The modulation of mitochondrial nitric-oxide synthase activity in rat brain development.

Different mitochondrial nitric-oxide synthase (mtNOS) isoforms have been described in rat and mouse tissues, such as liver, thymus, skeletal muscle, and more recently, heart and brain. The modulation of these variants by thyroid status, hypoxia, or gene deficiency opens a broad spectrum of mtNOS-dependent tissue-specific functions. In this study, a new NOS variant is described in rat brain with an M(r) of 144 kDa and mainly localized in the inner mitochondrial membrane. During rat brain maturation, the expression and activity of mtNOS were maximal at the late embryonic stages and early postnatal days followed by a decreased expression in the adult stage (100 +/- 9 versus 19 +/- 2 pmol of [(3)H]citrulline/min/mg of protein, respectively). This temporal pattern was opposite to that of the cytosolic 157-kDa nNOS protein. Mitochondrial redox changes followed the variations in mtNOS activity: mtNOS-dependent production of hydrogen peroxide was maximal in newborns and decreased markedly in the adult stage, thus reflecting the production and utilization of mitochondrial matrix nitric oxide. Moreover, the activity of brain Mn-superoxide dismutase followed a developmental pattern similar to that of mtNOS. Cerebellar granular cells isolated from newborn rats and with high mtNOS activity exhibited maximal proliferation rates, which were decreased by modifying the levels of either hydrogen peroxide or nitric oxide. Altogether, these findings support the notion that a coordinated modulation of mtNOS and Mn-superoxide dismutase contributes to establish the rat brain redox status and participate in the normal physiology of brain development.

Different mitochondrial nitric-oxide synthase (mt-NOS) isoforms have been described in rat and mouse tissues, such as liver, thymus, skeletal muscle, and more recently, heart and brain. The modulation of these variants by thyroid status, hypoxia, or gene deficiency opens a broad spectrum of mtNOS-dependent tissuespecific functions. In this study, a new NOS variant is described in rat brain with an M r of 144 kDa and mainly localized in the inner mitochondrial membrane. During rat brain maturation, the expression and activity of mt-NOS were maximal at the late embryonic stages and early postnatal days followed by a decreased expression in the adult stage (100 ؎ 9 versus 19 ؎ 2 pmol of [ 3 H]citrulline/min/mg of protein, respectively). This temporal pattern was opposite to that of the cytosolic 157-kDa nNOS protein. Mitochondrial redox changes followed the variations in mtNOS activity: mtNOSdependent production of hydrogen peroxide was maximal in newborns and decreased markedly in the adult stage, thus reflecting the production and utilization of mitochondrial matrix nitric oxide. Moreover, the activity of brain Mn-superoxide dismutase followed a developmental pattern similar to that of mtNOS. Cerebellar granular cells isolated from newborn rats and with high mtNOS activity exhibited maximal proliferation rates, which were decreased by modifying the levels of either hydrogen peroxide or nitric oxide. Altogether, these findings support the notion that a coordinated modulation of mtNOS and Mn-superoxide dismutase contributes to establish the rat brain redox status and participate in the normal physiology of brain development.
In recent years, the occurrence of nitric-oxide synthase (NOS) 1

variants located in mitochondria (mtNOS) has been
reported: an iNOS protein was detected in rat liver and thymus and pig heart (1-3), eNOS was found in rat skeletal muscle and liver (4,5), and nNOS was described in rat skeletal muscle (6) and mouse heart (7). Thus, the mtNOS family appears to cover a broad spectrum of structurally and immunologically different proteins that could result from transcriptional or translational modifications that allow them to be targeted to mitochondria. However, differences between mtNOS and the classic NOS isoforms were reported: liver mtNOS exhibits an amino acid sequence similar to iNOS (8,9), albeit with a distinctive acylation pattern (8); mtNOS activity invariably depends on Ca 2ϩ and calmodulin, even when it may be immunologically related to iNOS (1,6); and a distinct kinetic pattern has been observed for mtNOS (1,8).
mtNOS seems to possess functions adapted to tissue-specific needs. In support of this notion, mtNOS activities in liver and skeletal muscle are modulated by the thyroid status (6), and those in brain and liver are modulated by hypoxia (5). Moreover, the level of mtNOS expression is modified by the activity or deficiency of specific genes, such as dystrophin in the heart (7). The regulation of mtNOS provides a new insight into the physiological significance of ⅐ NO and mitochondria in cell biology. It is well known that ⅐ NO binds to cytochrome oxidase with high affinity, modulates O 2 uptake (10,11), and increases the mitochondrial production of superoxide anion (O 2 . ), and depending on mitochondrial matrix Mn-superoxide dismutase levels, of hydrogen peroxide (H 2 O 2 ) (10,12). Accordingly, activation of mtNOS could be followed by an increase in the rate of mitochondrial H 2 O 2 production rate (13). H 2 O 2 , generated in this manner, is involved in the regulation of different cellular processes: in brain, it could participate in cell signaling networks during early synaptogenesis and plasticity (14); furthermore, following mtNOS activation, an exacerbated oxidant production may play a role in apoptotic signaling (15). Characteristically, nNOS and splice variants of the nNOS gene with distinctive subcellular localizations are differentially expressed during the embryonic life (16) and are related to the normal development of the brain (17). These studies were aimed at analyzing the occurrence of an NOS activity localized in brain mitochondria, its developmental profile, and its influence on redox metabolism. The physiological significance of these studies was strengthened by an experimental approach aimed at providing a relationship between mitochondrial production of ⅐ NO and H 2 O 2 and neuron proliferation. Biochemicals-3-Amino-1,2,4-triazole, aprotinin, Larginine, bovine serum albumin, calmodulin, catalase, CHAPS, Cu,Zn  superoxide dismutase, cytochrome c, dithiothreitol, EDTA, FAD, FMN,  glutathione, HEPES, horseradish peroxidase, hydrogen peroxide, KCN,  leupeptin, NADPH, N G -monomethyl-L-arginine (L-NMMA), ovomucoid, Percoll, phenylmethylsulfonyl fluoride, p-hydroxy-phenyl acetic acid, poly-D-lysine, sucrose, tetrahydrobiopterin, Tris, trypsin, xanthine, and xanthine oxidase were from Sigma. Mn(III) tetrakis (4-benzoic acid) porphyrin chloride (Mn(III) TBAP) was from Cayman Chemical (Ann Arbor, MI), and DNase was from Roche Molecular Biochemicals. N-Acetyl cysteine was a gift from Dr. Daniel Colombari (Poen Laboratories, Buenos Aires, Argentina). Monoclonal antibodies antineuronal NOS (1095-1289), polyclonal anti-endothelial NOS, and polyclonal anti-macrophage iNOS were from Transduction Laboratories (Lexington, KY); polyclonal anti-neuronal NOS (1409 -1429) and monoclonal anti-neuronal NOS (1-181) were from Sigma. L-[2,3,-3 H]Arginine was from PerkinElmer Life Sciences. Acrylamide solutions, polyvinylidene difluoride membranes, anti-mouse and anti-rabbit IgG antibodies, and Immun-Star kit were from Bio-Rad. Nitric oxide solutions (1.2-1.8 mM) were obtained by bubbling ⅐ NO gas to 99.9% purity (AGA Gas Inc., Maumee, OH) in water degassed with helium for 30 min at room temperature and stocked for a week at 4°C. Subcellular Fractionation-Adult Wistar rats were sacrificed by decapitation, and the brain or cerebellum was rapidly removed and homogenized in 320 mM sucrose/20 mM HEPES, pH 7.2, containing 1 mM EDTA, 1 mM dithiothreitol, 10 g/ml leupeptin, 2 g/ml aprotinin, and 10 g/ml phenylmethylsulfonyl fluoride. For developmental experiments, brain or cerebellum dissected from embryos (n ϭ 15-30) or neonate animals of different ages (n ϭ 5-10) was pooled. Mitochondria and synaptosomes were purified by Percoll gradient centrifugation as described (1) in a buffer consisting of 0.23 M mannitol, 0.07 M sucrose, 0.5 mM EGTA, and 2 mM HEPES, pH 7.4. The cytosolic fraction was obtained after 100,000 ϫ g centrifugation of the 8,000 ϫ g supernatant of brain homogenates. Except for experiments in the presence of proteinase K, which required intact mitochondria, and the isolation of submitochondrial particles, all determinations were performed utilizing once frozen/thawed mitochondria.

Chemicals and
Submitochondrial Fractionation-Percoll-purified mitochondria were frozen and thawed three times and then sonicated twice at 40 watts with a Cole-Parmer sonicator (WPI, Sarasota, FL). Subsequently, samples were centrifuged for 10 min at 8,000 ϫ g to precipitate unbroken mitochondria, and the supernatant was then centrifuged for 30 min at 100,000 ϫ g. The supernatant (soluble fraction) was separated from the pellet (membranes), and the latter was suspended in 0.25 M sucrose, 10 mM HEPES, pH 7.4, containing 2 mM dithiothreitol and 20 mM CHAPS and incubated for 20 min at 4°C under gentle shaking. The preparation was then centrifuged for 30 min at 100,000 ϫ g, and the supernatant (CHAPS-soluble fraction) and pellet (CHAPS-insoluble fraction) were separated and stored at Ϫ70°C until used.
Nitric-oxide Synthase Activity-⅐ NO production was assessed by the conversion of [ 3 H]L-arginine to [ 3 H] L-citrulline as described previously (18)  Immuno-electron Microscopy-Purified mitochondria were suspended in 4% formaldehyde freshly prepared from paraformaldehyde in PBS, pH 7.4, for 2 h at 4°C and then dehydrated in 70, 96, and 100% ethanol (30 min for each step) and embedded in LR White. Slides were obtained with glass knives and placed on nickel grids over Formvar membranes. Immunocytochemistry was performed using a primary rabbit anti C-terminal nNOS (1409 -1429) at a dilution of 1:100 in PBS, pH 7.4. After washing with PBS, a goat anti-rabbit serum conjugated with colloidal gold (10-nm diameter particles) was used at a dilution of 1:100 in PBS, pH 7.4. Grids were washed again in PBS and slightly counterstained with 1% uranyl acetate in water for 5 min. Nonspecific background was blocked by incubating the grids with 5% normal goat serum in PBS at the beginning of the procedure. Specimens were observed in a Zeiss EM-109-T transmission electron microscope at 80 kV.
Mitochondrial Production of H 2 O 2 -H 2 O 2 production was continuously monitored in a Hitachi F-2000 spectrofluorometer (Hitachi Ltd., Tokyo, Japan) with excitation and emission wavelengths at 315 and 425 nm, respectively (19). The assay medium consisted of 25 mM sucrose, 5 mM MgCl 2 , 20 mM KCl, 10 mM PO 4 K 2 , and 20 mM HEPES, pH 7.4, and it was supplemented with 12.5 units/ml horseradish peroxidase, 50 M p-hydroxyphenylacetic acid, and 0.1 mg of mitochondrial protein/ml with 6 mM succinate as substrate. To explore the effects of ⅐ NO and mtNOS activity on mitochondrial H 2 O 2 production rate, the assay was initiated by supplementation with either 0.1 mM L-arginine or a pulse of ⅐ NO or 2 M antimycin A. To assess specific mtNOS-dependent H 2 O 2 production rates, 1 mM L-NMMA was added to the mitochondrial preparations in the appropriate cases. To restrict oxidative decay of intramitochondrial ⅐ NO, assays were carried out at about 70 M O 2 . All fluorometric variations in the different conditions were attributable to H 2 O 2 as they were completely inhibited by supplementation with 3 M catalase. To make the maximal H 2 O 2 production rate uniform, mitochondrial preparations were supplemented with 1 mM of the superoxide dismutase mimetic Mn(III) TBAP.
Mitochondrial Complex Activities and Ubiquinol Content Determination-NADH-cytochrome c reductase and succinate-cytochrome c reductase activities (complexes I-III and complexes II and III, respectively) were assayed spectrophotometrically by determining cytochrome c reduction at 550 nm in the presence of 30 M cytochrome c, 1 mM KCN, and either 150 M NADH or 8 mM succinate as electron donors. Cytochrome oxidase activity (complex IV) was determined by recording the oxidation of 50 M reduced cytochrome c in a Hitachi 3000 spectrophotometer at 550 nm; ⑀ 550 ϭ 21 mM Ϫ1 cm Ϫ1 (20). The rate of the reaction was determined as the pseudo-first order reaction (kЈ) constant and expressed as min Ϫ1 mg of protein Ϫ1 . Ubiquinone content in isolated mitochondria was determined by high pressure liquid chromatography (Waters) with UV detection at 275 nm after extraction with cyclohexane:ethano1 (5:2) (6 mg of mitochondrial protein/ml and 7 ml of cyclohexane:ethanol) (21).
Superoxide Dismutase Activity-Mn-and Cu,Zn-superoxide dismutase activities were determined spectrophotometrically by measuring the inhibition of cytochrome c reduction by the xanthine/xanthine oxidase system, as described previously (22). Briefly, the reduction rate of 10 M cytochrome c by 3.5 microunits/ml xanthine oxidase and 50 M xanthine was followed in the absence (basal) or presence of 100 g of cytosolic or mitochondrial proteins. Both cytochrome oxidase and Cu,Zn-superoxide dismutase were inhibited by 1 mM KCN. Superoxide dismutase units in each fraction were calculated by extrapolation from a calibration curve with titrated commercial Cu,Zn-superoxide dismutase.
Western Blotting-Aliquots of 100 g of protein were separated by electrophoresis on reducing 6% SDS-polyacrylamide gels and transferred to polyvinylidene difluoride membranes. Membranes were washed with 0.1% Tween 20 in 20 mM Tris-buffered saline solution, pH 7.4, and blocked with 4% non-fat milk. Membranes were incubated with the first antibody (1 h, 1:1000 dilution), washed, and subsequently incubated with a secondary goat anti-mouse or anti-rabbit IgG antibody conjugated to alkaline phosphatase (1 h, 1:3000 dilution). Bands were detected by chemiluminescence using Immun-Star (Bio-Rad). Quantification of bands was performed by digital image analysis using a Hewlett-Packard scanner and Totallab analyzer software (Nonlinear Dynamics, Biodynamics).
Cell Cultures-Cerebellar granule cell cultures were obtained as described previously (23). Five-day-old Sprague-Dawley rats were decapitated, and cerebella were dissected in Krebs-Ringer solution supplemented with 6 g/l glucose. Meninges were eliminated, and tissue was cut into 1-mm pieces and incubated in saline containing 0.025% trypsin, 1.2 mM MgSO 4 , and 3 mg/ml bovine serum albumin for 15 min at 37°C with continuous agitation. The enzymatic digestion was stopped with ovomucoid (trypsin inhibitor), and the tissue was mechanically dissociated with Pasteur pipettes of two different diameters (25 strokes) in saline-containing ovomucoid and 0.01% DNase. The cell suspension was centrifuged at 150 ϫ g for 10 min, and the pellet was resuspended in cell culture medium (Neurobasal, Invitrogen). Cells were inoculated onto 96 multiwells (300,000 cells/well) coated with poly-D-lysine (molecular weight 300,000) in serum-free medium supplemented with B27 (Invitrogen).
[ 3 H]-Thymidine Assay-Cells were seeded on 96 multiwells, and after 2 h, they were incubated with [ 3 H]thymidine (1 mCi/ml) for 22 h. Drug treatments were performed simultaneously with the addition of [ 3 H]thymidine. The cells were osmotically disrupted through a 2-min incubation with bidistilled water and harvested with a Nunc cell harvester coupled to glass fiber filters (Whatman, GF/C) as described previously (23). Filters were washed seven times with bidistilled water and allowed to dry overnight. Radioactivity was counted in a liquid scintillation ␤-counter. [ 3 H]Thymidine incorporation was expressed as the ratio (experimental Ϫ control)/control.
Statistics-Student's t test and linear regression analysis were used as appropriate; statistical significance was accepted at p Ͻ 0.05.

Nitric-oxide Synthase Activity and Its Localization in Brain
Mitochondria-NOS activity was assessed in rat brain cytosol, synaptosomes, and purified mitochondria (Table I). In samples obtained from adult rats, the activity present in the mitochondrial fraction represented about 10% of that of cytosolic nNOS. The purification of mitochondria from organelles with similar sedimentation properties was confirmed by measuring the ratio between succinate-cytochrome c reductase (mitochondrial marker) and acidic phosphatase (lysosomal marker) activities. The mitochondrial preparations used throughout this study were enriched about 30-fold in mitochondrial markers. In addition, we confirmed the purity and the integrity of mitochondria by electron microscopy.
NOS activity was not affected by pretreatment of intact mitochondria with proteinase K, suggesting that it was not a product of cytosolic contamination and that the protein was localized inside the organelles. Accordingly, after fractionation of submitochondrial particles, NOS activity was enriched by 18-fold in the membrane fraction, which was mainly composed of inner mitochondrial membrane (CHAPS-insoluble fraction) ( Table I).
Brain mtNOS activity was dependent on tetrahydrobiopterin, Ca 2ϩ /calmodulin, and NADPH, and to a lesser extent, on flavins ( Fig. 1), as reported previously for nNOS. The mitochondrial NOS activity that represented the N G -monomethyl-L-arginine-sensitive activity was also inhibited by the NOS inhibitors NG-nitro-L-arginine, 7-nitroindazol, and N-iminoethyl-L-ornithine and by the flavoprotein inhibitor, diphenylene iodonium. Preliminary kinetics studies yielded a similar apparent K m for mtNOS and nNOS (2-12 M).
Ultrastructural Mitochondrial Nitric-oxide Synthase Detection-The specific detection of mtNOS by anti-nNOS 1409 -1429 was used to assess by means of immuno-electron microscopy the localization of mtNOS at the ultrastructural level. In these conditions, colloidal gold particles were clearly detected in mitochondria (Fig. 3A). Moreover, the background levels were low, and samples processed with normal rabbit serum as a primary antibody did not show any labeling, thus confirming the specificity of the reaction. A more detailed study of the ultrastructural localization of the enzyme in mitochondria using a higher dilution of the first serum showed that labeling was mainly located in the inner face of the mitochondrial inner membrane, and to a lesser extent, in the mitochondrial matrix. (Fig. 3B).
Developmental Modulation of Rat Brain Mitochondrial Nitric-oxide Synthase-The up-regulation of a 144-kDa nNOS variant in rat synaptosomes was reported to occur early after birth (17). In this study, the expression (Fig. 4) and activity (Fig. 5) of mtNOS in the brain of embryos and neonate animals in parallel to cytosolic nNOS were addressed as follows. (a) Cytosolic nNOS activity and expression were faintly detectable by embryonic day 15 and 19 and increased after birth to peak around the second and third postnatal weeks (Fig. 4A) (as reported previously (17)). (b) mtNOS activity showed a different developmental pattern: its expression and activity were elevated throughout the late embryonic period ( Fig. 4B; E15-E19) up to the first postnatal week ( Fig. 4B; P0-P6) and markedly decreased beyond this age. Thus, at P0, mtNOS protein expression and activity were about 6-fold higher than the corresponding adult values. The modulation of mtNOS activity was attributed to changes in protein expression, as inferred from the respective ratios (Fig. 5). Furthermore, mt-NOS followed a similar developmental pattern in cerebellum, although elevated activity persisted until postnatal day 15 (Fig. 5, inset). (c) Studies on the synaptosomal fraction permitted us to observe the co-modulation of both mtNOS and nNOS: in this preparation, mtNOS from synaptic mitochondria was apparently unique during the first 4 days; beyond this time FIG. 3. Immuno-electron microscopy of purified brain mitochondria. As shown in A, the organelles were first labeled with anti nNOS 1409 -1429 antibody followed by treatment with a goat antirabbit serum conjugated with colloidal gold (10-nm diameter particles); high intramitochondrial positive staining was observed. As shown in B, the use of a higher dilution of the first antibody allowed a more detailed localization of the gold particles, which are localized in the inner mitochondrial membrane (arrows). Slides were slightly counterstained with uranyl acetate. Magnification ϭ ϫ50,000. The numbers next to the E or P denote the day at the embryonic or postnatal stage, respectively. On the right-hand side, densitometries of cytosolic and mitochondrial protein bands with respect to nNOS protein expression in adult rats (P90 ϭ 1) are shown. Data are in arbitrary units (AU). point, both enzymes could be detected (Fig. 4C).
The Nitric Oxide-dependent Production of H 2 O 2 in Brain Mitochondria-In different rat tissues, ⅐ NO induces an increase of mitochondrial O 2 . and H 2 O 2 production rates by mechanisms involving inhibition of electron transfer at the bc 1 segment (10) and oxidation of membrane-bound ubiquinol (Reaction 1) (12, 13, 24) followed by ubisemiquinone autoxidation (Reaction 2).

REACTION 2
The production of H 2 O 2 by both P2-4 neonatal and adult brain mitochondria showed a biphasic response to ⅐ NO (Fig. 6) Comparable maximal H 2 O 2 production rates (0.4 -0.5 nmol/ min/mg of protein) were achieved at 0.1 M ⅐ NO in adults and at 0.2 M ⅐ NO in neonates (Fig. 6). In addition, total ubiquinone was ϳ20% lower in mitochondria from neonates than in those from adults. The ubiquinone-reducing activities of complexes I-III and complexes II-III as well as the activity of complex IV FIG. 5. Nitric-oxide synthase-specific activities during rat brain development. The nitric-oxide synthase-specific activities during development were measured in the brain cytosolic and mitochondrial fractions described in the legend for Fig. 4 and according to "Materials and Methods." Inset, mitochondrial nitric-oxide synthase activity during postnatal maturation of rat cerebellum. The arrows indicate the moment of birth.
FIG. 6. Effects of ⅐ NO on the H 2 O 2 production rate of rat brain mitochondria. The reaction mixture contained 0.1 mg of mitochondrial protein/ml from either P2-4 neonates (E) or adults (q) in buffer as described under "Materials and Methods." The assay temperature was 30°C. Data represent the mean of three experiments.
were ϳ45-50% lower in neonates than those in adult animals (Table II). These data may explain the 2-fold higher ⅐ NO concentration required in neonatal mitochondria to elicit an H 2 O 2 production rate similar to that in adult mitochondria.
The mtNOS-dependent Production of H 2 O 2 during Development-Supplementation of brain mitochondria from P2-4 animals with L-arginine, the universal substrate of NOS, elicited a selective and marked increase in the rate of H 2 O 2 production (Fig. 7). This yield was comparable with the increase in H 2 O 2 generation obtained by supplementation with exogenous ⅐ NO (Fig. 6) or antimycin A (Fig. 7), a complex III inhibitor that stimulates mitochondrial production of O 2 . and H 2 O 2 . The Larginine-mediated increase was prevented (ϳ80% lower value) by a molar excess of the NOS inhibitor L-NMMA, thus indicating that the activation of mtNOS was necessary for mediating the effect of L-arginine. Conversely, H 2 O 2 production by mitochondria isolated from P15 or adult animals was not significantly affected by L-arginine supplementation, although these mitochondria retained the response to antimycin A (Fig. 7). Considering the temporal expression (Fig. 4) and activity (Fig.  5) of mtNOS, it may be surmised that (a) mitochondrial H 2 O 2 production parallels the developmental modulation of mtNOS and (b) H 2 O 2 production by mitochondria from adult animals is more susceptible to ⅐ NO regulation, but the limited expression of mtNOS renders H 2 O 2 production by these mitochondria less sensitive to L-arginine activation. . , a study on superoxide dismutase activity through the period of high mtNOS expression was undertaken. Remarkably, the developmental time course of Mn-superoxide dismutase activity (the mitochondrial isoform) (Fig.8A) paralleled that of mtNOS (Fig. 5), whereas cytosolic Cu,Zn-superoxide dismutase activity increased with age to reach constant levels (Fig. 8A).

Co-modulation of mtNOS and Mn-Superoxide
The activities of Mn-and Cu,Zn-superoxide dismutases correlated with the respective variations in the activities of mt-NOS (Fig. 8B) and nNOS (Fig. 8C), respectively. Intramitochondrial formation of ONOO Ϫ was not detected during the studied period, as inferred by the absence of protein tyrosine nitration (data not shown). Considering the two alternative reactions of O 2 . utilization (Reactions 3 and 4 above) (25), these results suggest that the 2-fold increased Mn-superoxide dismutase activity in P2-4 (with respect to that in adults) (Fig.  8A) would support the decay of O 2 . to H 2 O 2 rather than to ONOO Ϫ , thus favoring the net flux of H 2 O 2 into cytosol.

Role of H 2 O 2 in Granular Cerebellar
Neuroblast Proliferation-At variance with brain neurons, which exhibit a low proliferation rate after birth, cerebellar granular cells from cerebellum proliferate during a relatively prolonged postnatal period. Therefore, these cells represent a suitable model to study the effects of mtNOS expression and mitochondrial H 2 O 2 in developing cerebellum. At P5, cerebellar tissue exhibited maximal expression and activity of mtNOS, which subsequently declined in adult life (P90) (Fig. 9A). Accordingly, the L-arginine-dependent mitochondrial production rate of H 2 O 2 was 3-fold higher in neonates at P5 than in adult rats at P90.  The steady-state levels of H 2 O 2 in proliferating cerebellum from neonates were 4-fold higher than those in the quiescent mature cerebellum of adult animals. These results suggest that the differences in neonatal and adult mitochondria are of physiological significance in vivo.
In support of this hypothesis, changes in the H 2 O 2 steadystate levels in immature granular cerebellar neurons isolated from P5 rats significantly affected proliferation rate in terms of [ 3 H]thymidine incorporation (Fig. 9B). In these experimental conditions, primary granular cells grow in culture by 24 h in an enriched medium with Dulbecco's modified Eagle's medium/B27 and in the absence of serum supplementation or growth factors (23). Cells are rounded, and few of them express tetanus toxin fragment C, a specific marker of postmitotic neuronal phenotype (23). In agreement with the different redox states of neonates and adults, the magnitude of spontaneous proliferation of the primary cerebellar neurons at P5 was maximal.

DISCUSSION
Submitochondrial fractionation and immuno-electron microscopy analysis indicated that the nitric-oxide synthase variant in rat brain mitochondria described here had a predominant localization at the inner mitochondrial membrane. Differences between brain mtNOS and classic nNOS suggest a variation in the N-terminal domain of the protein (Fig. 2). Moreover, apparent molecular weight, immunological detection, and developmental modulation resemble those of the pre-viously reported mouse synaptosomal 144-kDa nNOS (17). The two variants share a similarly reduced catalytic activity respect to the classic 157-kDa nNOS, although specific activity, Ca 2ϩ dependence, and requirements are similar to all other mitochondrial isoforms.
Taking into consideration the immunological characterization shown in Fig. 2, brain mtNOS appears to be an entity different from rat or mouse liver mtNOS (1,5). This discrepancy suggests the occurrence of tissue-and species-specific different mitochondrial variants. In this regard, mitochondrial variants of iNOS-like (3) and nNOS-like (7) isoforms in pig and rat heart organelles were described. Moreover, several years ago, an eNOS-like mtNOS in the rat brain was reported (27).
Immunological and functional observations are indicative of a developmental modulation of the brain mtNOS variant. This notion is supported by data showing that brain mtNOS is highly expressed and active in the late stages of fetal development and during the first postnatal days followed by a decreased expression in the adult brain. In agreement with previous reports, cytosolic nNOS was poorly detected in embryos or immediately after birth, and its expression increased sharply after postnatal day P6 (17).
The occurrence of mtNOS is consistent with a fine modulation of critical mitochondrial functions by ⅐ NO. In the brain, ⅐ NO inhibits respiration (11), promotes de-energization of synaptic mitochondria (28), and induces Ca 2ϩ release from the mitochondrial matrix (29). Furthermore, ⅐ NO and reactive oxygen species participate in significant neuronal events, such as dendritic growth and arborization (30) and apoptosis (31).
Most of mitochondrial O 2 . and H 2 O 2 production is a consequence of reactions of ⅐ NO with components of the respiratory chain (10,12). Accordingly, L-arginine is known to stimulate H 2 O 2 release through the activation of mtNOS (13), an observation extended to brain mitochondria (Fig. 7) and dependent on the mtNOS expression level. This study established a link between ⅐ NO metabolism and the generation of reactive oxygen species by mitochondria during development: mitochondria with different mtNOS content, such as those from newborn P2-4 and adults, yielded similar H 2 O 2 production rates in the presence of ⅐ NO (Fig. 6). Nevertheless, only neonatal organelles containing significant amounts of mtNOS responded to L-arginine stimulation (Fig. 7), thus leading to a temporal increase in H 2 O 2 production. Thus, the similar temporal pattern of brain mtNOS (Fig. 5) and Mn-superoxide dismutase (Fig. 8A) (32,33). A similar correlation between Cu,Zn superoxide dismutase and cytosolic nNOS is consistent with previous reports on protection against ⅐ NO toxicity in neurons by superoxide dismutase (34). Considering that at maximal mtNOS expression, the formation of ONOO Ϫ was not detected, this study supports the following notions: (a) mitochondrial production of ⅐ NO is an important factor in the developmental setup of brain H 2 O 2 steady-state concentration in parallel with Mn-superoxide dismutase activity and (b) as a corollary, the accurate temporal up-regulation of mtNOS may represent a redox signaling mechanism rather than a stressful event. This view is supported by recent observations that propose that ⅐ NO and H 2 O 2 participate in the development and maturation of the nervous system, particularly through potentiation of Ca 2ϩ signaling by redox changes (31). Moreover, ⅐ NO and H 2 O 2 may promote cell cycle arrest and quiescence in different tissues by mechanisms such as activation of p38 MAPK (35), inhibition of activity or expression of cyclin D1, FIG. 9. H 2 O 2 levels and cerebellar neuron proliferation. As shown in A, the expression and activities of mtNOS and their respective mitochondrial H 2 O 2 yields were compared in organelles from quiescent adult and proliferating P5 newborn rat cerebellum as described under "Materials and Methods." The contribution of mtNOS to cytosol H 2 O 2 steady-state concentration ([H 2 O 2 ] ss ) was estimated from the L-Arg-dependent mitochondrial H 2 O 2 production rate and the specific activities of cytosolic catalase and glutathione peroxidase (Equation 1 under "Results"). As shown in B, immature granular cerebellar cells were isolated from P5 rats. Two h after seeding in Dulbecco's modified Eagle's medium Neurobasal B27 without serum supplementation, increase of cyclin D1 ubiquitination (36), activation of phosphatases, dephosphorylation of Rb protein, and activation of p21 and p53 proteins (37). Some of these pathways had been confirmed in nervous tissue (14,38).
From this perspective, the sequential activation of mitochondrial and cytosolic isoforms of nNOS and superoxide dismutase in the brain might play a role in synaptogenesis and synaptic remodeling that follows proliferation arrest (39). The temporal correlation of mtNOS up-regulation with neuronal network remodeling is consistent with this perspective. In support of this notion, the persistence of mtNOS activity until postnatal day 15 in the cerebellum correlates with the longer period of proliferation and plasticity of this brain region. In this period, mtNOS was significantly expressed and 8-fold more active in immature cerebellum at P5 than in the quiescent adult organ at P90; this effect resulted in increased cerebellar mitochondrial H 2 O 2 production rate and estimated steady-state concentration in the newborns. The biological significance of these findings is emphasized considering that, at this precise H 2 O 2 concentration, proliferation of immature cerebellar granular cells isolated from P5 was maximal and that, moreover, it was clearly down-modulated by finely setting down H 2 O 2 or ⅐ NO concentrations (Fig. 9). Taken together, these effects suggest that maximal neuronal proliferation depends on a narrow range of H 2 O 2 concentrations, which is likely to be dependent on mtNOS activity.
The modulation of neuronal-like mtNOS variants is consistent with the diversity of nNOS gene expression (40) adapted to perform tissue-specific functions. For example, brain mtNOS regulates mitochondrial free radical production; cardiac mt-NOS modulates mitochondrial parameters in a beat-to-beat fashion, and its increased expression in muscular dystrophy in mouse models contributes to generate cardiac damage (7); and rat skeletal muscle neuronal mtNOS activity varies with the thyroid status, which has a central role in development (6). Likewise, the fine modulation of brain 144-kDa mtNOS and mitochondrial reactive oxygen species during the perinatal period suggests an essential role of ⅐ NO in the chronological phases of brain maturation and synaptic plasticity.