原著論文
Haplotype Transition at the Freestone-Melting flesh Locus during Japanese Peach Cultivar Improvement
2025 年 94 巻 2 号 p. 162-173
詳細
2025 年 94 巻 2 号 p. 162-173
The Freestone-Melting flesh (F-M) locus in peach controls two strongly linked traits that have significant effects on fruit quality: pit adhesion (freestone/clingstone) and flesh type (melting/non-melting). Recently DNA markers have been developed to identify major haplotypes of M0, M1, M2, M2b, and M3 at this locus. Japanese peach cultivars were historically developed using cultivars introduced from overseas; however, how the haplotypes have changed at the F-M locus during major cultivar improvement remains unknown. In this study, we aimed to clarify the transition of haplotypes at the F-M locus in major varieties of peach cultivars from the past to the present. First, we modified the F-M locus haplotype genotyping method by multiplexing the PCR to genotype many cultivars efficiently. Then, we genotyped the F-M locus of 63 cultivars, including major cultivars in 1934, 1980, and 2020 using this method. The results show that the number of cultivars harboring the M1 haplotype, which is a dominant freestone haplotype, decreased significantly, and cultivars homozygous for the M0 haplotype dominated. Haplotypes other than M0 and M1 were infrequent among major cultivars. Cultivars harboring M1 have not always been evaluated as freestone in the literature, while we found a strong relationship between the presence of the M1 haplotype and the freestone trait. All the genetically non-melting cultivars were evaluated as non-melting. There has been only one case in which a genetically melting cultivar was previously evaluated as semi-melting, although all other genetically melting cultivars were evaluated correctly. This result suggests that DNA markers could be effective for freestone breeding and there were no major problems in flesh texture prediction using this marker system. The haplotype transition at the F-M locus during Japanese peach breeding elucidated in this study could lead to new research on freestone/clingstone and flesh texture traits and their application in breeding.
The peach Melting flesh (M) locus determines two flesh types, melting or non-melting. Melting flesh softens rapidly in the late stages of fruit ripening, whereas non-melting flesh softens slowly, and is also known as “rubbery” (Bailey and French, 1949; Peace et al., 2005; Sandefur et al., 2013). Furthermore, this locus is strongly linked to a Freestone (F) locus that determines the adhesion of the flesh to the stone, i.e. whether the type will be freestone or clingstone (Bailey and French, 1932, 1949). The mesocarp of freestone peaches freely separates from the endocarp, whereas that of clingstone peaches does not separate from the endocarp (Sandefur et al., 2013). The two loci are strongly linked, with melting flesh and freestone being the dominant traits (Peace et al., 2005; Wang et al., 2024). To date, the phenotypes freestone melting flesh (FMF), clingstone melting flesh (CMF), and clingstone non-melting flesh (CNMF) have been classified, while cultivars with consistently freestone non-melting flesh (FNMF) have not been reported (Peace et al., 2005). However, the evaluation of pit adhesion is not always easy because some cultivars show an intermediate pit adhesion phenotype in which a portion of the flesh adheres to the pit (Kikuchi, 1955). Bailey and French (1949) reported that there are varying degrees of pit adhesion from year to year.
At the Freestone-Melting flesh (F-M) locus, endo-polygalacturonase (endo-PG) genes have been identified as candidate genes (Morgutti et al., 2006; Peace et al., 2005), and subsequent studies have reported multiple haplotypes for these genes (Gu et al., 2016; Morgutti et al., 2017; Nakano et al., 2020). Endo-PG randomly cleaves the pectate chain, a component of the cell wall, effectively reducing its molecular size (Pressey and Avants, 1978). Recently, Nakano et al. (2020) proposed a hypothesis using resequencing data from 412 breeding lines, which defined three endo-PG genes, PGM-M0, PGM-M1, and PGF, and identified haplotypes based on the presence or absence of these genes: the M0 haplotype harbors PGM-M0 only, M1 PGM-M1 and PGF, M2 and M2b PGM-M1 only, and M3 which does not harbor any functional PG in terms of flesh texture or pit adhesion (Table S1). M1 is the original haplotype, and the different deletions of the genomic region including PGF resulted in the M2 and M2b haplotypes. PCR-based DNA markers detected the difference in each deletion in M2 and M2b haplotypes (Nakano et al., 2020). According to the nomenclature of Gu et al. (2016), only the M1 haplotype carrying the gene PGF for pit adhesion has been considered the dominant freestone type. For flesh texture, the M0 haplotype with PGM-M0 and the M1 haplotype with PGM-M1 and PGF were the dominant melting types, whereas the M3 haplotype lacking effective endo-PG genes and the M2 haplotype with only the PGM-M1 gene were a recessive non-melting type (Nakano et al., 2020). The reason PGM-M1 is a non-melting type is unknown. For haplotype identification, a PCR-based genotyping method was developed using 19 Japanese varieties that can characterize the five major haplotypes found in Japanese cultivars M0, M1, M2, M2b, and M3 (Nakano et al., 2020). Among 393 cultivars with re-sequencing genotyping data, 24 minor cultivars harbor haplotypes other than the five major haplotypes, but they are all foreign cultivars.
Although peaches have historically been cultivated in Japan, the cultivation of old Japanese peaches almost disappeared by the end of the Meiji era because the peach varieties introduced from overseas during the Meiji era had far better fruit quality than the old Japanese cultivars in terms of taste and fruit size (Kikuchi, 1955). As a result of breeding efforts using introduced varieties, major domestic varieties after the Meiji era have undergone major changes (Kikuchi, 1955; Yoshida, 1989). As pit adhesion and flesh texture are important traits that determine peach quality, the transition of the F-M locus haplotypes in leading cultivars is an interesting example of the use of genetic variation in the fruit industry in Japan. Nakano et al. (2020) suggested that the M0 haplotype is widespread in current Japanese varieties, while the extent to which M0 and other haplotypes have spread in major varieties from the past to the present has not been studied in detail. Information on the haplotypes selected in major cultivars is also important for Japanese peach breeding.
Therefore, the objective of this study was to clarify the haplotype transition of the F-M locus in major peach cultivars from the past to the present. Genotyping using DNA markers is necessary to reveal the F-M locus haplotype in many peach cultivars. However, the developed PCR-based DNA markers required the amplification of individual haplotypes under different PCR conditions, and the close amplicon sizes of the most important marker required methods with higher separation accuracy such as electrophoresis on polyacrylamide gels. Therefore, in this study, we first improved the identification method for the major F-M locus haplotypes of the Japanese cultivars M0, M1, M2, M2b, and M3 so that the haplotypes could be identified by agarose gel electrophoresis in a single PCR. To clarify the transition of haplotypes in major peach cultivars from the past to the present, the genotypes at the F-M locus of peach cultivars with large production areas were identified based on statistics for 1934, 1980, and 2020, for which published statistical data were available. We also validated the DNA markers by comparing the genotypes with the results of pit adhesion and flesh texture in the literature.
All plant materials analyzed in this study (Table S2) were maintained at the Institute of Fruit Tree and Tea Science, NARO, Tsukuba, Ibaraki. English-translated cultivar names conformed to NARO Genebank or original breeding literature.
DNA extractionGenomic DNA was extracted from the young leaves or flower buds. An unfolding young leaf was sampled and kept in a 2-mL tube with a 5-mm stainless bead at room temperature. Dormant flower buds were sampled from dormant branches, and most brown bud scales were removed manually using tweezers immediately before DNA extraction. Five of these scale-removed buds were sampled in 2-mL tubes. Leaf or flower bud samples in tubes were disrupted and homogenized at room temperature three times at 25 Hz for 1 min using a TissueLyser II (QIAGEN, Dusseldorf, Germany) without sample freezing. Genomic DNA was extracted using a DNeasy Plant Mini Kit (QIAGEN), according to the manufacturer’s protocol. Because DNA could be extracted from flower buds, as well as the young leaves, DNA extraction from peaches is feasible almost year-round.
PCR and fragment detectionFragments for F-M locus genotyping were amplified by PCR using BIOTAQ DNA Polymerase (Meridian, OH, USA) according to a previously described method (Nakano et al., 2020) with modifications. PCR was performed in a 10 μL reaction volume containing 1 μL of 10× NH4 Reaction Buffer, 0.5 μL of 50 mM MgCl2, 0.8 μL of 2.5 mM of dNTP mixture, 0.5 μL of 10 μM PGM/F forward and reverse primers, 0.5 μL of 10 μM M2D forward and reverse primers, 0.15 μL of 10 μM M3D forward and reverse primers, 0.1 μL of 10 μM M2D2 forward and reverse primers, 0.1 μL of BIOTAQ DNA Polymerase, 1 μL of 10 ng·μL−1 genomic DNA and 4.1 μL of water. The sequences of all primer sets are shown in Nakano et al. (2020). PCR conditions were as follows: 95°C for 3 min; 30 cycles of 95°C for 20 s, 56°C for 15 s, 72°C for 45 s; 72°C for 7 min. Amplified fragments were separated on a 2% high-resolution agarose gel (PrimeGel Agarose PCR-Sieve HRS; Takara, Japan) with 0.5 × TBE and stained with GelRed Nucleic Acid Gel Stain (Biotium, CA, USA). The 100 bp DNA Ladder (Takara Bio Inc., Shiga, Japan) was used as the DNA size marker.
The primer sets of PGM/F, M2D, M3D, M2D2, NDP1, and NDP2 were developed for PCR based genotyping (Nakano et al., 2020). The five major haplotypes M0, M1, M2, M2b, and M3 that were found in 19 Japanese cultivars (Nakano et al., 2020) could be genotyped by using the four primer sets, PGM/F, M2D, M3D, and M2D2. The other minor haplotypes, M1b, M2c, M1r1, M1r2, M1r3, and M2r1, were found only in foreign cultivars (Nakano et al., 2020), so we used the four primer sets for PCR genotyping as we focused mainly on Japanese cultivars.
StatisticsStatistics on the major cultivar-producing areas in 1934 (Showa 9) were obtained from the Ministry of Agriculture (1936). They are available at the National Diet Library Digital Collection. Based on a preliminary survey conducted in 1933 (Showa 8), cultivars that were recognized as major cultivars distributed nationwide were classified as Agreement cultivars (“Kyotei-hinshu” in Japanese), while cultivars that were not distributed nationwide, but were recognized as major cultivars in some regions, were classified as Non-agreement cultivars (“Kyotei-gai-hinshu” in Japanese; Ministry of Agriculture, 1936). Within the Non-agreement cultivars, 10 cultivars available at the NARO Genebank were genotyped. The areas of each Non-agreement cultivar are shown separately in the columns for each prefecture where they were produced; hence, the total areas of genotyped Non-agreement cultivars are shown as the sum of the areas of each cultivar shown in different prefectures. The original area unit in the statistics is “cho”. This is almost the same as a hectare (ha), with less than 1% difference (one cho is 0.99174 ha, as shown by MAFF: https://www.maff.go.jp/e/data/stat/93th/attach/pdf/index-28.pdf, October 9, 2024). Hence, the numbers were not converted from the original values as shown in the statistics. Statistics for 1980 (Showa 55) were obtained from the Ministry of Agriculture, Forestry and Fisheries (MAFF) (1981). They are available at the National Diet Library Digital Collection. Those for 2020 (Reiwa 2) are available online at the MAFF statistics (https://www.maff.go.jp/j/tokei/kouhyou/tokusan_kazyu/index.html, October 9, 2024) as “Tokusankaju-seisan-dotai-tou-chosa Reiwa 2 nen-san”. Peach statistics for fresh consumption, processing, and nectarines are summarized in three different spreadsheets in the MAFF statistics; hence, all cultivars in the three sheets were pooled and summarized in this study for 2020 statistics. To show the transition of haplotypes among the leading cultivars in each year, we calculated the sum of the production areas of the M0 homozygous, M1 homozygous, M1 heterozygous, and other genotypes. The proportion of each genotype per year was calculated by dividing the sum of the areas of each genotype by the total production area. The area proportions of cultivars that could not be genotyped in this study were grouped as not genotyped.
Pit adhesion and flesh texture phenotype in the literatureThe detailed definitions of pit adhesion and flesh texture phenotypes, especially the intermediate phenotype, in the literature are not clear, partly because they include historical literature. In this paper, we consider pit adhesion as follows: freestone is a state in which pit and flesh are separated, clingstone is a state in which pit and flesh are not separated, semi-freestone is a state in which pit and flesh are mostly separated, but partially attached, and semi-clingstone is a state in which pit and flesh are partially attached. For flesh texture, we consider melting flesh as that which softens rapidly in the late stages of fruit ripening, non-melting flesh that softens slowly, semi-melting flesh that softens rapidly, but not as rapidly as melting flesh, and semi-non-melting that softens slowly, but not as slowly as non-melting.
There were 15 diploid genotypes of individual peaches possessing one or two of five haplotypes, M0, M1, M2, M2b, and M3: M0M0, M0M1, M0M2, M0M2b, M0M3, M1M1, M1M2, M1M2b, M1M3, M2M2, M2M2b, M2M3, M2bM2b, M2bM3, and M3M3. Among the primer sets developed by Nakano et al. (2020), the lengths of the amplified products derived from PGM/F, M2D, M3D, and M2D2 did not overlap (Fig. 1A). Multiplex PCR using these four primer sets revealed that the band patterns of all 15 genotypes differed (Fig. 1B). The results of the multiplex PCR using genomic DNA corresponding to each of the 15 genotypes are shown in Figure 1C. In some genotypes, a faint nonspecific amplified product slightly larger than the 1.5 kb M2D amplification product was observed; however, the two could be distinguished because the nonspecific product was clearly larger than the 1.5 kb size marker. Genotyping results for the cultivars analyzed in this study are shown in Figure S1.
Multiplex PCR genotyping at the F-M locus. A: Fragment size and schematic representation of the banding pattern in agarose gel electrophoresis. Fragment sizes (left) and haplotypes amplifying each fragment (right) are shown according to Nakano et al. (2020). The haplotypes shown here are the possible haplotypes among the five haplotypes, M0, M1, M2, M2b, and M3; the other possible minor haplotypes are not shown. B: Expected banding pattern of fragments for each genotype shown above. C: Agarose-gel electrophoresis of multiplex PCR products. M0M0: ‘Akatsuki’, M0M1: ‘Ookubo’, M0M2: ‘Sakuhime’, M0M2b: ‘Okayama 3’, M0M3: ‘Shang Hai Shui Mi Tao’, M1M1: ‘Tian Jin Shui Mi Tao’, M1M2: ‘Lovell’, M1M2b: ‘Zansetsu Shidare’, M1M3: a breeding selection, M2M2: ‘Nishiki’, M2M2b: ‘Kanto 5’, M2M3: a breeding selection, M2bM2b: a breeding selection, M2bM3: ‘Mochizuki’, M3M3: ‘Hinanotaki’. DNA size markers (from bottom to top): 200, 300, 400, 500 (stronger signal than others), 600, 700, 800, 900, 1000, 1500 bp.
To study the transition of F-M locus haplotypes in domestic peach cultivars, we genotyped the leading cultivars using publicly available statistics from 1934, 1980, and 2020, with an interval of approximately 40 years. In the 1934 statistics (Table 1), six varieties that were widespread throughout Japan were grouped as Agreement cultivars, and four of them, ‘Denjuurou’, ‘Tian Jin Shui Mi Tao’, ‘Rikaku’, and ‘Doyou’, were heterozygous or homozygous for the M1 haplotype. The other two ‘Tachibana Wase’ and ‘Hakutou’ were M0 homozygous. Cultivars other than Agreement cultivars, which were not widespread nationwide, but were regionally important at the time, are listed as Non-agreement cultivars. Table 1 lists the 10 varieties for which genotypes were analyzed. Seven of them were heterozygous or homozygous for the M1 haplotype, with only ‘Kintou’ being homozygous for M0 and ‘Meigetsu’ being homozygous for M3. M1-harboring cultivars in Table 1 accounted for a total of 51.8% (M1 homozygous: 13.3%, M1 heterozygous: 38.5%) of the total cultivated area, whereas the M0 homozygous varieties accounted for 23.6% (Fig. 2).
F-M locus haplotype, flesh type, pit adhesion, and producing area of leading cultivars in 1934 (Showa 9).
Proportion of producing area for cultivars of M1 homozygous, M1 heterozygous, M0 homozygous and other genotype at the F-M locus.
The genotypes of the varieties listed in Table 1 were all melting types, except for ‘Meigetsu’. All cultivars with flesh phenotypes that indicated melting in the literature were also genetically melted. The evaluation of pit adhesion varied among the literature even within the same cultivar (Table 1), but for the clingstone genotype, i.e., ‘Tachibana Wase’ (M0M0), ‘Hakutou’ (M0M0), ‘Shang Hai Shui Mi Tao’ (M0M3), and ‘Kintou’ (M0M0), all were clingstone except one reference that indicated a semi-freestone for ‘Tachibana Wase’. The other varieties were genetically freestone, heterozygous, or homozygous for the M1 haplotype, except for ‘Meigetsu’, for which pit adhesion evaluation was not available. No varieties were rated as clingstone; they were evaluated as either semi-clingstone, semi-freestone, or freestone in the literature.
F-M locus haplotype, flesh type and pit adhesion of leading cultivars in 1980To determine the F-M locus haplotypes of the leading cultivars in 1980, the haplotypes of the top 10 cultivars in terms of production area were genotyped (Table 2). Three varieties, ‘Ookubo’, ‘Sunago Wase’, and ‘Nunome Wase’, were heterozygous or homozygous for the M1 haplotype. Five cultivars, ‘Hakuhou’, ‘Kurakata Wase’, ‘Yamane Hakutou/Aichi Hakutou’, ‘Shuuhou’, and ‘Hakutou’, were M0 homozygous. The other haplotypes were M0M3 for ‘Shimizu Hakutou’ and M2M2b for ‘Kantou 5’, the only canning peach in this table. The three cultivars possessing M1 accounted for a total of 42.6% of the producing area due to the large area for ‘Ookubo’. Meanwhile, the five M0 homozygous varieties accounted for 36.4% of the total area (Fig. 2).
F-M locus haplotype, flesh type, pit adhesion, and producing area of leading cultivars in 1980 (Showa 55).
Of the cultivars listed in Table 2, ‘Kantou 5’ was the non-melting genotype (M2M2b) and the flesh type was evaluated as non-melting in several literatures. All other varieties were of the melting genotype, and all fruit phenotypes were also evaluated as melting, except for ‘Shuuhou’, which was evaluated as semi-melting in one literature. The pit adhesion phenotype also varied in the literature within the same cultivar, but the seven cultivars of the clingstone genotype (M0M0, M0M3, or M2M2b) were all evaluated as clingstone, except for ‘Hakuhou’ and ‘Shuuhou’, which were evaluated as semi-clingstone in some literature. Of the other three cultivars, ‘Ookubo’ (M0M1) was always rated freestone, while ‘Nunome Wase’ (M0M1) was rated either semi-clingstone, semi-freestone, or freestone. ‘Sunago Wase’ (M0M1) was rated semi-clingstone or clingstone.
F-M locus haplotype, flesh type, and pit adhesion of leading cultivars in 2020The top 10 cultivars in 2020 statistics were all clingstone, and freestone ‘Ougontou’ was in 11th position; hence, the haplotypes for the top 11 were genotyped (Table 3). Of the 11 varieties, only ‘Ougontou’ was heterozygous for M1. All other cultivars were M0 homozygous, except for ‘Shimizu Hakutou’, which had the M0M3 haplotype. The genotype of ‘Yumemizuki’ could not be genotyped, but its parents, ‘Asama Hakutou’ and ‘Gyousei’, were both M0 homozygous (Table S2; Fig. S1). Hence, its genotype was estimated to be M0 homozygous. ‘Ougontou’ with M1 accounted for only 1.6% of the total production area. Meanwhile, M0 homozygous cultivars including ‘Yumemizuki’ accounted for at least 69.8% of total producing area, and more M0 homozygous cultivars could be included among the cultivars not genotyped in this study (Fig. 2).
F-M locus haplotype, flesh type, pit adhesion, and producing area of leading cultvars in 2020 (Reiwa 2).
All the cultivars shown in Table 3 were of the melting genotype and were evaluated as melting phenotypes. Only freestone ‘Ougontou’ was genetically freestone (M0M1), while all the other cultivars were genetically clingstone. Of the genetically clingstone cultivars, all were evaluated as the clingstone phenotype, except for ‘Hakuhou’, which was evaluated as semi-clingstone in two papers. The current leading cultivars are dominated by melting and clingstone peaches, and nine of the top 11 F-M locus haplotypes were fixed as M0 homozygous, with only ‘Shimizu Hakutou’ (M0M3) and ‘Ougontou’ (M0M1) being heterozygous.
F-M locus haplotype, flesh type, and pit adhesion of cultivars released by the national instituteHaplotypes at the F-M locus were identified in 29 cultivars released by NARO or equivalent research institutes before renaming (Table 4). Nine cultivars were homozygous for M0, six were heterozygous or homozygous for M1, and five were either M0M2, M0M2b, or M0M3. These were melting genotypes, and all cultivars were evaluated for the melting phenotype. All samples were used for fresh consumption. Haplotypes of eight varieties were either M2M2, M2M2b or M2bM3. All had non-melting or semi-non-melting phenotypes and were canning peaches. ‘Hinanotaki’, which is mainly used ornamentally or in home gardening, was the only M3 homozygous cultivar among the released cultivars, and its flesh phenotype was evaluated as semi-non-melting.
F-M locus haplotype of cultivars released by NARO or equivalent research institute before renaming.
Among the cultivars shown in Tables 1–4, we found no major discrepancies between the F-M locus genotype and flesh type (melting, or non-melting), but there were considerable variations in the evaluation of phenotypes in the literature and discrepancies between genotypes and phenotypes regarding pit adhesion. Table 5 summarizes the M1 haplotype status and phenotype of pit adhesion for the 57 genotyped varieties in this study, for which information on the pit adhesion phenotype is available in the literature. Consequently, of the 38 varieties that do not possess the M1 haplotype, 35 were rated as clingstone, and the remaining three were rated as clingstone or semi-clingstone (‘Hakuhou’ and ‘Shuuhou’) or clingstone or semi-freestone (‘Tachibana Wase’), depending on the paper checked. In contrast, the 19 varieties harboring the M1 haplotype had a wide range of pit adhesion evaluations, ranging from clingstone to freestone.
Contingency table of pit adhesion and the presence of haploid M1 based on the cultivars genotyped in this study.
In this study, we first improved the genotyping method for the F-M locus haplotype so that the haplotype could be determined using a single PCR. This method required a high-resolution agarose gel and did not employ polyacrylamide gel electrophoresis. Identification can be performed using an ordinary thermal cycler and agarose gel electrophoresis, making this a simple and versatile method. It is also a co-dominant marker in which at least one product is detected in any of the 15 genotypes derived from the M0, M1, M2, M2b, and M3 haplotypes; if no signal is detected, the amplification fails. However, it should be noted that when genotyping the progeny of unutilized genetic resources, haplotypes other than M0, M1, M2, M2b, and M3 may exist, resulting in a pattern different from the one shown in Figure 1B. These minor haplotypes, such as M1b, M2c, M1r1, M1r2, M1r3, and M2r1, were found in minor foreign cultivars (Nakano et al., 2020), and the inability to genotype these haplotypes is a limitation of our multiplex PCR method. It is assumed these minor haplotypes were caused by mutations that occurred in foreign countries and did not affect Japanese cultivar improvement. However, if conflicts between genotype and phenotype are found by our multiplex PCR method, we may need to genotype them by using other primer sets such as NDP1 and/or NDP2.
Haplotype transition in the peach F-M locus in JapanWe analyzed the haplotypes at the F-M locus of leading domestic peach cultivars in 1934, 1980, and 2020. The major peach cultivars changed significantly during this period, with only ‘Ookubo’ and ‘Hakutou’ being found both between 1934 and 1980, ‘Hakuhou’ and ‘Shimizu Hakutou’ between 1980 and 2020, and no varieties being common between 1934 and 2020. In this major change in the leading cultivars, a major trend regarding haplotypes at the F-M locus was evident: the number of varieties possessing the M1 haplotype decreased dramatically, and the M0 haplotype became more widespread (Fig. 2). In addition, there were fewer haplotypes other than the M0 and M1 haplotypes throughout the study period (Tables 1–3).
Freestone peaches in JapanAs of 1934, the percentage of production area accounted for by cultivars possessing M1 shown in Table 1 totaled 51.8%. Sakaguchi (1929) evaluated 73 domestic and foreign varieties that existed at the time and reported on the pit adhesion phenotype. Of the varieties tested, only eight were clingstones, one was between semi-freestone and clingstone, 27 were semi-freestones, and 37 were freestones. The semi-freestone cultivars include ‘Tachibana Wase’, but other literature evaluated this cultivar as clingstone (Kajiura and Tanigawa, 1935; Kimura, 1959). Therefore, although some of the cultivars evaluated as semi-freestones by Sakaguchi (1929) may have been clingstones, it is likely that freestone peach production was common around 1929. However, the leading freestone cultivars have declined since then to the point where even the cultivar with the largest area, ‘Ougontou’, accounted for only 1.6% of the total production area in 2020 (Table 3).
Kajiura and Tanigawa (1935) stated that white peaches have the disadvantage of being clingstone, and Ono (1957) stated that ‘Ookubo’ is suitable for use as a peach for fresh consumption because of its freestone phenotype. Thus, it is natural to assume that the percentage of freestone cultivars is maintained or increases during peach cultivar improvement. However, the number of freestone cultivars decreased significantly over time (Tables 1–3). Recently, Wang et al. (2024) showed that even among stony hard peaches that do not soften after harvest, freestone stony hard peaches become mealy and soft after harvest. Although it is not clear how mealiness differs between freestone and clingstone peaches in the normal melting genetic background, if freestone peaches tend to be mealier than clingstone, this could be a reason for the decline in freestone cultivars in Japan. Determining the cause of the decline in the number of freestone cultivars, such as differences in the degree of mealiness, is an issue that needs to be addressed.
Relationship between M1 haplotype and freestone traitNakano et al. (2020) have stated that further studies are required to elucidate the role of PGF in regulating stone adhesion. This study did not show that the M1 haplotype always caused the freestone phenotype (Table 5). However, Bailey and French (1949) found that evaluation of pit adhesion is not always easy because of annual fluctuations and many cultivars show intermediate phenotypes between freestone and clingstone (Kikuchi, 1955). Yoshida (1989) stated that pit adhesion is an incompletely dominant trait and that some heterozygous individuals are semi-freestones or semi-clingstones. Thus, pit adhesion is a difficult trait to evaluate, and even within the literature cited in this study, the evaluations varied. The reasons for the phenotypic variation within genetically freestone cultivars are unknown; however, other loci with small effects may contribute to the pit adhesion trait even though F/M locus has the major effect. PGF, an endo-PG gene defined by Gu et al. (2016) as conferring a freestone, was found only in the M1 haplotype (Table S1). Of the 19 varieties heterozygous or homozygous for M1 (Table 5), only ‘Sunago Wase’, ‘Saotome’, Chiyodared’, and ‘Shizukured’ were evaluated as clingstone in the literature (Tables 2 and 4), while the others were rated semi-clingstone, semi-freestone, or freestone (Table 5). In addition, ‘Rikaku’, ‘Doyou’, ‘Ookubo’, ‘Daitouryou’, and ‘Ougontou’, were always evaluated as freestone in the literature, and they were M1 heterozygous or homozygous. On the other hand, of the 38 cultivars that do not possess M1 (Table 5), only three cultivars, ‘Tachibana Wase’, ‘Hakuhou’ and ‘Shuuhou’, were evaluated other than clingstone in the literature (Tables 1–3). Therefore, even though it is not always possible to obtain perfect freestone progeny by selecting M1-harboring individuals, the effect of PGF in M1 on pit adhesion is significant, and it is reasonable to select individuals with M1 to breed freestone cultivars. Another possibility for the discrepancy between genotype and phenotype in terms of pit adhesion is that the gene responsible for the trait is not PGF, but a different gene located close to PGF, and the discrepancy may occur due to recombination between the two genes. To verify this possibility, it may be possible to prepare the offspring of cultivars with discrepant phenotypes and confirm whether the recombinant genomic region is inherited.
Melting flesh has always been dominant in Japanese peach breedingThe genetically non-melting cultivars, for which flesh phenotypes are available in the literature, were evaluated as non-melting or semi-non-melting (Tables 1–4). Only one melting cultivar, ‘Shuuhou’ (M0M0), was evaluated as semi-non-melting in one paper, and all others were melting. Thus, the DNA marker developed by Nakano et al. (2020) can identify flesh type, melting or non-melting, and the use of this marker could be highly effective for breeding.
‘Early Gold’ released by NARO was genotyped as M1M1 by Nakano et al. (2020), which is inconsistent with its non-melting phenotype, and they suggested that the reason for this discrepancy could be due to a potential mix-up with another accession in the SRA (Sequence Read Archive) database. The genotype of this cultivar was non-melting M2M2 (Table S2; Fig. S1), which supports the possibility of an error in the database.
Non-melting major cultivars in JapanThe leading genetically non-melting cultivar in 1980 was ‘Kantou 5’ (Table 2), which was released as a canning peach. The breeding of canned peaches in Japan began in 1935 in the public sector, such as the National Horticultural Research Station, Ministry of Agriculture and Commerce, and prefectural agricultural experiment stations, such as Okayama and Kanagawa (Kobayashi, 1990; Yoshida, 1989). ‘Kantou 5’ was an achievement by the National Institute. Peaches for canning are generally required to be non-melting because the flesh can maintain its shape when heated (Yoshida, 1989), and because firm flesh is advantageous for machines such as pitters (Mori, 1969). The ‘Kantou 5’ genotype is M2M2b, but neither of these haplotypes is found in any of the other varieties listed in Tables 1–3, suggesting that they have not become widespread compared to other leading cultivars in Japan.
As a cultivar homozygous for M3, a haplotype that causes non-melting, ‘Meigetsu’ is included in Non-agreement cultivars in Table 1. Nakano et al. (2020) also showed ‘Meigetsu’ to be homozygous for M3. The only literature describing the fruit traits of ‘Meigetsu’ could be found in Togashi (1933), and there was no clear description of flesh type or pit adhesion traits, but it is possible that the phenotype is non-melting, since it is described as having a long shelf life. Although ‘Meigetsu’ peaches are large and the taste is described as being similar to that of ‘Hakutou’. It is a late-ripening cultivar that ripens in the first half of September and Togashi (1933) reported that it would not be an important cultivar for large-scale production. Indeed, its production has not spread and cannot be found in other statistics or in the literature. As another peach cultivar like M3, the late-ripening ‘Daijumitsuto’ has also been shown to be M3M3 (Nakano et al., 2020).
M3 haplotype in Japan‘Shang Hai Shui Mi Tao’, an important ancestral cultivar in Japanese peach breeding (Kikuchi, 1955; Yamamoto et al., 2003), was heterozygous for M0 and M3 (Table 1; Fig. S1). ‘Chinese Cling’, which is considered to be the same cultivar (Kikuchi, 1955), is M0M3 as reported by Nakano et al. (2020). The introduced foreign cultivar ‘Carman’, included in Non-agreement cultivars in the 1934 statistics, also has the M3 haplotype (M1M3: Table 1; Fig. S1). It is noteworthy that the production area of ‘Carman’ in 1934 was relatively large (188.4 ha), and was recorded in a wide range of prefectures, including Niigata (77.7 ha), Ishikawa (27.5 ha) Nagano (24.0 ha), Kyoto (37.6 ha), Hyogo (18.9 ha), and Okayama (2.7 ha) (Ministry of Agriculture, 1936). Therefore, the M3 haplotype could have spread from foreign introduced varieties such as ‘Shang Hai Shui Mi Tao’ or ‘Carman’ in the early stage of peach improvement in Japan. However, as far as we understand, the only cultivar with the M3 haplotype in 1934, other than ‘Carman’ and ‘Shang Hai Shui Mi Tao’, was ‘Meigetsu’, which had a limited production area (Table 1), and in 1980 and 2020, the only cultivar with the M3 haplotype was the ‘Shimizu Hakutou’ (M0M3). This is very different from M0, which dominates among Japanese cultivars, suggesting that, unlike M0, the M3 haplotype is not selected for in Japan. Although it is possible that M3 was lost by chance due to a bottleneck effect in the breeding process, such as repeated use of ‘Hakutou (M0M0)’ or its descendants as parents, it would be interesting to question whether M3 heterozygosity in melting peaches has a negative effect or whether homozygosity rather than M0 heterozygosity has a positive effect.
M0 haplotype dominance in major Japanese cultivarsIn 1980, as many as five cultivars were M0 homozygous, but all the top 10 varieties were M0 homozygous except for ‘Shimizu Hakutou’ (M0M3) in 2020 (Tables 2 and 3). The widespread occurrence of M0 haplotypes among Japanese cultivars, as shown by Nakano et al. (2020), was confirmed in this study. Thus, the striking trend at the F-M locus in Japanese peach cultivar improvement is the selection of the M0 haplotype and the elimination of all other haplotypes, especially the M1 haplotype.
Although the haplotype diversity at the F-M locus has been lost in leading modern cultivars and is dominated by M0 homozygotes, haplotypes other than M0 are maintained in the recently released cultivars by NARO (Table 4) such as ‘Tsukikagami’ (M0M1), ‘Sakuhime’ (M0M2), ‘Mochizuki’ (M2bM3), ‘Himemaruko’ (M0M3) and ‘Himekonatsu’ (M0M3). These new varieties and their progeny are valuable materials for breeding new cultivars with both good fruit quality and diverse flesh types and pit adhesion, as well as for studying the influence of endo-PG genes at the F-M locus on fruit traits such as pit adhesion and fruit softening.
The authors would like to thank Editage (www.editage.com) for the English language editing.
