As part of our horticultural program, we have crossed various cultivars of D. kaki and obtained their progenies segregating for fruit astringency. One of these crosses was made between two genetically distinct NA-types of D. kaki, cv. Luo Tian Tian Shi and cv. Taishu. Their F1 progenies were segregated into A- and NA-types, the latter of which appears to be determined by the presence of a dominant allele derived from cv. Luo Tian Tian Shi . In this report, we made use of these materials, attempting to elucidate the molecular mechanism of soluble PA accumulation in young persimmon fruits. We employed suppression subtractive hybridization to identify transcripts whose accumulation patterns were distinct between the segregated A- and NA-type fruits when their PA contents became distinct. Interestingly, only a few clones encoding Xavonoid biosynthetic enzymes were identified. Among cDNAs represented by multiple clones were those encoding a new member of the 1-Cys peroxiredoxin family and that of subgroup F of family 1 glycosyltransferases . A detailed sequence comparison and phylogenetic analysis revealed that the isolated 1-Cys Prx may be a novel type. In addition, UDP-galactose:anthocyanidin/Xavonol 3-O-galactosyltransferase activity of the GT homolog was conWrmed by using a bacterial expression system. These results may suggest complex mechanisms of PA accumulation in persimmon fruits.Generation of F1 progenies between the two distinct NAtype persimmon cultivars , their growth, and collection of their fruits were partly described by Ikegami et al. . Fruits usually become fully mature and ready to harvest for human consumption in the end of October to November. Immature green fruits used in this study were harvested considerably earlier, at three different dates , 25 liter plant pot during which PA accumulation started to decline in the NA-type but not in the A-type . The A- and NA-types used in this study were defined previously based on the size of PA-accumulating cells and the soluble tannin content of fully mature fruits . Fruit fresh was separated from seeds and peel, cut into small pieces , frozen with liquid nitrogen, and stored at ¡80°C before further analysis.

Total RNA was isolated from 5 g of the frozen sample using the hot borate method .Amino acid sequences were aligned with ClustalX . Phylogenetic analyses of the aligned amino acid sequences based on maximum parsimony were implemented in PAUP* with heuristic searches using the TBR branch-swapping algorithm 1,000 random taxon addition replicates and no limit on the number of trees saved. Relative support for clades was assessed using 1,000 bootstrap replicates with 10 random taxon addition replicates per bootstrap replicate. For the Prx sequences, 191 ambiguously aligned sites were excluded from the original alignment of 357 sites. A Wnal data set of 166 characters, of which 3 were constant, 6 were variable but parsimony-uninformative, and 157 were parsimony-informative, was subjected tophylogenetic analysis. The tree shown in Fig. 4a is one of four most parsimonious trees , arbitrarily rooted along a strongly supported internal branch. For the GT sequences, the complete alignment of 610 amino acid characters, of which 61 characters were constant, 62 were variable but parsimony-uninformative, and 487 were parsimony-informative, was used in phylogenetic analysis. The unrooted phylogram shown in Fig. 5a is one of three most parsimonious trees .The astringency of A-type fruits and that of NA-type fruits became distinct in those harvested in August. That is, the soluble PA content per dry weight of A-type fruits remained almost constant from June to August, whereas that of NA-type fruits dropped by more than 50% . By contrast, during this period, the concentrations of insoluble PAs remained largely unchanged and at comparable levels between the two fruit types ; and fresh weight of a fruit of both types increased similarly up to five times . These data suggest constant PA biosynthesis in A-type, but not in NA-type fruits.

It may also be possible that the level of Xavonoid oxidation, which has been shown to be negatively correlated with PA solubility , was lower in A-type than in NA-type. Hence, we decided to compare the transcript accumulation patterns in NA- and A-type fruits harvested in July and August. To identify differentially accumulating transcripts, we used suppression subtractive hybridization , which recently allowed us to isolate several transcripts including those encoding PA-biosynthetic enzymes whose accumulation levels were up- or down-regulated in persimmon fruits by ethanol treatment to remove astringency . We prepared RNAs from A- and NA-type fruits and generated reciprocal cDNA SSH libraries for samples of each data point. Among 4,800 recombinant clones that were randomly selected from these libraries, 37 clones showed significantly different accumulation patterns between the two fruit types by RNA dot blot assays, and their nucleotide sequences were determined . Based on the previous results , we had expected to find a number of PA-biosynthetic genes in A–NA libraries. Indeed, a total of nine independent clones involved in phenylpropanoid metabolism were isolated . However, none of them encodes an enzyme catalyzing one of the committed steps for PA biosynthesis, i.e., LAR or ANR. Some other clones identified in these libraries encode homologs to known proteins that do not play a direct role for PA biosynthesis . We also attempted to obtain full-length coding sequences of SSH clones isolated from A–NA libraries in order to facilitate further analysis. To this end, we screened a cDNA library from A-type persimmon fruits, and were able to obtain apparent full-length sequences for a subset of cDNAs . The presence of clones for the three Xavonoid biosynthetic enzymes, phenylalanine ammonia lyase , chalcone synthase , and dihydroXavonol 4-reductase , in A–NA libraries was consistent with the preliminary RNA-blotting data . Also identified in A-NA libraries were clones for two other Xavonoid biosynthetic enzymes, cinnamic acid 4-hydroxylase and Xavonoid 3 5 -hydroxylase .

Three truncated sequences may derive from nonoverlapping portions of a single mRNA encoding a C4H homolog . In addition, three SSH clones from the July sample were found to encode 3-dehydroquinate dehydratase/shikimate 5-dehydrogenase , which is involved in the biosynthesis of aromatic amino acids to fuel the phenylpropanoid pathway. Finally, represented by the highest number of clones for both the July and August samples were the transcripts encoding a protein with high sequence identities to various plant Xavonoid GTs, such as UDPGlc:anthocyanidin 3-O-GlcT from Xower buds of Lobelia erinus L. , anthocyanin 3-O-GalT from cell suspension culture of Aralia cordata Thunb. , and kaempferol 3-O-GalT from pollen of Petunia x hybrida . The group of cDNAs not directly involved in phenylpropanoid metabolism consists of a total of eight independent clones. Three of them were found in A–NA libraries, encoding proteins similar to glucose acyltransferase , 1-Cysperoxiredoxin , and glutathione S-transferase , respectively. SCPL might be involved in PA accumulation, since correlation of its gene expression and accumulation of PAs has been reported at least twice, one in persimmon fruits and another in hairy roots of grape overproducing MYB transcription factors . The clone encoding a 1-Cys Prx homolog was represented by five SSH clones, whereas GST, which was shown to have multiple functions including conjugation to anthocyanidins in the cytoplasm for vacuole sorting , was represented by a single clone . Among the five cDNAs found in NA–A library, three clones may encode non-overlapping portions of a single protein with a high sequence similarity to LATE BLOOMER 1 from pea, which plays roles in photoperiodic Xowering, de-etiolation, and circadian regulation . The other two appear to encode distinct chitinases, among which is one for a class II chitinase-like protein, black plastic plant pots which was represented by Wve SSH clones.The citrus fruit, termed hesperidium, is a fleshy fruit which, like all berry-type fruit, is characterized by a thick and fleshy pericarp . The pericarp is usually divided into three tissues: the exocarp, which is the outer skin, the mesocarp, which usually refers to the major fleshy, edible interior, and the endocarp, an internal tissue composed of one or several cell layers. In true fruit, which develop from the ovary, these three tissues are part of the ovary wall. The exocarp of citrus fruit is the outer colored peel, often referred to as the flavedo . Proceeding inward is the albedo, the spongy white part of the peel. Most cell layers of the albedo are considered to be mesocarpal tissue, and the two or three innermost cell layers are referred to as endocarp . In mandarins, the albedo disintegrates during fruit maturation, leaving only the vascular system , which gives this group its name, Citrus reticulata. The pulp, the edible part of the fruit, is composed of juice sacs/vesicles that develop from the endocarp at an early stage of fruit development . Some authors refer to the juice sacs as endocarp, while others consider them to be a separate tissue. The juice sacs develop into the ovary locule, defined as the section in which the ovary wall that develops into fruit. The carpel and the juice sacs are covered by the same epidermal layer of segment epidermis . The juice sac is connected to the wall by a stalk, which joins the segment epidermis, so the latter provides one continuous layer covering both the segment and the juice sac.

Three major vascular bundles, a dorsal and two side bundles, are found in each section. Most juice sacs initiate from the dorsal wall, but some develop from the side wall, adjacent to the side vascular bundle . When present, seeds develop in the inner side of the fruit, where the carpels merge or along the ovary wall. Nutrition is supplied by a specific bundle, termed seed bundle, reaching from the fruit pedicle to the center of the fruit.The juice sac is a unique structure, found only in fruit of the genus Citrus and its close relatives. It is often referred to as a “sac of juice,” but this is misleading; the juice sac is composed of various layers of cells, each with distinct morphology . The vesicle primordia emerge from the endocarp soon after fertilization and fruit set. In a few cases, juice sac primordia are visible even before fertilization and fruit set, mainly when fertilization does not occur and parthenocarpic fruit develop . During fruit development, the vacuole of the juice sac cell becomes greatly enlarged, occupying over 90% of the total cell volume, and releases its content as juice. At fruit maturity, the vacuole contains about 100, 75, and 90% of the total cellular sucrose, hexose and citrate, respectively . The juice sac is considered the major fruit sink; however, it is disconnected from the vascular system, which ends in the albedo . This unique characteristic determines photo assimilate translocation rate into the sink cells and therefore, rate of fruit development, and the time required to reach maturity.In many citrus cultivars, the major external change that marks the conversion of the citrus ovary into a fruit let is usually petal fall . Fruit development is divided into three overlapping stages: cell division , cell expansion , and fruit maturation . During stage I, fruit growth is relatively moderate, and the peel, especially the albedo, thickens by cell division. During this stage, juice sacs grow out via cell division into the locule. Stage II is characterized by rapid fruit growth, mostly due to juice cell expansion. During stage III, the rate of fruit volume increase is greatly reduced. Externally, the major change is color break, and internally, sugar and acid levels reach the desired levels for harvesting and consumption, as discussed further by Spiegel-Roy and Goldschmidt . Changes in secondary metabolites give the fruit its unique aroma and flavor . As there is no respiration burst or autocatalytic ethylene production, the citrus fruit does not undergo the classical ripening process, typical of climacteric fruits. For a given citrus cultivar, the final flavor quality of the fruit has to be determined empirically and depends, largely, on consumer preference . The completion of fruit development is cultivar-dependent, with some cultivars, such as Satsuma mandarin , being ready for harvest 5–6 months after flowering, whereas others, such as Valencia orange , are harvested 12–14 months after flowering . In hot climates, fruit development is accelerated, potentially reducing the time needed for fruit maturation by ca. 50% .