2017 Volume 86 Issue 2 Pages 200-207
We analyzed the pedigree records (1995–2010) of the Kuchinotsu Citrus Breeding Program (KCBP) at the National Institute of Fruit Tree Science (NIFTS) in Japan, abbreviated as NIFTS-KCBP, to reveal the genetic background and current status of inbreeding and genetic uniformity of the parental cultivars/genotypes and their F1 breeding progenies. The founding genotypes mostly used for crossing in NIFTS-KCBP were satsuma mandarin (Citrus unshiu Marcow.), sweet orange (C. sinensis [L.] Osbeck), king mandarin (C. nobilis Lour.), clementine (C. clementina hort. ex Tanaka), mediterranean mandarin (C. deliciosa Ten.), dancy tangerine (C. tangerina hort. ex Tanaka), and ponkan (C. reticulata Blanco). The intensive use of these seven genotypes and their progenies as crossed parents has led to a high degree of inbreeding in the breeding population. Moreover, these seven genotypes have dominated about 80% of the genetic composition of the breeding population. Although further studies are needed to reveal the influence of inbreeding and genetic uniformity on agronomically important traits, these data offer useful information for the selection of cross combinations and breeding strategies in the ongoing NIFTS citrus breeding program, Japan.
Inbreeding and genetic uniformity in breeding populations are major concerns for crop breeders. In most fruit breeding programs, the number of high-quality genotypes is restricted, and thus they are extensively used as crossed parents (Kajiura and Sato, 1990; Noiton and Alspach, 1996; Scorza et al., 1985). The common practice of using a limited number of genotypes with high quality as crossed parents contributes to the efficient genetic improvement in the initial stages of a breeding program. However, the recurrent use of few genotypes and their progenies can lead to inbreeding depression and genetic uniformity in breeding materials, resulting in the loss of breeding efficiency in the long term. Inbreeding depression downgrades fitness-related traits, such as tree vigor, whereas genetic uniformity causes genetic erosion and decreases the genetic gain per selection. Therefore, both inbreeding and genetic uniformity prohibit the future progress in genetic performance. For the continuous development of improved cultivars that meet market demands, the current status of inbreeding and genetic uniformity in the existing breeding populations needs to be evaluated.
Several studies have evaluated the inbreeding and genetic uniformity in breeding populations of fruit tree crops. For instance, Kajiura and Sato (1990) reported that only 15 cultivars out of the 1212 native varieties are used as crossed parents in Japanese pear breeding; Scorza et al. (1985) demonstrated that the selection for fruit quality in the freestone peach has led to a high degree of inbreeding and genetic uniformity in the eastern United States; and Yamada (1993) and Yamada et al. (1994) found that repeated crosses over generations within the narrow gene pool of persimmon have led to inbreeding depression, which reduces tree vigor, productivity, and fruit weight.
In Japan, the National Citrus Breeding Program is conducted at the National Institute of Fruit Tree Science (NIFTS) Kuchinotsu Citrus Research Station and NIFTS Okitsu Citrus Research Station. The Kuchinotsu Citrus Breeding Program (KCBP) at the NIFTS, which is abbreviated as NIFTS-KCBP, started in 1964 and has been recognized as one of the largest fruit breeding programs in the world (Matsumoto and Takahara, unpublished). NIFTS-KCBP has successfully developed 21 novel citrus cultivars such as ‘Shiranuhi’ (Matsumoto, 2001) and ‘Setoka’ (Matsumoto et al., 2003), which are widely grown in Japan.
Since the 1970s, fruit quality traits, including high sugar content, easy peeling, seedlessness, soft pulp firmness, and soft segment firmness, have been the main breeding targets of NIFTS-KCBP. These breeding objectives narrowed down the candidate parental cultivars, and thus may lead to inbreeding depression and genetic uniformity in the breeding populations. However, the genetic background, as well as the degree of inbreeding and genetic uniformity in NIFTS-KCBP populations, have not yet been investigated.
The objectives of this study were to reveal the genetic background and current status of inbreeding and genetic uniformity in NIFTS-KCBP parental cultivars/genotypes (hereinafter referred to as parental cultivars) and their F1 breeding progenies using their pedigree information. These estimates can offer useful information for the future selection of parental cultivars or plants for cross combinations and breeding strategies.
Parental cultivars and their F1 breeding progenies obtained in NIFTS-KCBP from 1995 to 2010 were considered in this study. Seventy-three seed parents and 69 pollen parents, a total of 126 parental cultivars, produced 12541 F1 progenies from 466 cross combinations in NIFTS-KCBP during the period. About 1000 F1 progenies from 20–40 cross combinations were planted in breeding fields in each year from 1995 through 2010. From these 12541 F1 progenies, we selected a prominent early-maturated cultivar ‘Mihaya’ (Nonaka et al., 2012), and two prominent genotypes were selected and are evaluated in a national trial as of June 2016. They were also used as parental cultivars, and were included in the 126 parental cultivars in this study.
Genetic backgroundThe pedigree records of 126 parental cultivars and their 12541 F1 breeding progenies that contained the names of individual genotypes and their seed and pollen parents were obtained from NIFTS-KCBP. The pedigrees of all genotypes were traced back to genotypes with unknown parents, which were regarded as founding genotypes, and all the founding genotypes and ancestral genotypes were added to the pedigree records to obtain the full pedigree charts that were constructed using Pedimap (Voorrips et al., 2012). Such full pedigree information was used to reveal the genetic background and founding genotypes of the parental cultivars and their F1 progenies.
Inbreeding and genetic uniformityThe inbreeding coefficients of 126 parental cultivars and their 12541 F1 progenies, as well as the coefficients of co-ancestry between them and their founding genotypes, were computed based on the assumption that the founding genotypes were genetically unrelated and non-inbred. The inbreeding coefficient was defined as the probability that two alleles of an individual at a given locus were identical by descent (IBD) and reflected the genetic similarity of a mating pair to generate the individual. To investigate the changes in the inbreeding coefficients of F1 progenies, we evaluated the mean inbreeding coefficients of F1 progenies obtained in each year from 1995 through 2010. Coefficient of co-ancestry between two individuals was defined as the probability that the two alleles sampled from respective individuals at a given locus were IBD. Thus, the coefficients of co-ancestry between parental cultivars or F1 progenies and the founding genotypes reflected the genome composition of a breeding population, each component of which was originated from the genomes of the founding genotypes. We calculated twice the coefficient of co-ancestry as a numerator relationship coefficient between the parental cultivars or F1 progenies and the founding genotypes, which can be interpreted as the genetic contributions of founding genotypes to individuals in a breeding population (Bulmer, 1980).
Inbreeding coefficients and coefficients of co-ancestry were calculated using Peditree (van Berloo and Hutten, 2005) based on the assumption that the founding genotypes were unrelated and non-inbred, and that breeding selections had no influence on the genetic contribution of descendant genotypes. In addition, all mutants included in the breeding population were regarded to have the same genetic composition as the original genotypes.
The pedigree chart of 126 parental cultivars used for crossing between 1995 and 2010 along with their founding genotypes is shown in Figure 1. The 126 parental cultivars were traced back to 21 founding genotypes, which were as follows: satsuma mandarin (Citrus unshiu Marcow.), clementine (C. clementina hort. ex Tanaka), sweet orange (C. sinensis [L.] Osbeck), iyo (C. iyo hort. ex Tanaka), hyuga-natsu (C. tamurana hort. ex Tanaka), grapefruit (C. paradisi Macfad.), dancy tangerine (C. tangerina hort. ex Tanaka), king mandarin (C. nobilis Lour.), mediterranean mandarin (C. deliciosa Ten.), ponkan (C. reticulata Blanco), ‘Mukaku-kishu’ (C. kinokuni hort. ex Tanaka), hassaku (C. hassaku hort. ex Tanaka), ‘Tanikawa-buntan’, ‘Ogon-kan (C. flavicarpa hort. ex Tanaka)’, ‘Kawachi-bankan’, ‘San-jacinto’ tangelo, ‘Soren-tangelo’, tankan (C. tankan Hayata), ‘Haruka’, yuge-hyokan (C. yuge-hyokan hort. ex Yu. Tanaka), and ‘Murcott’ tangor. We generated 12541 F1 progenies from crosses with the 126 parental cultivars in NIFTS-KCBP between 1995 and 2010 as described in Materials and Methods; therefore, these 21 founding genotypes composed the genetic background of NIFTS-KCBP parental cultivars and their F1 breeding progenies.
Pedigree chart of parental cultivars used in the Kuchinotsu Citrus Breeding Program at the National Institute of Fruit Tree Science (NIFTS-KCBP) from 1995 to 2010. Pedigrees are oriented in generations from left to right. The founding genotypes mostly used for crossing in NIFTS-KCBP are shown in the green box: satsuma mandarin, sweet orange, king mandarin, clementine, mediterranean mandarin, dancy tangerine, and ponkan. Genotypes in the yellow box represent the other 14 founding genotypes. Red lines represent seed parents. Blue lines represent pollen parents. Cross symbols connect parents to offspring.
Inbreeding coefficients ranged from 0 to 0.250 with the mean and standard deviation of 0.029 ± 0.062 in the 126 parental cultivars (Table 1), of which 30 were estimated to have inbreeding coefficients greater than 0, including 13 parental cultivars with inbreeding coefficients higher than 0.10. Parental cultivars released in later generations tended to have higher inbreeding coefficients, indicating the upward trend in the inbreeding coefficient with the progress of NIFTS-KCBP.
Inbreeding coefficient and genome composition of parental cultivars in the Kuchinotsu Citrus Breeding Program at the National Institute of Fruit Tree Science (NIFTS-KCBP) from 1995 to 2010.
Continued
In F1 progenies, inbreeding coefficients ranged from 0 to 0.500 (data not shown). The mean values of inbreeding coefficients in F1 progenies obtained in each year were relatively high and ranged from 0.076 to 0.169 in 1995–2010 (Table 2). The grand mean value of inbreeding coefficients in all F1 progenies was 0.116 (± 0.079).
Inbreeding coefficient and genome composition of F1 breeding progenies in the Kuchinotsu Citrus Breeding Program at the National Institute of Fruit Tree Science (NIFTS-KCBP) from 1995 to 2010.
Of the 21 founding genotypes, satsuma mandarin, sweet orange, king mandarin, clementine, mediterranean mandarin, dancy tangerine, and ponkan were extensively used as predominant parents for crossing in NIFTS-KCBP and dominated 77.1% of the genetic composition of parental cultivars. Satsuma mandarin showed the highest mean genetic contribution (25.7%), followed by sweet orange (17.3%), king mandarin (8.1%), clementine (7.9%), mediterranean mandarin (7.3%), dancy tangerine (5.8%), and ponkan (4.9%) in terms of numerator relationship coefficients (Table 1).
In F1 progenies, satsuma mandarin showed the highest mean genetic contribution (28.5%–45.2%) in every year of crossing (Table 2). Although the rank between the other six predominant founding genotypes slightly changed through the years, they steadily dominated a total of 54.5% of the genetic composition of F1 progenies. In total, the seven predominant founding genotypes dominated 88.7% of the genetic composition of F1 progenies. Therefore, an apparent genetic uniformity was observed in both the parental cultivars and F1 progenies of NIFTS-KCBP.
The evaluation of inbreeding and genetic uniformity is essential for ongoing fruit tree breeding programs. This study investigated the genetic background and current status of inbreeding and genetic uniformity in the NIFTS-KCBP parental cultivars and their F1 breeding progenies and revealed a relatively narrow genetic background and high levels of inbreeding and genetic uniformity.
We used the pedigree information to calculate inbreeding and co-ancestry coefficients. However, pedigree information may not be an accurate predictor of inbreeding and genetic contribution in artificially selected populations (Culp, 1998). Our parental cultivars were selected as the superior genotypes in NIFTS-KCBP and other citrus breeding programs, and thus the estimates of inbreeding and genetic uniformity may include some degree of bias. Additionally, a recent study using molecular markers indicated high probabilities of incorrect parentage for several parental cultivars (Ninomiya et al., 2015). Molecular markers also indicated genetic relationships between the founding genotypes, which were assumed to be genetically unrelated in this study; for instance, clementine is a hybrid between mediterranean mandarin and sweet orange (Ollitrault et al., 2012), and satsuma mandarin is a putative hybrid between ‘Kunenbo’ (C. nobilis Lour) and kishu-mikan (C. kinokuni hort. ex Tanaka) (Shimizu et al., 2016). Although these issues may slightly bias the estimates, our study provided a practical evaluation of the current status of inbreeding and genetic uniformity in NIFTS-KCBP parental cultivars and their F1 breeding progenies.
The intensive use of the seven predominant founding genotypes and their descendants as parental cultivars in NIFTS-KCBP could eventually result in the loss of genetic diversity, which could limit the genetic improvement and development of new citrus cultivars in the near future. A narrow genetic background and high degree of genetic uniformity were also observed in other fruit tree crops, such as the Japanese pear (Kajiura and Sato, 1990), apple (Bannier, 2011; Noiton and Alspach, 1996; Son et al., 2012), freestone peach (Scorza et al., 1985), almond (Lansari et al., 1994), and sweet cherry (Choi and Kappel, 2004), which could lead to the loss of potential gain from selection and genetic vulnerabilities to epidemic insects, diseases, and stress conditions. In NIFTS-KCBP, the influence of a high degree of genetic uniformity on agronomic traits has not yet been investigated in detail. However, some citrus industries have pointed out that the fruit appearance of some newly developed cultivars is fairly similar to the existing cultivars, probably due to the high degree of genetic uniformity in the breeding materials (Takahara, Personal communication). The uniform appearance is a critical disadvantage for citrus marketing; therefore, we need to elucidate the influence of a high degree of genetic uniformity on agronomic traits using our accumulated phenotypic records.
The mean inbreeding coefficient in parental cultivars (0.029 ± 0.062) was as low as that in the sweet cherry (0.033) (Choi and Kappel, 2004), apple (0.01–0.04) (Noiton and Alspach, 1996), Japanese-type plum (0.016–0.054) (Byrne, 1989), and almond (0.022) (Lansari et al., 1994), in which inbreeding depression has a negligible influence on agronomic traits. However, our results showed that the inbreeding coefficients of 13 parental cultivars in advanced generations exceeded 0.10, revealing an upward trend with the progress of NIFTS-KCBP. Additionally, the grand mean of the inbreeding coefficient in F1 progenies was as high (0.116) as that in the blueberry (0.13) (Hancock and Siefker, 1982) and raspberry (0.12) (Dale et al., 1993): These values were close to 0.125, which is expected in half-sib mating, suggesting the potential threat of inbreeding depression. Although the susceptibility to inbreeding is different among crops, careful consideration to inbreeding depression may have to be given regarding the breeding materials of NIFTS-KCBP. The influence of inbreeding on citrus is not well known, and in the breeding materials of NIFTS-KCBP, associations between inbreeding coefficients and agronomic traits are also unclear at present. However, Frost (1943) reported that crosses of ‘Ruby’ × ‘Valencia’ oranges and ‘Eureka’ × ‘Lisbon’ lemons produced progenies with a weak tree vigor. These two crosses are extreme examples because they produced progenies with an inbreeding coefficient of 0.5. However, they indicate that inbreeding depression can occur in citrus and that the sweet orange has genetic factors that cause inbreeding depression. Sweet orange was the second most frequently used founding genotype in NIFTS-KCBP, and as a result, its influence of inbreeding on agronomic traits, including tree vigor, needs to be taken into serious consideration if the upward inbreeding coefficient trend continues in NIFTS-KCBP.
Although the NIFTS germplasm collection includes more than 3000 citrus accessions, composed of local varieties, wild species, and related species from all around the world, the genetic background of parental cultivars and their F1 breeding progenies is very limited. It is, therefore, necessary to broaden the genetic base of cross breeding and use diverse citrus accessions as parents to enable future NIFTS citrus breeding program to stably progress.
We would like to thank all the parties that support NIFTS-KCBP, Japan.
