Nitric oxide emission from tobacco leaves and cell suspensions: rate limiting factors and evidence for the involvement of mitochondrial electron transport


(fax +0 931 888 6158; e-mail


Quantitative data on nitric oxide (NO) production by plants, and knowledge of participating reactions and rate limiting factors are still rare. We quantified NO emission from tobacco (Nicotiana tabacum) wild-type leaves, from nitrate reductase (NR)- or nitrite reductase (NiR)-deficient leaves, from WT- or from NR-deficient cell suspensions and from mitochondria purified from leaves or cells, by following NO emission through chemiluminescence detection. In all systems, NO emission was exclusively due to the reduction of nitrite to NO, and the nitrite concentration was an important rate limiting factor. Using inhibitors and purified mitochondria, mitochondrial electron transport was identified as a major source for reduction of nitrite to NO, in addition to NR. NiR and xanthine dehydrogenase appeared to be not involved. At equal respiratory activity, mitochondria from suspension cells had a much higher capacity to produce NO than leaf mitochondria. NO emission in vivo by NiR-mutant leaves (which was not nitrite limited) was proportional to photosynthesis (high in light +CO2, low in light −CO2, or in the dark). With most systems including mitochondrial preparations, NO emission was low in air (and darkness for leaves), but high under anoxia (nitrogen). In contrast, NO emission by purified NR was not much different in air and nitrogen. The low aerobic NO emission of darkened leaves and cell suspensions was not due to low cytosolic NADH, and appeared only partly affected by oxygen-dependent NO scavenging. The relative contribution of NR and mitochondria to nitrite-dependent NO production is estimated.


Based on a vast amount of literature, it is now beyond doubt that nitric oxide (NO) is an important second messenger in plants. NO appears involved in controlling various aspects of plant pathogen resistance, growth, development and senescence or stomatal movement (Beligni and Lamattina, 2000; Dangl et al., 1996; Delledonne et al., 1998; Durner and Klessig, 1999; García-Mata and Lamattina, 2002; Leshem and Haramaty, 1996; Neill et al., 2002). For understanding the multiple physiological roles of NO in plants, it is important to clarify by what reactions, in what situation and how much NO is produced and how NO production is controlled.

Probably plants have several enzyme systems to produce NO. The most intensively studied one is nitrate reductase (NR; EC: which catalyzes the reduction of nitrite to NO at the expense of NADH (Rockel and Kaiser, 2002, and literature cited therein). In certain situations like infection by pathogens, inducible nitric oxide synthase-like (NOS; EC: enzymes appear responsible for NO production, converting l-arginine into citrulline plus NO (Delledonne et al., 1998; Zeidler et al., 2004). Other NOS-type activities appear to participate in NO signaling of stomata (Guo et al., 2003), but NR has also been considered as an NO source, both in pathogen–plant systems (Yamamoto et al., 2003) and with respect to NO in stomata (Desikan et al., 2004). In addition, at least under anoxic conditions, xanthine oxidase/dehydrogenase (like NR a MoCo-enzyme) has been suggested to reduce nitrite to NO (Godber et al., 2000; Li et al., 2001; Millar et al., 1998). Primarily in roots, yet an additional, plasma membrane (PM)-associated enzyme system was suggested to reduce nitrite to NO at the expense of NADH (Meyer and Stöhr, 2002; Stöhr et al., 2001). More recently, non-enzymatic production of NO was suggested to occur in the acidic apoplast of some tissue types like aleurone layers (Bethke et al., 2004).

Various direct and indirect methods have been used to detect NO in tissues, cells, solutions and in the gas phase, which have been summarized recently (Lamattina et al., 2003). All of them have distinct advantages and disadvantages, and at least those based on fluorescing indicators are at best semiquantitative.

Nitric oxide is an uncharged gaseous free radical with moderate water solubility (1.9 mm in pure water at atmospheric pressure). At nanomolar concentrations of NO, the physical half-life in pure water is in the range of hours (Henry et al., 1997). However, as living cells contain and produce many compounds that react rapidly with NO, for example, reactive oxygen species, sulfhydryl groups and heme-compounds, the half-life of NO in vivo may be in the range of a few seconds only (Lancaster, 1997). Accordingly, much of the NO produced may be trapped within the cells, thus escaping detection based on emission. It is, therefore, not enough to just measure NO released from plants, tissues or cells. Attempts are required to estimate also how much of the produced NO is actually trapped inside the cells and how much is released, in order to quantify NO production by plants.

Here we summarize data on NO emission into purified air by tobacco leaves and cell suspensions, measured by the ozone-collision chemiluminescence method, which specifically detects NO in the gas phase down to the ppt range. Using different N-nutrition regimes and various tobacco mutants, we first show how this NO emission is modulated by internal and external factors, examine by which reactions NO is produced under a given condition and estimate internal NO scavenging.


NO emission from detached tobacco leaves

In previous work, we have shown that NO emission from leaves was due to the activity of NR, and that it was closely related to the NR activation state and to nitrite contents of the tissue (Morot-Gaudry-Talarmain et al., 2002; Rockel et al., 2002). Confirming and extending this, Figure 1 shows a typical time course of NO emission into the gas phase by detached leaves of WT and mutant tobacco plants. Consistent with previous results obtained with sunflower and spinach leaves (Rockel et al., 2002), tobacco leaves with the petioles in nitrate solution (10 mm) emitted only small amounts of NO in the dark, with rates in the range of 0.01–0.05 nmol g−1 FW h−1. Their NR activity (+Mg2+) was 1.6 μmol g−1 FW h−1 in the dark. Thus, NO emission rates in the dark (air) were more than 5 orders of magnitude lower than NR activity. NO emission increased strongly in the light. Return to darkness was typically accompanied by a transient increase in NO emission (‘light-off peak’, compare Rockel et al., 2002), and then came back to the very low dark level. After flushing WT leaves in the dark with nitrogen, the NO emission rate increased up to 1000-fold, reaching a steady-state of about 100 nmol g−1 FW h−1, or about 2% of NR activity in dark, nitrogen (4.44 μmol g−1 FW h−1). This increase was accompanied by a large increase in the nitrite concentration but only a minor increase in NR activity. NO emission thus at first sight appeared to be mainly determined by the tissue nitrite concentration, as previously described (Morot-Gaudry-Talarmain et al., 2002; Rockel et al., 2002). However, as will be shown below, this was not totally true, as other factors also controlled NO emission.

Figure 1.

NO emission from detached tobacco leaves during a dark-light (air) and air-N2 (dark) transient.
Leaves (1–3 g FW) from WT (Nicotiana tabacum cv. Xanthi) (bsl00066), or from an NR-deficient double mutant (nia30) (N. tabacum cv. Gatersleben) (○), and from tungstate-grown WT plants (bsl00001) were mounted in the chamber with petioles in their respective nutrient solutions. Leaves from WT ‘Gatersleben’ are not shown here, but their NO emission was identical with that from WT ‘Xanthi.’ Further, the measuring points from ammonium-grown plants were so similar to those shown for the nia-double mutant and for tungstate-grown plants that they were omitted. Data show one experiment, representative of four replicates. Note that the curve and the scale on the right y-axis for NO emission in nitrogen are interrupted to fit the figure.

Confirming previous observations on Arabidopsis (Magalhaes et al., 2000, 2002), NO emission was completely absent in tobacco leaves (Figure 1) having little or no NR activity. Neither leaves from ammonium-fed (nitrate-free) plants, nor from plants grown for 5 days on tungstate instead of molybdenum, nor from a nia double mutant emitted NO from nitrate (but see below for exceptions).

Factors limiting NO emission rates

We have previously suggested that the higher NO emission from leaves in the light is mainly a consequence of the light activation of NR and a resulting higher nitrite concentration (Rockel et al., 2002). Here, we show that this is only partly correct, using leaves of an NiR-free mutant (‘clone 271’), which contain very low NiR activity but normal NR activity (Vaucheret et al., 1992). These mutant leaves accumulate much more nitrite than WT leaves, and accordingly rates of NO emission are at least 1 order of magnitude higher than in WT leaves (Morot-Gaudry-Talarmain et al., 2002). Figure 2 shows NO emission, NR activity (+Mg2+) and activation state and nitrite concentrations in ‘clone 271’ during a dark-light-dark transient and in dark/anoxia. NO emission of the mutant increased drastically in the light with 100-fold higher rates than in WT. Importantly, this very high NO emission broke down almost immediately in a subsequent dark phase, although NR activity decreased only marginally and nitrite concentrations remained high for some time. This shows that other factors, aside from NR activity and nitrite concentrations, may strongly limit the rate of NO emission.

Figure 2.

NO emission, NR activity and nitrite content of leaves of the NiR-deficient tobacco mutant ‘clone 271’, after induction for 1 week on 10 mm nitrate (hydroponics).
The NO emission curve (○) is from a representative experiment. Samples (leaf disks) for determination of NR activity (columns) and nitrite content (bsl00001) were taken at the indicated times from separate leaves, kept under identical conditions as those used for the NO measurement (n = 5 ± SD). Maximum NR activity (+EDTA) is represented by light gray columns and the ‘actual NR activity’ (+Mg2+) by dark gray columns, which permits an estimation of the activation state. Light intensity was 250 μE m−2 sec−1 (PAR).

Nitric oxide emission from ‘clone 271’ in the dark increased under anoxia, as already described for WT leaves. However, under this condition, WT leaves produced as much NO as ‘clone 271’ (Figures 1 and 2), because under anoxia nitrite reduction was also blocked in WT leaves and nitrite accumulated to similar levels as in ‘clone 271’.

In order to distinguish between direct light effects on NO emission and effects of photosynthesis, in a second experiment leaves of ‘clone 271’ were kept in continuous light, but exposed to a transient from air to CO2-free air and back (Figure 3). NO emission dropped drastically when CO2 was removed, although not as rapidly as during an abrupt light/dark transient (Figure 2). NR is inactivated in spinach leaves illuminated in the absence of CO2 (Kaiser and Förster, 1989). However, in the above experiment with tobacco (Figure 3), NR activity changes and changes in the nitrite concentration were too small (NR) or too slow (nitrite) to be responsible for the large changes in the NO emission rate (see Discussion).

Figure 3.

NO emission, NR activity and leaf nitrite content of leaves of the NiR-deficient tobacco mutant ‘clone 271’ in continuous light during a transition from air or CO2-free air.
Other conditions are as in Figure 2.

Cell suspensions are very efficient tools for physiological studies, and have been used for measuring NO production in response to pathogens (Clarke et al., 2000; Delledonne et al., 1998; Gerber et al., 2004). Using non-green nitrate-grown tobacco cell suspensions, we found that normal NO emission in air was as low as in leaves or even lower on a FW basis (Figure 4). This low aerobic NO emission of suspension cells was completely insensitive to NOS inhibitors (l-NMMA, l-NAME and l-NIL, 2 mm each, data not shown). As in darkened leaves, NO emission increased strongly under nitrogen (anoxia). After return to aerobic conditions, the high NO emission dropped rapidly, although nitrite concentrations (medium plus cell) were still high, and NR activity decreased only marginally (Figure 4). Thus, nitrate-grown cell suspensions behaved very much like nitrite-accumulating mutant leaves. Obviously some unknown factor was strongly limiting NO emission in air in cell suspensions and darkened leaves. In order to check whether NADH could be this limiting factor, an estimation of the cytosolic NADH/NAD+ ratio was attempted by measuring the lactate/pyruvate ratio in leaves of WT and of ‘clone 271’, which is coupled with the cytosolic NADH/NAD+ via lactate dehydrogenase (LDH). For example, in illuminated WT leaves the cytosolic NADH/NAD+ ratio was 1.17 × 10−4, and this ratio was higher than for leaves kept in darkness (0.71 × 10−4). Unexpectedly, the ratio was slightly lower under dark-anaerobic conditions (Table 1).

Figure 4.

NO emission, NR activity and nitrite content of nitrate-grown tobacco cell suspensions during an air–nitrogen transition.
Cells (10 ml) were placed in a small petri dish in a cuvette mounted on a shaker under a stream of air or nitrogen (1.5 l min−1). At the time indicated, aliquots were taken for NR activity (columns) and nitrite determination (bsl00001). Maximum NR activity (+EDTA) is represented by light gray columns and ‘actual NR activity’ (+Mg2+) by dark gray columns. The NO emission curve is a representative example of four similar measurements. NR activity and nitrite contents are mean (n = 3) ±SD.

Table 1.  Pyruvate and l-lactate production and NADH/NAD+ ratio from tobacco WT and NiR -deficient mutant leaves under aerated and anoxic conditions
 Pyruvate (nmol g−1 FW)l-lactate (nmol g−1 FW)NADH/NAD+ ratio (×10−4)
  1. Leaves were harvested 3 h into the light phase. Detached leaves (petiole in water) were incubated in air for 15 min in dark or light, or for 20 min under nitrogen (in dark). Results are mean (n = 5) ±SD. For details, see Experimental procedures.

 Dark216.4 ± 29.7494.5 ± 45.80.71 ± 0.15
 Light217.3 ± 30.9859.6 ± 212.91.17 ± 0.26
 Nitrogen, dark402.3 ± 119.9839.2 ± 228.80.65 ± 0.18
NiR-deficient mutant
 Dark163.4 ± 11.6580.7 ± 75.61.01 ± 0.13
 Light142.0 ± 15.5712.6 ± 117.41.52 ± 0.37
 Nitrogen, dark304.1 ± 16.4825.2 ± 110.30.81 ± 0.03

Mitochondrial electron transport contributes to nitrite-dependent NO formation

Cell suspensions were also cultured in the absence of nitrate with ammonium (where NR expression is suppressed leading to NR-free cells), or they were grown on ammonium plus tungstate (not shown) to prevent activity not only of NR but also of other MoCo enzymes like xanthine oxidase/dehydrogenase (XDH) or aldehyde oxidase. These cell suspensions never produced NO when supplied with nitrate only, even under nitrogen. But when nitrite was added, even completely NR-free cells, and even tungstate-treated cells emitted NO. As an example, Figure 5 shows kinetics of NO emission from ammonium-grown cells in air (a) or nitrogen (b) and the response to subsequent additions of Myxothiazol, an inhibitor of the CytOX pathway of mitochondria, of SHAM (Salicylhydroxamic acid), an inhibitor of the AOX pathway, and finally of cyanide (KCN). NO production of NR-free ammonium cells was hardly (nitrogen) or slightly (air) inhibited by Myxothiazol. Stronger inhibition was achieved by SHAM and the inhibition was completed by 2 mm KCN, which unspecifically blocks all cytochrome moieties (Figure 5). Although such inhibitor studies in vivo may suffer from non-specific effects of the inhibitors, the data from Figure 5 suggest that mitochondrial electron transport could reduce nitrite to NO.

Figure 5.

NO emission from a tobacco cell suspension grown on ammonium in the absence of nitrate.
(a) NO emission under aerobic conditions.
(b) NO emission under anoxic conditions. No NR activity was detected in those cell extracts (not shown) and according to this, these cells do not emit any NO in nitrogen, in contrast to normal nitrate grown cells, because NR as a source for nitrite is lacking. When nitrite (300 μm) is added, aerobic or anoxic NO emission reaches normal rates, just as in NR-expressing cells. Myxothiazol (10 μm), SHAM (2.5 mm) and KCN (2 mm) were added to the cell suspension as indicated. Data show one typical experiment, representative of three replicates. Similar experiments were carried out with cells grown on ammonium plus tungstate, with practically identical curves.

Subsequently, we analyzed the capacity of isolated mitochondria to produce NO. As expected, following the addition of NADH and nitrite (Figure 6a), mitochondria purified from NR-free (ammonium-grown) suspension cells were able to produce NO under nitrogen with rates in the range of 0.49 nmol mg−1 protein h−1. This NO emission was inhibited by Myxothiazol and to a smaller part by SHAM. Oxygen uptake was about 4.8 μmol mg−1 protein h−1 (not shown). Thus, the capacity of mitochondria to reduce nitrite to NO under anoxia was about 0.1‰ of the respiration rate, whereas in air, mitochondrial NO emission was practically absent. Mitochondria purified from tobacco leaves produced only very little NO (0.053 nmol mg−1 protein h−1), although they had a similar respiration rate as mitochondria from suspension cells (Figure 6b). Mitochondria purified from tobacco roots gave even higher rates of NO emission under anoxia (up to 10 nmol mg−1 protein h−1, data not shown).

Figure 6.

NO emission from mitochondria purified from ammonium grown tobacco suspension cells (a) and tobacco wild-type leaves (b).
A suspension of purified mitochondria containing 1.5–2 mg protein in a total liquid volume of 10 ml was placed in a flat glass beaker mounted in the headspace cuvette at 22°C on a rotary shaker. The cuvette was flushed with purified air or nitrogen (1.5 l min−1). Nitrite (final concentration 500 μm), NADH (1 mm), ADP (0.5 mm), Myxothiazol (25 μm) and SHAM (2.5 mm) were injected with microliter syringes through a serum stopper, where indicated, without interrupting the measurement. Data show one experiment, representative of four replicates. The mean respiration rate in air was 4.8 (±1.2, n = 4) μmol mg−1 protein h−1.

In order to estimate the contribution of mitochondrial electron transport and of NR to total NO emission, we compared the fraction of NO emission that could be blocked by Myxothiazol plus SHAM (=NO produced by mitochondria), in NR containing (nitrate-grown) cells and NR-deficient (ammonium-grown) cells (Table 2). Generally, in ammonium cells the percentage of NO emission that could be blocked by Myxothiazol plus SHAM was larger than in nitrate-grown cells, which should be expected if NR would contribute to NO emission. However, even in ammonium cells which did not contain measurable NR activity, the inhibition of NO emission by Myxothiazol + SHAM was not complete, although respiratory oxygen uptake (not shown) was completely blocked.

Table 2.  Effect of inhibitors of mitochondrial electron transport on oxygen uptake and on NO emission of tobacco suspension cells
 NO emission in airNO emission in nitrogenO2 uptake (%)
NOinline image cellsNHinline image cellsNOinline image cellsNHinline image cellsNOinline image cells
  1. NO emission was followed with cell suspensions grown on nitrate (NOinline image cells = NR-containing) or on ammonium (NHinline image cells = NR-deficient). The NO emission rate (nmol g−1 FW h−1±SD, n = 3–5) was determined after that nitrite (300 μm) was added into the cells. After NO emission reached a steady-state, Myxothiazol (10 μm) or SHAM (2.5 mm) and finally KCN (2 mm) were injected into the cell suspension. Oxygen uptake is expressed as percentage of the control.

Control + nitrite0.57 ± 0.090.83 ± 0.3124.05 ± 4.6835.76 ± 13.01100.0
+Myxothiazol0.53 ± 0.040.75 ± 0.3128.83 ± 9.4028.62 ± 10.2976.0 ± 6.2
+SHAM0.37 ± 0.140.36 ± 0.1028.37 ± 5.4722.66 ± 7.4624.0 ± 6.2
+Myxothiazol + SHAM0.32 ± 0.070.28 ± 0.1030.73 ± 9.2315.52 ± 6.710.0
+Myxothiazol + SHAM + KCN0.05 ± 0.020.07 ± 0.031.35 ± 0.720.81 ± 0.520.0

An attempt to estimate NO scavenging

As already stated in the Introduction, NO has a rather short biological half-life because it may react with many cellular constituents. Chemiluminescence, like other methods (e.g., electron paramagnetic resonance or hemoglobin or mass spectrometry) detects only NO that has been released from the plant or plant cells. Thus, an estimation of NO scavenging by leaves or cells is important. We first tried to estimate NO scavenging by titrating cell suspensions with NO dissolved at known concentrations in buffer (Figure 7). The data suggest that NO quenching could be as high as 95% (cells in air). Surprisingly, however, quenching in air was only twice as high as under nitrogen, and a large part of the quenching appeared due to components of the media used for our suspension cells (Figure 7). Due to that complex response of NO emission to various components of the cell suspension, and because NO quenching could not be measured in the same way with leaves, we used yet another strategy to estimate NO quenching in leaves, with NR as a natural NO source inside cells. We first determined nitrate reduction of purified NR at substrate saturation (nitrate plus NADH), which was 11.4 nmol min−1 (Figure 8). Subsequently, NO emission from a solution of purified NR was measured with nitrite and NADH. NO emission from the aerated enzyme solution started immediately after NADH addition, and increased continuously to reach a steady-state (Figure 8). After switching to nitrogen, NO emission (but not nitrate reduction) was further increased by 10%. Thus, NO emission of purified NR in air was 0.84% of the corresponding NR activity, or 0.92% in nitrogen. Next, we compared the ratio of NR activity (+Mg2+) in leaf extracts with NO emission from photosynthesizing leaves or from darkened anoxic leaves of ‘clone 271’, which accumulate nitrite. NR in those leaves and under those conditions should be nitrite (and probably NADH) saturated, like NR in vitro. Assuming that NO emission as percentage of NR activity would be the same in vivo and in vitro, we could use these ratios to estimate scavenging of NO in vivo. NO emission by NR in leaves varied from 1 to 2% of total NR activity (compare Figure 2), which is within the range observed for purified NR in vitro.

Figure 7.

Scavenging of NO by a buffer solution (50 mm MES-KOH, pH 5.5) or by a cell suspension or by a cell-free medium.
Ten milliliters of the buffer solution was pre-flushed with 100 ppm NO in nitrogen (N2) for at least 10 min. Two milliliters of the solution (containing 380 pmol dissolved NO) was injected into 1 ml of buffer solution (bsl00001), or into 1 ml LS medium (○) or into 1 ml of a cell suspension (Nicotiana tabacum cv. Xanthi) flushed with nitrogen (bsl00066), or flushed with air (▿) in a small petri dish placed in a glass cuvette (total gas volume 1 l), mounted on a rotary shaker. The NO which was subsequently emitted from the liquid phase was continuously measured. Numbers (%) at the curves give the percentage of the injected NO that was recovered.

Figure 8.

NO emission from a solution of purified NR (maize) in air or nitrogen.
Two milliliters of a buffer solution was incubated in a 5-ml glass beaker mounted in the headspace cuvette at 22°C. Purified NR, nitrite (1 mm) and nitrate (100 μm) were added at T = 0. NADH (500 μm) was added where indicated. The cuvette was flushed with (1.5 l min−1) purified air or nitrogen. Nitrate reduction activity of added NR was 11.4 nmol min−1, both in air and in nitrogen. Thus, NO production was 0.84% (air) or 0.92% (nitrogen) of total NR activity.


Rates and determinants of NO emission

With a solution of purified maize NR, substrate-saturated NO production was between 0.84 and 0.92% of NR activity. A similar percentage could be expected in vivo only, if nitrite reduction would be absent, nitrate reduction would be substrate-saturated and NO quenching would be negligible. When meeting the first two conditions by using leaves of ‘clone 271’, we found that NO emission in vivo was around 1–2% of the extractable NR activity, similar to values obtained with purified NR in vitro. Conclusively, NO scavenging in vivo should be low, even in air.

As NO scavenging in leaves appeared negligible, tissue NO concentrations could be roughly estimated from NO emission data, assuming that NO in the gas stream around leaves would be in equilibrium with NO dissolved in the aqueous phase. With illuminated WT leaves in air, NO emission was typically 0.25 nmol g−1 FW h−1, and the NO concentration in the gas phase was 0.2 nl l−1 (not shown). According to the water solubility of NO (1.9 mm at atmospheric pressure), that should correspond to a minimum NO concentration in the water phase of 0.4 pm. Taking into account diffusional resistances, and some NO quenching during its way out, the real NO concentration may have been 10 times higher than calculated – which could sum up to 4 pm. Remarkably, these concentrations are more than 5 orders of magnitude below literature values for NO concentrations assumed to provoke a hypersensitive response in plant cells (Delledonne et al., 1998).

Tissue nitrite concentration was certainly a most important factor determining the rate of NO emission. However, it was very obviously not the only one. NO emission from leaves (clone 271) or cell suspension dropped rapidly upon light/dark or nitrogen/air although nitrite concentrations remained still high, at least initially (Figures 2 and 4), and NO emission was also affected by the presence or absence of photosynthesis (Figure 3). We were originally assuming cytosolic NADH to be another determinant of the NO emission rate, but our pyruvate/lactate determination as a measure for cytosolic NADH did not confirm this. One other possibility would be that the change in NR activity and NR activation state that we measure in crude extracts does not really reflect the in situ situation, at least in tobacco. Mutating the regulatory phosphorylation site on tobacco NR resulted in high nitrite formation and NO emission in vivo (Lea et al., 2004).

Reactions participating in nitrite-dependent NO production

We have shown that NR-free plants or cell suspensions supplied for short periods with nitrate never emitted NO. However, when supplied with nitrite, virtually NR-free ammonium-grown cell suspensions emitted NO under anoxia at almost the same rates as NR-induced cells (Figure 5). Thus, NR is obligatory for NO emission, because it is the only nitrite source.

Nitric oxide emission was also observed with tobacco cell suspensions grown on ammonium plus tungstate (data not shown), which express neither a functional NR nor other MoCo enzymes. Therefore, the MoCo enzyme XDH can probably be excluded as NO source, contrary to suggestions (Godber et al., 2000; Li et al., 2001; Millar et al., 1998). Moreover, in experiments with recombinant purified XDH, we found no NO production from nitrite plus NADH (C. Hesberg, R. Hänsch, E. Planchet, R. Mendel, Dept. Plant Biology, TU Braunschweig, Braunschweig, Germany, unpublished data). Further, as high NO emission was found with leaves of the NiR-deficient mutant, it appears very improbable that NiR itself is a source for NO.

Initially based on inhibitor studies with suspension cells, we assumed mitochondria to contribute to NO production from nitrite, at least in those cases were NR was absent. Confirming this, we show that higher plant mitochondria, like mammalian mitochondria (Kozlov et al., 1999) and algal mitochondria (Tischner et al., 2004), reduce nitrite to NO under anoxia. Interestingly, even the combined action of Myxothiazol plus SHAM never caused a complete inhibition of NO production, which was, however, abolished by KCN (which also blocks NR and other heme enzymes). Thus, while there is no doubt that all NO was produced enzymatically, some as yet undefined, cyanide-sensitive reaction appeared to contribute to some extent to NO formation. The PM-bound nitrite:NO oxidoreductase detected by Stöhr et al. (2001) may be one possible candidate.

Why is NO emission so low in air, and how much do mitochondria contribute to NO emission?

One of the most puzzling observations was the very strong depression of NO emission by oxygen (air) observed with almost all systems used (except with photosynthesizing leaves of clone 271, Figures 2 and 3). We originally assumed that cytosolic NADH could become rate limiting in air, at least under those conditions were nitrite was high (as in leaves of ‘clone 271’, or in cell suspensions fed nitrite). However, our pyruvate/lactate determinations suggested only slightly higher NADH/NAD+ in the light than in the dark. In addition, purified mitochondria from cell suspensions gave almost no measurable NO emission in air, even when supported with nitrite plus NADH. Accordingly, the low aerobic NO production could not be traced back to substrate limitation, and presently we have no satisfying explanation for that observation. In preliminary experiments not shown here, we found that 0.05% oxygen was sufficient for a 50% inhibition of NO emission from purified root mitochondria. In contrast to mitochondria, NO emission by purified NR was rather insensitive to air (compare Figure 7). Non-symbiotic hemoglobin might catalyze NADH-dependent oxidation of NO back to nitrite and nitrate (Igamberdiev and Hill, 2004; Perazzolli et al., 2004). Such a reaction would lead to an underestimation of NO production in air, but not in nitrogen. To what extent this reaction contributes to NO scavenging in tobacco leaves or cell suspensions is not known yet.

Prevailing mitochondrial NO production in nitrite-fed suspension cells was also suggested by inhibitor studies, where Myxothiazol + SHAM caused an almost complete inhibition of anoxic NO emission from NR-free cells, and a 57% inhibition with nitrate-grown cells (Figure 5, Table 2). In contrast, the very low NO emission from purified leaf mitochondria (Figure 6b) suggests that in leaves, NR should be the major NO source. This was confirmed by preliminary experiments (not shown) with leaf slices from WT and nia leaves in solution. When fed nitrite in the dark under anoxia, NR-free nia leaf slices emitted very little NO (below 0.3 nmol g−1 FW h−1), whereas NR-containing WT leaf slices emitted more than 100-fold more NO (up to 50 nmol g−1 FW h−1). This is in marked contrast to the response of suspension cells, where NO emission from NR-containing cells was only slightly higher than that from NR-free cells (Table 2). However, NR is always required as a source for nitrite, except perhaps in cases where microbial nitrite may become available for plants, as may be expected in soils.

Under anoxia, nitrite accumulated in all tissues (or was released to the medium) that contained NR and nitrate. This anoxic nitrite accumulation has two reasons: one is the well-known activation of NR, probably triggered by cellular acidification (Kaiser and Brendle-Behnisch, 1995), the second one is a decreased rate of plastidic nitrite reduction (Botrel et al., 1996). Using NR-free nia suspension cells fed nitrite, we found that nitrite reduction under anoxia was about 25% of that in air (not shown). Non-green plastids or darkened chloroplasts produce NAD(P)H by means of the oxidative pentose phosphate (OPP)-cycle (Kaiser and Bassham, 1979). Under anoxia, ATP and sugar phosphate levels are low (Stoimenova et al., 2003), eventually too low to fuel the OPP cycle. It is tempting to speculate whether reduction of nitrite to NO under anoxia, where nitrite accumulates, may represent a ‘nitrite respiration.’ However, the above measured rates of NO production were very low, probably far too low to be relevant for anoxic energy production.

Experimental procedures

Plant material and culture conditions

Experiments were performed with Nicotiana tabacum cv. Xanthi and cv. Gatersleben. Seeds from a nitrite-reductase antisense-transformant were obtained from INRA Versailles (France) and seeds of an NR-deficient nia30 mutant were obtained from R. Mendel (Braunschweig, Germany). Seeds were germinated on vermiculite in a day/night regime of 14/10 h, 24/20°C, a relative humidity of 80% and 350–400 μmol m−2 sec−1 photon flux density (PAR). After 3 weeks, the plants were transferred to hydroponic culture for an additional 4–8 weeks. Plastic pots, each containing 1.8 l nutrient solution, were kept in a growth chamber with artificial illumination (100W Hqi; Schreder, Winterbach, Germany) at a PAR of 300 μmol m−2 sec−1, and with 16 h daily light periods. The day/night temperature regime of the chamber was 24/20°C, respectively. The nitrate nutrient solution (pH 6.3) contained 10 mm KNO3, 1 mm CaCl2, 1 mm MgSO4, 25 μm NaFe-EDTA, 0.5 mm K2HPO4, 1 mm KH2PO4 and trace elements from a stock solution containing 92.5 mm H3BO3, 18.3 mm MnCl2, 1.5 mm ZnSO4, 0.64 mm CuSO4 and 0.24 mm Na2MoO4. For ammonia plants, the nutrient solution was 3 mm NH4Cl, 1 mm CaCl2, 2 mm MgSO4, 25 μm NaFe-EDTA, 0.5 mm K2HPO4, 1 mm KH2PO4 and trace elements. The composition of the nutrient solution for NR-deficient nia30 mutant was 1 mm KNO3, 3 mm NH4Cl, 1 mm CaCl2, 2 mm MgSO4, 25 μm Fe-EDTA, 2 mm KH2PO4/K2HPO4 and trace elements. NiR-deficient transformants (‘clone 271’) were pre-grown on ammonium nutrient solution, and 1 week before harvest for experiments, plants were transferred into the above nitrate nutrient solution. For all these conditions, nutrient solutions were changed three times a week. For tungstate-treated plants, the nutrient solution contained 500 μm sodium tungstate instead of molybdate. The solution was changed every day during 5 days of treatment before experiments.

Tobacco (N. tabacum cv. Xanthi) cells were grown in 300 ml conical flasks containing 100 ml of LS medium pH 5.8 (Linsmaier and Skoog, 1965) at a constant temperature of 24°C and a continuous illumination, on a rotary shaker (100 rpm). Ammonium cell suspension cultures totally devoid of nitrate were grown on LS medium with small modifications: 1.5 mm NH4Cl instead of NH4NO3 and KNO3, and addition of 2.35 mm MES in order to maintain the pH around 5.5. Cells were maintained in the exponential phase and subcultured 3 days before measurement. Prior to utilization, cells were washed with their LS medium, respectively, and resuspended at 0.1 g FW ml−1 and equilibrated for 30 min–1 h on a rotary shaker (100 rpm).

In vitro assay of NR activity

Following different treatments, disks of tobacco leaves were harvested, weighed and immediately quenched in liquid nitrogen. Aliquots of cell suspensions were vacuum-filtrated on membrane filters, and also frozen immediately in liquid nitrogen. The material was ground with liquid nitrogen and 2 ml of extraction buffer (100 mm HEPES pH 7.6, 3.5 mm mercaptoethanol, 10 μm FAD, 10 μm molybdate, 15 mm MgCl2, 2 mm pefabloc, 10 μm leupeptin, 50 μm cantharidine, 0.5% PVP, 0.5% BSA and 0.3% of Triton X100) was added to 1 g FW. Cantharidine (a PP2A inhibitor) was added in order to prevent dephosphorylation of NR. After continuous grinding until thawing the suspension was centrifuged (14 500 g, 10 min, 4°C). After this centrifugation, aliquots of the extract were directly used for the colorimetric determination of nitrite content (excepted for ‘271’ samples where the nitrite content was measured from desalted extracts). The remaining supernatant was desalted on Sephadex G 25 spin columns (1.5 ml gel volume, 650 μl extract, 4°C) equilibrated with the extraction buffer without the protease inhibitors. With aliquots of the supernatant the following assays were carried out:

  • (i) Determination of NR act (+Mg2+): 200 μl leaf extract was added to 800 μl reaction mixture (100 mm HEPES-KOH pH 7.6, 1 mm DTT, 10 μm FAD, 10 μm molybdate, 15 mm MgCl2, 5 mm KNO3 and 0.2 mm NADH). After 5 min (24°C), the reaction was stopped by adding 125 μl zinc acetate (0.5 m).
  • (ii) Determination of NRmax (+EDTA): 200 μl leaf extract as above, but containing in addition 20 mm EDTA and 5 mm AMP (final concentrations) were pre-incubated at 24°C. After 13 min, buffer (100 mm HEPES-KOH pH 7.6, 1 mm DTT, 10 μm FAD, 10 μm molybdate and 15 mm EDTA) was added to a final volume of 1 ml. After 2 min (total pre-incubation time 15 min), the reaction was started by adding 5 mm KNO3 and 0.2 mm NADH (final concentrations). Five minutes later, the reaction was stopped by adding 125 μl zinc acetate (0.5 m). Following centrifugation (16 000 g, 5 min), the supernatant was treated with 10 μm phenazine methosulfate to oxidize unreacted NADH, and the nitrite content was determined colorimetrically (Hageman and Reed, 1980).

Isolation and purification of mitochondria

Mitochondria were isolated from tobacco suspension cells and tobacco leaves by adapting, with slight modifications, the previously described methods of Nishimura et al. (1982) and Vanlerberghe et al. (1995). All procedures were carried out at 4°C. A total of 1.0–1.2 l of cell culture (3 days sub-cultured) was filtered through gauze to remove the media and was then ground by mortar and pestle in 250 ml of homogenization medium containing 300 mm sucrose, 100 mm HEPES buffer (pH 7.6), 0.1% (w/v) milk powder, 0.6% (w/v) PVP, 1 mm EDTA, 2 mm MgCl2, 4 mm cysteine, 5 mm KH2PO4 and a ready-to-use protease cocktail (‘Complete EDTA-free’; Roche, Mannheim, Germany: one tablet in 100 ml medium). A second step of homogenization was carried out with an Ultra Turrax (Janke and Kunkel, Staufen, Germany). The final homogenate was filtered through Miracloth. The filtered cell extract was separated by centrifugation for 1000 g for 15 min, and the supernatant was centrifuged again at 10 000 g for 20 min. The final pellet was resuspended in 4 ml of suspension medium containing 20 mm HEPES (pH 7.6), 300 mm sucrose, 0.1% (w/v) milk powder, 2 mm MgCl2 and 1 mm EDTA. The mitochondria were purified on discontinuous Percoll step gradient (bottom to top): 3 ml of 60% (v/v), 4 ml of 45% (v/v), 4 ml of 28% (v/v), and 4 ml of 5% (v/v), all containing 250 mm sucrose, 20 mm HEPES (pH 7.6) and 0.1% defatted milk powder. The mitochondria (4 ml) were layered on top, and the gradient was centrifuged at 30 000 g for 30 min. The mitochondrial fraction appeared at interface between 45 and 28% (v/v) Percoll layer, and was collected, diluted in suspension medium, centrifuged and resuspended twice (7 000 g, 10 min) in order to remove Percoll. The final pellet was suspended in 8 ml of suspension medium. Mitochondrial preparations were routinely checked for the respiratory activity (state 4), and for contamination with NR and catalase (peroxisomal). Contamination with catalase (measured as oxygen evolution from a 20 mm H2O2 solution in suspension buffer) was usually negligible. Mitochondria from nitrate-grown WT plants contained an NR activity (+EDTA) of 27.4 nmol mg−1 protein h−1 (±3.4, n = 3), which represents 0.6% of the leaf NR activity (based on a leaf protein content of 25 mg g−1 FW and assuming that leaf mitochondria contained 5% of the total leaf protein). Mitochondria from NR-free cells or tissues expectedly contained no detectable NR activity.

Gas phase NO measurements

For experiments with detached leaves, the leaves were cut off from the plant and immediately placed in nutrient solution, where the petiole was cut off a second time below the solution surface. The leaves (petiole in nutrient solution) were placed in a transparent lid chamber with 2 or 4 l air volume, depending on leaf size and number. A constant flow of measuring gas (purified air or nitrogen) of 1.5 l min−1 was pulled through the chamber and subsequently through the chemiluminescence detector (CLD 770 AL ppt; Eco-Physics, Dürnten, Switzerland; detection limit 20 ppt; 20 sec time resolution) by a vacuum pump connected to an ozone destroyer. The ozone generator of the chemiluminescence detector was supplied with dry oxygen (99%). The measuring gas (air or nitrogen) was made NO free by conducting it through a custom-made charcoal column (1 m long, 3 cm internal diameter, particle size 2 mm). Calibration was routinely carried out with NO free air (0 ppt NO) and with various concentrations of NO (1–35 ppb) adjusted by mixing the calibration gas (500 ppb NO in nitrogen; Messer Griesheim, Darmstadt, Germany) with NO-free air. Flow controllers (FC-260; Tylan General, Eching, Germany) were used to adjust all gas flows. Light was provided by a 400-W Hqi-lamp (Schreder) above the cuvette. Quantum flux density was adjusted within limits (150–400 μmol m2 sec−1 PAR) by changing the distance between lamp and cuvette. Air temperature in the cuvette was continuously monitored, and was usually about 20°C in the dark and 23–25°C in the light.

For the measurement of NO production from cell suspensions (10 ml), purified NR (2 ml), or purified mitochondria (10 ml), the solutions were placed in small glass beakers of suitable size, located in a transparent lid chamber (1 l gas volume) mounted on a shaker. NO was measured by chemiluminescence detection as described above.

Oxygen uptake

Oxygen uptake was measured in solution (2 ml of the suspension medium, containing in addition 0.5 mm KH2PO4, 0.2 mm ADP and mitochondria equivalent to 0.1–0.2 mg protein) with a Clark-type oxygen electrode system (Hansatech, Instruments Ltd, Bachofer, Reutlingen, Germany). The electrode was calibrated with Na2S2O4 (zero oxygen) and with air-saturated water (250 μm oxygen). All measurements were performed at 25°C.

Protein determination

The protein content of isolated mitochondria was determined after suitable dilution with BCA protein assay reagent (bicinchonimic acid; Pierce, Rockford, IL, USA). Bovine serum albumin (BSA) was used as a standard.

Determination of pyruvate, l-lactate and NADH/NAD+ ratio

For determination of NADH/NAD+ ratio, pyruvate and lactate contents were measured spectrophometrically (Sigma ZFP 22; Sigma Instrumente, Berlin, Germany). Frozen samples (1 g FW) were extracted with 2 ml of 7.5% perchloric acid. The extract was centrifuged (12 000 g, 10 min) and the supernatant was transferred to a new tube. For 1 ml supernatant, 20 μl of Tris (2 m) was added and the pH of this supernatant was neutralized to pH 7.0 with 5 m K2CO3. The extract was then recentrifuged and the supernatant collected and stored at −80°C until it was assayed. The enzymatic determination of l-lactate content was assayed using the glutamate pyruvate transaminase (GPT)–lactate dehydrogenase (LDH) system. Fifty micro liters of extract was added to 1 ml buffer solution (100 mm CHES, pH 10) containing already glutamate (10 μm), GPT (5 U), NAD+ (2 mm). The reaction was started by adding LDH (3 U). In the case of pyruvate determination, 50 μl extract was added to 1 ml buffer solution (100 mm Hepes, pH 7.6) containing NADH (0.5 mm) and LDH (0.15 U). After each reaction, an appropriated internal standard was added to the mixture for the determination of these two different metabolites.

Preparation of NO solution for titration experiments

Nitric oxide gas (100 ppm in nitrogen; Messer Griesheim) was flushed for 15 min at 24°C through a glass vial containing buffer (MES-KOH 50 mm, pH 5.5). According to the solubility of pure NO in water at atmospheric pressure (1.9 mm), the NO equilibrium concentration of the solution flushed with 100 ppm NO should be 190 pmol ml−1. Aliquots of the solution were sampled with a plastic syringe (1 ml) fitted with a stainless steel needle, and immediately injected into the NO cuvette through a gas-tight serum stopper, without interrupting the NO measurement.


This work was supported by the DFG (SFB 567, Ka 456 13-1 and 15-1). We gratefully acknowledge the technical assistance of Maria Lesch. M. Sonoda was a recipient of fellowships by AvH and JSPS. Seeds for the tobacco NiR-antisense transformants were a kind gift from J.F. Morot-Gaudry, INRA Versailles (France). We also thank Christiane Loeffler and Barbara Dierich (Pharmaceutical Biology, University of Wuerzburg) for kindly supplying us with tobacco cell suspensions.