Introduction
While climate change is considered to have mostly adverse effects on crop production through heat stress (Zhao et al. 2017) and heat spells, rising nocturnal temperatures (Jagadish et al. , 2015, 2016; Sadok and Jagadish, 2020; Impa et al. , 2021), more frequent periods of drought or floods (Rohde, 2023), and more variable climatic conditions in general, the rise of atmospheric CO2 concentration ([CO2]) may have beneficial effects. For photosynthesis of C3 plants such as rice, CO2 is a limiting resource, and free-air CO2 enrichment experiments (FACE) in Japan (Hasegawa et al. , 2013, 2016; Kumaret al. , 2017) and China (Liu et al. , 2017; Cai et al. , 2020; Lv et al. , 2020) have established that yield gains are substantial under an increase of [CO2] from current levels (around 400 μmol mol-1) to those expected for 2050 (550 to 600 μmol mol-1). The same studies indicated that varietal differences are large for this response.
Dingkuhn et al. (2020) reviewed the probable physiological causes of varietal differences in CO2 response in C3 crops. A major determinant is the carbon sink capacity that should be commensurate with the increased source under elevated [CO2] (e-CO2). Thus, breeding for optimum C source-sink relationships might provide gains in photosynthesis response to e-CO2 and thereby increase biomass and yield.
Comparing various rice genotypes having contrasting local source-sink ratio (LSSR), defined as the ratio between flag leaf area and grain number of the adjacent panicle on the main stem, Fabre et al. (2019 & 2020) demonstrated that this trait is related to rice yield and photosynthetic response to e-CO2. Genotypes with larger sink capacities during grain filling (low LSSR) benefited more from CO2 enrichment while having increased photosynthetic capacity (A max) of flag leaves. Under severe sink limitation caused by panicle pruning and e-CO2treatment, triose-phosphate utilization (TPU) was identified as the main biochemical driver of photosynthesis down-regulation, also called acclimation (Fabre et al. , 2019, 2020; McClain et al. , 2023). Down-regulation of A max mainly occurred in the afternoon. A negative correlation was found between TPU and markers of sink limitations, such as leaf sucrose accumulation and LSSR. It is becoming increasingly evident that ignoring TPU in situations of source-sink imbalance can cause errors when modeling crop photosynthesis (Sharkey, 2019; Yin et al. , 2021)
A consequence of these findings is that A max is not solely determined by constitutive properties of the photosynthetic apparatus of a given plant, in which case its value would be constant during the day. Instead, A max decreases throughout the day, particularly when the local carbon source-sink ratio is high (Fabre and Dingkuhn, 2022). This indicates that maintaining highA max requires efficient evacuation of photosynthates from the leaves. This has consequences for strategies to improve plants for growing under future e-CO2environments.
Given that photosynthesis and yield response to e-CO2can be substantial and is in large part genotypic in rice (FACE trials: Hasegawa et al., 2013; Lv et al., 2020), this trait can be potentially enhanced by breeding. Breeding now for this trait, however, would be unrealistic if requiring selecting for the trait under e-CO2 conditions, as FACE experiments are expensive, limited in size allowing to test only a small number of genotypes, and available for rice at only a few sites worldwide. The same applies to indoor controlled environments, which in addition would be unacceptable to breeders. The true source-sink ratios during grain filling are difficult to measure. This situation calls for the identification of more easily measurable proxy traits for showing e-CO2response if they exist. If a genotype’s morphological sink-source ratio is indeed a major physiological determinant of e-CO2response and sufficiently constitutive to be conserved across different CO2 levels, predictive proxy traits might indeed be found. The LSSR (Fabre et al. , 2020) may be such a trait, although it was proposed, based on the experimental results in only a small set of rice varieties.
The LSSR is only a crude indicator. On the source side, the flag leaf, although the most light-exposed and located close to the panicle, represents only a fraction of the plant’s photosynthetic potential. Furthermore, its specific leaf area (SLA) and chlorophyll and nitrogen content also contribute to photosynthetic potential (Seneweera et al. , 2011; Xiong et al. , 2015; Wang et al. , 2022) and are not captured by LSSR. On the sink side, the panicle’s spikelet number is widely considered a measure of its overall sink capacity (Sheehyet al. , 2001; Fabre et al. , 2016; Nakano et al. , 2017; Mai et al. , 2021) but does not inform on the current sink strength of any given spikelet, nor does it take into account the genotypic variation of the attainable grain weight. There may thus be room to improve LSSR as an indicator of a rice plant’s source-sink ratio.
The present study aims at i) validating LSSR as a potential proxy trait to predict rice genotypic photosynthetic and yield response to e-CO2 for larger samples of cultivars; and ii) exploring possible improvements of the proxy trait, in terms of predictive power and practical considerations for plant selection or phenotyping. We present the results of controlled-environment experiments and discuss them with respect to further research needed to enable breeding for improved rice CO2 response.