A Novel Cation-Dependent O- Methyltransferase Involved in Anthocyanin Methylation in Grapevine

Anthocyanins are major pigments in colored grape ( Vitis vinifera ) berries, and most of them are monomethoxylated or dimethoxylated. We report here the functional characterization of an anthocyanin O -methyltransferase (AOMT) from grapevine. The expression pattern in two cultivars with different anthocyanin methylation proﬁles (Syrah and Nebbiolo) showed a peak at start ripening ( ve´raison ), when the concentrations of all methylated anthocyanins begin to increase. The puriﬁed recombinant AOMT protein was active on both anthocyanins and ﬂavonols in vitro, with K m in the micromolar range, and was dependent on divalent cations for activity. AOMT showed a preference for 3 # ,5 # methylation when a 3 # ,4 # ,5 # hydroxylated anthocyanin substrate was tested. In order to assess its in planta activity, we performed transient expression of AOMT in tobacco ( Nicotiana benthamiana ) leaves expressing the Production of Anthocyanin Pigment1 (PAP1) transcription factor from Arabidopsis ( Arabidopsis thaliana ). PAP1 expression in leaves induced the accumulation of the nonmethylated anthocyanin delphinidin 3-rutinoside. The coexpression of PAP1 and AOMT resulted in an accumulation of malvidin 3-rutinoside. We also showed that AOMT localized exclusively in the cytoplasm of tobacco leaf cells. These results demonstrate the ability of this enzyme to methylate anthocyanins both in vitro and in vivo, indicating that AOMT plays a major role in anthocyanin biosynthesis in grape berries.

Methylation by S-adenosyl-L-methionine (SAM)dependent O-methyltransferases (OMTs; EC 2.1.1) is a common modification in plant secondary metabolism, which modulates the physiological properties and the chemical reactivity of phenolic groups (Zhu et al., 1994). Several classifications of plant OMTs have shown that they can be categorized into two major classes (Joshi and Chiang, 1998;Noel et al., 2003;Lam et al., 2007). In the nomenclature proposed by Noel et al. (2003), type 1 OMTs consist of homodimeric OMTs with subunit sizes of about 38 to 43 kD, which do not require divalent cations for activity. This large family includes caffeic acid, flavonoid, coumarin, and alkaloid OMTs (Frick and Kutchan, 1999;Dong et al., 2003). Type 2 OMTs represent a group of lower molecular mass (23-27 kD) and cationdependent OMTs. Most members of this family have been shown to be specific for CoA esters of phenylpropanoids and have been suggested to be key enzymes in lignin biosynthesis (Ye et al., 1994). Among flavonoids alone, hundreds of O-methylated compounds have been characterized, ranging from monomethylated to polymethylated compounds and belonging to the chalcone, flavone, isoflavone, flavonol, and anthocyanin families (Wollenweber and Dietz, 1981;Markham, 1989;Andersen and Markham, 2006).
Anthocyanins are responsible for the characteristic red, blue, and purple colors of many plant tissues (Harborne and Williams, 2000). They play important roles as pigments of flowers and fruits to attract animals for pollination and seed dispersion and act as protectants against UV irradiation (Winkel-Shirley, 2001). Anthocyanins contribute to the distinctive color and the aesthetic quality of many plant-derived products such as wine. In addition to their direct role in red wine color, anthocyanins also participate in the organoleptic and chemical attributes of wine, because of their interaction with other phenolic and aroma compounds as well as with proteins and polysaccharides (Mazza and Miniati, 1993).
Red grape (Vitis vinifera) varieties usually accumulate 3-O-monoglucoside, 3-O-acetylglucoside, and 3-O-p-coumaroylglucoside derivatives of five anthocyanidins: delphinidin, cyanidin, peonidin, petunidin, and malvidin. The total amount of anthocyanins and relative abundance of single anthocyanins are extremely variable among red-to blue-skinned cultivars; however, monomethylated and dimethylated derivatives are usually largely predominant (Mazza and Miniati, 1993). Methylation is an important aspect of grape anthocyanin biosynthesis, as anthocyanin O-methylation has a slight reddening effect and is known to reduce the chemical reactivity of phenolic hydroxyl groups (Sarni et al., 1995). This modification is proposed to increase the stability of anthocyanins and modify their water solubility, thus significantly contributing to the accumulation of color, a pivotal quality parameter for the red wine industry. Anthocyanin methylation in grapes is affected by genetic factors, malvidin 3-glucoside being the predominant anthocyanin in most cultivars and peonidin 3-glucoside appearing in a more restricted group of genotypes (Mattivi et al., 2006), as well as by environmental and cultural conditions (Downey et al., 2006). Although many structural and regulatory genes involved in anthocyanin biosynthesis have been cloned from a wide variety of plant species (Mol et al., 1998;Forkmann and Heller, 1999;Winkel-Shirley, 2001), genes responsible for the formation of methylated anthocyanins remain to be characterized. Genetic analyses of anthocyanin biosynthesis in petunia (Petunia hybrida) showed that two pairs of duplicate genes, Mt1/Mt2 and Mf1/Mf2, were responsible for the methylation of the anthocyanin molecule at the 3# and 5# positions, respectively (Wiering and de Vlaming, 1977). Genetic relationships among these four genes suggested that if one or both of the Mt genes were dominant, mainly 3#-O-methylated anthocyanins (peonidin or petunidin) accumulated in the flower. The 3#,5#-O-methylated malvidin accumulated as the main pigment only when one or both of the Mf genes were dominant. However, additional investigation of the relationship between the four methylation genes and anthocyanin O-methyltransferase activity in vitro revealed that both the Mt enzymes and the Mf enzymes were capable of methylating the oxygen in the 3# and 5# positions of the anthocyanin molecule (Jonsson et al., 1983). The same authors could then separate the four different OMTs in petunia flower extracts using chromatofocusing techniques (Jonsson et al., 1984). A detailed characterization of these enzymes showed that they exhibit different kinetic properties and that their combination reflects the methylation patterns of anthocyanins in flowers in different lines of petunia. However, the genes encoding these OMTs from petunia have not been cloned to date. Bailly et al. (1997) partially purified an OMT from anthocyanin-producing grape cell suspensions. This enzyme could methylate cyanidin 3-O-glucoside but showed no activity with cyanidin 3-p-coumaroyl-Oglucoside, suggesting that methylation occurred before acylation. Furthermore, the lack of OMT activity when delphinidin was used as a substrate indicated that more than one OMT may be involved in anthocyanin methylation in grape (Bailly et al., 1997). EST and grapevine genome sequencing (Da Silva et al., 2005;Jaillon et al., 2007) have identified a large number of putative OMTs, and expression of some of them is correlated with the accumulation of methylated anthocyanins in berries (Ageorges et al., 2006;. In this work, we describe the identification and the biochemical characterization of a novel Mg 2+dependent anthocyanin O-methyltransferase (AOMT) from grapevine. Combining in vitro characterization of recombinant AOMT enzyme with Agrobacterium tumefaciens-mediated transient transformation of tobacco (Nicotiana benthamiana), we show that this grapevine AOMT catalyzes both 3# and 5# O-methylation of anthocyanins in vitro and in planta.

Isolation of a Candidate AOMT cDNA from Grapevine
Recently, a high-throughput transcriptomic screening identified a set of genes coexpressed during anthocyanin biosynthesis in grape (Ageorges et al., 2006). One of these genes, corresponding to the partial grape EST BQ796057, was highly similar to the multifunctional phenylpropanoid and flavonoid OMT (PFOMT) from Mesembryanthemum crystallinum (Ibdah et al., 2003), suggesting that this gene could be involved in anthocyanin methylation. A BLAST search of Vitis databases using the grapevine EST BQ796057 as the query sequence identified three overlapping ESTs, named EC932040.1, EC953324.1, and CF603338.1. Their sequences were assembled in a contig, and the full-length cDNA was amplified by reverse transcription-PCR from mature Syrah berries. The predicted amino acid sequence of this candidate AOMT cDNA was the same in the two cultivars and corresponded to a 26.4-kD protein, which contained a conserved domain identified as Methyltransf_3: O-methyltransferase (Pfam 01596). The candidate AOMT belonged to the type 2 family of low-M r and cation-dependent OMTs. Most members of this family have been shown to be active with CoA esters of phenylpropanoids. However, some type 2 OMTs are also active with flavonoids and/or phenylpropanoid conjugates (Ibdah et al., 2003;Kim et al., 2006;Lee et al., 2008). The candidate AOMT presented 56% and 54% identity with PFOMT from M. crystallinum (Ibdah et al., 2003) and caffeoyl-CoA-OMT (CCoA-OMT) from grapevine (Busam et al., 1997), respectively (Fig. 1A). Phylogenetic analysis of selected members of the plant OMT gene family showed that the grapevine AOMT belongs to a small clade of type 2 OMTs, together with the flavonoid-OMTs PFOMT from M. crystallinum and OMT1 from Arabidopsis (Arabidopsis thaliana; Fig. 1B).

Analysis of AOMT Gene Expression during Grape Berry Development
Transcriptomic analyses have shown that AOMT gene expression was induced during anthocyanin accumulation in grape berries, leading to the hypothesis that AOMT could be involved in anthocyanin methylation (Ageorges et al., 2006;. In order to gain insight into AOMT gene regulation, Syrah and Nebbiolo grape berries were harvested at different ripening stages and analyzed for both anthocyanin content and AOMT gene expression (Fig. 2). No anthocyanin was detected before véraison, the onset of ripening. From this stage on, anthocyanins started to accumulate in berry skin. Malvidin accumulated in Syrah berries as the major aglycone, reaching 1.12 mg equivalent malvidin per berry (69% of total anthocyanins) in fully mature berries ( Fig. 2A). Peonidin content increased significantly during Syrah berry development, reaching 17% of total anthocyanins ( Fig.  2A). In Syrah berries, AOMT transcripts were hardly detectable before the onset of ripening (Fig. 2C). After véraison, AOMT expression level increased dramatically during the first weeks and then the expression decreased and remained stable until the end of ripening (Fig. 2C). In Nebbiolo berries, the accumulation of anthocyanins followed the same developmental pattern as in Syrah berries (Fig. 2B). In Nebbiolo, the main anthocyanin was the monomethylated peonidin 3-Oglucoside, which accumulated at concentrations about 2-fold higher than the sum of petunidin 3-O-glucoside and malvidin 3-O-glucoside (Fig. 2B). However, the expression of AOMT in this variety followed the same profile as it did in Syrah (Fig. 2D). Given the fact that AOMT resembled the previously characterized CCoA-OMT (Busam et al., 1997), the expression of the CCoA-OMT gene was investigated during grape berry development as a comparison (Supplemental Fig.  S1). Unlike AOMT, the CCoA-OMT gene was strongly expressed in berries a few weeks after flowering, and then its expression decreased rapidly until véraison. During the ripening process, CCoA-OMT transcript accumulation was very weak in grape berries.
AOMT gene expression was then evaluated in several tissues of Syrah. In vegetative organs, AOMT transcripts could be detected at very low levels in young leaves (Fig. 3). Before véraison, AOMT gene expression was barely detectable in the different berry tissues, although a weak expression could be detected in seeds at the green stage (Fig. 3). During ripening, AOMT transcripts were mainly detected in berry skin, where anthocyanin biosynthesis takes place. As a summary, the expression of AOMT appeared rather fruit specific, with a coincidental timing corresponding with the accumulation of anthocyanin pigments.

Characterization of Recombinant AOMT
The AOMT coding sequence was cloned into the Gateway-compatible entry vector pDONR207 and then transferred into the destination vector pHNGWA (Busso et al., 2005) to express the corresponding protein in Escherichia coli as a His-tagged NusA fusion protein. The fusion protein was purified by metal affinity chromatography and cleaved on the resin with thrombin to yield purified recombinant AOMT (Supplemental Fig. S2). The activity of AOMT enzyme was then tested in vitro with a number of potential substrates (Table I; Fig. 4), including flavan 3-ols, flavonols, anthocyanidins, and anthocyanins, in the presence of SAM. HPLC-diode array detection (DAD)/electrospray ionization (ESI)-mass spectrometry (MS) was used to analyze reaction products obtained after incubation of the potential substrates without or with AOMT enzyme (Fig. 5). First, AOMT could not utilize flavan 3-ols (catechin and epicatechin) as substrates (Table I). On the other hand, AOMT did not exhibit any methylation activity with the 4#-hydroxylated anthocyanin pelargonidin 3-Oglucoside (Fig. 5,A and B). Indeed, HPLC analysis of the reaction products obtained with pelargonidin 3-Oglucoside detected one major peak, the unmodified substrate (98%), and also 2% of the corresponding aglycone, pelargonidin ( G and H). A second product was detected (m/z 285) with a l max at 487 nm ( Fig. 5H, peak P1). This product was not identified but may result from the instability of anthocyanidins under experimental conditions (pH 7.5 in aqueous solution; Awika et al., 2004). In addition to anthocyanin and anthocyanidin substrates, AOMT could methylate the flavonol quercetin 3-O-glucoside to yield its 3#-O-methyl ether isorhamnetin (Fig. 5,I and J). Detailed quantifications of reaction products are presented in Supplemental  Table S1. Taken together, these results indicated that recombinant AOMT catalyzed in vitro the 3# O-methylation of anthocyanins and flavonols with a catechol B-ring (3#,4# di-OH) and both 3# and 5# O-methylation for those showing a pyrogallol (3#,4#,5# tri-OH) B-ring.
Kinetic analyses (Table I) confirmed that AOMT exhibited only background levels of activity with pelargonidin 3-O-glucoside, catechin, and epicatechin. AOMT exhibited equivalent kinetic parameters with cyanidin 3-O-glucoside, delphinidin 3-O-glucoside, quercetin 3-O-glucoside, and the flavonol aglycones quercetin and myricetin (Table I). AOMT activity with cyanidin aglycone was slightly lower than with cyanidin 3-O-glucoside. K m values for these substrates ranged from 20 to 75 mM. Recombinant AOMT was active within a broad range of pH values, with an optimum around 7.5 on both substrates tested (cyanidin 3-O-glucoside and delphinidin 3-O-glucoside; Supplemental Fig. S3). It retained over 25% of its activity at pH 9.4 and at pH 6.25, but activity was dramatically reduced when more acidic conditions were tested (Supplemental Fig. S3).
The influence of different divalent cations on AOMT activity was tested (Tables II and III). AOMT activity was extremely low in the absence of Mg 2+ ions, confirming the Mg dependence of AOMT (Table II). The best activity was obtained in the presence of 10 mM MgCl 2 (Table II). Incubation of AOMT in the presence of EDTA greatly reduced its activity with both cyanidin 3-O-glucoside and delphinidin 3-Oglucoside substrates. The presence of other divalent cations in addition to Mg 2+ strongly decreased activity in the case of Zn 2+ and Mn 2+ , while little effects were observed with Ca 2+ (Table III).

Characterization of AOMT Activity in Planta
Recombinant AOMT exhibited flavonol and AOMT activity in vitro. However, in vitro data should be taken with caution for the prediction of the physiological substrates and products of OMTs (Liu and Dixon, 2001). We used Agrobacterium-mediated transient transformation of tobacco to investigate AOMT activity in planta. This approach allowed us to suc-cessfully characterize OMT activities in planta (Scalliet et al., 2006;Schmidlin et al., 2008). In order to provide AOMT with anthocyanin substrates, AOMT was coexpressed with Production of Anthocyanin Pigment1 (PAP1), an R2R3 MYB transcription factor from Arabidopsis involved in anthocyanin biosynthesis (Borevitz et al., 2000). Constitutive expression of PAP1 in Arabidopsis and other plant species, including tobacco, petunia, and tomato (Solanum lycopersicum), has been shown to induce anthocyanin accumulation, due to the coordinated up-regulation of genes in the anthocyanin biosynthetic pathway (Borevitz et al., 2000;Xie et al., 2006;Ben Zvi et al., 2008;Zuluaga et al., 2008). No anthocyanin was detected in extracts from leaves expressing the GFP control (Fig. 6A). As expected, Agrobacterium-mediated transient expression of PAP1 in tobacco resulted in significant levels of anthocyanin accumulation in infiltrated leaves (Fig.  6, B and C). The flavonoid composition of transformed tobacco leaves was then determined using both HPLC/ESI-MS/DAD and HPLC/DAD in order to compare the flavonoids produced in PAP1 leaves with those accumulated in leaves coexpressing PAP1 and AOMT. The metabolites were putatively identified from their UV-visible absorption spectra, and comprehensive analysis of mass fragmentation patterns obtained by tandem MS spectroscopy was compared with those of known compounds.
In addition to anthocyanins, two quercetin glycosides and two kampferol glycosides were identified in control tobacco leaves expressing GFP as well as in both PAP1-expressing and PAP1-and AOMTexpressing leaves (Fig. 6, D-F). Quercetin O-rutinoside O-hexoside was the major flavonol detected in all of the transformed tobacco leaves, accounting for approximately 70% of total flavonols in control leaves and 50% in the leaves expressing PAP1. Nevertheless, the same composition in flavonol compounds was observed in leaves expressing PAP1 or coexpressing PAP1 and AOMT. Moreover, two hydroxycinnamic acids, caffeoyl quinic acid and its corresponding methylated form, feruloyl quinic acid, each with three isomers, were identified at 320 nm in 35S-GFP tobacco leaves as well as in both PAP1-expressing and PAP1and AOMT-expressing leaves (Fig. 6, D-F, peaks c1, c2, c3, f1, f2, and f3). Two isomers of feruloyl quinic acids (Fig. 6, D-F, peaks f2 and f3; Supplemental Table S2) accumulated more in PAP1-transformed and PAP1and AOMT-transformed leaves than in GFP control leaves. Taken together, these analyses indicated that coexpression of AOMT and PAP1 did not modify significantly the flavonol and phenolic acid profile, in comparison with leaves expressing PAP1 alone.

Subcellular Localization of AOMT
To investigate AOMT subcellular localization, a construct encoding AOMT fused to the N terminus of GFP (AOMT-GFP) was expressed in tobacco using Agrobacterium-mediated transient transformation. Nontargeted GFP control and AOMT-GFP were coexpressed with PAP1 transcription factor in order to investigate both the cellular localization and the enzyme activity of the AOMT-GFP fusion protein. Confocal microscopy imaging of tobacco cells expressing AOMT-GFP showed a cytosolic fluorescence pattern, very similar to that obtained with nontargeted GFP (Fig. 7). HPLC analysis of anthocyanins accumulated in tobacco leaves coexpressing PAP1 and AOMT-GFP (Supplemental Fig. S4) showed that the anthocyanin profile closely resembled that of leaves coexpressing PAP1 and AOMT (Fig. 6C). Taken together, these results indicated that the AOMT-GFP fusion protein retained AOMT activity and was located in the cytosol of transformed cells.

Functional Characterization of a Candidate AOMT from Grapevine
Several recent transcriptomic studies suggested that a putative OMT could be involved in anthocyanin methoxylation in grape berries (Ageorges et al., 2006;Pilati et al., 2007;Cutanda-Perez et al., 2009). In order to characterize this gene, called hereafter AOMT, its cDNA was cloned from Syrah and Nebbiolo grape berries. Phylogenetic analysis of selected members of the plant OMT family showed that AOMT belonged to a subset of type 2 plant OMTs and was distinct from the clade of classical CCoA-OMTs, presumably involved in lignin biosynthesis. In recent years, several OMTs active with flavonoid substrates have been characterized from different plant species, especially OMTs catalyzing the 3# or 3#,5# O-methylation of the B-ring of flavonoid substrates. Interestingly, such OMTs have been characterized in both type 1 and type 2 plant OMT families (Muzac et al., 2000;Kim et al., 2006;Lee et al., 2008). Thus, AOMT represents a new member of a small subfamily of type 2 plant OMTs involved in the methylation of flavonoid sub- strates (Ibdah et al., 2003). Determination of AOMT kinetic parameters showed that K m values and specific activities for all flavonol and anthocyanin substrates tested were similar to values found for other OMTs with their primary substrate (Ibrahim et al., 1998). AOMT exhibited a strict specificity for 3# and 5# hydroxyl group methylation in both flavonol and anthocyanin substrates, showing no activity with pelargonidin 3-O-glucoside. Furthermore, AOMT showed roughly equivalent activity with quercetin and quercetin 3-O-glucoside and with cyanidin and cyanidin 3-O-glucoside, suggesting that AOMT could methylate efficiently both glycosylated and aglycone substrates. In grape berries, major anthocyanins are both glycosylated and methylated, raising the question of the order in which these two reactions occur in vivo. Ford et al. (1998) characterized a UDP Glcflavonoid glucosyl transferase (UFGT) catalyzing the glycosylation of both anthocyanidin and flavonol substrates in vitro. However, UFGT enzyme exhibited a greatly reduced activity with malvidin aglycone compared with petunidin aglycone, even though malvidin 3-O-glucoside is usually the most abundant anthocyanin in grape berry. This suggests that anthocyanin glycosylation may occur prior to methylation in grape berries.
In order to determine the subcellular localization of AOMT, an AOMT-GFP fusion protein was expressed in tobacco. Like the nontargeted GFP control, the AOMT-GFP fusion protein exhibited a clear cytosolic localization, which is very likely to reflect the physiological localization of AOMT, as discussed below. These results suggest that both glycosylation and methylation of anthocyanins take place in the cytosol, although the precise in vivo chronology of these reactions remains to be determined. Several enzymes involved in flavonoid biosynthesis have been shown to be organized into macromolecular complexes optimizing the cooperation between its members (Burbulis and Winkel-Shirley, 1999;Liu and Dixon, 2001;Winkel, 2004). The cytoplasmic localization of the AOMT-GFP fusion protein has led to the suggestion that AOMT may be associated with the endoplasmic reticulummultienzyme complexes, facilitating rapid methylation and stabilization of its product. Glycosylation and methylation reactions may take place in such complexes, ensuring an efficient biosynthesis of anthocyanins in vivo, although the activity of UFGT with some anthocyanin substrates is low in vitro (Ford et al., 1998).

The AOMT Gene Expression Profile Mirrors Anthocyanin Biosynthesis in Berry
Five major anthocyanins exist in grape berries, which differ in their hydroxylation and methylation patterns at the phenolic B-ring. These anthocyanins include the nonmethylated cyanidin (3#,4#-OH) and delphinidin (3#,4#,5#-OH), the monomethylated peonidin (3#-OCH3 4#-OH) and petunidin (3#-OCH3 4#,5#-OH), and the dimethylated malvidin (3#,5#-OCH3 4#-OH), pelargonidin (4#-OH) being notably absent in all known Vitis genotypes (Macheix et al., 1990). The expression of the AOMT gene was monitored in two genotypes, characterized either by high total anthocyanin content and predominance of the dimethylated malvidin (Syrah; Mattivi et al., 2006) or by low anthocyanin content and predominance of the monomethylated peonidin (Nebbiolo; Guidoni et al., 2008). The AOMT gene was expressed in both varieties, exhibiting similar expression profiles, with a maximum at the onset of ripening, when methylated anthocyanins start to accumulate.  reported similar expression profiles for this OMT gene during grape berry development in other cultivars presenting different anthocyanin composition. The evolution of the transcriptional level of AOMT through ripening and the relative abundance of methoxylated anthocyanins are compatible with a major role in anthocyanin methylation in grape varieties, irrespective of their different anthocyanin composition. Figure 5. Analysis of AOMT in vitro reaction products. HPLC/ESI-MS/DAD analysis of reaction products produced following incubation of potential substrates without (A, C, E, G, and I) or with (B, D, F, H, and J) recombinant AOMT enzyme. Reactions were carried out for 60 min in a total volume of 200 mL, with 200 mM anthocyanin or flavonol substrate, 200 mM SAM, and 5 mg of purified AOMT. Substrates were as follows: A and B, pelargonidin 3-glucoside; C and D, cyanidin 3-glucoside; E and F, delphinidin 3-glucoside; G and H, cyanidin; I and J, quercetin 3-glucoside. Anthocyanins were monitored at 520 nm and flavonols at 360 nm. Reaction products were identified according to their mass fragmentation, UV/visible absorption spectra, and retention times. The data shown are representative of three independent experiments. Pg3G, Pelargonidin 3-glucoside; Cy3G, cyanidin 3-glucoside; Pn3G, peonidin 3-glucoside; Dp3G, delphinidin 3-glucoside; Pt3G, petunidin 3-glucoside; Mv3G, malvidin 3-glucoside; Cy, cyanidin; Pn, peonidin; Q3G, quercetin 3-glucoside; Q, quercetin; IsoR3G, isorhamnetin 3-glucoside; IsoRh, isorhamnetin; P1, unknown product. While the expression pattern and the substrate specificities of AOMT correlated well with the anthocyanin profiles found in grape berries, the situation was less clear for flavonols. Similar to anthocyanins, dihydroxylated or trihydroxylated flavonol compounds on the B-ring (e.g. quercetin and myricetin) can be monomethylated (quercetin to isorhamnetin and myricetin to laricitrin) or dimethylated (myricetin to syringetin). The comparative study of anthocyanin and flavonol methylation patterns over different grape genotypes showed that the ratio of flavonol monomethylation versus total flavonol was positively correlated with the ratio of anthocyanin monomethylation versus total anthocyanin, and the same was true for dimethylation (Mattivi et al., 2006). These results agree with a model where AOMT would also operate on flavonols, with the same methylation patterns as on anthocyanins. The AOMT enzymic activity measured in vitro on both substrates corroborates this hypothesis. However, in grapevine, flavonols and anthocyanins accumulate at different time points in the berry skins: flavonols accumulate at the early stage of development and during ripening, while anthocyanins accumulate only during ripening (Downey et al., 2003). In this context, it is expected that the gene encoding the flavonol methyltransferase would have a different expression pattern from the AOMT gene. Furthermore,  have shown that AOMT is not expressed in Pinot Gris, although this white genotype was shown to contain methylated flavonols (Mattivi et al., 2006). The possibility that an alternative flavonol-OMT is active in grape berries is thus still open.

Coexpression of PAP1 and AOMT in Tobacco Leaves Induces the Accumulation of Methylated Anthocyanins
In this work, we used tobacco leaves in which anthocyanin biosynthesis was induced by transient expression of the R2R3 MYB transcription factor PAP1 from Arabidopsis to characterize the activity of AOMT in planta. Overexpression of PAP1 in transgenic plants was shown to enhance the accumulation of lignin, hydroxycinnamic acid esters, and flavonoids, including various anthocyanins (Borevitz et al., 2000;Tohge et al., 2005). In the species used in our study (N. benthamiana), the most abundant anthocyanin induced by the overexpression of PAP1 was delphinidin 3-Orutinoside. The strong accumulation of this nonmethylated anthocyanin in PAP1-expressing tobacco leaves allowed the investigation of AOMT activity in vivo. The malvidin 3-O-rutinoside accumulation following coexpression of PAP1 and AOMT in tobacco leaves confirmed that AOMT alone catalyzed both 3# and 5# methylation steps of delphinidin. The predominance of malvidin 3-O-rutinoside indicates a good efficiency for the double meta methylation in vivo, even if residual delphinidin was still detected. These data confirmed the in vitro results that AOMT was able to catalyze two sequential methylations of delphinidin derivatives at the 3# and 5# positions to yield malvidin derivatives. However, petunidin 3-O-rutinoside was absent, suggesting a higher efficiency for the second methylation step in planta, which prevented the accumulation of the monomethylated intermediate. In contrast to anthocyanins, coexpression of AOMT and PAP1 did not modify significantly the flavonol and phenolic acid contents in tobacco leaves, in comparison with expression of PAP1 alone. Although AOMT was active with flavonols in vitro, the tobacco leaves, which accumulated only minor amounts of flavonols, did not allow us to confirm this activity in vivo.
Moreover, the coexpression of the AOMT-GFP fusion protein with PAP1 transcription factor in tobacco leaves resulted in the accumulation of methylated anthocyanins, identical to those accumulating after coexpression of AOMT and PAP1. These similar anthocyanin profiles indicated that the AOMT-GFP fusion protein retained AOMT activity and that its cytosolic localization likely reflected the physiological localization of the native AOMT protein.

AOMT Is Likely to Play a Major Role in Anthocyanin Methylation in Grape Berries
Previous works in petunia and grape suggested that different OMTs were responsible for the O-methylation at the 3# and 5# positions, respectively, of anthocyanin molecules (Wiering and de Vlaming, 1977;Bailly et al., 1997). In contrast with these predictions, the AOMT described here behaved as a bifunctional 3#,5#-OMT both in vitro and in vivo, suggesting that AOMT alone would be sufficient to perform both methylation steps in grape berries. The absence of cyanidin derivatives in PAP1-expressing tobacco leaves did not allow the investigation of AOMT activity with 3#,4#-dihydroxylated anthocyanins in the tobacco system. However, the AOMT gene was strongly expressed after véraison in Nebbiolo and in Grignolino grapes , which accumulate mainly the monomethylated peonidin 3-O-glucoside as the major anthocyanin. Given the fact that AOMT could efficiently methylate cyanidin 3-O-glucoside in vitro, these results suggest that AOMT Table III. In vitro AOMT activity in the presence of 10 mM MgCl 2 and 10 mM EDTA, ZnCl 2 , CaCl 2 , and MnCl 2 using cyanidin 3-O-glucoside and delphinidin 3-O-glucoside as substrates Activity is expressed as a percentage of the activity with 10 mM MgCl 2 . is also involved in the methylation of cyanidin derivatives in grape berries. These results can thus be accommodated in a model where AOMT would be the major methyltransferase acting on anthocyanins in grape berries or the main member of a family of enzymes with similar kinetic properties. This model would also explain the fact that among several grape cultivars whose berry anthocyanin composition has been analyzed, accumulation of petunidin in the absence of malvidin was never reported (Wenzel et al., 1987;Mattivi et al., 2006). The predominance and preferential 3#,5# methylating activity of this AOMT would also explain the common occurrence of malvidin in the genotypes where trihydroxylated anthocyanins are available and the existence of a positive correlation between the 3# and 3#,5# anthocyanin methoxylation activity, as estimated by metabolomic data over 64 grape genotypes (elaborated from Mattivi et al., 2006). This model is further supported by several transcriptomic studies, which identified independently the same AOMT as a major candidate for anthocyanin methylation in grape berries (Ageorges et al., 2006;Pilati et al., 2007;Cutanda-Perez et al., 2009). Finally, this model does not rule out the possibility that other OMTs may also participate in anthocyanin methylation in grapevine. Analysis of the grape genome sequence (Jaillon et al., 2007;Velasco et al., 2007) shows that the AOMT gene characterized in this work belongs to a small family of four similar putative OMT genes. Future work will be needed to investigate the roles of the other members of this gene family and their potential role in anthocyanin and/or flavonol methylation in grapevine.
In conclusion, we characterized a grapevine anthocyanin OMT that efficiently catalyzes the methylation of anthocyanins both in vitro and in planta. In addition, we provide evidence that AOMT is likely to play a major role in anthocyanin methylation in grape berries. Methylation tends to shift berry color toward purple, and these changes are transmitted to the wines, where anthocyanins form relatively stable complexes with other grape-derived flavonoids (Heredia et al., 1998;Jensen et al., 2008). Thus, identification of AOMT constitutes an important step in understanding the biosynthesis of major anthocyanins in grape and the determination of one of the key quality parameters of red grapes and wine.

Plant Material, RNA Extraction, and Grape Anthocyanin Analysis
Roots, shoots, leaves, and berries were harvested from grapevine plants (Vitis vinifera 'Syrah' [malvidin 3-glucoside as the predominant anthocyanin] and 'Nebbiolo' [peonidin 3-glucoside as the predominant anthocyanin]). Syrah plants were grown in the SupAGRO-INRA vineyard (Montpellier, France) on 30 vinestocks of homogeneous vigor. Average daily temperatures in the growing season ranged between 22.4°C (June) and 23.6°C (August), and mean relative humidity values ranged from 55.9% (June) to 67.6% (August). Nebbiolo plants were from a 15-year-old experimental vineyard at Grinzane Cavour (Piedmont, Italy) on east-west oriented rows. Average daily temperatures in the growing season ranged between 17.1°C (May) and 22.8°C (August), and rainfall summed up to 411 mm in the period from May to September. Young leaves were from the third rank, counted from the apex, with mean weight of 0.3 g per leaf. Old leaves were fully expanded leaves with mean weight of 2.8 g per leaf. Berries were collected at nine developmental stages as described previously (Terrier et al., 2005). All collected samples were quickly frozen in liquid nitrogen, ground under liquid nitrogen to a fine powder with a Dangoumau blender, and stored at -80°C until use. Total RNA from Syrah tissues was extracted using the RNeasy Plant Mini kit (Qiagen) following the manufacturer's protocol from 200 mg of starting tissue, and RNA from Nebbiolo tissues was extracted using a modification of the Chang protocol (Carra et al., 2007). Grape anthocyanin analyses of Syrah berries were performed by HPLC on 200 mg of ground frozen tissue as described by Ageorges et al. (2006). In Nebbiolo berries, anthocyanin analyses were done on 250 mg of frozen tissue by HPLC following the protocol of Guidoni et al. (2002).

Cloning of AOMT cDNA and Phylogenetic Analysis
The full-length cDNA of AOMT was amplified from cDNA of mature berries (Syrah) with high-fidelity Taq polymerase (Advantage-HF 2 PCR kit; Clontech) using the forward primer 5#-TTTTCTTGTACGGCAGGCTTA-3# and the reverse primer 5#-TGAGAATGGATTTAGGCTAATAGAG-3#. The amplified cDNA for AOMT was cloned into the pGEM-T Easy vector (Promega), and the resulting plasmid was sequenced. OMT sequences were obtained from GenBank. Nucleic acid and protein sequences were aligned using ClustalW (Thompson et al., 1994). The Phylo_win program (Galtier et al., 1996) was used to construct phylogenetic trees, using the neighborjoining method, with 500 bootstrap replicates.

Quantitative PCR Analysis of Gene Expression
In the case of Syrah, RNA was quantified with Ribogreen (Molecular Probes) and reverse transcription was performed in triplicate from each sample from 500 ng of purified RNA using the SuperScript II Reverse Transcription kit (Life Technologies). Triplicates were further pooled for PCR. Gene-specific primers were as follows: AOMT-F (5#-CTCTGCA-GGCGCCTCTATTA-3#) and AOMT-R (5#-CCCAAAACAGAGTCTGGACA-3#). Specific annealing of the oligonucleotides was controlled on dissociation kinetics performed at the end of each PCR run. The efficiency of the primer pair was measured on a plasmid serial dilution. PCR was performed in triplicate as described by Fernandez et al. (2007). Gene transcripts were quantified upon normalization to VvEF1a (Terrier et al., 2005) by comparing the cycle threshold of the target gene with that of VvEF1a. Gene expression was expressed as mean and SE calculated over all biological and technical replicates. In the case of Nebbiolo, the following modifications to the protocol were applied. Ten micrograms of RNA was reverse transcribed in duplicate using the High Capacity cDNA Reverse Transcription kit (Applied Biosystems) and PowerSYBR Green (Applied Biosystems) for quantitation of amplification results. The gene-specific primers for Nebbiolo were AOMT-F# (5#-GATGAATGTCCCTGTCGATGAG-3#) and AOMT-R# (5#-GCAAGAGCT-GTTGCCAAGAGA-3#).

Characterization of Recombinant AOMT
AOMT cDNA was amplified by PCR using the upstream primer 5#-GGGGACAAGTTTGTACAAAAAAGCAGGCTTGGTTCCGCGTGGATCC-ATGTCCAGCTCAAGTCATAGG-3# and the downstream primer 5#-GGGGAC-CACTTTGTACAAGAAAGCTGGGTTCACTAATAGAGGCGCCTGCAGAG-3# and cloned into pDONR207 Gateway-compatible vector (Invitrogen). AOMT cDNA was sequenced to verify that no mutation had been introduced and transferred into the pHNGWA destination vector (accession no. EU680842; Busso et al., 2005). AOMT enzyme was expressed as His-tagged NusA-fusion protein. Recombinant AOMT was purified using TALON metal affinity resin (Clontech) and characterized after cleavage of the NusA moiety using thrombin (GE Healthcare-Amersham). Purified recombinant AOMT (5 mg) was assayed in a final volume of 200 mL with 200 mM SAM and 200 mM anthocyanin or flavonol substrate in 0.1 M Tris, pH 7.5, containing 10% (v/v) glycerol, 10 mM MgCl 2 , 14 mM 2-mercaptoethanol, and 1 mg mL 21 bovine serum albumin. Incubations were stopped with 800 mL of methanol containing 0.2 N HCl. For kinetic studies, recombinant AOMT (200 ng) was incubated in a final volume of 25 mL as above, with the exception that 100 mM [ 14 C]SAM (8 mCi mmol 21 ) was used. Ranges of anthocyanin and flavonol substrate concentrations between 3 and 500 mM were used for K m determination. Reactions were stopped by the addition of 75 mL of 2 N HCl and incubated at 95°C for 30 min for hydrolysis of glucosides. Reaction products were extracted with 100 mL of isoamyl alcohol, and the incorporated radioactivity was measured by liquid scintillation. K m and V max values were calculated from Lineweaver-Burk plots.

Subcellular Localization of AOMT
The AOMT coding sequence was amplified by PCR using the upstream primer 5#-GGGGACAAGTTTGTACAAAAAAGCAGGCTTGGTTCCGCGT-GGATCCATGTCCAGCTCAAGTCATAGG-3# and the downstream primer 5#-GGGGACCACTTTGTACAAGAAAGCTGGGTCATAGAGGCGCCTGCA-GAGCGCGAC-3#. The amplified DNA fragment was first introduced into pDONR 207 (Invitrogen) and then into the Gateway-compatible pK7FWG2 vector (Karimi et al., 2002) in order to obtain a C-terminal GFP fusion protein (AOMT::GFP). The pK7FWG2-AOMT construct was transformed into tobacco using agroinfiltration as described above. A nontargeted GFP construct (mGFP4) was used as a control (Haseloff et al., 1997). Tobacco leaves were examined 48 h after transformation using a Zeiss LSM 510 confocal imaging system attached to a Zeiss Axioplan 2 microscope (Carl Zeiss). GFP was visualized by excitation with the 488 line of a krypton/argon laser and use of a band-pass 505 to 550 emission filter. Projections of optical sections were accomplished using LSM image-processing software (Zeiss).

Analysis of AOMT Reaction Products
The in vitro enzymatic reactions (200 mL) were diluted with 800 mL of methanol containing 0.2 N HCl. The transformed tobacco tissues were resupended in methanol containing 0.2 N HCl (1 mg tissue fresh weight mL 21 solvent). Samples were stirred on a Stuart Tube Rotator SB3 (Bibby Sterilin) in the dark at room temperature for 1 h, centrifuged (13,000g, 15 min, 4°C), and then dried under vacuum at 35°C for 2 h (Genevac). Dried residues were dissolved in 100 mL of 20% methanol containing 1% HCl before HPLC analyses.
Sequence data from this article have been deposited with the GenBank/ EMBL data libraries under accession number FJ460168 (AOMT from Syrah). The Nebbiolo sequence differs from Syrah in two single nucleotide polymorphisms (A216C and T441C).

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Quantitative real-time PCR expression profiling of VvCCoAOMT during Nebbiolo berry development.
Supplemental Figure S2. SDS-PAGE of the recombinant AOMT.
Supplemental Table S1. In vitro substrate specificity of the recombinant AOMT.