Diel O₂ Dynamics in Partially and Completely Submerged Deepwater Rice: Leaf Gas Films Enhance Internodal O₂ Status, Influence Gene Expression and Accelerate Stem Elongation for ‘Snorkelling’ during Submergence

Abstract

Deepwater rice has a remarkable shoot elongation response to partial submergence. Shoot elongation to maintain air-contact enables ‘snorkelling’ of O₂ to submerged organs. Previous research has focused on partial submergence of deepwater rice. We tested the hypothesis that leaf gas films enhance internode O₂ status and stem elongation of deepwater rice when completely submerged. Diel patterns of O₂ partial pressure (pO₂) were measured in internodes of deepwater rice when partially or completely submerged, and with or without gas films on leaves, for the completely submerged plants. We also took measurements for paddy rice. Deepwater rice elongated during complete submergence and the shoot tops emerged. Leaf gas films improved O₂ entry during the night, preventing anoxia in stems, which is of importance for elongation of the submerged shoots. Expressions of O₂ deprivation inducible genes were upregulated in completely submerged plants during the night, and more so when gas films were removed from the leaves. Diel O₂ dynamics showed similar patterns in paddy and deepwater rice. We demonstrated that shoot tops in air enabled ‘snorkelling’ and increased O₂ in internodes of both rice ecotypes; however, ‘snorkelling’ was achieved only by rapid shoot elongation by deepwater rice, but not by paddy rice.

Introduction

Flooding is an abiotic stress which has adverse impacts on crops and is of growing concern for some regions as climate change is predicted to result in increased rain and floods (IPCC, 2007, Pedersen et al. 2017). In water, the diffusion rates of dissolved gases are about 10⁴-fold slower than gases in air, so that submerged plant tissues can become low in O₂ restricting respiration (Armstrong 1979). In addition, restricted access to CO₂ can also limit photosynthesis of submerged leaves (Mommer and Visser 2005, Colmer et al. 2011). Even rice (Oryza sativa L.), which is the only staple crop that grows in waterlogged soils, suffers growth restrictions and yield losses when floods occur which submerge the shoots (Septiningsih et al. 2009, Ismail et al. 2013, Colmer et al. 2014).

Plants possess various traits which enable them to cope with, or avoid, complete submergence (Voesenek et al. 2006, Colmer and Voesenek 2009). Of particular importance are two contrasting responses of either suppression of shoot extension (‘sit-and-wait’) or rapid extension of shoot organs to ‘escape’ by maintaining at least some leaves above water (Bailey-Serres and Voesenek 2008). These contrasting responses can occur for rice ecotypes/genotypes; e.g. ranging from paddy type SUB1 rice with little elongation of submerged shoots to the fast internode elongation of deepwater rice (Hattori et al. 2009, Ismail et al. 2013, Kuroha et al. 2018). Indeed, rice ecotypes cultivated in environments of contrasting hydrological regimes have been described (Kirk et al. 2014) and the rapid internode elongation response specific to deepwater rice during submergence is regulated through deepwater rice specific Quantitative trait loci (QTLs) (Kuroha et al. 2018).

Shoot extension to either reestablish, or maintain, contact of leaves with air enables ‘snorkelling’ of O₂ to the submerged tissues, provided that a porous or even hollow tissues/organs are present along the petioles or leaf sheaths and/or stems (Raskin and Kende 1983, Beckett et al. 1988, Herzog and Pedersen 2014). Interconnected gas-filled spaces provide a low resistance pathway for diffusion of O₂ to the respiring submerged tissues (Armstrong 1979). In addition to porous tissues, superhydrophobic leaf surfaces which retain a thin layer of gas aid internal aeration during submergence of many plants, including rice (Raskin and Kende 1983, Pedersen et al. 2009, Winkel et al. 2013, Kurokawa et al. 2018). These leaf gas films promote O₂ and CO₂ exchange with the surrounding floodwater (Supplementary Fig. S1), so that nighttime O₂ uptake can support respiration in a greater proportion of the plant body and during the day underwater photosynthesis is enhanced (Raskin and Kende 1983, Colmer and Pedersen 2008, Winkel et al. 2014, Kurokawa et al. 2018).

Most paddy rice genotypes respond to submergence by some shoot elongation, but this is often insufficient to restore air-contact and the plants can perish due to carbohydrate starvation (Fukao et al. 2006, Fukao et al. 2012) or post-submergence stress including reactive oxygen species (ROS) damages (Fukao et al. 2011) and adverse water relations (Setter et al. 2010). In stark contrast with paddy rice, deepwater rice is normally capable of keeping up with rising floodwaters by rapid elongation of its internodes (Raskin and Kende 1984, Hattori et al. 2009). The internode elongation of deepwater rice is promoted by the gaseous plant hormone ethylene, the biosynthesis of which is upregulated under low O₂ and high CO₂ concentrations (3% O₂ and 6% CO₂) as compared with in air equilibrium (Raskin and Kende 1984); ethylene accumulates in the submerged tissues owing to the slow outward diffusion when under water. With the upper part of the shoot in air-contact, the porous stem with hollow pith and cortical lacunae can act as a ‘snorkel’ and sustain internal aeration of the tissues under the water (Bleecker et al. 1986, Kende et al. 1998, Nagai et al. 2010, Hattori et al. 2011) (Supplementary Fig. S2A–C). If the water rises quickly and completely submerges the shoot of deepwater rice, surprisingly, shoot elongation then was stunted Raskin and Kende (1984), but this appears not to be a general response since a different variety of deepwater rice did elongate rapidly when completely submerged [video in supplementary materials of Hattori et al. (2009)]. However, this phenomenon had not been studied further for deepwater rice.

Under partial submergence, the ‘snorkelling’ of deepwater rice and O₂ partial pressure (pO₂) dynamics of the internodes have been monitored by discrete sampling of gases from the pith cavity (e.g. Setter et al. 1987, Stünzi and Kende 1989). At night, internodal pO₂ declined to hypoxic levels (6.1 kPa), and recovered to levels around air equilibrium (≈20.6 kPa) during the day. Similarly, deepwater rice in a field in Thailand had pith cavity pO₂ in internodes close to the water surface of near air equilibrium whereas those closer to the soil surface became hypoxic (6–7 kPa) toward the end of the night (Setter et al. 1987). Moreover, gas films on the submerged parts of the leaves, which had tips in air, enhanced O₂ and CO₂ supply as the surface gas layers provided a continuous gas-path connection (Raskin and Kende 1983, Raskin and Kende 1985, Beckett et al. 1988). Internal aeration in completely submerged deepwater rice had not been studied, but in completely submerged paddy rice gas films greatly facilitated entry of O₂ in dark and CO₂ in light periods (Pedersen et al. 2009, Winkel et al. 2013).

Tissue O₂ status will impact on metabolism, and various metabolic fluxes can also be modulated as cells acclimate to O₂ deprivation stress (Bailey-Serres and Voesenek 2008). Sugars are catabolized to pyruvate in the glycolysis pathway and followed, when O₂ is available, by the tricarboxylic acid (TCA) cycle, whereas in the absence of O₂ the TCA cycle is inhibited (Gibbs and Greenway 2003, Lasanthi-Kudahettige et al. 2007, Bailey-Serres and Voesenek 2008, Mustroph et al. 2013). Under O₂ deprivation, fermentation pathways are induced (Gibbs and Greenway 2003) and some tissues exhibit an increase in the rate of glycolysis linked to fermentation (Bailey-Serres and Voesenek 2008, Shingaki-Wells et al. 2014), the so-called ‘Pasteur Effect’ (Gibbs and Greenway 2003).

The abovementioned studies of paddy rice pO₂ dynamics have shed light on the importance of leaf gas films for internal aeration of plants during complete submergence. In the present experiments, we used O₂ optodes inserted into the internodal pith cavity of deepwater rice variety C9285 to monitor the O₂ status of partially or completely submerged plants in light-dark cycles, and in some experiments with leaf gas films intact or experimentally removed. We tested the hypothesis that leaf gas films enhance pO₂ in internodes and found that avoidance of nighttime anoxia (or severe hypoxia) enhances the stem elongation of intact submerged deepwater rice. Indeed, experiments using an excised system demonstrated that although internode extension can proceed at 3 kPa O₂, anoxia inhibited the extension relative to that of internodes in air (Raskin and Kende 1984). In addition, we also compared the diel dynamics in pO₂ in the internode pith cavity of deepwater rice when partially or completely submerged to quantify the beneficial influence of ‘snorkelling’, and compared the ‘snorkelling’ also with that by a paddy rice variety T65. Finally, we assessed the diel gene expression dynamics associated with O₂ deprivation in the internodes of submerged plants to further understand the potential consequences of the distinct O₂ conditions on plant functioning.

Results

Internode elongation in deepwater rice and paddy rice during complete submergence

To assess the submergence-induced shoot elongation responses of the deepwater rice variety C9285 and the paddy rice variety T65, 3-month-old plants were submerged in 150 cm of water in deep tanks for 10 d in natural day–night cycles. Deepwater rice responded with substantial shoot elongation and the uppermost leaf tip emerged above the water after 4 d of submergence and by 10 d many leaves were above the water (Fig. 1A). Total plant height of deepwater rice increased from about 1 m at the initiation of submergence to almost 2 m after 10 d of submergence (Fig. 1B) and the vast majority of the vertical growth was due to stem elongation (Fig. 1C). By stark contrast, paddy rice only elongated slightly during complete submergence (Fig. 1A–C). The shoot elongation rates of paddy rice were only about one-tenth of those of deepwater rice (16 and 133 mm d⁻¹, respectively). These results confirm for this study the contrasting capacities for submergence-induced stem elongation of the deepwater and paddy rice varieties used.

Fig. 1

Responses of shoots of 3-month-old plants of deepwater rice and paddy rice to complete submergence. (A) Example of shoot appearance of deepwater rice and paddy rice after 10 d with shoots in air or shoots initially completely submerged in 150-cm deep tanks. Deepwater rice elongated and emerged typically after 4 d of submergence, whereas paddy rice remained under water; dashed lines indicate water surface. Elongation rates during the first 4 d of complete submergence were 80 and 9.5 mm d^-1 for deepwater rice and paddy rice, respectively [measured in the photo cuvettes in (A)]. (B) Plant height and (C) stem height (total internode length) both after 10 d with shoots in air or shoots initially completely submerged (mean ± SD, n = 4). Different letters indicate significant differences between treatment means (P < 0.05, Tukey’s test).

Leaf gas film retention, underwater net photosynthesis and dark respiration and leaf surface structure

For some paddy rice genotypes, leaf gas films persist only for the first several days of submergence (Winkel et al. 2014). Gas film volumes and retention time had not previously been assessed for a deepwater rice and thus we submerged both the deepwater rice and paddy rice for a period of 2 weeks and measured gas film thickness on leaves every second day. Emergence from the tanks was prevented by netting just beneath the water surface. Leaf gas films thickness did not differ between the two rice types, and for both the gas films decreased in thickness with time and were no longer detected after 6–8 d of submergence (Fig. 2A).

Fig. 2

Leaf gas film retention with time of submergence and underwater net photosynthesis (P_N), in deepwater rice or paddy rice. (A) Gas film thickness on the lamina of the second youngest fully expanded leaf measured with time of submergence (no emergence of the target leaf occurred during the 2 weeks of submergence); gas film retention time did not differ between deepwater rice and paddy rice (P > 0.05 for genotype × time in two-way ANOVA). (B) Underwater P_N of leaf segments taken from the second youngest fully expanded leaf of 3-month-old plants measured at 30°C. The segments were either brushed with DI water so that gas films formed during submergence or they were brush with dilute Triton-X100 to remove hydrophobicity to prevent formation of gas films. Underwater P_N was significantly higher in the presence of leaf gas films for both deepwater rice and paddy rice. Data are means ± SE (n = 4) for both leaf gas film retention time and underwater P_N. Different letters indicate significant differences between treatment means (P < 0.05, Tukey’s test).

Underwater net photosynthesis (P_N) was significantly higher in leaf segments with gas films present compared with those where gas films were prevented from forming under water due to prior brushing with dilute Triton X-100 (Fig. 2B). The positive effect of leaf gas films upon underwater P_N was evident both for the deepwater rice and the paddy rice as rates of leaf segments with gas films present were 2.7- and 2.5-fold higher, respectively, when measured with 100 µM CO₂ in the medium (six to seven times air equilibrium of CO₂). By contrast, underwater respiration in the dark (R_D) of leaf segments was not significantly influenced by the presence or absence of leaf gas films when O₂ was initially at air equilibrium (238 µM O₂ at 30°C) in the medium (Supplementary Fig. S6). Moreover, the leaf surface structure did not appear to differ between deepwater rice and paddy rice (Supplementary Fig. S7).

In conclusion, leaf gas films seem equally important for both deepwater rice and paddy rice in terms of facilitating gas exchange with the surrounding floodwater as underwater P_N was enhanced to the same extent by the presence of leaf gas films for both ecotypes of rice. Similarly, the gas film retention time did also not differ between deepwater rice and paddy rice.

Diel pO₂ dynamics measurement of paddy rice under partial or complete submergence

Diel pO₂ dynamics had been measured in the base of roots of a paddy type, but these plants were 4-week-old and had not developed stem internodes (Winkel et al. 2013). We therefore used O₂ optodes to measure the pO₂ dynamics in the pith cavity of internodes of 4-month-old paddy rice when partially or completely submerged Fig. 3A, so as to enable a later comparison with the pO₂ dynamics in deepwater rice internodes. The paddy rice was older than the deepwater rice since the two ecotypes then each had a stem with internodes so that comparable measurements of pith cavity pO₂ could be taken.

Partially submerged paddy rice had a pith cavity pO₂ at around 10 kPa throughout the dark period and upon sunrise, pO₂ sharply increased and fluctuated around air equilibrium (20.6 kPa) for some hours before declining again (Fig. 3B). For the completely submerged paddy rice, nighttime pO₂ in the pith cavity leveled out at around 2–3 kPa and then steeply rose to hyperoxic levels (i.e. well-above the 20.6 kPa in air) after some hours in the light (Fig. 3B). The higher pO₂ levels in the pith cavity during the darkness in the partially submerged vs. the fully submerged plants demonstrate ‘snorkelling’ (i.e. inward diffusion of O₂ from the air surrounding the emergent leaves) in partially submerged paddy rice.

Fig. 3

Diurnal pO₂ dynamics in the pith cavity of 4-month-old paddy rice during partial or complete submergence. (A) Diagram showing the experimental conditions with the shoot partially submerged or completely submerged; the O₂ optode was inserted into internode No. 3 from the apex and numbers in brackets correspond to pO₂ traces in (B) showing mean pith cavity pO₂ of a diurnal cycle (SD of three replicate plants are shown as bands around the mean); the dashed lines (A) indicate that the panicle was severed and in (B) it indicates atmospheric level of pO₂; mean temperatures were 27°C (day) and 25°C (night). Additional details of the experimental approach are shown in Supplementary Figs. S3, S4.

Diel pO₂ dynamics in deepwater rice during partial or complete submergence

For detached stem segments of deepwater rice, it has been shown that the observed stem elongation as a response to submergence (Fig. 1) is not only promoted by ethylene but also requires O₂ and ceases during anoxia (Raskin and Kende 1984). This prompted us to assess diel pith cavity pO₂ during partial or complete submergence of deepwater rice (Fig. 4A).

Fig. 4

Diel pO₂ dynamics in the pith cavity of 3-month-old plants of deepwater rice during partial or complete submergence. (A) Diagrams of the experimental conditions with the shoot in air, partially submerged or completely submerged. An O₂ optode was inserted into internode No. 3 from the apex and numbers in brackets correspond to pO₂ traces in (B) showing mean pith cavity pO₂ of consecutive diurnal cycles (SD of three replicate plants are shown as bands around the means); mean temperatures were 34°C (day) and 30°C (night). Additional details of the experimental approach are shown in Supplementary Figs. S3, S4.

Plants with the shoot in air showed very little diel variation in pith cavity pO₂ which remained around 19–20 kPa which is slightly below the 20.6 kPa in air (Fig. 4B). During partial submergence, pith cavity pO₂ showed characteristic responses to light and darkness, similar to the dynamics established for the partially submerged paddy rice. At the onset of darkness, pith cavity pO₂ initially declined but soon leveled out and remained at a quasi-steady state of 10 kPa throughout the dark period although with some variation between the replicate plants (Fig. 4B). At sunrise, pith cavity pO₂ increased from approximately 10 kPa to air equilibrium in 4–5 h, leveled out and remained around 20 kPa for another 4 h with some fluctuations likely caused by a fluctuating light environment (see e.g. Rich et al. 2013, Winkel et al. 2013). During the mid-afternoon, pith cavity pO₂ again declined and typically reached 12–15 kPa before the dark period. The night pO₂ was higher in these partially submerged plants (next paragraph) showing the beneficial effect of ‘snorkelling’ via the portion of the shoot above the water.

Completely submerged deepwater rice had pith cavity pO₂ dynamics with similar diel patterns with response to changes in light availability as for the partially submerged plants, but with substantially greater amplitudes. Daytime pO₂ peaked at around 30 kPa never reaching a quasi-steady state but showed the same characteristic fluctuations during the light period likely caused by fluctuating light (Fig. 4B). The accumulating O₂ in the pith cavity leading to hyperoxic conditions strongly indicates photosynthetic production of O₂ during complete submergence, as the outward movement of the O₂ produced would be slower in the submerged plants than for those with some portion of the shoot above water. Interestingly during darkness, pith cavity pO₂ fell significantly below the levels of that in partially submerged plants but generally the pith cavity remained oxic with around 3–5 kPa pO₂ (Fig. 4B).

In conclusion, submergence, whether partial or complete, introduces dramatic diel fluctuations in pith cavity pO₂ due to resistance phenomena caused by the slow diffusion of gases in the floodwater surrounding the shoots. Nevertheless, some O₂ was always present within the pith cavity pO₂ albeit significantly lower in completely submerged plants. The diel patterns of pO₂ in the pith cavity for deepwater rice (Fig. 4) were similar to those of paddy rice (Fig. 3).

Influence of leaf gas films on diel pith cavity pO₂ dynamics in deepwater rice

Gas films retained by superhydrophobic leaf cuticles of submerged wetland plants, including rice, are an important feature which enhances leaf gas exchange with the surrounding water (Raskin and Kende 1983, Pedersen et al. 2009, Kurokawa et al. 2018). We tested the hypothesis that leaf gas films aid internal aeration of the pith cavity of completely submerged deepwater rice, in an experiment where we experimentally removed leaf hydrophobicity by brushing a subset of plants with dilute Triton X-100 (Fig. 5A); this prevented retention of gas films.

Fig. 5

Diurnal pO₂ dynamics in the pith cavity of 3-month-old plants of completely submerged deepwater rice with or without leaf gas films. (A) Diagram of experimental conditions showing (1) control plants with sheaths and leaves brushed with DI water and the shoot in air; (2) plants with sheaths and leaves brushed with dilute Triton-X100 and rinsed with water and the shoot in air; (3) plants with sheaths and leaves brushed with DI water so that gas films were formed when completely submerged and (4) plants with sheaths and leaves brushed with dilute Triton-X100 (and rinsed) to remove hydrophobicity so that gas films did not form on the surface of these plants during complete submergence. An O₂ optode was inserted into the pith cavity of internode No. 3 from the apex. The numbers in brackets correspond to pO₂ traces in (B) showing mean pith cavity pO₂ of a diurnal cycle (SD of three replicate plants are shown as bands around the mean); mean temperatures were 27°C (day) and 26°C (night). Additional details of the experimental approach are shown in Supplementary Figs. S3, S4.

Internode pith cavity pO₂ was measured during a diel cycle and the data show that plants without leaf gas films suffered from severe hypoxia as pith cavity pO₂ fell below the detection level (0.005 kPa) only 2–3 h after the onset of darkness (Fig. 5B). By comparison, the pO₂ within the pith cavity of submerged plants that possessed gas films remained oxic with levels around 1–3 kPa throughout the night (Fig. 5B). The influence of the presence or absence of leaf gas films was also evident during the daytime. At the onset of light, plants with intact gas films soon reached a pith cavity pO₂ of around air equilibrium whereas for plants without gas films the quasi-steady state pO₂ leveled out at around only 10 kPa (Fig. 5B). Moreover, two sets of plants remained with the shoots in air; one set was brushed with dilute Triton X-100 and another with groundwater. Regardless of treatment, these plants with the shoots remaining in air showed very minor diel fluctuation in pith cavity pO₂ and there were no differences between plants brushed with dilute Triton X-100 and plants brushed with deionized (DI) water, showing that the dilute detergent did not adversely influence stomatal function.

In summary, the plants manipulated so as to lack gas films during submergence suffered severe hypoxia (pO₂ < 0.005 kPa) for 9–10 h during the night, whereas plants with intact gas films did not. Leaf gas films also influenced plant pO₂ during the day, being around 20 kPa in the pith cavity of plants with gas films and 10 kPa in those without gas films.

Diel gene expression dynamics in stems of deepwater rice

To assess the influence of severe hypoxia during nighttime caused by removal of gas films on plant functioning, we measured the diel gene expression (transcript abundances) dynamics associated with relevant components of plant metabolism, in stems of submerged deepwater rice (Fig. 6).

Fig. 6

Diel transcript abundance dynamics of genes associated with plant metabolism relevant to low O₂ stress, in stems of deepwater rice in various submergence treatments (controls in air; partial or complete shoot submergence; and complete submergence with gas films removed, -GF). (A) Metabolism pathways associated with energy production with emphasis on the genes investigated in this study. (B to E) heat maps of diel gene expression dynamics associated with (B) sucrose synthesis (sucrose phosphate synthase 1, SPS1), (C) TCA (isocitrate dehydrogenase a subunit, IDHa; pyruvate dehydrogenase E1 A subunit, PDHE1A), (D) glycolysis (sucrose synthase 1 and 2, SUS1 and SUS2; cytosolic pyruvate phosphate dikinase, cyPPDK; phosphofructokinase 5, PFK5; phyrophosphate-fructose-6-phosphate-phosphotransferase A3, PFPA3) and (E) fermentation (pyruvate decarboxylase 1, 2 and 3, PDC1, PDC2 and PDC3; alcohol dehydrogenase 1 and 2, ADH1 and ADH2). Deepwater rice plants were submerged partially or completely with or without gas films, then sampled at each indicated submergence timepoints. Plants with shoots in air served as controls. Gene expressions associated with glycolysis and fermentation were most highly upregulated in completely submerged deepwater rice when gas films were removed during the night. Log 2-fold changes of each gene expression were indicated by representative colors using three biological replicates with two technical replicates. Light intensity and temperature dynamics in air and water throughout the measurement are shown in Supplementary Fig. S5. Values of averages and standard deviation of each gene expression analysis are shown in Supplementary Table S2.

Gene expression associated with sucrose synthesis (SPS1) and the TCA cycle (IDHa and PDHE1A) exhibited similar diel patterns in the stems of all plants regardless of time of day or submergence treatment (Fig. 6B, C). By contrast, expression of genes associated with glycolysis (SUS1, SUS2, PFK5, PFP3A and cyPPDK) and fermentation (PDC1, PDC2 PDC3, ADH1 and ADH2) were gradually upregulated during the night but the magnitude depended on submergence regime (Fig. 6D, E). The expression levels of these genes were highest during the night for completely submerged deepwater rice when gas films were removed. Similar expression patterns were also observed for anoxia-induced genes (NIA2, nsHb1, nsHb2 and OVP3) (Supplementary Fig. S8). These results suggest severe hypoxia induced anaerobic metabolism during the night in completely submerged plants when gas films were removed.

Influence of leaf gas films on stem elongation of deepwater rice

Our finding that gas films affect plant metabolism during the night prompted us to test the hypothesis that leaf gas films also influence the ability of deepwater rice to elongate during complete submergence. Submerged plants with intact gas films elongated 70% more than plants where the gas films had been removed (calculated from Fig. 7). These results indicate that stem elongation of submerged deepwater rice is influenced by the O₂ status of the internodes, which in turn is enhanced by the presence of gas films on the leaves of submerged deepwater rice.

Fig. 7

Stem height (total internode length) of 1-month-old plants of deepwater rice as influenced by submergence and the presence or absence of leaf gas films. Plants with shoots in air or submerged had been brushed with either DI water or dilute Triton-X100 and rinsed (the latter removed hydrophobicity and prevented formation of gas films during submergence) and initial total internode length was measured. Complete submergence lasted 7 d (no reemergence occurred) where after total internode length was again recorded. The box-whisker plot shows mean (+), median (horizontal line), 50% of the observations (box) and minimum or maximum (bars); n = 6–8 depending on treatment. Different letters indicate significant differences between treatment means (P < 0.05, Tukey’s test).

Discussion

Our findings on shoot elongation of completely submerged deepwater rice add knowledge to previous work on partially submerged deepwater rice and the associated elegant studies of excised internodes under defined conditions (see Introduction section and below). Surprisingly, in an earlier study, the deepwater rice did not elongate when intact and completely submerged (Raskin and Kende 1983), whereas in the present experiments fast shoot elongation occurred and moreover the presence of leaf gas films enhanced this underwater shoot growth. Comparisons of diel O₂ status within the pith cavity of internodes of completely and partially submerged deepwater rice demonstrated the importance of leaf gas films (completely submerged rice) or ‘snorkelling’ (partially submerged rice) for nighttime O₂ supply. These O₂ supply differences between respective conditions affected diel gene expression dynamics in the stems during the night, deprivation of gas films dramatically induced anoxia associated genes in particular. Similar patterns of internode pO₂ and of leaf gas film retention in deepwater rice and paddy rice when submerged emphasize that the major difference in the submergence responses between these two rice ecotypes is the amount of shoot elongation.

The shoots of completely submerged deepwater rice elongated at 80 mm d⁻¹ (caption of Fig. 1) and at 163 mm d⁻¹ in earlier work [assessed by us for plants when still under water in the video in Hattori et al. (2009)]. For comparison, reported shoot elongation rates of partially submerged deepwater rice are 85 mm d⁻¹ (Raskin and Kende 1983) and with a maximum rate of 250 mm d⁻¹ (Vergara et al. 1976). Thus, the shoot elongation rates of completely submerged deepwater rice appear to be similar to those of partially submerged plants with the shoot tops above water. Internode elongation of partially submerged deepwater rice has been studied in detail (e.g. Keith et al. 1986, Stünzi and Kende 1989, Kende et al. 1998, Lorbiecke and Sauter 1999, Hattori et al. 2009), but genotypic variation in the elongation response of completely submerged deepwater rice should be evaluated since the variety used here (C9285) had a fast elongation response whereas cv. Habiganj Aman II did not (Raskin and Kende 1983). This difference in underwater shoot elongation response between C9285 and Habiganj Aman II is somewhat surprising as both are Bangladeshi Aman season deepwater rice varieties. A recent study, however, also found stem elongation rates can vary substantially for deepwater rice varieties (Kuroha et al. 2018). The influence of environmental conditions (especially light and water chemistry) on the elongation responses of completely submerged deepwater rice should be investigated, including the influence on tissue sugars and the signaling networks [summarized by Kuroha et al. (2018)], for a number of deepwater rice varieties.

Responses of deepwater rice internodes to various gas compositions was studied by Raskin and Kende (1984) using excised stems and showed that internode elongation occurs at 3 kPa O₂ but ceases in anoxia. Pith cavity pO₂ in partially or completely submerged deepwater rice both showed large diel fluctuations (highest during the day, lowest during the night), although with greater amplitudes in the completely submerged plants. Even in the completely submerged plants, the pith cavity pO₂ never fell below 3–5 kPa (Figs. 3, 4) which should be sufficient for internode elongation (Raskin and Kende 1984). Leaf gas films were essential for the maintenance of these O₂ levels within the internodes during the night, and thus also influenced stem elongation of completely submerged deepwater rice (Figs. 5, 7). A previous study had demonstrated the beneficial effect of leaf gas films on stem elongation of partially submerged deepwater rice (Raskin and Kende 1983). Here, we also demonstrate that influence of the measured dynamics in O₂ status on gene expressions in stem tissues of completely submerged deepwater rice; transcript abundances of O₂-deprivation associated genes were low during the day but increased during the night, consistent with the much reduced pith cavity pO₂ during the night (Figs. 4, 6D, E and Supplementary Fig. S8). Therefore, trends toward aerobic metabolism during the day and at least some anaerobic metabolism, in addition to aerobic respiration (R_D), during the night are observed in the stems of completely submerged rice plants. Furthermore, gene expressions associated with O₂ deprivation were more extensively upregulated in the stems during the night when gas films were removed from the leaves (Fig. 6 and Supplementary Fig. S8). Hence, the adverse effect on stem elongation caused by removal of leaf gas films from completely submerged deepwater rice was likely due to the nighttime periods of severe hypoxia (or anoxia) in the internodes and the resulting restriction of R_D. This is evidenced by the increased expression of genes associated with anaerobic metabolism, as compared with deepwater rice with intact leaf gas films that would have greater O₂ entry to sustain significant R_D in internodes due to pith cavity pO₂ being at 1–5 kPa.

Leaf gas films eventually disappeared within 8 d of submergence in both rice ecotypes (Fig. 2A); such loss of gas films would reduce underwater P_N (Fig. 2B) and although R_D could continue in leaves in water with O₂ near air equilibrium (Supplementary Fig. S6) [which is above the critical O₂ pressure for leaf R_D (Verboven et al. 2014)], lower amounts of O₂ reached the distal parts of the internal diffusion path of plants lacking gas films, such as the stems (Fig. 5) and roots (Winkel et al. 2013). The loss of leaf gas films with time of submergence has previously been observed for rice (Winkel et al. 2014, Herzog et al. 2018, Kurokawa et al. 2018), wheat (Konnerup et al. 2017, Winkel et al. 2017) and wild wetland plants (Winkel et al. 2016). A mechanistic understanding of the process by which leaf gas films diminish over time is lacking; but, Konnerup et al. (2017) speculated, based on findings of reduced abundance of wax platelets on leaves when in a high humidity condition (Koch et al. 2006), that the otherwise continuous movement of wax precursors to the cuticle as occurs in transpiring leaves could cease when evaporation from leaves stops during submergence.

Submergence depth influenced pith cavity pO₂ and gene expressions in stems of deepwater rice, demonstrating the capacity for (and importance of) ‘snorkelling’ in the plants with air-contact by the shoot tops. Partially submerged plants maintained pith cavity pO₂ at around air equilibrium during the day (Figs. 3, 4) in accordance with observations for upper internodes of partially submerged deepwater rice in a field situation (Setter et al. 1987) and also in a tank experiment in a controlled environment (Stünzi and Kende 1989). Pith cavity pO₂ in partially submerged plant remained at around 10 kPa during the night (Figs. 3, 4). The diel amplitudes (minimum to maximum) of 12 kPa in the partially submerged plants in the present study were similar to the 15 kPa reported in Stünzi and Kende (1989) whereas the field measurements showed little or no diel amplitudes in the upper internodes (Setter et al. 1987). These differences between studies are likely due to variation in the depths below the water surface of the measured internodes and the leaves emanating from each node. The internode measured and the associated two leaves were under water in the present study, whereas diagrams in the two earlier publications indicate emergence of the upper leaf (Stünzi and Kende 1989) and even the top part of the measured internodes (Setter et al. 1987). Furthermore, anaerobic metabolism associated genes were gradually upregulated in the stems with gradual decreases in pith cavity pO₂ during the night, as influenced by submergence treatments (Figs. 4, 6 and Supplementary Fig. S8). These results further demonstrated the importance of ‘snorkeling’ by the emerged shoot tops for maintaining aerobic respiration within the internodes of deepwater rice.

The data on pO₂ in internodes of paddy rice were taken to enable a comparison of its internal aeration with that of deepwater rice. Previous work on completely submerged paddy rice had used much younger plants and with a focus on roots (Pedersen et al. 2009, Winkel et al. 2013). Our comparison here of deepwater rice and paddy rice shows that apart from the large difference in the response of stem elongation to submergence, the shoots of the two ecotypes were similar for other key characteristics of leaf gas film retention, underwater P_N, leaf R_D, leaf surface structure and internode pO₂ dynamics. Both rice ecotypes retained leaf gas films for 6–8 d of submergence (Fig. 2A), which compares with 4–7 d in four genotypes of paddy rice submerged in a field pond (Winkel et al. 2014). For both ecotypes the rates of underwater P_N were greater when leaves possessed gas films (Fig. 2B), as found also for other genotypes of rice (Pedersen et al. 2009, Winkel et al. 2013, Winkel et al. 2014, Kurokawa et al. 2018). The underwater P_N values were, however, toward the lower end of rates measured for submerged rice which was likely due to the external CO₂ being 100 µM compared with 200 µM or more in previous studies (Pedersen et al. 2009, Winkel et al. 2013, Winkel et al. 2014, Kurokawa et al. 2018) (Fig. 2B).

In conclusion, deepwater rice elongated its shoots during complete submergence and emerged after 4 d so that some leaves were above the water surface. Stem elongation of completely submerged deepwater rice was enhanced by the presence of leaf gas films and we propose that the stimulating effect would be due to nighttime O₂ entry so that anoxia (or severe hypoxia below 0.005 kPa) is avoided within the stems. This oxic status within stems would enable aerobic R_D during the night even when under water. The removal of gas film from leaves diminishes O₂ entry and would inhibit R_D so that anaerobic metabolism must then occur, but consequently internode elongation is also retarded or suppressed. In this context, rice varieties with long gas film retention may have greater tolerance of floods. The ‘snorkelling’ effect of shoot tops above the water and O₂ diffusion through interconnected gas-filled spaces along the body of the plant substantially increased pith cavity pO₂ in both deepwater and paddy rice. However, the much lower capacity for stem elongation of paddy rice precludes ‘snorkelling’ by this ecotype during deep prolonged floods (Fig. 8), further highlighting the importance of the stem elongation trait for deepwater rice which has also recently been emphasized. in genetic analyses by Kuroha et al. (2018).

Fig. 8

Conceptual model showing the responses of deepwater rice and paddy rice to a flood event. Both ecotypes possess leaf gas films during submergence, a feature which can greatly enhance O₂ and CO₂ exchange between tissues and floodwater. Some of the O₂ which enters the leaves can diffuse to the stems and then down to the roots, which can sustain aerobic respiration within the tissues that the O₂ can reach, even when plants are completely under water and in darkness. However, the leaf gas films do not persist beyond 6–8 d of submergence (see Fig. 2A) and to sustain internal aeration of submerged tissues through interconnected gas-filled spaces during prolonged floods, extensive shoot elongation is necessary in order to restore shoot contact with the atmosphere (see Fig. 1 and the effect of ‘snorkelling’). Deepwater rice responds to submergence by rapid internode elongation, so that shoot tips emerge above the water. Paddy rice can only elongate to some extent and in deep floods the elongation is insufficient to restore shoot contact with the atmosphere. As a consequence, paddy rice perishes during prolonged deep floods after the leaf gas films have been lost, which results in poor gas exchange with the floodwater. When shoot extension occurs, this increases the rate of carbohydrate consumption and eventual ‘carbohydrate starvation’ in submerged paddy rice (Fukao et al. 2006, Bailey-Serres and Voesenek 2008).

Materials and Methods

Plant materials and growth condition

We used a deepwater rice ecotype C9285 (O. sativa L., var. C9285 syn. Dowai38/9) and a paddy rice ecotype T65 (O. sativa, var. Taichung 65). C9285 and T65 exhibit similar growth pattern when they are grown in air, however only C9285 shows sufficient internode elongation in response to submergence regulated through deepwater rice specific QTLs (Raskin and Kende 1984, Hattori et al. 2009). Because the abovementioned two genotypes have been used as representative ecotypes as a comparative set to study deepwater rice traits (Raskin and Kende 1984, Hattori et al. 2009), in the following, we will refer to these as ‘deepwater rice’ and ‘paddy rice’. Seeds were sterilized by heating at 60°C for 10 min and were then stored at 4°C for 24 h. The seeds were imbibed for germination in DI water at 30°C for 72 h. Germinated seeds were sown in trays (30 seeds per cell, cell volume 4.0 × 4.0 × 4.5 cm) filled with a soil suitable for rice (with N, P and K at 0.25, 0.3 and 0.25 g/kg soil, respectively, Mikawabaido, Medel Ltd.). After 1 month in a glasshouse at 23°C, seedlings of both rice ecotypes were transplanted into a field with shallow (10 cm) standing water and grown for 2 months (June and July, Aichi, Japan). Deepwater rice at the late vegetative stage and paddy rice at the reproductive stage were transplanted in ‘blocks’ of soil from the paddy field into pots (15.9 cm in diameter, 19.8 cm in depth) using additional soil from the paddy field to fill any gaps within the pots. The stages were selected based on the two ecotypes of rice each having a stem, so that comparable measurements of pO₂ in the pith cavity of internodes could be taken. After 1 week of recovery and continued growth, the plants were used in experiments.

Plant height and total internode length measurement

Thirteen-week-old plants were used for evaluation of plant height and stem length (total internode length) as a response to complete submergence. Plant height was measured from soil surface to the tip of the youngest leaf and stem length was measured from soil surface to the youngest node (Supplementary Fig. S2). The measurements of stem lengths were destructive as the sheaths were sliced away from around the stems so that the position of each stem node was visible. Initial measurements were taken on 10 plants and then two additional batches of 10 plants were each either grown with the shoots in air or with the shoots initially completely submerged in 3,000-L tanks, 180 cm deep containing groundwater (alkalinity = 0.7 mmol L⁻¹; pH in air equilibrium = 7.9). Plants of deepwater rice that were initially completely submerged typically emerged after 4 d of submergence. After 10 d, plant height and stem length were again measured.

One-month-old plants of deepwater rice (6–7 leaf stage) were raised in pots of soil in a glasshouse, and then used to assess the influence of leaf gas films on internode elongation. The entire shoot body surface (sheaths and leaves) was brushed with 0.05% v/v Triton-X100 in groundwater to remove hydrophobicity in order to prevent gas film formation during submergence (Supplementary Fig. S1); plants brushed with groundwater served as controls. After brushing, plants were either grown in the air or completely submerged in 1,250-L tanks containing groundwater. The plants remained completely submerged throughout the 7 d of submergence treatment. After 7 d, total internode length (i.e. stem height) was again measured.

Diel pO₂ dynamics during various submergence regimes

Three-month-old plants of deepwater rice and 4-month-old plants of paddy rice were used to measure diel O₂ dynamics in the stem pith cavity during partial or complete submergence (Supplementary Figs. S3, S4). Measurements were carried out using potted plants placed into 1,250-L tanks in a glasshouse (Togo field of Nagoya University, Aichi, Japan) from August to September 2015 (Supplementary Fig. S3A) and largely followed the approach of Pedersen et al. (2016). In detail, O₂ optodes embedded in 500 µm hypodermic needles (OXF500PT, Pyroscience, Aachen, Germany) were calibrated in DI water at atmospheric equilibrium (20.6 kPa pO₂) and in 0.1 M sodium ascorbate in 0.1 N KOH (0.0 kPa pO₂) immediately prior to use; the drift in signal was <1% per 24 h and hence we did not perform any correction for drift in signal even after prolonged measurements lasting up to 72 h. Before insertion of the O₂ sensor, a 0.45 µm hole was made in the internode at the point of insertion using a hypodermic needle. The O₂ sensor was inserted into the pith cavity of the target internode and sealed with silicone rubber at the insertion point (Supplementary Fig. S4); the position of the internodes measured is given in each figure caption. The signal of each O₂ sensor was logged using an optode meter (FireStingO₂, Pyroscience, Aachen, Germany) with a frequency of one sample per minute. Submergence was imposed by adding groundwater (alkalinity = 0.7 mmol L⁻¹; pH in air equilibrium = 7.9) to the tank (Supplementary Fig. S3). For the partial submergence treatment, approximately 70% of the shoot height was submerged and with the upper leaf blades in atmospheric contact (Fig. 4A). For complete submergence, the entire shoot was completely submerged with the uppermost leaf tips 10 cm below the surface of the water (Fig. 4A). During light periods (7 am to 7 pm), supplementary light was provided using metal halide lamps in addition to natural light in the glasshouse. During dark periods, the area with the tanks inside the glasshouse was covered with black curtains (7 pm to 7 am), so that the plants were in darkness (Supplementary Fig. S3).

Leaf gas film thickness

Three-month-old plants of deepwater rice and 4-month-old plants of paddy rice were used to assess the gas film thickness and its retention during submergence. Gas film thickness on the second youngest fully expanded leaf at the time of submergence (target leaf) was assessed every 2 d for completely submerged plants; the target leaf remained completely submerged for the whole of the 2-weeks duration. Gas film thickness was measured using the buoyancy method of Raskin (1983). In brief, the buoyancy of 5-cm long leaf lamina segments taken approximately one-third from the tip of the target leaf was measured in DI water with the sample mounted on a hook underneath a four-digit balance with measurements before and after brushing both surfaces of the lamina with 0.05% v/v Triton-X100 in DI water. Leaf gas film thickness (m) was calculated as gas volume (m³) divided by two times the projected area of the lamina samples used (m²).

Underwater net photosynthesis and dark respiration

Three-month-old plants of deepwater rice and paddy rice were used to assess the underwater net photosynthesis (P_N) and dark respiration (R_D) of leaf segments according to the methods of Pedersen et al. (2013). In detail, for each replicate, one 2.5 cm leaf lamina segment of the second youngest fully expanded leaf was placed in a transparent glass vial (25 mL) with artificial floodwater prepared according to the medium of Smart and Barko (1985). pO₂ of the artificial floodwater was initially reduced to approximately half air equilibrium (10–11 kPa pO₂) by purging 1:1 volumes with air or N₂; this procedure was applied to prevent increase in O₂ above air equilibrium levels during measurements, that might have led to photorespiration and thus decreased P_N (Setter et al. 1989). The medium was then adjusted to an alkalinity of 1.0 mM at pH 7.35 yielding 100 µM dissolved CO₂. Half of the sample vials contained a leaf segment with gas films (control) whereas the other half contained leaves which had been brushed on both sides with 0.05% v/v Triton-X100 in DI water and then rinsed in artificial floodwater, so that these did not possess gas films; vials without tissues served as blanks. The glass vials were mounted on a vertical disc which rotated inside a water bath at 30°C with light (PAR = 1,000 µmol m⁻² s⁻¹). After incubation of approximately 1 h (the exact time was taken), dissolved O₂ in each bottle was measured using an O₂ mini-electrode (OX-500, Unisense A/S, Aarhus, Denmark). Rates of underwater net photosynthesis (µmol O₂ m⁻² s⁻¹) were calculated based on one-sided lamina area. Underwater R_D was measured as O₂ consumption in darkness following the same procedure as for P_N except that samples were incubated in artificial floodwater that was initially at air equilibrium (20.6 kPa pO₂), and they were incubated for approximately 3 h in order to obtain sufficient differences in pO₂ in vials with or without tissues.

Scanning electron microscopy of leaf surfaces

Surface structures on the adaxial surface of leaf blade of paddy rice, T65, and deepwater rice, C9285, were visualized using a scanning electron microscope, S-3000N (Hitachi, Japan). Approximately 1 cm² of the middle part of a young fully expanded leaf blade was sampled. Scanning was operated at an accelerating voltage of 15 kV under high vacuum.

Real-time quantitative PCR

Submergence treatments of deepwater rice were imposed using the same methods as for the experiments in which diel pO₂ dynamics were measured (see above). Light intensity and temperature dynamics in air and water throughout the measurements are shown in Supplementary Fig. S5; data were logged every 15 min during the experiment using HOBO Pendant (Onset, Bourne, MA, USA). Measured light intensity was converted from lux to photosynthetic photon flux density (µmol m^–2 s^–1) using the conversion factor by Thimijan and Heins (1982). We sampled stem segments (internode and nodes) of deepwater rice when shoots were in air or following submergence, in a glasshouse. After sampling a stem section at the indicated submergence time points (Fig. 6), total RNA was isolated using the Maxwell RSC plant RNA kit (Promega, Madison, WI, USA). Stem sections from shoots in air were used as control. First-strand cDNA was generated using Omniscript RT kit (Qiagen, Hilden, Germany) and real-time quantitative PCR analysis was performed using StepOnePlus (Thermo Fisher, Waltham, MA, USA). The gene-specific primers used are listed in Supplementary Table S1. Ubiquitin was used as the reference gene. Each 10-µL reaction mixture contained THUNDERBIRD SYBR qPCR Mix (TOYOBO, Osaka, Japan), 0.3-µM gene-specific primers and 1× ROX reference dye (TOYOBO, Osaka, Japan), and reactions were performed according to the manufacturer’s instructions. The relative mRNA expression levels of each gene of interest were normalized to Ubiquitin gene expression levels in three biological replicates each consisting of two technical replicates.

Statistical analyses of data

Graphpad Prism 7.0 was used for statistical analyses. Figure captions provide details on the various tests used along with levels of significance.

Funding

The research was supported by JSPS KAKENHI [grant No. JP16K18565], and a MEXT Grant-in-Aid for Scientific Research on Innovative Areas [grant No. 17H06473] and SATPEPS by JST and JICA. O.P. was supported by The Carlsberg Foundation [grant No. CF17-0067] and the Scandinavia-Japan Sasakawa Foundation.

Acknowledgments

C9285 was kindly provided by the National Institute of Genetics in Japan (http://www.shigen.nig.ac.jp/rice/oryzabase/top/top.jsp) (January 21, 2019, date last accessed).

Disclosures

The authors have no conflicts of interest to declare.