DNA Markers and the Molecular Mechanism of Self-(in)compatibility in Japanese Pear (Pyrus pyrifolia Nakai)

Abstract

Japanese pear (Pyrus pyrifolia Nakai) is an important Rosaceous fruit tree in Japan. This species exhibits gametophytic self-incompatibility (GSI), which is controlled by a single S-locus with multiple S-haplotypes. Although GSI is a genetic mechanism to prevent inbreeding and promote outcrossing to maintain genetic diversity, it can be problematic in fruit trees because it causes unstable fruit set. Therefore, orchardists interplant Japanese pear with other lines as pollinizers or conduct artificial pollination to ensure fruit set. To achieve stable fruit production and to reduce or eliminate the need for artificial pollination, research on the GSI of Japanese pear has been conducted by pollination experiments and by characterization of self-compatible (SC) mutants. Additionally, breeding programs have progressed to produce SC cultivars with high fruit quality. Recently, molecular analyses of GSI and SC mutants in Japanese pear have provided new information that is relevant to the stable fruit production and efficient breeding of Japanese pear. This review focuses on studies of the GSI of Japanese pear, and especially on the recent development of DNA markers for S-genotyping and marker-assisted selection of SC trees. In addition, the candidate genes controlling pollen S specificity and a model of the molecular mechanism of GSI in Japanese pear are described.

Introduction

Flowering plants have evolved various genetic mechanisms to prevent inbreeding and promote outcrossing to maintain genetic diversity within a species. One such mechanism is self-incompatibility (SI), which allows the pistil to distinguish between self (genetically related) pollen and non-self (genetically unrelated) pollen. The self pollen is rejected, while the non-self pollen is accepted for fertilization. To date, SI has been recorded in more than 50% of flowering plant species (Kao and McCubbin, 1996). In many species, SI is controlled by a single polymorphic locus, the S-locus. This locus consists of at least two linked genes: the pistil S gene (female determinant gene) and the pollen S gene (male determinant gene), so variants of the gene complex are called S-haplotypes (Kao and McCubbin, 1996; Takayama and Isogai, 2005).

Among the Rosaceous fruit trees, Malus (apple), Pyrus (pear), and Prunus (almond, cherry, plum, and apricot) exhibit SI (Yamane and Tao, 2009). However, SI in fruit trees can be problematic because it causes unstable fruit set. Therefore, to ensure fruit set in orchards, cross-compatible cultivars are interplanted as pollinizers, or artificial pollination is conducted.

Japanese pear (Pyrus pyrifolia Nakai) is an important Rosaceous fruit tree in Japan, and its self- and cross-sterility have been studied for a long time. Kikuchi (1929) revealed that Japanese pear exhibited gametophytic SI (GSI) controlled by a single S-locus with multiple S-haplotypes. In GSI, the SI phenotype of pollen is determined by its own S-genotype. The growth of the pollen tube is inhibited at the style when its S-haplotype matches one of the S-haplotypes of the pistil. When two different cultivars have the same S-genotype, they exhibit cross-incompatibility. Therefore, S-genotype assignment is important to select compatible pollinizers in orchards and compatible parents in breeding programs. In contrast, self-compatible (SC) cultivars are useful because they save on artificial pollination labor costs and achieve stable fruit set. Therefore, an important objective of Japanese pear-breeding programs is to produce SC cultivars with high fruit quality. In Japanese pear, the first SC cultivar ‘Osanijisseiki’ was discovered as a bud mutation from ‘Nijisseiki’ (Furuta et al., 1980). ‘Osanijisseiki’ has been used as a parent for cross-breeding of SC cultivars in Japan and Korea, producing new SC cultivars such as ‘Akibae’, ‘Akikansen’, ‘Shinmizuki’, and ‘Shino’ (Kim et al., 2004; Matsumoto et al., 2014; Tanabe et al., 2001). However, because it is time-consuming to select SC trees by pollination experiments, there is a need to develop DNA markers to select SC trees at the seedling stage. This review focuses on GSI research on Japanese pear, especially the recently developed DNA markers for S-genotyping and marker-assisted selection of SC trees. In addition, candidate genes controlling pollen S specificity and a model for the molecular mechanism of GSI in Japanese pear are described.

1. Analyses of S-genotype

1)  Determination of S-genotypes by pollination experiments

In early pollination experiments on Japanese pear, Kikuchi (1929) found that cross combinations of some cultivars exhibited cross-incompatibility. In addition, Kikuchi (1929) found that approximately half of the first filial generation (F₁) plants in cross combinations of some cultivars exhibited cross-incompatibility with their pollen parent; this was named paterclinal incompatibility. On the basis of the patterns of cross-incompatibility and paterclinal incompatibility, Terami et al. (1946) identified seven S-haplotypes (S₁ to S₇) and 10 S-genotypes in 23 cultivars for the first time. Using these S-genotypes as first cross indicators, the S-genotypes of other cultivars were determined in pollination experiments; when a cultivar was cross-incompatible with the cross indicator, they had the same S-genotype (Hiratsuka et al., 1998; Machida et al., 1982; Ogaki, 1958).

The S-genotype of the SC cultivar ‘Osanijisseiki’ was assigned as follows: the cross-pollination of ‘Nijisseiki’ (S₂S₄) × ‘Osanijisseiki’ was cross-incompatible, whereas that of ‘Osanijisseki’ × ‘Nijisseiki’ (S₂S₄) was cross-compatible. These results suggested that ‘Osanijisseki’ was a stylar-part mutant (Sato, 1993). Most of the F₁ plants from ‘Kosui’ (S₄S₅) × ‘Osanijisseiki’ were self-incompatible, suggesting that the mutation was in the S₄ gene, and that the S-genotype of ‘Osanijisseiki’ was S₂S₄^sm (where sm indicates stylar-part mutant) (Sato, 1993).

In this way, conventional pollination experiments have been used to assign the S-genotypes of approximately 40 Japanese pear cultivars. However, pollination tests can be affected by environmental conditions, and are time-consuming and labor-intensive. Additionally, the cross indicators currently available do not cover the 21 S-genotypes consisting of combinations of seven S-haplotypes.

2)  Identification of pistil S gene

To identify the pistil S gene of Japanese pear, researchers attempted to identify S-haplotype-specific proteins (S-proteins) in the style. Two-dimensional gel electrophoresis analyses of proteins in style extracts identified seven S-proteins corresponding to the S₁- to S₇-haplotypes. These S-proteins were identified as S-RNases, that is, basic glycoproteins with ribonuclease activity (Ishimizu et al., 1996; Sassa et al., 1993). S-RNases are developmentally expressed in the style during flower maturation (Ishimizu et al., 1996), but are not expressed in the leaf, pollen, or germinating pollen (Sassa et al., 1993).

The stylar-part SC mutant ‘Osanijisseiki’ (S₂S₄^sm) was shown to produce S₂-RNase, but not S₄-RNase (Ishimizu et al., 1996). The cDNA encoding the S₂-RNase was cloned from the stylar cDNA library, but that encoding the S₄-RNase was not (Norioka et al., 1996). Finally, it was revealed that ‘Osanijisseiki’ lacked a region more than 4 kb long, including the full-length S₄-RNase sequence (Sassa et al., 1997). These findings suggested that the S-RNase is involved in pistil SI, but not pollen SI, in Japanese pear.

3)  Development of S-genotyping systems using the pistil S gene

The cDNAs encoding seven S-RNases (S₁- to S₇-RNases) were cloned from styles of Japanese pear (Ishimizu et al., 1998; Norioka et al., 1996). Alignment of the deduced amino acid sequences of these S-RNases revealed five conserved regions (C1, C2, C3, RC4, and C5) and one hypervariable (HV) region with an S-haplotype-specific sequence (Ishimizu et al., 1998; Ushijima et al., 1998) (Fig. 1A). The HV region was thought to interact with the pollen S gene product(s) to recognize self and non-self pollen (Matsuura et al., 2001). One intron located within the HV region also showed considerable diversification in terms of sequence and length among the S-RNases (Ishimizu et al., 1999).

Fig. 1

Analysis of S-RNase fragments using the CAPS systems in Japanese pear. (A) Structural features of S-RNase in Japanese pear. Boxes show exons and a bar shows an intron. Orange, blue, and red regions in boxes correspond to signal peptide (SP), conserved regions (C1, C2, C3, RC4, and C5), and hypervariable (HV) region, respectively. Arrows indicate primers “FTQQYQ”, “anti-IIWPNV”, and “anti-(I/T)IWPNV”. (B) S-RNase fragments amplified from genomic DNA of Japanese pear cultivars by PCR with primers “FTQQYQ” and “anti-IIWPNV” (Takasaki et al., 2004). (C) S-RNase fragments amplified from genomic DNA of Japanese pear cultivars by PCR with primers “FTQQYQ” and “anti-(I/T)IWPNV” (Takasaki et al., 2004). Fragments were subjected to 2% agarose gel electrophoresis. 1: ‘Imamuraaki’ (S₁S₆), 2: ‘Nijisseiki’ (S₂S₄), 3: ‘Hosui’ (S₃S₅), 4: ‘Okusankichi’ (S₅S₇), 5: ‘Ichiharawase’ (S₁S₈), 6: ‘Shinkou’ (S₄S₉).

On the basis of the variations in the HV region and the intron, the cleaved amplified polymorphic sequence (CAPS) (S₁ to S₇) system for discriminating S₁- to S₇-haplotypes was developed as a rapid method to identify the S-genotypes of Japanese pear cultivars (Ishimizu et al., 1999). In this system, S-RNase fragments including the intron are amplified from genomic DNA by polymerase chain reaction (PCR) with the consensus primers “FTQQYQ” (5'-TTTACGCAGCAATATCAG-3') and “anti-IIWPNV” [5'-AC(A/G)TTCGGCCAAATAATT-3'] (Fig. 1A). Electrophoresis of PCR products can discriminate the 1347 bp S₂-RNase fragment, but not the S₁-, S₃- to S₇-RNase fragments of ca. 350 bp (Fig. 1B, lanes 1–4). These ca. 350 bp fragments are subsequently discriminated by digestion with six S-haplotype-specific restriction endonucleases: SfcI (S₁-specific), PpuMI (S₃-, S₅-specific), NdeI (S₄-specific), AlwNI (S₅-specific), HincII (S₆-specific), and AccII (S₆-, S₇-specific). Castillo et al. (2001b) confirmed that the assignment of S-genotypes by the CAPS (S₁ to S₇) system coincided with that by pollination experiments. Thus, the CAPS (S₁ to S₇) system was shown to be highly reliable for identifying the S-genotypes of cultivars with S₁- to S₇-haplotypes. However, the S-genotypes of some cultivars could not be identified using this system. It was thought that these cultivars had other uncharacterized S-RNases.

Castillo et al. (2001b) amplified a new 436 bp S-RNase fragment from ‘Ichiharawase’ (Fig. 1B, lane 5), ‘Meigetsu’, and ‘Heiwa’ (‘Nijisseiki’ × ‘Ichiharawase’) using the CAPS (S₁ to S₇) system. The length of this fragment and the deduced amino acid sequence in the HV region differed from those of S₁- to S₇-RNases. The new S-RNase was designated as S₈-RNase, and the S-genotypes of ‘Ichiharawase’, ‘Meigetsu’, and ‘Heiwa’ were identified as S₁S₈, S₁S₈, and S₄S₈, respectively. Castillo et al. (2002) determined the complete sequence of the S₈-RNase, and selected NruI as an S₈-haplotype-specific restriction endonuclease. This established the CAPS (S₁ to S₈) system for discriminating S₁- to S₈-haplotypes.

In Japan, three main cultivars of Japanese pear accounted for 78% of the total growing area (11745 ha) in 2011: ‘Kosui’ (S₄S₅), ‘Hosui’ (S₃S₅), and ‘Nijisseiki’ (S₂S₄) and its bud mutation (Statistics Bureau of Japan, http://www.e-stat.go.jp/SG1/estat/List.do?lid=000001115710, February 16, 2014). The next most common cultivars were ‘Niitaka’, ‘Akizuki’ (S₃S₄), ‘Shinkou’, and ‘Nansui’, occupying 9.5%, 2.7%, 2.4%, and 2.1% of the total growing area in Japan, respectively. However, S-genotypes of ‘Niitaka’, ‘Shinkou’, and ‘Nansui’ had not been determined by pollination experiments or using a CAPS system [(S₁ to S₇) or (S₁ to S₈)] (Fig. 1B, lane 6). Sawamura et al. (2002a) amplified a new S-RNase fragment from ‘Shinkou’ and ‘Shinsei’ (‘Suisei’ × ‘Shinkou’) using the consensus primers “FTQQYQ” and “anti-F(I/R)(N/D)CP(H/R)”. The new S-RNase was designated as S₉-RNase, and the S-genotypes of ‘Shinkou’ and ‘Shinsei’ were identified as S₄S₉. Takasaki et al. (2004) isolated the full-length cDNA encoding the S₉-RNase from pistils of ‘Shinkou’ and ‘Shinsei’, and established the CAPS (S₁ to S₉) system for discriminating S₁- to S₉-haplotypes. In this system, the nine S-RNase fragments are amplified from genomic DNA by PCR using the primer pair “FTQQYQ” and “anti-(I/T)IWPNV” [a mix of the primer “anti-IIWPNV” and an S₉-RNase-specific primer “anti-TIWPNV” (5'-ACGTTTGGCCAAATAGTT-3')] (Fig. 1A): 1347 bp (S₂), 1307 bp (S₉), 436 bp (S₈), 367 bp (S₁), 376 bp (S₃), 368 bp (S₄), 376 bp (S₅), 347 bp (S₆), and 352 bp (S₇) (Fig. 1C). The S-RNase fragments with similar lengths are subsequently discriminated by digestion with S-haplotype-specific restriction endonucleases (Table 1). The ca. 1.3 kb fragments are discriminated by digestion with AflII (S₂-specific) and BstBI (S₉-specific). The 436 bp fragment is discriminated by digestion with NruI (S₈-specific). The ca. 350 bp fragments are discriminated by digestion with SfcI (S₁-specific), PpuMI (S₃-, S₅-specific), NdeI (S₄-specific), AlwNI (S₅-specific), HincII (S₆-specific), and AccII (S₆-, S₇-specific). Using this system, the S-genotypes of ‘Niitaka’, ‘Nansui’, ‘Nangetsu’, ‘Amanogawa’, and ‘Ishiiwase’ were identified as S₃S₉, S₄S₉, S₄S₉, S₁S₉, and S₃S₉, respectively (Okada et al., 2004; Takasaki et al., 2004).

Table 1

S-genotype assignments of Japanese pear cultivars using CAPS (S₁ to S₉) system.

Moreover, these CAPS systems were used to reconsider the S-genotypes of five cultivars for which there was some uncertainty. The S-genotypes of ‘Akaho’, ‘Tanzawa’, ‘Kimizukawase’, ‘Choju’, and ‘Kumoi’ were previously determined by pollination experiments as S₁S₂, S₃S₅/S₃S₄, S₂S₅, S₂S₅, and S₃S₄, respectively (Ogaki, 1958; Terami et al., 1946; Yasunobu et al., 1977). Using the CAPS systems, the S-genotypes of ‘Akaho’, ‘Tanzawa’, ‘Kimizukawase’, ‘Choju’, and ‘Kumoi’ were reassigned as S₃S₅, S₄S₅, S₁S₅, S₁S₅, and S₁S₃, respectively (Castillo et al., 2001b; Okada et al., 2004). The reassigned S-genotypes of ‘Akaho’, ‘Tanzawa’, and ‘Kumoi’ were confirmed by pollination experiments (Castillo et al., 2001b; Okada et al., 2004). The reassigned S-genotypes of ‘Kimizukawase’ and ‘Choju’ [‘Asahi’ (S₄S₅) × ‘Kimizukawase’] were consistent with those according to the pedigree: ‘Shinsui’ (S₄S₅) [‘Kikusui’ (S₂S₄) × ‘Kimizukawase’], ‘Hayatama’ (S₁S₂) [‘Gion’ (S₂S₄) × ‘Kimizukawase’], and ‘Aikansui’ (S₄S₅) [‘Choju’ × ‘Tama’ (S₄S₅)] (Hiratsuka et al., 1998; Kajiura and Sato, 1990).

According to the pedigree, recent commercial cultivars of Japanese pear (46 cultivars that occupied a growing area of ≥ 1 ha in 2011) are genetically derived from 10 local cultivars: ‘Chojuro’ (S₂S₃), ‘Shinkozo’ (unknown), ‘Doitsu’ (S₁S₂), ‘Taihaku’ (S₄S₅), ‘Wasekouzou’ (S₂S₅), ‘Nijisseiki’ (S₂S₄), ‘Akaho’ (S₃S₅), ‘Okusankichi’ (S₅S₇), ‘Imamuraaki’ (S₁S₆), and ‘Amanogawa’ (S₁S₉) (Kajiura and Sato, 1990; Sawamura et al., 2004, 2008). These local cultivars were used as the first parents in the Japanese pear-breeding program initiated in 1915. The S-genotype of ‘Shinkozo’ has not yet been identified, but ‘Shinkozo’ has only been used once as a parent, for ‘Kimizukawase’ (S₁S₅) [‘Shinkozo’ × ‘Doitsu’ (S₁S₂)] (Kajiura and Sato, 1990). Therefore, the S-genotypes of recent commercial cultivars in Japanese pear comprise various pairs of only eight S-haplotypes (S₁ to S₇, and S₉) derived from 10 ancestral local cultivars, and can be identified using the CAPS (S₁ to S₉) system.

To date, the CAPS systems have been used to identify the S-genotypes of 77 Japanese pear cultivars (Castillo et al., 2001a, b, 2002; Ishimizu et al., 1999; Okada et al., 2004; Sawamura et al., 2002b, 2008; Takasaki et al., 2004) (Table 2); 20 cultivars were identified as S₄S₅ and 9 cultivars as S₃S₄. This result indicates that there are many cross-incompatible combinations among Japanese pear cultivars. The S₄-, S₅-, and S₃-haplotypes are distributed in 43, 34, and 21 cultivars, respectively. One reason for the imbalance of S-haplotypes among cultivars is because ‘Nijisseiki’ (S₂S₄), ‘Kosui’ (S₄S₅), ‘Hosui’ (S₃S₅), and their progeny have often been used in Japanese pear-breeding programs (Sawamura et al., 2002b). Recent breeding programs have tended to use a few excellent cultivars as parents. Consequently, the new cultivars show a narrow range of S-genotypes. Therefore, S-genotype assignment is becoming more and more important for stable fruit production and efficient breeding.

Table 2

S-genotypes of Japanese pear cultivars identifed using the CAPS system.

As mentioned above, use of the CAPS systems is a rapid and reliable method for identifying the S-genotypes of recent commercial cultivars and new cultivars selected from their progeny. However, the S-genotypes identified using these systems should be confirmed by pollination experiments to increase the reliability of the results. Accurate S-genotypes determined using both CAPS systems and pollination experiments make it possible to select compatible pollinizers for orchards and compatible parents in breeding programs.

4)  S-genotype assignments of local cultivars

In Japanese pear, many local cultivars appear to have been selected from chance seedlings or bud mutations that arose during cultivation. To date, 1212 cultivars of Japanese pear have been reported (Kajiura and Sato, 1990). Although the recent commercial cultivars have only eight S-haplotypes (S₁ to S₇, and S₉), it is unclear whether this low number is because fewer genetic resources were used in the breeding program, or because there is a low diversity of S-haplotypes in Japanese pear. Kim et al. (2007) identified a new Sk-RNase in a local cultivar of Japanese pear, ‘Kinchaku’ (S₄S_k), which is an important breeding material for resistance to pear scab disease (Abe and Kotobuki, 1998). Okada et al. (2009) analyzed the S-genotypes of three local Japanese pear cultivars, ‘Senryo’, ‘Kuroki’, and ‘Hogyoku’, and identified the S_e-RNase of European pear in ‘Senryo’ (S₃S_e), the S₁₂-RNase of Chinese pear in ‘Kuroki’ (S₂S₁₂), and the S₃₀-RNase of Chinese pear and Sk-RNase in ‘Hogyoku’ (S_kS₃₀). These results revealed that Japanese pear has more S-haplotypes than the nine previously identified (S₁ to S₉). The number of S-haplotypes is expected to increase as more local cultivars are analyzed.

5)  Development of an S-genotyping system using a pollen S candidate

Although the details of the pollen S gene are described later in section 3, multiple pollen S candidates, S-locus F-box brothers (SFBB) genes, have been identified in Japanese pear. Kakui et al. (2007) and Sassa et al. (2007) isolated 10 SFBB^-γ genes (SFBB^1-γ to SFBB^9-γ, and SFBB^k-γ) from S₁- to S₉-, and S_k-haplotypes. On the basis of the sequence polymorphisms of the SFBB^-γ genes, Kakui et al. (2007) developed a CAPS/derived cleaved amplified polymorphic sequence (dCAPS) system for S-genotyping of Japanese pear harboring S₁- to S₉-, and S_k-haplotypes. In this system, SFBB^-γ fragments are amplified from genomic DNA by PCR with the specific primers “PpFBXf7” and “PpFBXr3”, and are subsequently discriminated by digestion with nine S-haplotype-specific restriction endonucleases: TaqI (S₁-specific), BbvCI (S₃-, S₅-specific), NspI (S₄-specific), AflII (S₅-specific), DdeI (S₆-specific), PsiI (S₇-specific), SmaI (S₈-specific), HaeIII (S₉-specific), and ApoI (S_k-specific). The SFBB^2-γ-specific fragment is detected at 156 bp using the dCAPS system, in which SFBB^-γ fragments are amplified with 1 bp mutated primer “GdCAPSS2f1-Rsa” and “PpFBXr11” and are subsequently digested with RsaI.

2. Development of a DNA marker specific for the S₄^sm-haplotype

The S₄^sm-haplotype derived from the SC cultivar ‘Osanijisseiki’ (S₂S₄^sm) has been used as a genetic resource to breed new SC cultivars with high fruit quality. When this cultivar was used as a parent for breeding SC cultivars, SC and self-incompatible trees segregated in most of the cross-progeny (Sato, 1993). The selection of SC trees harboring the S₄^sm-haplotype by pollination experiments takes at least 5 years because the trees must grow to the flowering stage. Therefore, to save time, the CAPS systems have been used for S-genotype assignment and the early selection of SC trees (Ishimizu et al., 1999; Kim et al., 2004). In the CAPS systems, the lack of an S-RNase fragment in the PCR analysis indicates the S₄^sm-haplotype because a sequence of > 4 kb including the full-length S₄-RNase is missing from this haplotype (Sassa et al., 1997). However, the absence of an S-RNase fragment may be due to experimental errors, or could be because the cultivar harbors novel S-RNase alleles that were not amplified by the consensus primers. Therefore, the development of an S₄^sm-haplotype-specific DNA marker was necessary to confirm the S₄^sm-haplotype.

To identify the deleted region in the S₄^sm-haplotype, Okada et al. (2008) prepared a bacterial artificial chromosome (BAC) library from an S₄-homozygote (S₄S₄) in Japanese pear, and constructed a 570 kb BAC contig around the S₄-RNase. Using primers designed from the contig sequence, the deleted region in the S₄^sm-haplotype was analyzed by PCR using the genomic DNA from S₄- and S₄^sm-homozygotes (S₄^smS₄^sm) as the template, and by DNA sequencing of the BAC contig. These results revealed that the S₄^sm-haplotype lacks a 236 kb sequence spanning from 48 kb upstream to 188 kb downstream of the S₄-RNase (Fig. 2A). To develop a DNA marker specific for the S₄^sm-haplotype, Okada et al. (2008) designed a new primer pair, “SM-F” and “SM-R”, from the regions outside both of the deletion breakpoints in the S₄^sm-haplotype (Fig. 2A). PCR using “SM-F” and “SM-R” primers amplified a 666 bp fragment from genomic DNA of the S₄^sm-homozygote, but not from that of the S₄-homozygote or six SI cultivars harboring nine S-haplotypes (S₁ to S₉) (Fig. 2B). This result indicated that the 666 bp fragment can be used as an S₄^sm-haplotype-specific marker. Its reliability was confirmed by the selection of SC trees from the cross seedlings of ‘Osanijisseiki’ (S₂S₄^sm) × ‘Nansui’ (S₄S₉) (Okada et al., 2008).

Fig. 2

DNA marker specific for S₄^sm-haplotype in Japanese pear. (A) Schematic genomic structures around S₄-RNase in S₄- and S₄^sm-haplotypes. Arrowhead indicates S₄-RNase. S₄^sm-haplotype lacks 236 kb spanning from 48 kb upstream to 188 kb downstream of S₄-RNase. Arrows indicate S₄^sm-haplotype-specific primers “SM-F” and “SM-R”. (B) S₄^sm-haplotype-specific DNA marker amplified from genomic DNA of Japanese pear by PCR with primers “SM-F” and “SM-R” (Okada et al., 2008). Fragments were subjected to 2% agarose gel electrophoresis. SC and SI indicate self-compatibility and self-incompatibility, respectively. 1: S₄^sm-homozygote, 2: S₄-homozygote, 3: ‘Imamuraaki’ (S₁S₆), 4: ‘Nijisseiki’ (S₂S₄), 5: ‘Hosui’ (S₃S₅), 6: ‘Okusankichi’ (S₅S₇), 7: ‘Ichiharawase’ (S₁S₈), 8: ‘Shinkou’ (S₄S₉).

However, trees with the S₄^sm-haplotype are not always SC. The S₄^sm pollen was accepted by the pistils of ‘Osanijisseiki’ (S₂S₄^sm), but rejected by those of ‘Nijisseiki’ (S₂S₄) and ‘Doitsu’ (S₁S₂). This observation suggested that the S₄^sm pollen is rejected by pistils harboring S₄- or S₁-haplotypes (Saito et al., 2012; Sato, 1993). (The details of this phenomenon are described later in section 3.) That is, S₁S₄^sm and S₄S₄^sm trees exhibit SI. To select SC trees efficiently, therefore, S-genotypes of cross seedlings must first be predicted based on the S-genotypes of both parents assigned using the CAPS system, and by the S₄^sm-haplotype-specific marker. If the predicted S-genotypes are not S₁S₄^sm or S₄S₄^sm, SC trees can be selected from the cross seedlings using only the S₄^sm-haplotype-specific marker. If the predicted S-genotypes contain S₁S₄^sm or S₄S₄^sm, SC trees can be selected by eliminating seedlings assigned to S₁- or S₄-haplotypes using the CAPS system from those identified using the S₄^sm-haplotype-specific marker. The combination of the S₄^sm-haplotype-specific marker and the CAPS system provides an early and reliable system to select SC Japanese pear, and will facilitate the breeding of SC cultivars with the S₄^sm-haplotype (Okada et al., 2008).

3. Analysis of the pollen S gene and model for the molecular mechanism of GSI

The GSI of Japanese pear is controlled by two genetically linked genes: the pistil S gene and the pollen S gene at the S-locus. As mentioned above, the pistil S gene of Japanese pear is the S-RNase. In Japanese pear, Liu et al. (2007) showed that self S-RNases inhibited pollen tube growth in vitro, but non-self S-RNase did not. In the Solanaceae family, which also exhibits S-RNase-based GSI, both self and non-self S-RNases enter pollen tubes growing in the style (Luu et al., 2000). In the Solanaceae, the RNase activity of S-RNases is essential for the rejection of self pollen, and pollen rRNAs are degraded after self-pollination (Huang et al., 1994; McClure et al., 1990). On the basis of these results in the Solanaceae, it has been speculated that the self S-RNase of Japanese pear inhibits the growth of the self pollen tubes in the style by degrading pollen rRNAs, but the RNase activity of all non-self S-RNases is inhibited by the pollen S gene product. Additionally, recent studies have shown that, in vitro, the S-RNase of Japanese pear induces alterations in mitochondria (membrane potential collapse, cytochrome c leakage, and swelling), disruption of tip-localized reactive oxygen species (ROS; which are required for tip growth), depolymerization of the actin cytoskeleton, and degradation of the nuclear DNA of incompatible pollen tubes. These findings suggested that programmed cell death (PCD) may specifically occur in incompatible pollen tubes (Liu et al., 2007; Wang et al., 2009, 2010).

In contrast to the pistil S gene (S-RNase), the pollen S gene of Japanese pear remained unknown for a long time. To identify the pollen S gene in apple (Malus × domestica Borkh.) and Japanese pear, Sassa et al. (2007) analyzed the S-locus in apple, and identified two homologous F-box genes from the S₉-haplotype and two from the S₃-haplotype. These multiple and related F-box genes were named SFBB of M. domestica: MdSFBB^9-α and MdSFBB^9-β from the S₉-haplotype, and MdSFBB^3-α and MdSFBB^3-β from the S₃-haplotype. Using primers designed from the MdSFBB sequences, six cDNAs were isolated by RT-PCR from pollen of the Japanese pear cultivar ‘Kosui’ (S₄S₅): PpSFBB^4-α, PpSFBB^4-β, and PpSFBB^4-γ from the S₄-haplotype, and PpSFBB^5-α, PpSFBB^5-β, and PpSFBB^5-γ from the S₅-haplotype. Because these SFBBs showed linkage to the S-RNase, S-haplotype-specific sequence divergence, and pollen-specific expression, they seemed to be a good candidate for the pollen S gene in apple and Japanese pear (Sassa et al., 2007). However, a later linkage analysis of PpSFBB^4-α and S₄-RNase in Japanese pear detected two recombinants among 59 segregants, suggesting that PpSFBB^4-α (PpSFBB6-S₄’) was not involved in pollen S specificity (Kakui et al., 2011). Generally, F-box proteins function as one of the four major subunits (CUL1, SKP1, RBX1/ROC1, and an F-box protein) of the SCF complex that catalyzes the attachment of polyubiquitin chains to target proteins for their subsequent degradation by the 26S proteasome (Lechner et al., 2006). Therefore, SFBBs are speculated to mediate ubiquitination and degradation of all non-self S-RNases through the ubiquitin/26S proteasome system.

For further exploration of the pollen S gene, Okada et al. (2008, 2011) constructed two BAC contigs around the S₄- and S₂-RNases of Japanese pear. A 649 kb region around the S₄-RNase and a 378 kb region around the S₂-RNase were sequenced, based on the prediction that the pollen S gene was tightly linked to the S-RNase. As a result, six new pollen-specific F-box genes [PpSFBB^4-u1, PpSFBB^4-u2, PpSFBB^4-u3, PpSFBB^4-u4, PpSFBB^4-d1 (formerly named S₄F-box0; Okada et al., 2008), and PpSFBB^4-d2] were identified around the S₄-RNase, and 10 new pollen-specific F-box genes (PpSFBB^2-u1, PpSFBB^2-u2, PpSFBB^2-u3, PpSFBB^2-u4, PpSFBB^2-u5, PpSFBB^2-d1, PpSFBB^2-d2, PpSFBB^2-d3, PpSFBB^2-d4, and PpSFBB^2-d5) were found around the S₂-RNase (Fig. 3). PpSFBB^4-α, PpSFBB^4-β, and PpSFBB^4-γ were not within the sequenced region, and were located much further away from the S₄-RNase. Moreover, Kakui et al. (2007, 2011) isolated 31 new PpSFBBs from different S-haplotypes of Japanese pear by PCR or RT-PCR. These studies on the pollen S gene (Kakui et al., 2007, 2011; Okada et al., 2008, 2011; Sassa et al., 2007) revealed that multiple PpSFBBs are clustered around the S-RNase of Japanese pear: 7 in the S₁-haplotype, 11 in the S₂-haplotype, 8 in the S₃-haplotype, 10 in the S₄-haplotype, 5 in the S₅-haplotype, and 7 in the S₆-haplotype. Interestingly, the deduced amino acid sequence identities of some PpSFBBs were higher between different S-haplotypes than within each S-haplotype (Okada et al., 2011). For example, there were high (> 90%) pairwise deduced amino acid sequence identities between nine PpSFBB⁴ genes (PpSFBB^4-u1–u4, PpSFBB^4-d1–d2, and PpSFBB^4-α–γ) and 11 PpSFBB² genes (PpSFBB^2-u1–u5, PpSFBB^2-d1–d5, and PpSFBB^2-γ); 92.1% identity between PpSFBB^4-u2 and PpSFBB^2-u1, 93.1% identity between PpSFBB^4-u3 and PpSFBB^2-u2, 94.9% identity between PpSFBB^4-β and PpSFBB^2-u4, and 99.0% identity between PpSFBB^4-γ and PpSFBB^2-γ. Conversely, the deduced amino acid sequence identities ranged from 62.3% to 86.2% among the nine PpSFBB⁴ genes, and from 63.0% to 86.0% among the 11 PpSFBB² genes. In a phylogenetic tree, 42 PpSFBBs isolated from S₁- to S₆-haplotypes clustered into eight types (SFBB1–SFBB8), and each type contained similar PpSFBBs isolated from different S-haplotypes (Kakui et al., 2011).

Fig. 3

Schematic genomic structures and locations of PpSFBBs around S₄-RNase and S₂-RNase in Japanese pear. Arrows indicate directions of transcription of S-RNases and PpSFBBs.

Similarly, 16–20 S-locus F-box (SLF) genes were identified in each S-haplotype in Petunia, a member of the Solanaceae family that exhibits the same S-RNase-based GSI as Japanese pear (Hua et al., 2007; Kubo et al., 2010, 2015; McCubbin et al., 2000; Wang et al., 2003, 2004; Williams et al., 2014a). These SLFs were classified into 18 types (SLF1–SLF18) based on their phylogenetic relationships (Kubo et al., 2015). Transformation experiments showed that at least eight of the SLFs (SLF1–SLF6, SLF8, and SLF9) were involved in pollen specificity (Kubo et al., 2010, 2015; Sijacic et al., 2004; Williams et al., 2014b). On the basis of these findings, a collaborative non-self recognition model was proposed. In this model, the pollen S gene in Petunia comprises multiple types of F-box genes. Each type of F-box protein recognizes and degrades only some of the non-self S-RNases, and multiple types of F-box proteins collaboratively inactivate all of the non-self S-RNases (Kubo et al., 2010).

As pollen S candidates, PpSFBBs in Japanese pear show characteristics that fit well with the collaborative non-self recognition model reported for Petunia, suggesting that this system may also exist in Japanese pear. This idea was supported by the results of a recent study on Japanese pear. For a long time, it was believed that S₄ pollen and S₄^sm pollen in Japanese pear have the same function in pollen S specificity. Interestingly, however, pollination experiments showed that S₄^sm pollen was rejected by the pistils harboring not only the S₄- but also the S₁-haplotype, although S₄ pollen was accepted by the pistils harboring the S₁-haplotype (Saito et al., 2012). Moreover, the progeny of ‘Kimizukawase’ (S₁S₅) × ‘Osanijisseiki’ (S₂S₄^sm) did not inherit the S₄^sm-haplotype, but the progeny of ‘Hosui’ (S₃S₅) × ‘Osanijisseiki’ (S₂S₄^sm) did, indicating that S₄^sm pollen was accepted by the pistils harboring the S₃- and S₅-haplotypes but rejected by those harboring the S₁-haplotype. On the other hand, the progeny of ‘Kimizukawase’ (S₁S₅) × ‘Nijisseiki’ (S₂S₄) inherited the S₄-haplotype, showing that S₄ pollen was accepted by the pistils harboring the S₁-haplotype (Kakui et al., 2011). These results suggested that the S₄^sm-haplotype also has a mutation in the pollen S₄ gene. As the S₄^sm-haplotype lacks not only the S₄-RNase but also PpSFBB^4-d1 (S₄F-box0) (Okada et al., 2008) (Fig. 3), the collaborative non-self recognition model would predict that PpSFBB^4-d1 is an element of the pollen S gene in Japanese pear that specifically recognizes the S₁-RNase (Kubo et al., 2010) (Fig. 4).

Fig. 4

Collaborative non-self recognition model by multiple PpSFBBs for GSI in Japanese pear. Red ellipse indicates S-RNase as pistil S gene product and blue box shows multiple PpSFBBs as pollen S gene product. For simplicity, it is assumed that PpSFBB^4-x, PpSFBB^4-y, PpSFBB^4-z, and PpSFBB^4-d1 encoded in the S₄- or S₄^sm-haplotype recognize and degrade non-self S_x-, S_y-, S_z-, and S₁-RNase, respectively (S_x-, S_y-, S_z-RNase ≠ S₁-RNase and self S₄-RNase). (A) In the S₄ pollen tube, S_x- and S₁-RNase entering the pollen tube are recognized and degraded by PpSFBB^4-x and PpSFBB^4-d1, respectively. As a result, rRNAs in the pollen tube remain intact, and the pollen tube can grow through the style. (B) In the S₄^sm pollen tube, S_x-RNase entering the pollen tube is recognized and degraded by PpSFBB^4-x, but S₁-RNase entering the pollen tube is not recognized and degraded because S₄^sm pollen lacks PpSFBB^4-d1. Consequently, the S₁-RNase can degrade the rRNAs in the pollen tube, and pollen tube growth is arrested.

Very recently, another SC mutant 415-1 was produced by crossing ‘Kosui’ (S₄S₅) with pollen from gamma-irradiated ‘Kosui’. The S-genotype of 415-1 was identified as S₄S₅ using the CAPS system. Cross-pollinations of 415-1 (S₄S₅) × ‘Shugyoku’ (S₄S₅) and 415-1 (S₄S₅) × ‘Oushuu’ (S₄S₅) were cross-incompatible, whereas cross-pollinations of ‘Shugyoku’ (S₄S₅) × 415-1 (S₄S₅) and ‘Oushuu’ (S₄S₅) × 415-1 (S₄S₅) were cross-compatible. These results suggested that 415-1 is a pollen-part mutant (Sawamura et al., 2013). A ploidy analysis, and the segregation in the progeny of S-haplotypes and simple sequence repeat (SSR) markers around the S-locus, indicated that the pollen-part self-compatibility of 415-1 was not caused by a loss of function of pollen S genes in the S₄- or S₅-haplotype, but by segmental duplication of the chromosome containing the S₅-haplotype. Consequently, 415-1 was able to produce S₄S₅ pollen with two different S-haplotypes (S-heteroallelic pollen). This pollen is capable of breaking down SI by competitive interaction between the two different S-haplotypes in a single pollen grain (Mase et al., 2014). Because the pollen of a pollen-part SC mutant would be cross-compatible with cultivars of any S-genotype, 415-1 can serve as additional useful breeding material to produce new SC cultivars with high fruit quality.

Matsumoto et al. (2006) reported that two tetraploids arising as a bud mutation from ‘Shinkou’ (S₄S₉) exhibited self-compatibility. The cross-pollination of the tetraploid × ‘Shinkou’ (S₄S₉) was cross-incompatible, whereas that of ‘Shinkou’ (S₄S₉) × the tetraploid was cross-compatible. These results suggested that the tetraploids are also pollen-part mutants.

According to the collaborative non-self recognition model, mutations that disrupt the function of the pollen S gene, such as the deletion of PpSFBB^4-d1 in the S₄^sm-haplotype (Okada et al., 2008), cause pollen to be incompatible with more S-haplotypes because decreasing the repertoire of the pollen S gene eliminates the ability to recognize and degrade its target S-RNase(s). Conversely, mutations to increase the repertoire of the pollen S gene, such as segmental duplication of the chromosome containing the S-locus in 415-1 (Mase et al., 2014) or tetraploidization (Matsumoto et al., 2006), allow pollen to recognize and degrade more S-RNases. Therefore, inducing mutations to increase the repertoire of the pollen S gene would be one effective method to create new pollen-part SC mutants of Japanese pear.

4. Conclusions

Japanese pear is an important fruit tree with a long history of cultivation in Japan. Since almost all cultivars of Japanese pear exhibit SI, S-genotype assignment is essential for stable fruit production and efficient breeding. Studies on the pistil S gene have led to the establishment of the CAPS system for S-genotyping based on sequence polymorphisms of the S-RNase allele (Castillo et al., 2002; Ishimizu et al., 1999; Takasaki et al., 2004). Similarly, studies on the pollen S gene have led to the development of the CAPS/dCAPS system for S-genotyping based on sequence polymorphisms of the PpSFBB^–γ allele (Kakui et al., 2007). The use of the CAPS systems has allowed more S-genotypes to be determined quickly and accurately. Consequently, the S-genotypes of the main commercial cultivars and important breeding materials have been revealed. This information has already been used to select compatible pollinizers for orchards and compatible parents in breeding programs. In addition, analyses of the SC mutant ‘Osanijisseiki’ have led to the development of a DNA marker specific for the S₄^sm-haplotype (Okada et al., 2008). This marker has allowed the early and accurate selection of SC trees, and it has already been used in breeding programs to develop new SC cultivars with high fruit quality. Thus, studies on the GSI of Japanese pear have contributed to the development of new techniques to identify S-genotypes and to select SC trees efficiently.

In addition to its importance as a fruit tree, Japanese pear is an important model plant for elucidating the molecular mechanism of S-RNase-based GSI in the Rosaceae. The recent discovery of PpSFBBs as a candidate for the pollen S gene suggested that Japanese pear may use the collaborative non-self recognition system that has been identified in Petunia of the Solanaceae. On the other hand, surprisingly, recent studies have reported that the features of the GSI differ between the tribe Pyreae (Pyrus and Malus) and the tribe Amygdaleae (Prunus) in the Rosaceae (for reviews, see De Franceschi et al., 2012; Meng et al., 2011; Sassa et al., 2010; Yamane and Tao, 2009). These findings suggested that the role of the pollen S gene may differ between the tribe Pyreae (Pyrus and Malus) and the tribe Amygdaleae (Prunus) in the Rosaceae, although Prunus also has an S-RNase as the pistil S gene and an F-box gene [called SFB (S-haplotype-specific F-box) or SLF (S-locus F-box)] as the pollen S gene (Entani et al., 2003; Ushijima et al., 2003). In Japanese pear, however, pollen S candidates have only been found recently, and further experiments are required to confirm that multiple PpSFBBs are the pollen S gene in the GSI. Future research on PpSFBBs should determine how many PpSFBBs exist in each S-haplotype; which PpSFBB recognizes each S-RNase; how PpSFBBs degrade non-self S-RNases; and how a single S-RNase and multiple PpSFBBs coevolved to maintain GSI. Further research is also required to determine the mechanism of self pollen tube growth arrest. Resolving these issues will provide a better understanding of the mechanism and the evolution of GSI, and will contribute to the development of new techniques for stable fruit production and efficient breeding of Japanese pear.

Acknowledgements

The studies of the author described in this review were conducted with Dr. Takeshi Takasaki-Yasuda, Dr. Tetsu Nakanishi, Dr. Carlos Castillo, Dr. Yuki Moriya-Tanaka, and members of the Laboratory of Pomology at Kobe University; Toshihiro Saito and Dr. Yutaka Sawamura at NARO Institute of Fruit Tree Science; Dr. Naoko Norioka and Dr. Shigemi Norioka at Osaka University; and Dr. Tatsuya Matsumoto at Niigata Agricultural Research Institute. The author greatly appreciates their instruction and assistance.

Literature Cited

•  Abe,  K. and  K.  Kotobuki. 1998. Inheritance of high resistance to Venturia nashicola Tanaka et Yamamoto in Japanese pear (Pyrus pyrifolia Nakai) and Chinese pear (P. ussuriensis Maxim.). J. Japan. Soc. Hort. Sci. 67: 677–680.
•  Castillo,  C.,  T.  Takasaki,  T.  Saito,  S.  Norioka and  T.  Nakanishi. 2001a. Characterization of S₉-RNase (S₉-allele) in Japanese pear cultivars, ‘Amanogawa’, ‘Niitaka’, ‘Shinko’, ‘Shinsei’, ‘Echigo’, ‘Nansui’ and ‘Nangetsu’. J. Japan. Soc. Hort. Sci. 70 (Suppl. 2): 134 (In Japanese).
•  Castillo,  C.,  T.  Takasaki,  T.  Saito,  S.  Norioka and  T.  Nakanishi. 2002. Cloning of the S₈-RNase (S₈-allele) of Japanese pear (Pyrus pyriforia Nakai). Plant Biotechnol. 19: 1–6.
•  Castillo,  C.,  T.  Takasaki,  T.  Saito,  Y.  Yoshimura,  S.  Norioka and  T.  Nakanishi. 2001b. Reconsideration of S-genotype assignments, and discovery of a new allele based on S-RNase PCR-RFLPs in Japanese pear cultivars. Breed. Sci. 51: 5–11.
•  De Franceschi,  P.,  L.  Dondini and  J.  Sanzol. 2012. Molecular bases and evolutionary dynamics of self-incompatibility in the Pyrinae (Rosaceae). J. Exp. Bot. 63: 4015–4032.
•  Entani,  T.,  M.  Iwano,  H.  Shiba,  F. S.  Che,  A.  Isogai and  S.  Takayama. 2003. Comparative analysis of the self-incompatibility (S-) locus region of Prunus mume: identification of a pollen-expressed F-box gene with allelic diversity. Genes Cells 8: 203–213.
•  Furuta,  O.,  T.  Imai,  T.  Miyoshi,  N.  Yatsumi,  S.  Ueki,  S.  Hayashi and  S.  Hiragi. 1980. Properties of Japanese pear ‘Osanijisseiki’. Abstr. Japan. Soc. Hort. Sci. Autumn Meet.: 70–71 (In Japanese).
•  Hua,  Z. H.,  X. Y.  Meng and  T. H.  Kao. 2007. Comparison of Petunia inflata S-locus F-box protein (Pi SLF) with Pi SLF-like proteins reveals its unique function in S-RNase-based self-incompatibility. Plant Cell 19: 3593–3609.
•  Huang,  S.,  H. S.  Lee,  B.  Karunanandaa and  T. H.  Kao. 1994. Ribonuclease activity of Petunia inflata S proteins is essential for rejection of self-pollen. Plant Cell 6: 1021–1028.
•  Hiratsuka,  S.,  T.  Kubo and  Y.  Okada. 1998. Estimation of self-incompatibility genotype in Japanese pear cultivars by stylar protein analysis. J. Japan. Soc. Hort. Sci. 67: 491–496.
•  Ishimizu,  T.,  K.  Inoue,  M.  Shimonaka,  T.  Saito,  O.  Terai and  S.  Norioka. 1999. PCR-based method for identifying the S-genotypes of Japanese pear cultivars. Theor. Appl. Genet. 98: 961–967.
•  Ishimizu,  T.,  Y.  Sato,  T.  Saito,  Y.  Yoshimura,  S.  Norioka,  T.  Nakanishi and  F.  Sakiyama. 1996. Identification and partial amino acid sequences of seven S-RNases associated with self-incompatibility of Japanese pear, Pyrus pyrifolia Nakai. J. Biochem. 120: 326–334.
•  Ishimizu,  T.,  T.  Shinkawa,  F.  Sakiyama and  S.  Norioka. 1998. Primary structural features of rosaceous S-RNases associated with gametophytic self-incompatibility. Plant Mol. Biol. 37: 931–941.
• Kajiura, I. and Y. Sato. 1990. Recent progress in Japanese pear (Pyrus pyrifolia Nakai) breeding, and descriptions of cultivars based on literature review. Bull. Fruit Tree Res. Stn. Extra No.1 (In Japanese with English summary).
•  Kakui,  H.,  M.  Kato,  K.  Ushijima,  M.  Kitaguchi,  S.  Kato and  H.  Sassa. 2011. Sequence divergence and loss-of-function phenotypes of S locus F-box brothers genes are consistent with non-self recognition by multiple pollen determinants in self-incompatibility of Japanese pear (Pyrus pyrifolia). Plant J. 68: 1028–1038.
•  Kakui,  H.,  T.  Tsuzuki,  T.  Koba and  H.  Sassa. 2007. Polymorphism of SFBB-gamma and its use for S genotyping in Japanese pear (Pyrus pyrifolia). Plant Cell Rep. 26: 1619–1625.
•  Kao,  T. H. and  A. G.  McCubbin. 1996. How flowering plants discriminate between self and non-self pollen to prevent inbreeding. Proc. Natl. Acad. Sci. USA 93: 12059–12065.
•  Kikuchi,  A. 1929. Investigations in 1927 and 1928. 1. Paterclinal incompatibility in the Japanese pear. Okitsu Hort. Soc. 24: 1–6 (In Japanese).
•  Kim,  H. T.,  Y.  Hirata,  Y. U.  Shin,  H. S.  Hwang,  J. H.  Hwang,  I. S.  Shin,  D. I.  Kim,  S. J.  Kang,  H. J.  Kim,  D. Y.  Shin and  I. S.  Nou. 2004. A molecular technique for selection of self-compatible varieties of Japanese pear (Pyrus pyrifolia Nakai). Euphytica 138: 73–80.
•  Kim,  H.,  H.  Kakui,  T.  Koba,  Y.  Hirata and  H.  Sassa. 2007. Cloning of a new S-RNase and development of a PCR-RFLP system for the determination of the S-genotypes of Japanese pear. Breed. Sci. 57: 159–164.
•  Kubo,  K.,  T.  Entani,  A.  Takara,  N.  Wang,  A. M.  Fields,  Z. H.  Hua,  M.  Toyoda,  S.  Kawashima,  T.  Ando,  A.  Isogai,  T.-h.  Kao and  S.  Takayama. 2010. Collaborative non-self recognition system in S-RNase-based self-incompatibility. Science 330: 796–799.
•  Kubo,  K.,  T.  Paape,  M.  Hatakeyama,  T.  Entani,  A.  Takara,  K.  Kajihara,  M.  Tsukahara,  R.  Shimizu-Inatsugi,  K. K.  Shimizu and  S.  Takayama. 2015. Gene duplication and genetic exchange drive the evolution of S-RNase-based self-incompatibility in Petunia. Nature Plants 1: 14005. DOI: 10.1038/nplants.2014.5. <http://www.nature.com/articles/nplants20145>.
•  Lechner,  E.,  P.  Achard,  A.  Vansiri,  T.  Potuschak and  P.  Genschik. 2006. F-box proteins everywhere. Curr. Opin. Plant Biol. 9: 631–638.
•  Liu,  Z. Q.,  G. H.  Xu and  S. L.  Zhang. 2007. Pyrus pyrifolia stylar S-RNase induces alterations in the actin cytoskeleton in self-pollen and tubes in vitro. Protoplasma 232: 61–67.
•  Luu,  D. T.,  X.  Qin,  D.  Morse and  M.  Cappadocia. 2000. S-RNase uptake by compatible pollen tubes in gametophytic self-incompatibility. Nature 407: 649–651.
•  Machida,  Y.,  Y.  Sato,  I.  Kozaki and  K.  Seike. 1982. S-genotypes of several cultivars of Japanese pear and the question of the parents of ‘Hosui’. Abstr. Japan. Soc. Hort. Sci. Autumn Meet.: 58–59 (In Japanese).
•  Mase,  N.,  Y.  Sawamura,  T.  Yamamoto,  N.  Takada,  S.  Nishio,  T.  Saito and  H.  Iketani. 2014. A segmental duplication encompassing S-haplotype triggers pollen-part self-compatibility in Japanese pear (Pyrus pyrifolia). Mol. Breed. 33: 117–128.
•  Matsumoto,  T.,  H.  Kameyama and  K.  Kojima. 2006. Self-compatibility of bud mutation strain in Japanese pear ‘Shinko’. J. Japan. Soc. Hort. Sci. 75 (Suppl. 2): 147 (In Japanese).
•  Matsumoto,  T.,  T.  Nomizu and  K.  Nedu. 2014. New self-compatible Japanese pear cultivars ‘Shinmizuki’ and ‘Shino’. Hort. Res. (Japan) 13 (Suppl. 1): 45 (In Japanese).
•  Matsuura,  T.,  H.  Sakai,  M.  Unno,  K.  Ida,  M.  Sato,  F.  Sakiyama and  S.  Norioka. 2001. Crystal structure at 1.5-angstrom resolution of Pyrus pyrifolia pistil ribonuclease responsible for gametophytic self-incompatibility. J. Biol. Chem. 276: 45261–45269.
•  McClure,  B. A.,  J. E.  Gray,  M. A.  Anderson and  A. E.  Clarke. 1990. Self-incompatibility in Nicotiana alata involves degradation of pollen rRNA. Nature 347: 757–760.
•  McCubbin,  A. G.,  X.  Wang and  T. H.  Kao. 2000. Identification of self-incompatibility (S-) locus linked pollen cDNA markers in Petunia inflata. Genome 43: 619–627.
•  Meng,  X.,  P.  Sun and  T.-h.  Kao. 2011. S-RNase-based self-incompatibility in Petunia inflata. Ann. Bot. 108: 637–646.
•  Norioka,  N.,  S.  Norioka,  Y.  Ohnishi,  T.  Ishimizu,  C.  Oneyama,  T.  Nakanishi and  F.  Sakiyama. 1996. Molecular cloning and nucleotide sequences of cDNAs encoding S-allele specific stylar RNases in a self-incompatible cultivar and its self-compatible mutant of Japanese pear, Pyrus pyrifolia Nakai. J. Biochem. 120: 335–345.
•  Ogaki,  C. 1958. The sterility-factors in new varieties of the Japanese pear (Pyrus serotina Rehd. var. culta Rehd.). Res. Rep. Kanagawa Agr. Sta. Hort. Inst. 5: 23–26 (In Japanese with English abstract).
•  Okada,  K.,  C.  Castillo,  Y.  Sawamura,  T.  Nakanishi and  T.  Takasaki-Yasuda. 2009. S-genotype assignments of local cultivars in Japanese pear ‘Senryo’, ‘Kuroki’, and ‘Hogyoku’. J. Japan. Soc. Hort. Sci. 78: 55–60.
•  Okada,  K.,  T.  Takasaki,  T.  Saito,  Y.  Moriya,  C.  Castillo,  S.  Norioka and  T.  Nakanishi. 2004. Reconsideration of S-genotype for a Japanese pear ‘Kumoi’. J. Japan. Soc. Hort. Sci. 73: 524–528.
•  Okada,  K.,  N.  Tonaka,  Y.  Moriya,  N.  Norioka,  Y.  Sawamura,  T.  Matsumoto,  T.  Nakanishi and  T.  Takasaki-Yasuda. 2008. Deletion of a 236 kb region around S₄-RNase in a stylar-part mutant S₄^sm-haplotype of Japanese pear. Plant Mol. Biol. 66: 389–400.
•  Okada,  K.,  N.  Tonaka,  T.  Taguchi,  T.  Ichikawa,  Y.  Sawamura,  T.  Nakanishi and  T.  Takasaki-Yasuda. 2011. Related polymorphic F-box protein genes between haplotypes clustering in the BAC contig sequences around the S-RNase of Japanese pear. J. Exp. Bot. 62: 1887–1902.
•  Saito,  T.,  Y.  Sato,  Y.  Sawamura,  M.  Shoda,  T.  Takasaki-Yasuda and  K.  Kotobuki. 2012. Dual recognition of S₁ and S₄ pistils by S₄^sm pollen in self-incompatibility of Japanese pear (Pyrus pyrifolia Nakai). Tree Genet. Genomes 8: 689–694.
•  Sassa,  H.,  H.  Hirano and  H.  Ikehashi. 1993. Identification and characterization of stylar glycoproteins associated with self-incompatibility genes of Japanese pear, Pyrus serotina Rehd. Mol. Gen. Genet. 241: 17–25.
•  Sassa,  H.,  H.  Hirano,  T.  Nishio and  T.  Koba. 1997. Style-specific self-compatible mutation caused by deletion of the S-RNase gene in Japanese pear (Pyrus serotina). Plant J. 12: 223–227.
•  Sassa,  H.,  H.  Kakui and  M.  Minamikawa. 2010. Pollen-expressed F-box gene family and mechanism of S-RNase-based gametophytic self-incompatibility (GSI) in Rosaceae. Sex. Plant Reprod. 23: 39–43.
•  Sassa,  H.,  H.  Kakui,  M.  Miyamoto,  Y.  Suzuki,  T.  Hanada,  K.  Ushijima,  M.  Kusaba,  H.  Hirano and  T.  Koba. 2007. S locus F-box brothers: Multiple and pollen-specific F-box genes with S haplotype-specific polymorphisms in apple and Japanese pear. Genetics 175: 1869–1881.
• Sato, Y. 1993. Breeding of self-compatible Japanese pear. p. 241–247. In: T. Hayashi, M. Omura and N. S. Scott (eds.). Techniques on gene diagnosis and breeding in fruit trees. Fruit Tree Research Station, Tsukuba.
•  Sawamura,  Y.,  N.  Mase,  N.  Takada,  A.  Sato,  C.  Nishitani,  K.  Abe,  T.  Masuda,  T.  Yamamoto,  T.  Saito and  K.  Kotobuki. 2013. A self-compatible pollen-part mutant of Japanese pear produced by crossing ‘Kosui’ with pollen from gamma-irradiated ‘Kosui’. J. Japan. Soc. Hort. Sci. 82: 222–226.
•  Sawamura,  Y.,  T.  Saito,  M.  Shoda,  T.  Yamamoto,  Y.  Sato,  T.  Hayashi and  K.  Kotobuki. 2002a. A new self-incompatible allele in Japanese pear ‘Shinsei’ and ‘Shinkou’. J. Japan. Soc. Hort. Sci. 71: 342–347.
•  Sawamura,  Y.,  T.  Saito,  N.  Takada,  T.  Yamamoto,  T.  Kimura,  T.  Hayashi and  K.  Kotobuki. 2004. Identification of parentage of Japanese pear ‘Housui’. J. Japan. Soc. Hort. Sci. 73: 511–518.
•  Sawamura,  Y.,  T.  Saito,  Y.  Yamamoto,  T.  Sato,  K.  Hayashi and  K.  Kotobuki. 2002b. Identification of self-incompatible genotype and accumulation of S₄ and S₅ allele in Japanese pear. J. Japan. Soc. Hort. Sci. 71 (Suppl. 2): 220 (In Japanese).
•  Sawamura,  Y.,  N.  Takada,  T.  Yamamoto,  T.  Saito,  T.  Kimura and  K.  Kotobuki. 2008. Identification of Parent-offspring Relationships in 55 Japanese Pear Cultivars using S-RNase allele and SSR markers. J. Japan. Soc. Hort. Sci. 77: 364–373.
•  Sijacic,  P.,  X.  Wang,  A. L.  Skirpan,  Y.  Wang,  P. E.  Dowd,  A. G.  McCubbin,  S.  Huang and  T. H.  Kao. 2004. Identification of the pollen determinant of S-RNase-mediated self-incompatibility. Nature 429: 302–305.
•  Takasaki,  T.,  K.  Okada,  C.  Castillo,  Y.  Moriya,  T.  Saito,  Y.  Sawamura,  N.  Norioka,  S.  Norioka and  T.  Nakanishi. 2004. Sequence of the S₉-RNase cDNA and PCR-RFLP system for discriminating S₁- to S₉-allele in Japanese pear. Euphytica 135: 157–167.
•  Takayama,  S. and  A.  Iosgai. 2005. Self-incompatibility in plants. Annu. Rev. Plant Biol. 56: 467–489.
•  Tanabe,  K.,  F.  Tamura,  A.  Itai and  S.  Hayashi. 2001. New Japanese pear cultivars, ‘Akibae’, ‘Zuishyu’ and ‘Shinjyu’. J. Japan. Soc. Hort. Sci. 70 (Suppl. 1): 220 (In Japanese).
•  Terami,  H.,  H.  Torikata and  Y.  Shimazu. 1946. Analysis of the sterility-factors existing in varieties of the Japanese pear (Pyrus serotina Rehd. var. culta Rehd.). Studies Hort. Inst. Kyoto Imp. Univ. 3: 267–272 (In Japanese with English abstract).
•  Ushijima,  K.,  H.  Sassa,  A. M.  Dandekar,  T. M.  Gradziel,  R.  Tao and  H.  Hirano. 2003. Structural and transcriptional analysis of the self-incompatibility locus of almond: Identification of a pollen-expressed F-box gene with haplotype-specific polymorphism. Plant Cell 15: 771–781.
•  Ushijima,  K.,  H.  Sassa,  R.  Tao,  H.  Yamane,  A. M.  Dandekar,  T. M.  Gradziel and  H.  Hirano. 1998. Cloning and characterization of cDNAs encoding S-RNases from almond (Prunus dulcis): primary structural features and sequence diversity of the S-RNases in Rosaceae. Mol. Gen. Genet. 260: 261–268.
•  Wang,  Y.,  T.  Tsukamoto,  K. W.  Yi,  X.  Wang,  S. S.  Huang,  A. G.  McCubbin and  T. H.  Kao. 2004. Chromosome walking in the Petunia inflata self-incompatibility (S-) locus and gene identification in an 881-kb contig containing S₂-RNase. Plant Mol. Biol. 54: 727–742.
•  Wang,  Y.,  X.  Wang,  A. G.  McCubbin and  T. H.  Kao. 2003. Genetic mapping and molecular characterization of the self-incompatibility (S) locus in Petunia inflata. Plant Mol. Biol. 53: 565–580.
•  Wang,  C. L.,  J.  Wu,  G. H.  Xu,  Y. B.  Gao,  G.  Chen,  J. Y.  Wu,  H. Q.  Wu and  S. L.  Zhang. 2010. S-RNase disrupts tip-localized reactive oxygen species and induces nuclear DNA degradation in incompatible pollen tubes of Pyrus pyrifolia. J. Cell Sci. 123: 4301–4309.
•  Wang,  C. L.,  G. H.  Xu,  X. T.  Jiang,  G.  Chen,  J.  Wu,  H. Q.  Wu and  S. L.  Zhang. 2009. S-RNase triggers mitochondrial alteration and DNA degradation in the incompatible pollen tube of Pyrus pyrifolia in vitro. Plant J. 57: 220–229.
•  Williams,  J. S.,  J. P.  Der,  C. W.  dePamphilis and  T.-h.  Kao. 2014a. Transcriptome analysis reveals the same 17 S-locus F-box genes in two haplotypes of the self-incompatibility locus of Petunia inflata. Plant Cell 26: 2873–2888.
•  Williams,  J. S.,  C. A.  Natale,  N.  Wang,  S.  Li,  T. R.  Brubaker,  P.  Sun and  T.-h.  Kao. 2014b. Four previously identified Petunia inflata S-locus F-box genes are involved in pollen specificity in self-incompatibility. Mol. Plant 7: 567–569.
•  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.
•  Yasunobu,  Y.,  C.  Ogaki and  T.  Watanabe. 1977. A new Japanese pear (Pyrus serotina REHDER var. culta REHDER) variety ‘Choju’. Res. Rep. Kanagawa Hort. Inst. 24: 22–25 (In Japanese with English abstract).