Introduction

Pears (Pyrus communis, European pear, P. bretschneideri, P. ussuriensis, Chinese pears, and P. pyrifolia, Asian pear or Nashi) are important pome fruits, since they are favorable foodstuff due to their delicious flavor and their manifold cultivars. World production of pear fruits is about 17 million tons per year. Currently, pear cultivation is continuously rising worldwide and drastically expanding in Asia (Fischer and Weber 2005). Despite this, they were only rarely investigated at the molecular level yet.

Flavonoids, polyphenols in general, are of great importance in plants. Apart from their many biological functions such as pollinator attraction, pollen fertility, UV protection, regulation of polar auxin transport, establishment of microbial symbioses, and pathogen defense, polyphenols contribute to or even determine additional features that are of special relevance in fruit crops (Schijlen et al. 2004). These include color and flavor, which both strongly influence fruit attractiveness of ripe fruits for herbivores. With respect to polyphenols, color is chiefly determined by anthocyanins, but also by flavonols, which act as their co-pigments. Fruit flavor is influenced by, e.g., bitter-masking flavanones (Ley et al. 2005) and astringent proanthocyanidins (polymeric flavan 3-ols), the latter are primarily present in unripe fruits and contribute to herbivore deterrence (Harborne 1993). Furthermore, the contribution of secondary metabolites such as polyphenolic compounds to pathogen resistance is of special interest in fruit crops (Treutter 2005). Currently, the health-promoting effects of flavonoids in food are a highly debated topic (e.g., Punyasiri et al. 2004; Mendoza-Wilson and Glossmann-Mitnik 2006; Yilmaz 2006).

Pears contain a broad spectrum of phenolic compounds comprising different flavonoid classes [anthocyanins, flavonols, monomeric (‘catechins’), and polymeric flavan 3-ols (proanthocyanidins, syn. condensed tannins), and flavanones], hydroxyphenolic acids (mostly hydroxycinnamic acids derived from caffeic acid and p-coumaric acid) and the p-hydroquinone-glucoside arbutin (Fig. 1) (Challice and Williams 1968; Macheix et al. 1990; Amiot et al. 1995; Andrade et al. 1998; Schieber et al. 2001; Petkou et al. 2002; Andreotti et al. 2006). In leaves arbutin, its derivatives, and hydroxycinnamic acids are dominating. In fruits hydroxycinnamic acids are the dominating polyphenols. Among flavonoids of leaves and fruits B-ring dihydroxylated flavonol derivates (quercetin and isorhamnetin) and monomeric and polymeric flavan 3-ols (epicatechin and proanthocyanidins) are dominant. These compounds contribute to color, fruit quality, and plant resistance. Cloning of pear flavonoid cDNAs was performed to provide the molecular biological information on this subject for fundamental knowledge and for practical application.

Fig. 1
figure 1

Biosynthesis of the main constitutive flavonoids, phenols and polyphenols in European pear (Pyrus communis). End products are shown in bold. Beyond the specific flavonoid pattern, the occurrence of arbutin and the lack of dihydrochalcones (abundant in Malus) are Pyrus specific. Enzyme abbreviations: ANR anthocyanidin reductase, ANS anthocyanidin synthase, C4H cinnamate 4-hydroxylase, CHI chalcone isomerase, CHS chalcone synthase, 4CL 4-coumarate : CoA ligase, CoumCoA3H p-coumaroyl-CoA 3-hydroxylase, DFR dihydroflavonol 4-reductase, F3GT UDP-glycosyl : flavonoid 3-O-glycosyltransferase, F7GT UDP-glucose : flavonoid 7-O-glucosyltransferase, F3′H flavonoid 3′-hydroxylase, FHT flavanone 3-hydroxylase (=F3H), FLS flavonol synthase, FNR flavanone 4-reductase, LAR leucoanthocyanidin reductase, OMT O-methyltransferase, PAL phenylalanine ammonia lyase

Full size image

Cloning of genes via sequence homology is a widely applied approach. Selected pear flavonoid genes [flavanone 3β-hydroxylase (FHT), dihydroflavonol 4-reductase/flavanone 4-reductase (DFR/FNR)] have previously been cloned for their involvement in inducible fire blight and scab resistance (Fischer et al. 2003; Halbwirth et al. 2006). This work now describes the cloning of pear main flavonoid cDNAs in general [phenylalanine ammonia lyase (PAL), chalcone synthase (CHS), chalcone isomerase (CHI), flavonol synthase (FLS), leucoanthocyanidin reductase (LAR1, LAR2), anthocyanidin synthase (ANS), anthocyanidin reductase (ANR), and UDP-glucose : flavonoid 7-O-glucosyltransferase (F7GT)], which have been isolated via homology with apple sequences to elucidate gene functions, gene copy numbers, and gene relationships within the Maloideae. With the exception of UDP-glycosyl : flavonoid 3-O-glycosyltransferase (F3GT), which could not be obtained, the sequences of the most important pear flavonoid genes are now available. In future, these sequence data can serve as a base for the development of molecular markers for marker assisted selection or for metabolic engineering of pear. It is also demonstrated here that homologous pear genes generally can be obtained straightforward from the extensive sequence information available from apple.

Materials and methods

Comparative Southern blot analysis

The comparative Southern blot analysis for Pyrus and Malus flavonoid genes was performed with existent Malus cDNAs, which were derived from previous work (cited below). Genomic DNA was prepared from young leaves of P. communis cv. Pyrodwarf and of M. x domestica (syn. Malus pumila) cv. M9 using the DNeasy® Plant DNA Kit (Qiagen, Crawley, UK). In each case 5 μg DNA were digested with 50 U restriction enzyme (EcoRI, HindIII, SalI, SacI, and XbaI) in the respective buffer for 5 h at 37°C. The restricted DNA was ethanol precipitated, redissolved in 20 μl TE at 65°C and used for agarose gel electrophoresis (1% agarose, 1 mg ethidium bromide/l, TAE buffer, 30 V). The gel was soaked in 0.25 N HCl for 15 min afterwards, rinsed with water, washed twice in 0.5 N NaOH/1.5 M NaCl for 20 min, rinsed with water and washed two times in 1 M Tris pH 7.4/1.5 M NaCl for 20 min. Blotting transfer was done overnight with 0.4 N NaOH on Immobilon Nylon N+® membrane (Millipore, Bedford, MA, USA). After blotting, the membrane was briefly washed with 5 × SSPE, air-dried and baked for 30 min at 80°C. Hybridization probes were PCR amplified from plasmid inserts of full length Malus cDNAs [Table 1, Fischer et al. 2003 (DFR), Halbwirth et al. 2006 (FHT and FLS), Fischer et al. 2006 (ANS), Pfeiffer et al. 2006 (LAR1, LAR2, and ANR), this paper (CHS and F7GT)]. In case of F3GT a cDNA amplified from Malus cv. M9 relying on AF117267 (full size F3GT) was used. For CHS and PAL, fragments of cDNAs amplified from Malus cv. M9 were used relying on X68977 (CHS) and X68126 (PAL). Only for CHI the Pyrus cDNA described here was used as a probe.

Table 1 Flavonoid genes cloned and reported from Pyrus communis, their cDNA sequence accessions, and the sources used for the Malus/Pyrus cDNA comparisons
Full size table

Hybridizations were performed with 32P-labeled DNA obtained with the Rediprime™ II kit (RPN 1633, Amersham, Freiburg, Germany) in the hybridization buffer (3 × SSPE/0.02% Ficoll/0.02% Polyvinylpyrrolidone/0.1% SDS/50 mg/l preboiled calf thymus DNA) at 63°C for at least 15 h. The blot was washed twice with 2 × SSPE/1% SDS and once with 2 × SSPE/0.1% SDS at 63°C for 5 min each. Bands were revealed by exposition with a phosphor-imager (Fuji BAS 1000 Bio-Imaging Analyzer, screens: BAS-MS 2040, Fuji, Kanagawa, Japan).

Cloning of pear flavonoid cDNAs

For cloning of the pear flavonoid cDNAs (PAL, CHI, FHT, FLS, DFR, LAR1, LAR2, ANS, and ANR) total RNA was obtained from young leaves of the P. communis cv. ‘Conference’ and the cv. ‘Pyrodwarf’ using the RNeasy Plant Mini kit (Qiagen). For the cloning of CHS and F7GT, mRNA was isolated from young leaves of P. communis cv. ‘Abbe Fetel’ with the μMACS mRNA Isolation Kit (Miltenyi Biotec, Auburn, CA, USA). Reverse transcription was performed with the SuperScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA) and the oligo(-dT) anchor primer GACCACGCGTATCGATGTCGAC(T)16V. PCR conditions were 1.5 mM MgCl2, 200 μM dNTPs, 500 nM each primer, with 10 ng cDNA and 1 U Taq polymerase in 1 × buffer (MBI). Cycling conditions were: 94°C for 1.5 min, 30 × (94°C for 30 s, 45-64°C for 1 min, 72°C for 2 min), 72°C for 7 min. For each cloning approach, three 5′ primers and three 3′ primers were derived from the non-coding regions of the Malus sequences listed in Table 2, where the primer combinations and annealing temperatures which were finally used for the cloning process are presented.

Table 2 Apple (Malus) basic sequence information (Davies 1993; Honda et al. 2002; Kim et al. 2003; Takos et al. 2006) used for pear (Pyrus) flavonoid cDNA cloning, derived primer sequences and annealing temperatures (T ann) finally used for obtaining the Pyrus cDNAs
Full size table

All 5′ and 3′ primer combinations were tested for PCR amplification of expected DNA fragments at three to five different annealing temperatures in the range of 45-64°C. The optimal primers and conditions were chosen and the PCR was repeated with proof-reading polymerase (Taq/Pwo-polymerase Expand High fidelity PCR System®, Roche, Nutley, NJ, USA) with the optimal annealing temperature and other PCR conditions according to the manufacturer’s instructions. The proof-reading PCR products were directly cloned into the yeast expression vector pYES2.1 (Topo TA Cloning Kit, Invitrogen), selected for sense insert orientation by PCR and commercially sequenced using vector primers.

Pyrus/Malus sequence comparisons

The Pyrus sequences obtained were compared with orthologous Malus sequences (Table 1) using CLUSTALW Multiple Alignment algorithm. Sequences from the M. x domestica cultivars ‘Golden Delicious’, ‘Rewena’, and the cv. ‘M9’ had been obtained previously making use of the sequence accessions listed in Table 2. The Malus CHI sequence was assembled from ESTs. Generally, all cDNAs of Malus and Pyrus were derived from young leaves.

Recombinant enzymes

For heterologous expression, the yeast expression vectors pYES2.1 harboring the RT-PCR-cloned flavonoid cDNAs in sense orientation were transformed into the yeast strain InvSc1 (Invitrogen) using the S.c. Easy Comp™ Transformation Kit (Invitrogen). The enzymes were prepared from galactose-induced yeast cultures as described by Urban et al. (1997). Transformation of yeast and preparation of yeast-derived enzymes was also done with empty pYES2-vector (pYES2 vector kit, Invitrogen) to provide a negative control for enzyme activities. Protein content was determined by a modified Lowry procedure (Sandermann and Strominger 1972) using BSA as a standard.

Enzyme assays

Enzyme assays were carried out according to Halbwirth et al. (2002) (CHS, FHT, FLS, and DFR/FNR), Pfeiffer et al. (2006) (ANR and LAR), S. Martens (unpublished, ANS), and Stich et al. (1997) (F7GT). In brief, [14C]-labeled substrates were incubated under conditions described with the respective recombinant enzyme. Products were extracted with ethylacetate and subjected to thin layer chromatography (TLC). Scanning radiography of TLC plates was used for qualitative and quantitative analysis of enzyme reaction products. Only anthocyanidins for ANR reaction were used in unlabeled form, epicatechin products being detected with dimethylaminocinnamicaldehyde (DMACA). [14C]-labeled flavonoids were synthesized via [2-14C]-malonyl-coenzyme A (2 GBq/mmol), which was purchased from Amersham International. Phenylalanine for the PAL assays was also [14C]-labeled (127 pmol/μl, 5.25 kBq/pmol; Sigma, St. Louis, MO, USA). p-Coumaroyl-CoA (Coum-CoA) and caffeoyl-CoA (Caff-CoA) were obtained according to Stockigt and Zenk (1976). [14C]-Naringenin (NAR), [14C]-eriodictyol (ERI), [14C]-dihydrokaempferol (DHK), and [14C]-dihydroquercetin (DHQ) were synthesized as described previously (Halbwirth et al. 2006). Unlabeled naringenin, eriodictyol, apigenin, luteolin, kaempferol (Km), quercetin (Qu), pelargonidin, cyanidin, and delphinidin for use as reference substances were purchased from Extrasynthesis (Genay, Lyon-Massieux, France). Enzymatic formation of the flavonoid 7-O-glucosides by F7GT was confirmed with HPLC and TLC with authentic standards according to Stich et al. (1997). CHI activity was assayed (pH 8) photometrically by observing the isomerization of naringenin chalcone at 385 nm with controls for yeast enzymes (empty expression vector) and the slower non-enzymatic isomerization to racemic naringenin.

Results

Comparative Southern blot analysis of Malus and Pyrus genomic DNA hybridized with cDNAs of Malus flavonoid genes revealed related band patterns (Fig. 2), and, hence, correspondingly similar gene copy numbers for some, but not all flavonoid gene classes (Table 3, columns 1–3). Especially for UDP-glycosyl : flavonoid- glycosyltransferases (FGTs) significant discrepancies were observed. Even so, this result indicated suitability of a homology-based cloning approach for most Pyrus flavonoid genes.

Fig. 2
figure 2

Comparative Southern blot analysis of Malus and Pyrus genomic DNA with Malus probes: examples shown for PAL, FHT, and F3GT cDNA hybridizations. Lanes for each blot from left to right: restriction enzymes EcoRI, HindIII, SalI, SacI, and XbaI

Full size image
Table 3 Comparison of Pyrus and Malus flavonoid cDNAs/genes with respect to gene copy numbers by Southern analysis, number of variable positions of the coding sequences (S substitutions, D deletions, I insertions, Pyrus, and Malus sequences given in Table 1 were compared), and enzyme substrate acceptance
Full size table

Cloning of the Pyrus flavonoid cDNAs was performed by deriving for each target sequence three 5′ primers and three 3′ primers from the non-coding 5′ and 3′ regions of available corresponding Malus flavonoid cDNA sequences (Table 2). Full size coding sequences of PAL, CHS, CHI, FLS, LAR1, LAR2, ANS, ANR, and F7GT (Table 3) were successfully amplified with at least one primer combination in each case, DFR/FNR (Fischer et al. 2003) and FHT (Halbwirth et al. 2006) had been obtained previously. All sequences have been submitted to GenBank (Table 1). The cDNA of LAR2 could not be obtained from the cv. ‘Conference’ by this approach, but only from the cv. ‘Pyrodwarf’.

The sequences of the Pyrus cDNAs obtained were different from those of Malus, both on nucleotide and amino acid level (Table 3, columns 4–6), being represented on the amino acid level by substitutions and also by insertions/deletions in many cases. The amplified 11 cDNAs, each representing the complete open reading frame, were directly cloned into a yeast expression vector. The recombinant enzymes were used to determine the function of the respective sequences (examples CHS, FHT, DFR, and FLS are shown in Fig. 3) and to analyze the substrate specificities of the Pyrus flavonoid enzymes (Table 3, column 7). For all assays, yeast enzyme preparations from an empty pYES2 line were used as a negative control. Product identification was done by co-chromatography with authentic standards.

Fig. 3
figure 3

Examples of gene-function verifications with selected recombinant Pyrus enzymes: Radioscans of TLCs on cellulose with solvent system CAW (chloroform/glacial acetic acid/water 10/9/1) from incubation of: a [14C]-malonyl CoA and p-coumaroyl-CoA with enzyme preparations from yeast expressing the CHS cDNA, b [14C]-NAR in the presence of 2-oxoglutarate, ascorbate and Fe2+ with enzyme preparations from yeast expressing the FHT cDNA, c [14C]-DHQ in the presence of NADPH with enzyme preparations from yeast expressing the DFR cDNA, d [14C]-DHQ in the presence of 2-oxoglutarate, ascorbate, and Fe2+ with enzyme preparations from yeast expressing the FLS cDNA. Abbreviations: NAR naringenin, DHK dihydrokaempferol, DHQ dihydroquercetin, LCy leucocyanidin, Qu quercetin

Full size image

Comparison of substrate specificities of recombinant Pyrus enzymes with known metabolic end product analyses of Pyrus fruits and leaves (cited above) allows to deduce the metabolic route realized in Pyrus tissues (Fig. 1). PAL converts phenylalanine to cinnamic acid. After 4-hydroxylation to p-coumaric acid and CoA conjugation, the product p-coumaroyl-CoA together with 3 malonyl-CoA is converted to the flavanone naringenin by the enzyme pair CHS and CHI. Alternatively, in some plants the dihydroxylated caffeoyl-CoA is produced from p-coumaric acid/p-coumaroyl-CoA. Also caffeoyl-CoA may thus be used to form a flavanone by CHS/CHI, resulting in the B-ring dihydroxylated eriodictyol, as Pyrus CHS possesses both substrate specificities. However, in most plant species with dihydroxylated flavonoids, like Malus, second B-ring hydroxylation is performed by a flavonoid 3′hydroxylase acting on flavanones (or dihydroflavonols), but this enzyme is not demonstrated for Pyrus yet. The major physiological substrate of FHT in Pyrus is eriodictyol which is 3-hydroxylated to dihydroquercetin, the last common precursor of flavan 3-ols (catechins), flavonols and anthocyanins. Eriodictyol is also the natural substrate of the F7GT, which catalyzes the formation of eriodictyol 7-O-glucoside. Despite possessing F7GT activity, its sequence identity to F5GTs is also high (data not shown), but it does not possess a F5GT activity. Dihydroquercetin is oxidized by FLS to the flavonol quercetin. Only minor amounts of the B-ring monohydroxylated flavonol kaempferol are formed from dihydrokaempferol as another substrate for FLS. Kaempferol is the only B-ring monohydroxylated flavonoid end product in Pyrus. Dihydroquercetin is also substrate for DFR producing leucocyanidin. (The side activity of DFR as FNR converting eriodictyol to luteoforol is only observed in Pyrus tissue with artificially blocked FHT activity (Fischer et al. 2003; Halbwirth et al. 2006)). Leucocyanidin is either converted by LAR 1 or LAR 2 to catechin, or by ANS to form cyanidin. The latter is reduced by ANR to epicatechin, or glycosylated to build various cyanidin 3-O-glycosides.

Discussion

Flavonoid enzymes

Differences in the substrate acceptance of recombinant DFR enzymes from Pyrus and Malus were observed although only a few exchanges at the amino acid sequence existed (Fischer et al. 2003). When more than one cultivar from each genus was considered, as for DFR (Fischer et al. 2003) and FHT (Halbwirth et al. 2006), some seemingly genus specific amino acid substitutions could be identified in addition to the sequence variation observed between cultivars of either genus. Also for gene copy numbers variation was found between Pyrus and Malus, most obvious for either F3GT or F7GT, which each show quite different gene copy numbers and band patterns in Southern analyses probed with Malus cDNAs. Furthermore, only F3GT cDNA could not be cloned by the approach described, even if no alternative approaches were undertaken in comparison to the other cloning tasks. In contrast, the distantly related F7GT could be cloned, relying on a Malus F7GT sequence. Correspondingly to the differences at the gene level, glycosylation patterns of flavonoids of Pyrus and Malus are divergent.

Pyrus polyphenols

Beyond flavonoids, there are considerable differences in the polyphenol composition of Pyrus compared to Malus. Most important, pears contain the p-hydrochinone arbutin as a glucoside, but lack the dihydrochalcone phloretin of Malus; many anthocyanins and flavonol glycosides differ, too. These differences were also exploited for chemotaxonomic differentiation (Challice 1981) and are practically applied for juice authenticity control (Schieber et al. 2001). The presence of different polyphenols indicates that, in addition, several Pyrus-specific polyphenol enzymes or enzyme functions have to exist.

Pyrus gene cloning

The applicability of a homology cloning approach for Pyrus was indicated by the positive result of the Southern analyses with existent Malus cDNAs. The cloning approach solely relied on sequence similarities in transcribed, but non-coding parts (5′ leader and 3′ trailer) of known Malus sequences. When the various 5′ and 3′ primer combinations were tested for each gene, mostly 50% of the RT-PCR reactions resulted at least in low amounts of RT-PCR products of the expected size. Homology approaches relying on Malus sequences will thus be valuable for cloning other Pyrus genes. Considering the current number of accessions of the genera Pyrus, Malus, Prunus, Fragaria, and Rosa in GenBank (Table 4), it is obvious that Malus sequences are the best source of sequence information for Pyrus gene cloning, especially by assembling Malus expressed sequence tags (ESTs) (Newcomb et al. 2006) as demonstrated in this work for PAL and CHI. Even if less related, Prunus sequences might become a valuable source of information as well.

Table 4 Current numbers of sequence accessions for the genera Pyrus, Malus, Prunus, Fragaria, and Rosa in GenBank (27/11/2006)
Full size table

Marker development/transgenic plants

Genetic markers for marker-assisted selection are especially valuable with respect to the long generation time of Pyrus, demanding highly efficient selection measures. Recently, synteny between Malus and Pyrus chromosomes was reviewed (Arus et al. 2006), and simple sequence repetition (SSR) markers have been reported to be transferable from Malus to Pyrus (Yamamoto et al. 2001; Pierantoni et al. 2004). The Pyrus cDNA sequences of flavonoid genes obtained can serve as a base to develop non-anonymous markers for relevant alleles of the structural flavonoid genes in breeding for color, for improved flavonoid content with respect to pathogen resistance or for health-promoting fruit components.

In addition, the cloned cDNAs of Pyrus ANR, LAR1, and LAR2 allow metabolic engineering aimed at plants with higher contents of catechins and proanthocyanidins (Xie et al. 2006), making use of Pyrus cDNAs rather than foreign sequences for overexpression or silencing strategies. It was reported that public acceptance of transgenic crops is better when genes of the same species are used (Fischer and Weber 2005).