Endogenous Superoxide Production and the Nitrite/Nitrate Ratio Control the Concentration of Bioavailable Free Nitric Oxide in Leaves*

We have quantitatively measured nitric oxide production in the leaves of Arabidopsis thaliana and Vicia faba by adapting ferrous dithiocarbamate spin tapping methods previously used in animal systems. Hydrophobic diethyldithiocarbamate complexes were used to measure NO interacting with membranes, and hydrophilic N -methyl- D -glucamine dithiocarbamate was used to measure NO released into the external solution. Both complexes were able to trap levels of NO, readily detectable by EPR spectroscopy. Basal rates of NO production (in the order of 1 nmol g (cid:1) 1 h (cid:1) 1 ) agreed with previous studies. However, use of methodologies that corrected for the removal of free NO by endogenously produced superoxide resulted in a significant increase in trapped NO (up to 18 nmol g (cid:1) 1 h (cid:1) 1 ). Basal NO production in leaves is therefore much higher than previously thought, but this is masked by significant superoxide production. The effects of nitrite (increased rate) and nitrate (decreased rate) are consistent with a role for nitrate reductase as the source of this basal NO production. However, rates under physiologically achievable nitrite concentrations never approach that reported following pathogen induction of 1 (cid:2) 1 were different for the two types of experiment. This is in the case of the water-soluble MGD trap, the of NO produced by the leaf was dissolved in the incubation (1-ml Eppendorf vial), of which only an aliquot was taken as an EPR sample, whereas in the case of the lipid soluble DETC experiments it was assumed that all NO produced by the leaf sample is trapped inside the leaf tissue, all of which was then trans-ferred to the EPR tube.

The phenomenon of NO production by plants was first reported by Klepper in 1979 (1), significantly earlier than the discovery of NO generation by animals (1985)(1986)(1987). Nevertheless research in the plant NO field has lagged behind the animal field for two important reasons: the absence of a physiological phenomenon elicited by NO and a lack of understand-ing of the molecular mechanism underpinning any such phenomenon. In animals the role of NO as the endotheliumderived relaxing factor and the discovery of the L-arginine/ nitric-oxide synthase/guanylate cyclase signaling pathway (2,3) provided the impetus for the explosion of research in this area in the 1990s (4). However, recently there has been a similar expansion of interest in the role of NO in plants.
Although a complete molecular model of the physiological role of NO in plants remains to be determined, a number of normal physiological processes are now known to be modulated by NO production. These include: production of the hormone ethylene (5), germination (6), programmed cell death (7,8), senescence (5), and stomatal closure (9). NO is also involved in pathogen defense, either directly (10) or via production of antimicrobial phytoalexins (11). There are several recent reviews covering the role of NO in plants (12)(13)(14)(15)(16).
Although the molecular target eliciting the NO response has proved elusive, there has been considerable progress in identifying the enzymes that produce NO in vivo. A consensus has emerged that basal levels of NO production are controlled by the heme enzyme, nitrate reductase (NR). 1 (9,(17)(18)(19)(20)(21). This enzyme has a primary activity of converting nitrate to nitrite; however, it can also convert nitrite to NO, presumably via a single electron reduction similar to that occurring in the well characterized bacterial heme proteins (e.g. cd1) that catalyze nitrite reduction to NO as part of the nitrogen cycle (22). The hypothesis that NR plays a significant role in NO generation in plants is supported by the recent finding that NO synthesis in plants is stimulated by nitrite, and this stimulation is attenuated in an Arabidopsis thaliana NR double mutant (at nia1 and nia2) that has a diminished NR activity (9).
Recently, however, two other enzyme systems have been characterized as being involved in plant NO synthesis. The first is induced in response to pathogens (23). This enzyme, a variant of the P protein of the glycine decarboxylase complex, catalyzes similar chemistry to the mammalian nitric-oxide synthase (converting L-arginine to citrulline and NO) but has very little sequence homology with it. Thus, it has been difficult to detect via genome mining or antibody screening. The second new enzyme, AtNOS1, is involved in hormone signaling by NO and also converts arginine to citrulline (24). It is also not homologous to mammalian NO synthases, although it shares many of their functional characteristics (including calcium activation and inhibitor sensitivity).
Nitrogen oxides have a range of chemical reactivity, making difficult the quantitative analysis of their production, destruction, and physiological/pathophysiological effects in complex biological solutions. Yet without these hard numbers it becomes difficult to determine which of the wide range of NO chemical/biological activities are important physiologically. However, one advantage of the relatively late arrival of plant research into the nitric oxide field is that a number of methods for measuring NO have been optimized on mammalian systems (25,26).
Superoxide reacts with NO at a diffusion-controlled rate to produce peroxynitrite (27). One of the early indicators that free NO was involved in mammalian physiology was the finding that superoxide completely attenuated both endogenous (bradykinin) and exogenous nitric oxide-induced vessel relaxation (28,29). Because plants have the potential for very high rates of superoxide production (30) both basally and, especially, in response to light or water stress (31)(32)(33), it is likely that this is the main obstacle in accurately measuring NO production rates. Our recent studies in animals have demonstrated that the addition of exogenous NO complexes of iron-dithiocarbamate complexes can compete with superoxide anions for NO molecules, ensuring a more quantitative formation (34) of the EPR-detectable mononitrosyl iron-dithiocarbamate complexes. This allows for a far more accurate measure of NO production in the system (35).
NO spin traps have not been extensively used in plant systems. The aim of the present study was to verify the efficiency of NO spin trapping as a method of quantitatively measuring NO production rates in plant leaves. The hydrophobic or hydrophilic ferrous-dithiocarbamate complexes, ferrous-DETC and ferrous-N-methyl-D-glucamine dithiocarbamate, were used respectively to monitor NO generation in cell membranes or to trap NO escaping from the cells to the medium. In agreement with suggestions from other studies, we show that basal levels of NO production are catalyzed by the nitrite reductase activity of NR. However, we also show that endogenous superoxide production has a significant effect on the levels of bioavailable free NO in the plant.

EXPERIMENTAL PROCEDURES
Materials-Sodium diethyldithiocarbamate (DETC), sodium ascorbate, sodium citrate, sodium nitrite, sodium nitrate, L-cysteine, and Hb from Sigma; FeSO 4 (NH 4 ) from Fluka (Buchs, Switzerland); and L-nitroarginine methyl ester acetate and NO-proline (NO-Pr) from Alexis (San Diego, CA) were used in the experiments. Sodium N-methyl-Dglucamine dithiocarbamate (MGD) was a generously gift from Dr. Yu. V. Khropov (Chemical Department of Moscow State University). Gaseous NO was obtained by the reaction of FeSO 4 with NaNO 2 in 0.1 M HCl and then purified by the method of low temperature sublimation. The paramagnetic dinitrosyl iron complexes with cysteine were obtained by treating the FeSO 4 and L-cysteine solutions with gaseous NO at molar ratio 1:30 and at an NO pressure of 100 -200 mm Hg. The synthesis was performed in a Thunberg tube apparatus containing 100 ml of gasphase volume. A 0.5-ml aliquot of FeSO 4 solution in distilled water (pH 5.5) and 4.5 ml of thiol in 10 mM HEPES buffer (pH 7.4) were placed in the upper and bottom part of the apparatus, respectively. The Thunberg tube was evacuated, and NO was added. Both solutions were then mixed in the presence of NO and shaken for 5 min followed by evacuation of NO from the tube. A 0.5 mM concentration standard for the EPR measurements was made by the addition of an excess of NO-Pr (final concentration, 10 mM) to the solution of 0.5 mM Fe 2ϩ and 5 mM MGD in 10 mM HEPES, pH 7.4 (resulting in 0.5 mM MNIC-MGD standards).
Experiments on Plant Leaves-A. thaliana (variety Columbia) was grown at low light in John Innes seed and potting compost at a light intensity of 300 mol photons m Ϫ2 s Ϫ1 at 80% relative humidity and 25°C with a 16-h light period in a controlled environment growth cabinet. The plants were used at 28 days post-germination. Vicia faba (broad bean) was greenhouse-grown in seed and potting compost until the third true leaves were mature. Chinese rose (Hibiscus rosa sinensis) was cultivated in a domestic flower border.
Rectangular cuts of whole bean leaves (4 ϫ 20 mm; total weight of 40 mg) or whole A. thaliana leaves (10 ϫ 10 mm; total weight of 40 mg) were put into 1.5-ml Eppendorf vials with 1 ml of 5 mM DETC (or MGD) solution with 5 mM sodium ascorbate. 100 mM potassium HEPES buffer was used (pH 7.4). A 10-l aliquot of the 50 mM stock solution of FeSO 4 (NH 4 ) in distilled water was then added to the Eppendorf vial to give a final Fe 2ϩ concentration of 0.5 mM. The leaves (cut or whole) were kept in the solution for 1-3 h at ambient temperature before either leaf tissue (the DETC series) or the solution (the MGD series) were frozen in liquid nitrogen for EPR assays. The addition of exogenous NO donor, NO-Pr resulted in a 6 -8-fold increase of MNIC concentration in the preparations (see "Results"), indicating that the concentrations of Fe 2ϩdithiocarbamate complexes used were sufficient to ensure scavenging of the majority of NO molecules.
The use of cut leaves in these experiments ensured that the NO trap only had access to the leaf via the cut edge. The buffer would also enter the leaf via the same route; this resulted in an increase of the buffer capacity in the intercellular space and therefore decreased the likelihood of NO being formed nonenzymatically from any nitrite inside the "acidic" cell compartment.
EPR Assays-Most of the EPR spectra were measured at 10 K on a Bruker EMX EPR spectrometer equipped with an Oxford Instruments liquid helium system. A spherical high quality Bruker resonator SP9703 was used. The instrumental parameters were: modulation frequency, m ϭ 100 kHz; microwave frequency, ϭ 9.47 GHz; microwave power, p ϭ 0.05 mW; modulation amplitude, A m ϭ 3 G; spectrum sweep rate, V ϭ 3.57 G/s; time constant, ϭ 82 ms; and number of scans/ spectrum, NS ϭ 1. The EPR spectra of the Chinese rose leaves were recorded at liquid nitrogen temperature using an X-band Bruker ESC-106 EPR spectrometer ( m ϭ 100 kHz, ϭ 9.4 GHz, p ϭ 0.05 mW, A m ϭ 3 G). Wilmad SQ EPR tubes (Wilmad Glass, Buena, NJ) were used for EPR samples. To minimize the effect that slightly different sizes of tubes might have on the quantitative results, only selected tubes were used with outer diameters of 4.05 Ϯ 0.07 mm and inner diameters of 3.12 Ϯ 0.04 mm (mean Ϯ range). To estimate the concentration of paramagnetic centers in the EPR tubes, a set of 0.5 mM standard samples of MNIC-MGD in HEPES was used. The observed concentrations of MNIC-MGD and MNIC-DETC in the EPR tubes are illustrated on each figure, as is the resultant calculation of the rate of NO production in leaves (in nmol/g of wet weight of leaves/h). Although the leaf weight and the time of NO accumulation were the same in the MGD and DETC experiments, the transition coefficients from the observed concentrations of trapped NO in the EPR tubes (in M) to the NO production in leaf (in nmol g Ϫ1 h Ϫ1 ) were different for the two types of experiment. This is because in the case of the water-soluble MGD trap, the whole amount of NO produced by the leaf sample was dissolved in the incubation volume (1-ml Eppendorf vial), of which only an aliquot was taken as an EPR sample, whereas in the case of the lipid soluble DETC experiments it was assumed that all NO produced by the leaf sample is trapped inside the leaf tissue, all of which was then transferred to the EPR tube. Fig. 1 demonstrates that isolated leaves produce NO that can be trapped either in the incubation medium or directly in the leaf tissue. The concentration of both MNIC-dithiocarbamate complexes increased approximately linearly with time, within the interval of 0 -3 h (illustrated for the MNIC-MGD complexes in Fig. 2). Both kinds of MNIC complexes, the water-soluble MNIC-MGD that accumulates in the medium around the cuts of the leaves and the lipid soluble MNIC-DETC that accumulates in the intracellular membranes, gave similar EPR signals with a g factor symmetry close to axial, spreading from g ϭ 2.049 (peak) to g ϭ 2.025 (trough) and with a triplet hyperfine structure in the area of the perpendicular (low field) components with a distance between components of ϳ10 Gauss (Fig. 1). The hyperfine structure resolution in the low field component in the MNIC-MGD signals is not as good as in the MNIC-DETC signals, probably because of the effect of saltification of MNIC-MGD complexes during freezing that results in an increase of the spin-spin interaction. In some MNIC-MGD samples, an asymmetric signal with g ϭ 1.98 was detected at 10 K registration temperature (Fig. 1). The parameters of this signal are close to those that can occasionally result from the freezing of NO molecules in solutions (36). An EPR signal of a low intensity from the Cu 2ϩ -MGD (and Cu 2ϩ -DETC) complexes was a characteristic of all preparations. This signal was particularly distinctly observed in the EPR spectra of the preparations without MNIC-MGD (MNIC-DETC) complexes (Fig. 1d). The signal is characterized by a quartet hyperfine structure arising from the nuclear spin of copper. The doublet splitting in the free radical EPR signal at g ϭ 2.0 in the MNIC-DETC spectra is mainly due to the "dark" free radicals in the photosynthetic system of the leaves (this signal is missing in the MNIC-MGD spectra when the medium of the incubation mixture was analyzed rather than the leaves themselves).

Isolated Leaves from Beans and A. thaliana Can Spontaneously Generate NO in Vitro -
The addition of 100 mM NaNO 3 to the leaf medium resulted in a 3-4-fold decrease of formation of MNIC-DETC (Fig. 3, a and c, and Table I), whereas L-nitroarginine methyl ester acetate (a nitric-oxide synthase inhibitor) added at the same concentration did not have a noticeable effect (not shown). When a 10 mM nitrate concentration was used, the amount of MNIC-DETC decreased by 60 -70% of the control value (Fig. 3, a and b, and Table I). A similar effect of nitrate addition on the MNIC-MGD formation was also observed (data not shown). This result favors the idea that NR is a significant source of NO production in plant leaves under basal conditions. We suggest that, in agreement with other studies (18,20), nitrate added at high millimolar concentrations inhibits the nitrite reductase activity of the enzyme by a competitive mechanism, thereby attenuating NO production in the leaves.
Nitrite and Nitrate as NO Sources in Leaves-The possibility that NR plays a crucial role in NO generation in leaves points to nitrite as the ultimate substrate of the process. The addition of 1 mM nitrite alone resulted in a sharp increase of formation of MNIC-MGD ( Fig. 4, a* and b*, and Table I). Even in the preparations where spontaneous formation of the MNIC-MGD complexes was not detected (Fig. 4a), the addition of 1 mM nitrite resulted in a readily quantifiable measure of these complexes ( Fig. 4b and Table I). When 1 mM nitrite was added together with 100 mM nitrate, the effect of the complex induction was diminished ( Fig. 4, b, c, b*, and c*, and Table I). Curiously, nitrate could induce formation of MNIC-DETC in some leaf preparations that were apparently not capable of spontaneously producing these complexes (Fig. 5). The effect was more pronounced at lower concentrations of nitrate (Fig. 5,  b and c). This paradoxically stimulatory effect of nitrate on NO generation can be explained by a rapid transformation of nitrate to nitrite, catalyzed by NR, followed by a catalytic reduction of nitrite to NO. The rate of the NR induced production of NO must therefore depend on the nitrite/nitrate ratio and the total concentration (nitrite ϩ nitrate) in a complex way; at low levels of nitrite, adding nitrate will increase nitrite levels to those necessary to produce detectable NO, and at higher levels of nitrite, the primary effect of nitrate will be to compete with nitrite at the active site and decrease the NO production rates. This effect obviously needs further experimental study and associated kinetic modeling, but it does suggest that the relative and absolute levels of nitrite and nitrate are key determinants in the basal level of NO production in leaves.
It has been shown before that NO can be synthesized nonenzymatically from nitrite through its direct interaction with Fe 2ϩ -dithiocarbamate complexes (37). Clearly, that cannot be the case for hydrophobic complexes, such as Fe 2ϩ -DETC, localized in intracellular membrane compartments to which nitrite does not have access. However, the reaction was shown to take place for water-soluble Fe 2ϩ -dithiocarbamate complexes, such as Fe 2ϩ -MGD (37). Thus, the possibility exists that the MNIC-MGD complexes in the leaf preparations are formed via this nonenzymatic mechanism of nitrite reduction by the Fe 2ϩ -MGD complexes in the medium and not via the intracellular NO generation by NR. Nitrite, in turn, could be produced enzymatically by the leaf. However, this hypothesis can be dismissed if quantitative considerations were taken into account.
The steady state concentration of nitrite in the leaf tissue is about 10 M (18). If, hypothetically, we could instantly release all that concentration to the incubation medium to allow nitrite to interact with Fe 2ϩ -MGD, nitrite would have been diluted to   (Fig. 1, e and f). Therefore intercellular nitrite concentrations of the order of 10 mM would be necessary to induce this level of MNIC-MGD nonenzymatically. This is only possible under extremes of conditions (180 min of dark anoxia) not present in these studies (18). Therefore we conclude that the observed level of formation of the MNIC-MGD complexes in the leaf incubation medium cannot be explained exclusively, or indeed significantly, by the formation of NO formed via nitrite reduction by Fe 2ϩ -MGD.
Why Is the Amount of Spontaneously Generated NO So Variable in Isolated Leaves?-The method of NO trapping by ferrous-dithiocarbamate is characterized by a large variation in the concentration of detected complexes. There are two probable explanations for this. The first one is that variations in the nitrite/nitrate ratio in the leaves can have a very complex effect on NO production, as described previously. However, small variations in the rate of superoxide anion production can also cause variations in the basally detected level of MNIC. The presence of superoxide in the leaves will transform MNIC-MGD (and MNIC-DETC) complexes to an EPR-silent state (34,35). It is well known that plants respond to a stressful environment with an activation of their superoxide production (30). The extent of the activation depends on changes in plant cultivation, for example on the change of light condition, temperature, humidity, etc. All of these factors could result in a dramatic decrease in the concentration of detected MNIC-MGD (MNIC-DETC) complexes, sometimes down to a zero level. We have previously shown, for example, that increased superoxide production accompanies leaf wounding (33,38), such as the cuts used in the leaf preparations in this study to allow access of the NO scavengers to the plant cells.
We have previously proposed a novel method (termed "ABC") that allows us to correct for this superoxide effect when evaluating the level of endogenously produced MNIC-MGD (MNIC-DETC) complexes (34,35). In this method, a known concentration of exogenous paramagnetic MNIC-dithiocarbamate complexes is added to the biological cells or tissues. This quantity, being in a significant excess, acts as a very efficient scavenger of superoxide anions (34). To trap endogenous NO, iron and dithiocarbamate (DETC or MGD) are also added to the system. The MNIC-dithiocarbamate EPR signal measured in such experiment is presented as A ϩ B Ϫ C where A is the signal from the exogenously added MNIC-dithiocarbamate complexes, B is the signal caused by endogenous NO generation, and C accounts for the concentration of MNIC-dithiocarbamate complexes transformed to the EPR-silent state by superoxide. We will refer to the A ϩ B Ϫ C experiment as Type I. The experiment is then repeated in the presence of an inhibitor of NO synthesis (Type II study). Now the measured MNICdithiocarbamate EPR signal measured is represented as A Ϫ C (B is missing now because of the inhibition of endogenous NO synthesis). The subtraction of the signals measured in the two experiments results in a signal arising purely from the endogenously generated MNIC-dithiocarbamate complexes (B) with no reduction in concentration because of the interaction of superoxide and NO.
In the type I experiment (A ϩ B Ϫ C), 5 mM DETC, 5 mM sodium ascorbate, 0.5 mM Fe 2ϩ , and 0.3 mM NO-Pr (an exogenous NO donor with a half-life of 30 s) were added in the indicated sequence to a 1-ml vial with the bean leaf cuts in a 100 mM HEPES buffer (pH 7.4). The type II experiment (A Ϫ C) requires the additional presence of an inhibitor of endogenous NO synthesis. We used sodium nitrate at high (100 mM) concentrations to inhibit NO production from NR.
Nitrate, however, is not a perfect inhibitor, because it can be reduced by NR to nitrite, and the latter can be the source of additional NO production by NR. Therefore we also used an alternative method to that of inhibition of the endogenous NO production. Instead of nitrate, we added hemoglobin (Hb) at a concentration of 100 M to the medium 5 min after the introduction of NO-Pr; at this time all of the exogenous NO that has not reacted with Fe 2ϩ -DETC would have been removed. Hb is a good NO scavenger and provides an external sink for NO. In its presence it would be impossible for any endogenously formed NO molecules that escaped from the leaf tissue to the medium to return to the leaf and be trapped by the intercellular Fe 2ϩ -DETC complexes. Effectively, Hb competes with Fe 2ϩ -DETC for the endogenously formed NO. In this modification of the ABC method, partial inhibition of the MNIC-DETC complex formation from endogenous sources is achieved without the need for a specific inhibitor of NO synthesis.
The result of using of the ABC method in the two versions described above is shown in Fig. 6 and summarized in Table I. Whereas production of the MNIC-DETC complexes in the conventional experimental set-up (B Ϫ C) is low, sometimes undetectably so (0 -1.5 nmol g Ϫ1 h Ϫ1 ), the employment of the ABC method reveals that the value of the endogenous NO production (B) is actually significantly higher (17-18 nmol g Ϫ1 h Ϫ1 ). This level of NO production could have been readily detected by EPR spectroscopy in all leaves, if most of the MNIC-DETC complexes formed were not turned into EPR-silent species by, most probably, superoxide anion.
The fact that the two modifications of the ABC method (experiments 5 and 6) gave very close values of B (Table I)  have their shortcomings associated with different methods of inhibiting MNIC-DETC complexes formation from endogenously formed NO. Obviously, both methods are not 100% effective. Consequently, the real values of B in plants must be even greater than those obtained in this work. To support this statement, we can point to a single experiment (Fig. 1e) where we found the rate of spontaneous NO production in A. thaliana treated with MGD approximately twice as high (36 nmol g Ϫ1 h Ϫ1 ) when compared with the B ϭ 18 nmol g Ϫ1 h Ϫ1 determined by the ABC method.
Although the amount of spontaneously generated NO detected was very variable, in general greater values were observed in the MGD experiments, when compared with the DETC ones. Both the high variability of the detected MNIC-MGD (MNIC-DETC) and the higher average value for MNIC-MGD as compared with that for MNIC-DETC can be rationalized in terms of the superoxide anion effect. The level of NO detected is dominated by the destruction of the MNIC complexes by superoxide (as the "true" level, in the presence of superoxide scavenging, is over 10 times the level detected with no scavenging). Therefore relatively small variations in the superoxide status of leaves can result in very different concentrations of detectable MNIC. A significantly lower concentration of superoxide is expected, in general, in the MGD experiments because the EPR samples are taken from the incubation medium where superoxide is approximately eight times diluted when compared with its concentration in the leaf tissue (as in the DETC experiments).
The presence of ascorbate in the medium also contributes to the improved NO trapping by MGD. When added to the aerobic medium, the reduced MNIC are immediately oxidized to the ferric state. However, ascorbate re-reduces this to the ferrous state, enabling NO to bind and form a paramagnetic species. Although ferric MGD can bind NO, the resultant complex is EPR-silent. This same diamagnetic ferric NO complex is formed following superoxide oxidation of the paramagnetic MNIC. Ascorbate will re-reduce this to the paramagnetic state (39). Therefore ascorbate has multiple roles in enhancing NO detection by MGD.
The concentration of ascorbate used in the external leaves (5 mM), although high for mammalian systems, is unlikely to perturb the redox state of the plant leaves; the concentration of ascorbate in leaf cells (40) is well in excess of this (20 -50 mM). One problem that may arise, however, is that the ascorbate/ MGD/iron system could undergo redox cycling, whereby ferrous MGD reacts with oxygen to produce superoxide radicals in the medium; the superoxide so formed could oxidize MNIC and thus reduce the efficiency of NO detection. Although, the experiments are not formally anaerobic, the presence of ascorbate and the dithiocarbamate decrease the oxygen tension considerably. Although there could still be a small superoxide flux because of oxygen diffusion into the system, the combination of ascorbate and the iron-MGD(DETC) complexes are not likely to be major producers of superoxide because lowering the ascorbate concentration decreases, not increases, the amount of the paramagnetic NO complexes (by a factor or approximately five; results not shown). Clearly the ability of ascorbate to reduce diamagnetic complexes is more important than its ability to produce superoxide (and hence convert the paramagnetic species to diamagnetic). This was confirmed by the fact that we found no significant difference in the ability of MGD to trap NO in the presence (3.4 Ϯ 0.9 M, mean Ϯ S.D., n ϭ 3) or absence (3.8 Ϯ 1.6 M, mean Ϯ S.D., n ϭ 3) of 60 units/ml superoxide dismutase.
Endogenous Spin Traps: the Formation of Dinitrosyl Iron Complexes in Leaf Preparations-The formation of paramagnetic dinitrosyl iron complexes (DNIC) from endogenous free iron and thiol-containing ligands has been demonstrated before for isolated bean and China rose leaves treated with exogenous gaseous NO (41). The complexes are characterized by an EPR signal with g values of g Ќ ϭ 2.037 and g ʈ ϭ 2.012 (the "2.03 signal" is so called because of the value of the average g factor,  ). B represents the MNIC-DETC concentration arising from the NO fraction generated by the leave tissue (the endogenous component). C represents the concentration of the MNIC-DETC complexes transformed to the diamagnetic (EPR-silent) state by the superoxide. a and d, the leaf preparations were treated with 5 mM DETC in the presence of 5 mM ascorbate and 0.5 mM ferrous sulfate (added last, the leaf cuts added second last); b and e, as for a and d, but 0.3 mM NO-Pr was added straight after addition of ferrous sulfate; c, as for b, but 100 mM nitrate was added just before the leaf cuts were placed into the solution, and the cuts were incubated for 10 min before the addition of ferrous sulfate; f, as for e, but 100 M Hb was added 5 min after NO-Pr addition. All of the leaf preparations were incubated at room temperature for 3 h starting from the moment of iron addition. g av ϳ 2.03). Therefore it is reasonable to expect formation of DNIC in plant leaves treated with nitrite, when the production of NO might be expected from NR. In agreement with this, when isolated Chinese rose leaves were incubated at ambient temperature with 10 mM nitrite in a 150 mM HEPES (pH 7.4) for 3 h under anaerobic conditions (the buffer was bubbled with argon for 30 min), an EPR signal at g ϭ 2.04 appeared (Fig. 7a). Incubation of the leaves with a mixture of 10 mM nitrite and 200 mM nitrate resulted in a 4-fold decreased signal intensity (Fig. 7b). The signal was not detected in the non-nitrite-treated leaf preparations (data not shown). Detection of DNIC at a lower nitrite concentration of 1 mM nitrite was not possible because of the masking effect by the third hyperfine splitting component of EPR signal from manganese complexes (Fig. 7). The subtraction of the EPR spectra a and b (Fig. 7c), results in an EPR signal similar to the standard EPR spectrum from DNIC with cysteine in frozen solution (Fig. 7d). The latter is characterized by g factor values of gЌ ϭ 2.04 and g ʈ ϭ 2.014 (g av ϭ 2.03). The similarity of the signals makes it possible to suggest that the signal shown in Fig. 7c is due to DNIC formation from thiol-containing ligands in the leaves. The intensity of the signal corresponds to a rate of DNIC formation of 8.0 Ϯ 2.0 nmol g Ϫ1 h Ϫ1 . Similar EPR signals of DNIC were observed in isolated Chinese rose leaves treated with gaseous NO for 20 -30 min under a pressure of 200 mm Hg, consistent with the results described in Ref. 41 (data not shown).

DISCUSSION
In this study we applied EPR spin trapping techniques developed in measurements of mammalian NO synthesis to the production of NO in healthy leaves, unchallenged with pathogens. The results can be summarized as follows: (i) Nitrate and nitrite are the main nitrogen sources for the basal level of NO production in leaves of V. faba, A. thaliana, and H. rosa sinensis. This finding is consistent with the view that nitrite reduction by NR is responsible for NO generation in leaves (1,7,9,18,20,21,42). (ii) Superoxide anion production interferes with the detection of endogenous NO. Two possible mechanisms for this interference are suggested; either it can decrease the concentration of NO in leaves by direct interaction producing peroxynitrite, or it can transform the MNIC-DETC (MNIC-MGD) complexes, detectable by EPR, to the EPR-silent species. (iii) The high variation in the production of NO, as evaluated by the ferrous-dithiocarbamate method, can be explained by the variation in the production of superoxide anions in the leaves. (iv) NO trapping in the incubation medium using the hydrophilic Fe 2ϩ -MGD complexes turns out to be more efficient, on average, than NO trapping inside the leave tissue using the hydrophobic Fe 2ϩ -DETC complexes. The main reasons for this are probably a combination of the lower concentration of superoxide present in the external medium available to react with the NO-trapped species and the relative difficulty of endogenous ascorbate to reduce the hydrophobic DETC species, compared with the hydrophilic MGD.
Control over this basal level of NO production is likely to be complex. NR levels and activity are controlled by phosphorylation; it has been shown that, in addition to increasing the "normal" reduction of nitrate to nitrite, activation of NR by dephosphorylation increases NO production from nitrite (18). It has also been demonstrated that cytosolic NR gene expression and the activity of the enzyme in higher plants is controlled by the signals from photosynthetic electron transport chains in chloroplasts, themselves linked to superoxide production (19). Therefore the functional state of chloroplasts can affect the level of NO production in the leaves via modulating the expression and activity of NR. In the dark uncouplers of oxidative phosphorylation activate NR. This suggests a possible link between mitochondrial function and NO production. This is interesting in the light of the well characterized NO inhibition of mammalian mitochondrial respiration (43). Plant mitochondrial respiration is intrinsically less sensitive to NO inhibition because the alternative cyanide insensitive terminal oxidase is less sensitive to NO than the aa 3 -type oxidase (44). Indeed NO induces the transcriptional activation and synthesis of the alternative oxidase (19,45). However, switching oxidases under conditions of enhanced NO production would modulate the efficiency of plant respiration. The ATP produced per oxygen consumed (P/O ratio) is lower if oxygen is reduced via the alternative oxidase, and this reduction in the efficiency of ATP production can cause significant physiological effects (46).
One criticism of exogenous spin traps is that they can perturb the system they are measuring. In the case of dithiocarbamates, both NO and superoxide production will be modulated by adding such a dual NO/superoxide trap. However, the data supporting nitrite as a substrate for NO production do not rely solely on the addition of exogenous spin traps. NO formation from nitrite was confirmed via the formation of paramagnetic DNIC from endogenous iron and thiol-containing ligands in Chinese rose leaves. The concentration of the complexes accumulated over 1 h was found to be 8 nmol g Ϫ1 h Ϫ1 . It is possible that this level of formation of DNIC from NO production by NR has an effect on the non-heme iron metabolism in plant tissues, particularly by making iron more bioavailable. Recent data suggest iron availability inside plant leaves is directly dependent on the presence of NO (47); NO treatment completely prevented the leaf interveinal chlorosis normally associated with iron deprivation in maize growing in a low iron medium (Ͻ10 M Fe-EDTA). A similar protection was observed when MNIC with cyanide, nitroprusside anions, S-nitrosothiol, or S-nitroso-N-acetylpenicillamine was added to the nutrient solution. It is reasonable to suggest that all of these treatments resulted in DNIC formation inside leaves that effectively protected the internal non-heme iron from precipitation in the form of hydroxide complexes, thereby ensuring that iron stays available for various intracellular components, particularly the chloroplasts.
One of the most significant findings in this work is that other methods, which do not correct for superoxide destruction of NO, significantly underestimate the true rate of basal NO production in plants. The values reported in previous studies range from 0.5 to 1 nmol g Ϫ1 h Ϫ1 (18). This is consistent with our base-line values. However, once corrected for superoxide production our values increase to 18 nmol g Ϫ1 h Ϫ1 . These values, although higher, are still significantly lower than those achievable following maximal activity of the inducible form of the enzyme in tobacco (23) (1,200 nmol g Ϫ1 h Ϫ1 ) or rat liver (48) (2,000 nmol g Ϫ1 h Ϫ1 ). However, nitrite levels in the range of 5 mM induced by anoxia have been reported to increase NO production to levels as high as 200 nmol g Ϫ1 h Ϫ1 (18). Taken as a whole this quantitative study supports the recent idea (23) of a "functional homology" between NO production in plants and animals. Basal levels of NO are low and support normal physiological signaling ("neuronal" NOS/"endothelial" NOS in animals and NR/AtNOS1 in plants); pathogen induction or major physiological stress induces high levels of production ("inducible" NOS in animals, pathogen inducible iNOS or NR in plant, the latter only if nitrite concentrations reach the millimolar range).
The significant effect of superoxide on detectable NO levels in leaves is another reason why it has been more difficult than in animals to measure NO production. It is also likely to have functional significance. The rate of superoxide production from photoreduction of oxygen in leaves is variable and is dependent on the light intensity and capacity for transfer of electrons from Photosystem I to NADP (49). Superoxide (and consequent hydrogen peroxide) production are also increased in response to a range of stimuli in plants (30). NO production has been shown to decrease superoxide concentrations in a range of plant systems (21, 50 -52). However, the converse is also likely to be true. Increased superoxide production is likely to diminish the concentration of NO, resulting in a complex interaction between oxygen and nitrogen radical signaling, as has been recently suggested to occur in mammalian systems (53,54). However, there are possible strategies that the plant can use to maintain NO levels in the presence of superoxide. Clearly, given the limited membrane permeability of superoxide and its major production inside the chloroplast, spatial heterogeneity may be one such strategy. Although its subcellular localization is yet to be determined, it is perhaps significant in this context that plant variant P/"inducible" NOS is derived from a mitochondrially targeted enzyme (glycine decarboxylase). Interestingly we have recently described an alternative means of maintaining functional NO levels in the presence of superoxide. Ascorbate is able to convert peroxynitrite into NO, thus regenerating free NO from the product of its reaction with superoxide (55,56). This reaction is further catalyzed by the presence of the mitochondrial enzyme, cytochrome c oxidase (56). We therefore suggest that an additional function of the high ascorbate concentrations in plants may be to preserve NO signaling pathways from variations in superoxide flux.
A consequence of the interplay of superoxide and NO is likely to be peroxynitrite-catalyzed protein modification. Indeed increased tyrosine nitration has been seen in leaves, following increases in endogenous NO production or exogenous peroxynitrite addition (17). The increasing number of novel plant-based peroxynitrite scavengers recently identified with possible therapeutic benefits (57)(58)(59) is perhaps a testament to the requirement for plants to survive in a high peroxynitrite environment, the inevitable consequence of producing NO in a photosynthetic organism.