The harvest index (HI) is used to measure biological success in the formation of harvestable products. However, the genetic basis of HI in rice is not well understood because it is a complex trait comprising a number of yield-related traits and physiological attributes. Under the IRRI-Japan Collaborative Research Project, 334 introgression lines with the genetic background of IR 64 were developed to enhance yield potential by backcrossing between nine NPT varieties and one Japanese high-yielding cultivar as the donor parents and IR 64 as the recurrent parent (Fujita et al. 2009). Among those introgression lines, YTH183 showed remarkably higher yield and greater adaptability to both tropical and temperate regions than IR 64 (Takai et al. 2019). In this study, we genetically analyzed the yield potential of YTH183 from the perspective of HI. Our results showed that high correlations were found among the 3 years of HI at Ishigaki, whereas low correlations were found between the 3 years of HI at Ishigaki and the HI at Tsukuba (Table 3), suggesting that RILs showed different performances between Ishigaki and Tsukuba. HI is a complex trait and susceptible to environmental conditions, and it is difficult to find a genetic factor. Nonetheless, our results clarified the fact that two major QTLs for HI (qHI5.1 and qHI8.1 detected on chromosomes 5 and 8, respectively) contributed to the yield potential of YTH183. It is noteworthy that qHI5.1 was consistently detected under the different climatic conditions of Ishigaki (subtropical) and Tsukuba (temperate) (Fig. 3, Table 4), suggesting that qHI5.1 may be a useful genetic factor for the genetic improvement of HI in Indica Group varieties grown in both temperate and subtropical regions. In the same region on chromosome 5, QTLs for panicle weight (PW), grain weight (GWt), and grain width (GWd) were also identified, indicating that this QTL increased the total sink size through increasing grain size. This QTL did not have any negative effect on yield-related traits, such as culm weight (CW) or fertility rate (FR). Our results are consistent with those of previous research showing that the improved HI of YTH183 is attributable to the enlarged sink capacity due to large grain size (Kato et al. 2011). Further, previous studies have shown that an introgressed segment in YTH183 contains a major QTL for grain size (qSW5) on chromosome 5 (Shomura et al. 2008), and using the 334 introgression lines identified a QTL for grain weight (qGW5) in the same region as qSW5 (Fujita et al. 2009). Recently, Duan et al. (2017) and Liu et al. (2017) identified a previously unrecognized gene (GSE5/GW5) at the qSW5/qGW5 locus. They showed that natural variations in the promoter of GSE5/GW5 contribute to grain size diversity. In this study, we found a DNA polymorphism between IR 64 and YTH183 (an insertion in YTH183) in the promoter region of the GSE5/GW5 gene (Supplemental Fig. 2). The sib lines derived from the same parent (YP5) as YTH183 with the insertion significantly showed heavier grain weight than those without the insertion (Supplemental Fig. 4). We therefore consider that qHI5.1 is identical to GSE5/GW5.
Shomura et al. (2008) compared the rice varieties of the Japonica and Indica Groups, and they then identified a major QTL for grain size as qSW5, which is identical to GSE5/GW5, and the Japonica allele clearly contributed heavier grain weight. YTH183 was developed by the backcross breeding between a new plant type variety, IR69093-41-2-3-2 (YP5) harboring chromosome segments from the Tropical Japonica Group rice varieties and IR 64 as the recurrent parent. We therefore consider that qHI5.1 is identical to GSE5/GW5, which might be widely harbored in Japonica Group varieties from tropical to temperate, and that the allele of YTH183 contributes to the wide grain, resulting in the higher GWt, PW, and HI in YTH183 than in IR 64. Further, because qHI5.1 was stably identified under different environmental conditions, it is possible that this QTL could become one of the widely useful genetic factors to increase HI without the negative effect of environmental conditions, especially high temperatures during the maturing stage such as in tropical or subtropical regions. It still needs to be considered that qHI5.1 might be co-located with GSE5/GW5 in the same chromosome region, and the identity will have to be confirmed based on the gene isolation of qHI5.1.
At the qHI8.1 region, no other QTLs were identified other than a QTL for leaf weight (LW) in 2018. Kato et al. (2011) demonstrated that the source supply for grain growth, i.e., the concentration and amount of non-structural carbohydrate in vegetative organs at anthesis, was largely not improved in YTH183. On the other hand, Ishimaru et al. (2017) suggested that the higher HI in YTH183 is supported by greater photosynthate translocation from sources to sinks during the maturing stage. They showed greater non-structural carbohydrate depletion during the grain-filling stage and a higher number of vascular bundles in the panicle neck of YTH183, suggesting enhanced capacity for assimilate transport to the developing panicles during the maturing stage. Laza et al. (2006) identified QTLs for HI on chromosomes 8 and 11 by using RILs derived from a cross between IR 72 and an NPT variety that was a sibling of the donor variety (YP5) of YTH183. They indicated that these QTLs might be involved in the difference in grain-filling abilities between the Indica and Japonica Groups. Therefore, qHI8.1 might be involved in the ability of the photosynthate to be translocated from the source to the sink organs. Interestingly, qHI8.1 did not have a significant effect on HI in the cropping at Tsukuba or in the second cropping at Ishigaki. This result might be explained by differences in the environmental conditions during the maturing stage: The daily average temperature during the maturing stage was higher in the first cropping season at Ishigaki than in the second cropping at Ishigaki and the cropping at Tsukuba (Supplemental Table 3); therefore, the effectiveness of qHI8.1 might be temperature dependent. Welch et al. (2010) reported that a higher minimum temperature reduced yield by analyzing data from 227 intensively managed irrigated rice farms in six important rice-producing countries. Peng et al. (2004) also reported direct evidence of decreased rice yields from increased nighttime temperature associated with global warming. In the metabolic process of grain filling, Yamakawa et al. (2007) showed that the expression level of genes encoding starch- or carbohydrate-metabolizing enzymes and translocators was diminished to 89% of the control on average by exposure to high temperature, suggesting that ripening under high temperature induced the occurrence of grains with various degrees of chalky appearance and decreased weight. Therefore, qHI8.1 is expected to enhance source ability or the translocation of photosynthesis products from shoots to grains under conditions of higher temperature during the maturing stage. Introgression of qHI8.1, in addition to qHI5.1, into Indica Group varieties would contribute to enhancement of HI in tropical regions, and in temperate regions where air temperatures during the maturing stage have been progressively increasing (Iizumi et al. 2011). The effects of the two QTLs, qHI5.1 and qHI8.1, under various environmental conditions need to be confirmed using the isogeneic lines for qHI5.1 and qHI8.1 and the accumulated line for qHI5.1 and qHI8.1 with an IR 64 genetic background.
Studies on inheritance of the HI trait are limited, because it is a complex trait that incorporates the number of grains, grain weight, and number of panicles as the sink capacity, and also incorporates culm and leaf weight, photosynthetic rate, and translocation of the assimilated carbon as the source activity. In addition, although some of these traits are highly genetically controlled, most are environment sensitive. Consequently, it is difficult to identify the QTLs for HI independent of environmental effects. Multi-location or multi-year trials are essential for identifying effective and stable QTLs. Hittalmani et al. (2003) identified eleven QTLs for HI at nine locations in Asia representing different environments; however, no common QTL was detected in all nine locations. Li et al. (2012) identified genetic markers associated with HI in both temperate (Arkansas) and subtropical (Texas) climates, but detected no associated markers in common. In our study, two stable QTLs (qHI5.1 and qHI8.1) for HI were identified by three multi-year field trials in the first cropping season at Ishigaki. Among them, qHI5.1 was found to contribute to the grain size independent of the environmental conditions, whereas qHI8.1 was efficient for maturing under high temperatures. Further, YTH183 shows superior performance when cultivated under water-saving conditions and shows high plasticity in root elongation in response to re-watering after drought (Kato et al. 2011; Kano-Nakata et al. 2013). Furthermore, Obara et al. (2014) identified a QTL on chromosome 6 for root length under N-deficient conditions in YTH183. Thus, YTH183 is remarkable for its excellent response to water- and N-deficient conditions in addition to its enhanced sink capacity owing to its enlarged grain size. Our current findings and previous work suggest that YTH183 will likely be a noteworthy source of breeding material that can be exploited in rice-breeding efforts in the near future for breaking through the ceiling of maximum yield in response to environmental change under global warming.