Tyrosine nitration as mediator of cell death

Abstract

Nitrotyrosine is used as a marker for the production of peroxynitrite and other reactive nitrogen species. For over 20 years the presence of nitrotyrosine was associated with cell death in multiple pathologies. Filling the gap between correlation and causality has proven to be a difficult task. Here, we discuss the evidence supporting tyrosine nitration as a specific posttranslational modification participating in the induction of cell death signaling pathways.

Introduction

The perception of oxidants and free radicals in biological systems underwent important transformations in the last few years. Growing empirical evidence challenges the old dogma that free radicals and oxidants are simple unwanted byproducts of oxygen metabolism. Reactive species produced at low rates, leading to low concentrations within the cells can act as messengers involved in a number of cellular functions [1]. In contrast, overproduction of reactive species leads to oxidative damage associated with many pathological conditions including heart disease, inflammation and neurodegeneration [2, 3]. However, how oxidants and free radicals induce cell death remains highly controversial. It is pending to determine whether these reactive species trigger cell death by massive unspecific damage or through the modification of a limited number of specific cellular targets. Even more basic questions such as whether oxidation of macromolecules plays a role in the induction of cell death or is simply a consequence of the process of cell death is still under debate. In spite of their identity, it is accepted that reactive species exert their action through the oxidative modification of macromolecules in the cells. The nature, identity, selectivity and consequence of these modifications are the matter of profuse research and speculation [4–10]. In fact, the relevance of some oxidants and oxidative modifications in cell signaling has been dismissed due to the oxidants high reactivity and the irreversible nature of the oxidative modifications [10]. On the other hand, the prolonged half-life of these modifications could be advantageous in some conditions. There are permanent post-translational modifications that are widely accepted as events contributing to cell signaling. Some kinases are activated by partial proteolysis and inactivated by degradation [11–14]. Similarly, the endoproteases of the caspase family are activated also by partial proteolysis, which leads to a highly regulated proteolytic cascade responsible for apoptosis [15, 16]. These long-lived posttranslational modifications have profound consequences in the signaling pathways carrying the cellular machinery in one defined direction. However, even the activation of caspases is not always irreversible since the cells have inhibitor of apoptosis proteins (IAP) that can inhibit caspase activation and prevent cell death [15, 17]. In the case of tyrosine nitration, there is some evidence that the modification can be removed in certain conditions, though the mechanism remains unknown [18].

Defining the role of oxidative stress in cell death has been extremely difficult because of the multiplicity of potential targets that can be damaged by oxidation. Yet, oxidants can be surprisingly selective and act as second messengers, as described by Carl Nathan in his essay on “Specificity of a third kind” [19]. Every macromolecule in the cell is a possible target of oxidative damage, and massive oxidative damage to any macromolecule is enough to induce cell death. It is generally accepted that massive unspecific damage to DNA, proteins or lipids is enough to produce necrosis. Nonetheless, oxidative damage also activates stereotyped mechanism of cell death such as autophagy and apoptosis [20–23]. The question that remains to be answered is whether the cell senses unspecific damage and then activates the cell death program or the modification of key cellular components by oxidation is what leads to the activation of the programs.

In this review we revisit 20 years of literature, from the original description of peroxynitrite-induced cell death to the role of tyrosine nitration and the discovery of the first nitrated protein responsible for the activation of specific cell death pathways.

Reactive oxygen and nitrogen species

Oxidants and free radicals are frequently named as reactive species and grouped based on the atom that is more reactive and/or that originates the molecule. Free radicals are molecules with one or more unpaired electrons, while oxidants are molecules with a high potential for taking electrons from other molecules. Typical examples of free radicals with one unpaired electron are superoxide and nitric oxide. Oxygen on the other hand has two orbitals with one unpaired electron, which make it a di-radical [24]. The reactivity of members of the different “families” varies greatly and depends on the environment. In biological relevant conditions, some molecules such as superoxide are poor oxidants. However, in the presence of nitric oxide, superoxide forms the strong oxidant peroxynitrite in a diffusion-limited reaction [38]. The oxidant hydrogen peroxide in the absence of transition metals oxidizes only specific cysteine residues. However, in the presence of transition metals it forms the strong oxidant hydroxyl radical. In spite of the differences of reactivity, oxidants and free radicals are classified into several families of reactive species. The best-known examples are the reactive oxygen species (ROS), which originate by reduction of oxygen, and reactive nitrogen species (RNS), which are derived from the free radical nitric oxide.

Nitric oxide is produced from arginine by the three isoforms of nitric oxide synthase (NOS). Nitric oxide reacts with ROS to generate a variety of compounds with a very complex chemistry, previously reviewed in depth [2, 25, 26].

Most of these radicals and oxidants are produced in the cell during normal conditions and were traditionally considered aberrant byproducts of normal metabolism. In recent years it has been shown that some of them are not the byproduct of metabolism, but produced by the cell as intracellular and intercellular messengers. Nitric oxide was the first free radical with incontrovertible signaling functions in the regulation of blood vessel relaxation [27–29]. Further research showed that hydrogen peroxide also has important signaling functions [1]. Nonetheless, the literature and the experimental designs call for careful analysis due to the use of high concentrations of hydrogen peroxide as well as the lack of specificity of the methods employed to detect the oxidant [30, 31].

The identification of non-damaging regulatory oxidative modifications of proteins led to the concept of oxidative regulation of cellular processes and oxidative signaling. Oxidation of protein residues such as cysteine is now an accepted posttranslational modification. Nitrosation involves the oxidation of thiol groups on cysteine residues by nitric oxide. Cysteine nitrosation regulates the function of a number of proteins and is probably one of the more common mechanisms by which nitric oxide regulates cell metabolism [10, 32–35]. Similarly, hydrogen peroxide production leads to the oxidation of cysteine and other residues. Proteins with specific oxidized cysteine residues are now considered signaling molecules responsible for the regulation of a variety of normal cellular functions [1, 10, 30, 36]. In contrast, a more general oxidation of thiolates and other residues is present in pathological conditions. Although in most cases the mechanisms that give specificity to the reactions are not fully understood, even in normal physiological conditions not every residue that is oxidized participates in signal transduction. Many residues act like decoys or scavengers for the oxidants, as is the case of the cysteine in the active site of the peroxiredoxin family of peroxidases [37]. Growing evidence strongly suggests that the oxidant peroxynitrite is also an intra-cellular messenger [38–40].

Tyrosine nitration

Nitrotyrosine is a footprint left by peroxynitrite and other reactive nitrogen species [2, 41–44]. The mechanisms of tyrosine nitration are well understood, even though the relative participation of each mechanism in vivo is unknown (Fig. 1). Nitrotyrosine is present in a variety of human diseases and animal models of human diseases [4, 8, 45–49]. It is also present in inflammatory conditions such as multiple sclerosis [50–53], spinal cord injury, stroke and ischemia/reperfusion [54–59]. Although amply described in pathological settings, the presence of nitrotyrosine is not limited to disease since this oxidative modification can also be found in physiological conditions [40, 44, 60].

Fig. 1.

Mechanisms of protein tyrosine nitration in the cell and the effect on protein function. In the cells, the incorporation of a nitro group on a tyrosine residue can occur through different mechanisms. All mechanisms of nitration but one involve a double hit by radicals, the first forming the radical tyrosyl. The second reaction is a radical–radical reaction between the radical tyrosyl and nitrogen dioxide (^.NO₂). Nitration could be mediated by the radical products of decomposition of peroxynitrite nitrogen dioxide and hydroxyl radical (^·OH). The exception for the two radical hit is the nitration catalyzed by transition metals or peroxidases. Peroxynitrite can react also with bicarbonate (${\text{HCO}}_3^{-}$) to form oxoperoxocarbamato, whose radical products of decomposition efficiently nitrate tyrosine residues. Nitrite (${\text{NO}}_2^{-}$) and hydrogen peroxide (H₂O₂) can also nitrate tyrosine residues in the presence of transition metals or peroxidases. Depending on the location of the tyrosine residue, nitration of that residue may not have a functional consequence, or the protein may be inactivated, activated or may gain a new function

Dozens of nitrated proteins in vivo have been identified by mass spectrometry [2, 45]. The lack of tools to identify specific targets and methods to recapitulate the specific forms of oxidative modification or damage has greatly limited the comprehension of the role that nitrative stress plays in pathology. Not all tyrosine residues are equally probable to be nitrated. Location and microenvironment also play an important role. For nitration to occur the tyrosine residue must be ionized, exposed and the conditions that stabilize the tyrosyl radical as well as the intermediates that allow the nitrogen dioxide attack must be present [61, 62]. In addition, the amount of a protein and the tyrosine content are not predictors of the level of nitration [4, 61, 63, 64]. We have shown that in a simple tetrapeptide, replacement of arginine by alanine is enough to decrease 20 % the efficiency of nitration by peroxynitrite in the presence of bicarbonate, suggesting that some of the specificity is the consequence of the tyrosine environment [65].

Peroxynitrite and cell death

Shortly after the discovery of peroxynitrite and nitrotyrosine in biological systems [38, 66], several independent groups almost simultaneously reported that incubation of various cell types with exogenous peroxynitrite induces apoptosis or necrosis depending on the concentration of the oxidant [67–70]. Endogenous production of peroxynitrite by downregulation of Cu, Zn superoxide dismutase in PC12 cells, or trophic factor deprivation and Fas receptor activation in motor neurons results also in the activation of apoptosis [71–74]. In all these conditions, nitration of tyrosine residues always precedes peroxynitrite-induced apoptosis. One important observation is that peroxynitrite activates specific pathways for the induction of apoptosis. In PC12 cells, the incubation with peroxynitrite has different effects on cell death depending on the interaction with specific signal transduction pathways (Table 1). Pre-incubation of PC12 cells with nerve growth factor (NGF) or insulin protects the cells from peroxynitrite-induced apoptosis. Similarly, the incubation with insulin immediately after exposure to peroxynitrite is also protective. Conversely, pre-incubation of the cells with fibroblast growth factor (FGF) or incubation with NGF immediately after exposure to peroxynitrite results in increased cell death [69, 75–77]. These results reveal specific and selective interaction of peroxynitrite with signaling pathways activated by the different trophic factors. In fact, FGF added immediately after peroxynitrite or endothelial growth factor added at anytime does not modify the effects of peroxynitrite on cell death. These results suggest that peroxynitrite activates specific intracellular signaling pathways to stimulate apoptosis rather than just causing unspecific molecular oxidation.

Table 1.

Interaction of trophic factors with peroxynitrite in the induction of cell death

Trophic factor	Effect on survival after incubation with peroxynitrite
	Before	After
NGF	↑↑	↓↓
Insulin	↑↑	↑↑
IGF1	↑↑	↑↑
FGF1	↓	=
FGF2	↓	=
EGF	=	=

↑ Prevents cell death; ↓ enhances cell death; = No effect

Tyrosine nitration as mediator of peroxynitrite-induced cell death

As mentioned before, tyrosine nitration is present under pathological conditions but whether it plays a causative role or is merely collateral damage has been difficult to determine. Tyrosine does not react directly with peroxynitrite, but with peroxynitrite-derived radicals such as carbonate radical and nitrogen dioxide [44, 65, 78–80] (Fig. 1). However, in the presence of transition metals peroxynitrite can directly nitrate tyrosine residues [81]. In addition, tyrosine residues can also be nitrated by myeloperoxidase using hydrogen peroxide and nitrite as substrates [82, 83]. To directly assess the role of tyrosine nitration on cellular physiology and cell death, we developed tyrosine-containing peptides that act as a “decoy” for the nitrating species derived from peroxynitrite. The amino acid sequences of the peptides were based on known nitration sites from endogenously nitrated proteins. In vitro, these peptides did not scavenge peroxynitrite, but prevented nitration of proteins in cell and tissue homogenates. When delivered intracellularly, peptides containing tyrosine, but not phenylalanine or proline, prevented motor neuron apoptosis induced by trophic factor deprivation and peroxynitrite-induced PC12 cell apoptosis [65] (Fig. 2). From the peptides tested, only those containing tyrosine or tryptophan residues were protective against peroxynitrite-induced cell death but had no effect on hydrogen peroxide- or staurosporine-induced apoptosis. In addition, the peptides did not prevent glutathione oxidation by peroxynitrite in PC12 cells and in vitro [65], in agreement with being unable to scavenge peroxynitrite. As expected, thiol-containing compounds such as N-acetyl cysteine and glutathione prevented peroxynitrite-induced apoptosis. The direct reaction of peroxynitrite with thiols makes these compounds excellent scavengers of the oxidant [79, 84]. Moreover, the scavengers of lipid peroxyl radicals, vitamin E and the soluble analogue trolox, did not provide protection against peroxynitrite-induced apoptosis, suggesting that lipoperoxidation is not involved in peroxynitrite-induced apoptosis. Thus, tyrosine-containing peptides do not prevent thiol oxidation but tyrosine nitration. Collectively, these results suggest that peroxynitrite-induced apoptosis is mediated by the nitration of tyrosine residues in proteins.

Fig. 2.

Role of tyrosine nitration in peroxynitrite-induced apoptosis in PC12 cells and motor neurons. Nitration of tyrosine residues has been implicated in the induction of cell death upon exogenous addition or endogenous production of peroxynitrite. In PC12 cells, peroxynitrite treatment leads to the simultaneous inactivation of the PI3K/Akt pathway and the activation of the p38-MAPK and JNK pathways, ultimately inducing cell death. In motor neurons, the endogenous production of peroxynitrite upon trophic factor deprivation leads also to cell death. In both cases, intracellular delivery of tyrosine-containing peptides is protective, suggesting that peroxynitrite-induced cell death is mediated by tyrosine nitration

Tyrosine nitration, protein function, and cell death: correlation or causality?

Not all the tyrosine residues on a protein are prone to nitration. In fact, a very small number of tyrosine residues are nitrated in inflammatory conditions. During inflammation only 100–500 μmol of nitrotyrosine per mole of tyrosine residue can be detected [85]. Depending on the surrounding microenvironment and accessibility, some tyrosine residues will be preferential targets of nitration. In addition, constitutive proteins that are highly expressed in the cells, such as actin and heat shock proteins, constitute primary targets probably due to their abundance [2, 65]. Modifications of enzymes that reduce the total activity by 1 % are most likely not significant, but modifications that significantly reduce protein activity, activate an enzyme or cause gain-of-function may be highly significant. As detailed below, nitration of specific tyrosine residues causing protein loss or gain-of-function has been correlated with cell death.

The best-described effect of tyrosine nitration on proteins is the inactivation or loss of function. Twenty years ago manganese superoxide dismutase (MnSOD) was the first protein described as a target of tyrosine nitration and inactivation in vivo during human kidney allograft rejection [86]. This protein is particularly susceptible to nitration because the manganese atom in the active site catalyzes the nitration of the tyrosine residue, which results in the loss of activity [86, 87]. Nitration of tyrosine 34, located in the active site of the protein, completely inhibits MnSOD enzymatic activity upon treatment with peroxynitrite [88, 89]. Interestingly, shortly after this tyrosine was identified it was shown that its conservative replacement by phenylalanine, an amino acid resistant to nitration, failed to protect MnSOD from peroxynitrite-mediated inactivation, suggesting that tyrosine nitration was not the only oxidative modification involved in MnSOD loss of function [87].

One of the most important barriers for the study of oxidative modifications is that the incubation of proteins with any oxidant produces a variety of oxidative products that are not homogenously modified. Tyrosine nitration may be achieved by different methods, but other residues such as tryptophan and cysteine will also be oxidized. In addition, even if a protein contains only a few tyrosine residues, there is no certainty that all the molecules will be homogeneously nitrated in all sites. Then, it is difficult to assign an effect to the modification of one residue. The recent development of a method for the incorporation of the non-natural amino acid nitrotyrosine into recombinant proteins provides a new powerful tool to solve all these pitfalls [90]. Using this methodology, it was demonstrated that the presence of nitrotyrosine in position 34 as the sole modification on MnSOD inhibits 97 % of the superoxide dismutase activity of the enzyme. This finding raises the question as to how reliable the substitution of amino acids is when studying oxidative modifications, considering that protein folding and amino acid microenvironment is key to the final result.

Over the years, MnSOD tyrosine nitration has been observed in several pathologies and animal models, including renal and hepatic ischemia/reperfusion, acute rejection of cardiac transplants, cardiovascular disease, aging, human brain injury, and cyclosporine A-related vascular toxicity [91–96]. This enzyme is in the mitochondrial matrix at 10–20 μM concentrations, and is particularly relevant to the mitochondrial metabolism [97, 98]. The genetic deletion of the MnSOD gene is early lethal [99]. However, mice heterozygous for the deletion of the MnSOD, which has 50 % MnSOD activity, show increased susceptibility to mitochondrial oxidative damage and increased apoptosis at younger ages than wild type mice [100]. In the setting of a pathological condition, this means that for MnSOD inactivation to be relevant as cause of cell death, more than 50 % of the enzyme should be nitrated and inactivated. The nitrating products from one molecule of peroxynitrite can nitrate only one tyrosine residue and the efficiency of nitration in vitro is between 10 and 30 %, but in some cases can achieve 100 % [80], meaning that the nitration of one tyrosine residue requires the production of 1–3 molecules of peroxynitrite. Applying these calculations to the nitration of only one tyrosine in MnSOD, the concentration of peroxynitrite-derived radicals produced in a cell should be at least equal to the concentration of MnSOD, considering the more efficient nitration conditions [80]. In this context, it is possible to speculate that peroxynitrite should be either acutely produced at over 10 μM concentrations, or the steady-state concentration of peroxynitrite should be sustained over time to compensate for the MnSOD protein turnover. Given that these conservative calculations do not consider endogenous scavengers of peroxynitrite such as thiols and competing tyrosine residues in other proteins, achieving conditions in which more than 50 % of the MnSOD is inactivated would require extreme circumstances most probably associated with catastrophic pathological conditions.

Nitration and inactivation of other enzymes involved in antioxidant mechanisms, such as thioredoxin-1 (Trx-1), has also been associated with cell death. Trx-1 is a small, 12 kDa oxidoreductase protein involved in redox metabolism. Trx-1 is ubiquitously expressed in cells. Human Trx-1 contains a single tyrosine residue at position 49 [101]. Nitrated Trx-1 is detected in cardiac tissue subjected to ischemia/reperfusion, and nitration of this tyrosine residue in vitro decreases Trx-1 activity by 80 % [102]. Administration of recombinant human Trx-1 proved to be cardioprotective in a myocardial ischemia/reperfusion animal model [102, 103]. Interestingly, administration of nitrated Trx-1 failed to protect myocardial tissue from apoptosis. Replacement of tyrosine 49 by phenylalanine was enough to restore the cardioprotective role of Trx-1 upon treatment of the protein with the NO and superoxide donor 3-morpholino-sydnonimine (SIN-1) [102], suggesting that nitration of tyrosine 49 and inactivation of Trx-1 contributes to postischemic cardiomyocyte apoptosis. However, the authors used exogenous enzyme to test the protective effects and inactivation of the enzyme. There is neither quantification of endogenous nitration nor investigation of how much enzyme should be inactivated for the cells to die reported in the manuscript. In these conditions, it can only be established a correlation between the presence of the nitrated protein in disease conditions, but not a causal relation between the nitration of the enzyme and the disease.

The induction of structural changes by tyrosine nitration on proteins involved in apoptotic processes may also play a role in the regulation of cell death. Nitration of the mitochondrial protein cytochrome c occurs in vitro upon treatment with peroxynitrite [104] and in vivo when reactive nitrogen species are endogenously produced [91, 105]. Nitration of the solvent-exposed tyrosine 74 induces structural changes that may have an impact on cytochrome c function [106]. Interestingly, in the presence of nitrite and hydrogen peroxide, cytochrome c can catalyze its own nitration as well as nitration of adjacent proteins such as MnSOD [107]. In in vitro studies, nitrated cytochrome c fails to assemble a functional apoptosome [108, 109]. Similarly, replacement of all but one tyrosine residue at either position 46, 48, or 74 by phenylalanine in the human cytochrome c prevents caspase 9 activation [110, 111]. However, nitration of tyrosine 46 and 48 has not been described to occur in vivo and nitration of these residues seems to induce degradation of the nitrated protein [112]. Whether nitration of cytochrome c is relevant to the regulation of cellular apoptotic processes in vivo is still being investigated.

Tyrosine hydroxylase (TH) is another protein whose loss of function by tyrosine nitration was linked to cell death several years ago. This enzyme is the rate-limiting enzyme in dopamine synthesis. Decreased dopaminergic function in neurons that undergo neurodegeneration is a hallmark of Parkinson’s disease (PD). In a mice model of PD, the loss of TH activity correlates with an increase in tyrosine nitration and a decrease in dopamine levels [113]. Indeed, peroxynitrite treatment induces TH loss of enzymatic activity in vitro. The enzymatic inhibition can be prevented by the conservative replacement of tyrosine 423 on TH by phenylalanine [114]. Although these results make a strong case for the role of tyrosine nitration in the lost functionality of dopaminergic neurons in Parkinson’s disease, two key issues remain to be elucidated. The first is the proportion of TH that is inactivated in pathological conditions to determine whether it would produce a significant alteration in the production of dopamine. The second is to establish a relation between the nitration of TH and the induction of cell death.

As with other post-translational modifications, tyrosine nitration on certain proteins can induce a gain-of-function, one that the unmodified protein cannot perform. This is the case of α-synuclein and heat shock protein 90 (Hsp90). The presence of nitrated α-synuclein in inclusion bodies in PD, dementia, and Alzheimer’s disease was described more than 10 years ago [115]. Nitration of α-synuclein has been proposed to be responsible for the aggregation and toxicity of this protein in these synucleinopathies [115, 116]. Indeed, nitrated α-synuclein was detected in vivo in the substantia nigra of old primates (>16 years of age) but rarely in adult mature animals (<10 years of age) [117]. In addition, nitrated α-synuclein evades immunological tolerance, activating an immune response that in turn accelerates nigral dopaminergic neuron degeneration [118]. The recombinant protein used in this report was nitrated using a final peroxynitrite concentration of 5 M in the presence of 5 mM iron. These extreme oxidative conditions should surely produce other oxidative modification in the recombinant protein that can participate in the gain of toxic function. Important controls using α-synuclein with the tyrosine residues replaced by phenylalanine show that tyrosine nitration is necessary. However, whether tyrosine nitration is sufficient to induce the gain-of-function is not evident from these experiments.

Nitration and oxidation of α-synuclein induce highly stable oligomer formation in vitro and stabilize pre-assembled α-synuclein filaments increasing the formation of aggregates [119]. Dimers and polymers of α-synuclein nitrated at all tyrosine residues (tyrosine 39, 125, 133 and 136) trigger cell death when exogenously added to the human neuroblastoma cell line SH-SY5Y at 5 μM concentration. Nitrated α-synuclein binds to integrin on the cell membrane, increases inducible NOS expression and activity, and decreases phosphorylation of focal adhesion kinase (FAK), followed by activation of caspase 3 [120]. The nitration of α-synuclein in this report was performed incubating with NaNO₂ for 7 days by a non-described mechanism. Alpha-synuclein nitrated by this method was characterized using western blot and mass spectrometry. However, there is no information showing the complete mass spectrometric analysis of the protein to dismiss other possible oxidative modifications or other type of modification due to the long incubation.

Nitrated Hsp90 is an effector in peroxynitrite-induced apoptosis

As detailed above, nitration of tyrosine residues on several proteins has been associated with induction of cell death. Nonetheless, demonstrating causality rather than a correlation between nitration and cell death has proven difficult to achieve. The main reasons are the number of potential oxidative and nitrative modifications and the lack of tools to assess the role of tyrosine nitration as the sole modification on a protein. By intracellularly delivering tyrosine-containing peptides as nitration decoys we were the first to demonstrate that tyrosine nitration did play a role in peroxynitrite-induced cell death (Fig. 2). At that time, the nitrated proteins that were the actual effectors in the induction of cell death remained unknown. Recently, we identified one of such nitrated proteins. We showed that nitration of a single tyrosine residue on Hsp90 turns this pro-survival chaperone into a toxic protein [121]. This is the first demonstration that tyrosine nitration on a specific protein plays a role in the induction of cell death, establishing causality rather than correlation.

Hsp90 is a ubiquitous molecular chaperone present in all eukaryotic cells [122]. It is expressed at high concentrations constituting 1–2 % of the total cytosolic protein [123]. Most proteins in the cell require Hsp90 for folding, stabilization, activation or repression, degradation, and translocation to subcellular compartments [124, 125]. Due to this vast array of functions, Hsp90 is necessary for cell survival and proliferation [126]. Structurally, Hsp90 has three distinctive domains: N-terminal domain (ND), where the ATP-binding pocket resides, middle domain (MD), responsible for the binding to client proteins, and C-terminal or dimerization domain (CD) [127]. Functionally, Hsp90 is a dimer that binds client proteins and hydrolyze ATP in a complex cycle with several intermediate conformations. Amino acids from the different domains interact and are crucial to this Hsp90 ATPase cycling [125, 127]. The human Hsp90 sequence has 24 tyrosine residues, five of which are target of nitration [121]. Two of these residues, tyrosine 33 and 56 in Hsp90β are located in the ND, near the ATP binding pocket (Fig. 3). By using a new, state-of-the-art technology that allows the genetically encoded incorporation of nitrotyrosine into proteins at specific positions [90], we were able to express a recombinant human Hsp90 with nitrotyrosine at position 33 or 56. The presence of nitrotyrosine at either of these positions as the sole modification on Hsp90 proved to be necessary and sufficient to turn the chaperone from a pro-survival to a pro-apoptotic protein. Similar results were obtained by intracellularly delivering a peroxynitrite-treated Hsp90 in which 4 tyrosine residues were replaced by phenylalanine, maintaining either tyrosine 33 or 56 intact [121].

Fig. 3.

Induction of cell death by nitrated Hsp90. The intracellular delivery of peroxynitrite-treated Hsp90 induces cell death by the activation of a novel P2X7/Fas pathway. The genetically directed incorporation of nitrotyrosine at either position 33 or position 56 confirmed that the incorporation of a single nitrotyrosine on Hsp90 is sufficient to induce cell death

The prediction from 3D modeling analysis is that nitration of tyrosine 56 should impair ATP binding to Hsp90, while tyrosine 33 (tyrosine 24 in yeast) plays a crucial role during Hsp90 ATPase cycle [128]. Indeed, peroxynitrite-treated Hsp90 retains only 20 % of the ATPase activity compared to unmodified Hsp90 [121]. Intracellular delivery of nitrated Hsp90 at a concentration as low as 5 % of the total endogenous Hsp90 is sufficient to induce cell death, implying that the nitration of Hsp90 causes a gain-of-function. Interestingly, trophic factor deprivation stimulates the nitration of 5 % of total Hsp90 in cultured motor neurons before the induction of apoptosis, but at the time the cells are committed to die. In these cells, nitrated Hsp90 induces cell death by activating P2X7 receptor, which in turn activates the Fas/Fas ligand cell death pathway [121] (Fig. 3). By developing an antibody that recognizes Hsp90 when nitrated at position 56 we detected the toxic protein in vivo in pathological conditions such as amyotrophic lateral sclerosis and spinal cord injury [65, 121]. This is the first demonstration that nitration of specific tyrosine residues on a specific protein activates apoptotic pathways and induces cell death.

Concluding remarks

There is the general, widely accepted idea that RNS induce cell death by simultaneously causing general oxidative damage to multiple cellular components. Depending on the concentration at which they are produced and the cellular context, RNS can indeed cause such damage, leading to cell death by necrosis. However, there is a more subtle and relevant biological aspect to RNS that has only recently started to surface. Over the years, others and we have shown that RNS can stimulate apoptosis through the activation of tightly regulated pathways. More and more evidence supports the concept of RNS acting as second messengers in the regulation of cellular metabolism. In the last 20 years, tyrosine nitration has been detected in numerous pathologies in the cell types or tissues most affected by the disease. In addition, tyrosine nitration can lead to protein inactivation. As a consequence, tyrosine nitration was naturally associated with cell death without empirical evidence showing causality rather than correlation. Only recently, the first nitrated protein playing an actual role in the induction of cell death was described. Tyrosine nitration is the switch turning a pro-survival protein such as Hsp90 into an executioner of cell death. Several reports have shown that tyrosine residues can be reduced and the nitro group removed, though the enzymatic activity responsible for this activity has not been determined [18, 129]. These observations suggest that tyrosine nitration can also be regulated by means other than protein degradation. It can be hypothesized that some tyrosine residues are functional targets for nitration, while the vast majority of nitration is just a protective mechanism to secure specificity of the signals. The same concept could be applied to other oxidative modifications. In this context, in order to prevent the formation of specific and potentially ‘toxic’ nitrated proteins, most of the nitration targets may act as decoy scavenging the nitrating agents. Thus, only a small amount of any given protein would be nitrated, preserving the total activity and function of that protein. In addition, it is possible that as some residues in some proteins evolved to be nitrated as a signaling event, some residues in other proteins evolved to act as a decoy for nitration without any effect on protein function. In this context normal protein turnover would be responsible for removing unwanted nitration. It is the modification of specific residues on specific proteins which gives powerful oxidants such as peroxynitrite, the specificity required for them to be second messengers involved in the control of cellular processes. These are some of the new challenges ahead of us to understand the biology of reactive nitrogen species.