Molecular identification of 1-Cys peroxiredoxin and anthocyanidin/flavonol 3-O-galactosyltransferase from proanthocyanidin-rich young fruits of persimmon (Diospyros kaki Thunb.)

Abstract

Fruits of persimmon (Diospyros kaki Thunb.) accumulate large amounts of proanthocyanidins (PAs) in the early stages of development. Astringent (A)-type fruits remain rich in soluble PAs even after they reach full-mature stage, whereas non-astringent (NA)-type fruits lose these compounds before full maturation. As a first step to elucidate the mechanism of PA accumulation in this non-model species, we used suppression subtractive hybridization to identify transcripts accumulating differently in young fruits of A- and NA-type. Interestingly, only a few clones involved in PA biosynthesis were identified in A-NA libraries. Represented by multiple clones were those encoding a novel 1-Cys peroxiredoxin and a new member of family 1 glycosyltransferases. Quantitative RT-PCR analyses confirmed correlation of the amount of PAs and accumulation of transcripts encoding these proteins in young persimmon fruits. Furthermore, the new family 1 glycosyltransferase was produced in Escherichia coli and shown to efficiently catalyze galactosylation at 3-hydroxyl groups of several anthocyanidins and flavonols. These findings suggest a complex mechanism of PA accumulation in persimmon fruits.

Fig. 1

Chemical structures of flavonoid aglycones relevant to this study. Flavan-3-ols are monomers of proanthocyanidins (PAs), whereas anthocyanidins are aglycones of anthocyanins, which contain sugar moieties attached to various hydroxyl groups at A, B, and C rings

Fig. 2

Estimation of PA contents in non-astringent- and astringent-type persimmon fruits at the early developmental stages. Fruits from three independent trees of non-astringent (NA)- and astringent (A)-type persimmon were collected on three different dates (June 12, July 12, and August 3, 2001) and the amount of 80% methanol-soluble and insoluble PAs in these fruits were measured as (+)-catechin equivalent by DMACA method (a) or cyanidin equivalent by n-butanol–HCl (b) method. The bar graphs indicate the average PA contents of NA- (white bars) or A- (black bars) type fruits. Error bars are standard deviations. Asterisks indicate that the differences are significant (P < 0.002)

Fig. 3

Accumulation of a subset of transcripts in non-astringent and astringent-type persimmon at the early developmental stages. Accumulation of a subset of transcripts identified by SSH analysis (Table 1) as well as a that of ANR was examined by real time PCR with RT products from fruits of three independent non-astringent (NA)- and astringent (A)-type persimmon trees collected on three different time points (June 12, July 12, and August 3, 2001). The bar graphs indicate the average level of amplified cDNA, normalized to the amount in NA-type fruits of June 12, in NA- (white bars) or A- (black bars) type fruits. Error bars are standard deviations. An asterisk and double asterisks indicate that P values to indicate the significance of differences between the transcript levels in NA- and A-type are below 0.002 and 0.02, respectively

Fig. 4

Phylogenetic analysis of peroxiredoxins from various plant species. a One of four most parsimonious trees from phylogenetic analysis of partial amino acid sequences of peroxiredoxins listed in Table S3. Numbers above branches are bootstrap support values >50%. Asterisks indicate branches that collapsed in the strict consensus tree. Six distinct groups are indicated on the right. b Alignment of amino acid sequences in the middle region of 1-Cys Prxs (top six) and one protein each from five other groups in A. thaliana as shown in A (bottom five). The catalytic cysteine residue conserved in all the Prxs is indicated with an asterisk, whereas the residue conserved only in 1-Cys Prx is indicated with a plus symbol. The second cysteine residues conserved in PrxQ and type II Prx (Rouhier and Jacquot 2002) are indicated as white characters on a black background. c Alignment of C-terminal region of 1-Cys Prxs. Basic residues, which may play roles in nucleus-targeting, are indicated as white characters on a black background

Fig. 5

Phylogenetic analysis and comparison of glycosyltransferases from various plant species. a One of three most parsimonious trees from phylogenetic analysis of complete amino acid sequences of GT listed in Table S4. Numbers along branches are bootstrap support values >50%. Asterisks indicate branches that collapsed in the strict consensus tree. Fourteen groups (A–N) defined by Bowles et al. (2005) are indicated. b Alignment of the regions near C-termini of GTs encompassing the Plant Secondary Product Glycosyltransferase (PSPG) box. The top panel figure was generated with Weblogo (Crooks et al. 2004), based on the sequences analyzed in a. The bottom panel shows the sequence alignment of group F proteins. Residues conserved in at least five proteins are shown as white characters on a black background. The histidine residue conserved among FGalTs is indicated with an asterisk in the bottom. Three conserved residues unique to this group are indicated with arrows

Fig. 6

Biochemical characterization of DkFGT. a Production of GST-DkFGT by E. coli. Proteins in the soluble fraction from E. coli transformed with pGEX-KG-DkFGT (lane 1) and those in the fraction bound to and eluted from glutathione agarose column (lane 2) were separated by SDS-PAGE and visualized by Coomassie blue staining. The 77-kD band, which cross-reacted with the αGST antibody by immunoblotting (data not shown), and thus corresponded to the intact GST-DkFGT, is indicated with an arrow. The smaller bands of 30–40 kD, which also cross-reacted with the αGST antibody, are indicated with an asterisk. The black line separates images from different portions of the same gel. b HPLC elution profiles of GT assay mixtures with various combinations of proteins and substrates. The reactions were done for 75 min at 30°C. GST protein extracts purified with a glutathione column from E. coli transformed with pGEX-KG, GST-DkFGT fraction 2 in a, Q quercetin, Q3Gal quercetin 3-O-galactoside, K (IC) kaempferol (as an internal control). c Temperature optimum of DkFGT activity. d pH optimum of DkFGT activity. e, f HPLC elution profiles of GT assay mixtures. The reactions were done with GST-DkFGT for 5 min (e), or with recombinant proteins indicated for 15 min (f)