2018 年 87 巻 2 号 p. 288-296
Actinidin is a major protein contained in kiwifruit (Actinidia spp.). While uptake of actinidin is beneficial to help gastric protein digestion with cysteine protease activity, the protein is also recognized as a major elicitor of allergy which can induce tingling in the oral cavity and occasionally severe anaphylactic reactions. Given that consumption of fresh kiwifruit has increased globally, development of Actinidia cultivars with lower level of actinidin is required to reduce the risk of allergenicity. In the present study, we examined variations in the actinidin level in Actinidia varieties. Among several varieties having trace amounts of actinidin, A. chinensis ‘Kohi’ was targeted to be analyzed for the molecular basis for the phenotype. ‘Kohi’ had below the detectable transcript level of Act1a, a critical gene for actinidin level. The upstream region of Act1a in ‘Kohi’ constituted different sequences from that of A. deliciosa ‘Hayward’, which has an active promoter for high expression of Act1a. The ‘Kohi’ sequence in the diverged region (upstream from −873 b) was rich in cytosine residues methylated at a higher level than in ‘Hayward’. Our data suggest the possibility of novel epigenetic regulation to reduce the actinidin level. The molecular mechanism for the phenotype in ‘Kohi’ was differentiated from ‘Hort16A’, a globally popular cultivar with a low level of actinidin. This cultivar could be a choice as a genetic resource in breeding to develop cultivars with controlled actinidin levels.
Actinidin is a predominant protein found in kiwifruit (Actinidia spp.). The content in green-fleshed kiwifruit (A. deliciosa) comprises 40 to 50% of soluble proteins (Lewis and Luh, 1988; Praekelt et al., 1988). Because of its significant content, the protein has been extensively studied from biochemical and application aspects (Boland, 2013; Nishiyama, 2007). Actinidin possesses cysteine (Cys) protease activity and belongs to a papain-like subfamily that includes papain from papaya (Carica papaya), bromelain from pineapple (Ananas comosus), and ficin from fig (Ficus carica) (Beers et al., 2004; Nieuwenhuizen et al., 2007). The actinidin level is increased as fruit mature (Lewis and Luh, 1988; Lin et al., 1993), whereas Cys proteases in the fruits of other species decrease during the ripening progress (Dunne and Horgan, 1992; Moyle et al., 2005; Raskovic et al., 2016), resulting in kiwifruit having the highest protease activity among a wide range of fruits and vegetables in the edible phase (Sun et al., 2016).
Despite such characterizations, the physiological function of actinidin has not been fully elucidated. Plant Cys proteases play roles not only in programmed cell death in response to developmental cues and pathogen infection but also in a range of developmental processes (van der Hoorn, 2008). However, actinidin isoforms form a distinct kiwifruit-specific clade among Cys protease family members based on amino acid sequences (Nieuwenhuizen et al., 2007), suggesting its unique role in kiwifruit. While actinidin is not thought to be an essential protein for fruit development because no phenotypic defect has been found in actinidin-deficient varieties (Boland, 2013), some studies suggested its potential roles in pathogen defense through processing of a class IV chitinase (Nieuwenhuizen et al., 2012) and insect resistance (Malone et al., 2005).
Actinidin-derived proteolytic activity in kiwifruit could give some benefits. It is reported that dietary actinidin accelerates gastric protein digestion (Montoya et al., 2014). The efficient hydrolyzing activity against myofibrillar proteins could be potentially useful as a meat tenderizer (Bekhit et al., 2014). In New Zealand, for instance, actinidin-containing supplements have actually been marketed as digestion enhancers. However, a serious issue remains in actinidin uptake from fresh kiwifruit. It is well known that actinidin is a major elicitor of allergy (Pastorello et al., 1998). Of 13 kiwifruit allergens (Act d 1–13) defined by the Allergen Nomenclature Subcommitte of the International Union of Immunological Societies (www.allergen.org), actinidin (Act d 1) is established as a major allergen and is often recognized in kiwifruit-monosensitized patients (Bublin et al., 2010; Palacin et al., 2008). The symptoms range from mild ones in the oral cavity to severe anaphylactic reactions (Le and Knulst, 2015). Since kiwifruit is an excellent source of nutrients including vitamins (especially C and E), minerals, and fibers, it is important to develop cultivars containing low amounts of actinidin to address the allergic issue. A. chinensis ‘Hort16A’, commercially known as “Zespri Gold”, has become globally popular because of its yellow-fleshed sweet and tropical flavored fruit. Another big advantage of this cultivar is its low actinidin level (Nieuwenhuizen et al., 2007). Given that global production of kiwifruit has increased over the last few decades (Ward and Courtney, 2013), introduction of low actinidin cultivars will help to grow the market share of kiwifruit.
Multiple studies have reported considerable variations in the actinidin level among Actinidia fruits. For instance, A. deliciosa cultivars in general contain actinidin as a major protein, but many of the A. eriantha and A. rufa genotypes do not accumulate a lot of the protein (Maddumage et al., 2013; Nishiyama, 2007). A great diversity in one species has been found in yellow-fleshed A. chinensis (Maddumage et al., 2013; Nieuwenhuizen et al., 2007; Zhang et al., 2017). For instance, while the A. chinensis advanced selection EM4 (parentage [CK-01×CK-02]×[CK-02×CK-03]) contains an equivalent level of actinidin to A. deliciosa ‘Hayward’, ‘Hongyang’, and ‘Hort16A’ have only trace amounts of actinidin (Nieuwenhuizen et al., 2007). A recent genetic study using a segregating population of A. chinensis determined a region containing Act1a, an actinidin-coding gene, as a major quantitative trait locus for Cys protease activity (Nieuwenhuizen et al., 2012). They also revealed that ‘Hort16A’ possesses two non-functional Act1a alleles; one is non-functional because of a large insertion and the other encodes an inactive unstable actinidin, resulting in a lower actinidin level in the cultivar. Such a molecular basis for actinidin accumulation will be an important issue in future kiwifruit breeding to control the actinidin level. In the present study, we examined actinidin levels in a range of Actinidia varieties and found that A. chinensis ‘Kohi’ accumulates little actinidin like ‘Hort16A’. Here, we found that transcriptional suppression of Act1a results in the phenotype in ‘Kohi’, which is presumably caused by a disabled Act1a promoter because of sequence variations. We also demonstrate the possibility of epigenetic regulation for Act1a transcription in ‘Kohi’. The mechanism for the lower actinidin accumulation in ‘Kohi’ was clearly distinguished from ‘Hort16A’.
Fruits of Actinidia chinensis (Planch.) ‘Kohi’ were obtained from Kuyuna farm in Tokushima prefecture in Japan. A. chinensis ‘Hort16A’ was purchased at local markets. Actinidia deliciosa (A. Chev.) C.F. Liang & A.R. Ferguson cultivars (‘Hayward’, ‘Abbott’, ‘Monty’, and ‘Bruno’) and the other A. chinensis cultivars (‘Sensational apple’, ‘Golden king’, ‘Yellow queen’, ‘Shonan yellow’, ‘Kataura yellow’, and ‘Koshin’) were grown in a field in Nebukawa in Kanagawa prefecture, Japan. A. arguta and A. eriantha were grown in a field at Nihon University (Fujisawa, Kanagawa, Japan). When kiwifruits were not completely ripe, they were stored at room temperature until reaching eating ripeness. Ripe kiwifruits were peeled and excised into pieces, followed by freezing in liquid nitrogen and storage at −80°C until use. At least three fruit were used in each experimental unit.
Antibody preparationTo raise actinidin antibodies, actinidin was purified from the fruits of A. deliciosa ‘Hayward’ as previously described (Pastorello et al., 1998) with some modifications. Actinidin was purified by anion exchange chromatography with AKTA10S (GE healthcare life sciences, UK). Fruit crude extract was loaded onto a Hi-Trap Q HP (16 mm × 26 mm; GE healthcare life sciences) previously equilibrated with 20 mM Tris-HCl (pH 7.5). Proteins were eluted with a linear gradient of NaCl concentration (0–0.5 M) and fractionated. Fractions having high protease activity were pooled and then subjected to SOURCE 15Q (4.6 mm × 100 mm; GE healthcare life sciences) previously equilibrated with 20 mM sodium citrate buffer (pH 4.5), followed by elution with a linear gradient of NaCl concentration (0–0.5 M). Following protease assay, fractions with high activity were analyzed by sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli, 1970) to confirm a single band of actinidin. Lyophilized actinidin (5.6 mg) was submitted to Sigma-Aldrich for antibody production in rabbits.
Immunoblot analysisFruit tissues except for peels were homogenized at 4°C using a polytron PT10-35GT (Central Scientific Commerce, Japan) in two volumes of homogenization buffer (50 mM Sodium phosphate [pH 7.0], 2% [w/v] polyvinylpyrrolidone (PVP), and 5 mM ethylenediamine tetraacetic acid (EDTA) with Complete Protease Inhibitor Cocktail [Roche Diagnostics, Switzerland]) and centrifuged at ×8500 g for 15 min at 4°C. The supernatant was passed through Miracloth (Merck Millipore, USA) and subjected to PD10 column (GE healthcare life sciences) previously equilibrated of 20 mM Tris-HCl (pH 7.0). The sample was mixed with an equal volume with SDS sample buffer (125 mM Tris-HCl [pH 6.8], 2.5% [w/v] SDS, 25% [v/v] Glycerol, 2.5% [v/v] β-mercaptoethanol, and 0.002% [w/v] bromophenol blue) and heated at 95°C for 1 min. An aliquot (20 μL) was subjected to SDS-PAGE (12.5% acrylamide gel). After electrophoresis, proteins were transferred to ClearTrans Nitrocellulose Membranes (Wako chemicals, Japan). Immunoblot analysis was performed as previously described (Kamiyoshihara et al., 2012). Antiserum was used at 1/2000 dilution. To visualize all the proteins, replicated gels were stained with SimplyBlue SafeStain (Thermo Fisher Scientific, USA).
Enzyme-linked immunosorbent assay (ELISA)The samples from the PD10 column were diluted 25-fold with coating buffer (5 mM Na2CO3, 10 mM NaHCO3 [pH 9.5]). An aliquot (50 μL) was applied to each well of a 96F MAXISORP BlackMicrowell SH (Thermo Fisher Scientific) plate and incubated at 37°C for 1 h. After washing with Dulbecco’s PBS (Thermo Fisher Scientific) containing 0.05% Tween 20 (PBST), wells were blocked with PBS containing 1% bovine serum albumin. An anti-actinidin antibody diluted (1/4000) in 1% bovine serum albumin (BSA)/PBST was applied, followed by incubation at 37°C for 1 h. After a wash with PBST, a peroxidase-conjugated goat anti-rabbit secondary antibody diluted (1/4000) in 1% BSA/PBST was applied and incubated as above. Following a wash with PBST, antigen-antibody binding was detected with SureBlue (Funakoshi, Japan) according to the manufacturer’s instruction. Reactions were stopped by adding 1N HCl and 450 nm absorbance was monitored using Synergy 2 (Biotek, USA). The experiments were done in triplicate. Purified actinidin from ‘Hayward’ was used to make a standard curve.
Genomic DNA extractionDNA extraction was performed as previously described (Xu et al., 2004) with some modifications. 1.0 g fruit tissue ground in liquid nitrogen was briefly mixed with 5 mL of wash buffer (100 mM Tris-HCl [pH 8.0], 5 mM EDTA, 0.35 M glucose, 2% (w/v) PVP, and 4% (v/v) 2-mercaptoethanol), followed by incubation on ice for 30 min. After centrifuging at 6000 × g for 10 min at 4°C, the sediment was resuspended in 5 mL of extraction buffer (100 mM Tris-HCl [pH 8.0], 5 mM EDTA, 1.5 M NaCl, 3% (w/v) cetyltrimethylammonium bromide, and 4% (v/v) 2-mercaptoethanol) and incubated at 65°C for 30 min with vortexing several times. After centrifuging at 6000 × g for 10 min at 4°C, the supernatant was mixed with 0.1 volume of 5 M potassium acetate (pH 4.8) and an equal volume of chloroform/isoamylalcohol (24:1). The supernatant after brief centrifuge was mixed with 0.1 volume of 3 M sodium acetate and an equal volume of isopropanol to allow DNA precipitation. DNA was washed with 70% ethanol three times and reconstituted in 200 μL sterile water. The DNA samples were incubated at 37°C for 3 h in the presence of 1 μg RNase to degrade contaminated RNA.
Total RNA extractionRNA extraction was performed as previously described (Wan and Wilkins, 1994) with some modifications. Fruit tissue (1 g) ground in liquid nitrogen was incubated at 42°C for 90 min in 5 mL of preheated extraction buffer (0.2 M sodium borate [pH 9.0], 30 mM EDTA, 1% sodium deoxycholate, 1% SDS, 2% PVP, 1% 4-nonylphenyl polyethylene glycol, 10 mM dithiothreitol) containing 10 mg⋅mL−1 Protease K. After adding 3 M KCl (0.16 M final concentration) and incubation on ice for 1 h, the sample was centrifuged at 7500 rpm for 10 min. The supernatant was recovered and mixed with one third volume of 10 M LiCl, followed by incubation at −20°C overnight. The pellet resulting from centrifugation at 13000 rpm for 30 min was resuspended in 500 μL of 2 M LiCl. After repeated washing with 2 M LiCl three times, the pellet was resuspended in 400 μL of 10 mM Tris-HCl (pH 7.5). DNA was digested with DNase I Amplification Grade (Thermo Fisher Scientific).
Reverse transcript-PCRcDNA synthesis was performed using PrimeScript Reverse Transcriptase (Takara, Japan) with Not I-Oligo d(T)18 primer (5'-AACTGGAAGAATTCGCGGCCGCAGGAATTTTTTTTTTTTTTTTTT-3') according to the manufacturer’s instruction. PCR was done in a 20 μL mixture consisting of 2 x GoTaq Green Master Mix (Promega, USA), 0.5 μM primers, and cDNA. PCR conditions were as follows: step 1, 95°C for 2 min; step 2, 95°C for 1 min, 55°C for 1 min, 72°C for 1 min per 1 kb length; step 3, 72°C for 5 min. Step 2 was repeated 30 times. The following sets of gene-specific primers were used to amplify DNA fragments: Act1a (sense, 5'-GAAGTCAAGGCCATGTACGAG-3', antisense, 5'-AACCGCTTGTAAATCCCAAGT-3'), Act2a (sense, 5'-GTCAGCCATCGACATCGAAAACTCAGT-3', antisense, 5'-GAAACGGTTCAAGCCCAAAGA-3'). PCR products were subjected to agarose gel electrophoresis, followed by detection with ethidium bromide under UV illumination.
Real-time PCRReal-time PCR was performed using SYBR Premix Ex Taq (Takara) with cDNAs as templates using 0.4 μM primers. The reaction and detection were performed using Mx3000P (Agilent, USA). Reaction conditions were as follows: step 1, 95°C for 10 s; step 2, 95°C for 5 s, 60°C for 20 s. Step 2 was repeated 40 times. The expression levels of Act1a were normalized with the constantly expressed UbiC9 level. The following sets of gene-specific primers were used to amplify DNA fragments: Act1a (sense, 5'-GGCACAGAGGGAGGTATCGA-3', antisense, 5'-CAATTCCACATGTTCCAGCA-3'), UbiC9 (sense, 5'-CCATTTCCAAGGTGTTGCTT-3', antisense, 5'-TACTTGTTCCGGTCCGTCTT-3').
Thermal asymmetric interlaced (TAIL)-PCRTAIL PCR was performed to obtain the 5'-flanking region of ‘Kohi’ Act1a according to the procedure shown by Liu et al. (1995). Three specific (SP) primers to the promoter ‘Hayward’ sequence and five arbitrary degenerate (AD) primers were designed: SP primers (SP1, 5'-GTTGGAACTGCAGATTCAGCAG-3', SP2, 5'-CTTTAAATCCTTATCCAACCACAAC-3', SP3, 5'-GAATGCAAGTAGCTAACCCTGTGG-3'), AD primers (5'-NGTCGASWGANAWGAA-3', 5'-GTNCGASWCANAWGTT-3', 5'-WGTGNAGWANCANAGA-3', 5'-AGWGNAGWANCAWAGG-3', 5'-NTCGASTWTSGWGTT-3'). Following three rounds of PCRs using EmeraldAmp MAX PCR Master Mix (Takara) coupled with a SP primer (0.5 μM) and mixed AD primers (0.8 μM each), an amplified product was cloned into T-Vector pMD20 (Takara). PCR conditions followed instructions by Liu et al. (1995).
Sequence analysisSequences of the Act1a gene and the promoter region were determined by Sanger sequencing using a BigDye Terminator v3.1 Cycle Sequencing Kit (Thermo Fisher Scientific) according to the manufacturer’s instruction. Multiple sequence alignment was performed with ClustalW.
Bisulfite sequencingExtracted genomic DNA was treated with a MethylEasy DNA Bisulphite Modification Kit (Takara) to convert unmethylated cytosine to uracil while methylated residues remained as cytosine. KOD-Multi & Epi- (Toyobo, Japan) was used to amplify DNA segments of the promoter region. Reaction conditions were as follows: step 1, 94°C for 2 min; step 2, 98°C for 10 s, 52°C for 30 s, 68°C for 30 s. Step 2 was repeated 40 times. For the conserved sequence regions of ‘Kohi’ and ‘Hayward’, the same primer sets were used; the region of −822 b to −520 b (sense, 5'-ATTTGTTTTAYGAAGAAGATAGTTTTGTGATTTG-3', antisense, CCTAATAAATCTAAACAAAACAACTAAAAAATCC-3'), the region of −603 b to +67 b (sense, 5'-YAAAATTGYAYTTGTATTAAGGTTG-3', antisense, 5'-CAAACCCATTTTTRTTCTCTCTCRTTTTTTRAAT-3'). For the diverged region in ‘Hayward’, the region of −1188 b to −807 b was amplified with a sense primer (5'-TTAAATTTTATGGGGTTATTGAGAAAT-3') and an antisense primer (5'-AAAAACAACAAATCACAAAACTATCTTC-3'). For the diverged region in ‘Kohi’, two primer sets were used: the region of −1449 b to −1084 b (sense, 5'-GTTGAATATAAYGAGGTGGTGGCAT-3', antisense, ACCATACTAACCCATTTACTTAATC-3'), the region of −1108 b to −853 b (sense, 5'-GATTAAGTAAATGGGTTAGTATGGTA-3', antisense, 5'-AAATCTAACACACTCCACTATTAAAAATAT-3'). E. coli TOP10 (Thermo Fisher Scientific) was transformed with plasmid pZero2.1 (Thermo Fisher Scientific) ligated with the amplified product. At least 13 independent clones were subjected to sequence analyses for each section to determine the methylation percentage.
To determine the actinidin levels in the Actinidia varieties, immunoblot analysis was performed using polyclonal antibodies raised to native actinidin purified from the fruit of A. deliciosa ‘Hayward’. The total soluble proteins extracted from whole ripe fruit tissue were analyzed. When all the proteins were subjected to SDS-PAGE followed by coomassie brilliant blue (CBB) staining, a major protein was detected at the 26 kDa-position in all four cultivars in A. deliciosa and six out of eight cultivars in A. chinensis, and A. arguta (Fig. 1). Immunoblot analysis with an anti-actinidin antibody confirmed that this abundant protein was actinidin (Fig. 1). On the other hand, actinidin was almost absent from two A. chinensis cultivars, ‘Kohi’ and ‘Hort16A’, and A. eriantha. Although the lower actinidin level in ‘Hort16A’ has already been reported (Maddumage et al., 2013; Nieuwenhuizen et al., 2007), ‘Kohi’ was newly found to contain only a little actinidin in this study. In the CBB-stained gel, a protein with slightly lower molecular mass than actinidin was detected in ‘Koshin’, ‘Kohi’, and ‘Hort16A’ (Fig. 1). This protein did not react with the anti-actinidin antibody, indicating the presence of another major protein in these cultivars. We then performed ELISA for quantification of actinidin in the fruits. The actinidin content in most cultivars of A. deliciosa and A. chinensis ranged from about 1–2 mg per gram fresh weight with the maximum in ‘Sensational Apple’ (1.95 ± 0.14 mg⋅g−1 FW) (Fig. 2). Consistent with the immunoblot analysis, ELISA confirmed much lower levels of actinidin in ‘Kohi’ (0.06 ± 0.01 mg⋅g−1 FW) and ‘Hort16A’ (0.09 ± 0.01 mg⋅g−1 FW).
Analysis of soluble proteins and immunoblotting of actinidin in Actinidia varieties. Soluble proteins extracted from kiwifruits were subjected to polyacrylamide gel electrophoresis, followed by detection with coomassie brilliant blue staining (CBB) and immunoblot analysis with an anti-actinidin antibody (Actinidin Ab.). Molecular masses of marker proteins are indicated at the sides. An arrowhead indicates actinidin at 26 kDa. Several cultivars were shown in abbreviated names: ‘S. Apple’, ‘Sensational apple’; ‘Y. queen’, ‘Yellow queen’; ‘Shonan Y’, ‘Shonan yellow’; ‘Kataura Y.’, ‘Kataura yellow’.
Actinidin levels in Actinidia varieties. ELISA with an anti-actinidin antibody was conducted using soluble proteins to quantify actinidin levels in kiwifruits. Purified actinidin from ‘Hayward’ was used to make a standard curve. Assays were performed in triplicate, and values shown are means ± SE (n = 3).
Since ‘Kohi’ bears a unique yellow-fleshed fruit with an anthocyanin-derived red color around the core (Fig. S1), this cultivar could be a valuable source for breeding. To address the molecular basis for the lower actinidin level in ‘Kohi’, we first analyzed the gene expression of the actinidin-coding genes, Act1a and Act2a, because they have been shown to be detected in the ripe fruits of A. deliciosa ‘Hayward’ and A. chinensis ‘Hort16A’ (Nieuwenhuizen et al., 2007). Genomic PCR first confirmed the presence of Act1a and Act2a in ‘Kohi’ as well as in ‘Hayward’ and ‘Hort16A’ (Fig. 3A). However, reverse transcription PCR with cDNAs from each ripe fruit amplified the Act1a fragment in ‘Hayward’ and ‘Hort16A’ but not in ‘Kohi’, while the Act2a fragment was detected in all three cultivars (Fig. 3B). Thus, we confirmed the Act1a expression levels with quantitative PCR. Although the expression level in ‘Hort16A’ was comparable to ‘Hayward’, almost no expression was detected in ‘Kohi’ (Fig. 3C). Therefore, it appeared that differential factors caused the lower actinidin phenotype in ‘Hort16A’ and ‘Kohi’.
Genomic and transcriptional analyses of Act1a and Act2a in ‘Hayward’, ‘Hort16A’, and ‘Kohi’. A, B: PCR for Act1a and Act2a was performed using Genomic DNA and cDNA as a template. Primers were designed based on the conserved regions in the three cultivars. Parts of exon 1 of Act1a (109 b to 326 b of open reading frame) and Act2a (66 b to 260 b of open reading frame) were amplified. C: Quantitative PCR analyses were performed using the same cDNA used for B. Expression values of Act1a were normalized by ubiquitously expressed UbiC9. The relative values are expressed compared with ‘Hayward’. The experiment was performed in triplicate, and values shown are means ± SE (n = 3).
Since genomic PCR suggested the presence of the Act1a gene in the ‘Kohi’ genome, we sequenced the Act1a genomic region (accession No; LC330913). Although multiple sequence variations were found in both the exons and introns in ‘Kohi’ compared with the ‘Hayward’ sequence, it appeared that a complete Act1a gene was present (Fig. S2). Therefore, it was more likely that the causal factor for Act1a suppression resides in the upstream promoter region. The Act1a promoter in ‘Hayward’ has been previously reported (Lin et al., 1993; Snowden and Gardner, 1990). Using primers based on the ‘Hayward’ sequence, the corresponding regions of ‘Kohi’ were tested by amplification (Fig. 4). Region (I) (+5 b to −867 b in ‘Hayward’ sequence) was amplified with both ‘Hayward’ and ‘Kohi’ DNAs. However, region (II) (−728 b to −1294 b) was not amplified with ‘Kohi’ DNA, meaning that the Act1a upstream region constitutes different sequences in ‘Kohi’ in region (II). We thus conducted TAIL PCR to obtain the upstream region of ‘Kohi’ Act1a. Three rounds of PCR with combinations of specific primers and arbitrary primers gave a 1.5 kb-fragment. Sequencing of this fragment revealed that the proximal region to the gene (0 b to −873 b) almost shared the sequence with ‘Hayward’ (Fig. 5). However, the further upstream region (−874 b to −1449 b) was barely similar to the ‘Hayward’ sequence. Although alignment searches with the sequence of the diverged region against various genome databases did not find any similar annotated sequence, it was possible that this sequence variation disrupted the original promoter function.
Compositional comparison of upstream regions of Act1a of ‘Hayward’ and ‘Kohi’. Two sections of the Act1a upstream region, designated (I) and (II), were amplified with genomic PCR using primers based on the sequence of ‘Hayward’ previously reported (L07552, Lin et al., 1993). As controls to ensure DNA quality, UbiC9 and EF1a were amplified.
Sequence comparison of Act1a upstream regions of ‘Hayward’ and ‘Kohi’. Sequence alignment was performed using ClustalW with Act1a promoter of ‘Hayward’ (L07552) and the upstream region of ‘Kohi’ obtained by TAIL PCR. Common sequences are enclosed by squares. +1 indicates a major transcription initiation site suggested by Lin et al. (1993). The asterisk indicates a translation initiation codon.
It was noted that the sequence of the diverged region of ‘Kohi’ was characteristic in terms of numbers of cytosine residues. Cytosine methylation in promoter regions is often found and the methylation levels influence downstream gene expressions (Vanyushin and Ashapkin, 2011). We thus suspected an epigenetic regulation for silencing Act1a. In plants, methylation occurs in the contexts of CG, CHG, and CHH (H indicates for A, T, or C) while in animals only CG sites are targeted. The number of possible methylation sites found in the upstream regions is summarized in Table 1. Notably, the diverged region of ‘Kohi’ was rich in CG sites compared with the ‘Hayward’ sequence (44 sites in ‘Kohi’ versus 4 sites in ‘Hayward’) while the other sites had comparable frequencies, resulting in a total of 1.6-fold potential methylation sites in ‘Kohi’. We performed bisulfite sequencing to analyze the methylation status in each region in both cultivars (Fig. 6). In the conserved regions, methylation was hardly found except for a few positions with low percentages (‘Hayward’, position −666 b, −664 b, −631 b, and −621 b with 25% methylation, respectively; ‘Kohi’, position −666 b and −664 b with 20% methylation, respectively. Fig. 6). On the other hand, the diverged region was significantly methylated both in ‘Hayward’ and ‘Kohi’. Interestingly, several features in the methylation pattern were found between the cultivars (Table 1; Fig. 6). First, the averaged methylation percentage in the diverged region was higher in ‘Kohi’ (80.5%) than in ‘Hayward’ (65.7%). Second, the diverged region was entirely methylated in ‘Kohi’ while higher percentages of methylation were restricted in the region of −1100 b to −1000 b in ‘Hayward’. Third, CG sites were more preferentially methylated than CHG and CHH sites in ‘Kohi’, but CHH sites were most highly methylated in ‘Hayward’. Such differential methylation patterns may affect the promoter functionality that controls the downstream Act1a transcription.
Number of possible methylation sites and percentage of methylated cytosines in the upstream regions of Act1a.
Distribution of cytosine methylation in Act1a promoters of ‘Hayward’ and ‘Kohi’. Following bisulfite sequencing, the percentage of cytosine methylation was calculated for each site. The regions from −1200 b to −600 b are indicated. More than 13 clones were sequenced for each position to determine the methylation rates. Possible methylation sites are distinguished by color: Black, CG; Gray, CHG; White, CHH.
Actinidin has received scientific attention because of its allergenicity, as well as its high content in kiwifruit with unknown biological functions. Although allergen issues are present in many fruits to some extent, kiwifruit is recognized as a major elicitor of plant food allergies (Lucas et al., 2003; Mills et al., 2004). Multiple allergic proteins are contained in kiwifruit and the fruit is principally consumed fresh (Nishiyama, 2007). Therefore, it would be ideal to reduce the levels of allergens, especially actinidin. Many of the commercialized cultivars tested here contained abundant actinidin, which was clearly observed as the major protein in soluble fractions (Fig. 1). While all the A. deliciosa cultivars were rich in actinidin, A. chinensis cultivars showed significant variations in the actinidin level; ‘Kohi’ contained actinidin as a minor protein as low as in ‘Hort16A’, but the other cultivars tended to have even more abundant actinidin than A. deliciosa cultivars (Fig. 2). ‘Kohi’ was first found to have the lower actinidin level in the present study. In addition to this phenotype, ‘Kohi’ fruit has characteristics including anthocyanin accumulation and early ripening. Thus, this cultivar has potential for the development of new cultivars.
We addressed the molecular basis for the low actinidin phenotype in ‘Kohi’. Although the two actinidin-coding genes, Act1a and Act2a, were present in the ‘Kohi’ genome, reverse transcription-PCR showed significant suppression of Act1a while Act2a was expressed. Since the copy numbers of Act1a and Act2a mRNAs were not quantified, the contribution of each gene to actinidin accumulation cannot be assessed. The low level of actinidin detected in ‘Kohi’ by ELISA could be Act2a-derived (Fig. 2). However, it was clear that transcriptional suppression of Act1a is most likely the factor causing the low actinidin phenotype in ‘Kohi’. This mechanism is different from the observations in ‘Hort16A’; the Act1a transcript level is comparable to ‘Hayward’, but only a trace amount of actinidin can be detected (Nieuwenhuizen et al., 2007; Figs. 1 and 3). Inactive Act1a protein derived from a mutated allele in ‘Hort16A’ is subject to rapid protein turnover (Nieuwenhuizen et al., 2012). Therefore, both ‘Kohi’ and ‘Hort16A’ commonly showed the lower actinidin phenotype caused by functional defects in the Act1a gene, but the underlying mechanisms appeared to be different. Since multiple varieties with lower actinidin have been found other than ‘Kohi’ and ‘Hort16A’ (Maddumage et al., 2013; Nishiyama, 2007), it is of interest to investigate the causal factors for this phenotype. For instance, A. chinensis ‘Hongyang’ has been found to have a trace amount of actinidin (Maddumage et al., 2013). Given that ‘Hongyang’ fruit has phenotypic similarities to ‘Kohi’ (i.e. anthocyanin accumulation, Wang et al., 2003), they could be relative cultivars with lower actinidin levels under similar regulation.
We found differential constitutions in the Act1a upstream region in ‘Kohi’ compared with ‘Hayward’ (Fig. 4). TAIL PCR gave a fragment of the upstream region of Act1a in ‘Kohi’ which partially shared the sequence with the ‘Hayward’ Act1a promoter at the proximal side of the gene (Fig. 5). However, ‘Kohi’ had totally different sequences in a further upstream region (upstream from −873 b). Although this region did not show similarity to any annotated sequence, it was predicted that DNA deletion or insertion occurred in this region during evolution or breeding processes. Analysis of the ‘Hayward’ Act1a promoter using transgenic petunia plants suggested that a 1.3 kb promoter is necessary for the high level expression of Act1a while a shorter 173 b region (proximal to Act1a gene) appears to be involved in the spatial and temporal regulation (Lin et al., 1993). Thus, the sequence variation found in ‘Kohi’ could disrupt the original Act1a promoter functionality for strong gene expression. As one possibility, a cis-element in the promoter to be recognized by a transcription factor (TF) could be simply eliminated through DNA deletion or insertion. Alternatively, epigenetic regulation was also suspected because more possible methylation sites were present in the diverged region in ‘Kohi’. Bisulfite sequencing revealed high level DNA methylation in the diverged regions both in ‘Hayward’ and ‘Kohi’, while the conserved regions were hardly methylated (Fig. 6; Table 1). Interestingly, the methylation patterns in the diverged regions were clearly different between the cultivars in terms of methylation level and distribution. It is possible that that the entire and more frequent DNA methylation that occurred in ‘Kohi’ could cause Act1a silencing.
In conclusion, we found an A. chinensis ‘Kohi’ with a low actinidin level, which is caused by Act1a suppression. This is presumably derived from sequence variations occurring in the upstream region of the gene. To our knowledge, a TF for Act1a transcription has not yet been identified. Given that actinidin gradually accumulates as fruit mature in ‘Hayward’ (Lewis and Luh, 1988; Lin et al., 1993), the unknown TF would be induced as fruit develops and bind to a cis-element in the Act1a promoter to induce gene expression. Our results suggest that such TF binding is disrupted by the modified promoter sequence in ‘Kohi’, possibly via epigenetic regulation involving DNA methylation. Although understanding the detailed mechanism underlying the lower actinidin phenotype in ‘Kohi’ requires further studies, it is apparently different from the mechanism in ‘Hort16A’, in which the regulation is at the post-translational level. Utilization of multiple low actinidin cultivars with differential genetic mechanisms may provide new strategies for breeding.
The authors wish to thank Maki Miyazawa in Kanagawa Prefectural Institute of Public Health for sharing experimental materials and providing helpful information about kiwifruit allergens. We also thank members of the Laboratory of Horticultural Sciences in College of Bioresource Sciences at Nihon University for their support on this work.
