IR64: a high-quality and high-yielding mega variety

Breeding history, including parentage and selection history, evaluation and release

The breeding history of IR64 is summarized in Khush and Virk (2005). The breeding program at IRRI focused on combining the different traits desired by farmers, including high yield, resistance to biotic and abiotic stresses, early maturity, and improved grain quality. The Genetic Evaluation and Utilization (GEU) program was formulated at IRRI to focus research efforts on combining these traits through interdisciplinary collaboration (Khush and Coffman 1977). In segregating generations and in evaluation of fixed lines, plants and breeding lines were evaluated for these traits either in the pedigree breeding nursery or special screening nurseries. This improved the chance of combining multiple traits into a single variety.

The cross of IR5657-33-2-1/IR2061-465-1-5-5, was made in early 1977 and was designated IR18348. The female parent was noted for its good cooking quality (intermediate amylose), as well as having some tolerance to salinity. The male parent was a high-yielding breeding line derived from a highly productive cross that resulted in a number of other varieties from sister lines, including IR28, IR29 and IR34. The full pedigree (Fig. 1) shows the derivation of IR64 from 19 traditional rice varieties.

The F2 population was evaluated in 1978, and the pedigree method of breeding was followed in the F3 and F4 populations, grown in 1979. The breeding line IR18348-36-3-3 resulted from the bulk harvest of a F₅ family in 1980, and was subsequently evaluated in yield trials in 1981–83 at IRRI, as well as in the Philippine National Trials. It out-yielded IR36 by 21% in these trials (IRRI 1986). It was released by the Philippine Seed Board with the designation ‘IR64’ in 1985, and this designation has been subsequently used by IRRI (IRRI 1986).

The early varieties developed at IRRI were high yielding and had disease and insect resistance, but their grain quality was inferior to the best available varieties. At the time, the standard for grain quality in the Philippines was the variety BPI-76 and its sister line BPI-121, derived from the cross Fortuna/Seraup Besar 15 (Cada and Escuro 1972). The parent variety Fortuna is from the USA, and Seraup Besar 15 was originally introduced from Malaysia. BPI-76 was released in 1960 in the Philippines, and non-photoperiod sensitive selections, BPI-76 (n.s.) and BPI-76-1, were also released (Dalrymple 1978). Another popular variety noted for its quality was C4-63, and it was derived from the cross Peta/BPI-76. Before IR64, C4-63 was considered one of the main high-quality varieties in the Philippines. In 1970, a green-base C4-63 was released to replace the original seed stocks of C4-63. Because of its good eating quality (intermediate amylose content) it had spread in Indonesia, Malaysia, and Burma (Yoshida 1981). Rice was even widely marketed under the C4 name up to the 1980s; however, many of the market samples actually turned out to be other varieties or mixtures of varieties with inferior quality that were being grown by farmers (Juliano et al. 1989).

BPI-121-407 is considered the most likely contributor of superior quality in the pedigree of IR64, although BPI-76 is also in its pedigree. BPI-121-407 is a short-statured breeding line with superior quality, and was selected as an induced mutant of the original BPI-121 (Cada and Escuro 1972). In the advanced evaluation of the breeding line IR18348-36-3-3 at IRRI, taste panels were used to ensure that the quality of BPI-76 (or BPI-121) was captured. Taste panels were also applied in the government of the Philippines rice program to assist in providing the data for release of this variety.

Adoption and impact

The major breakthrough in the development of IR64 was the combination of the high yield and disease and insect resistance of earlier IRRI varieties with the superior grain quality associated with varieties like BPI-76 and C4-63. It’s first release was by the Philippine government in 1985. It was also released in the following countries: Bhutan, Burkina Faso (as FKR42), Cambodia, China, Ecuador (as NIAP11), Gambia, India, Indonesia, Mauritania, Mozambique, and Vietnam (as OM89) (Khush and Virk 2005). Its wide adaptation is notable, and it became widely grown in Southeast and South Asia. It is also noted to be well adapted to the Sahelian regions of West Africa (Devries et al. 2011; Julia and Dingkuhn 2013).

By 1995, IR64 was already estimated to be grown on 8 million ha (Khush 1995), and by the turn of the century this rose to over 10 million ha. The long persistence of IR64 in farmers’ fields after its release was attributed to its excellent eating quality (Champagne et al. 2010). The area of production gradually declined in the Philippines during the period 2000–2007, partly due to pressure from tungro disease (Laborte et al. 2015). Indonesia was a major producer of IR64, which was grown on more than 40% of its total area for around a decade (Fig. 2), and was still popular in 2009 (Brennan and Malabayabas 2011). It is also widely grown in India. During 1998–2006 IR64 accounted for over 10% of the breeder seed produced in India and was still above 3% in 2015, suggesting that it was grown on 2–3 million ha annually during the period (data provided by A. K. Singh).

A specific estimate of the impact of this variety has not been attempted, even though it is known as the most popular variety in terms of area, particularly in tropical Asia. IR64 has contributed greatly to farmer incomes not only through higher yields, but through improved quality that results in higher price and earlier maturity that allows higher cropping intensity.

As with other IRRI varieties, seed of IR64 was distributed freely to researchers and farmers and no intellectual property protection was sought on it or any progeny developed from it.

Major characteristics, including key traits

IR64 is a semidwarf indica rice variety, with average mature plant height of approximately 100 cm in the Philippines (Fig. 3). It is a relatively early duration variety, with total growth duration of about 117 days (Khush and Virk 2005). It inherits the same semidwarf sd1 allele as other IRRI semidwarf varieties, ultimately derived from Dee-geo-woo-gen. According to Wei et al. (2016) it has the loss of function alleles for Hd1 and Ehd1, which confer earlier duration and insensitivity to photoperiod. At the time of its release, IRRI (1986) listed the valuable traits as resistance to brown planthopper (BPH) biotypes 1 and 3, green leafhopper (GLH), white backed planthopper (WBPH), bacterial blight, grassy stunt virus; and moderate resistance to blast, BPH biotype 2, and stem borer. It was the first IRRI variety to combine intermediate amylose content and intermediate gelatinization temperature (GT).

IR64 has high yield, especially compared to earlier-released IRRI varieties, but not as high as some of the subsequently released varieties like IR72. Peng et al. (2000) list its grain yield as 8.76 and 8.28 t ha^− 1 compared to 9.50 and 9.06 t ha^− 1 for IR72 in the dry season at IRRI in 1996 and 1998, respectively. It has high grain filling percentage and grain weight, but a relatively low spikelet number per m² (Peng et al. 2000). According to Ujiie et al. (2016), IR64 possesses a gene GS3 for grain size and the narrow-leaf gene (NAL1) that improve grain yields. IR64 is considered a typical high-tillering indica type cultivar, in contrast to the “new plant type” NPT varieties that have lower tillering but large tiller and panicle size (Okami et al. 2015).

IR64 has relatively durable resistance to BPH, and it is known to carry the major gene Bph1. However, it is reported to have better resistance than other varieties carrying Bph1 and has good field resistance to the pest, exhibiting antiobiosis, antixenosis and tolerance (Cohen et al. 1997). This is partly attributed to its possessing additional QTLs controlling BPH resistance which confer greater durability of the resistance (Alam and Cohen 1998).

IR64 has very good levels of resistance to blast disease, including major-gene resistance and partial resistance (Bastiaans and Roumen 1993; Grand et al. 2012; Roumen 1992). Sallaud et al. (2003) reported six resistance genes, designated Pi25(t), Pi-27(t), Pi29(t), Pi30(t), Pi31(t), and Pi32(t). It also has resistance genes Pita (Lee et al. 2011) and Pi20 (Khush and Virk 2005), as well as Pi33 on chromosome 8 which was derived from O. rufipogon (accession IRGC101508, sometimes referred to as O. nivara) in its pedigree (Ballini et al. 2007; Berruyer et al. 2003). Sreewongchai et al. (2010) found that IR64 has broad resistance to blast disease in Thailand and this was conferred mainly by QTLs on chromosomes 2 and 12. It was also resistant in a blast hotspot in India (Thakur et al. 2013). Kongprakhon et al. (2010) identified QTLs that may correspond to Pi25 (chromosome 2), Pi29 (chromosome 8), and Pi28 (chromosome 10).

IR64 is resistant to Bacterial Blight (BB) disease (caused by Xanthomonas oryzae pv. oryzae) and possesses the major gene Xa4 for resistance (Adhikari et al. 1994; Khush and Virk 2005). The gene is thought to confer additional agronomic benefits in addition to BB resistance (Hu et al. 2017). It is also resistant against African strains of X. oryzae, and several QTLs for resistance were identified (Djedatin et al. 2016).

IR64 is susceptible to tungro disease, including Rice Tungro Spherical Virus (RTSV) (Lee et al. 2010) and Rice Tungro Bacilliform Virus (RTBV) (Zenna et al. 2006).

IR64 was developed primarily for irrigated rice production, and abiotic stress resistance was not an objective. While it is generally considered susceptible to abiotic stresses, it has been widely grown in more favorable rainfed situations and under mildly unfavorable soils.

IR64 is considered susceptible to drought stress and yield reductions can be considerable (Anantha et al. 2016). Yields under aerobic conditions (favorable upland) are also relatively low (Zhao et al. 2010). Vikram et al. (2015) attributed drought susceptibility of many modern high-yielding varieties to linkage between a drought susceptibility QTL and the semidwarf gene sd1. IR64 has a relatively shallow root system and low root length density (Henry et al. 2011; Shrestha et al. 2014). Under water stress, IR64 has relatively low water uptake rate (Gowda et al. 2012).

IR64 is considered sensitive to heat (Coast et al. 2016; Gonzalez-Schain et al. 2016; Shanmugavadivel et al. 2017; Ye et al. 2015), although this has not been reported as a problem with previous or current production environments. However, Jagadish et al. (2008) indicated that IR64 is actually moderately tolerant of high temperature at flowering. It is also susceptible to Fe toxicity (Wu et al. 1997), anaerobic germination (Miro and Ismail 2013), and low temperature at the vegetative stage (Chawade et al. 2013). IR64 lacks the P deficiency gene PSTOL1 like a number of other high-yielding varieties (Gamuyao et al. 2012). IR64 was reported to be sensitive to low P conditions (Mori et al. 2016; Vejchasarn et al. 2016). It is also relatively sensitive to Zn deficiency (Impa et al. 2013). While IR64 does not possess the submergence tolerance allele of SUB1, it has some moderate tolerance to submergence during the vegetative stage (Singh et al. 2010; Singh et al. 2013).

Grain and market characteristics

IR64 grain has good physical appearance and is a typical long-grain variety with high head rice yield (IRRI 1986). As mentioned above, IR64 was the first IRRI variety to have both intermediate amylose content and intermediate GT. These traits are considered important for the ideal texture of cooked rice, especially for many rice consumers in South and Southeast Asia. However, these two traits alone do not account for the superior cooking quality of the variety, and methods to evaluate cooking quality are still inadequate aside from laborious sensory methods (Concepcion et al. 2015). This is why the use of taste panels was essential to identify the superior quality of IR64.

Rice texture is mostly controlled by the allele at the waxy locus (Wx) on chromosome 6. Among the major alleles at this locus, IR64 carries the Wx ⁱⁿ allele at the waxy locus signifying intermediate amylose content (Zhang et al. 2012). However, various sequence features of the Wx gene are associated with grain quality of rice, including the number of CT repeats in the 5′ untranslated part of the gene and the SNPs at specific sites in intron 1, exon 6, and exon 10 (Bligh et al. 1998; Bligh et al. 1995; Larkin and Park 2003). Based on the DNA sequence, IR64 has the (CT)₁₇ allele of Wx (Roferos et al. 2008; Teng et al. 2012). (CT)₁₇ and (CT)₂₀ are associated with intermediate amylose and soft to medium hardness (Roferos et al. 2008). The SNPs at the Wx locus are G-C-C for intron 1, exon 6 and exon 10, a common haplotype for intermediate amylose varieties (Chen et al. 2010).

Azucena and IR64 both have intermediate amylose and intermediate GT, and the doubled haploid population of the cross between the two was used to identify QTLs for several grain quality related traits, including starch properties measured by Rapid Visco Analyzer (RVA) (Bao et al. 2002). The study showed that there are genetic differences for these traits despite both varieties having similar amylose content and GT.

Calingacion et al. (2014) surveyed grain quality preferences in different countries and regions based on the most popular varieties in each area. However, two countries where IR64 dominated have different preferences (Indonesia and Philippines). These preferences can change over time.

Genetic and genomic characteristics

IR64 has been used extensively in genetic studies of rice, mainly because it represents a high-yielding and high-quality indica variety that is widely adapted to tropical lowland growing conditions. The most well-known mapping population was a doubled haploid population of about 146 lines derived from the cross IR64/Azucena (Guiderdoni et al. 1992), and first used by Huang et al. (1997) to map important agronomic traits and by Wu et al. (1997) to map tolerance to Fe toxicity. Other mapping populations with IR64 as a parent have been developed using both indica and japonica parents. A recombinant inbred line (RIL) population of 171 families was developed in a cross between IR64 and the wild relative O. rufipogon and tolerance to Al toxicity was mapped (Nguyen et al. 2003). Reciprocal chromosome segment substitution lines (CSSL) were also developed in IR64 and Koshihikari backgrounds to study the inheritance of grain shape (Nagata et al. 2015).

The original genome sequence of rice used the japonica variety Nipponbare (IRGSP 2005). In the first report of resequencing multiple varieties, IR64 was used among 20 diverse varieties, although this only included 100 Mb of the unique fraction of the genome (McNally et al. 2009). Schatz et al. (2014) reported whole genome de novo assembly for IR64 as well as the japonica variety Nipponbare and the aus variety DJ123. The genome coverage was 88.5% and included 37,758 genes. IR64 was estimated to have 381 genes not present in either of the other two varieties, and DJ123 and Nipponbare had 297 and 786 genes not found in the other varieties. Jain et al. (2014) also reported whole genome sequencing of IR64 along with Pokkali and N22, achieving 84.5% coverage. Methylation pattern has also been studied in IR64 and compared with a japonica variety Dianjingyou1 and two wild ancestors (Li et al. 2012). Gene expression and identification of functional roles of genes have been carried out with this model variety, for example drought responsive genes (Ray et al. 2011), gene expression changes during different stages of development (Sharma et al. 2012), and salinity tolerance (Wang et al. 2016).

Important progeny

The excellent grain quality of IR64 has become the standard for rice quality requirements in a number of countries. Because of its popularity with farmers, IR64 has been used widely as a parent in rice breeding, as a recipient of new genes through marker-assisted backcrossing and genetic transformation, and as a standard check for basic studies by many rice researchers. IR64 figured prominently in the mapping of many QTLs when genome-wide markers became available.

In the Philippines, IR64 was replaced by newer varieties in the early 2000s mainly due to its susceptibility to tungro disease. However, the breeders have attempted to retain the quality traits of IR64. PSB Rc82 is an example of a variety that became very popular, and IR64 is one of its grandparents. In India as well, the variety MTU 1010 became very popular, and it is a cross of Krishnaveni/IR64.

IR64 was a dominant variety in Indonesia for over two decades. In the last 10 years, it has been replaced by the variety Ciherang, which has very similar grain quality and improved yields. This variety is from the cross IR18349-53-1-3-1-3/IR19661-131-3-1//IR19661-131-3-1/IR64/IR64, and has high genetic similarity with IR64 (IRRI 2015; Septiningsih et al. 2014). It has very similar grain quality to IR64 and is also morphologically and genetically similar (Muhamad et al. 2017).

The availability of genome-wide molecular markers for marker assisted selection enabled the transfer of important traits into popular varieties like IR64 through marker-assisted backcrossing (MABC) (Collard and Mackill 2008). Early examples of this included varieties developed by pyramiding BB resistance genes into IR64, including Angke and Conde in Indonesia, and NSIC Rc142 in the Philippines (Verdier et al. 2012).

The rice submergence tolerance gene SUB1 was introduced by marker assisted backcrossing into IR64 and several other popular varieties (Septiningsih et al. 2009) (Fig. 4). Because IR64 had moderate tolerance to submergence, IR64-Sub1 tended to perform better under submergence than some of the other Sub1 lines, and its relatively early maturity allowed it to recover and produce yields after prolonged flooding and before the onset of low temperatures (Singh et al. 2010; Singh et al. 2013). IR64-Sub1 was released as Submarino 1 in the Philippines in 2009. The genetic similarity of Ciherang with IR64 was exploited to develop a submergence-tolerant version of Ciherang with only one backcross (Septiningsih et al. 2014). IR64-Sub1 was also used to develop submergence tolerant varieties in Vietnam (Lang et al. 2015).

IR64 is generally considered a restorer line for hybrid rice, particularly in the WA CMS system widely used in indica “three-line” hybrid breeding (Xie et al. 2014). However, Toriyama and Kazama (2016) developed a new IR64 cms termed CW type (Chinese wild rice), restored by the Rf17 gene.

At IRRI, induced mutation was used to generate a large collection of mutants in IR64. This collection has been used to discover many genes controlling important traits in rice (Wu et al. 2005). Some of these mutants show favorable phenotypes that could be used directly in rice breeding. Examples include tolerance to salinity (Nakhoda et al. 2012), resistance to blast disease (Madamba et al. 2009), resistance to tungro disease (Zenna et al. 2008), and drought tolerance (Cairns et al. 2009).

Transgenic applications have been limited for rice and are currently not in commercial cultivation. However, Agrobacterium mediated transformation protocols are well developed and used for IR64 (Ignacimuthu and Raveendar 2011). For example, it has been transformed with genes conferring higher Fe and Zn content in the endosperm (Oliva et al. 2014; Trijatmiko et al. 2016) and herbicide tolerance (Chhapekar et al. 2015).