Introduction

Persimmons (Diospyros sp.) are distributed in tropical, subtropical and temperate regions of the world (Yonemori 1997). Oriental persimmon or kaki (Diospyros kaki Thunb.) is an important fruit in Asian countries such as Japan, China and Korea. In Japan, persimmon ranks fourth in production after citrus (satsuma mandarin and other citrus), apple, and Japanese pear (Yonemori 1997; Kanayama 2006).

Persimmon fruits are rich in vitamins. The vitamin A and C contents are almost equal to those of satsuma mandarin and strawberry, respectively. The vitamin A is derived from carotenoids, the pigments found in skin and flesh. Quantitative analysis of carotenoids shows that the main pigment in yellow- to orange-colored fruit is beta-cryptoxanthin, and that lycopene is the major pigment in red-colored fruit. These colors give the fruit an attractive appearance, which is preferred by consumers. Polyphenols (tannins) are abundant in persimmon fruit, and have an astringent flavor. Carbon dioxide and ethanol treatments are needed for the removal of astringency (Kitagawa and Glucina 1984; Taira 1996). Epicatechin, epigallocatechin, epicatechin-3-gallate, and epigallocatechin-3-gallate are components of soluble tannin, which occurs at 1% in the ripening astringent persimmon fruit (Matsuo and Ito 1978). Tannins have been shown to function as inhibitors of cancer (Fujiki 1999, 2005), heart disease (Chang et al. 2001), hypertension (Sasakawa1955; Mekhfi et al. 2006) and can ameliorate the symptoms of a hangover (Ogata 1976; Okuma et al. 1995). They are therefore in the spotlight as components of functional foods. Persimmon has been used not only as a fresh fruit, but also for many other purposes. For example, a tannin extract known as “kaki-shibu” has been used as water-proofing for fishing nets, paper and umbrellas, and as a fining agent to remove proteins in the production of Japanese sake (rice wine). Also it has also been found to have an antimutagenic effect (Matsuo et al. 1991).

Persimmon is a climacteric fruit that has a short shelf-life due to accelerated fruit softening. Ethylene plays a critical role in fruit softening of persimmon (Itamura et al. 1991). Ethylene may alter the gene expression of cell wall-degrading enzymes (Gray et al. 1992). However there is little information concerning molecular mechanisms controlling gene regulation. This softening process is accelerated after the removal of astringency, and significantly influences the acceptability of the fruit. Among persimmon cultivars, “Saijo” has the shortest shelf-life and begins to soften 4 days after removal of astringency (Itamura et al. 1997). This is the major problem in marketing ‘Saijo’ persimmon.

Analyses using ESTs have been used for a number of fruits to discover genes and reveal gene expression patterns. Through such studies, a set of ripening-related genes associated with cell walls, pigments, aroma, and ethylene, were identified in grape berry (Terrier et al. 2001; Moser et al. 2005; da Silva et al. 2005), tomato (Fei et al. 2004), pineapple (Moyle et al. 2005), apricot (Grimplet et al. 2005), apple (Newcomb et al. 2006), water melon (Levi et al. 2006) and papaya (Devitt et al. 2006). However, there is little information on genes associated with fruit development and ripening in persimmon. We report distribution of genes identified from two persimmon fruit cDNA libraries using EST sequencing and describe putative genes involved in proanthocyanidin and carotenoid synthesis.

Materials and methods

Plant materials

Ovaries and young fruit of persimmon (D. kaki Thunb. cv. Saijo) grown in Tsukisaka’s orchard at Matsue, Shimane prefecture, Japan were harvested at six different early stages of development: 1 week before flower opening (May 20, 2005); flowers just opened (May 27, 2005); 1 week after flower opening (June 1, 2005); 2 weeks after flower opening (June 8, 2005); 3 weeks after flower opening (June 15, 2005) and 4 weeks after flower opening (June 22, 2005) (Fig. 1a). These samples (each about 1 g FW) were mixed and used to construct the KA cDNA library as described below. For construction of the KC library, fruit were prepared at six different stages of ripening: stage 1 (October 7, 2005); stage 2 (October 20, 2005, commercial harvest time); stage 3 (November 3, 2005); stage 4 (November 17, 2005); stage 5 (December 1, 2005); and stage 6 (December 15, 2005) (Fig. 1b). These samples (each about 2 g FW) were mixed and used as described below.

Fig. 1
figure 1

Development and ripening of persimmon (Diospyros kaki Thunb. cv. Saijo) fruit used for construction of cDNA libraries. a Early developmental stage for KA cDNA library: 1 1 week before flower opening (May 20, 2005); 2 just opened flower (May 27, 2005); 3 1 week after flower opening (June 1, 2005); 4 4 weeks after flower opening (June 22, 2005). b Ripening stage for KC cDNA library: 1 stage 1 (October 7, 2005); 2 stage 2 (October 20, 2005, commercial harvest time); 3 stage 3 (November 3, 2005); 4 stage 4 (November 17, 2005); 5 stage 5 (December 1, 2005); 6 stage 6 (December 15, 2005)

Full size image

Preparation of cDNA libraries

Isolation of poly (A)+ RNA from persimmon samples, and construction of cDNA libraries using the Uni-ZAP XR system (Stratagene, La Jolla, CA) were carried out by a commercial service provider (Eugentech, Taejon, Korea). In brief, total RNA was prepared from ovaries and young fruit (for KA cDNA library) by the hot phenol method, and total RNA from fruit (for KC cDNA library) was prepared by the pine tree method (Chang et al. 1993). Poly (A)+ RNA was purified using a PolyATtract mRNA isolation system (Promega, Madison, WI, USA). Double stranded cDNA was produced from approximately 5 μg of poly (A)+ RNA using a cDNA synthesis kit (Stratagene). The cDNAs larger than 0.5 kb were selected by gel filtration and unidirectionally ligated into the EcoRI–XhoI sites of the UNI-ZAP XR vector. Libraries were packaged using GigaPackIII Gold packaging extract (Stratagene). The titers of the primary cDNA library were 2 × 107 pfu/ml for KA, and 1.5 × 106 pfu/ml for KC. The percentage of recombinants was approximately 90% in each library. Because the cDNA inserts were inserted in the same direction, sequencing from the 5′-end was carried out with T3 primer.

PCR and sequencing

Primary phage libraries were plated using E. coli strain XL1-Blue MRF’ as a host. Each plaque was suspended in 50 μl of SM (100 mM NaCl, 8 mM MgSO4, 50 mM Tris-HCl pH 7.5, 0.01% gelatin) in a 96-well microplate (stock plate). A 2 μl aliquot of each phage solution was transferred to the corresponding position of new 96-well microplate (reaction plate) compatible with the Genetic Analyzer 3100 DNA sequencer (Applied Biosystems, Foster City, CA, USA) and used as a template for amplification of cDNA inserts. The primers used for PCR were T3 (5′-GAATTAACCCTCACTAAAGG-3′) and T7 (5′-TAATACGACTCACTATAGGG-3′). Cycling parameters were as follows: (1) 95°C for 5 min; (2) 35 cycles of 94°C for 1 min, 56°C for 1 min, 72°C for 3.5 min; (3) 72°C for 7 min. Taq DNA polymerase with Thermal Pol Buffer (New England Biolabs, Ipswich, MA, USA) was used for amplification. After PCR, an equal volume of polyethylene glycol solution (13% PEG 6,000 and 1.6 M NaCl) was added and stored at 4°C for at least 3 h. Amplified products were precipitated by centrifugation at 2000g for 30 min, washed with 70% ethanol and used for sequence reactions. PCR products were sequenced with the T3 primer, the BigDye Terminator ver 3.1 Cycle Sequencing kit (Applied Biosystems) and an ABI PRISM Genetic Analyzer 3100 (Applied Biosystems). All the reactions were carried out using the reaction microplate described above. In total, 6,000 ESTs were obtained from KA and 6,100 ESTs from KC.

Sequence analysis

Vector-derived and ambiguous sequences (PHRED quality value <20) were eliminated using a combination of the Phred program (Ewing et al. 1998) and cross-match software (http://www.phrap.org). Poly (A) tail sequences in the ESTs processed by our Perl script were at most 10 bases. Subsequently, ESTs with sequences of less than 20 bp were omitted from the final data set. The ESTs were assembled into non-redundant sequences using the phrap program (http://www.phrap.org). Using the BLAST program (Altschul et al. 1990), the sequences of the non-redundant sequences were searched against the Arabidopsis protein database in TAIR and non-redundant protein database obtained from NCBI with a search threshold of 1e-10. Results from the BLAST program were parsed with PHP and stored in a PostgreSQL database management system. A publicly accessible web-based database (http://bio14.ipc.shimane-u.ac.jp/kakiest/) was also built on PHP and PostgreSQL. Gene ontology (GO) terms for each non-redundant sequence were also assigned according to those for the Arabidopsis genes (Rhee et al. 2003; Berardini et al. 2004) with sequence similarities. ESTs greater than or equal to 50 bp were submitted to the DDBJ database (accession numbers DC583395–DC592851).

Results and discussion

EST analysis

Using the Phrap program (http://www.phrap.org), we assembled 9,952 ESTs into a set of 6,700 non-redundant sequence groups that included 5,211 singletons. The length of the EST sequences ranges from 20 to 637 bp, with an average of 394 bp, and the length of non-redundant sequences ranges from 20 to 1945 bp, with an average of 420 bp. These average lengths of sequences are shorter than those of Vitis vinifera (463 bp: Moser et al. 2005; 527 bp: da Silva et al. 2005). Most of the contigs obtained (96.1%) were made up by 1–5 ESTs, and the remained contigs contained 6–10 (2.8%), 11–15 (0.5%) and more 15 ESTs (0.6%). Average number of ESTs per contig was 1.48. Of the 6,700 non-redundant sequences evaluated, the deduced amino acid sequences of 4,356 (65%) showed significant homology to known proteins and 2,344 (35%) showed no significant similarity to any known proteins in Arabidopsis database. When non-redundant sequences were assigned GO terms of the homologous Arabidopsis genes in the TAIR database, the molecular function for KA and KC libraries, respectively, were classified as follows: hydrolase activity (10.8 and 9.7%), transferase activity (8.1 and 7.5%), protein binding (7.6 and 7.6%), nucleotide binding (5.5 and 4.6%), transporter activity (5.3 and 5.9%), DNA or RNA binding (5.3 and 5.3%), kinase activity (4.2 and 3.9%) (Table 1). There was no significant difference in the population of each molecular function between KA and KC libraries.

Table 1 Distribution of GO-based molecular function for each library
Full size table

Distribution of highly expressed genes in each library

Non-redundant sequences which were assembled with more than 10 ESTs (cluster) in early and late stages are shown in Tables 2 and 3. Of the classes of ESTs with assigned functions, elongation factor 1-alpha, copper/zinc superoxide dismutase, glyceraldehyde-3-phosphate dehydrogenase and aspartyl protease family protein, were present abundantly at both stages. The copper/zinc superoxide dismutase that is known to be a reactive oxygen scavenger was abundant in both early and late stages. Sequences with similarity to the reactive oxygen scavengers catalase and glutathione S-transferase were also found abundantly in ripening fruit cDNA libraries of papaya (Devitt et al. 2006). In early stages, 12 different types of ESTs were found in multiple copies; the most abundant was polyubiquitin with 26 copies, except common ESTs between KA and KC library (Table 2). Other highly abundant sequences in early stages included the vitamin C defective protein, mutase family protein, water channel, histone H3.2, chalcone synthase (CHS), fructose-bisphosphate aldolase and anthocyanidin synthase (ANS). CHS and ANS were genes of proanthocyanidin synthesis pathway as mentioned below. There were 18 different types of EST found in multiple copies in late stages, with 220 basic chitinase ESTs as the most abundant clone (Table 3). At the late stages there were abundant sequences of basic endochitinase, osmotin-like protein, glutaredoxin, protease inhibitor, beta-fructofuranosidase, invertase/pectin methylesterase inhibitor, lactoylglutathione lyase, homology-dependent gene silencing 1, tonoplast monosaccharice transporter 2, pathogenesis related protein, acidic endochitinase, ubiquitin-protein ligase, tubulin alpha-2 chain and fibrillin except common ESTs between KA and KC library. Four plant defense-like proteins, basic endochitinase, osmotin, acidic endochitinase and pathogen-related protein were highly represented in late stages with 220, 104, 15 and 14 ESTs, respectively. Basic endochitinase was also reported to be the most abundant EST in ripening fruit cDNA libraries of papaya (Devitt et al. 2006). A sequence similar to the invertase/pectin methylesterase inhibitor that was abundant in the late stage cDNA library was also found in cDNA libraries of half-ripe and fully ripe apricot (Grimplet et al. 2005). This protein inhibits pectin methylesterases and invertases through the formation of a non-covalent complex (Giovane et al. 1995). It is thought to be involved in the regulation of carbohydrate metabolism, cell wall extension and susceptibility to pathogen attack (Camardella et al. 2000; Chen et al. 2000). At the late stage, 30 ESTs encoded the beta-fructofuranosidase protein, which converts sucrose consistent with increased contents of glucose and fructose during ripening in “Saijo” fruit (Zheng and Sugiura 1990). Fibrillin, encoded by 11 ESTs at the late stage, is a plastid-localized lipid-binding protein that is regulated by abscisic acid (ABA) and involved in photoprotection (Yang et al. 2006). The ABA content of persimmon fruit increases during ripening (Hirano et al. 1995), therefore fibrillin gene expression may be correlated with elevated ABA content in persimmon fruit. Interestingly, the KC library contained many ESTs that show no homology to genes of Arabidopsis (data not shown). These ESTs may be specific to fruit ripening in persimmon. Gene expression during development and ripening in persimmon fruit should be investigated in future studies.

Table 2 Assembled clusters that contain more than ten ESTs at early stages in persimmon
Full size table
Table 3 Assembled clusters that contain more than ten ESTs at late stages in persimmon
Full size table

Genes controlling softening pigmentation, and astringency in persimmon fruit

Both the KA and KC libraries from persimmon contained ripening-related EST clones, including 1-aminocyclopropane-1-carboxylate (ACC) synthase and ACC oxidase, both of which are associated with the ethylene biosynthesis (data not shown). They also both contained cellulose synthase, expansin, pectinesterase, polygalacturonase, xyloglucan:xyloglucosyl transferase and beta-galactosidase, which are all implicated in cell wall metabolism (biogenesis, modification, loosening, etc.). However, each EST differed in its make up of clones related to proanthocyanidin (tannin) and carotenoid synthesis. In “Saijo” persimmon the ovary and young fruit produce high amounts of tannins at the early stage (KA), and the pulp accumulates beta-cryptoxanthin and lycopene at the late stage (KC). The ESTs related to proanthocyanidin and carotenoid pathways in persimmon are shown in Table 4, and the biosynthetic pathways are shown in Figs. 2 and 3. These ESTs related to proanthocyanidin synthesis, including clones with similarity to two phenylalanine ammonia-lyases (PAL), one cinnamate 4-hydroxylase (C4H), two 4-coumarate: CoA ligases (4CL), one CHS, two chalcone isomerases (CHI), one flavone 3-hydroxylase (F3H), one dihydroflavonol 4-reductase (DFR), one ANS, one flavonoid 3′,5′ hydroxylase (F3′5′H), one leucoanthocyanidin reductase (LAR, BANYULS homolog), and one anthocyanidin reductase (ANR). These ESTs also contained clones related to carotenoid biosynthesis with similarity to one geranylgeranyl pyrophosphate synthase (GGPS), one phytoene desaturase (PDS), one zeta-carotene desaturase (ZDS), one beta-carotene hydroxylase (HYD-B), and one 9-cis-epoxycarotenoid dioxygenase (NCED).

Table 4 Diospyros kaki ESTs of proanthocyanidin and carotenoid synthesis
Full size table
Fig. 2
figure 2

Proanthocyanidin biosynthesis pathway. PAL Phenylalanine ammonia-lyase; C4H cinnamate 4-hydroxylase; 4CL 4-coumarate: CoA ligase; CHS chalcone synthase; CHI chalcone isomerase; F3H flavanone 3-hydroxylase; DFR dihydroflavonol 4-reductase; ANS anthocyanidin synthase; F3′H flavonoid 3′-hydroxylase; F3′5′H flavonoid- 3′,5′-hydroxylase; LAR leucoanthocyanidin reductase; ANR anthocyanidin reductase. Closed boxes show genes encoding enzymes that were isolated from persimmon ESTs in this study

Full size image
Fig. 3
figure 3

Carotenoid biosynthesis pathway. GGPS Geranylgeranyl pyrophosphate synthase; PSY phytoene synthase; PDS phytoene desaturase; ZDS ζ-carotene desaturase; CRTISO carotenoid isomerase; LCYb lycopene β-cyclase; LCYe lycopene ε-cyclase; Hyb β-ring hydroxylase; Hye ε-ring hydroxylase; ZEP zeaxanthin epoxidase; NCED 9-cis-epoxycarotenoid-dioxygenase. Closed boxes show genes encoding enzymes that were isolated from persimmon ESTs in this study

Full size image

Eleven genes of the flavonoid pathway were included in the KA library but in the mature persimmon library (KC) only 4CL, CHS and F3H were present. Recently, seven genes (PAL, C4H, CHI, F3H, F3′H, ANS and ANR) have been isolated from an astringent-type cultivar using suppression subtractive hybridization (Ikegami et al. 2007). Ikegami et al. (2005a, b) reported that transcription of PAL, C4H, 4CL, CHS, CHI, F3H, F3′H, F3′5′H, DFR and ANR genes is high until the middle of August, and declined in October compared to transcript levels in June and July in the astringent-type cultivar. These results correspond to the relative abundance of proanthocyanidin-related genes between the KA and KC libraries. Clones with similarity to PDS, ZDS HYD-B and NCED were isolated from yellowish and orange pulp in persimmon (Table 4). Niikawa et al. (2007) isolated partial fragments encoding seven genes by RT-PCR technique; phytoene synthase, PDS, ZDS, lycopene beta cyclase, HYD-B, lycopene epsilon cyclase (LCY-E) and zeaxanthin epoxidase (ZEP) from “Fuyu” persimmon. They reported that expression of all genes except ZEP and LCY-E increased during fruit ripening. This result also corresponds to differences in the number of carotenoid-related genes from the KA and KC libraries. Also, library from papaya ripened fruit, which produce beta-cryptoxanthin, contained orthologs of GGPS, ZDS and HYD-B (Devitt et al. 2006).

We successfully isolated unique ESTs from persimmon fruit, including two new genes, LAR and NCED. We are now constructing a microarray set that includes the cDNAs from the KA and KC cDNA libraries. These persimmon microarrays for the fruit ESTs will enhance our knowledge of fruit development and ripening.