Characterization and QTL Analysis of Oryza longistaminata Introgression Line, Plia-1, derived from a Cross between Oryza longistaminata and Oryza sativa (Taichung 65) under Non-fertilized Conditions

To meet and sustain the food demands of an ever-increasing world population, improving the yield of major cereal crops such as rice is necessary with sustainable cultivation harmonized with the environment. It is useful to utilize wild rice species as reservoirs of novel traits for breeding low-input adaptable (LIA) crops. Oryza longistaminata, a wild species of rice native only to Africa, possesses the vigorous biomass needed under low-input conditions. Thus, a potential LIA (pLIA) candidate, pLIA-1, showing large biomass, tall culm, large panicle with many primary and secondary branches and thick culms was selected from a selfed progeny of the cross between O. longistaminata and Taichung 65 (T-65), a japonica variety, under non-fertilized conditions. The pLIA-1 performance was superior to that of Koshihikari, Norin 18, T-65 and Nipponbare under fertilized and non-fertilized conditions suggesting that pLIA-1’s characteristics might be useful for breeding low-input adaptable varieties. QTL analysis in F2 of the cross between pLIA-1 and Norin 18 detected 31 QTLs for yield-elated traits under non-fertilized conditions. The pLIA-1 allele had a positive contribution in 20 of the QTLs detected. Importantly, many of the QTLs were identified around regions where O. longistaminata chromosome segments were introgressed into pLIA-1. These results suggests that the QTLs detected in the F2 are important to improve modern varieties for adaptability to low-input conditions. *Corresponding author: Masahiko Maekawa, Institute of Plant Science and Resources, Okayama University, Japan, Tel: +81864341214; Fax: +81864341249; E-mail: mmaekawa@rib.okayama-u.ac.jp Received June 16, 2016; Accepted August 18, 2016; Published August 30, 2016 Citation: Gichuhi E, Himi E, Takahashi H, Maekawa M (2016) Characterization and QTL Analysis of Oryza longistaminata Introgression Line, pLIA-1, derived from a Cross between Oryza longistaminata and Oryza sativa (Taichung 65) under Nonfertilized Conditions. J Rice Res 4: 174. doi:10.4172/2375-4338.1000174 Copyright: © 2016 Gichuhi E, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.


Introduction
In the near future, food insecurity will be a serious global problem, due to an explosively growing population that is predicted to reach 9 billion by 2050 [1]. Therefore, increasing crop productivity of major cereal crops such as rice; to meet the rising demand of an increasing population is mandatory. Rice is a staple food and it accounts for more than 21% of the calorific needs of the world's population [2].
A major milestone during the green revolution was the development and extensive adoption of semi-dwarf rice cultivars that almost tripled worldwide rice production and greatly enhanced food security [3]. Their high yields were primarily due to their improved harvest index and responsiveness to high inputs, especially nitrogen and water [4]. The extensive usage of semi-dwarf cultivars was, therefore, accompanied by increased inputs of chemical fertilizers and pesticides. This tripled the global grain yield, yet at a high environmental cost. The negative impacts of the current high input agriculture include soil acidification; pollution of rivers, lakes and ground water and the emission of greenhouse gases that may influence global warming [5].
To secure enough food without environmental degradation and excess dependency on inputs, it is necessary to breed new varieties with comparatively high yields and adaptability to low-input conditions. This can be achieved by maximizing crops' ability to produce biomass and tolerance of various abiotic and biotic stresses through the utilization of genetic resources and distantly related wild relatives. Wild species are reservoirs of latent useful traits for domesticated rice improvement [6]. Although many desirable alleles in wild relatives have, however, not yet been fully exploited, several QTLs for yield-related traits have been identified from Oryza rufipogon [6,7] and Oryza glumaepatula [8]. Of the wild species carrying the AA genome in rice, Oryza longistaminata, native only in Africa, is most distantly related to Oryza sativa [9]. It is a perennial species with resistance to bacterial leaf blight [10][11][12][13] and is characterized by strong rhizomes, long anthers and allogamy which are important traits for hybrid seed production [14]. Further, it shows a vigorous biomass under low-input conditions [15,16]. The large biomass of O. longistaminata is considered to be an important trait for breeding low-input adaptable rice. However, its utilization in breeding programs for rice improvement has been very limited, due to developed crossing barriers and hybrid sterility [15,[17][18][19].
We successfully crossed O. longistaminata, with Taichung 65 (T-65), an O. sativa, and successive self-fertilized plants of the progeny were selected based on large biomass production at a non-fertilized paddy field and called potential low-input adaptable (pLIA) rice line. In this study, we report the characterization of the introgression line, pLIA-1, carrying chromosome segments of O. longistaminata and QTLs for some important yield-related traits detected in F2 of the cross between pLIA-1 and Norin 18, a japonica variety, under non-fertilized conditions.

Breeding of the introgression line
Oryza longistaminata, locally known as Mpunga wa Majani (MwM), 18 were genotyped using 35 SSR markers found to be polymorphic between pLIA-1 and Norin 18, out of 111 SSR markers genome-widely distributed. To validate the QTLs identified in the F2 population, 230 F3 plants were genotyped using additional 8 SSR markers in the vicinity of the QTL identified on the distal end of the long arm of chromosome 8. The genetic linkage map was constructed using Mapmaker/Exp version 3.0. Composite interval mapping was performed for QTL analysis using the software Window QTL Cartographer 2.5 [22]. For each of the traits, the LOD threshold was determined at significant probability level of 5% by computing 1000 permutations, in both populations.

Results and Discussion
The breeding process Crossing barriers and hybrid sterility highly developed between O. longistaminata and O. sativa [17] hindered the direct analysis of some important traits in the F2 population. Therefore, many researchers bred backcrossed populations for genetic analyses. Maekawa et al. [23] reported that O. longistaminata from Ethiopia has comparatively good crossing ability with O. sativa and an F2 population could be produced. Iwamoto et al. [24] demonstrated that O. longistaminata from Kenya is closely related to that from Ethiopia, based on the Catalase gene structure. In this study, O. longistaminata from Kenya was successfully crossed with T-65 as a pollen-donor parent and an F2 population was obtained. A single F1 plant of the cross between MwM and T-65 showed low spikelet fertility of 27%. Although 107 of the 169 plants showed extremely high sterility in the F2, a few highly fertile plants segregated, as shown in Figure 1A. In order to obtain large biomass plants derived from the cross, descendants from fertile F2 plants were grown in the non-fertilized field. Thus, one promising plant was selected based on its large biomass production and large panicle. The relationship between the culm length and the number of spikelets per panicle of F4 plants selected from the F3 plant demonstrated that plants carrying large numbers of spikelets per panicle had relatively long culm lengths ( Figure 1B). Furthermore, the segregation of F5 plants was examined and the segregation pattern was found to be similar to that of F4 plants ( Figure 1C). Twelve plants (#1 to #12) showing large numbers of spikelets per panicle were, therefore, selected. These selected plants were characterized by large panicles, large numbers of spikelets per panicle, long culms, many elongated internodes, thick culm base diameters and few tillers, compared with those of Norin 18 ( Figures 1D  and 1E). Although plants carrying large panicles and long culms tend to lodge easily, these selected plants were highly tolerant to lodging due to the very thick culm-base diameter. Of the 12 selected plants, #1 plant with more spikelets per panicle was selectively grown for further self-fertilization at the non-fertilized field. After 11 generations of selffertilization, this line was named potential low-input adaptable (pLIA) line, pLIA-1.

Characterization of the pLIA-1
Recently, it has been reported that most of the semi-dwarf rice varieties have high numbers of unproductive tillers, small panicles and are susceptible to lodging in directly seeded conditions [25]. These traits are identified as the major constraints on improving yields in these varieties; whose yield potential has stagnated due to their plant type. Therefore, the concept of ideal plant architecture (IPA) has been proposed and demonstrated using plants with low tiller numbers with few unproductive tillers, more grains per panicle than in the currently cultivated varieties and thick and sturdy stems [26,27]. As shown in Figure 2, pLIA-1 had a very large number of spikelets per panicle with a large number of primary and secondary branches, compared with collected in a valley 50 km away from Mombasa, Kenya, was crossed with T-65, a japonica variety, as a female parent. A single unmatured crossed seed was then subjected to half-strength MS medium [20] culture. F2 population of 169 plants was grown in a greenhouse at the Institute of Plant Science and Resources (IPSR), Okayama University, Kurashiki, Japan. The progeny from the F3 generation was then grown at a non-fertilized paddy field that had been kept without any fertilizer application for more than 20 years at IPSR. Plants showing a large biomass were selected at the F3 generation at the non-fertilized paddy field. Twelve vigorous plants selected at F5 generation were continuously self-fertilized every year. At the F11 generation, a line showing a large biomass under non-fertilized conditions was bred and called potential low-input adaptable (pLIA) rice line, pLIA-1.
To evaluate the contribution of introgressed segment of O. longistaminata (MwM) into pLIA-1 to yield-related traits, F2 plants and F3 plants derived from the cross between pLIA-1 and Norin 18 were grown under non-fertilized conditions and used for QTL mapping.

Characterization of pLIA-1 and its progeny
The agronomic performance of pLIA-1, under both non-fertilized and fertilized conditions, was evaluated with two replicates and compared to T-65, Norin 18, Nipponbare and Koshihikari in 2012 and T-65 and Norin 18 in 2013. In both experiments, 3-week-old seedlings were transplanted with a spacing of 40 cm between rows and 15 cm between plants. Fifty kg/ha of N, P and K were supplied in fertilized field as basal dressing. The following agronomic traits were measured: the number of panicles per plant, total weight of panicles, total weight of the biomass of the plant shoot, 100-grain weight and grain yield. Culm length, flag leaf length, panicle length, fresh panicle-base diameter, fresh culm-base diameter (culm diameter at 5 cm above ground), number of primary branches per panicle, number of secondary branches per panicle, number of spikelets per panicle and spikelet fertility were measured on the tallest culm of a given plant. Grain length and grain width were measured in 30 grains with 3 replicates. The rough harvest index was calculated as the panicle weight/total plant weight in 2012. Days to heading were from sowing date to the emergence of the first panicle.
F2 and F3 plants derived from the cross between pLIA-1 and Norin 18 were only grown under non-fertilized conditions in 2011 and 2012, respectively. A total of 14 yield-related traits were measured and evaluated.
Analysis of variance (ANOVA) was performed using R software. The Tukey's test (P=0.05) was used to test the differences between means, where significant differences were detected by ANOVA.

Graphical genotyping of pLIA-1
Leaf samples from seedlings of MwM, Norin 18, T-65 and pLIA-1 were collected and dried at 50°C overnight. Genomic DNA was extracted from the leaf tissues using a modified procedure as described by Kawasaki [21]. PCR reaction was prepared by mixing 3.5 μl of distilled water, 20 pmol of 0.5 μl forward primer, 20 pmol of 0.5 μl reverse primer, 5 μl of Quick Taq (TOYOBO, Japan) and 5 ng of the extracted DNA. Amplification was performed as follows: the initial denaturing step at 95°C for 7 min, 30 cycles for 45 sec at 95°C, 30 sec at 55°C and 30 sec at 72°C. Electrophoresis was done in a 3% agarose gel for 90 minutes. The band pattern of the samples was observed in UVlighting after staining with Ethidium bromide.

QTL analysis
The

Graphical genotyping of pLIA-1
In order to utilize pLIA-1's characteristics in breeding programs, introgressed segments in pLIA-1 were needed to be revealed. Tian et al. [28] found that QTLs derived from O. rufipogon introgressed segments in a set of 159 introgressed lines were usually associated with improvement of the target trait (panicles per plant, grains per panicle and filled grains per panicle). In this study, pLIA-1 was found to possess introgressed chromosome segments of O. longistaminata and T-65 by using genome-wide SSR markers. As shown in Figure 5, MwM showed specific polymorphisms in 85 (76.6% frequency) of 111 markers used. Thirteen markers were not amplified in MwM. These non-amplified markers were distributed in 10 chromosomes, except chromosomes 5 and 12 in MwM. Since O. longistaminata is predicted to carry a highly rearranged DNA sequence (Maekawa, unpublished), it is likely that 13 of the markers were not amplified. Additionally, other 13 markers were not polymorphic among the markers used. However, it was observed that pLIA-1 showed MwM-specific band patterns in 18 of 85 markers in 7 chromosomes, except chromosomes 4, 5, 7, 9 and 12 ( Figure 5). Ten markers showed T-65-specific band patterns on chromosomes 1, 4, 5, 7, 8 and 10. Consequently, pLIA-1 was found to carry 20 MwM-specific markers including 2 non-amplified markers on chromosomes 1, 2, 3, 6, 8, 10 and 11 with a frequency of 18.0%. In particular, the short arm of chromosome 6 of O. longistaminata is presumed to be introduced into pLIA-1. Relatively few segments of O. longistaminata were introgressed into pLIA-1 chromosomes. Of 111 genome-1 wide SSR markers used, only 18 O. longistaminata specific markers were observed with a frequency of 16.2%. However, pLIA-1 was shown to exhibit the large biomass characteristic under non-fertilized conditions. This suggests that the large biomass and lowinput tolerance characteristics may be controlled by small segments of the O. longistaminata chromosome segments. On the other hand, novel introgressed chromosome segments observed on chromosomes 1, 5, 8, 9, 10 and 11 may have been caused by changed short sequence repeat numbers through successive self-fertilization or more likely outcrossing during the early stages of the breeding process of pLIA-1.

Correlations between yield-related traits
To reveal the important QTLs for yield-related traits, segregation patterns of the traits and correlations between them were examined in F2 of the cross between pLIA-1 and Norin 18. In most of the traits measured, segregation patterns of normal distribution were observed (Supplementary Figure 1) and transgressive segregations were found in all the traits. In order to understand QTL cluster, correlations among agronomic traits were examined in the F2. It was found that culmbase diameter was significantly positively correlated with panicle traits (panicle length, panicle weight, number of primary branches, number of secondary branches and number of spikelets per panicle) and flag leaf length ( Table 1). The panicle traits were further observed to be positively correlated to each other. The number of spikelets per panicle was strongly correlated to the number of secondary branches (Table  1). Further, significant positive correlations between flag leaf length and number of primary branches, number of secondary branches and number of spikelets per panicle were observed (Table 1).

QTL analysis
To explore the genetic resources from wild rice, several populations  qCL6.1, qPW6.2,  qWG6.1, qDH6.1,  qGW6.1   qCL6.2   qPL6.1, qSB6.1,  qSF6.1, qNSP6.1,  qGW6.2, qWG6.2  between pLIA-1 and Norin 18 ( Table 2 and Figure 5). These included 3 QTLs for culm length, 2 QTLs for panicle length, 1 QTL for number of panicles per plant, 1 QTL for culm-base diameter, 2 QTLs for flag leaf length, 2 QTLs for panicle weight, 1 QTL for number of primary branches per panicle, 4 QTLs for number of secondary branches per panicle, 3 QTLs for number of spikelets per panicle, 1 QTL for spikelet fertility, 3 QTLs for 100 grain weight, 4 QTLs for days to heading and 4 QTLs for grain width ( Table 2). The QTLs were distributed on chromosomes 3, 5, 6, 8, 10 and 11( Figure 5). In 20 of the QTLs identified, pLIA-1 had a positive contribution to the trait. QTLs for strongly correlated traits were observed to be localized near the same region on the chromosome. These clusters of QTLs were observed on chromosomes 3, 5, 6 and 8 ( Figure 5). Further analysis using an F3 population revealed that the region around RM6976 on chromosome 8 carried a crucial QTL cluster for culm-base diameter, panicle-base diameter and number of primary branches ( Figure 5 and Table 3). This is considered to be caused by a strong positive correlation observed between the number of primary branches, number of secondary branches and number of spikelets per panicle. Hence, it is plausible that QTLs for number of primary branches and number of spikelets per panicle were identified in the same QTL cluster in the F2 population. These results strongly suggest that O. longistaminata has great potential for utilization in yield improvement and especially under low-input conditions despite the low spikelet fertility. Among the traits transferred in interspecific crosses using wild rice relatives, spikelet sterility is especially a serious constraint.
A QTL for spikelet fertility was identified on chromosome 6 with a very high LOD score of 30.9 and 60% contribution to the total phenotypic variation ( Table 2), suggesting that this QTL might be the major cause of the low spikelet fertility of pLIA-1. In fact, this QTL was located near the same chromosome region where a QTL for pollen and spikelet fertility was previously identified in a cross using O. longistaminata [29]. QTL clusters of functionally related genes are of great interests in crop improvement. A total of 4 QTL clusters on chromosome 3, 5, 6 and 8 were observed ( Figure 5). Highly significant correlations were also observed between the yield-related traits whose QTLs' were observed to cluster in the same chromosome locations (Table 1). In previous QTL analysis it has been observed that QTL for significantly correlated traits usually had same chromosome location [8,28]. The QTL cluster on chromosome 8 was identified near the same chromosome region where the WFP (Wealthy Farmer's Panicle) was found to be located [30]. This result suggest that these traits are either controlled by strongly linked genes or are as a result of pleiotropism of a single gene locus located at the regions where QTL clusters were observed. Previously, Ookawa et al. [31] reported that the APO1 gene of Habataki on chromosome 6 increased spikelet number together with thicker culm through increased size of inflorescence meristem, hence, the higher spikelet number induced culm thickness pleiotropically. The WFP gene of ST-12 which encodes OsSPL14 (Squamosa Promoter Binding Protein-Like14) drastically increases primary branch number, resulting in increased number of spikelets per panicle [30]. Since in the report, inflorescence meristem of ST-12 was found to be larger than that of Nipponbare, ST-12 was presumed to have thick culms. Thus, it is plausible that the QTL of culm-base diameter was located near the QTL for primary branch number on chromosome 8 based on significantly positive correlation between culm-base diameter and panicle traits observed in this study. Further, there was no correlation between culmbase diameter and culm length in this study suggesting that different culm length plants could be bred with thick culm and larger panicles. Taken together, O. longistaminata is therefore suggested to carry several useful traits under low-input conditions.
The pLIA-1 line reported here is considered to have high potential for low-input adaptability. Identification of QTLs for yield-related traits under non-fertilized conditions further proves its potential for utilization in yield improvement of rice. This line could therefore be  utilized to introgress high productivity under low-input conditions to elite rice varieties. Improvement of nitrogen use efficiency has been proposed as a target for the Second Green Revolution [32]. Hence, "low-input and high output" agriculture is required for sustainability. Further analysis of this line and identification of genes governing the QTLs of agronomic importance is necessary for better understanding of its tolerance to low-input conditions.

Conclusion
In order to utilize O. longistaminata as gene resources, some selected plants which showed vigorous biomass under non-fertilized conditions were developed from the F2 of the cross between MwM, O. longistaminata collected in Kenya and T-65, O. sativa. Out of the selected plants, pLIA-1 showed tolerance to non-fertilized conditions compared to the other varieties, hence was named potential Low-input Adaptable-1. The pLIA-1 was subjected to polymorphic analysis against Norin 18 using SSR markers. Although pLIA-1 carried 20 MwMspecific segments, very important QTLs for panicle-related traits were especially found to be intensively located on the distal region of long arm of chromosome 8 in F2 of the cross between pLIA-1 and Norin 18.