Cereal crop yield is determined by three yield components, namely, the number of grains per unit of land area, grain weight, and the ratio of filled grains. In rice, single grain weight is genetically constant, irrespective of growth environments5. This character of rice is largely different from that of other cereal crops. For example, in wheat, single grain weight varies depending on growth conditions23,24, and a negative correlation is frequently observed between grain weight and number23. Therefore, in rice, an important target for achieving a high yield is to increase the number of grains with a high ratio of ripened grains. At the same time, this means that genetic enlargement of grain size has another great impact on increase in yields in rice. However, the relationship between grain size and yield has remained uncertain. Meanwhile, we found that a large-grain cultivar, Akita 63, exhibited a high yield (983 g m−2 of brown rice yield = 1,230 g m−2 of rough rice)10. The single grain weight of Akita 63 was 35% larger and the yield was 20–60% higher than that of the reference cultivars. According to our analysis, in spite of the large grain, the number of grains of Akita 63 did not differ from the common japonica cultivars at any crop-N content10,25,26. Therefore, a large grain size without reduction of the number of grains directly enhances the sink capacity, leading to high yield potential. However, although Oochikara, the maternal cultivar of Akita 63, has 80% larger grains, the yield was not necessarily high (560 g m−2 of brown rice yield9,27). The results in Fig. 2 clearly show that a large grain in Oochikara is associated with reduction of the number of grains and consequently, Oochikara has the same yield as that of reference cultivars with normal grains.
Among the fine-mapped major genes determining grain size, we examined the large-grain alleles of GS3, GW2, TGW6 and qSW5 in the present work. The results show that Akita 63 and Oochikara have the large-grain alleles of GS3 and qSW5 (Fig. 3; Supplementary Fig. S3). Lu et al.28 surveyed natural variation and artificial selection in major genes determining grain size among 127 varieties of rice cultivars, and reported that GS3 and qSW5 are major genes controlling grain size and that japonica cultivars with a nonfunctional GS3 and qSW5 genotype combination show the largest grain weight. Regarding this point, our results clearly coincide with their conclusion. However, as qSW5 with a 1,212-bp deletion was also found for Akita 39 with normal grains, the effects of this allele on single grain weight may be limited. Actually, functional qSW5 mainly originated from indica cultivars and leads to enlarged grain-length. The qSW5 with a 1,212-bp deletion mainly originated from japonica cultivars and has an effect on the enlargement of grain width14,29.
The single grain weight of Oochikara is appreciably greater than that of Akita 63 (Table 1; Fig. 2C). Nevertheless, we did not find a difference in the large-grain alleles between them in our investigation. This indicates that Oochikara has other genes/alleles contributing to large grain size. At the same time, our results also indicate the possibility that Oochikara has other genes/alleles which function as a negative regulator(s) of the number of grains or no genes/alleles which function as a positive regulator(s). Although it is not known whether a trade-off between single grain weight and the number of grains in Oochikara is determined by the same gene(s), the breeding from Oochikara to Akita 63 overcomes such a trade-off trait. This means that the large-grain allele of GS3 and qSW5 combination does not affect the number of grains and can be one of major determinants for a further increase in yield. These two genes are widely observed in Oryza sativa species14,21,28, and GS3 has stronger effects on grain weight in japonica cultivars28. Thus, although there still remains a possibility that other unidentified genes also come into play, it is suggested that the nonfunctional GS3 and qSW5 combination mainly contributes to the large grain size of Akita 63.
The 4-year average yields of Akita 63 were about 20% higher than those of the parents Oochikara and Akita 39, and 40–60% higher than those of the reference cultivar, Akitakomachi (Table 1). These results indicate that when Akita 63 was compared with Akitakomachi, the large grain of Akita 63 is not the sole determinant for high yield. Another factor was N uptake capacity. For the same N application, total crop-N content at the harvest stage tended to be higher in Akita 63 than in Akitakomachi (Table 1). Actually a significant difference in crop N content between them was found at an application of 0 g and 6 g N m−2 (Supplementary Table S2). This indicates that Akita 63 has superior N uptake capacity. This trait may have been inherited from the parental lines, Oochikara and Akita 39.
As already discussed above, to achieve a high yield, it is important to enhance the number of grains with a high ratio of ripened grains. In many cases, however, a negative correlation between the number of grains and the ratio of filled grains is frequently observed, especially when rice is cultivated with heavy N application30. This trend was clearly found for our data in Table 1. The ratio of filled grains to total grains decreased with increasing N application in all cultivars. Among them, the ratio of filled grains of Akita 63 was the lowest at an application of 13 g N m−2 for 3 years (Supplementary Table S2). As we previously pointed out, we think that this is caused by source limitation relative to yield potential26. Although the number of grains was linearly correlated with crop-N content passing through the origin, total biomass was curvilinealy correlated (Fig. 2A,B). Of course, the curvilinear correlation between biomass and crop N content was simply caused by a decline in canopy photosynthesis due to an excessive leaf area that may cause mutual shading at high N application. Grain mass (rough rice) in Akita 63 reached 60% of the total aboveground biomass at harvest, while that of other varieties reached about 45% (Table 1; Fig. 2B,D). This is the highest level of all cereal crops31,32. Thus, yield potential of high-yielding cultivars such as Akita 63 may surpass source capacity, leading to a decline in the ratio of filled grain. This indicates that a further increase in sink capacity is no longer effective and that improvements in source capacity will be essential for maintenance of high ratio of filled grains.
Many recent trials conducted at free-air CO2 enrichment (FACE) facilities have shown a highly positive correlation between enhanced photosynthesis, biomass and yield32,33. These results indicate that enhancement of photosynthesis by elevated [CO2] directly leads to an increase in yield when genetic factors besides photosynthesis are not altered32. Therefore, to examine the effects of source enhancement on yield, we conducted FACE experiments on several rice cultivars, including Akita 6334. The results showed that Akita 63 had the greatest enhancement of yield by CO2 enrichment among all rice cultivars grown at FACE facilities. Furthermore, the absolute yield of Akita 63 was also highest and the ratio of filled grains remained at higher level. These results indicate that enhancement of photosynthesis is of the greatest importance for a further increase in the yield of high-yielding cultivars with a large sink size. While there has been a dispute as to whether photosynthesis improvement leads to an increase in cereal crop yields35, we have actually shown that an increase in photosynthesis by overproducing Rubisco results in increased rice yields under field conditions36,37 Thus, improving photosynthesis is a possible target for realizing a further increase in yield of today’s high-yielding cultivars.
Source: Ecology - nature.com
