Genetic Performance of the Semidwarfing Allele sd1 Derived from a Japonica Rice Cultivar and Minimum Requirements to Detect Its Single-Nucleotide Polymorphism by MiSeq Whole-Genome Sequencing

The influence of the semidwarfing gene sd1 derived from the rice cultivar Jukkoku (Jukkoku_sd1) and IR8 (IR8_sd1), which contributed to the Green Revolution, d60 from Hokuriku 100, as well as the combination of sd1 and d60 (Jukkoku_sd1 plus d60 and IR8_sd1 plus d60), was investigated using isogenic lines raised by backcrossing with the cultivar Koshihikari. The isogenic lines carrying Jukkoku_sd1, IR8_sd1, d60, Jukkoku_sd1 plus d60, and IR8_sd1 plus d60 had considerably shorter culm lengths than Koshihikari by 19.2%, 22.8%, 26.0%, 45.1%, and 43.4%, respectively. The sd1 plus d60 lines showed additively reduced culms, indicating that the function of d60 was different from sd1. In contrast to the culm reduction, Jukkoku_sd1 showed productive merit with a panicle length of 2.5% greater than the origin. MiSeq next-generation sequencer was used to optimize a minimum scale to detect Jukkoku_sd1 in practical breeding. Mapping with the reference genome of Nipponbare gained the average depths of Koshihikari Jukkoku_sd1 and Koshihikari being 9.17 and 7.29, respectively. Comparing the vcf files of the entire genomes of Koshihikari Jukkoku_sd1 and the virtual Koshihikari revealed a G to T SNP at position 38,382,746 in the sd1 locus on chromosome 1 of Koshihikari, causing a loss-of-function mutation of GA20-oxidase.


Introduction
Semidwarfism of rice improves phenotype (light-interception properties and harvest index), nitrogen responsiveness, and lodging resistance and is thus an important trait worldwide. The International Rice Research Institute produced a cross between Peta, a tall indica variety (culm length, 150-180 cm), characterized by abundant, long hanging leaves, commonly grown in tropical Asia, and Dee-geo-woo-gen (DGWG), a Taiwanese indigenous semidwarf variety, to improve its lodging resistance and light-interception properties. The resulting semidwarf rice variety IR8 (culm length, 90-100 cm), developed in 1966, has dramatically improved rice yields and brought the Green Revolution to tropical Asia [1].
The semidwarf trait has also been introduced into rice cultivars grown in other countries. In Japan, Hoyoku, Shiranui, and Reihou, which are representative main cultivars grown in the Kyushu region, were developed using the indigenous semidwarf cultivar Jukkoku [2]. A series of cultivars, such as Akihikari and Niigata Wase, were developed in the 1970s in the Tohoku region using the semidwarf cultivar Reimei, which was induced by gammaray irradiation of Fujiminori [3]. In the USA, Calrose 76 (culm length, ∼90 cm) was developed in 1976 also by gammaray irradiation of the japonica variety Calrose (culm length, ∼120 cm) [4][5][6]. In Korea, d47 derived from IR8 was also introduced into Tongil by japonica-indica hybridization [7].
to be allelic to d47 [4,11,12]. Moreover, allelic relationships of the semidwarf gene have been found in Taichung Native 1, derived from DGWG; Shiranui, derived from Jukkoku; and d49 in the mutant cultivar Reimei [13,14]. Finally, all these semidwarf cultivars carry alleles at the sd1 locus, despite their different parentage [14][15][16][17][18]. The sd1 alleles, on the long arm of chromosome 1, encode loss-of-function mutations in GA20-oxidase (OsGA20ox2), which regulates the synthesis of biologically active gibberellins (GAs), which catalyze three steps in the GA biosynthesis pathway [15][16][17][18]. Taken together, a single semidwarf gene, sd1, solely confers the semidwarf phenotype of cultivars commonly grown around the world due to its nondetrimental effects on grain yield [19]. This narrow genetic base of current semidwarf rice cultivars has led to reduced genetic diversity in rice [20][21][22]. Thus, it is necessary to identify a novel semidwarf gene alternative to sd1 and to utilize it to extend genetic diversity in semidwarf cultivars. A novel dwarf gene, d60, which was found in the semidwarf mutant Hokuriku 100, developed by exposing Koshihikari to 20 kR of gamma radiation [23,24], is thus of particular importance. Unlike sd1, which is inherited as a single recessive gene, d60 complements the gametic lethal gene gal, which is carried by many rice varieties. Consequently, the cross between Hokuriku 100 and Koshihikari exhibits a unique genotype ratio of 4D60D60 : 4D60d60 : 1d60d60 [23][24][25].
Although many semidwarf genes allelic to sd1 have been identified in different cultivars (DGWG, Jukkoku, Reimei, and Calrose 76), differences in their influences have not been elucidated. Thus, investigating the differences in ecological and phenotypic traits in relation to yield between d60-and sd1-carrying plants, as well as those among sd1 allelic variants of different origins, will be beneficial for future use of d60 and multiple allelic variants of sd1. The first author developed isogenic semidwarf lines by continuously backcrossing sd1 of Jukkoku (Jukkoku sd1) and sd1 of Kinuhikari (IR8 sd1) and d60 and both genes into the Koshihikari background and maintained the lines to investigate the resulting traits. Hence, the study was conducted with the following specific objectives: (a) investigate and quantify the differences between d60-and sd1-carrying NILs, in relation to yield, to assess the utility of sd1 and d60 semidwarf genes and (b) sequence NILs by MiSeq to detect the differences between Sd1/sd1 target sites. Figure 1, the following rice cultivars were examined: Koshihikari, Koshihikari Jukkoku sd1 (Koshihikari * 6//Koshihikari/Jukkoku B 6 F 4 ), Koshihikari IR8 sd1 (Koshihikari * 3//Koshihikari/ Kinuhikari B 3 F 3 , Kinuhikari carrying sd1 of IR8 [26]), Koshihikari d60 (Koshihikari * 3//Koshihikari/Hokuriku 100 B 3 F 3 ), Koshihikari Jukkoku sd1 plus d60 (Koshihikari Jukkoku sd1/Koshihikari d60), and Koshihikari IR8 sd1 plus d60 (Koshihikari IR8 sd1/Koshihikari d60). In each backcross experiment, short-culm BC F 2 plants were backcrossed with Koshihikari as the recurrent female parent in each generation. Each semidwarf phenotype was fixed by the BC F 3 generation. The percentage of genetic materials from Koshihikari in the isogenic semidwarf lines was calculated as follows: For Koshihikari carrying Jukkoku sd1

Cultivation.
Rice seeds were collected from stocks kept in a refrigerator. Seeds of each line were immersed in just enough water to cover them. Water was changed every day for 7 days (May 2 to May 8) during germination stimulation. Seedlings were grown in special boxes (30 × 15 × 3 cm) for approximately 20 days in a greenhouse: two seeds were planted in each cell (2 × 2 × 3 cm) in the box on May 11 and 12 and watered twice a day (07:00 and 12:00). Seedlings were then individually transplanted into a paddy field (120 m 2 : 6.0 × 20.0 m) of the University Farm on June 8. The paddy field received 4.0 kg of basal fertilizer containing nitrogen, phosphorus, and potassium (weight ratio, nitrogen : phosphorus : potassium = 2.6 : 3.2 : 2.6). Except for the period of mid-season drainage (July 10 to July 17), the water level was maintained at 5-7 cm above ground to prevent seedlings from submersing. A herbicide (Joystar L floable) was applied on June 20 to kill weeds growing uncontrollably and the water was then kept at a high enough level to cover the weeds for 1 week. Koshihikari, Koshihikari carrying Jukkoku sd1, Koshihikari carrying IR8 sd1, Koshihikari carrying d60, Koshihikari carrying d60 and Jukkoku sd1, and Koshihikari carrying d60 and IR8 sd1 were grown in 4 m 2 plots (2 × 2 m) (two instances per line). Basal fertilizer was applied to give 0.43 g/m 2 nitrogen, 0.53 g/m 2 phosphorus, and 0.53 g/m 2 potassium. After ripening, 10 plants per genotype were sampled twice to assess the following traits: panicle heading time, culm length, internode interval, and panicle length.

Investigation of Agronomic Traits.
The time when the tip of the panicle first emerged from the flag leaf sheath was recorded as the heading time for all plants. Ten plants typical of each line were selected. The sampling procedure was performed twice. Sampled plants were air-dried and the following phenotypic traits were assessed or measured: (1) culm length (the length between the ground surface and the panicle end of the main culm), (2) internode length (the lengths between two neighboring internodes of the upper five internodes), and (3) panicle length (the length between the panicle base and the tip of the panicle). Differences in phenotypic traits between each line carrying a semidwarf gene (or genes) and Koshihikari were obtained using the following equation: percent difference = [(measurements of each line) − (measurements of Koshihikari)]/(measurements of Koshihikari) × 100.

Next-Generation Sequencing (NGS)
Analysis. An advantage of genomics is the development of a next-generation sequencer that can decode DNA sequences at the giga level. Development of the next-generation DNA sequencer was advanced under the Affordable Care Act aims to realize societal implementation of medical genomics [27,28]. The desktop-type next-generation sequencer MiSeq has the ability to read 15 million DNA sequences in one run. Generally, whole-genome sequencing analysis required 30 times of the given genome size (30× genome coverage). According to this standard scale, the MiSeq treats only one rice sample per run, because the 15 million DNA amount is just enough to reconstruct a single rice genome. However, it is a big problem that the running cost is excessively high to use the MiSeq at the practical breeding to detect target genes. So, in this study, to achieve a minimum scale to detect target genes with a reasonable cost, we try to detect Jukkoku sd1 at only 5× rice genome coverage by using the MiSeq. The semidwarf gene sd1 (a loss-of-function mutation of the GA20-oxidase encoding gene) was transferred to Koshihikari by consecutive backcrosses to prepare a semidwarf Koshihikari named Koshihikari Jukkoku sd1. The Koshihikari Jukkoku sd1 backcross was used to detect single-nucleotide polymorphisms (SNPs) in Jukkoku-derived sd1 by NGS. Whole-genome analysis was conducted using Koshihikari Jukkoku sd1 and Koshihikari. Genomic DNA was extracted from each cultivar by the CTAB (hexadecyltrimethylammonium bromide) method. Genomic DNA was tagged and fragmented to an average length of 500 bp using the Nextera5 transposome-based approach. After purification of the transposome with DNA Clean & Concentrator6-5 (Zymo Research, Irvine, CA, USA), adaptors for fixation on the flow cell were synthesized at both ends of each fragment using a polymerase chain reaction (PCR). Then, the DNA fragments were subjected to size selection using AMPure XP magnetic beads (Beckman Coulter, Inc., Brea, CA, USA). Finally, qualitative and quantitative measurements were performed using a Fragment Analyzer6 (Advanced Analytical Technologies, Inc., Ankeny, IA, USA) and a Qubit5 2.0 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA) to prepare a DNA library for NGS. Aiming to achieve 5× rice genome coverage, a MiSeq next-generation sequencer was used to simultaneously analyze five lines; namely, 4-5 ng of five DNA libraries was applied in each MiSeq run. Clusters then were formed on the flow cells by bridge-PCR and each pair-end of a 250 bp read was sequenced. Resulting sequenced reads were mapped using BWA software with the Nipponbare genome [29] as a reference. Then, SNPs and Indels were detected using SAMtools software (http://samtools.sourceforge.net/).

Effects of Semidwarf Genes sd1 and d60 on Heading Time.
The days to heading were compared (Figure 2). The earliness of varieties of Koshihikari and donors of semidwarf genes is as follows: Koshihikari, an early-medium maturing variety; Jukkoku, a medium-late maturing variety; Kinuhikari, an early-medium maturing variety; and Hokuriku 100, an early-medium maturing variety. The heading panicles were observed first in Koshihikari carrying IR8 sd1 (83 days to heading) and the latest heading panicles were observed in Koshihikari carrying Jukkoku sd1 or Jukkoku sd1 plus d60 (93 days to heading). However, the difference in the average number of days to heading was the largest between lines carrying IR8 sd1 (86.5 days) and those carrying d60 (90.5 days), but this 4-day difference was thought to be minor ( Figure 2). Thus, it was concluded that the time required for maturing was comparable among lines and the differences  in traits, such as panicle length, were attributed to genetic factors.  third and fourth and between the fourth and fifth internodes was markedly large in lines carrying IR8 sd1 (16.4%, 31.5%, 37.0%, 42.6%, and 45.0%: percent difference, from upper to lower internode intervals). Reductions in internode intervals were relatively uniform in lines carrying d60 (24.2%, 32.6%, 27.4%, 34.5%, and 45.7%: percent difference, from upper to lower internode intervals), while percent differences were larger than those observed in lines carrying Jukkoku sd1. In the sd1 plus d60 lines, marked reductions were observed in length between neighboring internodes, probably attributed to the additive effect of sd1 and d60 (Table 1 and Figure 3). The interval between the fourth and fifth internodes often disappeared in plants carrying IR8-sd1 and in sd1 plus d60 lines, which showed a marked reduction in length between the lower internodes.

Effects of Semidwarf Genes sd1 and d60 on Traits Related to
In contrast, as shown in Figure 4, the effect of a semidwarf gene (or genes) on panicle length (−3.0 to +2.5%), which  was smaller than that for internode intervals (−16 to −92%), was negligible in practical agriculture. Particularly, the mean length of panicles in the line carrying Jukkoku sd1 was 16.2 cm, which was solely 2.5% longer than that of Koshihikari. The panicle length of the line carrying Jukkoku sd1 was longer than that of the original cultivar Koshihikari: not only the mean value but also the value including the standard deviation caused by circumstance. Therefore, there was certain merit in increasing the production of Jukkoku sd1 as compared to other semidwarf genes.

NGS Analysis to Detect Jukkoku-Derived sd1 with a
Minimum Scale. Using the MiSeq sequencer, we obtained a total read length of 2.79 × 10 9 bp from the total read number of 9.53 × 10 6 in Koshihikari Jukkoku sd1, while a total read length of 1.92 × 10 9 bp was obtained from a total read number of 6.13 × 10 6 in Koshihikari. By mapping the read sequences obtained by NGS using Nihonbare genomic DNA as a reference, sequence coverage rates of 93.5% and 88.5% were attained for Koshihikari Jukkoku sd1 and Koshihikari, respectively, while average depths were 9.17 and 7.29, respectively. Furthermore, vcf files of the entire genomes were prepared and the whole-genome sequence of Koshihikari Jukkoku sd1 was compared with that of the virtual Koshihikari genome. As a result, three reads were obtained, including Jukkoku sd1 from Koshihikari Jukkoku sd1, while three reads of Sd1 were obtained from Koshihikari. The Sd1/sd1 locus (Os01t0883800-01) was localized at positions 38,382,385-38,385,469 from the end of the short arm of chromosome 1 in the Koshihikari genome. The difference observed between the Koshihikari Sd1/Jukkoku sd1 alleles was only one SNP from G to T in exon 1 of the GA20oxidase gene ( Figure 5), as reported by Sasaki et al. [17]. In this study, the G to T SNP of the defective GA20-oxidase gene was localized at position 38,382,746 from the end of the short arm of chromosome 1 of Koshihikari. Using this scale of the NGS approach with a coverage of only 5×, the targeted SNP in Jukkoku sd1 was successfully identified with three reads, with a cost reduced to 1/5 that of the ordinal whole-genome sequencing with a coverage of 30×.

Discussion
Environmental degradation caused by global warming, postearthquake salt damage, and radioactive contamination and globalization of agricultural markets due to the Trans-Pacific Partnership are serious issues that call for the innovation of new cultivars to overcome the shortcomings of current crops. The results of this study showed that all tested semidwarf lines had shorter culm lengths than Koshihikari, indicating improved lodging resistance. Furthermore, the leaves were straighter (pointing upwards) in the semidwarf lines than in Koshihikari (Figure 1), indicating improved light-interception properties attributed to the introduction of semidwarf gene(s). Panicle length was solely longer by 2.5% in Koshihikari carrying Jukkoku sd1 and shorter in lines carrying IR8 sd1 by 2.4% or d60 by 3.0%, as compared with the original cultivar Koshihikari (Table 1 and Figure 4). However, the reduction in panicle length was considerably less than that in culm length (Table 1 and Figure 3; a 22.8% decrease in lines carrying IR8 sd1 versus a 26.1% decrease in lines carrying d60), suggesting that the negative effects of the semidwarf genes sd1 and d60 on panicle length were negligible (Table 1 and Figure 4). Ogi et al. [30] and Murai et al. [31] reported that an isogenic line carrying sd1 was derived from DGWG, which was constructed with the genetic background of Norin 29 and Shiokari, respectively. However, IR8-derived sd1 did not show effects of increasing panicle length, which was also observed with the genetic background of Koshihikari in this study. On the other hand, the results showed that Jukkoku-derived sd1 solely increased panicle length by more than 2.5%, as compared to the original cultivar Koshihikari and other semidwarfing alleles. The first author reported transcription of the sd1 gene derived from Jukkoku [32]. RT-PCR of root RNA showed that the 779 bp fragment derived from the sd1 locus was clearly cleaved into 613 bp and 166 bp fragments by PmaCI digestion on the SNP in Jukkoku sd1, but there was no cleavage of the Sd1 locus in Koshihikari ( Figure 5). This is the first evidence of the transcription of sd1, a defective gene of GA20ox-2. Accordingly, Jukkoku sd1, which is substituted by only a single nucleotide against the 385 bp-deficient IR sd1 [17] and is actively transcribed, has no deteriorative effect on the panicle and may have a positive effect on panicle elongation by the overflow of nutrition due to culm reduction.
In Japan, Koshihikari suffers considerable lodging damage as a result of frequent powerful typhoons and, thus, the development of lodging-resistant cultivars has been a longstanding challenge. The first author transferred the semidwarf gene Jukkoku sd1 to Koshihikari to develop a semidwarf form of Koshihikari which could withstand a typhoon by backcrossing with Koshihikari eight times [33], with more than 99.8% of the background of the Koshihikari genome, except for Jukkoku sd1 [32]. This cultivar, which was about 20 cm shorter than Koshihikari, was named Hikarishinseiki (rice cultivar number 12273) [33,34]. Hikarishinseiki is the first cultivar to be a Koshihikari-type semidwarf with sd1 registered in Japan and USA [35,36].
In this study, a MiSeq next-generation sequencer was used to achieve 5× rice genome coverage; namely, 4-5 ng of five DNA libraries was applied in one MiSeq run. Using this approach, the targeted gene mutation in Jukkoku sd1 was successfully detected as the SNP (G to T) in the defective GA20-oxidase gene, which had reduced the culm length by a loss-of-function mutation of GA synthesis, localized at position 38,382,746 from the end of the short arm of chromosome 1 of Koshihikari. This would be the minimum scale to detect Jukkoku sd1 in practical breeding. In Japan, genetically modified organisms are not acceptable to consumers; thus the target SNP in Jukkoku sd1 would be effectively tracked by using NGS of 5× coverage scale in each backcrossed generation with Koshihikari. On the other hand, as the DNA sequence and function of sd1 have been deciphered [15][16][17][18]37], new breeding methods, such as RNAi gene silencing [38] and genome editing [39], are available in other countries to retard culm length by knockout of Sd1.
The effects on phenotypic traits of rice differed between the two semidwarf genes (sd1 and d60) and also between sd1 loci of different origins (Jukkoku sd1 and IR8 sd1). The effect on culm length was more pronounced in plants carrying d60 than in those carrying sd1 (culm length: Jukkoku sd1, 71.8 cm; IR8 sd1, 68.5 cm; and d60, 65.7 cm). The reduction in internode intervals was relatively uniform in lines carrying Jukkoku sd1 or d60 (uniform reduction type), while reductions of the third and lower intervals were larger than in the upper intervals in lines carrying IR8 sd1 ( Figure 3 and Table 1). Thus, the center of gravity will be lower in the lower internode reduction type, suggesting higher lodging resistance in lines carrying IR8 sd1 than in those carrying Jukkoku sd1 or d60. Although the function of d60 is not known, it is clearly distinct from that of sd1, as the additive double-dwarf effect of sd1 and d60 appeared in Jukkoku sd1 plus d60 and IR8 sd1 plus d60, respectively. The results of this study demonstrated that d60 has a stronger effect than sd1 on culm length and exerts similar effects on other phenotypic traits as with sd1. Although many semidwarf genes are associated with a reduction in panicle length, d60 does not exert such negative effects on phenotypic traits of rice plants. Taken together, these results showed that d60 is useful for adding genetic diversity to semidwarf varieties and is thus of particular importance in the field of plant bleeding.

Conflicts of Interest
The authors have no conflicts of interest to declare.