Anthocyanidin Reductase Gene Expression and Accumulation of Flavan-3-ols in Grape Berry

Plant materials.

Berries of Vitis vinifera cv. Cabernet Sauvignon at different stages of development and ripening were sampled at random from four vines. After being peeled and deseeded, the berry skins and seeds were immediately frozen in liquid nitrogen and kept at −80°C until used. The grapevine used was grown in our experimental vineyard in Higashi-Hiroshima, Japan.

Genomic sequencing of grape ANR gene.

A genomic fragment of VvANR was amplified by polymerase chain reaction (PCR) from the total DNA of Cabernet Sauvignon. For PCR, the sense (5′-ACTTCAGATGTCCCCAGCAG-3′) and antisense (5′-CCCTTGGCCTTGAAATACTC-3′) primers were designed in the coding region of the grape cDNA clone, VitiA765 (VitiGen, Siebeldingen, Germany), annotated as BAN. PCR was performed under the cycle conditions of 0.5 min at 94°C, 0.5 min at 52°C, and 1.5 min at 72°C. An amplified fragment of approximately 900 bp was cloned into the pCR2.1-TOPO vector (Invitrogen, Carls-bad, CA) and sequenced using ABI PRISM 3100-Avant Genetic Analyzer (Applied Biosystems, Foster City, CA). The upstream and downstream sequences of the amplified fragment then were obtained by DNA walking using the Universal GenomeWalker Kit (BD Biosciences Clontech, Palo Alto, CA). The obtained sequences were analyzed using the GENETYX-MAC software (GENETYX, Tokyo, Japan).

Southern blot analysis.

Total DNA (10 μg) of Cabernet Sauvignon was digested with BglII, BamHI, SacI, and DraI, electrophoresed on 1.0% agarose gel and then blotted on a positively charged nylon membrane (Roche Diagnostics, Mannheim, Germany). The membrane was hybridized with a 392-base probe. The probe was amplified from the VitiA765 clone and labeled with digoxigenin using a PCR DIG Probe Synthesis Kit (Roche Diagnostics). The sequences of PCR primers were 5′-ACTTCAGATGTCC CCAGCAG-3′ (610 to 629 from the start codon) and 5′-CCCTTGGCCTTGAAATACTC-3′ (complements of 982 to 1001 from the start codon). Hybridization was performed in a hybridization solution, DIG Easy Hyb (Roche Diagnostics), at 40°C. The membrane was washed twice at room temperature for 5 min with 2 x SSC and 0.1% SDS and then washed twice at room temperature for 15 min with 0.1 x SSC and 0.1% SDS. The hybridized fragments were detected using a DIG-CSPD system (Roche Diagnostics) and a LAS-1000-plus image analyzer (Fuji Photo Film, Tokyo, Japan).

RNA extraction and mRNA quantification.

RNA was extracted from the berry skins and seeds according to the procedure described by Geuna et al. (1998). The RNA concentration was determined by its absorbance at 260 nm.

For the determination of the mRNA level of VvANR, real-time quantitative-PCR (Q-PCR) was performed using a QuantiTect SYBR Green PCR Kit (Qiagen, Hilden, Ger-many) and a GeneAmp 5700 sequence detection system (Applied Biosystems) as described in the manufacturer manuals. cDNAs of all the samples were prepared by reverse transcription of 0.5 ng/μL of total RNA with 0.125 μM of an oligo dT primer using AMV reverse transcriptase XL (Takara Bio, Otsu, Japan). Primer sequences were VvANR Forward, 5′-GCTGCTGTTACCATCAATCA-3′ (1774 to 1793 from the start codon, located in the third exon), and VvANR Reverse, 5′-GCAGGATAGCCCCAAG TAGG-3′ (complements of 1867 to 1993 from the start codon, which ranges from the third exon to the fourth exon) to amplify a 113-bp fragment. With reference to the method described by Downey et al. (2003), an ubiquitin gene (Ubiquitin) was amplified as an internal control of Q-PCR using the primers Ubiquitin Forward, 5′-TCTGAG GCTTCGTGGTGGTA-3′, and Ubiquitin Reverse, 5′-AGGCG TGCATAACATTTGCG-3′, which gave a 99-bp amplicon. The Ubiquitin primers were designed from a tentative consensus sequence TC38636 in an expressed sequence tag (EST) database of grape (http://www.tigr.org/tigr-scripts/tgi/T_index.cgi?species=grape; release 4.0 [21 Sept 2004]) constructed by The Institute for Genomic Research (TIGR). The Q-PCR mixture (20 μL) contained 1.0 μL of the cDNA template and 0.3 μM each of the forward and reverse primers, and Q-PCR was performed under the following conditions: 95°C for 15 min followed by 45 cycles at 94°C for 15 sec, at 57°C for 30 sec, and at 72°C for 30 sec. Standard DNA for the calibration curve was prepared as described by Jeong et al. (2004). Real-time Q-PCR was carried out in triplicate per total RNA sample. The average data were presented as the molar ratio of the mRNA level of VvANR to that of Ubiquitin and therefore had no units.

Extraction, fractionation, and quantification of flavan-3-ols.

The frozen skins and seeds were ground using Multi-beads Shocker (Yasui Kikai, Osaka, Japan). Phenolic compounds were extracted from the ground skins and seeds with methanol, methanol/water (80/20, v/v), methanol/water (50/50, v/v), distilled water, and acetone/water (75/25, v/v) successively as described by Bourzeix et al. (1986). The extracts were then combined. Using tC₁₈ and C₁₈ Sep-Pak Plus cartridges (Waters, Milford, MA), the seed and skin extracts were fractionated into three fractions (FI: flavan-3-ol monomers; FII: oligomeric PAs; and FIII: polymeric PAs) following the method described by Sun et al. (1998a).

The concentration of total flavan-3-ols in each fraction was determined by modified vanillin assay (Sun et al. 1998b) using a standard curve prepared with (+)-catechin (Sigma Chemical Co., St. Louis, MO). Each sample fractionation and analysis was carried out in triplicate. The flavan-3-ol concentration of each fraction was shown as (+)-catechin equivalents per gram of fresh seed or skin weight. The total amount in the skins was also shown per berry, except in the last sample.

Sequencing and characterization of the grape ANR gene.

Because ANRs are known to have structures similar to those of DFRs, PCR primers were designed on the consensus nucleotide sequences of ANR genes and, at the same time, on the sequence not found in the grape DFR gene (VvDFR, GenBank accession no. X75964). After PCR cloning and genome walking, the obtained genomic ANR sequence of Cabernet Sauvignon (VvANR, AB199315) contained five introns and 5′ and 3′ untranslated regions of 678 bp and 264 bp, respectively. Five introns of VvANR were located in the same positions as those of the Arabidopsis thaliana ANR gene (AtANR, AC005882 REGION: complement [43221 to 44652]) and VvDFR.

The nucleotide sequence of exons and 3′ and 5′ untranslated regions of VvANR showed 98.8% homology to VitiA765 and 99.8% homology to a Chardonnay cDNA, which was annotated as an ANR gene (BAN) (BN000166; Tanner et al. 2003) during our sequencing of VvANR. The deduced amino acid sequence of VvANR consisted of 338 amino acid residues and showed 98.5% and 99.7% homology to that of VitiA765 and Chardonnay ANR, respectively. Thus, VvANR was regarded as the genomic sequence of VitiA765 and the Chardonnay ANR cDNA. VvANR was also regarded as the same gene as a tentative consensus sequence TC45542, which was annotated as an ANR gene in the grape EST database of TIGR. The EST database shows that TC45542 has been detected in berries and flowers as well as in abiotic-stressed leaves.

VvANR showed about 65 to 75% homologies to ANRs of A. thaliana and Medicago truncatula and lower homologies to DFRs (Table 1⇓). ANR and DFR both are known to be NADPH/NADH-dependent reductases. VvANR, like other ANRs and DFRs, has the classical Rossman NADPH/NADH-binding domain, which conserves a consensus GXXGXXG/A sequence (Xie et al. 2004) at the 5′ amino terminal region. VvANR, as well as other ANRs, has “A” at the last position of the consensus sequence, while DFR has “G” at this position (Table 2⇓). Therefore, VvANR is highly homologous to other ANRs and conserves the characteristic sequence of ANR.

A homology search using PLACE (http://www.dna.affrc.go.jp/PLACE/signalscan.html), a database of plant cis-acting regulatory DNA elements, showed the presence of several putative cis-acting regulatory elements in the VvANR promoter region (Figure 2⇓). Some types of elements which are conserved in light-regulated genes were found in the promoter of VvANR: one G-box (CACGTG), four GT-1 sites (GRWAAW), one I-box (GATAA) (Terzaghi and Cashmore 1995), and four GATA boxes (Gilmartin et al. 1990). The presence of many elements related to light suggests that the expression of VvANR is regulated by light.

• Download figure
• Open in new tab

Figure 2

Nucleotide sequence of the grape ANR gene (VvANR) between -678 and -1 relative to the transcription start site. Several putative cis-acting regulatory elements in the VvANR promoter region are indicated.

Four Myc recognition sites (CANNTG; Abe et al. 2003) and two Myb-binding sites (CNGTTR) were also found in the promoter of VvANR. Anthocyanin biosynthesis is controlled by Myc and Myb family transcription regulators (Dooner et al. 1991, Roth et al. 1991). A petunia Myb protein (MYB.Ph3) is involved in the regulation of flavonoid biosynthesis (Solano et al. 1995). Grape Myb-related genes regulate anthocyanin biosynthesis (Kobayashi et al. 2002). Baudry et al. (2004) reported that a basic helix-loop-helix Myc-like protein and a Myb protein of A. thaliana were necessary for the correct expression of BAN. Thus, the transcription of VvANR and PA biosynthesis are probably regulated by Myc and Myb proteins.

To determine the copy number of the ANR gene in the grape genome, Southern blot analysis was performed. The hybridization probe was designed in the 3′-end region of ANR cDNA because this region shows relatively low homologies to the corresponding sequences of DFR cDNAs. The hybridization gave one or two fragments in the digests of the individual restriction enzymes (Figure 3⇓). Assuming that the two bands generated by DraI digestion are allelic, the ANR gene is a single copy in its haploid genome. At the least, VvANR is the major ANR gene, if not the only ANR gene of grape, because we have not found any other ANR isoforms in available EST databases of grape.

• Download figure
• Open in new tab

Figure 3

Southern blot analysis of the ANR gene from the grape genome. Each lane shows 10 μg of DNA of Cabernet Sauvignon digested with BglII, BamHI, SacI, and DraI. Positions of the molecular size markers are shown at left.

Accumulation of VvANR mRNA in the berry skins and seeds of Cabernet Sauvignon.

The mRNA level of VvANR was analyzed by real-time Q-PCR using the berry skins and seeds of Cabernet Sauvignon. Accumulation of VvANR mRNA was detected in the skins and seeds at the early stage of development; however, this accumulation decreased toward veraison and was hardly detected at the harvest stage (18 Sept) (Figure 4⇓). In the berry skins and seeds of Pinot noir after veraison, accumulation of the mRNA of the ANR gene was also hardly detected (data not shown). Thus, the mRNA level of VvANR is high in the skins and seeds of young berries but decreases considerably toward the ripening stage.

• Download figure
• Open in new tab

Figure 4

mRNA accumulation of the ANR gene (VvANR) in berry seeds (A) and skins (B) of Caber-net Sauvignon, 2003. The mRNA level of VvANR was measured by real-time quantitative-PCR analysis. Results are expressed as the molar ratio of the mRNA level of VvANR to that of the ubiquitin gene. Each bar represents the mean of triplicate analysis, and error bars represent the standard deviation of the mean.

Flavan-3-ol concentrations during berry development and ripening.

Concentrations of flavan-3-ol monomers and oligomeric and polymeric PAs in the skins and seeds of Cabernet Sauvignon during berry development and ripening are shown in Figure 5⇓. In all samples, the concentrations of flavan-3-ol monomers were relatively low, and most of the flavan-3-ols were polymerized and accumulated as oligomers and polymers. The proportion of polymeric fractions in the berry skins was higher than that in the seeds. The same tendency was reported by Monagas et al. (2003). The monomeric fractions from the seeds contained (+)-catechin, (-)-epicatechin, and (-)-epicatechin-3-O-gallate (HPLC analysis, data not shown). In the skins and seeds of small berries, flavan-3-ol monomers and PAs had already accumulated. From the small berry stage (25 June) until the preveraison stage (24 July), the total flavan-3-ol concentration (mg/g) increased in the berry seeds (Figure 5A⇓), while it decreased in the skins (Figure 5B⇓). However, the total amount of flavan-3-ols per berry increased during this period because the volume of whole berries increased (Figure 5C⇓). From the preveraison stage to the harvest stage, the concentration of polymeric PA fractions decreased in the skins, which might have been caused by the increasing volume of berries, degradation, and/or a structural change which made PAs unextractable.

• Download figure
• Open in new tab

Figure 5

Accumulation of flavan-3-ols in the berry skins and seeds of Cabernet Sauvignon, 2003. Flavan-3-ol concentrations of the monomeric, oligomeric, and polymeric fractions were quantified by vanillin assay. Results are expressed in (+)-catechin equivalents per gram of fresh seed or skin weight (A, B) or per berry skin (C). Each bar represents the mean of triplicate analysis, and error bars represent the standard deviation of the mean. The data of 18 Sept is missing from C.

Although we do not have detailed data of PA concentrations during ripening, it was previously reported that the concentrations of flavan-3-ol monomers and oligomers decreased after veraison in the berry skins and seeds (Jordão et al. 2001) and in whole berries (De Freitas et al. 2000). In addition, Mateus et al. (2001) reported that total extractable PAs as well as monomers, dimers, and trimers of flavan-3-ols in berry skins and seeds decreased after veraison.

Thus, our analysis and the reports from other groups show that flavan-3-ols accumulate before veraison, especially during the early stage of development, and then decrease in concentration during ripening. The change in flavan-3-ol concentration does not contradict the change in the mRNA accumulation of VvANR. These results indicate that at least a part of (-)-epicatehin and (-)-epigallocatechin biosynthesis is probably controlled by the transcription of VvANR.

Flavonoid biosynthetic enzymes from PAL to LDOX (Figure 1⇑) are shared in the synthetic pathway of 2,3-cis-flavan-3-ols and anthocyanins. Former studies of anthocyanin biosynthesis in grape berry skins showed that the mRNA levels of these genes are high at the early stage of development, decrease toward veraison, and increase again during ripening (Boss et al. 1996, Kobayashi et al. 2001). At the early stage of development, mRNA accumulation of the enzymes from PAL to LDOX coincides with PA accumulation in the berry skins. Thus, some of the mRNA accumulated at this stage likely participates in 2,3-cis-flavan-3-ol biosynthesis.

Recently, the synthetic mechanism of 2,3-trans-flavan-3-ols (such as (+)-catechin), another component of PAs, was revealed using Desmodium. Biosynthesis of (+)-catechin and that of (-)-epicatechin and (-)-epigallocatechin seems to be controlled coordinately. Further studies are needed to clarify the biosynthesis and the control mechanisms of grape flavan-3-ols and proanthocyanidins.