiments ±se.

Detailed characterization of PaAIS1 showed that the enzyme was active in the pH range of 6.0 to 7.5 with optimal activity at pH values between 6.0 and 6.5 with Bis/Tris buffer. At the pH range of 6.5 to 7.5, a small amount of eugenol was detected. PaAIS1 activity was strongly inhibited by 2.5 mm Zn²⁺, Cu²⁺, and Fe²⁺ (63%–100% inhibition) and moderately inhibited at 2.5 mm of K⁺, Mn²⁺, and Ca²⁺ (10%–18% inhibition). In contrast, no inhibitory effect on the activity was observed by incubation with 2.5 mm Na⁺ and Mg²⁺. PaAIS1 was stable at 20°C for 30 min and retained 85% activity after incubation for 30 min at 37°C. However, it was almost completely inactivated when incubated for 30 min at 50°C (less than 4% activity remaining).

Isolation and Identification of Anise cDNA Encoding an Enzyme Capable of Synthesizing t-Anethole from t-Anol

To isolate anise cDNAs encoding an MT capable of methylating t-anol to produce t-anethole, we designed redundant oligonucleotides based on conserved amino acid sequences in phenylpropene MTs (see Supplemental Table S1) and used them in an RT-PCR experiment with leaf tissue mRNA. After nested PCR, clear PCR products (approximately 180 bp) were obtained and sequenced. Three MT-like sequences were identified and 5′- and 3′-RACE experiments led to the isolation of three complete and distinct cDNAs. One of the full-length cDNAs contained an open reading frame encoding a protein of 358 amino acids. The protein encoded by this cDNA, produced in E. coli and purified by nickel-nitrilotriacetic acid agarose affinity chromatography, catalyzed the formation of t-anethole from t-anol (Fig. 6) as well as the formation of methylisoeugenol from isoeugenol (see below), and therefore this cDNA was designated as PaAIMT1. The other two MT-like cDNAs encoded proteins that did not show appreciable levels of activities against t-anol. Sequence analysis indicated that PaAIMT1 was 43.2% and 41.4% identical to Medicago sativa COMT and C. breweri IEMT, respectively, and only 28.2% and 27.3% identical to basil CVOMT and EOMT, respectively (Fig. 7).

Figure 6.

Product analysis by GC-MS of the reaction catalyzed by PaAIMT1. A, An authentic t-anethole standard. B, The reaction product of anise AIMT1 using t-anol and SAM.

Figure 7.

A phylogenetic tree showing the relatedness of PaAIMT1 to selected MTs. An unrooted tree was constructed by using the neighbor-joining method. Bootstrap values from a minimum of 1,000 trials are shown. MpF8OMT, Mentha piperita flavonoid 8-OMT; RcOMT1, Rosa chinensis OMT1; ObCVOMT, O. basilicum chavicol OMT; ObEOMT, O. basilicum eugenol OMT; CrF4OMT, Catharanthus roseus flavonoid 4-OMT; MsIOMT, M. sativa isoflavone OMT; HvF7OMT, Hordeum vulgare flavonoid 7-OMT; HlOMT1, Humulus lupulus OMT1; HlOMT2, Humulus lupulus OMT2; EcRT7OMT, Eschscholzia californica reticuline 7-OMT; PsRT7OMT, Papaver somniferum (R,S)-reticuline 7-OMT; PaAIMT1, P. anisum t-anol/isoeugenol MT; McI4OMT, Mesembryanthemum crystallinum inositol 4-OMT; CbIEMT, C. breweri isoeugenol/eugenol MT; CbCOMT, C. breweri caffeic acid OMT; MsCOMT, M. sativa caffeic acid OMT; CaCOMT, Chrysosplenium americanum caffeic acid OMT; PamCOMT, Prunus armeniaca caffeic acid OMT; ObCOMT, O. basilicum caffeic acid OMT; ZeCOMT, Zinnia elegans caffeic acid OMT; RhOOMT1, Rosa hybrida orcinol OMT1. Clones biochemically characterized in C. breweri, basil, and rose (Rosa spp.) are underlined. Anise AIMT1 cloned in this study is indicated with an oval. The accession numbers of the sequences analyzed in this figure are given in “Materials and Methods.”

Enzymatic Characterization of PaAIMT1

We tested the activities of PaAIMT1 with a number of potential substrates at a final concentration of 10 μm, since some of the substrates, including t-anol, are not completely soluble in aqueous solution at higher concentrations (Fig. 8; Table III). The results of these assays indicated that PaAIMT1 methylated isoeugenol at the highest rate in an S-[methyl-¹⁴C]adenosyl-l-Met (SAM)-dependent fashion, followed by dihydroeugenol and t-anol. Neither chavicol, eugenol, nor 5-methoxyeugenol, which all possess an allylic propenyl side chain, served as efficient substrates. In addition, compounds with either no alkene side chain (guaiacol), with hydrophilic functionalities on the end of the side chain (e.g. caffeic acid), or with an ortho-hydroxy group on the phenyl ring (2-propylphenol) were not substrates for PaAIMT1. Interestingly, the activity of PaAIMT1 with dihydrochavicol was approximately half of that with t-anol, and a similar ratio of activity was observed with dihydroeugenol versus isoeugenol. CbIEMT, on the other hand, showed the opposite behavior for these substrates and preferred dihydroeugenol over isoeugenol (and eugenol) by a factor of 2 (t-anol, chavicol, and dihydrochavicol could not be methylated by CbIEMT).

Figure 8.

Substrates used to measure the relative specific activities of anise AIMT1 and Clarkia IEMT. A, Compounds that served as substrates for anise AIMT1 enzyme. B, Compounds that did not serve or showed less than 5% relative activities for anise AIMT1 enzyme as substrates.

Table III.

Relative specific activities of anise AIMT1 and Clarkia IEMT with t-anol, isoeugenol, and a variety of related substrates

All substrates were tested at 10 μm of a final concentration, since some of the substrates are not soluble in aqueous solutions at higher concentrations.

Substrates	Relative Activity (%) for Each Enzyme Assayed
	AIMT1	IEMT
Chavicol	5.9 ± 7.1a	NDb
t-Anol	45.3 ± 18.1	ND
Dihydrochavicol	22.8 ± 3.8	ND
Eugenol	9.1 ± 3.8	108.3 ± 1.7
Isoeugenol	100.0c	100.0c
Dihydroeugenol	48.5 ± 3.8	211.1 ± 10.6
5-Methoxyeugenol	17.8 ± 4.8	115.0 ± 4.5

^aValues are averages of three independent experiments ±se.

^bND, No activity detected.

^cSpecific activities for AIMT1 and IEMT were 5.83 and 1.30 pkat/mg protein, respectively.

Detailed characterization of PaAIMT1 showed that the enzyme was active in the pH range of 6.0 to 9.0 with optimal activity at pH values between 7.5 and 8.0 with Tris buffer. PaAIMT1 activity was strongly inhibited by 2.5 mm Zn²⁺ and Cu²⁺ (97% inhibition, respectively) and moderately inhibited at 2.5 mm of Mn²⁺ and Fe²⁺ (35% and 47% inhibition, respectively). In contrast, no inhibitory effect on the activity was observed by incubation with 2.5 mm K⁺, Na⁺, Ca²⁺, and Mg²⁺. PaAIMT1 was stable at 20°C to 37°C for 30 min and retained 88% activity after incubation for 30 min at 50°C. However, it was almost completely inactivated when incubated for 30 min at 65°C (less than 5% activity remaining). The apparent K_m values of 19.3 ± 2.9 and 54.5 ± 6.5 μm (n = 3) for isoeugenol and SAM, respectively, were calculated with an apparent k_cat value of 0.015 ± 0.001 per second (n = 3) for isoeugenol (Table II). Kinetic parameters for t-anol could not be obtained because of the lack of solubility of the substrate at concentrations above 75 μm.

Temporal Changes in Levels of t-Anol MT Activity in Anise Plants

Crude protein extracts from different organs of anise plants and during different stages of development were incubated with t-anol and [¹⁴C]SAM to measure t-anol MT activity. The results indicated that the highest levels of such activity were found in mature buds, followed by flowers and developing fruits (Fig. 9). Similar measurements for t-anol-forming activity could not be obtained because of the lack of radiochemical substrate to follow newly synthesized product; nonradioactive assays were not sensitive enough because of the low solubility of the substrate and because any t-anol formed was quickly converted to t-anethole, which was already present in some tissues at high concentrations.

Figure 9.

Levels of t-anol MT activities during the lifespan of anise plants. The specific activity with t-anol in buds is set at 100%.

Expression Patterns of PaAIS1 and PaAIMT1

We determined the steady-state levels of PaAIS1 and PaAIMT1 transcripts in different organs of anise plants and during different stages of the development by quantitative RT-PCR analysis (Fig. 10). The patterns of steady-state transcript levels for the two genes were very similar to each other in all organs tested, but the levels of PaAIMT1 transcripts were 15-fold higher than those of PaAIS1 throughout the plant. Transcript levels of both PaAIS1 and PaAIMT1 were lower in mature leaves than in young leaves. On the other hand, transcript levels of both genes increase as buds matured into flowers and were highest at the stage of fruit development. Intermediate levels of PaAIMT1 transcripts were detected in roots.

Figure 10.

Relative expression levels of PaAIMT1 and PaAIS1 transcripts in the different stages of anise development. A, Transcript levels of PaAIMT1. B, Transcripts levels of PaAIS1. The level of expression of PaAIMT1 in developing fruit was set as 100%. Results for each transcript in each sample were standardized with 18S rRNA transcripts levels. Results represent the average of three replications. Bar indicates se.

DISCUSSION

Synthesis of t-Anol Is Catalyzed by an Enzyme Most Closely Related to PhIGS1

We previously examined the biosynthesis of eugenol and isoeugenol from coniferyl acetate in basil, petunia, and C. breweri (Koeduka et al., 2006, 2008). Recently, Vassão et al. (2007) showed that a PCBER-like enzyme in Larrea tridentata is capable of catalyzing the formation of chavicol from coumaryl acetate as well as from coumaryl coumarate, although no evidence was presented that this plant actually synthesizes chavicol/methylchavicol. Here we show that anise plants contain an enzyme closely related to petunia IGS1, C. breweri IGS1, and C. breweri and basil EGS1 enzymes. PaAIS1 is capable of converting coumaryl acetate to t-anol, a compound that serves as the precursor to the highly abundant t-anethole, and can also convert coniferyl acetate to isoeugenol, a compound that is not found in anise. The kinetic parameters of PaAIS1 indicate that it has a slightly higher affinity for coumaryl acetate than for coniferyl acetate, but a turnover rate for the former is 14-fold slower than the latter. Thus, coniferyl acetate, at least in vitro, is a preferred substrate and the formation of t-anethole in anise, rather than isoeugenol, may be determined earlier in the overall biosynthetic pathway (i.e. by the lack of enzymes that act to introduce a methoxy group at the meta-position of the phenyl ring of the substrate).

AIMT1 Discriminates the Double-Bond Position in the Propenyl Side Chain of Its Substrates

The sequence of PaAIMT1 indicates that it is not closely related to either C. breweri IEMT or basil EOMT and CVOMT (Fig. 7), suggesting that its phenylpropene-methylating activity evolved independently of these former enzymes. PaAIMT1's substrate preference is also quite different from these former enzymes, having a 10-fold greater preference for substrates with a C7-C8 double bond including t-anol and isoeugenol. Its decreased activity with dihydrochavicol and dihydroeugenol, which lack a C7-C8 double bond, also suggests that this double bond is important for either enzyme binding or methyl group transfer by AIMT1. To identify the amino acids or regions that are responsible for the substrate specificities of AIMT1, additional structural studies of AIMT1 along with other phenylpropene MTs such as CbIEMT, are currently being carried out.

Biosynthesis of t-Anethole during Development of Anise

Many plants synthesize and accumulate high levels of volatiles, including phenylpropenes, in their vegetative tissue for defense (Pichersky and Gershenzon, 2002). Often, such volatile compounds are synthesized and stored in glandular trichomes localized on the surface of the leaf. For example, various cultivars of sweet basil synthesize and store methylchavicol, eugenol, and methyleugenol in peltate glandular trichomes (Gang et al., 2001; Iijima et al., 2004). These phenylpropene compounds, in addition to being volatile, are also quite cytotoxic to plants and animals alike, and storage in the trichomes, typically in the space between the cell wall and the cuticle (Turner et al., 2000; Gang et al., 2001), isolate these compounds from the rest of the plant (Schilmiller et al., 2008). When phenylpropenes are synthesized in flowers to attract pollinators, as in petunia and C. breweri, they do not accumulate in the floral tissue at high levels but are constitutively emitted into the ambient air (Wang et al., 1997; Orlova et al., 2006). However, our data indicates that in anise, which is in a different family than basil, the high levels of t-anethole (4.6 μg/mg FW) accumulate not in the trichomes or even in the epidermal cell layer, but instead inside the aerial organs and not in the trichomes and epidermal tissues unlike in sweet basil. How the tissue tolerates these high concentrations of t-anethole and whether internal specialized cells or storage structures exist in anise plants to synthesize and store this compound remain to be determined. Since phenylpropenes are found in many diverse plant families, it would also be of interest to determine how often these compounds are found in trichomes or internally.

The concentration of t-anethole in the aerial parts of the anise plant increases as the plant matures, and reaches a maximum in developing fruits (Fig. 3). The higher levels of t-anethole, a fungicide, in the reproductive tissues of the plant is a likely adaptation for increased protection of these structures, and its accumulation in the seeds allows for protection during germination and at the early stages of seedling growth. Comparisons of the changes in t-anethole levels, AIMT1 activity levels (PaAIS1 activity levels could not be measured), and PaAIMT1 and PaAIS1 transcript levels from young leaves to mature leaves and from buds to flowers to developing fruit suggest that product accumulation is regulated at multiple levels. The levels of t-anol methylating activity (Fig. 9) are higher in young leaves than in older leaves and transcript levels for both genes are also higher in young leaves, but t-anethole levels are higher in mature leaves (Fig. 3). This is consistent with observations in basil leaves (Gang et al., 2002) and is easily explained as the continuous synthesis and storage of t-anethole during leaf development leads to highest levels of accumulation in mature leaves even though synthesis has by then greatly diminished. A similar explanation can be invoked for the changes in enzyme activity levels and product accumulation in the reproductive parts of the anise plant, although in these parts transcript levels continue to increase, leading to very high levels in the fruit, although enzyme activity levels do not increase correspondingly. Similarly, transcripts of both genes at levels similar to those found in young leaves are also observed in roots, although PaAIMT1 activity and t-anethole levels in roots are very low. The observation of relatively high levels of transcripts without a corresponding level of enzymatic activity suggests posttranslational regulation. It is also possible that both in developing fruits and in roots, PaAIS1 and PaAIMT1 transcripts are stored for later use. For example, seedlings upon germination may use stored transcripts to quickly synthesize additional AIS1 and AIMT1 proteins to produce additional t-anethole for defense, and roots may also use stored AIS1 and AIMT1 transcripts to quickly increase their t-anethole biosynthetic capacity in case of herbivory.

MATERIALS AND METHODS

Plant Materials and Growth Conditions

Anise (Pimpinella anisum) plants were grown in a regular soil in a growth chamber under a 14-h light/10-h dark photoperiod. Temperature was set to 22°C during the light period and 18°C during the dark period. The root tissues for measuring metabolites, crude enzyme activities, and total RNA extraction were grown with vermiculite instead of soil.

Chemicals

All chemicals were from Sigma, unless otherwise indicated. Coniferyl acetate was synthesized as previously described (Koeduka et al., 2008). Coumaryl acetate was synthesized as described in the Supplemental Text S1.

Analysis of t-Anethole from Anise Plants

The leaf epidermal tissue was peeled out by the forceps and immediately incubated in hexane for the extraction of t-anethole. The area of leaf epidermis was calculated by weighing the footprint, where it was peeled off, and the sequential comparison of the weight for 1.1 × 1.1 cm (1.2 cm²) area as a control. In other tissues, t-anethole was extracted with hexane, as described in the text. GC-MS analyses of the extracts were preformed as described in Koeduka et al. (2008).

Enzyme Extraction

Crude protein extracts were prepared by homogenizing freshly excised plant tissues using a mortar and pestle in the presence of ice-cold extraction buffer (10:1 [v/w], buffer:tissue) containing 50 mm Tris-HCl, pH 7.0, 1 mm 2-mercaptoethanol, 5 mm Na₂S₂O₅, 1% (w/v) polyvinylpyrrolidone (PVP-40; Sigma), 10% glycerol, and 1 mm phenylmethylsulfonyl fluoride. The slurry was centrifuged at 12,000g for 10 min. The supernatant was used for enzyme assays immediately after clarification.

Enzyme Assays and Product Identification

The standard radiochemical assay mixture for the recombinant proteins consisted of 5 μL of 0.5 m Tris-HCl (pH 7.5), 3 or 5 μL of purified proteins (AIMT1; 0.8 μg or IEMT; 4.7 μg, respectively), 1 μL of 0.5 mm substrate dissolved in methanol (final concentration of substrates was 10 μm), and 0.25 μL of SAM (specific activity 48.8 mCi/mmol [Perkin Elmer Instruments]), and 40.75 or 38.75 μL of water to bring the assay volume to 50 μL. Reactions were incubated at 25°C for 30 min and stopped by the addition of 2 n HCl. The products were extracted with 200 μL ethyl acetate and counted in a scintillation counter as previously described (Wang et al., 1997). In addition, verification of the identity of the reaction products was performed using nonradiolabeled SAM as cofactor and 3 times larger reaction volumes. The products were analyzed by GC-MS as described previously (Koeduka et al., 2008). For kinetic analyses, appropriate enzyme concentrations were chosen so that the reaction velocity was linear in time during the reaction period. The kinetic parameters were determined as described by D'Auria et al. (2002).

Isolation, Characterization, and Expression in Escherichia coli of Anise AIMT1 and AIS1

Single-strand cDNAs were synthesized from poly-A RNA isolated from 3-week-old anise leaves with Oligo-dT primer and Superscript II reverse transcriptase (Invitrogen). These cDNAs were used as the template for the amplification of MT and phenylpropene synthase cDNA fragments. Degenerate primers (Supplemental Table S1) were designed to correspond to peptide sequences within phenylpropene synthases and MTs that are highly conserved in different proteins belonging to each group as identified by protein sequence comparisons. The identified conserved sequences included the regions encoding the catalytic domain and the SAM and NADPH binding sites, respectively. In the first PCR, the degenerated primers were used to amplify phenylpropene synthases and MTs partial cDNAs under the following PCR conditions: denaturation, 2 min at 94°C; five cycles, 15 s at 94°C, 1 min at 37°C, 1 min at 40°C, and 1 min at 72°C; five cycles, 15 s at 94°C, 1 min at 40°C, 1 min at 45°C, and 1 min at 72°C; five cycles, 15 s at 94°C, 2 min at 45°C, and 1 min at 72°C; five cycles, 15 s at 94°C, 1 min at 50°C, and 1 min at 72°C; 20 cycles, 15 s at 94°C, 1 min at 55°C, and 1 min at 72°C. Amplified PCR products at about 550 bp and 270 bp for phenylpropene synthases and MTs, respectively, were used as the template for the second PCR under the same PCR conditions as in the first PCR. The resulting PCR products at about 350 bp and 150 bp for phenylpropene synthases and MTs, respectively, were cloned into the pGEM-T easy vector (Promega) and sequenced. 5′- and 3′-RACE transcripts were performed to obtain the complete coding sequence of PaAIS1 and PaAIMT1 using SMART RACE cDNA amplification kit (CLONTECH) following the manufacturer's instructions with internal gene-specific primers. The resultant PCR products were subcloned into pGEM-T easy vector and their nucleotides were determined completely. The sequenced clones of 5′- and 3′-RACE covered the desired missing sequences. Based on these additional sequences, full-length cDNAs were cloned into the expression vector pEXP5-CT/TOPO (Invitrogen) to encode C-terminally His-tagged proteins. Resulting constructs were expressed in Escherichia coli. E. coli cells harboring PaAIS1 or PaAIMT1 genes were induced with isopropylthio-β-galactoside (final concentration; 0.5 mm), and the purification of induced proteins were performed using previously described methods (Koeduka et al., 2006).

Quantitative RT-PCR

Total RNA was isolated from each tissue with an RNeasy Plant Mini kit (Qiagen). The RNA was subjected to DNase treatment using the DNA-free kit (Ambion), and first-strand cDNA was synthesized by Superscript II transcriptase (Invitrogen) with poly-T primer in parallel with a negative control reaction in which no Superscript II reverse transcriptase was added. The qPCR reactions utilizing SYBR-Green I dye (Molecular Probes), Taq polymerase (New England Biolabs), fluorescein (Bio-Rad), the specific primers (Supplemental Table S2), and a dilution series of each cDNA as standards, were performed as previously described (Varbanova et al., 2007).

Sequence data from this article can be found in the GenBank/EMBL data libraries under the following accession numbers: CbIGS1, EF467238; CbEGS1, EF467239; CbEGS2, EF467240; ObEGS1, DQ372812; PhIGS1, DQ372813; PhEGS1, EF467241; FiPLR, AAC49608; TpPLR, AAF63507; ThPLR, AAF64184; ThPCBER, AAF64178; PtdPCBER, AAC32591; FiPCBER, AAF64174; PtPCBER, CAA06707; PsPCBER, ABF39004; CaIFR, Q00016; PsIFR, P52576; MsIFR, AAC48976; LjPTR, BAF34841; VvLAR, CAI26309; MdLAR, AAZ79364; DuLAR, Q84V83; AtPrR1, At1g32100; AtPrR2, At4g13660; PaAIS1, EU925388; MpF8OMT, AAR09600; RcOMT1, CAD29458; ObCVOMT, AAL30423; ObEOMT, AF435008; CrF4OMT, AAR02420; MsIOMT, T09254; HvF7OMT, S52015; HlOMT1, EU309725; HlOMT2, EU309726; EcRT7OMT, BAE79723; PsRT7OMT, AAQ01668; RcOMT1, BAC78826; McI4OMT, P45986; CbIEMT, AAC01533; CbCOMT, AAB71141; MsCOMT, P28002; CaCOMT, AAA86982; PamCOMT, AAB71213; ObCOMT, AAD38189; ZeCOMT, AAA86718; RhOOMT1, AAM23004; and PaAIMT1, EU925389.

Supplemental Data

The following materials are available in the online version of this article.

• Supplemental Table S1. Conserved sequences of phenylpropene synthases and phenylpropene MTs.

• Supplemental Table S2. Sequences of oligonucleotide for quantitative RT-PCR.

• Supplemental Text S1. Synthesis of coumaryl acetate.