Roles of the Hd5 gene controlling heading date for adaptation to the northern limits of rice cultivation
- First Online:
- Cite this article as:
- Fujino, K., Yamanouchi, U. & Yano, M. Theor Appl Genet (2013) 126: 611. doi:10.1007/s00122-012-2005-5
- 1.4k Views
During the diversification of cultivated rice after domestication, rice was grown in diverse geographic regions using genetic variations attributed to the combination of alleles in loci for adaptability to various environmental conditions. To elucidate the key gene for adaptation in rice cultivars to the northern limit of rice cultivation, we conducted genetic analyses of heading date using extremely early-heading cultivars. The Hd5 gene controlling heading date (flowering time) generated variations in heading date among cultivars adapted to Hokkaido, where is the northernmost region of Japan and one of the northern limits of rice cultivation in the world. The association of the Hd5 genotype with heading date and genetical analysis clearly showed that the loss-of-function Hd5 has an important role in exhibiting earlier heading among a local population in Hokkaido. Distinct distribution of the loss-of-function Hd5 revealed that this mutation event of the 19-bp deletion occurred in a local landrace Bouzu and that this mutation may have been selected as an early-heading variety in rice breeding programs in Hokkaido in the early 1900s. The loss-of-function Hd5 was then introduced into the rice variety Fanny from France and contributed to its extremely early heading under the presence of functional Ghd7. These results demonstrated that Hd5 plays roles not only in generating early heading in variations of heading date among a local population in Hokkaido, but also in extremely early heading for adaptation to northern limits of rice cultivation.
Heading date is a major determinant of rice adaptability to regional and environmental conditions and has been a major target in rice breeding programs. Due to its importance, many quantitative trait loci (QTLs) have been identified using various kinds of cross combinations and several genes have been identified by a positional cloning strategy (Izawa 2007a, b; Alonso-Blanco et al. 2009; Yamamoto et al. 2009; Miura et al. 2011; Tsuji et al. 2011). Previously, a series of QTL mapping using 12 diverse rice cultivars from Asia demonstrated that a large portion of the phenotypic variations in heading date could be generated by combinations of different alleles of QTLs for heading date (Ebana et al. 2011). QTL analysis for extremely late heading in an indica Nona Bokra revealed that its strong photoperiod sensitivity was generated mainly by the accumulation of additive effects of particular alleles at QTLs (Uga et al. 2007). These studies demonstrated that genetic variations can be attributed to the combination of alleles in loci involved in heading date.
Cultivated rice, Oryza sativa L., originating from a tropical region is a short-day plant and an example of adaptation to an extremely wide range of climatic conditions, from 53°N to 40°S latitude. Hokkaido, 41°29′–45°31′N latitude, is the northern-most region of Japan and one of the northern limits of rice cultivation in the world. Intensive selection on heading date by breeding programs has allowed the cultivation of rice in such higher latitudes with longer natural daylength and lower temperatures. Only varieties with extremely low photoperiod sensitivity were grown in such regions with long natural daylength conditions (Tanisaka et al. 1992; Okumoto et al. 1996). Furthermore, this unique heading behavior may contribute to the genetic differentiation of the local population in Hokkaido from those in other regions of Japan (Nagasaki et al. 2010).
To understand the genetic architecture of extremely early heading in rice, several studies have been performed using cultivars adapted to the northern limits of rice cultivation, such as Hokkaido and Europe (Fujino 2003; Fujino and Sekiguchi 2005a, b, 2008; Nonoue et al. 2008; Shibaya et al. 2011; Fujino and Iwata 2011). These results demonstrated that the loss of two photoperiod sensitivity genes, qDTH7-1/Ghd7 and qDTH7-2/Hd2, was significant to this adaptation (Fujino and Sekiguchi 2005a, b; Nonoue et al. 2008; Shibaya et al. 2011). Among the population with such extremely early heading, variations in heading date were observed. This variation is important to generate different maturity types of rice cultivation in a local region. Previously, we identified several QTLs controlling heading date among varieties from Hokkaido (Fujino 2003; Fujino and Sekiguchi 2008; Nonoue et al. 2008). However, their roles on variations in heading date among the local population were unclear.
A single gene involved in photoperiod sensitivity contributes to differences in heading date, early and medium maturity types, among commercial rice cultivars from Hokkaido (Fujino 2003; Nonoue et al. 2008). This gene collocated with a QTL controlling heading date, Hd5, which was identified in the population derived from the cross between Nipponbare and Kasalath (Lin et al. 2003). Results of map-based cloning of the genes DTH8 (QTL for days to heading on chromosome 8) and Ghd8 (QTL for grain yield, heading date, and plant height), which are allelic to Hd5 (Shibaya et al. 2011), revealed that this gene encodes the HAP3 subunit of CCAAT-box-binding transcription factor (Wei et al. 2010; Yan et al. 2011).
Here, we report on the roles of Hd5 on generating variations in heading date among cultivars in the northern limit of rice cultivation. Hd5 plays a role on generating early heading in variations of heading date among local populations from Hokkaido and the loss-of-function Hd5 was selected as a spontaneous mutation during rice breeding programs in Hokkaido. In addition, loss-of-function Hd5 could contribute to extremely early-heading behavior under the presence of functional Ghd7.
Materials and methods
A total of 204 rice varieties were used for analysis of the distribution of Hd5 alleles. Core collections for 63 diverse accessions (Kojima et al. 2005) and 37 landrace type varieties from Japan (Ebana et al. 2008) were included for the analysis (Table S1, S2). In addition, 85, 14, and 5 rice varieties were used from Hokkaido (the northern-most region of Japan), Honshu (the main island of Japan), and Europe, respectively (Table S3, S4). These rice varieties included landrace and modern varieties in each region. The 85 varieties from Hokkaido were selected to cover a wide area of cultivation across Hokkaido from the late 1800s to the present. Previously, we determined the population structure among rice varieties from Japan (Nagasaki et al. 2010; Yonemaru et al. 2012). These results demonstrated that varieties from Hokkaido were clearly assigned to a single population, which was genetically distinct from those from other regions of Japan. Eleven F2 populations (n = 76–100) derived from crosses between the nine varieties were developed for the genetic analysis of heading date. Seeds of rice varieties were provided by the Genebank of the National Institute of Agrobiological Science (Tsukuba, Japan) and Hokkaido Central Agricultural Experiment Station (Takikawa, Japan).
To confirm the effect of Hd5 on heading date and photoperiod sensitivity, an advanced backcrossed progeny for Hd5 was developed. First, a japonica Kitaibuki was crossed with a japonica Hoshinoyume. The resultant F1 plants were backcrossed with Kitaibuki to obtain BC5F1 seeds. Among the BC5F2 population derived from the self-pollination of BC5F1 plants carrying heterozygous of Hd5, a plant carrying homozygous for the functional Hd5 was selected as a near isogenic line (designated as KTm). In each generation, marker assisted selection (MAS) for Hd5 was conducted using the PCR primer set 19DEL to detect the 19-bp deletion found in Hayamasari; TCACATGAAGAGTAGGAAGAGCT and TGATGAACTCCGACACGCAC.
All rice varieties and F2 populations were cultivated in an experimental paddy field at HOKUREN Agricultural Research Institute (Naganuma, Hokkaido, Japan, 43°03′N latitude) in 2003, 2004, and 2006. Cultivation conditions were described in Fujino and Sekiguchi (2005a, b, 2008). Sowing and transplanting were performed in late April and late May, respectively. Days to heading (DTH) of the earliest heading panicle among individuals was recorded for each plant as the number of days required from sowing to heading. Leaf samples of each plant were collected for DNA extraction.
To test the function of Hd5 on heading date, an approximately 4.2-kb genomic fragment of Nipponbare, which was digested by HindIII and EagI, was transformed into the pPZP2H-lac binary vector (Fuse et al. 2001). The resultant plasmid was then introduced into Hayamasari by means of Agrobacterium-mediated transformation (Toki et al. 2006). The Nipponbare and Hayamasari alleles of Hd5 are function and loss-of-function, respectively (Shibaya et al. 2011). Two T2 homozygous lines were raised under LD conditions (14.5-h light) and scored their DTH.
Evaluation of photoperiod sensitivity
Kitaibuki, Hoshinoyume, Fanny, and KTm were grown under three different daylength conditions: 16-h day (16D; 16-h light; 27°C for 12 h and 23°C for 12 h), 14-h day (14D; 14-h light; 27°C for 12 h and 23°C for 12 h), and 10-h day (10D; 10-h light; 27°C for 12 h and 23°C for 12 h) in a controlled growth cabinet (Nihonika, Japan). DTH required from sowing to heading was scored for ten plants per lines.
Total DNA was extracted from the leaves of each plant individually according to the CTAB method (Murray and Thompson 1980). Genotyping with SSR and STS markers was performed as described in Fujino et al. (2004). For QTL analysis for heading date, SSR and STS markers listed in Table S5 were used, which were linked to the ten known QTLs for heading date (Yano et al. 1997; Lin et al. 1998, 2002; Yamamoto et al. 2000; Yano et al. 2000; Yano 2001; Doi et al. 2004; Xue et al. 2008).
The sequence of the Hd5 region including 1.3 kb of the 5′ upstream, 0.9 kb of the coding, and 1.0 kb of the 3′ downstream regions were determined in Kitaibuki and Hoshinoyume from Hokkaido, Italica Livorno from Italy, Arroz Da Terra from Portugal, and Dunghung Shali from Hungary. Amplified DNA fragments from genomic DNA with specific primers were purified and sequenced with the Sanger dideoxy terminator method on capillary sequencers (Applied Biosystems, USA). The sequences of Hd5 in these varieties were deposited in GenBank as Accession Nos. AB693200–AB693204.
In addition to the functional nucleotide polymorphism (FNP) of the 19-bp deletion, the genotype of three FNPs, the 1, 8, and 1,116-bp deletions (Wei et al. 2010; Yan et al. 2011), were determined in 45 varieties from Hokkaido and 32 varieties in the Japanese landrace core collection. To detect these FNPs, specific primer sets were designed for 1DEL (F: AAGGAGACGGTGCAGGAGT and R: TTGATGGTCTTCCGCTTCTC), 8DEL (F: CCGCTCAAGTCCTACCTCAA and R: TGACCATGGTGTGAGTGTGA), and 1116DEL (F: GTCAGGGAACAAGGCGTACT and R: AGGGCAGTACAACAGCATCC).
The association of the genotype of Hd5 with heading date among 67 varieties from Hokkaido was calculated by ANOVA. The genotype of Hd5 was determined using the PCR primer set 19DEL to detect the 19-bp deletion, deletion (Del) or non-deletion (Nd). For QTL analysis of heading date, the mean value of DTH in F2 populations was compared between each homozygous genotype class of parental type by ANOVA according to the procedure described in Fujino and Sekiguchi (2005a, b, 2008). The detection threshold for QTLs in this study was P < 0.001.
Hd5 controls heading date
To confirm that this gene is identical to Hd5, the functional Hd5 was introduced into Hayamasari by means of Agrobacterium-mediated transformation. Hayamasari is one of the parents for Kitaibuki and carries the 19-bp deletion in Hd5 (Shibaya et al. 2011). Transgenic T2 plants exhibited more delayed DTH than that of Hayamasari under LD conditions (Table S6). These results clearly show that allelic variations in Hd5 such as functional and loss-of-function are major determinants of differences in heading date in Hoshinoyume with Kitaibuki and Hayamasari.
Distribution of the 19-bp deletion in Hd5
Summary of the distribution of the loss-of-function Hd5 among different populations
No. of varieties
Japanese landrace core collection
Rice core collection
The loss-of-function Hd5 was found in only two varieties other than a local population in Hokkaido. Dianyu1 from China carries the loss-of-function Hd5 in the rice core collection over the world (Table S1). Dianyu1 is the progenitor of Jianjing6 carrying the loss-of-function Hd5 (Wei et al. 2010). Norin No. 34 is the progenitor of Dianyu1 and carries the loss-of-function Hd5, indicating that the loss-of-function Hd5 was introgressed into these varieties in China from Norin No. 34 in Hokkaido (Fig. S1B, C). Fanny from France carries the loss-of-function Hd5 among varieties from Europe (Table S4), in whom the pedigree was unknown. Because the same mutation rarely occurs several times within 10,000 years, we can consider this mutation to be a single event in rice evolution. These results strongly suggested that the loss-of-function Hd5 generated in Hokkaido was introduced into varieties from China and France by rice breeding programs.
Role of Hd5 on variations in heading date among rice varieties from Hokkaido
Relationships between the genotypes of Hd5 and heading date among rice varieties from Hokkaido
No. of varieties
Days to heading
In an F2 population of Bouzu No. 6/Kitaibuki, both with the 19-bp deletion of Hd5, no early transgressive segregation was observed, indicating that the loss-of-function Hd5 contributed to early heading in both varieties (Fig. 3a). The same results were also obtained in four F2 populations derived from crosses between landrace and modern varieties (Fig. 3d–g). QTL analysis revealed that only two loci corresponding to Hd1 and Hd5 contributed to variations in heading date (Table S5, S7). The loss-of-function Hd5 decreased DTH in these four populations (Table S7), and almost all of the early transgressive segregants carried the loss-of-function Hd5. Furthermore, bimodal distribution of DTH was clearly correlated with Hd5 genotypes in the F2 population of Sasahonami/Kitaibuki (Fig. 3h).
Role of Hd5 in the rice variety Fanny from France
Based on the QTL analysis with the sequence of Hd1, Hd5, and Ghd7 or FNPs typing for Hd5 (the 19-bp deletion) and Ghd7 (a premature stop codon), the genotypes of these genes in Kitaibuki (Hd1hd5ghd7), Hoshinoyume (Hd1Hd5ghd7), and Fanny (hd1hd5Ghd7) were identified. Due to the low resolution of QTL analysis including population size and epistatic interactions of QTLs, QTL could not be detected on the Hd1 locus in the F2 population of Kitaibuki (Hd1) and Fanny (hd1) (Table S8). Extremely late heading transgressive segregation was observed in the F2 population of Fanny with Hoshinoyume, but not in that with Kitaibuki (Fig. 4). These results revealed that extremely early heading in Fanny was dependent on the loss-of-function Hd5 under the presence of functional Ghd7.
Agronomic traits have been the targets of selection during crop domestication and diversification processes and breeding programs. Identification of the genes that control the range of these trait variations could present opportunities to study phenotypic diversification among cultivated crops. This study can give us some insights into the diversification of the traits in breeding programs for shaping adaptability to local regions. This study demonstrated that Hd5 has a role in generating early heading in variations of heading date among local populations from Hokkaido and loss-of-function Hd5 could contribute to extremely early-heading behavior under the presence of functional Ghd7.
QTLs for heading date including Hd5 have been identified using several populations between landrace and modern varieties from Hokkaido (Fujino and Sekiguchi 2008; Nonoue et al. 2008). However, it was difficult to understand how these QTLs contributed to variations of heading date among the local population due to the limited number of populations examined. QTL analysis can identify genes by differences between parental varieties. Therefore, once the corresponding gene to QTL has been identified at the molecular level, it would be possible to elucidate the role of the gene in the variation of traits among populations. A series of genetic analyses and a survey of the Hd5 genotype in this study clearly demonstrated that Hd5 has a role in generating early heading in variations of heading date among this local population.
Distribution of the loss-of-function Hd5 revealed that the mutation event of the 19-bp deletion occurred in the landrace Bouzu in the early 1900s. This spontaneous mutation exhibiting earlier heading has been selected as the variety Bouzu No. 6 (Fig. S1A). Cultivation areas of Bouzu were limited within south and south-western regions of Hokkaido, while Bouzu No. 6 could be grown in northern and eastern regions of Hokkaido, where rice never grew at that time. Human demands that enlarge rice cultivation areas had selection pressure for such mutations exhibiting earlier heading. In the process of shaping the adaptability to unique environmental conditions, desirable genes may have accumulated into the local population using standing alleles and newly occurred alleles. These varieties with unique adaptability would be useful in identifying novel genes for adaptability.
This study also demonstrated that the utility of the loss-of-function Hd5 selected in Hokkaido takes a new genotype for heading date under a different local population, France. Previously, we identified genes for extremely early heading and for variations in heading date among a local population in Hokkaido. Loss-of-function of qDTH7-1/Ghd7 has been shown to be important for extremely early heading in varieties from both Hokkaido and Europe (Fujino and Sekiguchi 2005b; Nonoue et al. 2008; Xue et al. 2008; Shibaya et al. 2011). Allelic variations in Hd5 are important for variations in heading date among a local population in Hokkaido, which carries the loss-of-function Ghd7, ghd7 (Fujino 2003; Nonoue et al. 2008). In this study, we identified a new genotype exhibiting extremely early heading other than ghd7. Under the presence of functional Ghd7, Fanny showed extremely early heading due to the loss-of-function Hd5. Because Hd5 exhibited epistatic interactions with Hd1 and Ghd7, Hd5 showed these different effects on heading date (Lin et al. 2003; Gu and Foley 2007; Nonoue et al. 2008; Xue et al. 2008; Yan et al. 2011).
For the adaptation of rice to the northern limits of rice cultivation, Hd5 has these roles in different genetic backgrounds. The understanding of genetic bases for adaptation to the local region and for controlling heading date in rice can design new genotypes with desirable heading dates. On the other hand, genes for heading date regulate agronomic traits including plant height and yield potentials (Xue et al. 2008; Wei et al. 2010; Yan et al. 2011; Endo-Higashi and Izawa 2011). Different genotypes for extremely early heading have the possibility to develop varieties with different agronomic traits and heading habits.
We are grateful to Kazuo Ise (Japan International Research Center for Agricultural Sciences, Japan) and Professor Dai Luyuan (Yunnan Academy of Agricultural Sciences, China) for information on the pedigree of rice varieties from China.