Identification of QTLs controlling harvest time and fruit skin color in Japanese pear (Pyrus pyrifolia Nakai).

Using an F1 population from a cross between Japanese pear (Pyrus pyrifolia Nakai) cultivars 'Akiakari' and 'Taihaku', we performed quantitative trait locus (QTL) analysis of seven fruit traits (harvest time, fruit skin color, flesh firmness, fruit weight, acid content, total soluble solids content, and preharvest fruit drop). The constructed simple sequence repeat-based genetic linkage map of 'Akiakari' consisted of 208 loci and spanned 799 cM; that of 'Taihaku' consisted of 275 loci and spanned 1039 cM. Out of significant QTLs, two QTLs for harvest time, one for fruit skin color, and one for flesh firmness were stably detected in two successive years. The QTLs for harvest time were located at the bottom of linkage group (LG) Tai3 (nearest marker: BGA35) and at the top of LG Tai15 (nearest markers: PPACS2 and MEST050), in good accordance with results of genome-wide association study. The PPACS2 gene, a member of the ACC synthase gene family, may control harvest time, preharvest fruit drop, and fruit storage potential. One major QTL associated with fruit skin color was identified at the top of LG 8. QTLs identified in this study would be useful for marker-assisted selection in Japanese pear breeding programs.

Introduction

Pears (Pyrus spp.) are commercially cultivated in over 50 countries of temperate climate areas, and have been one of the most important fruit trees in East Asia, Europe, and North America for more than 3000 years (Bell 1990, Bell ). The Japanese pear (Pyrus pyrifolia Nakai), the European pear (Pyrus communis L.) and the Chinese pears (Pyrus bretschneideri Rehd. and Pyrus ussuriensis Maxim.) are major edible species commercially grown for fruit production. Fruit development and ripening characteristics differ notably among Pyrus, which comprise complex processes (Nashima ). All the species of Pyrus are intercrossable and there are no major incompatibility barriers to interspecific hybridization in Pyrus (Westwood and Bjornstad 1971). Pear belongs to the subfamily Spiraeoideae, tribe Pyreae, like apple (Malus × domestica Borkh.); the two species share a basic chromosome number of x = 17, which indicates a polyploid origin. A relatively recent (>50 million years ago) genome-wide duplication has resulted in the transition from nine ancestral chromosomes to 17 chromosomes in the Pyreae (Velasco ).

Marker-assisted selection (MAS) can help breeders select for desirable traits and reduce the progeny size and the cost of raising individuals to maturity in the field (Luby and Shaw 2001). However, the use of MAS is just beginning, and remain limited for a few simply inherited traits in fruit trees, because molecular marker development for MAS via bi-parental quantitative trait locus (QTL) mapping is hampered by difficulties in phenotyping in traditional breeding. Molecular markers associated with self-incompatibility (Ishimizu ), ethylene production (the 1-aminocyclopropane-1-carboxylate [ACC] synthase gene; Itai ), fruit skin color (Inoue ), pear scab resistance (Terakami ) and black spot resistance (Terakami ) have been developed and some of them are used for selection in Japanese pear breeding programs. However, many fruit-related phenotypic traits are not amenable to MAS, because little genetic analysis (including QTL analysis) has been done to establish the molecular markers associated with these traits in pear.

A total of 19 QTLs have been detected for six fruit traits (weight, diameter, length, soluble solids content, shape index, and maturity date) in seedlings from a cross between ‘Bayuehong’, hybrid of the European pear ‘Clapp’s Favorite’ and the Chinese pear ‘Zaosuli’, and the Chinese pear ‘Dangshansuli’ (Zhang ). Four genomic regions associated with fire blight resistance have been identified in the progeny of a cross between European pear cultivars ‘Passe Crassane’ and ‘Harrow Sweet’, including two main QTLs in linkage group (LG) 2A and LG 4 of ‘Harrow Sweet’ (Dondini , Le Roux ). Four QTLs were detected for leaf length, two QTLs for leaf width, two for leaf length and leaf width, and three for petiole length by using F1 populations from a cross between Chinese pears ‘Yali’ and ‘Jingbaili’ (Sun ).

Since pear and apple, a closely related species, show co-linearity between all LGs (chromosomes) and have similar genic regions (Wu , Yamamoto ), functional synteny of fruit-related QTLs might exist. Similarly, Japanese pear and European pear show good co-linearity (Yamamoto ). Liebhard reported QTLs for number of fruit, size, flesh firmness, maturity time, sugar content, and acidity based on a segregating population of the cross between the apple varieties ‘Fiesta’ and ‘Discovery’. Kenis presented a comprehensive analysis of a population derived from a cross between the apple cultivars ‘Telamon’ and ‘Braeburn’ and identified a total of 74 different QTLs for all major fruit quality traits. Longhi reported comprehensive QTL analysis of complex fruit texture physiology and mapped a highly significant QTL cluster on chromosome 10, which was co-located with Md-PG1, a polygalacturonase gene. van Dyk detected a major QTL for time of initial vegetative budbreak in apple LG 9. Because of the synteny between apple and pear, the information about apple QTLs is useful for identifying pear QTLs for fruit traits.

In the present study, we evaluated seven fruit traits (harvest time, fruit skin color, flesh firmness, fruit weight, acid content, total soluble solids content and preharvest fruit drop) using F1 progeny derived from a cross between the new elite Japanese pear cultivar ‘Akiakari’ and the indigenous cultivar ‘Taihaku’. QTLs identified in this study were examined for the use of marker-assisted selection. Furthermore, functional synteny among Pyrus as well as between pear and apple were discussed.

Materials and Methods

Plant materials and DNA extraction

Ninety-three F1 individuals from a cross between Japanese pear (Pyrus pyrifolia Nakai) ‘Akiakari’ and ‘Taihaku’ were used for QTL analysis of fruit traits and for construction of genetic linkage maps. ‘Akiakari’ is an early-maturing russet-skin type cultivar released in 2000; this cultivar originated from crossing 162-29 (‘Niitaka’ × ‘Housui’, syn. ‘Hosui’) and ‘Nashi Hiratsuka 17’ (‘Kumoi’ × ‘Kousui’) in 1984 (Kotobuki ). ‘Taihaku’ is an indigenous smooth-skin (green skin color) cultivar, planting of which was documented in the late 19th century (Kajiura and Sato 1990). All plant materials were maintained at the NARO Institute of Fruit Tree Science (Ibaraki, Japan).

Genomic DNA was isolated from young leaves on Genomic-tip 20/G anion-exchange columns (Qiagen, Germany) as described by Yamamoto .

Evaluation of phenotypic traits

F1 individuals were sown in 2005. Harvest time, fruit skin color, flesh firmness, fruit weight, acid content, total soluble solids content, and preharvest fruit drop were recorded for all F1 individuals and their parents in 2010 and 2011. The measurement methods are summarized in Table 1, according to the modified procedures reported by Kotobuki (1999) and Nishio . Fruits were harvested twice a week at their respective fruit ripening date from August to October according to a color chart that indicates the optimum ground color for picking Japanese pear (Kajiura ). The optimum ground color for picking was determined by the time that fruit skin color in the calyx end changes from green to the color corresponding to plate 4 of the ground color chart. Average value of harvest days was evaluated for F1 individuals and their parents, and then harvest time was identified as the number of days after the first genotype reached harvest time.

Table 1

Phenotypic traits evaluated in this study

Trait	Measured parameter and unit
Harvest time	Number of days after fruit of the first genotype reached harvest time
Fruit skin color	Scores: 5, suberin formation on 100% of the surface area of mature fruit; 4, 95–99%; 3, 75–95%; 2, 20–75%; 1, 0–20%
Flesh firmness	Magness-Taylor pressure test (lb)
Fruit weight	Average weight of mature fruit (g)
Acid content	pH of juice
Total soluble solids content	Brix of juice (%)
Preharvest fruit drop	Ratio of preharvest fruit drop (%)

Fruit skin color was classified into five types according to the suberin formation area on the surface of mature fruit (see Table 1 for details). Fruit weight was calculated as the average weight of all collected mature fruits. Flesh firmness, acid content of the juice, and total soluble solids content were measured for up to five mature fruits on plural harvest days. Flesh firmness was measured for two section per one mature fruit with a hand-operated Magness-Taylor penetrometer (10-lb Fruit Tester Model 10B; D. Ballauf Mfg. Co., Inc., USA). Flesh firmness was calculated at the average value for all mature fruits of F1 individuals and their parents. Penetrated fruits were mixed and pressed with a hand-operated squeezer and the juice was collected. Acid content of the juice was determined with a pH meter (IQ240; Scientific Instruments, USA). Total soluble solids content was determined with a digital refractometer (DBX-55; Atago, Japan) by adding a few drops of juice onto the lens of the measuring device, and the results were recorded as Brix (%). Preharvest fruit drop was evaluated as the ratio of preharvest drop to all fruits. The Kolmogorov-Smirnov test was used to check the normality of the data distribution for each trait, with p > 0.05 indicating a normal distribution.

Molecular marker analysis

Designations of the simple sequence repeat (SSR) markers are listed in Table 2. To construct genetic linkage maps of ‘Akiakari’ and ‘Taihaku’, a total of 654 SSR markers that originated from genome and expressed sequence tag (EST) sequences was used: 349 pear markers (Fernández-Fernández , Inoue , Nishitani , Yamamoto , 2013), and 305 apple markers (Celton , Guilford , Liebhard , Moriya , Silfverberg-Dilworth , van Dyk ).

Table 2

Molecular markers mapped on genetic linkage maps of ‘Akiakari’ and ‘Taihaku’

Marker type	Designation	Reference
Pear genomic-SSR	TsuGNH	Yamamoto et al. 2013
	BG, KA, KU, NB, NH	Yamamoto et al. 2002
	EMPc	Fernández-Fernández et al. 2006
	IPPN	Inoue et al. 2007
Pear EST-SSR	TsuENH	Nishitani et al. 2009
Apple SSR	CH, MS	Liebhard et al. 2002
	AJ, AU, CN, Hi	Silfverberg-Dilworth et al. 2006
	NZ	Guilford et al. 1997
	NZms	Celton et al. 2009
	SA	van Dyk et al. 2010
	MEST	Moriya et al. 2012
	Mdo.chr	This study

Four SSR markers were newly developed based on the sequence at the top of apple chromosome 8 (Malus × domestica whole genome v1.0, Genome Database for Rosaceae, http://www.rosaceae.org/). The forward (F) and reverse (R) primers used for the developed SSR markers were as follows: Mdo.chr8.12 (F, CGTGTTTGTGTTTTTGCTGG; R, CCAAAGGACCATAGCAGCAT), Mdo.chr8.13 (F, GTTTTGATGGTGGCGTTTCT; R, CTCAGATGCGCTTTTGGTTT), Mdo.chr8.16 (F, ACTACCGCTCCCCCTAGTGT; R, TCAACAATCACCACGGAGAA) and Mdo.chr8.10 (F, TGCAGCCCTCAAACTTTTCT; R, CAACCCAACTCCAGCAATTT). The SSR markers had the following repeat motifs: (AG)11 in Mdo.chr8.12; (GA)13 in Mdo.chr8.13; (AT)12 in Mdo.chr8.16; and (GA)11 in Mdo.chr8.10. Mdo.chr8.12 was located 551 kb from the top of apple chromosome 8, Mdo.chr8.13 at 1012 kb, Mdo.chr8.16 at 1722 kb, and Mdo.chr8.10 at 3952 kb.

The random amplified polymorphic DNA-sequence-tagged site (RAPD-STS) marker OPH19 was designed based on OPH-19425 (Inoue ), which is significantly linked to fruit skin color in Japanese pear. The following primers were newly designed: F, CAATGCAGAAATGGTGAACG; R, GGTTTGGTTTCCTTGACTAG. The PPACS2 (ACC synthase gene) marker (Itai ) was used with primers (F, CAATGCAGAAATGGTGAACG; R, GGTTTGGTTTCCTTGACTAG) newly designed to amplify the AG-repeat of the 5′-untranscribed region instead of the original CAPS marker. Self-incompatible genotypes were identified according to the method described by Ishimizu . An expansin gene MdExp7 associated with fruit softening in apple and pear (Costa ) was used for genome mapping as a gene marker.

PCR amplification for SSR and gene markers was performed in a 5-μL reaction mixture containing 2.5 μL of Multiplex PCR Master Mix containing HotStar Taq Polymerase (Qiagen), 5 pmol of each primer (F fluorescently labeled with FAM, VIC, or NED, and unlabeled R), and 5 ng of genomic DNA. Multiplex PCR amplification was performed with three to six primer combinations per reaction. The PCR profile consisted of initial denaturation for 15 min at 95°C followed by 35 cycles of denaturation for 60 s at 94°C, annealing for 90 s at 55°C, and extension for 90 s at 72°C, with a final extension for 10 min at 72°C. The amplified PCR products were separated and detected using a PRISM 3100 DNA sequencer (Applied Biosystems, USA). The sizes of the amplified bands were scored based on internal standard DNA (400HD-ROX, Applied Biosystems) using the GeneScan software (Applied Biosystems).

Construction of genetic linkage maps

JoinMap version 4.0 (van Ooijen 2006) was used with the double pseudo-testcross mapping strategy (Grattapaglia and Sederoff 1994). The Kosambi function was used to convert recombination units into genetic distances. To determine the LGs, mapping analysis was conducted with a minimum LOD score of 4.0. The LG numbers for ‘Akiakari’ and ‘Taihaku’ were assigned based on those in the reference genetic linkage maps of ‘Bartlett’ and ‘La France’ (Yamamoto ).

QTL analysis

The MapQTL ver. 6.0 software (van Ooijen 2009) was used for interval mapping, multiple QTL model mapping (MQM), and the Kruskal-Wallis test. A genome-wide LOD significance threshold was determined for each trait based on a permutation test with 1000 replications, and then QTLs with an LOD score that was significant at p < 0.05 were identified. QTLs were initially identified by interval mapping, and then the marker with the highest LOD value was used for co-factor selection in subsequent rounds of MQM analysis. To obtain significant QTLs for fruit skin color and preharvest fruit drop traits, which were not normally distributed, interval mapping and MQM analysis were performed followed by the Kruskal-Wallis test to confirm the QTLs.

Results

Phenotypic trait evaluation

In 2010, the harvest time of the F1 individuals was distributed over 70 days (August 10 to October 19), that of ‘Akiakari’ was August 26, and that of ‘Taihaku’ was September 10. The distribution of fruit weight was 126–794 g (mean, 359 g). Flesh firmness was 3.70–10.30 lb (mean, 6.57 lb). Total soluble solids content was 12.55–16.65 Brix% (mean, 14.37 Brix%). Acid content was 4.33–5.34 (mean, 4.88). Fruit skin color of F1 individuals ranged from score 1 (as in ‘Taihaku’) to 5 (as in ‘Akiakari’). Preharvest fruit drop ranged from 0% to 100%, with an average value of 16%. Similar distributions were observed for all traits in 2011. All traits except fruit skin color and preharvest fruit drop were normally distributed.

Genetic linkage maps

The genetic linkage map of ‘Akiakari’ contained 208 loci, including 60 pear genomic SSR loci, 66 pear EST-SSR loci, 79 apple SSR loci and three other loci (MdExp7, OPH19, and S locus). Eighteen LGs were identified, which covered a genetic distance of 799 cM with an average marker density of 0.27 loci/cM (Table 3, Fig. 1). All 18 LGs were assigned to the reference genetic map; LG 6 was divided into two LGs, Aki6-1 and Aki6-2. The LG length ranged from 9.8 cM (Aki6-1) to 70.3 cM (Aki15). Large gaps or missing regions were observed in several LGs (Aki1, 4–7, 13, 15, and 17). Many SSR markers in LGs 4, 5, 6, and 13 of the reference linkage maps (Yamamoto ) were homozygous.

Table 3

Details of the genetic linkage maps of ‘Taihaku’(Tai) and ‘Akiakari’ (Aki)

Marker type	Linkage group																		Total
	Aki1	Aki2	Aki3	Aki4	Aki5	Aki6-1	Aki6-2	Aki7	Aki8	Aki9	Aki10	Aki11	Aki12	Aki13	Aki14	Aki15	Aki16	Aki17
Pear genomic SSRs	5	2	13	1	0	2	0	3	6	5	3	4	2	1	5	2	2	4	60
Pear EST-SSRs	3	4	4	2	0	1	1	6	1	4	8	4	4	2	8	6	4	4	66
Apple SSRs	0	4	5	3	3	0	3	4	5	7	14	6	7	1	9	2	2	4	79
Others	1	0	0	0	0	0	0	0	1	0	0	0	0	0	0	0	0	1	3
No. of mapped loci	9	10	22	6	3	3	4	13	13	16	25	14	13	4	22	10	8	13	208
Lengh of linkage groups (cM)	44.4	61.6	53.2	11.6	10.9	9.8	20.4	53.2	46.2	45.2	70.3	66.7	48.6	39.7	51.8	66.4	34.0	65.1	799.1
Marker density (loci/cM)	0.20	0.16	0.41	0.52	0.28	0.31	0.20	0.24	0.28	0.35	0.36	0.21	0.27	0.10	0.42	0.15	0.24	0.20	0.27

Fig. 1

Significant QTLs for seven fruit traits identified in genetic linkage maps of Japanese pears ‘Akiakari’ and ‘Taihaku’. Linkage groups (LGs) are designated as Aki1 to Aki17 for ‘Akiakari’ and as Tai1 to Tai17 for ‘Taihaku’. Marker loci and significant QTLs are shown on the left side of LGs for ‘Akiakari’ and on the right side for ‘Taihaku’. Boxes and range lines indicate 1-LOD and 1.5-LOD support intervals, respectively. Anchored marker loci mapped in both maps are underlined. The self-incompatibility locus in LG Aki 17 is denoted as the S locus. SSR multi-loci derived from the same primers are denoted as SSR name and “-m1/-m2”. Asterisks indicate markers showing distorted segregations in a chi-square test. Distortions at 5%, 1% and 0.1% levels of significance are indicated as *, ** and ***, respectively.

The linkage map of ‘Taihaku’ contained 275 loci, including 85 pear genomic SSR loci, 73 pear EST-SSR loci, 114 apple SSR loci and three other loci (MdExp7, OPH19, and PPACS2). Seventeen LGs were identified, which covered a genetic distance of 1039 cM with an average marker density of 0.27 loci/cM (Table 3, Fig. 1). The length of the LGs ranged from 31.7 cM (Tai10) to 92.5 cM (Tai5). Large gaps or missing regions were observed in three LGs (Tai6, 10, and 17). Because self-incompatible genotypes are S1/S5 for ‘Akiakari’ and S4/S5 for ‘Taihaku’, the S5 allele of ‘Taihaku’ was not inherited in F1 individuals. Therefore, a missing region was found at the bottom of LG Tai17 where the self-incompatibility locus is located. Several chromosomal regions carried markers showing a distorted segregation: LG Tai1, LG Tai12, and the bottom of LG Tai17.

Sixteen significant QTLs were observed for the seven phenotypic traits: four QTLs were detected in both 2010 and 2011 (counted as eight), whereas the other QTLs were detected either in 2010 or 2011 (Table 4, Fig. 1). For harvest time, two QTLs were detected at the bottom of LG Tai3 (BGA35 was the nearest marker) and at the top of LG Tai15 (PPACS2 and MEST050 were the nearest markers). The LOD values of QTLs in LG Tai3 were 5.00 in 2010 and 5.09 in 2011, corresponding to 22.0% and 22.5% of phenotypic variation explained, respectively. QTLs in LG Tai15 showed LOD values of 3.32–3.87 and explained 13.7–15.1% of phenotypic variation. For fruit skin color, one major QTL was found at the top of LG Aki8 near the SSR marker Mdo.chr8.10. The LOD values of the QTL were 8.99 in 2010 and 8.03 in 2011, corresponding to 36.9% and 33.1% of phenotypic variation explained, respectively. For flesh firmness, one QTL was observed at the top of LG Tai4 near TsuENH121-ml. Its LOD values were 3.66 in 2010 and 3.49 in 2011, corresponding to 16.9% and 16.0% of phenotypic variation explained, respectively. For fruit weight, two QTLs were detected. The QTL detected in 2010 was located at the top of LG Aki11, had an LOD value of 3.00, and explained 13.8% of phenotypic variation. The other QTL, detected in 2011, was found at the bottom of LG Tai3; it had an LOD value of 3.83 and explained 17.5% of phenotypic variation. For acid content, one QTL was detected. For total soluble solids content, two QTLs were found. For preharvest fruit drop, three QTLs were detected: in the middle of LG Tai15 (in 2010), in the middle of LG Aki1 (in 2011), and at the top of LG Tai15 (in 2011). The latter QTL overlapped with the QTL for harvest time.

Table 4

QTLs for fruit traits detected in F1 progeny from the ‘Akiakari’ × ‘Taihaku’ cross

Trait	QTL name	Year	Linkage group	Position (cM)	LOD score	Phenotypic variation explained	Nearest marker
Harvest time	HarT-2010-1	2010	Tai3	72.3	5.00	22.0	BGA35
	HarT-2010-2	2010	Tai15	18.0	3.32	15.1	MEST050
	HarT-2011-1	2011	Tai3	73.4	5.09	22.5	BGA35
	HarT-2011-2	2011	Tai15	4.0	3.87	13.7	PPACS2
Fruit skin color	FruC-2010-1	2010	Aki8	2.2	8.99	36.9	Mdo.chr8.10
	FruC-2011-1	2011	Aki8	2.2	8.03	33.1	Mdo.chr8.10
Flesh firmness	FruH-2010-1	2010	Tai4	3.3	3.66	16.9	TsuENH121-m1
	FruH-2011-1	2011	Tai4	3.3	3.49	16.0	TsuENH121-m1
Fruit weight	FruW-2010-1	2010	Aki11	0.0	3.00	13.8	CH04h02
	FruW-2011-1	2011	Tai3	66.2	3.83	17.5	CH01b12-m2
Acid content	Aci-2011-1	2011	Aki14	14.4	4.29	19.3	NH041a
Total soluble solids content	SugC-2010-1	2010	Tai4	22.5	2.95	11.3	CH02h11a
	SugC-2010-2	2010	Tai8	21.0	4.16	19.0	Hi01c11-m1
Preharvest fruit drop	FruD-2010-1	2010	Tai15	33.1	3.28	15.0	TsuENH143
	FruD-2011-1	2011	Aki1	20.8	2.94	13.6	TsuENH174
	FruD-2011-2	2011	Tai15	1.0	5.12	22.4	TsuGNH013

Discussion

Using 93 F1 individuals from a cross between Japanese pear cultivars ‘Akiakari’ and ‘Taihaku’, we constructed SSR marker-based genetic linkage maps of the parental cultivars. The ‘Akiakari’ linkage map consisted of 208 loci and covered 799 cM, whereas the ‘Taihaku’ map consisted of 275 loci and covered 1039 cM. Both linkage maps were assigned to the reference genetic maps of ‘Bartlett’ and ‘La France’ (Yamamoto ). We considered the length of our maps to be sufficient for QTL analysis, because it was similar to those of the previously reported pear maps: ‘Passe Crassane’ (912 cM; Dondini ), ‘Harrow Sweet’ (930 cM; Dondini ), ‘Bartlett’ (1000 cM; Yamamoto ), ‘La France’ (1156 cM; Yamamoto ), ‘Housui’ (1174 cM; Terakami ), ‘Bayuehong’ (1353 cM; Zhang ) and ‘Dangshansuli’ (1044 cM; Zhang ). Missing regions were observed in four LGs (Aki4, 5, 6, and 13) in the linkage map of ‘Akiakari’ and many SSR loci positioned in the LGs were homozygous. Terakami showed that three particular genome regions (LGs 4, 5, and 12) are homozygous in Japanese pear ‘Housui’. Since ‘Akiakari’ was bred by crossing closely related cultivars (Kotobuki ), selection may have increased the number of homozygous genome regions possibly related to fruit traits.

In this study, we identified 16 fruit trait-related QTLs: four QTLs for harvest time, two for fruit skin color, two for flesh firmness, two for fruit weight, one for acid content, two for total soluble solids content, and three for preharvest fruit drop. Two QTLs for maturity date were found in LG BYH8 of ‘Bayuehong’ (Zhang ), which may correspond to LG 11 in the pear reference map (Yamamoto ). Positions of these two QTLs are not consistent with our results, which indicate the presence of QTLs in LGs 3 and 15. Four QTLs related to fruit weight were found in LGs DS2, BYH7, BYH8, and BYH10 (Zhang ), which presumably correspond to LGs 17, 3, 11, and an unknown LG, respectively, in the pear reference map (Yamamoto ). In this study, two QTLs related to fruit weight were detected in LGs Aki11 and Tai3; it will be interesting to find out whether these QTLs correspond to those reported by Zhang . Because genetic linkage maps of ‘Bayuehong’ and ‘Dangshansuli’ were insufficiently anchored to the reference maps of pear and apple, the exact relationships between these QTLs remain unclear.

Some QTLs for fruit-related traits have been reported in apple (Kenis , Kunihisa , Liebhard , Longhi ). Co-linearity observed between pear and apple for all LGs (Pierantoni , Yamamoto ) makes comparison of their fruit-related QTLs a valuable approach. Using F1 individuals from a cross between apple cultivars ‘Fiesta’ and ‘Discovery’, Liebhard found one QTL for harvest date in LG 3, which originated from the early ripening parent ‘Discovery’, and explained 13% of the phenotypic variation. Two major QTLs for fruit acidity were identified in LGs 8 and 16, which explained 33 and 36% of the phenotypic variation, respectively. Several QTLs related to fruit flesh firmness were identified in LGs 3, 6, 11, and 12, and some QTLs related to sugar content were detected in LGs 4 and 8 (Liebhard ). The above data for apple and our current results for pear indicate that QTLs controlling harvest time observed in LG 3 are common for apple and pear, whereas QTLs for fruit acidity, flesh firmness, and sugar content are located in different LGs in the two species. Kenis reported harvest time QTLs in LGs 3, 9, 10, and 16 by using an apple population from a cross between ‘Telamon’ × ‘Braeburn’, whereas Kunihisa identified harvest time QTLs in LGs 3, 10, 15, and 16 by using an apple population from a cross between ‘Orin’ × ‘Akane’. These and our present results strongly suggest the presence of common QTLs for harvest time in LGs 3 and 15 of pear and apple. It would be useful to identify and compare the genes controlling harvest time in pear and apple.

In Japanese pear, fruit storage potential is closely related to the amount of ethylene, which is produced by ACC synthases (Itai ). Two ACC synthase genes (PPACS1 and PPACS2), both located in LG 15, were converted into CAPS markers. Itai pointed out that high ethylene producers tend to mature early, whereas low ethylene producers tend to mature late, which links ethylene production during fruit ripening with the harvest time. A similar tendency was observed for apple cultivars (Abeles 1992). In the present study, a QTL showing significant correlation with harvest time was found at the top of LG Tai15 near the PPACS2 locus. Kunihisa reported that two ACC synthase genes MdACS1 (Harada ) and MdACS3 (Bai ) were mapped in LG 15 in apple and that the harvest time QTL at the bottom of LG 15 around MdACS1 was tightly linked to preharvest fruit drop in apple. We suggest that ACC synthase genes (PPACS2, MdACS1) have multiple functions affecting harvest time, preharvest fruit drop and fruit storage potential both in Japanese pear and in apple. This information will be important for breeding new elite Japanese pear cultivars.

By using bulked segregant analysis of two F1 progenies of Japanese pear cultivars, Inoue identified a 425-bp fragment (OPH-19425, generated by OPH-19 RAPD amplification) linked to a gene controlling fruit skin color. OPH-19425, which showed a recombination rate of 7.3% with green skin trait. In this study, we converted OPH-19425 into a RAPD-STS marker OPH19 and identified a QTL peak for fruit skin color at 2.2 cM in LG Aki8, near the Mdo.chr8.10 locus. Mdo.chr8.12, Mdo.chr8.10, and OPH19 are located at 0, 2.2, and 4.3 cM at the top of pear LG 8, respectively, which corresponds to 551 kb, 3952 kb, and an unknown position on apple chromosome 8. Using hybrids between Japanese and Chinese pears, Song showed that the fruit russet skin trait is linked to apple SSRs CH01c06 located in LG 8 (or LG 2) and Hi20b03 in LG 8 in the apple linkage map (HiDRAS website, www.hidras.unimi.it). Our results confirm that one of the gene controlling fruit russet skin is located in LG 8 of the pear map.

The potential of genome-wide association studies (GWAS) and genomic selection was assessed with a combination of 76 Japanese pear cultivars and 162 markers for nine agronomic traits (Iwata ). Markers showing significant associations with three traits (harvest time, resistance to black spot, and the number of spurs) were detected, suggesting that these markers are linked to major QTLs controlling these traits. In the present study, two markers, BGA35 in LG 3 and PPACS2 in LG 15, showed significant association with harvest time. In this study, we identified two significant QTLs controlling harvest time near markers BGA35 and PPACS2, which is in good accordance with the GWAS results. These results imply that BGA35 and PPACS2 will be routinely applicable for MAS in Japanese pear breeding programs.

Iwata showed that genome-wide predictions for genomic selection were accurate at the highest level (0.75) for harvest time, at a medium level (0.38–0.61) for flesh firmness, fruit size, and acid content, and at a low level (<0.2) for soluble solids content. Iwata proposed genomic prediction of trait segregation in a progeny population on the basis of segregation simulation and Bayesian modeling of genomic selection. Information from bi-parental QTL analysis, GWAS, genomic selection, and genomic prediction and a combination of these techniques would increase the efficiency of future breeding programs in Japanese pear and fruit trees with long juvenility.