Localization of Nitric-oxide Synthase in Plant Peroxisomes*

The presence of nitric-oxide synthase (NOS) in peroxisomes from leaves of pea plants (Pisum sativum L.) was studied. Plant organelles were purified by differential and sucrose density gradient centrifugation. In purified intact peroxisomes a Ca2+-dependent NOS activity of 5.61 nmol ofl-[3H]citrulline mg−1 protein min−1 was measured while no activity was detected in mitochondria. The peroxisomal NOS activity was clearly inhibited (60–90%) by different well characterized inhibitors of mammalian NO synthases. The immunoblot analysis of peroxisomes with a polyclonal antibody against the C terminus region of murine iNOS revealed an immunoreactive protein of 130 kDa. Electron microscopy immunogold-labeling confirmed the subcellular localization of NOS in the matrix of peroxisomes as well as in chloroplasts. The presence of NOS in peroxisomes suggests that these oxidative organelles are a cellular source of nitric oxide (NO) and implies new roles for peroxisomes in the cellular signal transduction mechanisms.

Nitric oxide (NO) 1 is a widespread intra-and intercellular messenger which is involved in a wide range of functions such as the regulation of vascular tone, neuronal signaling, and immune response to infection (1). As a consequence of the physiological importance of the free radical NO, numerous studies have been focused on the enzyme responsible for its endogenous production. Nitric-oxide synthase (NOS; EC 1.14.13.39) catalyzes the NADPH-dependent conversion of Larginine to NO and citrulline in a complex reaction requiring molecular oxygen, calcium, calmodulin, FAD, FMN, and tetrahydrobiopterin (1). NOSs have been characterized as functional homodimers with a large subunit (Ն130 kDa) that contain iron-protoporphyrin IX, flavin adenine dinucleotide (FAD), fla-vin mononucleotide (FMN), and tetrahydrobiopterin (BH 4 ) as prosthetic groups (1)(2)(3). Three NOS isoforms have been identified in mammalian tissues: neuronal NOS (nNOS or type I), inducible NOS (iNOS or type II), and endothelial NOS (eNOS or type III). While types I and II are present in the soluble fraction, type III has been shown to be membrane-associated and none of the isoforms has a strict tissue-specific expression pattern (3). In organelles from animal systems, the presence of NOS activity has only been detected so far in mitochondria of rat liver (4).
In plants, several recent studies have shown that NO could function as a signal in disease resistance (5)(6)(7)(8). However, NOSlike activity has only been detected in plant extracts during the interaction of Rhyzobium-legume (9) and fungi plants (10), and in soybean cell suspensions induced with an elicitor from Pseudomonas syringae (8). Very recently, the presence of NOS in the cytosol and nucleus of maize cells has been reported (11). But no information is available concerning the occurrence of NOS in other plant cell organelle.
Peroxisomes are single-membrane bound organelles that contain catalase and H 2 O 2 -producing flavin oxidases as their basic enzymatic constituents, and are found in virtually all eukaryotic cell types (12,13). A characteristic property of these oxidative organelles is their metabolic plasticity since their enzymatic content can vary depending on the organism, cell/ tissue-type, and environmental conditions (13,14). In higher plants, peroxisomes contain a complex battery of antioxidative enzymes such as catalase, superoxide dismutase (15,16), the components of the ascorbate-glutathione cycle (17), and the NADP-dehydrogenases of the pentose-phosphate pathway (18). Likewise, the generation of superoxide radicals has also been reported in the matrices and membranes of peroxisomes (19 -21). All these findings evidence a relevant role for peroxisomes in the cellular metabolism of activated oxygen species (22).
In this work, we report the presence of NOS in peroxisomes (perNOS) of pea leaves using four complementary approaches: (i) distribution of NOS activity in cell organelle fractions purified by sucrose density gradient centrifugation, (ii) sensitivity of NOS activity to well characterized NOS inhibitors, (iii) crossreactivity of peroxisomes on Western blot with antibodies against iNOS, and (iv) electron microscopy immunocytochemical analysis of intact plant tissue.

EXPERIMENTAL PROCEDURES
Plant Material and Growth Conditions-Pea (Pisum sativum L., cv. Lincoln) seeds, obtained from Royal Sluis (Enkhuizen, Holland), were surface-sterilized with 3% (v/v) commercial bleaching solution for 3 min, and then were washed with distilled water, and germinated in vermiculite for 15 days. Healthy and vigorous seedlings were selected and grown in the greenhouse in nutrient solutions under optimum conditions (23) for 15 days.
Purification of Peroxisomes-All operations were performed at 0 to * This work was supported by the Dirección General de Enseñ anza Superior e Investigación Científica (DGESIC) Grant PB95-0004-01, the Junta de Andalucía (Research Groups CVI 0157 and CVI 0192), Spain, and European Union contract CHRX-CT94-0605. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Contributed equally to the results of this work. 4°C. Peroxisomes were purified from pea leaves by differential and sucrose density gradient centrifugation (35-60%, w/w) as described by López-Huertas et al. (24). Purified peroxisomes were practically free of chloroplasts and mitochondria, as confirmed by measuring the activities of the appropriate marker enzymes (24,25). Enzyme Activities and Other Assays-Catalase activity was assayed as described by Aebi (26) and fumarase activity was measured by the method of Walk and Hock (27). NO synthase activity was determined by monitoring the conversion of L-[ 3 H]arginine to L-[ 3 H]citrulline (28). The assays for total NOS activity were performed in duplicate for each sample for 20 min at 37°C in a reaction medium containing 40 mM Hepes buffer, pH 7.2, 0.2 mM CHAPS, 10 g/ml calmodulin, 1.25 mM CaCl 2 , 1 mM ␤-NADPH, 10 M FAD, 10 M FMN, 10 M tetrahydrobiopterin (BH 4 ), variable amounts of protein samples, and a variable concentration of L-arginine supplemented with L-[ 3 H]arginine (1.0 mCi ml Ϫ1 ) in a total volume of 220 l. The reaction was terminated by addition of 1.5 ml of a cation exchange resin (Dowex 50W, 8% crosslinkage, 200 -400 mesh, Na ϩ form) at 4°C which stops the reaction by removing the arginine substrate. Then, 5 ml of water was added and the resin was allowed to settle for approximately 10 min before aspirating 3.5 ml of supernatant which was analyzed by liquid scintillation counting. The level of L-[ 3 H]citrulline was computed after subtracting the blank value, which represented the nonspecific radioactivity in the absence of enzyme. In addition, the values obtained in the experiments performed in the absence of NADPH, which indicate the amount of L-[ 3 H]citrulline formation independent of a specific NOS activity, were also subtracted from the total level of L-citrulline. The activity was expressed as picomole of L-[ 3 H]citrulline min Ϫ1 mg Ϫ1 protein.
Protein levels were determined according to Bradford (29) using bovine serum albumin as standard. The density of the gradient fractions was calculated at room temperature using an Atago refractometer.
Antibodies-Two antibodies have been used in this study. A monoclonal antibody against the NADPH-binding region (residues 961 to 1144) of iNOS from mouse macrophage (Transduction Lab, Lexington, KY) and a polyclonal antibody against the peptide PT387 (Ac-Cys-(residues 1131 to 1144)) from the C terminus of the deduced amino acid sequence of murine iNOS (30).
Electrophoretic and Immunoblot Analyses-SDS-PAGE was carried out in 7.5% acrylamide slab gels. Samples were prepared in 62.5 mM Tris-HCl, pH 6.8, containing 2% (w/v) SDS, 10% (v/v) glycerol, and 10 mM dithiothreitol, and were heated at 95°C for 5 min. For immunoblot analyses, polypeptides were transferred onto polyvinylidene difluoride membranes (Immobilon P, Millipore Corp., Bedford, MA) using a Semi-Dry Transfer System (Bio-Rad) with 10 mM CAPS buffer, 10% (v/v) methanol, pH 11.0, at 1.5 mA cm Ϫ2 for 2 h (31). For slot blot studies, proteins were also transferred onto polyvinylidene difluoride membranes using a Bio-Dot SF Apparatus (Bio-Rad) following the instructions of the manufacturer. For immunodetections, the two antibodies against iNOS were used with an enhanced chemiluminescence kit (ECL-PLUS, Amersham Pharmacia Biotech) and were detected with a photographic film (Hyperfilm; Amersham Pharmacia Biotech).
Electron Microscopy and Immunocytochemistry-Pea leaf segments of approximately 1 mm 2 were fixed, dehydrated, and embedded in LR White resin as described by Corpas and co-workers (31). Ultrathin sections were incubated for 3 h with IgG against iNOS (30) diluted 1/250 in TBST buffer (10 mM Tris-HCl, pH 7.6, containing 0.9% (w/v) NaCl, 0.05% (v/v) Tween 20) containing 2% (w/v) bovine serum albumin and 1% (v/v) goat normal serum. The sections were then incubated for 1 h with goat anti-rabbit IgG conjugated to 15-nm gold particles (Bio Cell) diluted 1/50 in TBST plus 2% (w/v) bovine serum albumin. Sections were poststained in 2% (v/v) uranyl acetate for 3 min and examined in a Zeiss EM 10C transmission electron microscope. Preimmune serum, instead of the iNOS antibody, was used as control. A second control was used omitting the primary antibody.

RESULTS
The purification of peroxisomes from pea leaves by sucrose density gradient centrifugation is shown in Fig. 1. Peroxisomes (fractions 20 -25) were identified by the peak of catalase activity, used as peroxisomal marker enzyme. The peroxisomal fraction banded at an average equilibrium density of 1.24 g cm Ϫ3 , characteristic for these intact organelles in sucrose solutions (12,24). The absence of fumarase activity in these fractions indicated that peroxisomes were essentially free of contamination by mitochondria. The assay of NOS activity throughout the gradient fractions showed a peak of activity in the top of the sucrose gradient (fractions 1-3), corresponding to the zone of broken organelles, and another NOS peak in the peroxisomal fractions (fractions [21][22][23][24][25]. A maximum NOS activity of 170 pmol of L-[ 3 H]citrulline min Ϫ1 mg Ϫ1 protein was found in the peroxisomal fraction 23 which coincided with the maximum of catalase activity. The NOS activity of peroxisomes, measured as production of L-[ 3 H]citrulline, showed a linear increase with the amount of peroxisomal protein (data not shown). The specific activity of the peroxisomal NOS was reduced more than 70% in the absence of calcium (Table I) and was strictly dependent on NADPH.
The effect on the peroxisomal NOS activity of seven well characterized inhibitors of the three known mammalian NO synthases is shown in Table II. An inhibition of the NOS activity by 59 -100% was obtained with the inhibitors L-N 5 -(1iminoethyl)-ornithine, L-NMMA, L-NAME, L-thiocitrulline, 7-nitroindazole, diphenyliodonium, and aminoguanidine. When the NOS assays were performed using a 100 M L-Arg concentration, a reversion of about 30% of the inhibitory effect of 7-nitroindazole, L-NMMA, and L-NAME was observed (results not given). However, the inhibitory effect of aminoguanidine, which is an irreversible iNOS inhibitor, was not affected by 100 M L-Arg. The NOS activity was not sensitive to D-NAME, an enantiomer of L-NAME, which is used as a negative control.
The analysis by slot blot of both pea leaf crude extracts and pea leaf peroxisomal fractions with a monoclonal antibody against mouse macrophage NOS (iNOS) gave a positive reac-FIG. 1. Isolation of peroxisomes from pea leaves by sucrose density gradient centrifugation and localization of NOS activity. Gradient fractions of 1.5 ml were eluted with a gradient fractionator and assayed for specific marker enzymes, fumarase for mitochondria and catalase for peroxisomes. Catalase and fumarase activities are expressed as micromole min Ϫ1 ml Ϫ1 and NOS activity as picomole of L-[ tion (Fig. 2, panel A). SDS-PAGE and immunoblot analysis of the peroxisomal fractions using the polyclonal antibody against the peptide PT387 from the C terminus of the murine iNOS revealed an immunoreactive polypeptide of about 130 kDa (Fig.  2B, lane 4). Immunoreactive bands with similar mobility were obtained with the hepatic iNOS from lipopolysaccharide-induced rat and murine macrophage lysates (Transduction Lab, Lexington, KY) used as positive controls (Fig. 2B, lanes 2 and  3). The assay of the NOS activity after incubation of the peroxisomal fractions with the murine iNOS antibody, showed an inhibition of about 90% with a 1/200 dilution of this antibody (Fig. 3), and no inhibition was observed with a dilution of 1/100,000.
The ultrastructural immunolocalization of NOS with the antibody against murine iNOS showed that gold particles were localized inside two cell compartments, peroxisomes and chloroplasts, while no labeling was observed inside mitochondria (Fig. 4, panels B and C). No significant labeling was observed in the controls (Fig. 4, panel A). DISCUSSION In higher plants, recent data have shown that exogenous nitric oxide has an important role as messenger in resistance against diseases produced by plant pathogens (6,8,32), and several groups have detected NOS activity in plant extracts (8 -11). Considering that plant peroxisomes have a role in the metabolism of reactive oxygen species (22) and are also a source of H 2 O 2 which may function as a second messenger (33)(34)(35), the potential presence of NOS in peroxisomes was examined.
The first approach was to study the distribution of NOS activity throughout the sucrose density gradient fractions. Two peaks of L-[ 3 H]citrulline production were detected in the gradient fractions. One peak was in the top of the gradient, and corresponded to the broken organelles, and another peak of NOS activity was observed in the peroxisomal fractions. The peroxisomal NOS activity, measured as L-citrulline production, was protein-and calcium-dependent (Table I).
As part of the initial characterization of the NOS activity of leaf peroxisomes, the effect of seven archetype NOS inhibitors, including specific and unspecific inhibitors of different types of NOS isoforms was assayed. The results indicated a clear inhibition of the perNOS activity of 59 -100%, being the L-aminoguanidine the most effective inhibitor. In extracts of maize seedlings a 31% inhibition of NOS activity by 3 mM L-aminoguanidine plus 3 mM L-NAME was reported (11), and in extracts of lupin roots a 50% inhibition of NOS activity by 1 mM L-NMMA was described (9). However, the specific activity reported by these authors is much lower (0.2 and 0.7 pmol of L-citrulline mg Ϫ1 protein min Ϫ1 , respectively) than that obtained in pea leaf peroxisomes (5.6 nmol of L-citrulline mg Ϫ1 protein ⅐ min Ϫ1 ). Likewise, in this work the production of citrulline was specifically due to a NOS because an specific antibody against iNOS inhibited this reaction by more than 80% (Fig. 3).
The NOS activities described in crude extracts of roots of Lupinus albus (9), soybean cell suspensions (8), and maize seedlings (11) were calcium-dependent, moderately sensitive to    inhibitors, and were mainly found in the soluble fractions. In pea leaves, the perNOS was also calcium-dependent but was clearly sensitive to inhibitors of iNOS, nNOS, and eNOS. Nevertheless, none of the NOS activities detected so far in plants have been found identical to the NOS isoforms present in mammals, indicating that plant NOSs could be slightly different from the known NOS isoforms of mammals.
To get deeper insights into the presence of NOS activity in pea leaf peroxisomes, two immunological approaches were used, with two types of antibodies against NOS being assayed. The commercial monoclonal antibody against the NADPHbinding region of murine iNOS (a highly conserved region among the three known mammalian NOSs) gave an immunoreactive band with the peroxisomal fraction. Similar results have been obtained in extracts from maize tissues using the same antibody (11). However, this region also shares a high degree of identity with NADPH-cytochrome P450 reductases (36). To rule out the possibility that the cross-reactivity observed is due to NADPH-cytochrome P450 reductases, a polyclonal antibody against 14 residues from the C terminus of iNOS which are absent in the NADPH-cytochrome P450 reductases, was used. This type of anti-peptide antibodies provide powerful tools to discriminate among proteins with a high degree of homology but with differences in other regions of the sequence (37). In this way, an immunoreactive band with a molecular mass of approximately 130 kDa was detected, which is in the same molecular mass range described for most iNOS (38,39) and eNOS (40,41) but different from the molecular mass reported for NADPH-cytochrome P450 reductases (70 kDa) (4,42). In maize seedlings extracts, an immunoreactive protein band of approximately 166 kDa was detected using a commercial monoclonal antibody against mouse macrophage iNOS (11).
The electron microscopy immunolocalization of NOS demonstrated the presence of the enzyme in the matrix of peroxisomes and also in chloroplasts. However, no immunogold labeling of NOS was detected in mitochondria. This contrasts with results obtained in mammalian tissues where a NOS was found in mitochondria isolated from rat liver (4) which was characterized as a new isoform (mtNOS) (43). The mtNOS from rat liver is similar to the inducible form (iNOS) but is constitutively expressed and bound to the mitochondrial membrane (4).
The presence of NOS in peroxisomes suggests an interaction of this enzyme with other components of the metabolism of activated oxygen species of leaf peroxisomes. An important point is that the NADP-dehydrogenases, recently found in the peroxisomal matrix (18), could provide the necessary NADPH for the perNOS reaction. But on the basis of the recent results of Sakuma and co-workers (44), a potential function of NO inside peroxisomes can be proposed. NO could react with O 2 .
radicals generated in the peroxisomal matrix by xanthine oxidase (19) to form the powerful oxidant peroxynitrite which, according to Sakuma et al. (44) could regulate the conversion of xanthine dehydrogenase into the superoxide-generating xanthine oxidase. On the other hand, NO could also diffuse through the peroxisomal membrane, reacting with O 2 . produced in the cytosolic side of the membrane by a small electron transport chain (22), thus generating the oxidant peroxynitrite in the cytosol. Likewise, the NO diffused into the cytosol could also function as a cellular transduction signal, as it has been described for H 2 O 2 (33)(34)(35).
In summary, the biochemical, immunological, and immunocytochemical data reported in this work provide evidence that peroxisomes contain nitric-oxide synthase activity (perNOS) which is calcium-dependent, constitutively expressed, and immunorelated with the mammalian iNOS. The occurrence in peroxisomes of a NOS suggests that these organelles are a cellular source of NO and implies new roles for peroxisomes in the cellular signal transduction mechanisms.