2023 Volume 92 Issue 4 Pages 375-383
The pineapple (Ananas comosus (L.) Merr.) is an economically important tropical fruit crop. In this study, we performed quantitative trait locus (QTL) analysis using 168 individuals of the F1 population of ‘Yugafu’ × ‘Yonekura’ for 15 traits: leaf color (L*, a*, b*), harvest day, crown number, slip number, stem shoot number, sucker number, fruit weight, fruit height, fruit diameter, fruit shell color, soluble solid content, acidity, and ascorbic acid content. The constructed single-nucleotide polymorphism (SNP)-based genetic linkage map consisted of a total genetic distance of 2,595 cM with 3,123 loci, including 22,330 SNPs across 25 chromosomes. QTL analysis detected 13 QTLs for 9 traits: leaf color a*, harvest day, fruit weight, fruit height, fruit diameter, fruit shell color, soluble solid content, acidity, and ascorbic acid content. The causative gene for each QTL was predicted with two genes identified as candidate genes. The AcCCD4 gene on Aco3.3C08 was the predicted causative gene for the shell color QTL, which negatively controls shell color by carotenoid degradation. The Myb domain protein-encoding gene on Aco3.3C02 was the predicted causative gene for shell color and leaf color a* QTL, which positively regulates anthocyanin accumulation. The QTL and gene information provided here contributes to future marker-assisted selection for fruit quality.
The pineapple, Ananas comosus (L.) Merr., is one of the most economically important fruit species worldwide. In the 19th century, the ‘Smooth Cayenne’ cultivar was introduced in Europe from French Guiana and was subsequently distributed to tropical and subtropical regions. Currently, the proportion of ‘Smooth Cayenne’ of international fresh produce market is decreasing, displaced by the ‘MD-2’ variety (Coppens d’Eeckenbrugge et al., 2018). In Japan, 7,000 t of pineapple are produced with approximately 65% of this for fresh fruit consumption (Ogata et al., 2016; Takeuchi, 2022). Improving fruit quality via pineapple breeding is important to increase pineapple production and facilitate fresh fruit consumption in Japan. Pineapple breeding in Japan is primarily performed by hybridization and selection of F1 individuals. Approximately 4,000 pineapple seedlings are obtained from 20 to 30 artificial hybridizations every year. Each seedling is grown over three years, with evaluation of fruit quality and agronomic characteristics. During primary selection, 40–60 promising individuals are selected from 4,000 seedlings. To date, nine cultivars were bred at the Okinawa Prefectural Agricultural Research Center (OPARC) and registered as plant varieties (Ministry of Agriculture, Forestry, and Fisheries).
Owing to the three-year growth period between seedlings and fruit harvesting, pineapple development costs are high. Therefore, cost-effective breeding methods are needed. Marker-assisted selection (MAS) can assist in selecting plants with desirable traits, reducing progeny size and field-costs of raising individuals to maturity (Luby and Shaw, 2001). To perform quantitative locus (QTL) mapping or genome-wide association studies (GWAS), DNA marker development and/or genomic sequence information are required, leading to the development of molecular markers for MAS. DNA marker development and genomic sequence information are necessary to perform these analyses. For the pineapple, DNA marker development and genetic linkage analyses (Carlier et al., 2004, 2006, 2012; Nashima et al., 2020; Shoda et al., 2012; Sousa et al., 2013) are previously reported. And recently, pineapple cultivar genome sequences were determined using next-generation sequencing technologies (Chen et al., 2019; Ming et al., 2015; Nashima et al., 2022; Redwan et al., 2016). For the leaf margin phenotype, bulked restriction-site-associated DNA sequencing (Urasaki et al., 2015), GWAS (Sanewski, 2020, 2022), and QTL analysis (Nashima et al., 2022) were performed, revealing that it is determined by two loci on chromosomes 06 and 23. In addition, Nashima et al. (2022) identified a single locus on chromosome 08 related to fruit flesh color. From these studies, DNA markers for the leaf margin phenotype and fruit flesh color were developed (Nashima et al., 2022; Urasaki et al., 2015). However, many other molecular markers remain to be identified for the pineapple.
To proceed with MAS in pineapples, it is necessary to collect QTL information for various traits. The selection criteria for new cultivars in breeding programs are fruit weight of 1,000–1,500 g, soluble solid content of at least 16% in the juice, and juice acidity of less than 0.8% (Ogata et al., 2016). In addition to fruit quality traits, fruit appearance, postharvest storage properties, and logistical characteristics are important breeding targets. The main target phenotype for fruit appearance is the orange shell color. For postharvest storage, prevention of internal browning is important. Internal browning severely damages fruit quality and is a physiological disorder in which polyphenol oxidization occurs within pineapples during the postharvest storage stage (Zhang et al., 2016). As high ascorbic acid content is suggested to prevent internal browning (Song et al., 2022; Veltman et al., 2000), a high ascorbic acid content would be beneficial to reduce internal browning. For logistical characteristics, fruit size, especially fruit diameter, is a limiting factor when packing carton boxes. Excessively large fruit diameters hinder packaging pineapples into carton boxes. Despite the importance of these traits to pineapple cultivation and sale, their inheritance and QTL information is unreported.
In the present study, we evaluated 15 traits (leaf color (L*, a*, b*), harvest day, crown number, slip number, stem shoot number, sucker number, fruit weight, fruit height, fruit diameter, fruit shell color, soluble solid content, acidity, and ascorbic acid content) of F1 progeny derived from a cross between ‘Yugafu’ and ‘Yonekura’. Both ‘Yugafu’ and ‘Yonekura’ are frequently adopted as crossbreeding parents. In addition, as high-density genome-wide single nucleotide polymorphisms (SNPs) were detected for each F1 progeny previously (Nashima et al., 2022), QTL analysis with a saturated genetic linkage map was performed. In this study, QTLs were detected by analyses using SNPs data obtained from the double-digested restriction-site associated DNA sequence (ddRAD-Seq) analysis.
F1 individuals from ‘Yugafu’ × ‘Yonekura’ were used in this study. ‘Yugafu’ is derived from the parental lines, ‘HI101’ and ‘Cream pineapple’ (Plant variety protection database of the Ministry of Agriculture, Forestry and Fisheries, http://www.hinshu2.maff.go.jp). While ‘Yonekura’ is noted as a clonal selection of cultivar ‘MD-2’ (Nashima et al., 2020). Three slips from each F1 individual were planted in three rows in October 2013. F1 individuals were grown using standard culture techniques, common in Japanese commercial production orchards. To induce inflorescence formation, 30 mL of a solution containing 0.01% 2-chloroethyl phosphonic acid (ethephon) and 3% urea was added in November 2014. The fruit of each plant was harvested in 2015.
For QTL analysis, the mean of three replicates was used. Leaf color (L*, a*, b*) was determined as the mean value of the upper, middle, and lower regions of the maximum leaf length in June 2015. Each value was measured using the NF333 spectrophotometer (Nippon Denshoku, Tokyo, Japan). The crown number, slip number, stem shoot number, and sucker number were determined on harvest day. Fruit weight, fruit height, and fruit diameter were determined for each harvested fruit without crown shoots. Harvest day was defined as the day when 30% of the fruit shell color showed de-greening. Fruit skin color was determined by visual observation and classified into four types according to color: score 1, yellow; score 2, orange-yellow; score 3, yellowish-orange; and score 4, orange. The acid and soluble solid contents of the fresh juice were determined using NH-2000 (HORIBA Advanced Techno, Kyoto, Japan). The soluble solids content value indicates Brix value by reflective index. The acidity value refers to the citric acid equivalent value, as measured by electrical conductivity. The ascorbic acid content of the juice was determined using RQ Flex Plus 10 (Merck, Darmstadt, Germany) with Reflectoquant® ascorbic acid test strips (Merck). Fresh juice, extracted using a manual juicer, was used to measure the soluble solid content, acidity, and ascorbic acid content.
ddRAD-seq data used for SNP genotypingddRAD-Seq data of 168 of F1 individuals of ‘Yugafu’ × ‘Yonekura’ was obtained from the DNA Data Bank of Japan (DDBJ) Sequence Read Archive (DRA) with the accession number DRA007580 (Nashima et al., 2022). In summary, the extracted genomic DNA was digested with the six-base-cut restriction enzyme AseI (New England Biolabs, Beverly, MA, USA) and the four-base cutter NlaIII (New England Biolabs). Then, double-digested DNA was applied to the “End Repair” step in the TruSeq DNA LT Sample Prep Kit (Illumina, Inc., San Diego, CA, USA) protocol. Adaptor-ligated DNA, ranging from 300 to 1,000 bp, was selected and sequenced using a HiSeq2500 (Illumina, Inc.).
SNP genotyping and genetic map constructionThe ddRAD-Seq reads were filtered by trimming low-quality bases (< 10 quality value score) and adapter sequence (AGATCGGAAGAGC) with PRINSEQ (Schmieder and Edwards, 2011) and fastx_clipper from the FASTX-Toolkit <http://hannonlab.cshl.edu/fastx_toolkit/> and mapped to the ‘Cream pineapple’ haplotype sequences of the ‘Yugafu’ pseudomolecule genome (Aco_r3.3C.pmol, Nashima et al., 2022) with Bowtie2 (Langmead and Salzberg, 2012). SNP calling was performed using BCFtools (Li et al., 2009), with high-quality SNPs selected using the following filtering conditions: minimum read depth of 7 (-DP7), minimum SNP quality of 10 (-minQ 10), the maximum proportion of missing data of 0% (-max-missing 1), and no Insertion–deletion mutations (INDELs) with VCFtools (Danecek et al., 2011).
High-quality SNPs were grouped and ordered using Lep-Map3 (Rastas, 2017) to construct a genetic linkage map. Grouping was performed using the SeparateChromosomes2 function with the parameters lodLimit = 16, informativeMask = 123, and sizeLimit = 200. SNP ordering was performed using the OrderMarkers2 function.
QTL analysisThe MapQTL ver. 6.0 software (van Ooijen, 2009) was used for interval mapping. The logarithm of odds (LOD) significance threshold levels were determined using permutation Tests with 1,000 permutations at a significance level of P < 0.05. QTLs with LOD scores exceeding the genome-wide LOD threshold (P < 0.05) were considered statistically significant. The location of each QTL was determined based on the LOD peak location and the surrounding region. Linkage maps and QTL intervals were drawn using the MapChart 2.3 software (Voorrips, 2002). To obtain significant QTLs for harvest day, acid content, soluble solid content, and fruit shell color, which were not normally distributed, interval mapping was performed, followed by the Kruskal-Wallis test to confirm the QTLs.
To determine the haplotype for each F1 progeny, variant call format (VCF) files containing variation information of parental lines, the F1 population, and the output file from the OrderMarkers2 function of lepmap3 were compared. According to the output file from the OrderMarkers2 function of lepmap3, the map2gentypes.awk script (Rastas, 2017) was applied. Subsequently, phase data (1 or 2) for maternal parents on the LOD peak location of each QTL were obtained. Then, the QTL LOD peak or nearest SNPs, which were included in the linkage map, with heterozygous genotype in ‘Yugafu’ and homozygous genotype in ‘Yonekura’ was searched. Then, whether the reference SNP (from the ‘Cream pineapple’ haplotype) or alternative SNP (from the ‘HI101’ haplotype) was inherited was determined for each progeny.
Quantitative Reverse Transcription-polymerase chain reaction (qRT-PCR) analysis of the AcCCD4 geneFruit shells of ‘Yugafu’ and ‘Yonekura’ were used for qRT-PCR. Fruit shells of each cultivar were harvested at six stages (65, 80, 95, 110, 125, and 140 days after bolting) in three replicates. After harvesting, the shells of each sample were immediately frozen in liquid nitrogen and used for total RNA extraction. Total RNA was extracted using RNAiso Plus (Takara Bio, Kusatsu, Japan) with fruit-mate for RNA purification (Takara Bio) according to the manufacturer’s instructions. Total RNA was then subjected to reverse transcription using the PrimeScript RT Reagent Kit with gDNA eraser (Perfect Real Time) (Takara Bio) to obtain cDNA. Gene transcript levels were determined by qRT-PCR using a Thermal Cycler Dice Real-Time System III (Takara Bio). The reactions were performed in a final volume of 20 μL, consisting of 1 μL of diluted cDNA, 0.4 μM primers, and 10 μL of TB Green Premix Ex Taq II (Tli RNaseH Plus; Takara Bio). The PCR conditions were 95°C for 10 s, followed by 40 cycles of 95°C for 10 s, and 57°C for 30 s. The AcCCD4 primers have been described by Nashima et al. (2022). Expression levels were normalized to the 18S rRNA gene (Lv et al., 2012).
The distribution of F1 individuals, normality of data distribution, and parental phenotypes are listed in Table 1. Kolmogorov-Smirnov test shows no normal distribution for crown number, slip number, stem shoot number, sucker number, and fruit shell color, whereas other traits were normally distributed.
Distribution of traits in F1 population of ‘Yugafu’ × ‘Yonekura’.
A total of 8,326,726 SNPs were read using the bowtie2 program. After filtering, 69,327 high-quality SNPs were obtained. After grouping and ordering the high-quality SNPs using Lepmap3, the linkage map with a total genetic distance of 2,595 cM with 3,123 loci including 22,330 SNPs across 25 was obtained (Fig. 1), corresponding to the 25 ‘Cream pineapple’ haplotype pseudomolecule sequences. The average marker distance was 0.83 cM.
Genetic linkage maps of ‘Yugafu’ × ‘Yonekura’ population and significant quantitative trait loci (QTLs) identified. Numbering and orientation of linkage groups are based on the ‘Yugafu’ reference genome (Nashima et al., 2022). Significant QTLs are shown at the side of each linkage group; the 1–LOD (boxes) and 1.5–LOD (range lines) support intervals are shown.
Thirteen significant QTLs, with a LOD score significant at P < 0.05, were identified for the nine traits: one QTL associated with leaf color a*, one with fruit weight, two with fruit height, one with fruit diameter, two with fruit shell color, one with soluble solids content, two with acidity, and two with ascorbic acid content (Table 2; Fig. 1). As both ‘Cream pineapple’ and ‘HI101’ haplotype genome sequences of ‘Yugafu’ were determined by Nashima et al., (2022), inherited haplotype from ‘Yugafu’ in each F1 progeny were determined for each QTLs. Possessing ‘Cream pineapple’ haplotype from ‘Yugafu’ was indicated as (a), ‘HI101’ haplotype from ‘Yugafu’ as (b), one haplotype from ‘Yonekura’ as (c) and another haplotype from ‘Yonekura’ as (d) in Table 2.
QTLs detected in F1 population of ‘Yugafu’ × ‘Yonekura’.
The QTLs for fruit weight and fruit diameter were detected at the same locus on linkage group (LG)22; their relation corresponds to large diameter and high fruit weight. Fruit shell color and leaf color a* were detected at the same locus on LG02; their connection consists of individuals with high values of leaf color a* with a tendency to have an orange shell color. Fruit height and acidity were detected at the same locus on LG05; with the relationship that when fruit height was high, acidity also tended to be high.
Causative gene prediction for the QTLsFor the detected QTLs, the causative genes were predicted from the corresponding genomic region. Among the QTLs, possible candidate genes were detected for harvest days and shell color QTLs. For fruit shell color, one QTL was detected in the terminal region of LG08, which matched a previously reported flesh color locus (Nashima et al., 2022). It is suggested that the white or yellow fruit flesh color of pineapple is determined by AcCCD4, a carotenoid degradation enzyme gene (Nashima et al., 2022). High AcCCD4 expression levels are suggested to lead to decreased carotenoid degradation, resulting in white flesh (Nashima et al., 2022). Thus, the high gene expression levels of AcCCD4 was assumed to also decrease fruit shell carotenoid pigmentation. Therefore, AcCCD4 expression levels in fruit shells were compared between ‘Yugafu’ and ‘Yonekura’. AcCCD4 gene expression increased as the fruit ripened, and the expression level was higher in ‘Yugafu’ than in ‘Yonekura’ in all stages (Fig. 2).
Relative expression level of AcCCD4 in fruit shell. Harvest period is 140 days after bolting. Relative expression levels compared with 140 days after bolting of ‘Yugafu’ were shown. Error bar indicates standard error.
On the 18 Mb region of Aco3.3C02, QTLs were detected for leaf color a* and fruit shell color. When the leaf color a* was high, the leaf epidermis showed red pigment accumulation. As the red coloration of pineapple leaves is due to anthocyanin accumulation, anthocyanin-related genes were predicted to be causative genes for leaf color. In this region, the Myb domain protein-encoding gene (Aco3.3C02g1844.1, Aco3.3C02_17140106–17141679), designated Myb 5 by Zhou et al. (2021), is related to anthocyanin biosynthesis.
In this study, 15 pineapple QTLs were identified. This study is the first time that these QTLs were identified, and may contribute to MAS. The relationship between QTLs and breeding are discussed below.
For harvest day, individuals possessing the ‘Cream pineapple’ haplotype (shown as a genotype in Table 2) and one haplotype of ‘Yonekura’ (d) reached harvest day early. The Okinawa Prefecture is frequently visited by typhoons (intense tropical storms), mainly from August to September, occasionally breaking the pineapple fruit axis. Early harvest days minimize the period that pineapple crops are vulnerable to typhoons. QTL for harvest days is useful for breeding early ripening pineapples, thus avoiding typhoon events.
QTLs for fruit weight and fruit diameter were detected at the same locus on LG22. Individuals with haplotypes (a) and (d) tended to have heavy fruits. Our data suggest that only fruit diameter, and not fruit height, is linked to fruit weight. As large fruit diameter and fruit weight are desirable traits for processing canned pineapples, the QTL on LG22 contributes to pineapple breeding for processing. In shipping and logistics, on the other hand, pineapples of more than 1,500 g are not desirable for fresh consumption as general pineapple shipping carton boxes (40×27×30 cm) are sized to pack 6 pineapples approximately under 1,500 g. As pineapples for processing are not packed into carton boxes for shipping to the manufacturer, the shipping and logistics problems described above do not occur.
For fruit quality traits, soluble solid content, acidity, and ascorbic acid content were measured. For soluble solid content, 16% is the lower limit for the new cultivar in OPARC. QTL detection indicates higher soluble solid content in F1 individuals that inherited the ‘Cream pineapple’ haplotype from ‘Yugafu’ compared to others. Selection of the ‘Cream pineapple’ haplotype in F1 individuals among the population may increase soluble solid content from 14.9% (mean of F1 individuals) to 15.4% (mean of F1 individuals with ‘Cream pineapple’ haplotype). This QTL contributes to enabling an increase in individuals with high soluble solid content.
For acidity, less than 0.8% was the selection criterion for new cultivars. In the ‘Yugafu’ × ‘Yonekura’ population, 80% (136 of 168) of F1 individuals fulfilled the selection criteria. The largest QTL for acidity of LG05, possessing the ‘Cream pineapple’ haplotype from ‘Yugafu’ and d haplotype from ‘Yonekura,’ tended to show low acidities. Although selecting a and/or c haplotypes/s would result in the selection of individuals with low acidity, the contribution to breeding efficiency would not be as high as most F1 individuals showed sufficiently low acidity. However, this would be useful when crossbreeding parents with high acidity levels.
For the ascorbic acid content, two QTLs were detected in LG10 and LG17. Individuals inheriting the ‘Cream pineapple’ haplotype tended to show higher ascorbic acid content than those, which inherited the ‘HI101’ haplotype. Ascorbic acid has been suggested to prevent internal browning in pineapples and other horticultural crops. In European pears, ascorbic acid content declined under browning-inducing conditions, and internal browning was initiated when an ascorbic acid level threshold was passed (Veltman et al., 2000). In Raphanus roots, the occurrence of internal browning is closely correlated with intervarietal differences in the ascorbic acid content (Fukuoka and Enomoto, 2007). In pineapples, a relationship between ascorbic acid and internal is suggested. Exogenous treatment with ascorbic acid has been suggested to function as a buffer solution to scavenge excessive reactive oxygen species and to prevent the spread and deterioration of internal browning in pineapples (Hou et al., 2022). Song et al. (2022) suggested that a decline in internal ascorbic acid content is involved in the development of internal browning. Although it has not been proven that internal varietal differences in ascorbic acid contribute to internal browning, high levels of ascorbic acid scavenges excessive reactive oxygen species, prevent internal browning, and eventually would extend shelf life.
Fruit shell color is an important trait that affects consumers’ willingness to buy. Orange is considered more attractive than yellow. In the present study, the fruit shell color ranged from yellow to orange at the harvest stage. Pineapple fruit shell color is mainly determined by carotenoids, and red shell cultivars have been reported to accumulate anthocyanins (Brat et al., 2004). In this study, fruit shell color QTLs were found on Aco_r3.3C08 and Aco_r3.3C02. Previously, the fruit flesh color locus was identified in the terminal region of Aco_r3.3C08, which was the same locus as the fruit shell QTL. Additionally, the carotenoid degradation enzyme-coding gene, AcCCD4, was the predicted causative gene (Nashima et al., 2022). Individuals possessing the AcCCD4 allele on the ‘Cream pineapple’ haplotype showed a dominant decrease in fruit flesh carotenoid content and white flesh color (Nashima et al., 2022). We believe that AcCCD4 decreased carotenoid content in fruit shells, as well as in fruit flesh, resulting in yellowish and low-carotenoid fruit shell color. Due to higher expression levels of the dominant AcCCD4 allele than the recessive allele for fruit flesh color (Nashima et al., 2022), gene expression levels in fruit shells were expected to be different for shell color. AcCCD4 expression was significantly high in the mature fruit shell of ‘Yugafu,’ which possesses both the dominant and recessive alleles, while low expression levels were observed in ‘Yonekura,’ which possesses only the recessive allele. Therefore, AcCCD4 is the key gene for both, fruit shell color and flesh color, by affecting carotenoid content.
On LG02, the leaf color a* QTL was detected in addition to the shell color QTL. For shell color and leaf color a*, the Myb-encoding gene Aco3.3C02g1844 was considered a candidate QTL gene because of its role in anthocyanin biosynthesis. Although most Myb domain proteins function as transcription factors, their roles in regulatory networks are diverse; including development, metabolism, and responses to biotic and abiotic stresses in plants (Dubos et al., 2010). One reason why the Aco3.3C02g1844 gene could be anthocyanin-related may be its orthologous nature to the Myb domain protein ZmC1 gene in Zea mays, which positively controls anthocyanin accumulation (Paz-Ares et al., 1990; Riaz et al., 2019). Other reasons are anthocyanin-correlated gene expression patterns in the leaves of Ananas comosus var. bracteatus (Zhou et al., 2021). Zhou et al. (2021) suggested that the Myb gene Aco3.3C02g1844 (same as the Myb5 gene in Zhou et al. (2021)) is an anthocyanin-related gene because of its mRNA expression level in green and red leaves. Both QTL analysis in this study and previous transcriptome analysis suggests the Myb gene Aco3.3C02g1844 as a candidate causative gene for anthocyanin accumulation. Although the fruit shell color QTL was located in the same region, it is possible that Myb also affects shell color. For example, the pineapple cultivar ‘FLHORAN41’ has a red shell color due to the accumulation of high levels of anthocyanins and carotenoids (Brat et al., 2004). In this study, according to the genotypes of the F1 population, the genotype for leaf color a* led to a yellowish, low-score shell color. Therefore, the Myb gene Aco3.3C02g1844 did not seem to synchronically accelerate anthocyanin accumulation in shells and leaves. Because there are not enough experimental data to judge the effect of Myb gene Aco3.3C02g1844 on fruit shell coloring, future experiments are required on anthocyanin content, carotenoid content, color values, and Myb gene Aco3.3C02g1844 mRNA expression data in fruit shells.
QTLs for important traits were also identified in the present study. QTLs related to fruit quality (fruit weight, soluble solids content, and acidity), fruit appearance quality (fruit shell color), and postharvest preservation and logistics (fruit diameter and ascorbic acid content) help future breeding with MAS. However, some issues remain unresolved. Phenotypic data were obtained from only one crossbreeding population, without confirmation of QTL applicability to other crossbreeding populations. Because various genetic resources can be used as breeding parents in pineapple breeding, further analysis is needed to identify which allele should be selected from various genetic resources. In addition, our data consisted of only one year of phenotypic data. As phenotypic data for quantitative traits are known to be largely affected by environmental variation, annual differences and false-positive QTLs could be included in ensuing experiments. In the future, QTL analysis using other breeding populations, QTL analysis with annual repetition, GWAS with multiple breeding populations, and genetic resources should be performed to identify reliable QTLs applicable to various crossbreeding populations.
