2.1 Rice varieties and genetic material

Two rice varieties, YD 6 and Koshihikari (KOS), were used in this experiment. YD 6 and KOS showed different yield and biomass response to eCO₂ (Chen et al., 2015; Hasegawa et al., 2013; Liu et al., 2009; Zhang, Sakai, et al., 2013). According to Liu et al. (2009), free-air CO₂ enrichment (FACE) significantly increased the grain yield and total biomass of YD 6 by 31% and 21.3% than those under aCO₂. In the rice cultivar KOS, the yield and above-ground mass of KOS were increased by about 16% and 15% under FACE compared with aCO₂ (Hasegawa et al., 2013; Zhang et al., 2013a). The OsNCED3 RNAi lines were generated from the Oryza sativa ssp. Japonica cv. Nipponbare background. The relative expression level of OsNCED3 genes in the root of WT and RNAi lines was detected. The total RNA was extracted with Trizol reagent (Invitrogen). The relative transcript level of each gene was measured by real-time quantitative RT-PCR. The protocol was carried as described before (Xu et al., 2019), and the primers were listed in Table S1.

2.2 Growth condition in growth chamber and pot experimental

The experiment was conducted in two growth chambers (Saifu DRX-680 E-DG-CO₂). The growth chamber environment was set at 26/22°C day/night temperature with a 14 h/10 h day/night photoperiod, 60% relative humidity, and 300 μmol m⁻² s⁻¹ photosynthetically active radiation. Each pot (17 cm in height and 16 cm in diameter) was filled with 1.8 kg sandy loam soil. The soil was collected from a paddy rice field in the town of Yangzhong, Fujian Province, China (26°10′N, 118°21′E). Relevant soil properties are as follows: soil organic carbon 37.2 g/kg, total N 1.65 g/kg, total P 0.41 g/kg, total K 25.9 g/kg, Olsen-P 53.1 mg/kg, and exchangeable K 98.8 mg/kg. The soil was air-dried and passed through a 4 mm sieve to remove coarse material. Nitrogen, potassium, and phosphorus fertilizers were applied at 0.2 g N kg⁻¹ soil, 0.15 g K₂O kg⁻¹ soil, and 0.2 g P₂O₅ kg⁻¹ soil, respectively.

Rice seeds were surface sterilized in 10% (v/v) H₂O₂ for 30 min and rinsed with double-distilled water. After incubation in Petri dishes on filter paper at 30°C for 3 days, the uniform germination seedlings were selected and transplanted to pots with two hills per pot. After transplanted into pots for 14 days, the rice seedlings were subjected to two CO₂ concentrations and two irrigation treatments.

2.3 CO₂ and irrigation treatments

The CO₂ concentration was set as aCO₂ (400 ppm) and eCO₂ (800 ppm) in this experiment. The irrigation treatment was set as WW and AWD irrigation. The WW treatment was maintained at 1–3 cm water depth (soil water potential = 0 kPa). Under AWD irrigation, the pots were not irrigated until the soil water potential was −15 kPa at a depth of 10 cm. The soil water potential was measured with tensiometers (SP-11, Institute of Soil Science of Chinese Academy of Science).

2.4 The plant biomass and root morphological parameters measurement

After being treated with different CO₂ concentration and water irrigation for 15 days, rice shoot and root samples were harvested. The root morphological parameters (total length, total surface area, average diameter, and total volume) measured with an Epson scanner (Epson) and winRHIZO software (Regent Instruments). Then, shoot and root samples were dried at 80°C for 72 h for biomass measurement.

2.5 Measurement of root ABA concentration

ABA content measurement was followed the method described by Li et al. (2020). Briefly, 0.1 g root sample was homogenized under liquid nitrogen, then mixed with 1 ml pre-chilled 80% methanol and extracted overnight at 4°C. After centrifugation, the supernatant was collected. The residue was continuously extracted in 0.5 ml of 80% methanol for 2 h. The supernatant was collected and mixed with the first supernatant. The crude extracts were purified using a 0.22 μM membrane syringe filter. ²H₆-ABA was used as internal standard. ABA was quantified using high-performance liquid chromatography–mass spectrometry (Rigol L-3000, RIGOL Technologies, Inc.).

2.6 Statistical analysis

Data were analyzed using SPSS 18.0 software (SPSS Inc.). The three-way analysis of variance (ANOVA) was performed for the independent variables CO₂ concentration (CO₂), water irrigation (irrigation), and rice variety (variety), as well as for their interactions. Means were compared using Duncan test, and difference was considered significant at p < 0.05.

3.1 The effect of water irrigation and CO₂ concentration on rice growth

The shoot and root growth of rice was determined under two irrigation regimes and CO₂ concentrations (Figure 1; Table 1). Shoot dry weight was affected by CO₂ concentration and the interaction of irrigation×variety (p < 0.05; Figure 1a; Table 1). Irrespective of irrigation regimes, YD 6 and KOS under eCO₂ showed a significant increase in shoot dry weight of 44% and 28% than those under aCO₂. Furthermore, for YD 6, there was no significant difference in shoot dry weight between WW and AWD irrigations under aCO₂ or eCO₂ (Figure 1a). While for KOS, AWD treatment significantly reduced shoot dry weight by 22% compared with WW under aCO₂ and maintained shoot dry weight under eCO₂ (Figure 1a). The two rice varieties showed different trends when exposed to AWD (Figure 1). Compared with WW, AWD significantly reduced the shoot dry weight of KOS at aCO₂, while it tended to increased the shoot biomass in YD 6 at eCO₂ (Figure 1). Thus, ANOVA showed no significant difference in shoot dry weight between irrigation treatments by assessing the data of two rice varieties together (Figure 1; Table 1). Root dry weight was affected by all factors as well as CO₂ × irrigation and CO₂ × variety (p < 0.05; Figure 1b; Table 1). Under eCO₂, there was a significant increase in root dry weight by 41% in YD 6 and 26% in KOS compared with those under aCO₂. However, under aCO₂, the root dry weight of both YD 6 and KOS exhibited no difference between water treatments. After imposing to eCO₂, compared with WW, higher root dry weight was observed in both rice varieties when grown under AWD (Figure 1b).

3.2 The effect of water irrigation and CO₂ on root morphological parameters

The root length and total surface area were affected by all factors and an interaction of CO₂ × variety (p < 0.05; Table 2). Compared with aCO₂, eCO₂ increased root length and total surface area by 36%, 50% in YD 6 and 27%, 34% in KOS, respectively. Under aCO₂, YD 6 and KOS showed no significant difference in the root length and total surface area under AWD in relation to WW. Exposing to eCO₂, compared with WW, AWD increased root length and total surface area by 34%, 59% in YD 6 and 23%, 54% in KOS (Table 2). The root diameter was markedly affected by CO₂, irrigation, and a CO₂ × irrigation interaction (p < 0.05; Table 2). Under WW, the root diameter was not affected by CO₂ enrichment, but it was increased with CO₂ enrichment under AWD. Root volume was affected by all factors and interaction of CO₂ × irrigation, CO₂ × variety (Table 2). Under aCO₂, the root volume of YD 6 and KOS showed no significant. Root volume of YD 6 and KOS was notably higher under AWD than under WW when exposed to eCO₂.

3.3 The ABA is involved in rice root growth under water irrigation and CO₂ treatments

Under eCO₂, AWD treatment significantly increased rice root ABA content, with the increase by 83% in YD 6 and 58% in KOS than under WW (Figure 2). In addition, there was no significant difference between YD 6 and KOS under WW with eCO₂. To confirm whether ABA was involved in root growth under AWD with eCO₂ condition, the rice OsNCED3 RNAi lines (RNAi-1 and RNAi-2) were generated. The NCED3 transcripts level was detected in the root of WT and OsNCED3 RNAi lines. Compared with WT, the NCED3 expression level was reduced by 33% and 42% in RNAi-1 and RNAi-2 lines, respectively (Figure 3b). Next, the growth of WT and OsNCED3 RNAi lines was detected with same water irrigation and CO₂ concentration treatments (Figure 4; Table S2). The shoot dry weight was remarkably affected by all factors and the interaction of irrigation×variety (Figure 4a). The shoot dry weight of WT and OsNCED3 RNAi lines was significantly higher under eCO₂ than under aCO₂. AWD maintained the shoot dry weight of WT compared with WW across the two CO₂ concentrations. However, AWD reduced the shoot dry weight of OsNCED3 RNAi lines compared with WW (Figure 4a). The root dry weight was affected by CO₂, variety, and interaction of CO₂ × irrigation, irrigation × variety (Figure 4b; Table S2). WT and OsNCED3 RNAi lines produced greater root dry weight under eCO₂ than under aCO₂. AWD reduced the root dry weight of OsNCED3 RNAi lines relative to WW when grown under aCO₂, whereas they exhibited no difference between two irrigation regimes under eCO₂ (Figure 4b).

3.4 The effect of water irrigation and CO₂ on root morphological parameters of OsNCED3 RNAi lines

The root total length, total surface area, and root volume were significantly affected by CO₂, variety, and interaction of CO₂ × irrigation, irrigation × variety, but no significant difference in root diameter was observed among all treatments (Table 3). Compared with aCO₂, CO₂ enrichment significantly promoted the root length, total surface area, and root volume of WT and OsNCED3 RNAi lines (Table 3). Under aCO₂, WT showed a 20%, 17%, and 18% enhanced trend in root length, total surface area, and root volume under AWD in relation to WW. Under eCO₂, root length, total surface area, and root volume of WT were notably higher under AWD than under WW (Table 3). Under aCO₂, root length, total surface area, and root volume of OsNCED3 RNAi lines were 27%, 23%, and 25% lower under AWD in relation to under WW, while root length, total surface area, and root volume of OsNCED3 RNAi lines showed no difference between two irrigation regimes under eCO₂ (Table 3).

3.5 The effect of water irrigation and CO₂ on root ABA content of OsNCED3 RNAi lines

To confirm the effect of ABA content on root growth of OsNCED3 RNAi lines, the root samples were harvested. Under aCO₂ condition, the root ABA content of WT was significantly increased when exposed to AWD compared with WW (Figure 5). In addition, compared with WW, AWD significantly increased ABA content both in WT and RNAi lines, while the magnitude of increase was significantly higher in WT than in RNAi lines (Figure 5). Related to the aCO₂ plants, eCO₂ also significantly enhanced ABA content of WT under both WW and AWD (Figure 5).

4.1 CO₂ enhancement increases rice root growth under AWD

AWD is a water-saving irrigation technology in rice production with maintaining or increasing yield and can be an important adaptation strategy in the future climate (Carrijo et al., 2017; Yao et al., 2012). The drought and eCO₂ induced the stomatal closure and hence reduced transpiration rate, thus lowing leaf water potential (Song et al., 2018; Yan et al., 2017; Zhou et al., 2017). Young fully expended leaf was measured for leaf water potential using a psychrometer chamber (C-52 sample chamber, Wescor Inc.) connected to a microvoltmeter (HR-33 T, Wescor). Under WW or AWD conditions, the eCO₂ did not significantly change leaf water potential in both rice species (Figure S1). Under aCO₂ or eCO₂ conditions, AWD reduced leaf water potential of YD 6 and KOS in relation to WW treatment (Figure S1). Nevertheless, it was not always the case for plants grown under eCO₂ having the lower leaf water potential, even though the stomatal conductance and transpiration rate were found to be lower, but the plant hydraulic conductance could also be decreased under eCO₂ (Fang et al., 2019).

As the substrate of photosynthesis, the increase in atmospheric CO₂ had a positive effect on the growth of plant (Jiang et al., 2020; Wang et al., 2020). The biomass accumulation of rice under eCO₂ was remarkably increased at tillering, panicle initiation, heading, mid-ripening, and grain maturity (Yang et al., 2006). Wang et al. (2020) showed that eCO₂ dramatically increased total biomass of rice before heading stage. Regardless of irrigation regimes, eCO₂ increased the shoot dry weight of YD 6 and KOS (Figure 1a), which is consistent with previous studies in rice yield (Chen et al., 2015; Hasegawa et al., 2013; Liu et al., 2009; Zhang et al., 2013a). The stimulation effect of plant growth by eCO₂ is constrained by drought, depending on its severity and duration on rice growth (Leakey et al., 2006; Xu et al., 2007). In the continuously flooded, CO₂ enrichment resulted in a 29% increase in final aboveground biomass of rice while drought stress reduced biomass accumulation in both CO₂ treatment (Baker & Allen, 2005). Furthermore, AWD can reduce rice shoot growth owing to the lower root oxidation activity compared with WW (Chu et al., 2014). In this study, AWD reduced the shoot growth of KOS under aCO₂ (Figure 1), which may be due to a decrease in root oxidation activity under AWD than under WW. In our previous study, the rice shoot growth was lower under AWD than under WW at vegetative stage; however, it showed no significant difference between AWD and WW irrigations at maturity stage (Song et al., 2018). In addition, AWD reduced redundant vegetative growth and enhanced root growth, which was benefit to a higher grain yield compared with WW. For those reasons, the AWD technology should be used in rice production.

Plants can able to alter root phenotype in response to changing environment conditions (Benlloch-Gonzalez et al., 2014; Kano-Nakata et al., 2013). Our previous studies showed that root dry weight, root length, and root depth under AWD were significantly higher than those under continuous flooded (Song et al., 2019; Xu et al., 2020). Elevated CO₂ (eCO₂) can increase plant drought tolerance by contributing to root growth under drought stress (Li et al., 2020). Meanwhile, CO₂ enhancement can enhance cucumber drought resistance by improving the ability of antioxidant and osmotic adjustment (Cui et al., 2019). Here, under AWD, root dry weight of YD 6 and KOS was significantly enhanced under eCO₂ compared with aCO₂ (Figure 1; Table 2), which suggests that CO₂ enhancement can increase the positive effect of AWD on root growth. The result was consistent with the finding that CO₂ enhancement could increase root biomass under normal or moderate water stress (Li et al., 2020). Further, many studies have shown that AWD increases grain yield when compared with WW, and larger root biomass contributes to higher grain yield of rice under AWD irrigation (Chu et al., 2014; Lampayan et al., 2015; Song et al., 2019; Xu et al., 2020; Yang et al., 2004, 2012, 2017; Zhang et al., 2009b, 2013b). In addition, the eCO₂-induced root growth is also associated with the increase in grain yield (Wu et al., 2018; Yang et al., 2008). Those studies suggest that root growth under AWD or eCO₂ is beneficial to increase grain yield.

4.2 CO₂ enhancement increases rice root growth by involving root ABA response under AWD

Root ABA sharply increased when water deficit was imposed, thereby transporting to shoot for inducing stomatal closure (Brodribb & McAdam, 2013; Yang et al., 2002). In addition, stomata played a central role in leaf photosynthetic rate (Chater et al., 2015). Thus, the inhibition of shoot growth was possible due to the increased root ABA level under AWD. ABA rapidly accumulated in roots under mild and moderate drought stress and then stimulated root growth (Sengupta et al., 2011; Xu et al., 2013). Moreover, root ABA content was significantly increased under AWD irrigation compared with WW (Song et al., 2019). CO₂-enrichment-induced stomatal closure was associated with ABA signaling (Chater et al., 2015; Hsu et al., 2018). eCO₂ also increased leaf and xylem ABA concentrations in tomato (Fang et al., 2019). According to our previous study in pot experiment, root growth traits (total root length, total root volume, and root surface) were not significantly increased under AWD irrigation compared with WW in jointing stage; however, it was significantly enhanced in heading and maturity stages (Song et al., 2018). The results suggest that, besides ABA content, root growth increase may be related to treatment time of AWD in pot experiment. Thus, this may explain that there was no relationship of root ABA content and root growth between AWD and WW under aCO₂ in rice seedling stage (Figures 2 and 3). Under eCO₂, the root parameters of YD 6 and KOS were higher under AWD than WW (Figure 3), which may be attributed to the increase in ABA content. The cleavage of 9-cis epoxycarotenoids to xanthoxin, which is catalyzed by the enzyme 9-cis-epoxycarotenoid dioxygenase (NCED), is an important step in ABA biosynthesis (Nambara & Marion-Poll, 2005). OsNCED3 was a major ABA biosynthesis gene in plants (Shi et al., 2015). Rice OsNCED3 RNAi lines also exhibited a significant decrease in water stress tolerance and accumulate lower ABA than wild type under water stress condition (Huang et al., 2018). In the present study, under eCO₂ condition, AWD treatment significantly enhanced the root growth of WT but not the OsNCED3 RNAi lines (Figure 4). The ABA content of WT was higher under AWD than WW, but the ABA content of OsNCED3 RNAi lines had no difference between WW and AWD (Figure 5). These results suggest that CO₂ enhancement could increase rice root growth under AWD by involving ABA response.