Dynamic Aspects of Alcoholic Fermentation of Rice Seedlings in Response to Anaerobiosis and to Complete Submergence: Relationship to Submergence Tolerance

Abstract

Rice plants are severely damaged by complete submergence. This is a problem in rice farming and could be the result, in part, of tissue anoxia imposed by a reduced availability of oxygen. To investigate this possibility we monitored alcoholic fermentation products as markers for tissue anaerobiosis using sensitive laser‐based spectroscopy able to sense ethanol and acetaldehyde down to 3 nl l^–1 and 0·1 nl l^–1, respectively. Acetaldehyde emission began within 0·5 h of imposing an oxygen‐free gas phase environment followed closely by ethanol. As treatment progressed, ethanol output increased and came to exceed acetaldehyde emission as this stabilized considerably after approx. 3 h. On re‐entry of air, a sharp post‐anaerobic peak of acetaldehyde production was observed. This was found to be diagnostic of a preceding anoxic episode of 0·5 h or more. When anaerobiosis was lengthened by up to 14 h, the size of the post‐anaerobic acetaldehyde outburst increased. After de‐submergence from oxygen‐free water, a similarly strong but slower post‐anaerobic acetaldehyde upsurge was seen, which was accompanied by an increase in ethanol emission. Light almost, but not completely, eliminated fermentation in anaerobic surroundings and also the post‐anaerobic or post‐submergence peaks in acetaldehyde production. All photosynthetically generated oxygen was consumed within the plant. There was no substantial difference in acetaldehyde and ethanol output between FR13A and the less submergence‐tolerant line CT6241 under any submergence treatment. In some circumstances, submergence damaged CT6241 more than FR13A even in the absence of vigorous fermentation. We conclude that oxygen deprivation may not always determine the extent of damage caused to rice plants by submergence under natural conditions.

INTRODUCTION

Rice is the only major crop plant that can grow well in flooded conditions. It has a range of features that contribute to this ability. For example, seeds germinating in soil survive submergence by germinating anaerobically. Once germinated, the coleoptile may escape persistent shallow submergence by elongating more quickly and to greater length than normal in response to an absence of oxygen, to partial shortage of oxygen or to internally accumulated carbon dioxide and ethylene (reviewed in Jackson and Pearce, 1991). Older plants, especially of the deep‐water type, are able to survive in several metres of water depth because they elongate their stems in response to partial submergence enabling them to maintain some foliage above the water line (Catling, 1992). This fast underwater elongation requires the presence of oxygen and is hormonally driven, mainly by accumulated ethylene that sensitizes gibberellin‐dependent elongation, cell division and starch mobilization (Kende et al., 1998). In contrast to this growth‐promoting effect of partial submergence, the impact of complete submergence of whole plants is much more severe. Extensive foliar damage takes place if this stress persists for more than a few days (Yamada, 1959; Ellis and Setter, 1999) and, later, the plants may die. In the rain‐fed lowlands of south and south‐east Asia, complete submergence of young plants by temporary flash flooding from overflowing rivers or surface run‐off is the third most severe stress limitation to rice production, after weeds and drought (Mackill, 1986; Setter et al., 1997).

Intolerance to complete submergence is commonly ascribed to adverse reactions to reduced availability of oxygen (e.g. Ellis and Setter, 1999) caused by slow inward diffusion rates compared with those in air (Armstrong, 1979), leading to anoxia. The causes of injury to plants generally from anoxia are not entirely clear (Vartapetian and Jackson, 1997) but are linked to an imbalance between energy demands and the limited ability of fermentation to generate ATP. Alcoholic fermentation generates only four molecules of ATP from each glucose molecule entering glycolysis. Since two ATPs are consumed in steps that phosphorylate hexoses, the final yield is two ATPs from each glucose molecule. This is only about 5·5 % of that generated by the usual aerobic oxidation of glucose via the Krebs’ cycle and the mitochondrial electron transport chain. Although necessary for short‐term survival of anoxia, alcoholic fermentation is clearly an inefficient and limited energy source for plants in the longer term. However, it is by no means certain whether submergence damage and plant death are necessarily linked causally to oxygen deprivation and the associated alcoholic fermentation (Greenway and Setter, 1996). To investigate this point further, we first established the impact of oxygen deprivation, and also of a return to air, on the dynamics of fermentation by monitoring oxygen and the production of ethanol, acetaldehyde and carbon dioxide. This was done using sensitive, on‐line laser‐based or electrochemical trace gas detectors. Features indicative of fermentation were then sought in plants that has been submerged in water sufficiently long to be damaged. Although all rice types are damaged by complete submergence, some unusually tolerant cultivars are known. These are almost all genetically related to FR13A, a selection from a local Indian variety ‘Dhullaputia’ released to Indian farmers in the 1940s (Mackill, 1986). We therefore compared submergence‐tolerant cultivar FR13A with a highly intolerant cultivar CT6241. We used 14‐d‐old plants since other work has shown plants of this age are particularly susceptible to submergence (Adkins et al., 1990; Ram et al., 1999).

Materials and methods

Plant material, germination and plant culture

Seed of the submergence tolerant Oryza sativa L. cv. FR13A, and a susceptible line, CT6241, was supplied by Dr S. Sarkarung, IRRI Thailand Office, Bangkok, Thailand. Seeds were surface sterilized with 1 % sodium hypochlorite solution for 10 min, washed under running tap water for 5 min and placed in 110‐mm‐diameter glass Petri dishes lined with filter paper wetted with 15 ml of tap water. The Petri dishes were placed in dark at 30 °C and relative humidity of 65 %. Sprouted seedlings with 1‐cm‐long coleoptiles were transferred to culture trays (30 × 20 × 15 mm) filled with black ‘Lacqtene’ low‐density polyethylene grains and nutrient solution (major nutrients: 0·849 mm KH₂PO₄, 0·123 mm K₂HPO₄, 1·428 mm NH₄NO₃, 0·754 mm CaCl₂.2H₂O, 0·513 mm K₂SO₄, 1·644 mm MgSO₄.7H₂O; minor nutrients: 9·5 µm MnCl₂.4H₂O, 18·89 µm H₃BO₃, 0·156 µm CuSO₄.5H₂O, 0·152 µm ZnSO₄.7H₂O, 7·484 × 10^–5 µm (NH₄)₆Mo₇O₂₄.4H₂O and 35·75 µm FeEDTA; pH 5·0) (Yoshida, 1976). The strength of the nutrient solution was raised gradually from 25 % of full strength for the first 2 d, followed by 50 % for the next 2 d and finally 100% full strength after 8–10 d. The culture trays were aerated with air flowing through perforated silicone rubber tubing at the base of the tray. After 1 d, pH of the solution in the culture trays was adjusted to 5·0. On the following day, the solution was renewed and the procedure repeated every 2 d. The plants were grown under a 12 h light/12 h dark regime of 28/22 °C (PPFD: 700 µmol m^–2 s^–1; Philips SON‐T Agro400 source) and a relative humidity of 60–65 %. The pH of the solution around the roots increased from 5·0 to 5·8 over 16 h and typical shoot : root ratios (fresh weight) at the start of experiments were 1·72 ± 0·06 (FR13A) and 1·37 ± 0·04 (CT6241).

On‐line detection of acetaldehyde, ethanol, carbon dioxide and oxygen

Acetaldehyde and ethanol down to the nl l^–1 level (0·1 nl l^–1 and 3 nl l^–1, respectively) were measured with a laser‐based trace gas detector (Fig. 1) similar to that described by Bijnen (1995) and Oomens et al. (1998). A similar set‐up was used previously to investigate fermentation by cucumber seeds (Leprince et al., 2000) and red bell peppers (Zuckermann et al., 1997). Briefly, the evolved gases were detected via their absorption of rapidly chopped infrared light which generated pressure variations, resulting in acoustic energy detected by a miniature microphone (Bijnen et al., 1998). The intensity of the generated sound is proportional to the concentration of absorbing trace gas molecules.

Fig. 1. Experimental set‐up. The trace gases emitted by the rice seedlings in the sampling cuvette were transported with the gas flow (regulated by mass flow controllers MC) to the detection cells (PA cells), situated in the CO‐laser cavity. Before entering the detection cells, water vapour in gas flows was removed by a Peltier cooling element (–5 °C) and a cold trap (–45 °C). The cuvette outflow was split into two lines; one connected to the laser‐based detector and the other to the CO₂ analyser (URAS 14). Anaerobic conditions were imposed on the plants by replacing the inflowing air with humid nitrogen gas. Variations in the composition of the carrier gas changed the acoustic behaviour of the detection cell. Two flows, one flow (air) through the sample cuvette and the other flow (nitrogen) through a dummy cuvette were used to ensure that O₂ concentrations in the acoustic detector were always constant; the flows being recombined in a variable ratio using a valve system.

The trace gases emitted by rice plants in the sampling cell were transported by a gas flow to the detection cell, situated within the laser cavity. The source of infrared light was a CO‐laser that was line‐tuneable over a large frequency range in the infrared wavelength region (250 strong output lines between 5·0 and 7·7 µm wavelength) (Urban, 1988). In this frequency range, acetaldehyde and ethanol have a strong and characteristic absorption pattern (Persijn et al., 2000). Since there is a mixture of trace gases in the detection cell and each gas has a different absorption strength on every laser line, we unravelled the mixed absorption strength pattern using a multi‐component matrix calculation algorithm (Meyer and Sigrist, 1990). To determine the concentration of nitrogen gases, the microphone signal on N+1 laser lines was sampled (the additional laser line was used to eliminate instrument effects). The concentration was then calculated taking into account the presence of water vapour and carbon dioxide. The absorption coefficients of acetaldehyde, ethanol, carbon dioxide and water vapour for the six relevant laser lines are given in Table 1.

Table 1.

The absorption coefficients for acetaldehyde, ethanol, carbon dioxide and water vapour at the CO‐laser lines utilized in the analyses, together with their corresponding wavelengths

		Absorption coefficients (atm–1 cm–1)
Laser lines P(J′′)v′′	Wavelength (cm–1)	CO2	Acetaldehyde	H2O	Ethanol
P(7)7	1933·4265	4·74E‐3	0·116	0·00994	0·135
P(11)13	1765·4598	4·6E‐6	30·1	0·0264	0·0279
P(11)19	1616·0405	4·2E‐6	0·424	0·485	0·021
P(8)28	1406·9073	2·8E‐5	6·5	0·0163	2·35
P(8)29	1382·8562	9·57E‐5	5·95	0·0133	2·06
P(11)24	1493·8127	6·3E‐8	0·962	0·03597	0·221

Bold values indicate the highest absorption coefficients for the determination of trace gas amounts, at the corresponding wavelengths.

Water vapour was released in large amounts by the plants and thus had potential to interfere with the measurements. Thus, we dehydrated the gas flow by passing it over a Peltier cooling element at –5 °C and then through a –45 °C cold trap (Bijnen et al., 1996). The laser‐based system was equipped with three detection cells. Three independent samples (gas flows) were therefore handled simultaneously. In all the experiments, two cells were used to measure gas released by rice plants, while the third was used as a reference blank.

Carbon dioxide and oxygen measurements were made simultaneously with acetaldehyde and ethanol measurements using a commercial CO₂–infrared analyser (detection limit 1 µl l^–1), in which an electrochemical oxygen sensor (detection limit 0·01 % oxygen) was incorporated (URAS 14; Hartmann & Braun, Frankfurt, Germany). Oxygen dissolved in the water was measured with a hand‐held dissolved oxygen meter (Oxi 340; WTW GmbH, Weilheim, Germany) sensitive to ≥0·5 %.

Trace gas measurement procedure for gas‐phase tests

Gas‐phase treatments consisted of imposing anaerobic and post‐anaerobic treatments to 14‐d‐old rice seedlings (FR13A) in the dark or in the light using flows of nitrogen gas and air, as appropriate. For each measurement, batches of three plants rather than a single plant were used to generate more trace gas and to minimize effects of differences between individuals. Fresh weight was measured just before each experiment and was about 0·35 g per seedling. The seedlings were placed in a glass cuvette (300 ml) with the roots in 25 ml full‐strength nutrient solution (see Fig. 1). The inlet to the cuvette allowed gas flow treatment, either with pure nitrogen or with air. The outlet flow was divided into two. One gas line was connected to the laser‐based detector and the other to the carbon dioxide and oxygen analysers. In this way, gases of interest emitted or absorbed by the seedlings were monitored simultaneously, on‐line. Switching between nitrogen and air would change the acoustic behaviour of the detection cell because the velocity of sound in air is different to that in pure nitrogen. To overcome the problem, we used two flows, one flow (air) through the sample cuvette and one flow (nitrogen) through a dummy cuvette. After this, both flows were recombined prior to entering the detection system. A valve system was used to interchange the gas flow from the sample cuvette to the dummy cuvette. This maintained the same concentration of oxygen in the gas flow at the point of analysis. At the end of the measurements, plants were transferred back to the culture trays to recover and scored for survival and injury 7 d later. Anaerobic and post‐anaerobic experiments under dark conditions were performed at about 22 °C. When treatments were made in the light, the lamps unavoidably warmed the contents of the cuvette to 27 °C, while maintaining an irradiance of 500 µmol m^–2 s^–1.

Submergence tests

Submergence experiments of 16 h duration were performed in both anaerobic water and initially aerated water, in darkness and in light. To create oxygen‐free conditions, water was first bubbled vigorously with nitrogen gas for 6 h before being used to submerge the plants within a 300 ml cuvette. To prevent floating, plants were secured onto the bottom of the cuvette with fine thread. To achieve total submergence, sufficient water was added slowly into the cuvette by the inside wall to avoid any injury to the plants and to minimize mixing with air. During the 16 h of submergence the cuvette was completely closed and, when required, made dark by wrapping with aluminium foil. The inlet and outlet of the cuvette, closed during submergence, were connected to the measuring system just before the start of desubmergence. Desubmergence was achieved by introducing air from the gas flow system into the cuvette via an inlet; 25 ml of water being left in the cuvette for the roots. To avoid entry of contaminants, no laboratory air entered in the cuvette during the lowering of the water level. As in gas phase experiments, at the end of the measurements the plants were transferred back to the culture trays for recovery and were scored for survival and injury 7 d later. Plant survival was defined quantitatively as the percentage of desubmerged plants that were able to produced one or more new leaves.

Results

Gas emission and uptake under aerobic conditions

In air (21 % oxygen, v/v), very little acetaldehyde (0·04 µl h^–1 g^–1 f. wt) or ethanol (0·3 µl h^–1 g^–1 f. wt) was emitted by rice seedlings of FR13A kept either in the light (500 µmol m^–2 s^–1) or in the dark (Table 2). Carbon dioxide evolution in the dark was 300 µl h^–1 g^–1 f. wt, while under light conditions net carbon dioxide uptake was 300 µl h^–1g^–1 f. wt (Table 2). Total photosynthetic carbon fixation rate (600 µl h^–1 g^–1 f. wt) is estimated to be the sum of uptake of external carbon dioxide in the light and the emission rate at night.

Table 2.

Effect of up to 14 h anaerobic treatment in the dark or in the light on rates of acetaldehyde, ethanol and carbon dioxide production by batches of three 14‐d‐old FR13A rice plants measured at the end of the anaerobic period and again during recovery in air

	Production rates at the end of anaerobiosis (µl h–1g–1 f. wt)				Post‐anaerobiosis (µl h–1g–1 f. wt)
Duration of anaerobiosis	Acetaldehyde	Ethanol	CO2	CO2/ethanol ratio	Acetaldehyde	CO2
Dark conditions
0 h	0·04 ± 0·013	0·3 ± 0·17	300 ± 30	(1·0 ± 0·6) × 103
0·5 h	0·12 ± 0·03	0·38 ± 0·08	299 ± 1	(7·9 ± 1·7) × 102	0·028 ± 0·008	330 ± 20
1 h	1·10 ± 0·05	2·40 ± 0·04	206 ± 2	85 ± 2	0·5 ± 0·1	200 ± 10
2 h	1·00 ± 0·08	20·9 ± 0·2	144 ± 5	6·9 ± 0·2	2·5 ± 0·2	260 ± 15
4 h	0·95 ± 0·06	17·0 ± 1·4	120 ± 15	7·1 ± 1·0	1·40 ± 0·13	180 ± 10
6 h	0·55 ± 0·02	26 ± 3	120 ± 20	4·6 ± 0·9	2·0 ± 0·1	200 ± 50
8 h	0·90 ± 0·13	22 ± 4	100 ± 20	4·5 ± 1·4	2·2 ± 0·2	220 ± 10
10 h	1·07 ± 0·07	43 ± 4	126 ± 4	2·9 ± 0·3	3·3 ± 0·4	200 ± 30
14 h	0·90 ± 0·13	40 ± 7	100 ± 15	2·5 ± 0·6	4·6 ± 0·4	160 ± 40
Light conditions (500 µmol m–2 s–1)
0 h	0·04 ± 0·008	0·27 ± 0·06	–300 ± 15
2 h	0·075 ± 0·005	2·2 ± 0·13	8·1 ± 0·6		0·25 ± 0·02	–311 ± 10
12 h	0·05 ± 0·013	8·0 ± 0·9	11·0 ± 0·5		1·4 ± 0·3	–282 ± 8

Post‐anaerobic increase in acetaldehyde was calculated as the difference between the post‐anaerobic acetaldehyde peak value and the production rate at the end of anaerobiosis.

Post‐anaerobic carbon dioxide data are emissions or uptake values 2 h after the re‐introduction of air to the rice seedlings.

All values are means with standard errors of four or five individual experiments.

Gas emission and uptake under anaerobic conditions in the dark

Emissions of acetaldehyde and ethanol were recorded from seedlings of FR13A exposed to an oxygen‐free external gas phase (nitrogen gas) for 0·5, 1, 2, 4, 6, 8, 10 or 14 h (Fig. 2) in the dark. Each of these separate runs was repeated four or fove times. Their average values and standard errors at the end of the various anaerobic periods are shown in Table 2. Figures 2–4 detail the dynamic changes in emissions by displaying data points separated by only a few minutes. All measurements were on‐line and reveal the time of onset of fermentation by the release of the end product ethanol and its precursor acetaldehyde. In addition, detailed post‐anaerobiosis features were measured, where the time resolution of the photoacoustic technique readily identified a particularly sharp acetaldehyde peak (e.g. Fig. 2C).

Fig. 2. Effect of anaerobic treatment and subsequent relief from anaerobiosis on patterns of ethanol (open circles) and acetaldehyde (closed circles) emissions from single batches of three 14‐d‐old, FR13A rice seedlings measured by on‐line laser photoacoustics. The plants were placed in the dark and given an anaerobic treatment for 0·5 h (A), 1 h (B), 2 h (C), 4 h (D), 6 h (E), 8 h (F), 10 h (G) or 14 h (H) by enclosing them in a cuvette supplied with nitrogen gas flowing at 2 l h^–1 (grey horizontal bars). Afterwards, the plants were returned to a flow of air (2 l h^–1) while ethanol and acetaldehyde output continued to be measured.

Fig. 4. Effect of 2 h or 12 h of anaerobic treatment and subsequent relief from anaerobiosis on rates of acetaldehyde (filled circles), ethanol (open circles), CO₂ (continuous line) and oxygen (filled triangles) emission or uptake by single batches of three 14‐d‐old FR13A rice seedlings. Plants were exposed to light (500 µmol m^–2 s^–1) during the anaerobic treatment (grey bars) and during their return to air.

Thirty minutes from switching the gas flow from air to nitrogen there was little evident effect on emissions. This was principally because of the time taken for the nitrogen to displace oxygen present in the air originally within the cuvette (30 min to reach 0·22 %; Fig. 3A and B). However, after 1 h of nitrogen flow, production of acetaldehyde increased by up to 28‐fold. With longer anaerobic exposures, acetaldehyde emissions declined slightly to level off at approx. 1 µl h^–1 g^–1 f. wt by 3 h. The rise in ethanol output lagged behind that of acetaldehyde by approx. 10 min. The amount produced rose with increasing length of anaerobic treatment (up to 14 h) to over 40 µl h^–1 g^–1 f. wt, after 14 h without oxygen (Fig. 2). At this time, the rate of ethanol production exceeded acetaldehyde production 44‐fold. In contrast, 1 h without oxygen decreased carbon dioxide output by 31 % and by 50 % after 2 h. The decline in carbon dioxide output began within 0·5 h. After 8 h, production was one‐half to one‐third of the normal aerated value (Table 2; Fig. 3).

Fig. 3. Effect of anaerobic treatment and subsequent relief from anaerobiosis on net carbon dioxide production rates (continuous line) and oxygen concentration (filled circles) in the surrounding gas phase of three 14‐d‐old, FR13A rice seedlings. The plants were placed in the dark and monitored for 2 or 4 h in air and then switched to nitrogen gas for 4 h (A) or 12 h (B) (grey horizontal bars) before being returned to air.

If tissue is respiring entirely by alcoholic fermentation (C₆H₁₂O₆ → 2CO₂ + 2C₂H₅OH), the mol ratio of carbon dioxide and ethanol emission (and thus the ratio of the gas volumes produced) should be equal. When anaerobiosis was prolonged beyond 0·5 h, the observed CO₂ : C₂H₅OH ratio decreased from over 1000 initially, to 85 after 1 h, 4·5 after 8 h and 2·5 after 14 h, indicating the steadily increasing dominance of alcoholic fermentation (Table 2). The expected ratio of 1 was not reached, probably because more ethanol than CO₂ was retained in the plant and in the 25 ml nutrient solution that surrounded the roots. During transition from aerobic to anaerobic conditions induced by replacing air with nitrogen gas, we calculated that an oxygen concentration of 0·75 % would have been achieved around the plants after 30 min simply as a result of displacing air in the system by the stream of nitrogen gas. However, a smaller oxygen concentration than this was measured after 30 min (0·22 %; Figs 3 and 4B). The difference between the two values is attributed to oxygen consumption by the plants.

Gas emission and uptake in nitrogen gas and under illuminated conditions

We checked the influence of illumination (500 µmol m^–2 s^–1) since any resulting photosynthesis is likely to oxygenate the seedlings thus decreasing the impact of the anaerobic gas supply on respiratory pathways. Photosynthesis might also be expected to oxygenate the flow of nitrogen gas as it passed over the plants before making contact with the oxygen electrode. The dynamics of oxygen depletion of the gas flow as it was switched from air to nitrogen (Fig. 4B and D) were similar to those obtained in the dark (Fig. 3A and B). In both situations, oxygen dropped close to zero after 2 h of anaerobic treatment. This indicates no output of oxygen from the illuminated plants. This means either that no photosynthesis took place or that, in the absence of externally supplied oxygen, almost all internally generated photosynthetic oxygen was consumed within the plant. The latter explanation is supported by carbon dioxide output data. Thus, when plants were transferred from air to nitrogen gas, in the light, uptake of carbon dioxide by photosynthesis from air was replaced by a very small carbon dioxide emission into nitrogen gas of approx. 11 µl h^–1 g^–1 f. wt (Table 2). This small output was 89 % below anaerobic production of carbon dioxide in the dark (estimated to be 100 µl h^–1 g^–1 f. wt, taken after 14 h in nitrogen). Thus, photosynthetic consumption of all but 11 µl h^–1 g^–1 f. wt of this respiratory carbon dioxide (i.e. almost 90 %) is deduced to have taken place when anaerobically treated plants were illuminated. These results give evidence of strong photosynthetic activity by illuminated rice plants in oxygen‐free surroundings. Further evidence of photosynthetic activity in illuminated plants exposed to nitrogen gas was seen in acetaldehyde and ethanol output data. During 2 or 12 h in nitrogen gas treatment in the light, acetaldehyde emissions increased only slightly compared with aerobic conditions (Table 2) and was only 5 % of the value observed in an anaerobic cuvette in the dark. For ethanol, production was only about 20 % of the rate in darkness. Thus, light decreased alcoholic fermentation by 80 % when the seedlings were exposed to oxygen‐free nitrogen gas. This is the expected outcome if photosynthesis is actively generating oxygen that is utilized through the aerobic respiration pathway. A more detailed inspection of the pattern of ethanol emission in anaerobic surroundings in the light shows that ethanol emissions followed the curve generated under dark conditions for the first 3 h but then failed to increase further in the manner seen in darkness (Fig. 4C). This low, limited and constant production of ethanol (from the plant), and its internal depletion (via the gas flow) in the sampling cuvette were thus in balance during this period. If ethanol production had been completely arrested by light, a decrease—over hours—in ethanol concentration would have been observed. Since no such decrease was seen, some alcoholic fermentation was still present. This may have taken place at the non‐photosynthetic roots.

Post‐anaerobic gas emissions in the dark

Re‐introduction of oxygen after an anaerobic treatment resulted in a temporary upsurge in acetaldehyde emission (Fig. 2). As little as 0·5 h anaerobic treatment was sufficient to induce this post‐anaerobic peak (Fig. 2A), although because of oxygen displacement delays, the actual period of tissue anaerobiosis would have been shorter. This post‐anaerobic acetaldehyde peak occurred even though emissions of acetaldehyde and also ethanol had not increased during the preceding 0·5 h (note the contrasting vertical scales in Fig. 2). When the anaerobic treatment was extended up to 2 h, post‐anaerobic acetaldehyde peaks returned quickly to pre‐anaerobic level (Fig. 2A–C). The peak value and the total output of the post‐anaerobic acetaldehyde outburst (peak area) were increased when the preceding anaerobic treatment was lengthened up to 6 h (Fig. 2D and E). Following periods longer than 8 h of anaerobiosis, a clearly defined peak no longer characterized the post‐anaerobic upsurge (Fig. 2F–H). Instead, acetaldehyde emission remained elevated for many hours, extending beyond the period of monitoring. An upsurge (a peak, or broad maximum) of acetaldehyde that followed exposure to nitrogen gas in the dark is thus shown to be highly diagnostic of preceding tissue anoxia.

When re‐exposed to air, after 0·5 and 1·0 h without oxygen, ethanol production showed a slow increase (Fig. 2A and B). Unlike that for acetaldehyde, ethanol release reached a maximum after 0·5–1·0 h. These broad ethanol peaks indicate implementation and maintenance of fermentation extending beyond the anaerobic period itself (Fig. 2A and B). The post‐anaerobic rise in ethanol emission was not consistent (Fig. 2); for longer anaerobic periods a decline, from a high ethanol level, was typical. Within the first few minutes following transfer from nitrogen to air, a small dip in ethanol emission was usually seen that corresponded to the time of the post‐anaerobic acetaldehyde peak (Fig. 2C).

Re‐exposure to air after anaerobiosis generated a post‐anaerobic increase of carbon dioxide production. Carbon dioxide emissions after re‐exposure to air differed depending on the length of the preceding oxygen‐free treatment (Fig. 3). After short periods (4 h), a sharp outburst took place that levelled off about 2 h later to 60% of the pre‐anaerobic values (Fig. 3A). After longer periods of anaerobiosis (12 h) the initial burst was less prominent. Four hours later, CO₂ production recovered to rates similar to those seen after only 4 h of anaerobiosis (Fig. 3B).

Plant survival rates and leaf injury were recorded 7 d after anaerobic treatments. For 4 h and 12 h anaerobic treatment in the dark, survival was 100%. Leaf injury was not immediately visible on cessation of anaerobic exposure. However, foliar dehydration developed over the following days and the extent of necrosis injury was recorded visually after 7 d. This was 10–15 % of the whole shoot for plants given a 4 h anaerobic treatment and 20–30 % for plants recovering from 12 h of anaerobic treatment (data not shown).

Post‐anaerobic gas exchange in the light

On return to air after 2 h or 12 h in nitrogen gas (all in the light) a small outburst of acetaldehyde was detected (Fig. 4; Table 2). Detailed time course measurements revealed sharply defined peaks even though the absolute emission rates were extremely small from plants deprived of oxygen for only 2 h (Fig. 4A). In contrast to emissions following 10–14 h of anaerobiosis in the dark (Fig. 2), those following anaerobic treatment in the light decreased rapidly after the initial upsurge (Fig. 4C). Post‐anaerobic ethanol production in the light declined rapidly (Fig. 4A), indicating a fast decline of fermentation. Illumination during anaerobic treatment largely prevented direct or indirect damage to the respiratory or photosynthetic apparatus since after 2 h, carbon dioxide consumption rates returned to their initial aerobic value. After 12 h without oxygen, subsequent photosynthetic carbon dioxide uptake diminished only slightly (10 %) compared with the aerobic rate, while oxygen uptake was largely unchanged (Fig. 4B and D). Similarly, light prevented foliar damage. Following 2 h and 12 h of anaerobic treatment in the light, survival was 100 % with plants showing no visible damage to leaves after 7 d recovery in air.

Gas emission and uptake after submergence

We measured post‐submergence emission rates of acetaldehyde and ethanol (in air) following submergence for 16 h. Submergence‐tolerant (FR13A) and submergence‐susceptible (CT6241) genotypes were compared (Table 3). We used the presence of post‐anaerobic emissions of acetaldehyde and ethanol to indicate whether or not plants had actually experienced tissue anoxia and alcoholic fermentation, while submerged under three sets of conditions: (1) plants submerged in oxygen‐free water in the dark and de‐submerged in the dark; (2) plants submerged in oxygen‐free water in the light and de‐submerged in the light; (3) plants submerged in the dark in water initially in equilibrium with air and de‐submerged in the dark.

Table 3.

Average emission values (in µl h^–1 g^–1 f. wt) of acetaldehyde and ethanol over the first 4 h after de‐submergence for the submergence tolerant rice genotype FR13A and submergence intolerant rice genotype CT6241

	Acetaldehyde (µl h–1g–1 f. wt)		Ethanol (µl h–1 g–1 f. wt)		Ethanol/acetaldehyde ratio
Submergence conditions	FR13A	CT6241	FR13A	CT6241	FR13A	CT6241
16 h in anaerobic water, dark	0·87 ± 0·19	1·0 ± 0·3	4·3 ± 0·6	5·0 ± 0·8	4·9 ± 1·3	5·0 ± 1·7
16 h in anaerobic water, light	0·024 ± 0·005	0·04 ± 0·01	0·39 ± 0·19	1·1 ± 0·8	16 ± 8	30 ± 20
16 h in aerated water, dark	0·14 ± 0·08	0·05 ± 0·02	0·4 ± 0·3	0·28 ± 0·13	3 ± 2	6 ± 4
Anaerobic dark/anaerobic light ratio	36 ± 10	29 ± 11	11 ± 6	5 ± 4
Anaerobic dark/aerated dark ratio	6 ± 4	20 ± 10	11 ± 6	18 ± 9

All values are means of four or five individual experiments with standard errors.

Plants submerged in oxygen‐free water in the dark and de‐submerged in the dark.

Immediately on de‐submergence, plants showed a marked post‐submergence acetaldehyde increase (see Fig. 5A). The increase was rapid reaching a broad maximum within 4 h. Ethanol also showed a prominent post‐submergence rise that was initiated 1 h after that of acetaldehyde (Fig. 5B). This ethanol increase was not seen for plants after anaerobic gas phase treatments (e.g. Fig. 2G). Ten to twenty per cent of the leaf area was damaged in FR13A and 50–60 % in CT6241 when plants were de‐submerged and allowed to recover for 7 d.

Fig. 5. Effect of 16 h submergence under three contrasting sets of conditions on post‐submergence emissions of acetaldehyde (A) and ethanol (B) by single batches of three 14‐d‐old FR13A rice seedlings. Filled circles, submergence under dark conditions in water initially de‐oxygenated to < 0·1 % O₂ by bubbling with N₂ gas; filled triangles, submergence under dark conditions in water initially aerated with air (21 % O₂); open diamonds, submergence under light conditions (500 µmol m^–2 s^–1) in water initially de‐oxygenated to < 0·1 % O₂ by bubbling with nitrogen. Light conditions during post‐submergence were as during submergence.

Plants submerged in oxygen‐free water in the light and desubmerged in the light.

During submergence, oxygen was not detectable in the water, indicating the absence of any release of photosynthetically generated oxygen (Fig. 6). On desubmergence after 16 h, there was a very small but almost instantaneous increase in acetaldehyde release (36 times lower for FR13A and 29 times lower for CT6241 compared with submergence in oxygen‐free water in the dark; see Table 3 and Fig. 5A inset) and a slow rise in ethanol production from a very low base (11 and five times lower for FR13A and CT6241, respectively; see Table 3 and Fig. 5B). There was no sign of damage when plants were inspected after 7 d recovery.

Fig. 6. Changes in oxygen concentration in water surrounding batches of three 14‐d‐old FR13A rice seedlings submerged for 16 h under different conditions. Filled circles, initially aerated water in the dark, open triangles, initially oxygen‐free water in the light. For comparison, changes in dissolved oxygen in water initially in equilibrium with air but without plants are also shown (open circles). The inset gives enhanced detail of changes in oxygen concentration in the first hour.

Plants submerged in the dark in water initially in equilibrium with air and desubmerged in the dark.

This treatment was the closest to that likely to be experienced by rice plants in fields that become inundated by floodwater, especially if turbid. Oxygen in the water declined quickly within 30 min (inset to Fig. 6) and then steadied after 3 h around 25 % of the fully aerated value before falling almost to zero by 30–35 h (Fig. 6). After 16 h underwater, there was almost no notable post‐submergence rise in the output of acetaldehyde (six and 20 times lower compared with submergence in oxygen‐free water in the dark for FR13A and CT6241, respectively; see Table 3 and Fig. 5A). This was also the case for ethanol during the first 4 h out of water (11 and 18 times lower). Very small increases were seen then (Fig. 5 insets) but in absolute terms the rates of release were trivial. These data indicated little or no internal anoxia or fermentation had occurred within the plants during submergence. When the plants were allowed to recover for 7 d, the damage to the foliage was evident, being 5–10 % for FR13A and 15–20 % for CT6241.

These three types of experiment were repeated four to five times for FR13A and for the submergence‐intolerant CT6241 (Table 3). There were no statistically significant differences in post‐submergence output of either ethanol or acetaldehyde between the two cultivars. This was the case even though unpublished work elsewhere (V. P. Singh, P. C. Ram, R. K. Singh and B. B. Singh, Narendra Deva University of Agriculture Technology, Faizabad, India) has shown that the shoots of FR13A can contain up to 70 % more starch and respirable sugars than those of CT6241.

DISCUSSION

Rice plants are severely injured when submerged totally in water for several days (reviewed by Ito et al., 1999; Jackson and Ram, 2002; Ram et al., 2002). One of our aims was to examine whether this damage is necessarily linked to the development of anoxic tissue. Evidence for tissue anoxia was sought in emission patterns of ethanol and acetaldehyde that would point to fermentation replacing aerobic respiration with its attendant penalty of poor energy conversion from respirable sugars and potential toxicity effects of acetaldehyde. Because photosynthetic fixation of external carbon dioxide is largely prevented by submergence (Setter et al., 1989) the absolute rate of fermentation may also be important. For example, a slow rate of fermentation could prolong survival by conserving reserves or achieve the opposite by generating even less ATP for maintenance processes. One or other of these possibilities would be supported if cultivars with contrasting tolerance to submergence differ in their rates of fermentation when anaerobic or submerged. The influence of deprivation of oxygen on submerged plants was first established by studying responses to controlled external anaerobic conditions in a gas phase.

Responses to an anaerobic gas phase

Switching from air to nitrogen gas led to acetaldehyde and ethanol release under anaerobic conditions within 30 min (acetaldehyde) and 40 min (ethanol) (Fig. 2). These delays overestimate the time needed by the plants to commence fermentation in response to anoxia because most of this delay is needed for a combination of gas displacement and plant respiration to decrease oxygen concentration around the plants to near zero values. The period over which the plant runs the hypoxic transition from aerobic to the fully anaerobic state was marked by a particularly fast decline in carbon dioxide output (15–30 min). At the end of this period, the change to internal anoxia was probably completed. The subsequent vigour of fermentation, as estimated through ethanol production, increased as anaerobiosis was lengthened by up to 14 h. A notable feature throughout was a strong hangover effect of anoxia on fermentation, ethanol production often rising and remaining elevated for several hours after the re‐introduction of air.

Several physiologically informative features emerged from the measurements. Principal amongst these is the appearance of an immediate post‐anaerobic peak of acetaldehyde emission. This proved to be a highly sensitive diagnostic marker for the occurrence of a preceding period of tissue anoxia. It was more sensitive than emissions of acetaldehyde or ethanol during the anaerobic treatment itself. This is shown in Fig. 2B and C where a clear‐cut post‐anaerobic acetaldehyde peak was generated by dark‐grown plants after only 0·5, 1 or 2 h in nitrogen. When oxygen around the plants was decreased to only 0·15 % rather than to 0 % and prolonged for up to 18 h, no acetaldehyde peak was seen when the plants were returned to air (Fig. 7). In addition, we observed no leaf damage and a 100 % survival. This indicates that the acetaldehyde emission peak is diagnostic of anoxia and distinguishes this from the less damaging hypoxic state even when this is prolonged for many hours.

Fig. 7. Effect of 18 h treatment under low oxygen conditions (0·15 % O₂) on the pattern of acetaldehyde emission from single batches of three 14‐d‐old FR13A rice seedlings measured by on‐line laser photoacoustics (filled circles). The plants were placed in the dark, under 0·15 % O₂ conditions (2 l h^–1) at time t = –18 h. At time t = 0 h, the plants were returned to a flow of air (2 l h^–1). Note the absence of a post‐anaerobic acetaldehyde peak on return to air.

A post‐anaerobic increase in acetaldehyde emission has been observed in other species (Cossins, 1978; Monk et al., 1987a, Zuckermann et al., 1997). On the re‐entry of oxygen into the plant, a small fraction of the accumulated ethanol is seemingly converted back to acetaldehyde. This proposal is supported by our ethanol output data. This almost always showed a dip at the time of the acetaldehyde peak time, which agreed quantitatively (1 : 1) with the acetaldehyde output. Two main pathways have been proposed for the post‐anaerobic acetaldehyde peak: (1) the NAD⁺‐dependent re‐conversion of ethanol back to acetaldehyde catalysed by alcohol dehydrogenase; and (2) H₂O₂‐dependent catalase mediated peroxidation. The H₂O₂ will probably have super oxide radicals as its source, these being derived from the incoming oxygen and converted to H₂O₂ by enzymes in the super oxide dismutase complex Monk et al., 1987b). Indirect evidence for the generation of oxygen free radicals in post‐anaerobic tissue has come from our unpublished finding of a post‐anaerobic release of ethane from rice plants grown under the same conditions used in the present experiments (Santosa, 2002). Ethane is a degradation product of peroxidation of lipid membranes initiated by free radicals of oxygen (Halliwell and Gutteridge, 1989).

Carbon dioxide emission data revealed that quite short periods of anaerobiosis are sufficient to damage respiratory pathways. When rice plants were returned to air after as little as 4 h without O₂, the CO₂ output was little over half that shown before anaerobic treatment. Bertani et al. (1980) also observed this, which may have been a direct effect on biochemical mechanisms of respiration or an indirect outcome of structural tissue damage to mitochondria (Vartapetian and Andreeva, 1986) or tissues. The possibility that it was simply the outcome of time‐dependent substrate depletion can be eliminated since CO₂ production does not slow when well‐aerated plants are incubated in the dark for up to 16 h (result not shown). When our plants were grown‐on for a further 7 d in air, visible damage to the leaves developed. Thus, detection of a post‐anaerobic acetaldehyde peak after 4 h anoxia served as an effective marker for a short episode of anoxia that was demonstrably damaging to the respiratory apparatus and ultimately to leaf survival.

Light completely eliminated the damaging effects on the foliage for up to 12 h in oxygen‐free treatment. This was a consequence of photosynthesis that presumably acted through photosynthetic fixation of respiratory CO₂ to generate oxygen and permitted aerobic respiration while suppressing fermentation. Accordingly, much slower ethanol and acetaldehyde production was measured in the light than in the dark. It was deduced that over 90 % of aerobic respiratory carbon dioxide was taken up. Remarkably, all photosynthetic oxygen was apparently consumed by internal metabolism since no released oxygen gas could be measured. These findings highlight the importance of light for submerged plants under field conditions. However, the extreme sensitivity of our system revealed that light did not entirely eliminate fermentation. A small but measurable production of ethanol during 12 h in nitrogen gas under light was retained. Furthermore, on return to air, a distinct, albeit small, post‐anaerobic‐treatment acetaldehyde peak was discernible. Once more, this indicates the presence of an acetaldehyde peak acting as an extremely sensitive marker for even minor amounts of tissue anoxia. We assume that in the light, the residual anoxia was located in root tips or in the central stele of the roots (Thomson and Greenway, 1991). Here, any photosynthetic oxygen not immediately consumed by foliar respiration would be least likely to penetrate because of the long path length, minimal aerenchyma content and dense packing of small cells.

Responses to submergence

After submergence in oxygen‐free water in the dark, the pattern of acetaldehyde and ethanol release (Fig. 5) was reminiscent of that seen after a similar period of treatment in nitrogen gas (Fig. 2G and H). However, there were some differences. Instead of a rapid increase of acetaldehyde peaking within 2 h after the gas phase anaerobic treatment, the post‐submergence increase was slower, reaching a maximum after 4 h (Fig. 5A). Post‐submergence ethanol release shows a prominent increase 1 h after that of acetaldehyde, a rise that was absent in gas phase experiments (Fig. 2). The differences may be partly explained by impedance to outward diffusion of acetaldehyde imposed by stomatal closure that submergence induces (unpublished result). They may also arise because, in submerged plants, ethanol and acetaldehyde are readily washed‐out by submergence water (Setter and Ella, 1994). Immediately after desubmergence in air, internal levels would be low and time would be needed for internal concentrations to build‐up and be released into the atmosphere around the plants and on to the laser detector.

Nevertheless, submergence in oxygen‐free water in the dark induced a clear post‐stress acetaldehyde and ethanol emission, indicating that these plants experienced anoxia. Accordingly, the amount of damage sustained by the foliage was considerable and greater in the submergence‐susceptible cultivar CT6241 compared with more tolerant FR13A. However, the amounts of post‐anaerobic acetaldehyde and ethanol release were similar in both lines indicating that the basis of tolerance to submergence or anoxia tolerance was unlikely to be a difference in fermentation rate underwater. If plants submerged in anaerobic water were illuminated, damage to the foliage was undetectable 7 d after 16 h submergence treatment. As with anaerobic gas‐phase treatment, post‐treatment acetaldehyde and ethanol releases were almost but not entirely eliminated by light. The residual effect is ascribed to a small amount of anoxic tissue that persisted despite oxygenation by photosynthesis. This residuum is most likely to have been in root tips and vasculature.

In the field, floodwater will not necessarily be devoid of oxygen, as extensive surveys in India and Thailand have shown (e.g. Setter et al., 1987; Ram et al., 1999). Nevertheless, submergence in water containing some oxygen can still be damaging but with FR13A retaining its resilience compared with intolerant lines (e.g. Jackson et al., 1987). If internal tissue anoxia develops under these conditions and is responsible for submergence injury then the presence of a post‐anaerobic acetaldehyde peak would be expected when plants were desubmerged. However, no such peak was seen, even from submergence‐sensitive CT6241 after 16 h underwater. This is approx. 4 h longer than a typical tropical night and thus more likely to promote internal oxygen deficiency than night‐times in the tropics.

We conclude that (1) submergence in water devoid of oxygen and light is highly injurious to rice in association with vigorous fermentation and thus extensive tissue anoxia; (2) when submerged in oxygen‐free water, a differential fermentation response to the stress is unlikely to explain the differences in tolerance between lines such as FR13A and CT6241; (3) the presence of light minimizes submergence damage by depressing, but not completely eliminating, tissue anoxia; this being achieved by means of photosynthetic generation of oxygen utilizing respiratory CO₂; (4) submergence damage caused by water containing at least some oxygen (a common situation in the field) is unlikely to be a consequence of tissue anoxia since no evidence of fermentation was found after 16 h dark submergence. Alternative explanations may involve substrate shortage resulting from inhibited photosynthetic fixation of externally derived CO₂ (Setter et al., 1989) and a utilization of reserves by a well‐documented stimulation of underwater leaf extension that is especially strong in rice lines that are intolerant of submergence (Jackson et al., 1987; Setter and Laureles, 1996).

ACKNOWLEDGEMENTS

This work was supported by a Framework IV European Union INCO‐DC programme as part of the ‘Rice for Life Project’ (ERB3514‐PL95‐0708).

Received: 15 June 2001; Returned for revision: 8 February 2002; Accepted: 13 June 2002