Artificial Control of the Prunus Self-incompatibility System Using Antisense Oligonucleotides Against Pollen Genes

Abstract

Prunus (Rosaceae) includes many commercially important fruit crop species that exhibit self-incompatibility (SI), including sweet cherry (P. avium L.), Japanese apricot (P. mume Sieb. et Zucc.), Japanese plum (P. salicina Lindl.), apricot (P. armeniaca L.), and almond (P. dulcis [Mill.] D. A. Webb.). These species exhibit S-ribonuclease-based gametophytic SI, which prevents self-pollen tube growth in the pistil. The successful production of self-fertilized progeny accomplished by artificially overcoming the SI barrier has not been reported in Prunus, but self-compatible (SC) Prunus mutants with mutated pollen S determinant S haplotype–specific F-box (SFB) or pollen modifier M-locus encoded glutathione S-transferase-like (MGST) genes have been identified and used as SC cultivars and breeding stocks. In this study, we suppressed translation of SFB or MGST mRNA in self-pollen using antisense oligonucleotides to overcome the SI barrier in P. avium, P. mume, and P. salicina. Over the three years of the study, we obtained self-fertilized progeny of SI Japanese plum ‘Sordum’ only when SFB or MGST was knocked down. We also found that the average length of the self-pollinated pollen tube in the pistil of ‘Sordum’ was increased following treatment with an antisense oligonucleotide against SFB. This is the first report regarding the successful production of selfed progeny of Prunus obtained by disrupting SI. Our findings also provide evidence that the loss of function of SFB or MGST in Prunus pollen leads to SC.

Introduction

Plants have developed reproductive isolation mechanisms to maintain genetic diversity within a species. One such strategy, self-incompatibility (SI), prevents self-fertilization and fertilization between genetically related individuals. However, these reproductive barriers are an obstacle to efficient fruit production, plant breeding,and genetic analysis. To ensure cross-pollination and fruit set, growers therefore use pollinators, such as honey bees, or carry out artificial pollination. These practices are very time- and labor-intensive and increase production costs. In addition, global climate change and extreme weather events affect flowering time and pollinator behavior, resulting in erratic fruit set.

Extensive research has been conducted to overcome problems related to SI. Many reports have appeared that describe the successful breakdown of the sporophytic-type SI of Brassicaceae using a variety of techniques, such as bud pollination, old-flower pollination, and pre-treatment of the pistil with carbon dioxide, hot water, salt, or electricity (Kakizaki, 1930; Nakanishi and Hinata, 1973; Okazaki and Hinata, 1987; Roggen et al., 1972; Tlngdong et al., 1992). In addition, techniques such as old-flower pollination, pre-treatment of the pistil with hot water, and stigma excision are effective at overcoming gametophytic-type SI in Liliaceae (Ascher and Peloquin, 1966; Hopper et al., 1967; Okazaki and Murakami, 1992). S-ribonuclease (S-RNase)-based gametophytic SI, one of the most widespread SI mechanisms, can be overcome in Japanese pear (Pyrus pyrifolia [Burm. f.] Nakai) by bud pollination (Hiratsuka et al., 1985a, b) and in Nicotiana alata and apple (Malus × domestica Borkh.) by the mentor pollen method (Ramulu et al., 1979; Visser, 1981). In contrast, no reports have appeared regarding the successful breakdown of SI in Prunus (Rosaceae), a genus containing many commercially important SI fruit tree species, such as sweet cherry (P. avium), Japanese apricot (P. mume), Japanese plum (P. salicina), apricot (P. armeniaca), and almond (P. dulcis). Tsuruta and Mukai (2016) reported that stigma excision and bud pollination did not induce fruit set after self-pollination in ‘Somei-yoshino’ (Prunus × yedoensis Matsum.). Although hot water pre-treatment of the pistil of ‘Somei-yoshino’ promoted self-pollen tube growth in the pistil, the treatment also stimulated fruit drop (Tsuruta et al., 2020).

In the S-RNase-based SI system shared among Rosaceae, Solanaceae, Plantaginaceae, and Rutaceae, self- and non-self pollen are discriminated by the pistil-part S determinant S-RNase and the pollen-part S determinant F-box protein (Liang et al., 2020; McClure, 2009). In Prunus, the pollen-part S determinant is named S haplotype–specific F-box (SFB). In addition to S determinant genes, pollen-part SI-modifiers have been found in self-compatible (SC) mutants of sweet cherry and apricot and designated as M-locus-encoded glutathione S-transferase-like (MGST) (Ono et al., 2018) and M-locus DsbA-like oxidoreductase (ParMDO) (Muñoz-Sanz et al., 2017), respectively. Because these genes are orthologs, we use the term MGST in this paper. Given the many SC Prunus cultivars with mutated SFB or MGST genes that have been identified (Muñoz-Sanz et al., 2017; Ono et al., 2018; Sonneveld et al., 2005; Tao et al., 2007; Ushijima et al., 2004; Vilanova et al., 2006), the downregulation or knockdown of SFB or MGST should induce SC in Prunus. However, because of difficulties associated with genetic transformation in Prunus, a transgenic approach to produce SC has yet to be applied to this plant genus. Furthermore, the molecular functions of SFB and MGST in the SI reaction have not yet been investigated using transgenic engineering.

In this study, as an alternative approach to transgenic methods, we tested the use of antisense oligonucleotides (A-ONs) to inhibit translation of SFB and MGST mRNAs, the pollen-expressing genes supposedly indispensable for SI. Because pollen grains take up surrounding liquids by endocytosis during polar cell expansion, pollen may be an appropriate cell type for the efficient delivery of A-ONs into the cytoplasm. In one study, incubation of pollen in an artificial liquid medium supplemented with A-ONs was effective for inducing gene knockdown in Arabidopsis pollen even though a transfection reagent was not used (Mizuta and Higashiyama, 2014). An A-ON binds to its complementary sequence on the mRNA of a target gene. When the target lies in the upstream untranslated region (5' UTR) of an mRNA, cap-dependent translation is blocked, and the encoded protein is not produced. Phosphorothioate (S-) modified oligo DNA (SD), an ON backbone produced by substituting a non-bridging oxygen with sulfur, is resistant to various nucleases in cytoplasm and has been widely used in gene knockdown experiments. Examples include functional analyses of SI-related pollen-part genes in Malineae (Rosaceae) (Chen et al., 2018; Gu et al., 2015, 2019; Li et al., 2016, 2018; Meng et al., 2014) and pollen-expressed long non-coding RNAs in sweet cherry (Li et al., 2019). Morpholino oligonucleotide (MO), another nuclease-resistant ON backbone, may also be useful because of its advantages in terms of sequence specificity, water solubility, non-toxicity, stability, and knockdown effectiveness (Moulton, 2016).

The objective of the present study was the development of an artificial method for overcoming SI via the knockdown of SFB or MGST by A-ONs (A^SD-SFB, A^SD-MGST, A^MO-SFB, or A^MO-MGST) in pollen. We performed two different types of experiments applying on-tree and in vivo trials to test the effect of A-ONs on SI breakdown. In the on-tree trial, fruit set and the seed formation rate were evaluated after self-pollination of A-ON-treated pollen. To confirm whether the generated seeds were derived from self-pollination, we performed S genotyping (Beppu et al., 2002; Tao et al., 1997, 1999, 2000, 2002; Yamane and Tao, 2009; Yamane et al., 2001) and GRAS-Di analyses to determine kinship relationships (Enoki and Takeuchi, 2018). The mechanism underlying SI in Prunus is inhibition of pollen tube growth in the pistil after self-pollination. In the in vivo trial, SI breakdown was evaluated by measuring the length of the pollen tube in the pistil after self-pollination.

Materials and Methods

Plant materials

Pollination was conducted in the experimental orchard of Kyoto University in the 2019, 2020, and 2021 flowering seasons using 13 cultivars from three SI Prunus species: Japanese apricot ‘Nanko’ (S₁S₇), ‘Kairyo-Uchidaume’ (S₃S₄), ‘Kensaki’ (S_fS_f), ‘Shirokaga’ (S₂S₆), and ‘Aojiku’ (S₂S₅); Japanese plum ‘Sordum’ (S_aS_b), ‘Santarosa’ (S_cS_e), and ‘Karari’ (S_bS_t); and sweet cherry ‘Satonishiki’ (S₃S₆), ‘Rainier’ (S₁S₄), ‘Benishuho’ (S₄S₆), ‘Takasago (Rockport Bigarreau)’ (S₁S₆), and ‘Cristobalina’ (S₃S₆). Pollen-part SC cultivars, such as ‘Kensaki’, ‘Santarosa’, ‘Karari’, and ‘Cristobalina’, were used as seed parents or cross-pollination controls. ‘Nanko’, ‘Sordum’, and ‘Satonishiki’ were subjected to self-pollination using pollen treated with A-ON against SFB, whereas self-pollination of ‘Nanko’, ‘Kairyo-Uchidaume’, ‘Sordum’, ‘Rainier’, and ‘Benishuho’ used pollen pre-treated with A-ON against MGST. The other cultivars were mainly used as cross-pollination and non-treated self-pollination controls.

Anthers collected from flowers at the balloon stage were dried using silica gel desiccant for 24 h. Pollen grains released from dehisced anthers were stored until use at 4°C or −80°C for short- or long-term storage, respectively. Japanese apricot and Japanese plum trees used for pollination experiments were covered with insect-proof netting or isolated in a glasshouse or phytotron to prevent contamination from non-self pollen. In addition to the 13 cultivars described above, the following cultivars were subjected to DNA analysis: sweet cherry ‘Napoleon’; European plum (P. domestica L.) ‘Stanley’; peach (P. persica [L.] Batsch) ‘Shimizu Hakuto’ and ‘Okinawa’; apricot ‘Toa’; and ‘Tsuyuakane’, an interspecific hybrid of Japanese plum and Japanese apricot.

SFB and MGST sequence comparison

The high sequence similarity of MGST across Prunus was confirmed by aligning MGST genomic sequences of the following species in MAFFT v7 (https://mafft.cbrc.jp/alignment/server/) (Katoh and Standley, 2013): Japanese apricot (Pm4_19230131-19231977; Prunus mume_V1.0 GENOME ASSEMBLY) (Zhang et al., 2012), sweet cherry (PAV_r1.0chr3_18395861.18397694; Prunus avium Whole Genome Assembly v1.0 & Annotation v1) (Shirasawa et al., 2017), peach (Pp03:23936631.23939253; Prunus persica Whole Genome Assembly v2.0 & Annotation v2.1) (Verde et al., 2013), and the apricot ortholog of MGST, ParMDO (Muñoz-Sans et al., 2017) (GenBank accession no. KY429940.1). Sequence conservation among cultivars was confirmed by mapping Illumina mRNA sequencing reads of the pollen of sweet cherry ‘Satonishiki’, ‘Rainier’, ‘Cristobalina’, and ‘JI2434’ (Ono et al., 2018) using the Burrows-Wheeler Aligner v0.7.7 (https://github.com/lh3/bwa) and by Sanger sequencing (Eurofins Genomics, Luxembourg, Germany) of the target region using pollen cDNA of Japanese apricot ‘Nanko’ and ‘Kairyo-Uchidaume’. Total RNA was extracted with PureLink Plant RNA reagent (Thermo Fisher Scientific, Waltham, MA, USA) and used as a template for cDNA synthesis using ReverTra Ace qPCR RT Master Mix (Toyobo, Osaka, Japan). PCR conditions are listed in Table S1.

The mRNA sequences of Japanese apricot PmuS₁FB (GenBank accession no. AB101440.1) and PmuS₇FB (AB101441.1) were aligned using MAFFT to design an A-ON to inhibit translation of both PmuS₁FB and PmuS₇FB. The genomic DNA sequences of PavS₃FB (PAV_r1.0chr6: 21901650.21902780) and PavS₆FB (PAV_r1.0chr6: 21939079.21940206) including 50 bp upstream of the start codon, obtained from the reference genome (Prunus avium Whole Genome Assembly v1.0 & Annotation v1) (Shirasawa et al., 2017), and the mRNA sequence of PavS₃FB (GenBank accession no. AY805057.1) and PavS₆FB (EU077238.1) were aligned using MAFFT. The mRNA sequence of the region surrounding the PsaS_aFB start codon was determined by PCR and Sanger sequencing (Table S1) using cDNA obtained from ‘Sordum’ pollen.

A-ON design

ONs used in this study are listed in Table 1. We used two types of nuclease-resistant A-ONs: A^SD (Eurofins Genomics), which was phosphorothioated in the 3 bp of the both 5'- and 3'-termini, and A^MO (Gene Tools, Philomath, OR, USA). Sense-strand and reverse-antisense strand sequences were used as A^SD and A^MO controls, respectively (C^SD: control-SD; C^MO: control-MO). To determine the target sequence, we used Soligo software (http://sfold.wadsworth.org/cgi-bin/index.pl) to predict knockdown efficiency and used Mfold (Zuker, 2003) to avoid sequences with a high capacity to form secondary structures. To inhibit translation of target genes by ribosomes, we selected the target sequence from the 5' UTR or the sequence overlapping the start codon. Ssearch36 in the FASTA3 package (Pearson, 2016) was used to confirm ON sequence similarities and off-target possibilities based on coding sequence data of the following: Japanese apricot, sweet cherry, and peach (Shirasawa et al., 2017; Verde et al., 2013; Zhang et al., 2012), and the paralog of sweet cherry MGST (Pav_sc0000661.1_g340.1-2) (Ono et al., 2018), which is not annotated on the reference genome.

Table 1

Sequences, target genes, and concentrations of ONs used in this study.

ON treatment and pollination

ONs at a final concentration of 10 to 50 μM were dissolved in pollen germination medium (12.5% PEG6000, 10% sucrose, 0.03% casein, 1 mM Ca(NO₃)₂, 1 mM KNO₃, 0.8 mM MgSO₄, and 160 mM H₃BO₃) (Table 1). To introduce ONs into pollen cytoplasm, approximately 0.5 mg of pollen was incubated in 100 μL of pollen germination medium in the bottom of an upside-down, 1.5-mL microfuge tube (Fig. 1). After approximately 10 min incubation, the solution was centrifuged using a tabletop microcentrifuge at approximately 2,000×g for 1 min. The pollen paste was then collected and directly applied to the stigma to be pollinated. We also pollinated stigmas with two ON-untreated pollen controls: pollen incubated in germination medium without ONs (“paste” pollen), and non-incubated pollen grains (“powder” pollen), which was applied with a glass rod. ON delivery into the pollen tube was confirmed by fluorescence microscopic observation using fluorescein-conjugated ON dissolved in the pollen germination medium at an excitation wavelength of 460 to 495 nm (BX53, U-HGLGPS, U-FBW, cellSens; Olympus, Tokyo, Japan).

Fig. 1

Schematic representation of ON treatment and pollination methods used in this study.

On-tree SI evaluation experiment

We monitored floral pollination, fruit set, and seed formation in the flowering and harvesting seasons in 2019, 2020, and 2021 (Table S2). DNA was extracted from seed embryos and young leaves of cultivars in the experimental orchard using cetyltrimethylammonium bromide. Embryo S haplotypes were analyzed by PCR to determine if collected seeds were derived from self-fertilization (Table S1). The S haplotyping of Japanese apricot and sweet cherry embryos was carried out using general S-RNase primer sets that amplify most Prunus S-RNase alleles, while the S haplotypes of embryos of three Japanese plum cultivars, ‘Sordum’ (S_aS_b), ‘Santarosa’ (S_cS_e), and ‘Karari’ (S_bS_t), were distinguished using S-RNase allele-specific primer sets. Embryos harboring only one S haplotype were further analyzed by quantitative real-time PCR (qRT-PCR) on a LightCycler 480 instrument (Roche Diagnostics, Mannheim, Germany) to determine S-RNase allele copy numbers (Table S1), with ubiquitin-like and β-actin genes used as internal controls.

Some of the embryos obtained from self-pollination of Japanese plum ‘Sordum’ in 2019 and 2020 were further analyzed at the genome-wide scale. A paired-end 100-bp genomic DNA sequencing library was constructed with 63 primers on a GRAS-Di system (Toyota, Aichi, Japan) and then sequenced on the Hiseq 4000 platform (Illumina, San Diego, CA, USA). GRAS-Di software was used to identify dominant markers and their positions in the Japanese apricot reference genome (Zhang et al., 2012). To identify the candidate pollen parent having a similar marker combination to the progeny, principal component analysis of the marker data for each sample was conducted using the prcomp function in R.

Seeds obtained from self-pollination of ‘Sordum’ in 2021 were analyzed using two PCR primer sets. The first PCR analysis was conducted to exclude the possibility of interspecific hybridization with pollen from other Prunus species in the orchard where the experimental pollinations were conducted, as pollen grains from many Prunus species can readily grow in the pistils of Japanese plum (Morimoto et al., 2019). A Japanese plum-specific single nucleotide polymorphism (SNP) was identified in the early shoot regeneration (ESR) gene on chromosome 1 (data not shown) by mapping genomic DNA sequencing reads from four Japanese plum, 113 Japanese apricot, seven apricot (Numaguchi et al., 2019), three peach (Nakano et al., 2020), and seven sweet cherry cultivars (Ono et al., 2018). The Japanese plum-specificity of the SNP was further confirmed by sequencing the ESR gene of three Japanese plum, one Japanese apricot, one apricot, and two peach cultivars, and ‘Stanley’, the only European plum planted in the same orchard as described below. The second PCR primer set, for amplification of a ‘Sordum’-specific SNP on chromosome 7 of the Japanese apricot reference genome, was designed using the GRAS-Di data from 2019 and 2020. This PCR analysis was used to exclude the possibility of seed formation derived from non-self pollen of Japanese plum cultivars planted in the same orchard.

In vivo SI evaluation experiment

Japanese apricot and sweet cherry flower buds at the balloon stage of development were excised from branches and emasculated. After pollination, the emasculated Japanese apricot and sweet cherry flower buds were incubated for 70 h and 120 h, respectively, on 1% agar medium at 22°C. Japanese plum branches with emasculated flowers were also used in this experiment. After pollination, the basal portion of each branch was placed in water and incubated for 120 h at 22°C. Following incubation, pistils were fixed in a 1:3:1 mixture of chloroform, 90% ethanol, and glacial acetic acid overnight at 4°C. The solution was then replaced with 100% ethanol, and the samples were stored at 4°C until use. For the SI evaluation, samples were softened in 10 N NaOH for 1 h at 60°C, incubated for 5 h at room temperature, and stained overnight with 0.1% aniline blue in 0.1 N K₃PO₄. Stained pistils were immersed in 50% glycerol and then squashed between a slide glass and a cover glass for fluorescence microscopic observation at an excitation wavelength of 340 to 390 nm (BX53, U-HGLGPS, U-FUW, cellSens; Olympus). The length of the longest pollen tube in each pistil and the length of the squashed pistils were measured using ImageJ software (https://imagej.nih.gov/ij/). The Steel–Dwass test was used as a post-hoc test of the Kruskal–Wallis test to evaluate the effect of ONs on the pollen tube length. P-values < 0.01 were considered statistically significant.

Results

ON design

Genomic positions of the ONs designed in this study are shown in Figure S1. The same ONs, A^MO-MGST and C^MO-MGST, were used for the MGST genes of both Japanese apricot and Japanese plum. A^SD-SFB and C^SD-SFB were designed using the common sequence of the alleles of Japanese apricot ‘Nanko’: PmuS₁FB (AB101440.1) and PmuS₇FB (AB101441.1). A^MO-SFB and C^MO-SFB, the ONs for the Japanese plum ‘Sordum’ (S_aS_b) SFB, were designed solely according to the PsaS_aFB sequence, as sequence information was not available for PsaS_bFB. An equal mixture of the SDs targeting PavS₃FB and PavS₆FB, which differed by a single bp, was used for the knockdown of sweet cherry PavSFB. ON delivery into pollen tubes was confirmed by fluorescence microscopic observation (Fig. S2).

On-tree experiment

The number of pollinated pistils and number of fruits set under each condition are shown in Tables 2 and S3. Self-pollination with non-treated and control oligonucleotide (C-ON)-treated pollen produced no self-fertilized progeny, thereby confirming the robustness of Prunus SI. In contrast, 10 fruits were obtained after self-pollination of Japanese plum ‘Sordum’ when SFB or MGST were knocked down. However, no embryo could be obtained from two of these fruits, while only three formed mature seeds: two were obtained by A^MO-SFB-treatment in 2021, and one was obtained by A^MO-MGST-treatment in the same year (Figs. 2 and S3). According to PCR S haplotyping analyses, eight embryos obtained seemed to be derived from self-fertilization (Fig. 3). Any progeny bearing an S haplotype that was absent in the self-pollinated parent must have originated through contamination by non-self pollen, whereas progeny having the same S haplotypes as the parent are expected to have been derived from the breakdown of SI and self-fertilization. S haplotyping revealed that one, one, three, and one seed(s) of Japanese apricot ‘Nanko’, sweet cherry ‘Satonishiki’, ‘Rainier’, and ‘Takasago (Rockport Bigarreau)’, respectively, originated from contamination with non-self pollen (Fig. S4) (These fruits are not included in the counts in Tables 2 and S3). qRT-PCR confirmed that the ‘Sordum’ self-pollinated embryo with only the S_b allele was homozygous for S_b (Fig. S5). A^MO-SFB treatment also produced several S_b homozygous ‘Sordum’ progeny, which suggests that A^MO-SFB that was designed to knockdown PsaS_aFB, is also effective for the knockdown of PsaS_bFB.

Table 2

Fruit set rate and SI breakdown rate in the on-tree pollination experiment.

Fig. 2

Images of fruits, pits, seeds, and embryos obtained from self-pollination of Japanese plum ‘Sordum’ in 2021. Fruits, pits, seeds, and embryos obtained by cross-pollination of ‘Sordum’ with ‘Karari’ are also shown as cross-pollination controls.

Fig. 3

PCR S haplotyping of progeny obtained from self-pollination of ‘Sordum’. S haplotyping by PCR amplification of S-RNase was carried out using Japanese plum allele-specific primer sets. F₁ progenies obtained by pollination of ‘Sordum’ pistils with ‘Santarosa’ pollen are shown as cross-pollination controls.

Genome analysis to confirm the breakdown of SI

Although the SI of ‘Sordum’ exhibited rigidity in the above-mentioned experiment (Table 2) and in the study of Beppu et al. (2012), S_b pollen of ‘Karari’ has been reported to overcome the SI barrier through an unknown mechanism (Beppu et al., 2010). We therefore subjected embryos obtained by the self-pollination of ‘Sordum’ to further GRAS-Di analysis or two additional rounds of PCR and Sanger sequencing to exclude the possibility that the fruit set was due to contamination by S_b pollen of ‘Karari’ growing near the experimental ‘Sordum’ trees. Two embryos derived from the self-pollination of ‘Sordum’, one from A^MO-MGST treatment in 2019 and one from A^MO-SFB treatment in 2020, were analyzed by GRAS-Di (Table S4). The other ‘Sordum’ self-pollinated embryo obtained by A^MO-MGST treatment in 2020 could not be analyzed because the amount and quality of extracted DNA was low. We first performed a kinship analysis on 24,115 markers that were amplified in up to seven of the eight samples. Among these markers, 15,610 were uniformly mapped onto chromosomes 1 to 8 of the Japanese apricot genome (Zhang et al., 2012) (Fig. S6A). The first principal component (PC1) of the marker data divided the samples according to species (Fig. 4). PC1 scores of all ‘Sordum’ progenies were very similar to those of three Japanese plum cultivars, thus eliminating the possibility that the progenies were derived from interspecific hybridization. In addition, the PC2 scores of both progenies derived from self-pollination were very similar to that of ‘Sordum’, whereas the score of the control progeny derived from cross-pollination of ‘Sordum’ with ‘Santarosa’ was intermediate between those of the two parents. This result suggests that the two seeds were derived from the self-fertilization of ‘Sordum’. A second kinship analysis was performed on 15,548 markers that were amplified in at least one Japanese plum sample. Among them, 8,612 markers were uniformly mapped onto chromosomes 1 to 8 of the reference genome (Fig. S6B). Most of the amplified markers in the ‘Santarosa’ and ‘Karari’ genomes were not inherited by the two progenies obtained by self-pollination, whereas many of the ‘Santarosa’ markers were inherited by the ‘Sordum’ × ‘Santarosa’ progeny (Fig. 5). Taking all of these results into consideration, we conclude that the two seeds were derived from the breakdown of SI in ‘Sordum’.

Fig. 4

Principal component analysis plot of GRAS-Di markers used to identify the pollen parent of ‘Sordum’ progeny obtained after self-pollination using A^MO-treated pollen.

Fig. 5

Characterization of the presence/absence of GRAS-Di markers in ‘Sordum’ progeny obtained after self-pollination using A^MO-treated pollen. In the Venn diagram, each denominator and numerator indicate the number of markers present in each cultivar and each seed, respectively. (A) ‘Sordum’ seeds obtained after pollination with ‘Santarosa’, used as cross-pollination controls. (B–C) ‘Sordum’ seeds obtained by self-pollination with A^MO-treated ‘Sordum’ pollen.

In 2021, an additional three and four fruits were obtained by self-pollination of ‘Sordum’ using pollen treated with A^MO-SFB and A^MO-MGST, respectively. Although two fruits from the A^MO-MGST-treated pollen pollination could not be analyzed because embryo DNA could not be obtained, the other five embryos were confirmed to be homozygous for the Japanese plum-specific SNP on the ESR gene (Fig. 6A). In addition, a ‘Sordum’-specific marker on chromosome 7 was confirmed to be homozygously present in the genome (Fig. 6B). During the three years of the study, we obtained four ‘Sordum’ seeds apparently derived from SI breakdown due to A^MO-SFB treatment and at least three, and at most six, seeds resulting from SI breakdown following A^MO-MGST treatment. Fruit set rates (i.e., SI breakdown rates) after the self-pollination of ‘Sordum’ were 0.31% and 0.24–0.47% for A^MO-SFB and A^MO-MGST treatments, respectively.

Fig. 6

Sanger sequencing of a Japanese plum-specific SNP on the ESR gene (A) and a ‘Sordum’-specific SNP on chromosome 7 (B). (A) The position of the Japanese plum-specific SNP on the ESR gene is shown at the top. Sanger sequencing was performed to type the SNP genotype of the five individuals obtained by self-pollination of ‘Sordum’ in 2021 and eight Prunus cultivars. Japanese plum-specificity of the SNP genotype was confirmed using genomic sequencing data (Ono et al., 2018; Nakano et al., 2020; Numaguchi et al., 2019). Japanese apricot, apricot, peach, and sweet cherry were homozygous for “G”, whereas Japanese plum was homozygous for “A” (data not shown). (B) The position of the Japanese plum-specific SNP on chromosome 7 is shown at the top. The ‘Sordum’-specific SNP genotype of the five individuals obtained by self-pollination of ‘Sordum’ in 2021 was analyzed by Sanger sequencing. Three Japanese plum cultivars (‘Sordum’, ‘Santarosa’, and ‘Karari’) and one interspecific hybrid between Japanese apricot and Japanese plum (‘Tsuyuakane’), which were planted in the same orchard, were analyzed. An F₁ individual from a cross of ‘Sordum’ × ‘Santarosa’ was also analyzed as a control. Primer sequences used are listed in Table S1.

In vivo experiment

Because the length of the longest pollen tube in the pistil varied considerably (Fig. 7), a statistically significant promotional effect on the pollen tube length by the A-ONs treatment was found only in the A-SFB treatment in Japanese plum ‘Sordum’. In all three tested Prunus species, however, a tendency towards a longer pollen tube in the pistil were observed when SFB was knocked down than when C-ON-treated pollen was used (Fig. S7). Although A-MGST treatment had no discernable effect on pollen tube growth, this result may be consistent with the very low fruit set rate observed in the on-tree experiment.

Fig. 7

Effect of ON treatment on pollen tube growth in the pistil. The length of the longest pollen tube in each pistil of Japanese apricot (A), Japanese plum (B), and sweet cherry (C) is shown. Dark bars represent the average value. Lower case letters at the bottom of the graphs represent significant difference by Steel‒Dwass test (P < 0.01). The average length of the pistil is shown on the left.

Discussion

In this study, we demonstrated that A-ON-induced suppression of translation of SFB and MGST mRNA in pollen tubes may be used to artificially break down the rigid SI system of Prunus. This A-ON strategy for suppressing pollen S determinant gene(s) is applicable to Prunus, but cannot be used on other plants with an S-RNase-based SI system; this is because Prunus has a Prunus-specific self-recognition system, whereas the other species have a collaborative nonself-recognition system (Tao and Iezzoni, 2010). Mutation or dysfunction of the pollen S determinant gene only leads to SC in the Prunus system (Sonneveld et al., 2005; Tao et al., 2007; Ushijima et al., 2004; Vilanova et al., 2006). Similarly, MGST is a Prunus-specific gene, so our MGST knockdown strategy using A-ONs cannot be applied to Maloideae (Rosaceae), Solanaceae, Plantaginaceae, or Rutaceae (Muñoz-Sanz et al., 2017; Ono et al., 2018).

Although our data suggest that SFB and MGST are both indispensable for SI in Prunus, we found that the efficiency of SI breakdown following A^MO-SFB and A^MO-MGST treatments was very low, with fruit set rates of only 0.31% and 0.24–0.47%, respectively. Furthermore, self-fertilization was not confirmed after self-pollination with A-ON-treated pollen in Japanese apricot or sweet cherry, although tendencies towards promotional effects on self-pollen tube growth were observed after A-SFB treatment in the in vivo pollination experiment.

To enhance SI breakdown, several factors need to be considered. First, pollen tube growth varied only slightly between the A-ON and C-ON treatments in the in vivo analysis, which means that additional A-ON may be needed to fully suppress mRNA translation, but that an excess amount may be toxic to pollen. Optimization of ON backbone types, concentrations, and sequences, all of which affect gene suppression efficacy and toxicity, is therefore necessary. All the self-fertilized fruits obtained in this study were derived from A^MO-treated pollen, and A^MO thus appears to be more effective than A^SD. Given the results of the in vivo pollination experiment, however, we cannot clearly conclude that A^MO is more effective than A^SD for overcoming SI. A second consideration in regard to the efficacy of SI breakdown is the amount of ON-treated pollen on the stigma. Considerably fewer ON-treated paste-form pollen grains were observed on the stigma compared with normally pollinated powder-form pollen grains (Fig. S7). Our pollination methods, as well as A-ON treatment methods, need to be improved to increase the number of pollen grains attached to the stigma. In this study, only Japanese plum ‘Sordum’ bore self-fertilized fruits following A-ON treatment, but we observed a higher percentage of aborted seeds. This seed abortion after self-fertilization may be due to inbreeding depression. Further studies are needed to confirm whether inbreeding depression affected the fruit set of the sweet cherry and Japanese apricot cultivars used in this study and whether self-pollinated fruit were obtained from these species. Finally, the expression level of the target gene needs to be considered. In the in vivo pollination experiment, pollen tube growth in the pistil was relatively extensive following knockdown of SFB than that of MGST in all three analyzed species. This result may be attributed to the different expression levels of SFB and MGST because SFB expression is much weaker than that of MGST in pollen (Aguiar et al., 2015; Matsumoto and Tao, 2019; Ono et al., 2018). The strong expression of MGST may affect gene suppression efficiency and thus pollen tube growth in vivo.

In conclusion, we successfully produced self-fertilized progeny of an SI Prunus cultivar by treating pollen with A-ONs against SFB or MGST. Our results suggest that SFB and MGST are both essential to the SI of Prunus. Although the frequency of SI breakdown achieved in this study is not sufficient for practical use, our findings have important implications for the future development of an artificial SI breakdown technique.

Acknowledgements

We are grateful to Drs. Daiki Matsumoto and Takuya Morimoto for helpful discussions. We thank Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript. This work was partially supported by JSPS KAKENHI Grant Numbers 19J14774 to KO and 19H00941 to RT.

Literature Cited

•  Aguiar,  B.,  J.  Vieira,  A. E.  Cunha,  N. A.  Fonseca,  A.  Iezzoni,  S.  Van Nocker and  C. P.  Vieira. 2015. Convergent evolution at the gametophytic self-incompatibility system in Malus and Prunus. PLoS One 10: e0126138. DOI: 10.1371/journal.pone.0126138.
•  Ascher,  P. D. and  S. J.  Peloquin. 1966. Effect of floral aging on the growth of compatible and incompatible pollen tubes in Lilium longiflorum. Amer. J. Bot. 53: 99–102.
•  Beppu,  K.,  M.  Kumai,  H.  Yamane,  R.  Tao and  I.  Kataoka. 2012. Molecular and genetic analyses of the S-haplotype of the self-compatible Japanese plum (Prunus salicina Lindl.) ‘Methley’. J. Hortic. Sci. Biotechnol. 87: 493–498.
•  Beppu,  K.,  K.  Konishi and  I.  Kataoka. 2010. S-haplotypes and self-compatibility of the Japanese plum cultivar ‘Karari’. Acta Hortic. 929: 261–266.
•  Beppu,  K.,  H.  Yamane,  H.  Yaegaki,  M.  Yamaguchi,  I.  Kataoka and  R.  Tao. 2002. Diversity of S-RNase genes and S-haplotypes in Japanese plum (Prunus salicina Lindl.). J. Hortic. Sci. Biotechnol. 77: 658–664.
•  Chen,  J.,  P.  Wang,  B. H. J.  de Graaf,  H.  Zhang,  H.  Jiao,  C.  Tang,  S.  Zhang and  J.  Wu. 2018. Phosphatidic acid counteracts S-RNase signaling in pollen by stabilizing the actin cytoskeleton. Plant Cell 30: 1023–1039.
• Enoki, H. and Y. Takeuchi. 2018. New genotyping technology, GRAS-Di, using next generation sequencer. Proceedings of the Plant and Animal genome conference XXVI: P0153.
•  Gu,  Z.,  W.  Li,  J.  Doughty,  D.  Meng,  Q.  Yang,  H.  Yuan,  Y.  Li,  Q.  Chen,  J.  Yu,  C. S.  Liu and  T.  Li. 2019. A gamma-thionin protein from apple, MdD1, is required for defence against S-RNase-induced inhibition of pollen tube prior to self/non-self recognition. Plant Biotechnol. J. 17: 2184–2198.
•  Gu,  Z.,  D.  Meng,  Q.  Yang,  H.  Yuan,  A.  Wang,  W.  Li,  Q.  Chen,  Y.  Zhang,  D.  Wang and  T.  Li. 2015. A CBL gene, MdCBL5, controls the calcium signal and influences pollen tube growth in apple. Tree Genet. Genomes 11: 27. DOI: 10.1007/s11295-015-0853-2.
•  Hiratsuka,  S.,  N.  Hirata,  T.  Tezuka and  Y.  Yamamoto. 1985a. Self-incompatibility reaction of Japanese pear various stages of floral development. J. Japan. Soc. Hort. Sci. 54: 9–14.
•  Hiratsuka,  S.,  M.  Hirota,  E.  Takahashi and  N.  Hirata. 1985b. Seasonal changes in the self-incompatibility and pollen tube growth in Japanese pears (Pyrus serotina Rehd.). J. Japan. Soc. Hort. Sci. 53: 377–382.
•  Hopper,  J. E.,  P. D.  Ascher and  S. J.  Peloquin. 1967. Inactivation of self-incompatibility following temperature pretreatments of styles in Lilium longiflorum. Euphytica 16: 215–220.
•  Kakizaki,  Y. 1930. Studies on the genetics and physiology of self and cross-incompatibility in the common cabbage (Brassica oleracea L. var. capitata L.). Jpn. J. Bot. 5: 133–208.
•  Katoh,  K. and  D. M.  Standley. 2013. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30: 772–780.
•  Li,  W.,  D.  Meng,  Z.  Gu,  Q.  Yang,  H.  Yuan,  Y.  Li,  Q.  Chen,  J.  Yu,  C.  Liu and  T.  Li. 2018. Apple S-RNase triggers inhibition of tRNA aminoacylation by interacting with a soluble inorganic pyrophosphatase in growing self-pollen tubes in vitro. New Phytol. 218: 579–593.
•  Li,  W.,  Q.  Yang,  Z.  Gu,  C.  Wu,  D.  Meng,  J.  Yu,  Q.  Chen,  Y.  Li,  H.  Yuan,  D.  Wang and  T.  Li. 2016. Molecular and genetic characterization of a self-compatible apple cultivar, ‘CAU-1’. Plant Sci. 252: 162–175.
•  Li,  Y.,  C.  Wu,  C.  Liu,  J.  Yu,  X.  Duan,  W.  Fan,  J.  Wang,  X.  Zhang,  G.  Yan,  T.  Li and  K.  Zhang. 2019. Functional identification of lncRNAs in sweet cherry (Prunus avium) pollen tubes via transcriptome analysis using single-molecule long-read sequencing. Hortic. Res. 6: 135. DOI: 10.1038/s41438-019-0218-3.
•  Liang,  M.,  Z.  Cao,  A.  Zhu,  Y.  Liu,  M.  Tao,  H.  Yang,  Q.  Xu Jr,  S.  Wang,  J.  Liu,  Y.  Li,  C.  Chen,  Z.  Xie,  C.  Deng,  J.  Ye,  W.  Guo,  Q.  Xu,  R.  Xia,  R. M.  Larkin,  X.  Deng,  M.  Bosch,  V. E.  Franklin-Tong and  L.  Chai. 2020. Evolution of self-compatibility by a mutant S_m-RNase in citrus. Nat. Plants 6: 131–142.
•  Matsumoto,  D. and  R.  Tao. 2019. Recognition of S-RNases by an S locus F-box like protein and an S haplotype-specific F-box like protein in the Prunus-specific self-incompatibility system. Plant Mol. Biol. 100: 367–378.
•  McClure,  B. 2009. Darwin’s foundation for investigating self-incompatibility and the progress toward a physiological model for S-RNase-based SI. J. Exp. Bot. 60: 1069–1081.
•  Meng,  D.,  Z.  Gu,  W.  Li,  A.  Wang,  H.  Yuan,  Q.  Yang and  T.  Li. 2014. Apple MdABCF assists in the transportation of S-RNase into pollen tubes. Plant J. 78: 990–1002.
•  Mizuta,  Y. and  T.  Higashiyama. 2014. Antisense gene inhibition by phosphorothioate antisense oligonucleotide in Arabidopsis pollen tubes. Plant J. 78: 516–526.
•  Morimoto,  T.,  Y.  Kitamura,  K.  Numaguchi,  T.  Akagi and  R.  Tao. 2019. Characterization of post-mating interspecific cross-compatibility in Prunus (Rosaceae). Sci. Hortic. 246: 693–699.
•  Moulton,  J. D. 2016. Guide for morpholino users: Toward therapeutics. J. Drug Discov. Dev. Deliv. 3: 1023–1035.
•  Muñoz-Sanz,  J. V.,  E.  Zuriaga,  M. L.  Badenes and  C.  Romero. 2017. A disulfide bond A-like oxidoreductase is a strong candidate gene for self-incompatibility in apricot (Prunus armeniaca) pollen. J. Exp. Bot. 68: 5069–5078.
•  Nakanishi,  T. and  K.  Hinata. 1973. An effective time for CO₂ gas treatment in overcoming self-incompatibility in Brassica. Plant Cell Physiol. 14: 873–879.
•  Nakano,  R.,  T.  Kawai,  Y.  Fukamatsu,  K.  Akita,  S.  Watanabe,  T.  Asano,  D.  Takata,  M.  Sato,  F.  Fukuda and  K.  Ushijima. 2020. Postharvest properties of ultra-late maturing peach cultivars and their attributions to melting flesh (M) locus: re-evaluation of M locus in association with flesh texture. Front. Plant Sci. 11: 554158. DOI: 10.3389/fpls.2020.554158.
•  Numaguchi,  K.,  S.  Ishio,  Y.  Kitamura,  K.  Nakamura,  R.  Ishikawa and  T.  Ishii. 2019. Microsatellite marker development and population structure analysis in Japanese apricot (Prunus mume Sieb. et Zucc.). Hort. J. 88: 222–231.
•  Okazaki,  K. and  K.  Hinata. 1987. Repressing the expression of self-incompatibility in crucifers by short-term high temperature treatment. Theor. Appl. Genet. 73: 496–500.
•  Okazaki,  K. and  K.  Murakami. 1992. Effects of flowering time (in forcing culture), stigma excision, and high temperature on overcoming of self incompatibility in tulip. J. Japan. Soc. Hort. Sci. 61: 405–411.
•  Ono,  K.,  T.  Akagi,  T.  Morimoto,  A.  Wünsch and  R.  Tao. 2018. Genome re-sequencing of diverse sweet cherry (Prunus avium) individuals reveals a modifier gene mutation conferring pollen-part self-compatibility. Plant Cell Physiol. 59: 1265–1275.
•  Pearson,  W. R. 2016. Finding protein and nucleotide similarities with FASTA. Curr. Protoc. Bioinformatics 53: 3.9.1–3.9.25.
•  Ramulu,  K. S.,  G. M. M.  Bredemeijer,  P.  Dijkhuis,  D.  de Nettancourt and  H.  Schibilla. 1979. Mentor pollen effects on gametophytic incompatibility in Nicotiana, Oenothera and Lycopersicum. Theor. Appl. Genet. 54: 215–218.
•  Roggen,  H. P. J. R.,  A. J.  Van Dijk and  C.  Dorsman. 1972. “Electric aided” pollination: A method of breaking incompatibility in Brassica oleracea L. Euphytica 21: 181–184.
•  Shirasawa,  K.,  K.  Isuzugawa,  M.  Ikenaga,  Y.  Saito,  T.  Yamamoto,  H.  Hirakawa and  S.  Isobe. 2017. The genome sequence of sweet cherry (Prunus avium) for use in genomics-assisted breeding. DNA Res. 24: 499–508.
•  Sonneveld,  T.,  K. R.  Tobutt,  S. P.  Vaughan and  T. P.  Robbins. 2005. Loss of pollen-S function in two self-compatible selections of Prunus avium is associated with deletion/mutation of an S haplotype-specific F-box gene. Plant Cell 17: 37–51.
•  Tao,  R. and  A. F.  Iezzoni. 2010. The S-RNase-based gametophytic self-incompatibility system in Prunus exhibits distinct genetic and molecular feature. Sci. Hortic. 124: 423–433.
•  Tao,  R.,  A.  Watari,  T.  Hanada,  T.  Habu,  H.  Yaegaki,  M.  Yamaguchi and  H.  Yamane. 2007. Self-compatible peach (Prunus persica) has mutant versions of the S haplotypes found in self-incompatible Prunus species. Plant Mol. Biol. 63: 109–123.
•  Tao,  R.,  T.  Habu,  A.  Namba,  H.  Yamane,  F.  Fuyuhiro,  K.  Iwamoto and  A.  Sugiura. 2002. Inheritance of S^f-RNase in Japanese apricot (Prunus mume) and its relation to self-compatibility. Theor. Appl. Genet. 105: 222–228.
•  Tao,  R.,  T.  Habu,  H.  Yamane,  A.  Sugiura and  K.  Iwamoto. 2000. Molecular markers for self-compatibility in Japanese apricot (Prunus mume). HortScience 35: 1121–1123.
•  Tao,  R.,  H.  Yamane,  A.  Sugiura,  H.  Murayama,  H.  Sassa and  H.  Mori. 1999. Molecular typing of S-alleles through identification, characterization and cDNA cloning for S-RNases in sweet cherry. J. Amer. Soc. Hort. Sci. 124: 224–233.
•  Tao,  R.,  H.  Yamane,  H.  Sassa,  H.  Mori,  T. M.  Gradziel,  A. M.  Dandekar and  A.  Sugiura. 1997. Identification of stylar RNases associated with gametophytic self-incompatibility in almond (Prunus dulcis). Plant Cell Physiol. 38: 304–311.
•  Tlngdong,  F.,  S.  Ping,  Y.  Xiaoniu and  Y.  Guangsheng. 1992. Overcoming self-incompatibility of Brassica napus by salt (NaC1) spray. Plant Breed. 109: 255–258.
•  Tsuruta,  M.,  R.  Iwaki,  C.  Lian and  Y.  Mukai. 2020. Decreased RNase activity under high temperature is related to promotion of self-pollen tube growth in the pistil of the Japanese flowering cherry, Prunus × yedoensis ‘Somei-yoshino’. Hort. J. 89: 306–310.
•  Tsuruta,  M. and  Y.  Mukai. 2016. Overcoming self-incompatibility and pollen tube growth of Japanese flowering cherry, ‘Somei-yoshino’. Hort. Res. (Japan) 15: 341–345 (In Japanese with English abstract).
•  Ushijima,  K.,  H.  Yamane,  A.  Watari,  E.  Kakehi,  K.  Ikeda,  N. R.  Hauck,  A. F.  Iezzoni and  R.  Tao. 2004. The S haplotype-specific F-box protein gene, SFB, is defective in self-compatible haplotypes of Prunus avium and P. mume. Plant J. 39: 573–586.
•  Verde,  I.,  A. G.  Abbott,  S.  Scalabrin,  S.  Jung,  S.  Shu,  F.  Marroni,  T.  Zhebentyayeva,  M. T.  Dettori,  J.  Grimwood,  F.  Cattonaro,  A.  Zuccolo,  L.  Rossini,  J.  Jenkins,  E.  Vendramin,  L. A.  Meisel,  V.  Decroocq,  B.  Sosinski,  S.  Prochnik,  T.  Mitros,  A.  Policriti,  G.  Cipriani,  L.  Dondini,  S.  Ficklin,  D. M.  Goodstein,  P.  Xuan,  C.  Del Fabbro,  V.  Aramini,  D.  Copetti,  S.  Gonzalez,  D. S.  Horner,  R.  Falchi,  S.  Lucas,  E.  Mica,  J.  Maldonado,  B.  Lazzari,  D.  Bielenberg,  R.  Pirona,  M.  Miculan,  A.  Barakat,  R.  Testolin,  A.  Stella,  S.  Tartarini,  P.  Tonutti,  P.  Arús,  A.  Orellana,  C.  Wells,  D.  Main,  G.  Vizzotto,  H.  Silva,  F.  Salamini,  J.  Schmutz,  M.  Morgante and  D. S.  Rokhsar. 2013. The high-quality draft genome of peach (Prunus persica) identifies unique patterns of genetic diversity, domestication and genome evolution. Nat. Genet. 45: 487–494.
•  Vilanova,  S.,  M. L.  Badenes,  L.  Burgos,  J.  Martínez-Calvo,  G.  Llácer and  C.  Romero. 2006. Self-compatibility of two apricot selections is associated with two pollen-part mutations of different nature. Plant Physiol. 142: 629–641.
•  Visser,  T. 1981. Pollen and pollination experiments. IV. “Mentor pollen” and “pioneer pollen” techniques regarding incompatibility and incongruity in apple and pear. Euphytica 30: 363–369.
•  Yamane,  H. and  R.  Tao. 2009. Molecular basis of self-(in)compatibility and current status of S-genotyping in rosaceous fruit trees. J. Japan. Soc. Hort. Sci. 78: 137–157.
•  Yamane,  H.,  R.  Tao,  A.  Sugiura,  N. R.  Hauck and  A. F.  Iezzoni. 2001. Identification and characterization of S-RNases in tetraploid sour cherry (Prunus cerasus). J. Amer. Soc. Hort. Sci. 126: 661–667.
•  Zhang,  Q.,  W.  Chen,  L.  Sun,  F.  Zhao,  B.  Huang,  W.  Yang,  Y.  Tao,  J.  Wang,  Z.  Yuan,  G.  Fan,  Z.  Xing,  C.  Han,  H.  Pan,  X.  Zhong,  W.  Shi,  X.  Liang,  D.  Du,  F.  Sun,  Z.  Xu,  R.  Hao,  T.  Lv,  Y.  Lv,  Z.  Zheng,  M.  Sun,  L.  Luo,  M.  Cai,  Y.  Gao,  J.  Wang,  Y.  Yin,  X.  Xu,  T.  Cheng and  J.  Wang. 2012. The genome of Prunus mume. Nat. Commun. 3: 1318. DOI: 10.1038/ncomms2290.
•  Zuker,  M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res. 31: 3406–3415.