pH regulation in anoxic rice coleoptiles at pH 3.5: biochemical pHstats and net H+ influx in the absence and presence of NOFormula.

During anoxia, cytoplasmic pH regulation is crucial. Mechanisms of pH regulation were studied in the coleoptile of rice exposed to anoxia and pH 3.5, resulting in H(+) influx. Germinating rice seedlings survived a combination of anoxia and exposure to pH 3.5 for at least 4 d, although development was retarded and net K(+) efflux was continuous. Further experiments used excised coleoptile tips (7-10 mm) in anoxia at pH 6.5 or 3.5, either without or with 0.2 mM NO(3)(-), which distinguished two processes involved in pH regulation. Net H(+) influx (μmol g(-1) fresh weight h(-1)) for coleoptiles with NO(3)(-) was ∼1.55 over the first 24 h, being about twice that in the absence of NO(3)(-), but then decreased to 0.5-0.9 as net NO(3)(-) uptake declined from ∼1.3 to 0.5, indicating reduced uptake via H(+)-NO(3)(-) symports. NO(3)(-) reduction presumably functioned as a biochemical pHstat. A second biochemical pHstat consisted of malate and succinate, and their concentrations decreased substantially with time after exposure to pH 3.5. In anoxic coleoptiles, K(+) balancing the organic anions was effluxed to the medium as organic anions declined, and this efflux rate was independent of NO(3)(-) supply. Thus, biochemical pHstats and reduced net H(+) influx across the plasma membrane are important features contributing to pH regulation in anoxia-tolerant rice coleoptiles at pH 3.5.

Introduction

Acidosis of the cytoplasm has been proposed to lead to the death of anoxic maize root tips (Xia and Roberts, 1996). Although this conclusion has been questioned (Greenway and Gibbs, 2003; Felle, 2005), there is agreement that acidosis is a consequence of an inability to maintain cellular compartmentation during anoxia. In this paper, the mechanisms of pH regulation in anoxia-tolerant rice coleoptiles (Kulichikhin ) were further elucidated by exposing excised tips of coleoptiles to a combination of anoxia and pH 3.5. Assessment of functioning of the pHstats (i.e. systems that regulate the pH of the cytoplasm and vacuole to their set points, using biochemical and/or biophysical mechanisms) in rice coleoptiles under the challenging conditions of anoxia and an acid load is relevant to rice seeded directly into some flooded soils (Ismail ). An acid load is an environment that tends to decrease the pH of cells such as exposure to a weak acid or a low external pH. Rice grown in acid sulphate soils (de Datta, 1981; Lang ) will need to tolerate an acid load, and in other flooded soils an acid load may result from high concentrations of organic acids and/or H2CO3 (Ponnamperuma, 1984; Greenway ).

When plant tissues become anoxic, decreases in the pH of the cytoplasm are common (Greenway and Gibbs, 2003; Felle, 2005). In anoxia-tolerant tissues, although cytoplasmic pH declines, it soon stabilizes well above pH 7.0; examples are (i) stem segments of Potamogeton for at least 9 or for 7 h anoxia (Dixon ; Koizumi ), (ii) rice ‘shoots’ (presumably mainly coleoptiles) for at least 14 h anoxia (Menegus ), and (iii) excised tips of rice coleoptiles at pH 6.5, for which a stable cytoplasmic pH has been observed between 60 and 78 h anoxia (Kulichikhin ). Such observations on anoxia-tolerant tissues are consistent with the hypothesis that under anoxia the decrease in cytoplasmic pH will not be perceived as an error signal but reflects a decrease in the set point for cytoplasmic pH (i.e. the pH value that is maintained by a pHstat), consistent with the greatly altered metabolism during anoxia (Felle, 2005).

The putative regulation of intracellular pH can be investigated further by imposing acid loads. pH regulation despite an acid load may be achieved by: (i) a biochemical pHstat, (ii) a biophysical pHstat, i.e. H+ extrusion across the plasma membrane (Smith and Raven, 1979), and/or (iii) a decrease in H+ influx. Rice coleoptile tips are an excellent experimental system for studies of pH regulation during anoxia because these organs can survive at least 90 h anoxia, even when exposed to pH 3.5 for the last 30 h (Kulichikhin ). Evidence for the health of these tissues, even after exposure to pH 3.5 during anoxia, consists of the vigorous net uptakes of K+ and Cl− following a return to aerated solution (Kulichikhin ). When anoxic rice coleoptiles were transferred from pH 6.5 to pH 3.5, the cytoplasmic pH only dropped from 7.35 to 7.2 with a half-time of 0.5 h and then remained steady for at least 17 h, while vacuolar pH decreased from 5.7 to 5.25 with a half time of 1–1.5 h (Kulichikhin ). During the following 15 h, the vacuolar pH remained between 5.3 and 5.4 despite a continuous large net H+ intake of 2.8 μmol g−1 fresh weight h−1 for the first 12 h, and then of 1 μmol g−1 fresh weight h−1 during the last 6 h, at pH 3.5 (Kulichikhin ).

In the present paper, the participation of a biochemical pHstat in pH regulation of anoxic rice coleoptiles was explored by measuring changes in cellular concentrations of organic acids, amino acids, putrescine and selected inorganic ions, as well as net fluxes of H+, K+, NH4+ and NO3−, all when the tissue was anoxic and challenged with an acid load. The acid load was imposed by transfer to pH 3.5. In addition, treatment with or without a NO3− supply indicated a link between H+ and NO3− uptake, presumably via H+–NO3− symports, and showed a reduction in rate of this putative H+–NO3− co-transport with time after transfer to pH 3.5.

Materials and methods

Rice (Oryza sativa L. cv. Amaroo), an anoxia-tolerant cultivar, at least during germination (Huang ), was used in experiments to assess acclimation during anoxia to an acid load, consisting of exposure to pH 3.5 (Kulichikhin ). One experiment used intact seedlings and all others used excised 7–10 mm tips of coleoptiles from ∼3-d-old seedlings. For the excised coleoptile tips, the following were carried out: (i) time courses of ion net fluxes and degradation of organic acids; and (ii) experiments that established to what extent net H+ entry was linked to net NO3− uptake.

Common procedures included treatment in 50 ml conical flasks at 30°C in the dark. The flasks were sealed with rubber bungs and had one inlet and one outlet. All tubing was Tygon, which has a very low permeability to O2. Continuous flushing ensured anoxia in solutions bubbled with high-purity N2 gas. O2 in the outlet of the N2-flushed system was below the detection limit of 0.01% (gas chromatograph; Huang ). The volume of incubation solution depended on the objective. For the intact seedlings it was 50 ml, while for the excised coleoptile tips it varied between 10 and 40 ml, depending on the length of the interval and the requirement to measure appreciable differences over intervals in concentrations of ions and H+ to calculate net fluxes. Uniform bubbling rates of gases (see below) were achieved in the flasks by inserting high resistances in the tubes leading to the individual vessels by using hypodermic needles of 0.4 mm diameter (Teruma G 27). The basal nutrient solution contained (in mM): Ca2+ 0.5, K+ 0.2, NH4+ 0.1, NO3− 0.2, and SO42− 0.45. The basal solution also contained 50 mg l−1 ampicillin and 0.2 mM MES, and the pH was adjusted to 6.5 with Ca(OH)2. When NO3− was omitted from the solution, SO42− was 0.55 rather than 0.45 mM to maintain ion balance.

Solutions were flushed continuously with high-purity N2 and refreshed every 12–24 h; the used medium was removed, without interruption of anoxia, using a 150 mm long hypodermic needle (gauge 14) with a blunt tip and re-injecting solutions that had been pre-flushed with high-purity N2 so that anoxia was continuous. Loss of coleoptiles into the hypodermic syringe was avoided by inserting a thin stainless-steel wire into the tip of the hypodermic needle.

For the intact seedlings, ten seeds in each flask were exposed from the start of imbibition to high-purity N2-flushed nutrient solution at either pH 3.5 or pH 6.5 in both cases unbuffered.

For excised coleoptile tips, the time schedule, including the various treatments, is shown in Fig. 1. Raising seedlings and pre-treatment of excised tips have been described in previous papers (Huang ; Kulichikhin ). Briefly, after 48 h aeration followed by an 18 h hypoxic pre-treatment at 0.05 mM O2, 7–10 mm tips were excised from coleoptiles and 0.1–0.13 g fresh weight was each placed in a 50 ml conical flask. After 5 h healing at 0.05 mM O2, anoxia was imposed. The substrate supply of the seed was replaced by exogenous glucose. Glucose was at 20 mM after excision and was increased to 50 mM during anoxia, because at 20 mM glucose endogenous sugar levels still decrease over several days of anoxia (Huang ). All stages were in darkness.

Fig. 1.

Schematic presentation of the O2 regime, pH treatment, and composition of the medium. (This figure is available in colour at JXB online.)

Treatment at pH 3.5

In all but one experiment with excised coleoptile tips, the pH was changed from 6.5 to 3.5 after 60 h anoxia. Recently, we found that this long pre-treatment under anoxia at pH 6.5 was not required, as similar responses during anoxia and after return to aeration were obtained when the pH was changed to pH 3.5 after only 24 h anoxia at pH 6.5.

At pH 6.5, MES was at 0.2 mM, while at pH 3.5 there was no MES and usually no buffer, as the tissue/volume ratio was kept low enough to keep the external pH within 0.3 units or less from the starting pH. However, 0, 0.2 or 2 mM β-alanine was used in one experiment, as a check, as it was neecessary to relate the present findings to earlier experiments using 31P-nuclear magnetic resonance (NMR) spectroscopy, which used β-alanine to buffer the solutions at pH 3.5 (Kulichikhin ). During preparation of the pH 3.5 solution, the pH adjustment was with H2SO4.

Measurement of H+ net fluxes

Net fluxes of H+ were measured by titration of the medium with 1 mM NaOH or 1–2.5 mM H2SO4, after various periods of exposure of the tissues. For the first sampling period, the collected medium was titrated to the pH of a 10 ml subsample, taken ∼3 min after the change from pH 6.5 to 3.5. This procedure avoided variability due to the residual small volumes of solution used prior to the pH change, from the walls of the flasks and the free space of the tissue. Values were further corrected for blanks, which were flushed for the same period as the solutions containing the tissues. Titration of H+ within 1–2 h after collection of the solutions was required, due to a slow drift in pH of these weakly buffered solutions upon storage.

Preparation of extracts and measurements of ions in nutrient solutions and tissues

Selected ions were measured in the external medium and in tissue extracts. As the primary goal of the experiments was to evaluate the biochemical pHstat, tissues were extracted with ice-cold 5% perchloric acid (PCA) to preserve metabolites (Fan ). Aliquots of extracts were removed for ion analyses prior to the neutralization step, and K+, NH4+, and NO3− were measured in dilutions of these PCA extracts. For metabolite analyses (see below), the PCA extracts were neutralized with K2CO3, centrifuged at 18 000 g, and the supernatant collected.

K+ was determined by flame photometry (Model 410; Corning Medical and Scientific, Cambridge, UK). Cl− was measured using a Buchler–Cotlove chloridometer (Buchler Instruments, Model 4-2008, Fort Lee, NJ, USA). NO3− and NH4+ in most samples were measured with a Skalar Autoanalyzer (San System Plus, Netherlands), with reagents for NO3− (cupric sulphate, hydrazinium sulphate, naphthylethylenediamine dihydrochloride and sulphanilamide) and for NH4+ (sodium nitroprusside and sodium dichloroisocyanate). In some experiments, NH4+ determinations used prussiate phenol reagents (Dorich and Nelson, 1983) at 630 nm with a spectrophotometer (Shimadzu CPS 240a). Inorganic phosphate was also determined in some samples, according to the method of Motomizu and using the same spectrophotometer as above. Total N in the PCA-soluble extract was converted to NO3− by oxidation and then measured on the Skalar Autoanalyzer. In brief, a 1:1 ratio of digestion medium (2%, w/v, K2S2O8 and 3 g NaOH l−1 deionized water) to sample (the sample was first diluted with deionized water as required) was autoclaved at 105°C for 30 min. After cooling, 0.1 ml borate buffer was added ml−1 digest (Ebina ).

Total N and C in the PCA-insoluble fraction were measured with a Macro Elementar Analyser (Model vario Macro CNS, Hanau, Germany). Total protein in the extracted tissues was also measured in some experiments; the pellet remaining after the PCA extraction was washed with deionized water to remove residual PCA and then incubated for 24 h in 1 ml 1 M KOH and the protein analysed according to the method of Lowry and using the same spectrophotometer as above.

In one experiment, protein synthesis was measured by incubating coleoptile tips in anoxia with pH treatments of 6.5 or 3.5 for 4 h in 4 ml incubation solution also containing 0.15 mM leucine labelled with 37 KBq L-(U-C14)-leucine (Amersham, UK). PCA-soluble and -insoluble fractions were counted using Perkin Elmer Ultra cold scintillant (Perkin Elmer, Shelton, USA), at a ratio of 1:2 (extract:scintillant), and a Packard Tri-Carb Liquid Scintillation Counter. The PCA-insoluble pellet was digested in 0.2 ml 1 M KOH and this digest was diluted to 5 ml with deionized water before mixing with 10 ml scintillant.

Measurement of organic and amino acids, and putrescine

For compounds of interest detected in measurable quantities using high-performance liquid chromatography (HPLC) or gas–liquid chromatography/flame ionization detection (GLC-FID) (see below), standards were pipetted onto frozen samples followed immediately by ice-cold 5% PCA. The recoveries (% ±SE) were: malate 90±4, citrate 95±7, shikimic acid 87±5, fumarate 84±3, succinate 87±5, alanine 92±6, and putrescine 88±11. Data were corrected for these recoveries.

Organic acids in tissue extracts and in nutrient solution samples were analysed by HPLC [600E pump, 717 plus autoinjector, 996 photodiode array detector (PDA), Waters, Milford, MA, USA] by slightly adapting the method described by Cawthray (2003). Organic acid standards were: acetic, cis-aconitic, trans-aconitic, citric, iso-citric, fumaric, lactic, malic, maleic, malonic, shikimic, and succinic acids. In brief, separation was achieved at 22±0.5°C on a Prevail C-18 column [250×4.6 mm internal diameter (i.d.) with 5 μm packing; Alltech Associates, Deerfield, IL, USA] with a mobile phase consisting of 25 mM KH2PO4 at pH 2.5 at 1 ml min−1. A gradient-elution program using 60% methanol was used every fifth sample to flush the column of the more hydrophobic compounds and reduce carryover. Detection was at 210 nm, with PDA acquisition from 190 to 400 nm to enable positive identification of organic acids by comparing retention time and PDA peak spectral analyses, including peak purity, of standards with the unknowns. Calibration curves for each organic acid were generated from peak area versus the mass of standard organic acid injected. Data acquisition and processing were with Millenium 32 software version 3.05 (Waters, Milford, MA, USA).

Amino acid analysis by GLC-FID for tissue extracts and nutrient solution samples was conducted with a Phenomenex EZ:faast analysis kit (Lane Cove, NSW, Australia), adapted from the method of Nozal . Amino acid standards were: L-alanine, α-aminoadipic acid, γ-aminobutyric acid, α-aminoisobutyric acid, asparagine, aspartic acid, cystine, glutamic acid, glutamine, glycine, histidine, leucine, allo-isoleucine, isoleucine, lysine, methionine, ornithine, phenylalanine, proline, 4-hydroxyproline, sarcosine, serine, threonine, tryptophan, tyrosine, and valine. A GC-17A (Shimadzu, Kyoto, Japan) coupled with an AOC-20i autoinjector (Shimadzu, Kyoto, Japan) and controlled by GC Solution version 2.30 software was used. The column was a Zebro ZB-AAA amino acid GC column of 10 m×0.25 mm from Phenomenex (Lane Cove, NSW, Australia), with a carrier gas of high-purity He at 1.5 ml min−1. The oven temperature regime was as follows: initial temperature 110°C with a 32°C min−1 increase to 320°C and held at 320°C for 0.50 min. The injection port temperature was 250°C and the detector was held at 320°C. Samples were injected at 1 μl in split mode (1:15 or 1:5). Solutions of known standards at 200 μM were treated in the same way as samples. The peak areas of standard amino acid derivatives were used to quantify the concentrations of amino acids in samples using the internal standard (norvaline) method (see Nozal ).

The above GLC-FID method did not separate serine from γ-aminobutyric acid; thus, separation and quantification of these two compounds was achieved using HPLC, with the Waters AccQ.Tag package. The HPLC system consisted of a 600E dual-head pump, 717 plus autosampler, 470 scanning fluorescence detector, and a Nova-Pak C18 column (150 mm length×3.9 mm i.d.) with 4 μm particle size (all from Waters, Milford, MA, USA), held at 35°C. HPLC data were acquired and processed using Empower chromatography software (Waters, Milford, MA, USA) with a fluorescence detector (excitation: 250 nm; emission: 395 nm; filter: 0.5; gain: 100).

Putrescine analysis involved dansylation (Bartzatt, 2003) and subsequent separation by HPLC (600E pump, 717 plus autoinjector, 996 PDA and 470 fluorescence detectors; Waters, Milford, MA), adapted from the method of Linares . A 200 μl aliquot of standard or sample was vortexed with 200 μl 2 M Na2CO3 and 60 μl dansyl chloride in acetone solution (5 mg ml−1). Sealed samples were then incubated at 40°C, before extraction of the derivitized putrescine into ether (2×200 μl), which was then dried down and the residue redissolved in 600 μl acetonitrile for direct HPLC injection. Separation was achieved with an Alltima C-18 column (150 mm length×4.6 mm i.d.) with 5 μm packing (Alltech Associates, Deerfield, IL, USA), at 22±0.5°C with a flow rate of 1.2 ml min−1 of a mobile phase of 25% Milli-Q water and 75% acetonitrile. Detection was with both the PDA set at 216 nm for peak purity and spectrum matching, and the fluorescence detector (excitation: 320 nm; emission: 523 nm; gain: 100) with outputs used for quantification based on peak area.

Calculations

Data on tissue solutes are presented on a fresh-weight basis. However, for calculations involving relationships between ions and comparison with previously presented pH changes in vacuole and cytoplasm (Kulichikhin ), the solute concentrations on a fresh-weight basis were converted to a protoplast-water basis, by multiplication with a factor of 1.17. This factor is based on 9% dry weight (Alpi and Beevers, 1983, and unpublished data) and 5% free space (Pitman, 1963), and allowing for 1% water adhering to the tissue surface.

The amount of H+ that could be accommodated by the organic acid pHstat was calculated as shown in Appendix 1. The permeability of the plasma membrane to H+ (PH) was calculated according to Lüttge and Higinbotham (1979) and the free energy according to Nobel (1974), taking the previously measured cytoplasmic pH for excised coleoptiles as 3.5 (Kulichikhin ) and a plasma membrane potential (Em) of depolarized cells of rice coleoptiles (Zhang and Greenway, 1995), as the cells at 3.5 are likely to be depolarized as indicated by the K+ effluxes (see Fig. 5). The values of PH and free energy were therefore only approximate; deviations were, however, unlikely to be large (e.g. even if Em was 20 mV different from the assumed value of –80 mV, the free energy would only be ∼7% different from those given in the text). The PH calculations also used unpublished data on cell dimensions to estimate total cellular surface areas in anoxic rice (cv. Calrose) coleoptiles, provided by BJ Atwell (Macquarie University, Australia). The PH was based on net H+ influxes normalized to pH 3.5; normalization was needed for several reasons: the solutions were not buffered, the pH changed by 0.1–0.3 units during incubation, and starting pH differed by a small amount between different treatments and fresh solutions used at different times (described in Fig. 5).

Fig. 5.

Net fluxes of H+ (A), NO3−NO3− (B), K+ (C) and NH4+NO3− (D) in the presence or absence of 0.2 mM NO3−NO3− during anoxia at pH 3.5 or 6.5. Treatments were imposed after 60 h anoxia, before which time coleoptiles were at pH 6.5 and 0.2 mM NO3−NO3−. pH 6.5 with NO3−NO3−, open squares; pH 6.5 without NO3−NO3−, open circles; pH 3.5 with NO3−NO3−, closed squares; pH 3.5 without NO3−, closed circles. The y-axis presents the mean rate of net influx or efflux over the interval. The pH of the solutions was 3.54–3.68 for 0.2 mM NO3− and 3.4–3.5 for solutions without NO3−NO3−, except for the last period when the pH was 3.3–3.5. A second experiment gave similar results. For H+, the least significant difference (l.s.d.) (5%) was 0.124 for pH treatments. For K+, the l.s.d. (5%) was 0.19 for differences between pH and NO3−NO3− comparisons. For NO3−NO3− and NH4+NO3−, the l.s.d. (5%) was 0.114 and 0.087, respectively, for comparisons between pH treatments.

Statistical analyses of data

Data were analysed by calculation of means, standard error of the mean (SE) and analysis of variance where appropriate, using GenStat version 12.1.

Results

Effects of pH 3.5 during anoxia, imposed on germinating seedlings

This experiment: (i) tested whether the previously observed high tolerance of excised rice coleoptile tips to a combination of anoxia and pH 3.5 (Kulichikhin ) also exists in intact, germinating seedlings; and (ii) evaluated the effects on growth and survival.

At the end of 4 d anoxia, the length of the coleoptile was 3.5-fold smaller at pH 3.5 than at pH 6.5; hence, growth was reduced at the lower pH. Nevertheless, the seedlings survived the challenging combination of anoxia and pH 3.5, as shown by rapid elongation of the first leaf after a shift to an aerated solution at pH 6.5 (Table 1A). The shorter leaf length of the seedlings that had been at pH 3.5 during anoxia (Table 1A) was presumably mainly due to the delayed development during anoxia rather than to reduced elongation after the return to aeration at pH 6.5 (Table 1A).

Table 1.

Effects of pH 3.5 or 6.5 on germinating, intact rice seedlings during anoxia started at imbibition

A. Growth
Growth parameter	Treatment
	pH 6.5	pH 3.5
Length coleoptile (mm) after 4 d in anoxia	7.3±0.5	2.0±0.1
Aeration starting at d 5		Returned from pH 3.5 to pH 6.5
Length leaf (mm) after 2 d in aeration following 4 d of anoxia	19.5±1.1	9.6±0.9
B. K+ net fluxes under anoxia
Days after start of imbibition	K+ net uptake or loss (nmol per seedling h−1)a
	pH 6.5	pH 3.5
1	–4.8±2.8	–5.5±0.3
2	–1.6±0.9	–5.1±0.6
3	1.5±0.8	–5.5±0.3
4	1.0±1.6	–4.0±1.6
Total K+ loss over 4 d (nmol per seedling)	94	482

Recovery of shoot growth was measured by a shift to aeration after 4 d of anoxia, after which all solutions were at pH 6.5. Values given are means ±SE of three replicates.

Negative values represent loss to the solution.

The seedlings lost substantial K+ during the first d of anoxia, even at pH 6.5; however, at this pH these K+ losses declined on the second d and subsequently there was a small net K+ uptake during the last 2 d of the experiment (Table 1B). In contrast, at pH 3.5, the net K+ losses continued at the same rate throughout the 4 d of anoxia (Table 1B). The responses for net K+ fluxes in these intact seedlings were similar to those for excised coleoptile tips (Kulichikhin ). The severe decrease in elongation of the coleoptile at pH 3.5 under anoxia contrasts with the stimulatory effects of acidity on extension growth in aerobic oat coleoptiles (Cosgrove, 2001).

Having shown that an acid load can impair the growth of intact coleoptiles of seedlings in anoxia, the further investigations reported in this paper aimed at elucidating the mechanisms of intracellular pH regulation under anoxia, and hence excised coleoptile tips were used. Excised coleoptile tips provide a simpler system than intact seedlings and the data also assisted in elucidating the mechanisms involved in the regulation of cytoplasmic and vacuolar pH in coleoptile tips measured in our previous [31P]NMR spectroscopy experiments (Kulichikhin ).

Protein synthesis

The large effects on growth of the intact coleoptiles at pH 3.5 raised the question of whether in the excised system net protein synthesis was similarly affected. Net protein synthesis was never found at pH 3.5 (four separate experiments). At pH 6.5, there was net synthesis in two experiments, e.g. of 2.3±0.9 mg g−1 fresh weight between 0 and 90 h (Table 2), but no measurable net protein synthesis in two other experiments (data not shown). However, there was turnover of proteins, even in anoxia at pH 3.5, between 24 and 28 h after start of the pH 3.5 treatment: [14C]leucine incorporation, expressed as the percentage of total leucine uptake into the tissues, was 7.4% at pH 3.5 and 7.2% at pH 6.5 (only assessed in one experiment). Thus, overall there was little evidence of any substantial differences in protein synthesis between anoxic coleoptile tips at pH 3.5 and 6.5.

Table 2.

Total protein (mg g−1 fresh weight) in excised rice coleoptile tips before and after exposure to anoxia at pH 6.5 or 3.5

At 0 h, anoxia was imposed, while pH treatments were commenced at 60 h. The protein concentration in the coleoptile tips at the end of the hypoxic pre-treatment and healing (i.e. at time 0 when anoxia was imposed) was 6.9±0.3 mg g−1 fresh weight. For comparison, the protein concentration in coleoptile tips from seedlings germinated and grown continuously in aerated solution at time 0 was 11.3±0.3 mg g−1 fresh weight. Values given are means ±SE of three replicates.

Total protein (mg g−1 fresh weight) with time in anoxia
pH treatment	60 h	78 h	90 h
pH 6.5	8.0±0.9	–	9.2±0.5
pH 3.5 starting at 60 h	–	8.4±0.8	7.6±0.2
pH 3.5 between 60 and 78 h, then pH 6.5	–	–	7.1±0.6

Organic and amino acids

Our first priority was to ascertain to what extent the biochemical pHstat contributed to pH regulation. As mentioned before in an earlier paper, changes in cytoplasmic and vacuolar pH were evaluated in anoxic excised coleoptile tips at pH 3.5, and a substantial net influx of H+ was apparent at pH 3.5 (Kulichikhin ). The present experiments investigated possible changes in the concentrations of organic and amino acids with time of exposure at pH 3.5. The earlier experiments used β-alanine as a buffer at pH 3.5 (Kulichikhin ), while in the present experiments the pH of the external solution was retained within 0.3 units of the starting value by using suitable tissue-to-volume ratios. To facilitate linking of the results of these two investigations, it was therefore first established that β-alanine had no effect on the changes in endogenous organic and amino acids (data in Appendix 2).

We then measured endogenous organic solutes involved in a possible biochemical pHstat for coleoptile tips at pH 6.5 or 3.5 without β-alanine. During the first 60 h anoxia at pH 6.5, there were no consistent changes in malate concentration, which ranged between 9 and 10 μmol g−1 fresh weight, while there was net succinate formation of ∼2.7 μmol g−1 fresh weight (Fig. 2). Exposure to pH 3.5 during anoxia decreased the concentrations of malate from 9–10 to ∼3.8 μmol g−1 fresh weight during the first 18 h, with further decreases to ∼2 μmol g−1 fresh weight during the next 18 h at pH 3.5 (Fig. 2). Succinate also decreased at pH 3.5 although more gradually than malate (Fig. 2). Fumarate was at very low levels but followed a similar trend to malate and succinate (Fig. 2). Anoxia at pH 6.5 greatly increased the L-alanine concentration from 2 to ∼30 μmol g−1 fresh weight over the first 60 h of anoxia and then to ∼43 μmol g−1 fresh weight over the next 40 h (Fig. 3A). After transfer of coleoptile tips to pH 3.5 at 60 h anoxia, the means of three experiments showed no appreciable changes in L-alanine (Fig. 3A), while there was an appreciable increase in a fourth experiment (although this was only about half of the increase at pH 6.5; Fig. 4B). At pH 6.5, γ-aminobutyric acid increased from ∼1.5 to 2.8 μmol g−1 fresh weight during the first 60 h in anoxia; after 60 h the increase at pH 6.5 slowed but accelerated at pH 3.5, and this increase after 60 h was ∼8-fold greater at pH 3.5 than at pH 6.5 (Fig. 3B). Furthermore, the analyses showed that serine increased from 1.6 μmol g−1 fresh weight at the end of hypoxia to 3–4.6 μmol g−1 fresh weight during anoxia, with little difference between pH treatments.

Fig. 2.

Organic acids (μmol g−1 fresh weight) in excised tips of rice coleoptiles under anoxia in the presence of 0.2 mM NO3−NO3− at pH 6.5 and after 60 h at either pH 6.5 or 3.5. The start of anoxia is at 0 h. Malate, squares; succinate, circles; fumarate, triangles; pH 6.5, open symbols, thin line; pH 3.5, closed symbols, bold line. Data are means of three experiments with three replicates in each experiment. Citrate was below the detection levels in all tissues and both treatments (note: recovery of citrate spike was 96±7%). Values at 0 h are tissues sampled at the end of an 18 h pre-treatment at 0.05 mM O2. In one experiment, coleoptile tips were returned from pH 3.5 to 6.5 at 78 h while anoxia was maintained, and after another 18 h the malate concentration was not different to that in tips continued at pH 3.5. For aerated coleoptile tips at 0 h, the organic acids were (μmol g−1 fresh weight, means ±SE): malate, 8.6±0.5; succinate, below detection; fumarate, 0.5±0.2.

Fig. 3.

Amino acids (μmol g−1 fresh weight) in excised tips of rice coleoptiles under anoxia in the presence of 0.2 mM NO3−NO3− at pH 6.5 and after 60 h at either pH 6.5 or 3.5. (A) L-alanine, (B) γ-aminobutyric acid (GABA). The start of anoxia is at 0 h. pH 6.5, open squares, thin line; pH 3.5, closed squares, bold line. Data are the means of three experiments, each with three replicates. Values at 0 h are tissues sampled at the end of an 18 h pre-treatment at 0.05 mM O2. In one experiment, coleoptile tips were returned at 78 h from pH 3.5 to 6.5 while maintaining anoxia, and after another 18 h L-alanine had decreased from 30 to 21 (μmol g−1 fresh weight); no data for γ-aminobutyric acid are available after this shift.

Fig. 4.

Amino acids and total N in PCA-soluble extracts from anoxic coleoptiles, in the presence or absence of 0.2 mM NO3−NO3− and at pH 3.5 or 6.5. From 0 to 60 h, coleoptiles were at pH 6.5 and with 0.2 mM NO3−NO3−, and treatments were imposed after the 60 h anoxia. Open bars indicate incubation without NO3−NO3− and closed bars indicate with NO3−NO3−. Results on ion fluxes from the same experiment are presented in Fig. 5. pH treatments commenced at 60 h after starting anoxia, and the final sampling was 54 h later.

The changes in organic and amino acids at pH 6.5 were consistent with earlier experiments with rice coleoptiles (Fan ). Rates of L-alanine net synthesis observed here were of the same magnitude as in young shoots of intact rice seedlings exposed to anoxia at 3 d after imbibition, which synthesized 35 μmol g−1 fresh weight L-alanine over 48 h anoxia (calculated from Menegus ). The L-alanine accumulation would contribute ∼0.1 MPa to the osmotic pressure compared with the observed osmotic pressures of 0.4–0.6 MPa in anoxic rice coleoptiles (Atwell ; Menegus ). This contribution gains in importance when the tissues are transferred to pH 3.5, as the L-alanine accumulated previously assists in keeping the osmotic pressure at a tolerable level despite losses of other solutes, for example K+ by 0.125 MPa and organic acids by 0.065 MPa (estimated from the present data). Following the transfer of tips from pH 3.5 back to pH 6.5 with continuous anoxia, the organic acids remained at the levels attained at pH 3.5 (Appendix 3), while there was a possible decrease in L-alanine (see caption of Fig. 3).

Some other solutes also changed in concentration during anoxia, although the levels were rather low. Valine and putrescine, present at ∼0.3 μmol g−1 fresh weight at the start of anoxia, increased by 4- and 2-fold, respectively, over the first 60 h anoxia. The value for putrescine was approximately half of the value reported for coleoptiles of intact anoxic seedlings (calculated from Menegus ), but the role of putrescine in the pHstat suggested by Menegus was not evident in the present experiment, i.e. there was no appreciable net putrescine synthesis after transfer to pH 3.5. There were no appreciable changes in asparagine during anoxia or exposure to pH 3.5 (data not shown).

The three experiments on organic and amino acids (Figs 2 and 3) were with tissues supplied with 0.2 mM NO3−. In a subsequent experiment, treatments with or without NO3− were started after 60 h anoxia, and after the next 54 h at pH 3.5 the malate and succinate pools were not detectable with or without NO3− in the medium (data not shown; the detection limits were 8 and 17 nmol g−1 fresh weight for malate and succinate, respectively). The amino acid data for coleoptile tips with or without NO3− (Fig. 4) are relevant to the fate of the absorbed NO3− (rates of NO3− uptake are shown in Fig. 5 and discussed below). Total N in the PCA-soluble extract was higher in the presence of NO3− than in its absence, mainly due to higher levels of L-alanine and serine in the treatment with NO3− (Fig. 4). However, changes in γ-aminobutyric acid were more dependent on the external pH than on the NO3− supply, i.e. exposure for 54 h to pH 3.5 increased γ-aminobutyric acid levels by ∼85 and ∼120% in the absence and presence of NO3−, respectively; in contrast, there was little change in γ-aminobutyric acid at pH 6.5 (Fig. 4).

Importantly, absorbed N did not flow to PCA-insoluble compounds; analysis of the washed and dried pellet remaining after PCA extraction of the coleoptile tips, using a Macro Elementar Analyser, showed that both the amount of N (by dry weight) at ∼3.9% and the C:N ratio of 10.9–11.2 were similar in all treatments. Hydrolysis of some PCA-insoluble compounds is a probable contributor to the increases in PCA-soluble total N after 60 h, as such increases occurred even in the pH 3.5 treatment without NO3− (Fig. 4), for which there was no net N uptake (i.e. net NH4+ uptake was close to zero) (Fig. 5). However, whether hydrolysis occurred could not be ascertained from the present data, as there were only values for the percenatage of N, while no assessments were made for the total amount of N in the PCA-insoluble fraction.

Finally, experiments were carried out to establish what role the organic acid pHstat and the associated K+ fluxes played in the previously observed increase in vacuolar pH following the return from pH 3.5 to 6.5 (Kulichikhin ). Overall, there was remarkably little change in the levels of malate and succinate and K+ fluxes, as shown in Appendix 3.

Net fluxes of H+, K+, NH4+NO3−, and NO3−NO3− during anoxia at pH 6.5 or 3.5

Net H+ fluxes were required to assess to what extent the organic-acid-based pHstat could cope with an acid load. These experiments included treatments of 0 or 0.2 mM NO3−, as net NO3− uptake by anoxic coleoptile tips supplied with 0.2 mM NO3−, at pH 3.5 or 6.5, can be substantial (Fig. 5B). H+ influx into the cells at pH 3.5 (discussed below) would clearly be driven by the large electrochemical potential gradient across the plasma membrane, assessed at 30 kJ mol−1. Furthermore, no organic solutes leaked to the medium (solutions sampled, freeze dried to concentrate and assayed as described in Materials and methods for tissue extracts), indicating that the net H+ decrease in the medium was associated with net H+ influx into the cells, rather than with exchange of weak acids or bases.

Before describing the net ion fluxes during anoxia, we first considered possible injury of the coleoptile tips when challenged by anoxia in combination with pH 3.5, either with or without NO3−. Possible injury was tested by return to an aerated solution after anoxia and the two pH treatments. Based on rates of net uptake of K+ and Cl−, there was no indication of serious injury for either treatment; as in earlier experiments, which only had solutions with 0.2 mM NO3−, net K+ uptake by coleoptile tips with or without NO3− in the incubation medium was rapid after the return to aerated conditions, while net Cl− uptake was ∼30% lower in the tissues that had been exposed to pH 3.5 compared with pH 6.5 but only for the first 4 h after re-aeration (data not presented).

Ion net fluxes at pH 6.5 will be only briefly mentioned, as the central theme of this paper is the response to exposure at pH 3.5. At pH 6.5, net fluxes of H+ and K+ were variable, with net uptakes or losses at different time periods (Fig. 5A, C). There was some net K+ uptake at certain periods: these were between 0.1 and 0.2 μmol g−1 fresh weight h−1 (Fig. 5C), as has been observed previously (Colmer ; Kulichikhin ). Net NO3− uptake was ∼0.5 μmol g−1 fresh weight h−1 (Fig. 5B). Anion uptake under anoxia is not unique to NO3−, as the coleoptile tips in the present experiments showed a net Cl− uptake of ∼0.5 μmol g−1 fresh weight h−1 between 60 and 84 h after the start of anoxia (data not shown; Cl− was supplied in the absence of NO3−, and no measurements were taken after 84 h). Despite the substantial net uptake of NO3−, and for the pH 6.5 treatment also some net uptake of NH4+, tissue concentrations of these ions remained low and independent of pH, being ∼0.2 and ∼0.3 μmol g−1 fresh weight for NH4+ and NO3−, respectively.

At pH 3.5 with 0.2 mM NO3−, the net H+ influx between 0 and 24.5 h was ∼1.55 μmol g−1 fresh weight h−1, and then dropped to ∼0.5–0.7 μmol g−1 fresh weight h−1 between 24.5 and 54 h at pH 3.5; these time trends were similar to those in two other experiments in the present study and in an earlier investigation (Kulichikhin ). In contrast, without NO3−, the net H+ influx for the first 5 h at pH 3.5 was only half of that in 0.2 mM NO3−, and this net H+ influx did not appreciably slow with time (Fig. 5A). Rates of net NO3− uptake during the first 24 h at pH 3.5 were as high as ∼1.3 μmol g−1 fresh weight h−1, being ∼2.5 times higher than at pH 6.5. Similarly, high rates of net Cl− uptake during the first 8 h after exposure to pH 3.5 were observed (measured in the absence of NO3−, data not shown). At pH 3.5 the net NO3− uptake decreased substantially with time, reaching the rate of the tissue at pH 6.5 between 46 and 54 h after the start of exposure to pH 3.5 (Fig. 5B). Net losses of NH4+ were 0.24–0.35 μmol g−1 fresh weight h−1 when NO3− was supplied, with little net loss of NH4+ from tissues not supplied with NO3− (Fig. 5D).

Net K+ loss from the coleoptiles at pH 3.5 was higher in the presence than in the absence of NO3− during the first 5 h, but then became similar during the next 41 h of exposure to pH 3.5 (Fig. 5C). Rates of net K+ loss increased during the last 7 h of the experiment (i.e. between 107 and 114 h after the start of anoxia; in contrast, no such increases were observed in previous experiments).

Ethanol formation

The large differences in solute fluxes between treatments raised the question of whether ethanol formation, which is an indicator of energy production linked to glycolysis (Pradet ; Greenway and Gibbs, 2003), was increased in the anoxic coleoptile tips during an acid load. No substantial effects were found between coleoptile tips in treatments of pH 3.5 and 6.5 (Table 3).

Table 3.

Ethanol production (μmol g−1 fresh weight h−1) by anoxic coleoptile tips at pH 6.5 or 3.5, and with or without 0.2 mM NO3−NO3− in the incubation medium The treatments were started at 60 h after the start of anoxia; up to this point the coleoptiles had been at pH 6.5 and with 0.2 mM NO3−NO3−. Ethanol formation was measured for two 2 h periods during the treatments, the first period started at 6 and the second at 18 h after commencement of the pH treatments. Values given are means ±SE of three replicates (with the mean of the two periods of measurement used for each replicate).

Treatment
	pH 6.5 without NO3−	pH 6.5 with 0.2 mM NO3−	pH 3.5 without NO3−	pH 3.5 with 0.2 mM NO3−
Ethanol production rate (μmol g−1 fresh weight h−1)	5.7±0.7	6.8±1.1	6.1±0.5	5.8±0.3

Discussion

This discussion will argue that pH regulation in anoxic rice coleoptiles involves a pHstat based on organic acids as well as a second pHstat based on NO3− reduction to NH4+. Although anoxic soils typically lack NO3− (Ponnamperuma, 1984), NO3− reduction might be relevant, provided tissues contain substantial endogenous NO3− at the start of anoxia.

For the present metabolic studies, it was crucial to establish that the tissues did not suffer appreciable injury during anoxia lasting up to 120 h, particularly when also exposed to pH 3.5 in the medium. Increases in net K+ loss during the last time period in the experiment with or without NO3− (Fig. 5C), in all but the pH 6.5 treatment without NO3−, may reflect some deterioration of membrane semi-permeability. However, a more sensitive indicator of injury is inorganic phosphate (Pi) loss to the medium (Menegus et al., 1991), and in the present case the Pi net loss during the last 7 h of the experiment was between 0.04 and 0.065 μmol g−1 fresh weight h−1, which was only 0.9–1.4% of the Pi in the tissues. Thus, there was only minor injury, which is further supported by the rapid resumption of K+ and Cl− net uptakes after the resumption of aeration.

Upon exposure of the anoxic coleoptiles to pH 3.5 in the medium, cellular pH decreased during the first 3–4 h (Kulichikhin ), i.e. part of the net H+ intake was absorbed by cellular pH buffering. Cytoplasmic buffering would cope with ∼55% of the net H+ entry for the first 1.5 h, the time over which cytoplasmic pH dropped from 7.35 to 7.2 (Kulichikhin ; for calculation, see Appendix 1). This cytoplasmic buffering may arise for ∼60% from Pi (based on 9 mM Pi in the cytoplasm of anoxic sycamore cells; Gout ), while the other half of the cytoplasmic buffering was presumably associated with cytoplasmic proteins. Furthermore, net H+ influx would also result in H+ being transported across the tonoplast. Vacuolar Pi would not, however, participate in buffering in the vacuole, because at the vacuolar pH of 5.35–5.7 (Kulichikhin ) Pi is nearly entirely in the H2PO4− form (pK1 is 2.1 and pK2 is 7.2; Conn and Stumpf, 1972). We also estimated that, after 18 h exposure to pH 3.5, buffering involving organic acids would only have coped with 1.2–2.5% of the net H+ intake into the vacuole (assessment based on a reduction in vacuolar pH from 5.7 to 5.35 and organic acids remaining after 18 h exposure to pH 3.5; see Appendix 1). This small contribution to buffering was due to depletion of the organic acid pool by the biochemical pHstat over the first 18 h of anoxia (Fig. 2). Depending, however, on the rate of decrease in organic acids, their contribution to buffering may have been greater during the first 3 h of exposure to pH 3.5, i.e. before the bulk of the organic acids had been catabolized. This potential contribution was assessed at maximally 43% of the net H+ intake over the first 3 h of exposure to pH 3.5.

Thus, summing up: (i) internal pH decreases were mitigated by buffering during the first few hours after transfer to pH 3.5, but (ii) the prolonged net H+ intake must have been accommodated by biochemical pHstats, as both cytoplasmic and vacuolar pH remained stable between 3 and 18 h after transfer to pH 3.5 (Kulichikhin ).

Biochemical pHstat involving organic acid metabolism

The biochemical pHstat involving organic acid metabolism is discussed here first, while that based on NO3− reduction will be considered in the next section.

After 18 h exposure to pH 3.5, 25% of the net H+ entry had been neutralized by a pHstat, consisting of organic acids, mainly malate and to a lesser extent succinate. It has been known for decades that, in aerated tissues, malate metabolism is part of a pHstat (Davies, 1980); however, for anoxic plant cells only the switch between lactate and ethanol synthesis has been well established (Davies, 1980). No endogenous products of organic acid catabolism were detected in the tissues (see Results), so the likely fate of the catabolized organic acids is decarboxylation to pyruvate (Beevers, 1960). Under anoxia at pH 3.5, in three out of four experiments most of the pyruvate would have been converted to ethanol, as there was no net synthesis of other end products of anaerobic catabolism such as alanine (Fig. 3). However, in one of the four experiments there was an increase of 4.5 μmol g−1 fresh weight in L-alanine, equivalent to approximately half of the decrease in carbon that had been contained in the catabolized organic acids (Fig. 4B).

The second important difference between the pHstat in anoxic (the present study) and aerated (Hiatt, 1967) cells concerns the fate of the K+. In aerated cells, K+, which balances the organic anions, remains in the cell, contributing to the electrical balance for Cl− (Hiatt, 1967). In contrast, the principal purpose of the pHstat under anoxia at pH 3.5 is to neutralize H+, and catabolism of malate would leave no anion balancing the excess K+. Thus, the plasma membrane would depolarize, leading to opening of K+ channels and efflux of the K+ no longer required to balance the organic anions. This view is supported by the decreases in organic acids and K+ in tissues at the end of 48 h at pH 3.5 (means of three different experiments; Fig. 6), although the decrease in charge of organic anions was usually greater than that of K+, indicating that another cation must also have been involved in balancing the organic anions. Organic acid pools had nearly been depleted by the end of the experiment (Fig. 2) and K+ effluxes had also slowed down (Fig. 5). Further discussion on the ion balance in reference to K+, organic anions and Pi is given in Appendix 4. In keeping with previous findings with maize root tips (Roberts ), γ-aminobutyric acid also contributed to pH regulation in rice coleoptiles, although to a much lesser extent than the organic acids (Figs 3B and 4D).

Fig. 6.

Relationship between K+ and charge of organic anions (mEq l−1 tissue water) in excised rice coleoptile tips. The data combine results from treatments at pH 6.5 and 3.5 of three different experiments, which were all part of this investigation. Treatments were imposed after 60 h anoxia, before which time coleoptiles were at pH 6.5. All treatments were carried out in 0.2 mM NO3−NO3−. The assays were for samples taken between 18 and 48 h after transfer to pH 3.5. The charge of organic acids was calculated at the pH of the vacuole as measured using the method of Kulichikhin , using the concentrations in Fig. 2 and published pKa values.

The organic acid pHstat also probably operates in tissues exposed to pH 3.5 in the absence of NO3−. A strong indication of pHstat involvement was that, at 48 h after the start of the pH 3.5 treatment, malate was not detectable and succinate was greatly decreased in coleoptiles both with and without NO3−. Moreover, K+ net losses from the coleoptiles with and without NO3− were, respectively, 31 and 35 μmol g−1 fresh weight during 54 h of anoxia at pH 3.5. This similarity and the pronounced link between catabolism of organic acids and K+ net losses provide strong evidence for the operation of the organic acid pHstat, in tissues without and with NO3−. Whether the kinetic trends of the decreases of organic acids following exposure to pH 3.5 is also similar with and without NO3− still needs to be established.

Finally, measurements of organic acids and ion fluxes following a change of pH from 3.5 to 6.5 at 78 h with continuous anoxia, i.e. when the organic acid pHstat was nearly depleted, showed that there was no substantial regeneration of organic acids during anoxia (see Appendix 3).

pH regulation in the presence of exogenous NO3−

Nitrate reductase ‘pHstat’:

Reduction of NO3− involves H+ incorporation and so will also function as a biochemical pHstat. NO3− reduction to NH4+ in anoxic rice coleoptiles (Fan ) and in germinating rice seeds (Reggiani ) has been demonstrated by 15N-labelling experiments, and in the present study by NO3− uptake without increased tissue NO3− concentrations, whereas NH4+ was effluxed (Fig. 5, pH 3.5 treatment only) while PCA-soluble total N increased (Fig. 4, treatments with NO3−). Uptake of NO3−, if via a NO3−–2H+ symport, would have added 2×52=104 μmol H+ g−1 fresh weight during 54 h exposure to pH 3.5 (NO3− uptake calculated from Fig. 5B), whereas the overall potential for removing H+ by reduction of this NO3− was assessed at 368 μmol H+ g−1 fresh weight (see Appendix 5). This assessment was based on conversion of 1 mol NO3− to NH4+ with consumption of 10 mol H+ (Hopkins and Huner, 2004):and

Part of the absorbed NO3− reduced to NH4+ was presumably effluxed as NH4+ to the medium between 5 and 54 h anoxia at pH 3.5; this efflux over the whole period was 17 μmol g−1 fresh weight (Fig. 5D), which could account for 32% of the NO3− influx (calculated from Fig. 5B, D). Another 20% of the absorbed NO3− was presumably incorporated into soluble N compounds retained in the tissues (this percentage was calculated from Figs 5B and 4 by taking the difference in total soluble N in the absence and presence of NO3−). In the present experiments, the N flowed for a large part to L-alanine and serine, although at pH 3.5 in other experiments there were no large increases in these compounds (Fig. 3A). The flow of the remainder of the N from the absorbed NO3− might only proceed to NO2–, which may leak from the tissues, as observed for anoxic cereal roots in which ∼95% of the net production was in the external solution (Lee, 1979). At pH 3.5 in the medium, ∼50% of the NO2– would be in the HNO2 form (pKa=3.4) and hence lost during N2 flushing. There may also be loss of other volatile compounds such as NO (based on Igamberdiev and Hill, 2004). These assessments indicate a requirement for 132 μmol reduced nucleotides g−1 fresh weight (calculated from Appendix 5). Our data did not permit establishment of the source of these reduced nucleotides, but a substantial amount may be derived from partial operation of the tricarboxylic acid cycle in anoxic rice coleoptiles (Fan ; also discussed by Ratcliffe, 1997).

Reductions in H+–NO3− symport activity:

H+–NO3− symport presumably has reduced activity after long-term exposure to pH 3.5, as indicated by the concurrent decreases in net H+ and NO3− uptake with time (Fig. 4A, B). Such decreased transporter activity would be consistent with the known closure of ion channels in anoxic animal cells (Hochachka, 1991) and the inferred closure of K+ channels in excised, anoxic rice coleoptiles (Colmer ). A remaining enigma is the H+ influx in treatments with and without NO3− during the final hours at pH 3.5; these were similar in value even though there was still a substantial net NO3− uptake (Fig. 5A, B). The easiest suggestion we can offer is that, even in the absence of NO3−, anion–H+ symport(s) conducts some H+ (the less readily absorbed anion, SO42–, was the major anion present in the incubation solution).

The assessed PH of the plasma membrane of the anoxic rice coleoptile tips during the first 5 h after transfer to pH 3.5 ranged between 0.3×10−6 and 0.43×10−6 m s−1 (for calculation, see Materials and methods). Part of the net entry was presumably associated with an H+–NO3− symport; this influx would account for part of the increase in external pH, as the pH will be determined by the difference between the concentrations of strong cations and anions, i.e. the strong ion difference (Gerendás and Schurr, 1999; Greenway and Gibbs, 2003). NO3− influx through a channel is unlikely, as at the external concentration of 0.2 mM NO3− the influx would be against a steep free-energy gradient (McClure ), because in anoxic rice coleoptiles the steady-state plasma membrane potential is –120 to –130 mV (Zhang and Greenway, 1995). Furthermore, if transport occurred via a channel there would be no apparent reason why NO3− influx during the first period of exposure to pH 3.5 was 2–3-fold larger than at pH 6.5 (Fig. 5).

PH values in the absence of NO3− are most relevant to anoxic soil, as anaerobic soils typically lack NO3− (Ponnamperuma, 1984). There was no decrease over time of net intake of H+ following transfer to pH 3.5 when exogenous NO3− was absent, and the PH of the coleoptiles of 0.06×10−6–0.09×10−6 m s−1 compares with 0.65×10−6 m s−1 for plasma membrane vesicles of Elodea nuttallii (Miedema ) and 1.4×10−6 m s−1 for liposomes (Nichols and Deamer, 1980). Thus, the very low PH of membranes of the anoxic coleoptiles may be a composite of H+ fluxes through the lipid bilayer and embedded proteins in the membrane. The caveat of this suggestion is that PH for the coleoptiles was based on the net H+ flux and therefore would be underestimated if there was substantial H+ extrusion. Nevertheless, even a H+ influx three times higher than the measured net uptake would still give a value for the rice coleoptiles of the same order as the PH determined for the vesicles and liposomes (Miedema ; Nichols and Deamer, 1980).

Is there a contribution from a biophysical pHstat as well as from the biochemical pHstat?

Under aeration, H+ extrusion, i.e. the biophysical pHstat, is the main mechanism for coping with a long-term acid load, while the biochemical pHstat is considered a fine-tuning device (Smith and Raven, 1979). However, it is usually assumed that the plasma membrane H+-ATPase does not function under severe energy deficits, while residual maintenance of the transmembrane H+ gradient is likely to be associated with biochemical pHstats, coined ‘the battery’ (Felle, 2005), as the organic acids would have accumulated in aerobic tissues prior to anoxia.

The discussion in the previous section argued that the net H+ intake at pH 3.5 was accomodated by biochemical pHstats. Of course, it cannot be excluded that the total H+ influx was greater than the net H+ influx, i.e. that there was some H+ extrusion. For example, H+ extrusion (i.e. biophysical pHstat) may come into play once the organic-acids-based pHstat is exhausted. This paper has shown that Cl– can be taken up against a pronounced electrochemical gradient but that this can also be associated with ‘the battery-like system’ discussed by Felle (2005). Thus, testing whether the H+-ATPase at the plasma membrane can become operative would be best when the biochemical pHstat is exhausted.

Perspectives

Mechanisms of adaptation to anoxia:

The acclimation to the combination of pH 3.5 and anoxia by rice coleoptile tips in the present study contrasts with early death of root tips of intact maize and wheat plants, which are injured more rapidly when exposed to anoxia at pH 4.0 rather than pH 5–6 (maize, Xia and Roberts, 1996; wheat, Waters ).

For anoxic coleoptile tips at pH 3.5, the very steep electrochemical gradient for H+ entry across the plasma membrane enabled a surprisingly large net influx of NO3−. The ecological relevance of the pHstat depending on NO3− entry and reduction is not, however, immediately clear, as in anoxic soils NO3− is rapidly reduced and so is scarce or even absent (Ponnamperuma, 1984). Nevertheless, any NO3− in seeds or accumulated in tissues prior to anoxia would be balanced by cations and would provide a powerful pHstat, as indicated in this study by the large H+ consumption provided the NO3− was converted to NH4+ rather than only to NO2–, as happens in anoxia-intolerant species (Botrel and Kaiser, 1997). For anoxia-tolerant species such as rice, such a pool of NO3− may well substantially improve growth under anoxia, as the NO3− reduction would facilitate removal of reducing power, generated by some tricarboxylic acid cycle activity (Fan ) required for net protein synthesis, which is substantial in anoxic rice coleoptiles (Alpi and Beevers, 1983).

One of our initial objectives was to use the exposure to pH 3.5 as a means to impose extra demands for energy to test whether glycolysis linked to ethanol formation increased, or whether survival was compromised. Neither was the case, presumably because the observed increase in net influx of H+ at pH 3.5 was accomodated by a combination of the organic acid pHstat, a pHstat based on reduction of NO3−, and reduced PH at the plasma membrane. Hence, other treatments, likely to require substantial energy, are required to test whether ethanol formation in anoxic coleoptiles at pH 6.5 is below the maximum possible (as suggested by Huang ).

Possible relevance to rice in the field:

This study on pH regulation under anoxia is relevant to rice seeded directly into flooded soils but also to situations other than low soil pH, when an acid load may result from high concentrations of organic acids in the soil, including H2CO3 (Ponnamperuma, 1984; Greenway ). The sharp reduction in development of germinating seedlings in anoxia at pH 3.5 is relevant to the previously mentioned direct seeding of rice in flooded soils (Ismail ), particularly acid sulphate soils, which have a pH between 4.5 and 6.2 during the first 14 d of flooding (de Datta, 1981). These soils occur in many rice-growing areas in South-east Asia, for example in 40% of the rice fields in the Mekong Delta (Lang ).

The results presented here on the importance of the biochemical pHstat indicate that anoxia tolerance may depend on conditions prior to the imposition of anoxia, by increasing the capacity of the pHstat. There may be substantial potential for increasing this capacity; for example, organic acid accumulation in excised barley roots was favoured in the presence of K2SO4 rather than KCl (Hiatt, 1967). A high level of organic acids and/or NO3− would provide a substantial biochemical pHstat, so it would be of interest to establish whether conditions before anoxia, which increase the endogenous concentration of organic acids, would increase the tolerance to acid loads under anoxia.

Appendix Table A1.

Responses of some organic and amino acids (μmol g−1 fresh weight) in anoxic, excised coleoptile tips, to pH 3.5 without or with β-alanine as buffer in the incubation medium

Tissue solutes	pH 6.5			pH 3.5 started at 60 h and sampled at 78 h
(μmol g−1 fresh weight)	0 h	60 h	78 h	0 mM β-alanine	0.2 mM β-alanine	2 mM β-alanine
Malate	6.8±0.9	7.6±0.35	8.1±0.2	1.6±0.35	2.3±0.3	2.3±0.3
Succinate	ND	3.2±0.4	2.7±0.11	1.5±0.35	2.3±0.3	1.6±0.2
L-Alanine	2.2±0.2	31±2.0	33±1.3	27±6.2	35±1.4	28±2.5
Serine	2.5±0.5	5.3±0.8	5.3±0.1	6.7±2.2	8.5±0.5	6.1±0.2
β-Alanine	0.23±0.05	0.29±0.07	0.32±0.1	0.27±0.04	8.9±0.26	10.2±0.3

Time indicates hours after transfer to anoxia. As in the main experiments (Figs 2 and 3), there were only low levels of fumarate, shikimic acid and putrescine, and of these only fumarate was decreased at pH 3.5. ND, Not detected.