Dynamics of protein nitration in cells and mitochondria

Kulwant S. Aulak, Thomas Koeck, John W. Crabb, Dennis J. Stuehr


Nitric oxide is a precursor of reactive nitrating species such as peroxynitrite and nitrogen dioxide that modify proteins to generate 3-nitrotyrosine. Many diseases are associated with increased levels of protein-bound nitrotyrosine, and this is used as a marker for oxidative damage. However, the regulation of protein nitration and its role in cell function are unclear. We demonstrate that biological protein nitration can be a specific and dynamic process. Proteins were nitrated in distinct temporal patterns in cells undergoing inflammatory activation, and protein denitration and renitration occurred rapidly in respiring mitochondria. The targets of protein nitration varied over time, which may reflect their sensitivity to nitration, expression pattern, or turnover. The dynamic nature of the nitration process was revealed by denitration and renitration of proteins occurring within minutes in mitochondria that were subject to hypoxiaanoxia and reoxygenation. Our results have implications that are particularly important for ischemia-reperfusion injury.

  • nitrotyrosine
  • peroxynitrite
  • inflammation
  • ischemia reperfusion

nitric oxide (NO) is a relatively stable free radical that diffuses from the site of production, crossing cell membranes and interacting with targets without the need for special transporters or receptors. NO biosynthesis is involved in a growing number of biological functions (13, 15, 27, 44). NO is generated from three isoforms of nitric oxide synthase (NOS): neuronal NOS (nNOS), inducible (iNOS), and endothelial NOS (eNOS) (2, 54). Multiple splice variants of these proteins and multiple posttranslational modifications exist, and one such variant is believed to generate the mitochondrial-specific NOS (mtNOS) (20, 21, 60). The physiology of NO is complex and its role in disease is controversial with both beneficial and detrimental effects. NO may have direct effects such as signaling by interacting with soluble guanylate cyclase leading to vasodilation and causing alterations in gene expression and also may interact with other molecules to generate more reactive species (8, 28, 31, 64). One such reactivity is to generate potent nitrating species that result in the formation of 3-nitrotyrosine in proteins. The mechanism, regulation, and role of protein tyrosine nitration are controversial (10, 31).

Many mechanisms of tyrosine nitration have been proposed; however, the two that are widely believed to exist in vivo involve formation of peroxynitrite or mediation via hemeperoxidases (14). Both involve the use of NO or its by-products reacting with reactive oxygen species (ROS). Superoxide is a toxic anion that is generated in cells. It reacts with NO with a near diffusion-limited reaction rate and is much faster than its detoxification reaction with superoxide dismutase (SOD) (9). The product of superoxide and NO is peroxynitrite, which is a potent protein nitrating agent especially after further reacting with carbon dioxide (47). The level of superoxide in cells is kept low by antioxidants. However, certain states can lead to its increased production. Excessive generation of superoxide may be caused either by the overstimulation of tightly regulated processes such as NAD(P)H oxidases and NOSs or by processes that produce ROSs, such as the electron transport chain (ETC) in mitochondria or by xanthine oxidase (10, 31, 63). Peroxynitrite is highly reactive with a very short half-life, and therefore it would interact with proteins near the site of generation. A second mechanism of tyrosine nitration is mediated via hemeproteins. Nitrite, a breakdown product of NO, is used as a substrate and along with hydrogen peroxide to generate nitrogen dioxide. Nitrogen dioxide is also highly unstable and reacts close to the site of its generation. Therefore, locating the sites of nitration in a cell or tissue may help to indicate the mechanism.

The role protein nitration plays in cell physiology is unclear. This is complicated by the uncertain fate of the nitrated protein. Modified proteins are believed to be either degraded or subject to processes that could lead to enzymatic “denitration” (24, 30, 33). The latter possibility is intriguing because this would allow the process of tyrosine nitration to be reversible and thus enable a more dynamic physiological role. Protein nitration is observed under normal conditions in all tissues. Western blots using anti-nitrotyrosine antibodies demonstrate that multiple proteins are modified under normal conditions. Our understanding of protein nitration may be better defined under disease conditions. An increase in tyrosine nitration is observed in numerous vascular and neurological diseases (31). Protein nitration has been mainly considered as a disease marker (18, 25, 50) and has recently been suggested to be the best indicator for cardiovascular disease (52). Many diseases in which tyrosine nitration is increased have an inflammatory component that leads to the excessive or inappropriate generation of NO as well as ROS. Recent data have even suggested that protein nitration may be involved in the disease process itself. For example, Aslan et al. (5) studied intermittent vascular occlusions that lead to ischemia-reperfusion injury in sickle cell disease. This injury activated inflammatory processes, including the induction of iNOS and enhanced production of ROS, and resulted in nitration of many proteins including actin. Tyrosine nitration of actin caused a disorganization of the actin fibers that altered cytoskeleton dynamics and led to cell death.

To determine the role that protein nitration may play in cell physiology, it is necessary to identify the modified proteins in a comprehensive manner. We were the first to use proteomics to determine tyrosine-nitrated proteins and identified over forty proteins that were potentially modified in vivo (6, 42). That initial work allowed a platform to examine mechanistic and biological aspects of protein nitration. In this study we report on the dynamics of protein nitration. We show that proteins are nitrated in distinct temporal patterns in cells undergoing inflammatory activation and show that protein denitration and renitration can rapidly occur in respiring mitochondria. Our results have implications that may be particularly important for ischemia and reperfusion injury.


A549 cell extract. A549 cells were grown to 70% confluency in F12R medium, and fresh medium was added either alone (unstimulated) or with medium in the presence of cytokines (U/ml: γ-interferon 1,000; tumor necrosis factor-α 10; interleukin 1-β 10; BioSource). After 72 h, the medium was collected, and the cells were washed three times in phosphate-buffered saline, scraped into Eppendorf tubes, and spun down. For samples run on two-dimensional gels, a lysis buffer [8 M Urea, 4% CHAPS, 1% DTT; 2% IPG-ampholytes (pH 3–10)] was then added to the cell pellets, and the samples were isoelectric focused on a nonlinear pH 3–10 immobilized pH gradient strip.

Nitrite-nitrate determination. Nitrite-nitrate was determined using a modification of the procedure of Schmidt (48). Cell tissue culture medium (150 μl) was used for all assays. Nitrate was first reduced to nitrite using 0.1 U/ml nitrate reductase (from Aspergillus species; Boehringer Mannheim), 50 μM NADPH, and 5 μM FAD to a final volume of 160 μl, at 37°C for 2 h. Excess NADPH was then removed by the addition of 10 U/ml lactate dehydrogenase and 10 mM pyruvate and incubation at 37°C for 30 min. Total nitrite was then determined using the Greiss reagent. A standard curve using nitrate was used to determine the concentration of nitrate-nitrite produced. Nitrate-nitrite levels are determined by a minimum of three independent determinations.

Two-dimensional gel electrophoresis. Two-dimensional gel electrophoresis was performed with the IEF system (Bio-Rad; Hercules, CA). The first dimension used 7-cm nonlinear (pH 3–10), immobilized pH gradient (IPG) strips loading 135 μg protein. IPG strips were rehydrated with the sample at 50 V/14 h, and then isoelectric focusing was performed by a linear increase to 250 V over 30 min followed by a linear increase to 8,000 V over 60 min and then held at 8,000 V until a total of 15 kVh is reached. For the second dimension, the IPG strips were equilibrated for 12 min in 50 mM Tris·HCl (pH 8.8), 6 M urea, 30% glycerol, 2% SDS, 1% DTT, and bromophenol blue and then 15 min in 50 mM Tris·HCl (pH 8.8), 6 M urea, 30% glycerol, 2% SDS, 2% iodoacetamide, and bromophenol blue. The strips then were embedded in 1% (wt/vol) agarose on the top of 12% acrylamide gels containing 4% stacking gel. The second dimension SDS-PAGE was performed essentially according to Laemmli (37). After completion of the run, the acrylamide gels were soaked in transfer buffer (20 mM Tris·HCl, 96 mM glycine, and 20% methanol) and then partially transferred onto a polyvinylidene difluoride (PVDF) membrane by using a semidry transfer apparatus as suggested by the manufacturer. The gels were then stained with colloidal Coomassie blue (GelCode blue stain), and Western blot analysis was performed on the PVDF membrane.

Western blot analysis. PVDF membranes were blocked for 90 min by using a blocking buffer [20 mM Tris, 150 mM NaCl (pH 7.5), 0.2% Tween-20, and 1.5% BSA]. Membranes then were probed overnight at 4°C with a monoclonal antibody against nitrotyrosine (1:5,000; clone 1A6, Upstate Biotechnology; Lake Placid, NY) in blocking buffer. The membranes then were washed four times in washing buffer [20 mM Tris, 150 mM NaCl (pH 7.5), and 0.2% Tween-20] for 30 min each wash. The membranes were then probed with a goat anti-mouse antibody (horseradish peroxidase conjugate, 1:8,000, Sigma). After the membranes were washed four times in a washing buffer, the immunopositive spots were visualized by using ECL-Plus (Amersham Biosciences, Little Chalfont, Buckinghamshire, UK) as directed by the manufacturer. In experiments where samples needed to be compared, the membranes were placed on the same film and exposed simultaneously.

Protein identification. Proteins were identified by matrix-assisted laser desorption ionization-time of flight mass spectrometric analysis of in-gel tryptic digest of immunopositive spots (11). The two-dimensional gel spots were excised and cut into ∼1-mm3 cubes, and Coomassie blue was washed away with 250-μl aliquots of 60% acetonitrile (1× 30 min), 50% acetonitrile in 50 mM ammonium bicarbonate (1× 30 min), and 50% acetonitrile in 15 mM N-ethylmorpholin e(3× 15 min). After being destained, the gel pieces were dried in a Speed Vac, rehydrated in 15 μl 15 mM N-ethylmorpholine containing 0.1 μg modified trypsin (Promega), and incubated overnight at 37°C. The in-gel tryptic digest was extracted with 60% acetonitrile containing 0.1% trifluoroacetic acid. The extract was either used directly or extracted using a C18 Zip-Tip (Millipore), and the extract was subjected to MALDI-TOF mass spectrometric analysis by using a PE Biosystems Voyager DE Pro instrument equipped with a nitrogen laser (337 nm) and operated in the delayed extraction and reflector mode with a matrix of α-cyano-4-hydroxy-cinnamic acid (5 mg/ml in acetonitrile-water-3% trifluoroacetic acid, 5:4:1, vol/vol/vol). Internal standards were used for calibration and included two of the following three synthetic peptides G9I (MH+ 1015.579), K13L (MH+ 1543.859), and L20R (MH+ 2474.630). One microliter of sample was mixed with 1 μl of matrix and 0.5 μl of internal standard mix, and then 1.5 μl was applied to the sample plate and allowed to dry. Each spectrum was accumulated for ∼250 laser shots. Measured peptide masses were used to search the Swiss-Prot, translated EMBL database, and National Cancer Center for Biotechnology Information sequence databases for protein identifications. Each peptide map data set was searched using either MS-Fit (http://prospector.ucsf.edu/htmlucsf3.0/msfit.htm) or Mascot (http://www.matrixscience.com). All searches were performed with a mass tolerance of 0.005% error (50 ppm).

Isolation of mitochondria. All buffers and samples were kept on ice during the isolation. Rats were euthanized, and the livers were removed and weighed. The livers were then cut into small pieces and mixed with 9 volumes (wt/vol) of homogenization buffer (0.25 M sucrose in 10 mM HEPES, pH 7.5). The sample was then homogenized using a Teflon homogenizer. This was then centrifuged at 4500 g for 10 min at 4°C. The supernatant was then decanted into clean centrifuge tubes and recentrifuged for 30 min at 4°C at 16,000 g. The supernatant was then discarded and the pellet rinsed in homogenization buffer. The pellet was then resuspended in respiration buffer [(in mM) 250 sucrose, 20 d-glucose, 5 KH2PO4, 40 KCl, 0.5 EDTA, 3 MgCl2, 30 Tris (pH 7.4) plus protease inhibitors (μg/ml) 5 aprotinin, 1 leupeptin, 1 pepstain, and 24 pefabloc SC] and kept on ice until used. Mitochondria were stored for a maximum of 3 h before being discarded. The purity of the mitochondria was assessed using a small aliquot mixed with Janus Green B solution and viewing on a microscope. Protein concentrations were assessed using a modified Bradford regent and diluted at 20 mg/ml.

Respiratory activity of mitochondria. Oxygen levels in the mitochondrial samples were assessed using a YSI 5300A biological oxygen monitor (YSI, Yellow Springs, OH) connected to a computer through a RS-232 interface. Two Clark-type polarographic oxygen electrodes connected to two independent channels were used to record data from paired experiment simultaneously. Oxygen consumption rate (100% = 237 nmol O2/ml mitochondrial medium) is based on Campbell et al. (17). Briefly, 3 ml of a 3 mg/ml mitochondrial suspension in respiration buffer was placed in a stirred reaction chamber kept at 25°C (3 chamber isotemp circulator). After a 2-min incubation period to adjust to temperature, test reagents were added and then electrodes added at times indicated in the text. Respiration rates for states 2 and 3 were measured. To assess the respiratory control index for the mitochondria and so its integrity, the state 3-to-state 4 ratio was measured in the beginning of each experiment. After a period of hypoxia-anoxia, the electrodes were removed, and the sample was stirred under room air. The mitochondria were also subjected to a second phase of hypoxia-anoxia. At selected times, samples were removed from the chamber and immediately added to an equal volume of denaturing buffer (TM urea, 2 M thiourea, 1% Triton X-100, and 10 mM Tris·HCl; pH 7.5). These samples were then immediately frozen and stored at –20°C. After the sample was thawed at room temperature, they were sonicated once. For two-dimensional electrophoresis, 1% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 1% DTT, and 1% IPG-ampholytes (Bio-Lyte 3/10) were added. Samples were then used for the two-dimensional gel electrophoresis and Western blot analysis.


Protein nitration is a process that is found both in normal and diseased tissues. Here we show two models to illustrate the dynamics of protein nitration. The first involves using A549 lung adenocarcinoma cells, which undergo increased iNOS expression and protein nitration in response to cytokine stimulation (1). Because many of the targets we have identified are mitochondrial proteins (6), our second model illustrates how protein nitration in mitochondria may be altered in response to ischemic and reperfusion conditions.

Dynamics of protein nitration in A549 cells. A549 cells normally produce little NO in the basal state. On stimulation with tumor necrosis factor-α, interleukin 1-β, and interferon-γ, we detected increased NO breakdown products in the culture media (Fig. 1) as well as iNOS mRNA and protein within cells. Only a small amount of NO is produced over the first day, and the majority is produced during days 2 and 3. Over this time frame, protein nitration is also increased (Fig. 2). We have used two-dimensional gels to follow the nitration patterns in cells as a function of time after induction. Cellular proteins were determined to be nitrated by Western blotting with an anti-nitrotyrosine antibody and were then identified using MALDI-TOF mass spectrometry. Previously, we demonstrated that the immunoreactivity to the anti-nitrotyrosine antibody can be abolished in cells by adding either an iNOS inhibitor, the antioxidant ascorbate, or prereducing the membrane with dithionite, demonstrating specificity of the antibody (6, 42). The two-dimensional gels show that an increase in protein nitration occurred over time after induction. However, the nitration of proteins was not uniform. In A549 cells in the basal state, some protein nitration was apparent albeit to a smaller degree than after iNOS activity was induced. After 24 h, there was a group of proteins nitrated whose intensity either increased over time or remained constant. At the 24-h time point, only a modest NO generation occurred, yet many new proteins became nitrated. This suggests that these proteins are particularly sensitive to nitration and/or are close to the site of the nitrating source. Over the following days, intensities of these proteins increase along with new proteins being modified. This suggests that duration of oxidative stress may bring about cascading effects.

Fig. 1.

Production of nitrate/nitrate by stimulated A549 cells. A549 cells were stimulated with cytokines (1,000 U/ml γ-interferon; 10 U/ml tumor necrosis factor-α; and 10 U/ml interleukin-1-β) to induce inducible nitric oxide (NO) synthase (iNOS) expression for 0, 1, 2, or 3 days. At times indicated medium was collected and analyzed for total nitrate/nitrite.

Fig. 2.

Time course of tyrosine nitration in cytokine-stimulated A549 cells. A549 cells stimulated by cytokines were subjected to two-dimensional PAGE. A: Coomassie-stained gel of A549 cells stimulated for 3 days. Rectangle indicates the region shown in B of the Western blots probed with the monoclonal anti-nitrotyrosine antibody. B: Western blots from days 0, 1, 2, or 3 showing the immunopositive spots and identity of some of the proteins analyzed by matrix-assisted laser desorption ionization-time of flight mass spectrometry. 3α-HSD, 3α-hydroxysteriod dehydrogenase; MnSOD, manganese superoxide dismutase.

The proteins that were nitrated under basal conditions are actin, aldolase A, and glyceraldehyde 3-phosphate dehydrogenase. We and others have shown that these proteins are particularly sensitive to nitration. Actin was the most intense nitration signal in our basal A549 cell sample. This is consistent with the work by Aslan et al. (5), who reported that actin nitration occurred in normal mice and in a mouse model of sickle cell disease, with nitration being increased in the diseased animals. They suggested that the increased nitration may alter actin polymerization in cells. Regarding aldolase, we have found all isoforms of aldolase (A, B, and C) to be nitrated in various tissues under normal conditions (6, 42). For example, aldolase A and C are nitrated in the retina of rats that are kept in the dark with limited daily exposure to light. Interestingly, when these rats were exposed to light-induced retinal damage, nitration of aldolase proteins disappeared (42). Glyceraldehyde 3-phosphate dehydrogenase has previously been found to be highly sensitive to protein nitration. As little as 7 μM peroxynitrite caused 50% loss of activity of this enzyme (53). Therefore, our finding nitration of this protein in A549 cells under basal conditions is not surprising. A large increase in nitration under higher NO concentration does not occur and may suggest that it is denitrited or rapidly degraded (16).

Table 1 lists the identified proteins of which nitrotyrosine immunoreactivity increases with time. One protein that underwent early nitration is the mitochondrial protein manganese superoxide dismutase (MnSOD). This critical antioxidant enzyme detoxifies superoxide in the mitochondria. MnSOD is known to be rapidly induced and is nitrated under conditions of oxidative stress such as organ allograft rejection (40, 51). In vitro chemical nitration using peroxynitrite results in the modification of Tyr-34, which causes loss of activity. Early nitration of MnSOD suggests that mitochondria may be a prominent site for protein nitration. This is consistent with mitochondria being a major site of superoxide generation in cells, which should increase the likelihood of local peroxynitrite formation during NO synthesis.

View this table:
Table 1.

Anti-nitrotyrosine immunoreactive proteins identified in A549 cells

Two other proteins that appeared to be sensitive to early nitration are 3α-hydroxysteroid dehydrogenase (3α-HSD) and aldose reductase. Both belong to the aldo-keto reductase superfamily, which metabolizes a wide range of substrates and are potential drug targets (32). They all depend on nicotinamide cofactors for catalysis. 3α-HSD is a target for nonsteroidal anti-inflammatory drugs and may regulate levels of inflammatory prostaglandins. Aldose reductase is an enzyme that catalyzes the reduction of a wide variety of aromatic and aliphatic carbonyl compounds. It is the first enzyme in the polyol pathway. It has been implicated in the development of diabetes and galactosemic complications involving the lens, retina, nerves, and kidneys and is a major focus of drug development for therapeutic intervention (34, 36, 45). Because both of these aldo-keto reductases are nitrated in A549 cells, it would suggest that some key features may make them susceptible to protein nitration. The catalytic mechanism of aldo-keto reductases is conserved. A mechanism consistent with the crystal structure of 3α-HSD has been proposed (Fig. 3). All aldo-keto reductase structures retain the same spatial relationship between active site residues (Asp-50, Tyr-55, and His-117), the nicotinamide cofactor, and a water molecule. In this superfamily of proteins, Tyr-55 is absolutely conserved in all active members. The reductive reaction involves 4-pro-R hydride ion transfer from NAD(P)H to the substrate carbonyl and protonation of the oxygen by an enzyme residue acting as a general acid (46). The reverse process occurs in the oxidative reaction. Tyr-55 probably acts as the catalytic acid in the reaction mechanism (Fig. 3). Mutations of this reside lower enzyme activity. The electronic environment of the tyrosine may make it more susceptible to nitration. It is interesting to speculate that nitration of Tyr-55 may increase enzyme activity, because nitration of the tyrosine ring lowers the pKa of the hydroxyl group from 10 to 7, and so promote the proton transfer step that takes place in the reductive reaction.

Fig. 3.

Proposed mechanism of 3α-HSD. All ald-keto reductases follow this same mechanism. A direct hydride transfer from the C4 position of the nicotinamide ring to the ketone at the C3 position of the steroid substrate is required. Because there is no metal ion to polarize the acceptor carbonyl, a general acid is required to facilitate the reaction. The X-ray crystal structure of 3α-HSD implicated Tyr-55 as the general acid. Lys-84 is believed to help lower the pKa of the tyrosine to facilitate the transfer. Normally x is hydrogen; however, if this is replaced by a nitro group, as in nitrotyrosine, then the pKa of tyrosine should fall from 10 down to 7, and so possibly increasing the rate of reaction.

One protein that appeared substantially nitrated between days 2 and 3 is annexin 2. Annexins are a family of Ca2+ and phospholipid-binding proteins found to be evolutionary conserved throughout the plant and animal kingdom (22). All annexins contain a Ca2+-binding site and a central hydrophilic pore that may serve as a Ca2+ channel. In addition to this domain, annexins also contain a unique NH2-terminal domain that defines individual functions of each annexin. Annexin A2 is an F-actin binding protein that has a calcium-dependent filament bundling activity. It appears to play a role in the organization of membrane-associated actin sites of cholesterol-rich membrane domains. Also, annexin A2 has been proposed to function as a receptor for tissue-type plasminogen activator; therefore, its presence in a thrombolytic environment would be favored. Annexin A2 has been shown to be upregulated by hypoxia and by hyperoxidative stress in a renal cell carcinoma model. Therefore, increased protein expression may account for the late appearance of this particular nitration in the A549 cells because the basal concentration of the protein may be insufficient to be detected earlier.

Other identified targets of nitration include proteins involved in proteolysis. Two of these are the mitochondrial protein prohibitin and the proteosome activator PA28α. Prohibitin is a highly conserved protein that is an important component of the AAA protease, which is required for degradation of membrane proteins in mitochondria (4). PA28 activates hydrolysis by the 20S proteosome, a major component of protein degradative pathways in cells. It does not directly play a role in the degradative process but increases its rate (49). It is composed of two components, PA28-α and PA28-β (4 α-subunits and 3 β-subunits), in which only the α-subunit appears to stimulate the proteosome activity. A second activity of PA28-α is its involvement in refolding of proteins in the heat shock protein-90 pathway (41). The carboxy-terminal residue in PA28-α is a tyrosine, and removal of this residue results in loss of binding to the PA20 complex (61). If PA28-α is nitrated at this tyrosine, then this may enable PA28 dissociation from the proteosome and thus allow it to participate in the refolding activity as required when oxidative damage of proteins is occurring.

Dynamic protein nitration in mitochondria. Many targets that we have found to be nitrated in tissues and cells are mitochondrial or mitochondria-associated proteins (Fig. 4) (55). Diseases in which protein nitration and mitochondrial dysfunctions are observed include Parkinsons disease, diabetes, and amyotrophic lateral sclerosis (3, 26, 57). Mitochondria play a critical role in energy production and apoptosis, and these functions are affected by NO and related oxides (59). Mitochondria are expected to be a site of peroxynitrite generation when NO generation occurs inside the cells, in adjacent cells, or inside the mitochondria themselves (12, 23). We therefore chose to investigate protein nitration in isolated mitochondria using an ischemia-reperfusion model.

Fig. 4.

Involvement of protein nitration in mitochondria. Many of the proteins we have identified as being nitrated are of mitochondrial origin or are associated with mitochondria. Here we show a diagrammatic representation of energy production in mitochondria as well as mechanism of peroxynitrite generation. Pathways that are responsible for supply of substrates to the electron transport chain and apoptosis are shown. Proteins that are nitrated are in oval boxes or indicated with a solid star. TCA, tricarboxylic acid cycle; CoQ, coenzyme Q; Cyto C, cytochrome c; mtNOS, mitochondrial NO synthase; VDAC, voltage-dependent anion channel; ANT, adenine nucleotide translocase; SCOT, succinyl-CoA:3-oxoacid CoA-transferase.

The mechanism of cell death from ischemia-reperfusion is unclear (38). Studies contain contradictory findings as to whether cell death is induced during ischemia or reperfusion. In various animal models it was shown that ischemia for <3 h leads to no significant apoptosis but is greatly accelerated by reperfusion (58). However, others have found that signs of apoptosis appear as early as 10 to 15 min into ischemia with further damage occurring on reperfusion (29, 39). Apoptosis is a predominant mechanism of cell death in human acute myocardial infarcts and was shown to increase in reperfused myocardium (11).

In our experiments isolated mitochondria were subjected to rounds of ischemia and reperfusion in a stirred, sealed chamber. The experimental procedure is illustrated in Fig. 5. The state 4 oxygen consumption rate in the presence of l-arginine was 4.86 nmol/s. After hypoxia-anoxia was achieved and the sample was reoxygenated, the mitochondrial oxygen consumption decreased to 3.71 nmol/s (Table 2). Studies have demonstrated that resumption of blood flow to ischemic tissue leads to a burst of ROS, potentially leading to cell damage (7, 19). The ROS is most likely to be superoxide generated by the mitochondrial ETC in ischemia-reperfusion injury. Under normal physiological conditions, 2–5% of oxygen utilized by intact mitochondria is reduced to superoxide by the ETC (59). However, under ischemia-reperfusion, there is an increased likelihood of electron transfer to oxygen and thus a greater production of superoxide. Under normal conditions, the superoxide generated in mitochondria is detoxified by MnSOD. The importance of MnSOD was demonstrated by the early death of homozygous knockout mice. The clinical pathologies exhibited by the animals included myocardial injury and mitochondrial damage (56, 62). Additionally, protection from ischemia-reperfusion injury was conferred in transgenic MnSOD-overexpressing mice, demonstrating that the superoxide anion plays a role in tissue damage in this process (35).

Fig. 5.

Overview of the procedure of hypoxia-anoxia. Mitochondria were subjected to repeated hypoxia-anoxia for the periods described. Time points of substrate addition and aliquots taken are illustrated. RB, respiration buffer.

View this table:
Table 2.

Summary of mitochondrial respiratory rates

Superoxide can damage proteins and membranes directly or by reacting to form other bioactive molecules. For example, superoxide reacts with NO at a near diffusion-limited rate to generate the oxidant peroxynitrite (9). Indeed, NO is one of the only biological molecules that can out compete MnSOD for superoxide. Under normal conditions the levels of NO are low and therefore generation of peroxynitrite is small. During the first seconds to minutes of reperfusion, a burst of NO as well as superoxide occurs in the myocardium (38). This allows greater peroxynitrite formation, which may modify proteins, lipids, carbohydrates, and nucleic acids (31).

To determine whether NO was involved in our mitochondrial system, the experiment was repeated in the presence of d-arginine, which is not a substrate for NOS. With d-arginine, the initial rate of oxygen consumption was lower (4.18 nmol/s compared with 4.86 nmol/s with l-arginine). On a second cycle of hypoxia, the oxygen consumption had fallen to 3.12 nmol/s compared with 3.71 nmol/s in the l-arginine experiment. The ratio of O2 consumption from the first hypoxia to the second was 0.79 with l-arginine and 0.73 with d-arginine. These results imply that NO is not causing the decrease in O2 consumption rate and may instead be slightly protective.

To address whether protein nitration was involved in the process, samples of mitochondria were removed at different time points during the experiments and subjected to proteomic analysis. In these experiments, l-arginine was added to one chamber and d-arginine was added to a second chamber as a control. The results of the Western blot analysis are shown in Fig. 6. We observed nitrated proteins in the starting mitochondria sample (Fig. 6B), indicating a basal level of protein nitration. On subjecting the mitochondria to hypoxia-anoxia, protein nitration almost disappeared in the presence of either l-arginine or d-arginine (Fig. 6, C and F). This may be the first demonstration of rapid protein denitration under physiological conditions.

Fig. 6.

Anti-nitrotyrosine immunopositive proteins in mitochondria. Equivalent aliquots of mitochondria were taken during the hypoxia-anoxia experiment and subjected to two-dimensional PAGE and Western blotting (WB) using the monoclonal anti-nitrotyrosine antibody. A: Coomassie-stained two-dimensional PAGE gel. BG: WBs of the mitochondrial aliquots. B, C, and D: WBs using l-arginine; F and G: WBs with d-arginine added at the point indicted in Fig 5. B: aliquot taken at starting point after dilution of mitochondria in respiratory buffer. C and F: aliquots taken after 20 min of the first hypoxia-anoxia. D and G: aliquots taken after the reoxygenation phase. E: aliquot taken after the end of the second hypoxia-anoxia phase. In all experiments, two-dimensional PAGE gels, transfer to polyvinylidene difluoride membrane, and development of the WBs were run simultaneously to eliminate variability. No change in Coomassie staining was observed between the gels of all aliquots taken.

Removal of nitrated proteins has been suggested to occur either by proteolysis or by reductive enzymatic denitration (24, 30, 33). The apparent denitration that we observed was most likely due to an enzymatic process, given the rapid disappearance even in the presence of protease inhibitors. Coomassie staining of the mitochondrial protein samples showed that there was no obvious difference in the intensity of the protein spots (data not shown). Although it is still possible that the protease inhibitors were not completely effective, any potential degradation was insufficient to be visualized. Mitochondria may have evolved a rapid “denitration” mechanism in response to their being subjected to greater tyrosine nitration than in other organelles due to both NO and superoxide being produced locally. The presence of nitrated proteins in the starting mitochondrial sample suggests that either nitration and denitration are ongoing competing processes or that “denitration” only activates once oxygen levels begin to be depleted.

On reoxygenation, several mitochondrial proteins became renitrated (Fig. 6). Many of the same protein targets that were observed initially became renitrated. The process depended on the presence of l-arginine because very little renitration was seen in the mitochondrial experiment containing d-arginine. This result indicates that de novo NO synthesis was required and suggests that renitration of tyrosine may be involved. Indeed, NO synthesis should not be required if the initial denitration event was due to enzymatic reduction of nitrotyrosine to aminotyrosine. Mitochondria do not contain peroxidases; therefore, the process does not likely involve peroxidase-mediated nitration and is instead consistent with peroxynitrite as a potential nitrating source. This fits with the observation that mitochondria generate a burst of superoxide on transition from an anaerobic to aerobic environment, which in the presence of l-arginine-dependent NO synthesis could generate peroxynitrite.

After 20 min of reoxygenation, the levels of mitochondrial protein nitration were higher than in the starting sample. The mitochondria were then subjected to a second round of hypoxia-anoxia. During this time, protein “denitration” was again observed. However, the sample was not as fully “denitrated” as observed in the first cycle of hypoxia-anoxia. This may suggest that the denitration mechanism is less efficient due to enzymatic damage and/or depletion of a required cofactor. The mechanisms of protein denitration are currently under investigation in our laboratory.

It is not clear from our experiments what functions nitration may serve in mitochondria. The role that peroxynitrite plays in reperfusion injury is controversial where both beneficial and detrimental effect have been noted. Whereas it is clear that adding peroxynitrite to mitochondria induces protein nitration and leads to inhibition of mitochondrial energy production by causing specific lesions in the ETC, it is not clear what role protein nitration might play in normal physiology (43). Protein nitration could buffer changes in oxygen concentration under normal conditions; i.e., nitration may regulate ETC activity such that excessive superoxide generation is blocked. Protein nitration mediated by peroxynitrite may dampen activity at specific sites within the ETC, alter the flow of substrates feeding into the chain (Fig. 4), or act in a negative feedback loop. However, under pathogenic conditions where ischemia occurs, return of oxygen and the associated burst in superoxide generation may occur in the absence of nitrated proteins (due to their denitration during ischemia) and possibly lead to mitochondrial damage or cell death.

To date, protein nitration has typically been viewed as a cumulative, destructive process in which nitrotyrosine-containing proteins lose activity and damage cell function. This view stems from tyrosine nitration being observed in many diseases and from the demonstration that chemically nitrated proteins often lose their normal function. However, our current studies show that physiological tyrosine nitration can be selective, dynamic, and reversible, laying open the possibility of its being beneficial as well as detrimental. These concepts can now be further explored.



This work was supported by National Institutes of Health Grant R21 NS-41644 and American Heart Association (AHA) Grant AHA-01601598.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • * K. S. Aulak and T. Koeck contributed equally to this work.


View Abstract