Rice (Nipponbare, ZH11) and the published transgenic and mutant rice lines, including eto1-1 (ZH11), ein2-1 (Nipponbare) and OX-A8 (Nipponbare), were used in this study. The eto1-1 mutant was provided by Lizhong Xiong of the National Center of Plant Gene Research in Wuhan, China (Du et al., 2014). The OX-EIN2 and ein2-1 line were provided by Jinsong Zhang of the Institute of Genetics and Developmental Biology of the Chinese Academy of Sciences in Beijing, China (Ma et al., 2013). The OX-A8 line was provided by Yiyong Zhu of the Nanjing Agricultural University in Nanjing, China (Liu et al., 2012), and the expression of OsA8 in OX-A8 plants was significantly higher than WT (Supplementary Figure S1). The seeds were soaked in deionized water overnight at 30°C in darkness before transferring to a net floating on a 0.5 mM CaCl₂ solution. After 7 days of germination, the plants were grown hydroponically in black pots containing a pH 5.5 nutrition solution of 1.44 mM NH₄NO₃, 0.3 mM NaH₂PO₄, 0.5 mM K₂SO₄, 1.0 mM CaCl₂, 1.6 mM MgSO₄, 0.17 mM Na₂SiO₃, 50 μM Fe-EDTA, 0.06 μM (NH₄)₆Mo₇O₂₄, 15 μM H₃BO₃, 8 μM MnCl₂, 0.12 μM CuSO₄, and 0.12 μM ZnSO₄. After 7 days, the plants were transferred to a nutrition solution with a pH of 6, 7, or 8. Additionally, the rice plants were subjected to alkaline stress with or without the following additions: ethylene precursor ACC (1 μM, 10 μM), IAA (10 μM) and plasma membrane (PM) ATPase stimulator FC (10 μM). After 5 days in the various treatments, the phenotypes of plants were investigated. All experiments were performed with three replicates. The pH of the nutrition solution was measured daily using a pH electrode (METROHM) and regulated with 1 M HCl or KOH.

To study the effect of alkaline stress on root growth and proton release, we grew rice seedlings on a one-half-strength Murashige and Skoog medium (MS). The seeds were surface sterilized for 15 min using 0.5% NaClO (w/v) and rinsed completely with ultrapure water. After incubation at 30°C for 3 days, the plants were planted on one-half-strength agarose solid MS medium at pH 6 or pH 8. The medium acidification under an alkaline condition (pH 8) with the ethylene precursor ACC (1 μM, 10 μM), AgNO₃ (10 μM) and PM ATPase inhibitor VA (0.1 mM) was investigated by staining with the pH indicator bromocresol purple (0.04 g L^-1).

After 3 days of treatment in different pH solutions, the PM H⁺-ATPase activity at the root tip was determined following the method of Zhu et al. (2009). The activity of H⁺-ATPase was measured by the Pi amount after 30 min of reaction in 0.5 ml reaction volume with 30 mM BTP/MES (pH 6.5), 5 mM MgSO₄, 50 mM KCl, 1 mM NaN₃, 5 mM ATP, and Brij 58 (0.02% w/v). The reaction was initiated by adding 5 μg of membrane protein for 30 min at 30°C and stopped using 1 ml of stopping solution containing H₂SO₄ (2% v/v), sodium dodecyl sulfate (5% w/v), and sodium molybdate (0.7% w/v), followed by 50 μl ascorbic acid (10% w/v). The color development of the phosphomolybdate complex was completed after 30 min. The absorbance at 820 nm was measured using a spectrophotometer.

The rice root H⁺ fluxes were measured at 10 mm from the root apex. The scanning ion-selective electrode technique was used to measure the root H⁺ fluxes with a Non-invasive Micro-test Technology system (NMT, Xuyue Beijing Sci. & Tech. Co., Ltd., Beijing, China). After the roots were equilibrated in the measurement solution for 20 min, they were immobilized in the measuring chamber in 5 ml of fresh measuring solution.

After 5 days in various treatments, the new adventitious roots were imaged using a digital camera or an Olympus IX-70 microscope (Olympus, Japan). For anatomical analysis of the roots, toluidine blue staining of longitudinal cross sections (10 mm from apex) were performed to measure the cell sizes in the mature root zone. The size of the cortical cells was determined for 5 roots, with 10 cells each (n = 50).

The plants were grown hydroponically as described above for 3 days. The roots were harvested in three replicates, immediately frozen in liquid nitrogen and stored at -80°C. The total RNA was extracted independently using TRIZOL reagent (Invitrogen, Carlsbad, CA, United States). Reverse transcription was conducted using M-MLV Reverse Transcriptase (Promega, Madison, WI, United States). Real-time fluorescence quantitative PCR (RT-qPCR) for detecting the expression of the target genes was performed using a SYBR Green Real-Time PCR Master Mix Kit (TOYOBO, Japan) and CFX96TM Real-Time PCR Detection System (Bio-Rad, Hercules, CA, United States). The PCR conditions were as follows: 95°C for 5 min, 40 cycles of 95°C for 10 s, 60°C for 15 s and 72°C for 20 s. The relative mRNA levels for each gene in different samples were calculated as 2^-ΔΔC_t. Primer sequences used for the RT-qPCR are shown in Supplementary Table S1.

Each graphical plot represents the results of multiple independent duplicates from three replicates and the values are the means ± SE. Statistical significance between two treatments or two genotypes in a specific pH condition was used independent Student’s t-test. LSD test was used for multiple comparisons at the p < 0.05 level among three different treatments in a specific pH. Correlation coefficients were determined by Pearson’s bivariate correlation analysis at the p < 0.05 level.

In this study, a hydroponic system was used to investigate rice growth and root elongation under alkaline stress. After the seedlings were exposed to pH 6, 7 or 8 nutrition solutions for 30 days, the rice growth was severely inhibited by the alkaline stress (pH 8), with shorter plant height and root length than at pH 6 (Figures 1A,B), resulting in dramatic biomass reduction (Figure 1D). The plant root is the first organ that perceives soil alkalinity, which inhibits root elongation. The morphology of the newly grown roots was changed after short-term alkaline stress, with shortened root length, reduced apical unbranched zone and high density lateral roots in mature root zones (Figures 1C,E,F). These indicated that rice was more adaptive to an acidic condition.

The form of nitrogen (N) available to plants is an important factor that affects the plant rhizosphere pH by releasing H⁺ with ammonium uptake or OH^- with nitrate uptake. To verify the mediation of rhizosphere pH by the nitrogen form on rice growth, we investigated the phenotypes of rice growth on solid agarose solid medium supplemented with nitrate or ammonium at pH 6 or pH 8. After growing seedlings for 7 days, the medium with ammonium was acidified regardless of the medium pH treatments, and the medium with nitrate was alkalified by the roots (Supplementary Figure S2). After 15 days of growth on the medium, the rice plants grew well on the medium with ammonium, even at pH 8 (Supplementary Figure S2). These results suggest that the preference of rice to NH₄⁺ nutrition was partly due to the adaptation of rice to an acidic rhizosphere environment.

As ethylene plays a role in the inhibition of primary root growth by affecting cell elongation (Swarup et al., 2007), we analyzed the effect of ethylene on root growth inhibition under alkaline stress. 1-aminocyclopropane-1-carboxylic acid synthase (ACS) and 1-aminocyclopropane-1-carboxylic acid oxidase (ACO) are two key enzymes responsible for ethylene biosynthesis in plants. The rice genome has six genes encoding ACS (ACS1 to ACS6) and ACO (ACO1 to ACO5, and ACO7 (Iwai et al., 2006). Among the ACS and ACO genes, OsACS1, OsACS2, OsACS5, OsACS6 and OsACO3, OsACO4, OsACO5 showed higher expression in roots than other genes and their expression levels were strongly induced by high pH, except OsACS6 and OsACO5 (Figure 2). We investigated the effect of 1-aminocyclopropane-1-carboxylic acid (ACC) on rice growth at normal pH (6) or high pH (8). The growth of plants was inhibited by exogenous ACC and the inhibition was aggravated under the alkaline condition (Figure 3). Under the normal pH condition, the growth of plants was not affected by 1 μM ACC compared with the control and the growth inhibition occurred up to an ACC concentration of 10 μM (Figures 3A,B). The plant growth was greatly suppressed by 1 μM ACC at the high pH condition (Figures 3A,B). The decreased plant dry weight could be attributed to the severe root development inhibition (Supplementary Figure S3). The length of the root apical unbranched zone was decreased by 54% with 10 μM ACC at pH 6 and up to 73% at pH 8 compared with the roots at pH 6 (Figures 3C,D). The cell length in the unbranched zone was reduced in the high pH condition and was aggravated by exogenous ACC (Figures 3E,F). These results suggest that up-regulation of ACS and ACO genes promoted ethylene biosynthesis and induced the suppression of plant growth under the high pH condition.

To verify the effect of ethylene biosynthesis and signaling on alkaline stress-induced growth inhibition, a genetic approach was adopted using ethylene overproducer (eto1-1) mutants, and ethylene-insensitive mutants (ein2-1). In the eto1-1 mutant, the stability of ACS2 is increased, which leads to overproduction of ethylene (Du et al., 2014). In the present study, we found that the eto1-1 mutants showed similar growth to that of wild-type plants (ZH11) at normal pH (6). The growth of the eto1-1 mutants was severely inhibited under an alkaline condition and displayed shorter roots and more serious etiolation in the shoot than the WT plants at high pH (Figures 4A,B). We also grew rice seedlings on one-half-strength Murashige and Skoog medium (MS) at pH 6 as a control and pH 8 for alkaline stress. Under high pH (8) stress, the ein2-1 mutants had much longer roots than the WT plants (Nipponbare) and the roots were less inhibited than the roots at normal pH, but the root lengths of the eto1-1 mutants were dramatically reduced compared with that of WT plants (Figures 4C,D). These results indicate that increased ethylene biosynthesis could be responsible for the inhibition of rice growth under high pH stress.

The elongation of the primary root requires cell wall acidification via PM H⁺-ATPase (Cosgrove, 2005; Sánchez-Rodríguez et al., 2010). Under normal pH, transgenic plants overexpressing OsA8 (a rice PM H⁺-ATPase) showed a faster root elongation rate in 7-day-old seedlings and better growth than Nipponbare (WT) plants (Figures 5A,D). The root H⁺ efflux detected using the non-invasive micro-test technology revealed that overexpressing OsA8 in plants enhanced the H⁺ excretion under pH 6 (Figures 5B,C), indicating positive effect of H⁺-ATPase on root elongation and plant growth. However, the OX-A8 plants displayed as severe growth inhibition as the WT plants, which showed short roots and a lower dry weight under alkaline stress (Figures 5D–F). We considered that the suppression of rice growth may be associated with PM H⁺-ATPase activity. Using the non-invasive micro-test technology (Figure 6A), we found that the H⁺ flux rate was extremely influenced by external pH (Figure 6B). The mean rate of H⁺ efflux in the root was decreased from 100 to 0.3 pmol⋅cm^-2⋅s^-1 as the pH was increased from 4 to 8 (Figure 6C). We measured the H⁺-ATPase activity in response to pH 4, 6, and 8, and the results showed that the H⁺-ATPase activity was greatly reduced as the external pH increased from 4 to 8 (Figure 6D). We investigated the expression level of plasma membrane H⁺-ATPase genes under alkaline stress and the results showed that the expression of OsA1, OsA2, OsA3, OsA7 and OsA8 genes was not significantly influenced by external pH (Figure 6E). These results suggest that the H⁺ efflux mediated by H⁺-ATPase was inhibited under alkaline stress, which may be post-transcriptionally regulated by the H⁺-ATPase activity level.

Previous research reported that the interaction between PM H⁺-ATPase and 14-3-3 proteins was essential to the activation of H⁺-pump activity. Fusicoccin (FC), a phytotoxic metabolite produced by Fusicoccum amygdali, activates H⁺-ATPase activity by forming the H⁺-ATPase 14-3-3 complex that is stabilized by FC binding (Baunsgaard et al., 1998). Auxin promotes cell elongation by stimulating cell wall acidification that may be attributed to the auxin-induced activation of H⁺-ATPases according to the Acid Growth Theory. In the present study, rice growth was investigated under alkaline stress with exogenous auxin (IAA), the PM H⁺-ATPase inhibitor vanadate (NaVO₃, VA) and the PM H⁺-ATPase enhancer fusicoccin (FC). Under the normal condition, the plant growth was inhibited by exogenous VA treatment and root elongation was almost completely inhibited (Supplementary Figure S4), indicating that H⁺-ATPases are required for root elongation. In comparison with the simple alkaline stress treatment, exogenous FC and IAA significantly improved primary root elongation and strengthened the tolerance of rice to alkaline stress (Figures 7A–C), indicating the H⁺-ATPase activity was regulated by protein modification under alkaline stress. The results showed that the IAA efflux was decreased at high pH, and the correlation analysis revealed a strongly positive correlation (R²= 0.89, P = 0.08) between the Net H⁺ and IAA flux rates on rice roots 10 mm from the root tip (Figures 7D,E).

These results demonstrated that the inhibition of rice growth was associated with ethylene production and H⁺-ATPase activity was reduced at high pH. We analyzed the relationship between ethylene signaling and H⁺-ATPase activity in alkaline stress-mediated root inhibition by pH indicator staining to visualize rhizosphere acidification. The initial pH of the media was adjusted to 8, appearing purple due to the pH indicator bromocresol purple (0.04 g L^-1). When the medium turns acidic, it turns yellow due to the pH indicator. By applying an ethylene antagonist and ethylene precursor in the medium at pH 8, acidification was detected for the rhizosphere with exogenous 10 μM Ag⁺, an antagonist of ethylene perception, and almost no acidification was found in the rhizosphere with exogenous ACC (Figure 8A), suggesting that the acidification by H⁺-ATPase was negatively affected by ethylene. The rice growth was greatly recovered upon treatment with 10 μM Ag⁺ and the blocking the ethylene signal (ein2-1) under high pH stress (Figure 8B). The ein2-1 mutants, which were less sensitive to alkaline stress than WT plants, displayed equally severe inhibition of growth as the WT plants under an alkaline condition with 100 μM NaVO₃ (Figure 8B and Supplementary Figure S5). Under the normal pH, root growth and gravitropism were inhibited in the presence of exogenous ACC and the plants displayed curly roots (Figure 8C). We measured the root Net H⁺ flux rate and H⁺-ATPase activity with exogenous ACC, both of them were significantly decreased in the presence of ACC (Figures 8D,E and Supplementary Figure S6). These results verified that H⁺-ATPase is involved in ethylene-mediated inhibition of rice growth and ethylene acts as an upstream regulator that represses H⁺-ATPase activity.