Protein S-Nitrosylation and Cardioprotection

Abstract

Nitric oxide (NO) plays an important role in the regulation of cardiovascular function. In addition to the classic NO activation of the cyclic guanosine monophosphate (cGMP)-dependent pathway, NO can also regulate cell function through protein S-nitrosylation, a redox dependent, thiol-based, reversible post-translational protein modification that involves attachment of an NO moiety to a nucleophilic protein sulfhydryl group. There are emerging data suggesting that S-nitrosylation of proteins plays an important role in cardioprotection. Protein S-nitrosylation not only leads to changes in protein structure and function, but also prevents these thiol(s) from further irreversible oxidative/nitrosative modification. A better understanding of the mechanism regulating protein S-nitrosylation and its role in cardioprotection will provide us new therapeutic opportunities and targets for interventions in cardiovascular diseases.

I. Introduction

Nitric oxide (NO) plays an important role in the regulation of redox signaling and cellular function^1-3. NO can be generated by NO synthase (NOS) or by the breakdown of nitrite or other compounds to NO^4-6. NOS catalyzes the synthesis of NO by the conversion of L-arginine and oxygen to L-citrulline and NO, in the presence of nicotinamide adenine dinucleotide phosphate (NADPH) and tetrahydrobiopterin. Both endothelial NOS (eNOS) and neuronal NOS (nNOS) are constitutively expressed in distinct subcellular locations within cardiomyocytes: eNOS is generally thought to be predominantly localized to invaginations of the sarcolemma called caveolae^{7, 8} and nNOS is mostly found in the sarcoplasmic reticulum (SR)^{8, 9}. However, changes of expression and redistribution of these two constitutive isoforms of NOS have been found in some disease conditions such as myocardial infarction^{10, 11} and heart failure¹². The third isoform of NOS, the inducible NOS (iNOS), is barely detected in the myocardium under normal conditions, but expression of iNOS is stimulated by inflammatory mediators^{13, 14}. A mitochondrial NOS has been reported to be present in the inner mitochondrial membrane or matrix¹⁵, although there are conflicting results regarding mitochondrial NOS^16-19. An alternative source of mitochondrial NO could be nNOS associated with the SR, as the SR membrane has been shown to be attached to the outer mitochondrial membrane^20-22. This differential NOS localization provides organelle or subcellular generation of NO, which in turn provides a localized signal for protein S-nitrosylation^{23, 24}.

In addition to activating cyclic guanosine monophosphate (cGMP)-dependent signaling pathways, recent studies suggest that NO, by attachment of an NO moiety to a nucleophilic protein sulfhydryl resulting in S-nitrosothiol (SNO) formation, generates a post-translational modification known as protein S-nitrosylation, which plays an important role in biology^{25, 26}. Recent data suggests that protein SNO is important in cardioprotection. This review will focus on the role of S-nitrosylation in cardioprotection and the potential mechanism(s) involved.

II. NO and Ischemia/Reperfusion (I/R) Injury

NO is generated during I/R

A number of studies have shown that NO is produced during ischemia^{4, 27-29}. The NO that is generated appears to be due to both NOS dependent and NOS-independent pathways. During ischemiawhen oxygen is low, nitrite conversion into NO is likely to be an important source of NO. However there is likely to be some residual oxygen, particularly at the beginning of ischemia, and Zweier et al have shown that with short durations of ischemia (30 minutes) N-nitro-L-arginine methyl ester (L-NAME), a NOS inhibitor, blocks approximately 60-80% of NO generation²⁷. Thus it appears that both NOS dependent and independent (nitrite) mechanisms are involved in the generation of NO during ischemia. The exact proportion of NOS dependent and independent NO generation may depend on the model (the duration of ischemia, the diet, and the species). These data show that NO is generated during ischemia, which raised the question of whether the NO generated during ischemia is beneficial or detrimental?

Increasing NO protects

Addition of NO donors, particularly at concentrations that release NO at physiological (sub-micromolar) levels, has been shown to reduce I/R injury^{30, 31}. Similarly, cardiac specific overexpression of eNOS^{32, 33} or iNOS³⁴ has been reported to reduce I/R injury. Thus, there is reasonable agreement that a slight increase in NO is cardioprotective.

Consistent with a protective role for NO, a number of cardioprotective models have been shown to involve an increase in NOS³⁵. For example, protection in females^{36, 37} and erythropoietin-mediated protection^{38, 39} require NOS activation. Also in models of I/R injury in various tissues such as heart and liver, nitrite has been shown to reduce infarct size and to reduce apoptosis and necrosis^{5, 6}. The mechanism of nitrite-mediated cytoprotection appears to be NO-dependent, as a loss of cytoprotection occurred when animals were treated with a NO scavenger^{5, 6}.

Reducing NO below basal levels has no consistent effect on protection

Whether loss or inhibition of NOS exacerbates I/R injury has been controversial, especially the data involving eNOS knockout (eNOS-KO) mice. Studies comparing infarct size after I/R in wild type (WT) and eNOS-KO hearts report no difference^40-42, a reduction⁴³ and an increase in infarct size^{43, 44}. Part, but not all, of the discrepancy appears to be related to two different models of eNOS-KO mice. Lefer and coworkers⁴³ studied I/R in the eNOS-KO mice generated at Harvard compared to those generated at UNC. They found an 84% increase in infarct size in the Harvard eNOS-KO mice, but a decrease in infarct size in the UNC eNOS-KO mice. This difference was attributed to a compensatory increase in iNOS in the UNC eNOS-KO mice which resulted in protection in these mice⁴³. However, Guo et al⁴² also examined I/R injury in both of these mouse models and in contrast to the study by Lefer's group, reported no difference in infarct size in either UNC or Harvard eNOS-KO hearts compared to WT hearts. It is possible that the induction of iNOS in the UNC mice might depend on the housing and feeding conditions of the mice, and this might account for some of the differences observed. Most studies of infarct size show no difference between WT and nNOS knockout (nNOS-KO) hearts^{44, 45}. Most recent studies find that NOS inhibition does not alter infarct size or recovery of function following I/R^{42, 46, 47}. However, there are some older reports showing that NOS inhibition can be protective^{27, 48, 49}. NO can combine with reactive oxygen species (ROS) to generate peroxynitrite which can be detrimental particularly at high levels^{50, 51}. This discrepancy between the beneficial and detrimental effects of NO might depend on the relative availability of NO versus ROS. NOS can also become “uncoupled” resulting in the generation of ROS and this might also complicate the interpretation of the effects of NOS inhibitors. Thus the effect of NOS inhibition is variable and might depend on the level of oxidative stress. Taken together, the data suggest that under most conditions an increase in NO above a threshold level, as might occur with NO donors or overexpression of NOS, is protective; however lowering NO below basal levels does not appear to increase injury.

III. NO and Cardioprotection

As mentioned above, a slight increase in NO has been shown to be cardioprotective. Cardioprotection can be initiated acutely by activation of signaling pathways (acute protection) or by upregulation of new proteins (delayed protection), and NO has been shown to be important in both models. For simplicity we will discuss these separately and focus on acute protection. Acute protection can be initiated by signaling pathways activated prior to and during ischemia, i.e., preconditioning (PC), as well as pathways activated at reperfusion, such as postconditioning.

A. NO and acute PC

NO increases during PC

Brief episodes of ischemia and reperfusion, i.e., PC, render the heart more resistant to subsequent prolonged periods of ischemia⁵². PC has been shown to lead to an increase in NO and inhibitors of PC such at inhibitors of the PI3K pathway block the increase in NO production^53-55. Inhibitors of NOS have been shown to block protection in a number of different cardioprotective models. As will be discussed, inhibitors of NOS block protection associated with delayed PC and postconditioning. Also cardioprotection in females is blocked by inhibitors of NOS or loss of either eNOS or nNOS^{11, 56}, and erythropoietin mediated protection has also been reported to be blocked by NOS inhibitors⁵⁷. The ability of NOS inhibition or loss of NOS to block acute PC, however, has resulted in discrepant results. Some^{58, 59}, but not all studies^{60, 61} find that NOS inhibitors block acute PC. Interestingly, although Downey's group initially reported that NOS inhibitors do not block acute PC⁶⁰, Cohen, Downey and coworker⁵⁴ recently reported the ability of L-NAME to block PC depends on the relative contribution of signaling through the adenosine pathways compared to the bradykinin or opioid pathways. They suggest that signaling via the adenosine pathway is not blocked by L-NAME. Downey's group also reportedthat NOS inhibitors block bradykinin mediated protection⁶². Studies with mice lacking eNOS have also provided mixed results. Bell and Yellon reported that in eNOS-KO hearts, PC still reduced infarct size with 4 cycles of PC, but there was no reduction in infarct size with only 2 or 3 cycles⁶³. However, Guo et al in a very well established in vivo infarct model found that both Harvard and UNC eNOS-KO hearts exhibited a PC mediated reduction in infarct size that was comparable to that observed in WT hearts, suggesting the eNOS is not necessary for acute PC⁴².

So what do we conclude regarding the role of NOS and NO in acute PC? Data consistently show that PC results in an increase in NO. However, inhibition of NOS, by either pharmacological inhibitors or genetic ablation, results in inconsistent results; in some cases PC protection is blocked, but not in others. One explanation provided by the Downey, Cohen group is that L-NAME does not block under conditions where PC signaling is predominantly via adenosine, rather than bradykinin or opioids. Another possible explanation for this discrepancy might be a differential role provided by non-enzymatic production of NO during PC. NO can be generated from nitrite under acidic conditions that occur during PC⁶⁴, and this non-enzymatically generated NO would not be inhibited by NOS inhibitors. The levels of nitrite can vary depending on diet, thus leading to diet dependent differences in the NO that can be generated from nitrite⁶⁵. Thus, some of the differences regarding whether NOS generated NO is required for acute PC, might depend on the levels of non-enzymatically generated NO, which in turn might depend on diet. If sufficient NO is generated via non-enzymatic mechanisms, then NOS inhibitors would not be expected to block PC. This explanation might account for the more consistent role for NOS inhibitors in pharmacological PC (where there is no acidosis to generate NO nonenzymatically) and it would also be consistent with NOS inhibitors blocking after 2 or 3 cycles, but not after four cycles of PC^{54, 63}. This hypothesis would also be consistent with a recent study reporting that dietary nitrite restores cardioprotection in eNOS-KO hearts⁶⁵.

B. NO and postconditioning

It has been shown that following a sustained period of ischemia, brief intermittent periods of ischemia and reperfusion administered at the start of reperfusion (postconditioning) reduced infarct size⁶⁶. Tsang et al reported that postconditioning resulted in an increase in phosphorylation of eNOS and that the protective effects of postconditioning were blocked by the inhibition of the phosphatidylinositol 3-kinase (PI3K) with wortmannin, which blocked the subsequent eNOS phosphorylation⁶⁷. Consistent with a role for NOS in postconditioning, Yang et al found that the NOS inhibitor, L-NAME, administered just before reperfusion blocked the protective effects of postconditioning⁶⁸. In a recent study, addition of a mitochondria-targeted SNO at the start of reperfusion has also been found to be cardioprotective⁶⁹. Thus, NO appears to be an important mediator in postconditioning.

C. NO and delayed PC

As discussed, a trigger phase of brief intermittent periods of ischemia and reperfusion reduces infarct size following a subsequent sustained period of ischemia. The protection afforded by PC is lost if there is a delay between the brief intermittent episodes of ischemia and the sustained ischemia. Typically, if there is an hour or longer delay between the trigger preconditioning period and the sustained period of ischemia the protection is lost. However, a delayed protection occurs about 24 hours after the PC trigger phase. In contrast to acute PC⁷⁰, delayed PC involves alterations in gene expression^{13, 71}. Interestingly it has been shown that NO generated by eNOS during the trigger phase is required for the protection that is observed 24 hours later^{53, 58, 72, 73}. NO generated during the trigger phase activates a signaling cascade that leads to upregulation of iNOS, which is essential for the delayed PC, and that the delayed PC was blocked in mice lacking eNOS^{74, 75}. Furthermore, cardiomyocyte-specific overexpression of iNOS has been shown to attenuate reperfusion-induced ROS production and protect against I/R injury by preventing the mitochondrial permeability transition pore (MPTP) opening and cell death³⁴.

III. Mechanism of NO Protection

Initial studies indicated that many NO mediated effects were due to NO activation of soluble guanylyl cyclase (sGC) and the resulting increase in cGMP⁷⁶. However, NO can modify protein thiol groups resulting in the formation of protein SNO, and this protein modification can also mediate effects of NO^{25, 26}. Both sGC and SNO pathways have been reported to be important in cardioprotection. NO activation of sGC/cGMP/protein kinase G(PKG) is reported to activate the mitochondrial K_ATP channel, by a mechanism that has not been elucidated, but which presumably involves phosphorylation by PKG^{77, 78}. In support of this hypothesis, 1H-[1,2,4]-oxadiazole-[4,3-a]-quinoxalin-1-one (ODQ), an inhibitor of sGC, has been found to block the bradykinin mediated increase in ROS⁶². Lochner et al reported that ODQ significantly attenuated (to an intermediate level) the improved contractility that occurs in PC hearts⁵⁸.

Previous studies have shown that the cardioprotection of PC could be attenuated or abolished by antioxidants^79-81. Such a redox-based mechanism for cardioprotection of PC suggests that protein S-nitrosylation, a redox-reversible posttranslational protein modification, might play an important cardioprotective role in PC^{31, 82}. Recent studies have clearly shown that in addition to NO activation of sGC, NO can mediate protection by S-nitrosylation and a number of studies have found S-nitrosylated proteins that appear to be important in cardioprotection (see Table 1). An increase of protein S-nitrosylation has also been found to be involved in nitrite-mediated cardioprotection^{64, 83}. Before discussing the role of S-nitrosylation in cardioprotection we will first briefly discuss the mechanisms involved in regulating protein S-nitrosylation with an emphasis on mechanisms that would be important in I/R and cardioprotection.

Table 1.

Cardioprotective effects of S-nitrosylated protein

S-nitrosylated proteins	Cardioprotective effects
I. Mitochondrial proteins
Complex I	Inhibits activity, and attenuates ROS generation during I/R31, 64, 119
Cytochrome c oxidase	Inhibits its activity, and decrease oxygen consumption120
F1F0ATPase	Inhibits activity, and reduces ATP consumption during I/R31
Creatine kinase	Inhibits activity, and suppresses contractility under stress121
α-KGDH	Increases activity, might prevent oxidative inactivation upon I/R31
II. Ca2+ handling
L-type Ca2+ channel	Inhibits and reduces Ca2+ entry under oxidative stress11, 31, 122
RyR2	Hyponitrosylation increases SR Ca2+ leak and arrhythmogenesis123
SERCA2a	Increases SR Ca2+ uptake, reduces Ca2+ overload during PC31
III. Anti-apoptosis and anti-oxidative stress
Caspase-3	Inhibits activity, and elicits anti-apoptotic effect125
COX-2	Increases activity, and elicits preconditioning effect126
HIF-1α	Provokes stabilization in normoxic condition, against I/R injury97
NADPH oxidase	Inhibits activity, and suppresses ROS production127
Thioredoxin	Increases reductase activity, and elicits anti-apoptosis effect128
IV. Protein trafficking
Dynamin	Promotes endocytosis, and elicits anti-apoptosis effect131, 132
NSF	Inhibits exocytosis, and elicits anti-inflammatory effect133
GRK2	Decreases β-adrenergic receptor phosphorylation/desensitization134
β-arrestin 2	Promotes endocytosis and β-adrenergic receptor internalization135

Only proteins that have been shown to be involved in the cardiovascular system are included in this table.

IV. Mechanisms of Regulation of S-nitrosylation

A. SNO formation and stabilization

Protein S-nitrosylation occurs by endogenous NO-mediated nitrosylating agents such as dinitrogen trioxide (N₂O₃), by transition metal catalyzed addition of NO, or by transnitrosylation from low-molecular-weight SNO such as S-nitrosoglutathione (GSNO) or S-nitrosocystiene (CysNO)^{24, 84}. The S-nitrosylation of cysteine residue(s) on proteins depends on the precise conditions of NO, O₂, hydrophobicity, nucleophilicity, and redox surrounding the targeted thiol(s), which could change drastically during I/R. The actual redox state^{85, 86} and ultrastructural accessibility⁸⁷ of cysteine residue(s) under low oxygen tension such as hypoxia and ischemia might determine whether a particular thiol in a given protein is subjected to S-nitrosylation.

We are interested in mechanism(s) that might influence SNO in the setting of ischemia and PC. PC results in an increase in cytosolic Ca²⁺, which in turn activates constitutive NOS leading to an increase in NO for SNO formation. In addition, PC leads to formation of ROS and reactive nitrogen species (RNS), leading to a decreased intracellular reduced glutathione (GSH) pool, which might stabilize SNO formation by attenuating GSH-mediated trans-/denitrosylation^{88, 89}. The effects of oxygen on protein S-nitrosylation are complex and difficult to predict a priori. Autoxidation of NO yields N₂O₃ as a nitrosylating agent. This reaction is accelerated dramatically in a hydrophobic environment such as a lipid membrane^{90, 91}. PC involves intermittent ischemia and reperfusion, and reperfusion allows for generation of NO and nitrosylating agents such as N₂O₃. Furthermore, the stability of SNO appears to be favored by low ambient oxygen⁹². With increasing oxygen, the level of protein S-nitrosylation has been found to decrease while S-thiolation is promoted⁹³, and the intermediate thiyl radical is proposed to be involved in the decomposition SNO⁹⁴. Consistent with this concept, breathing low oxygen concentrations has been found to potentiate the ability of inhaled NO to increase cardiac albumin-SNO⁹⁵. Also, an increase of protein S-nitrosylation has been recently reported in a study of endothelial cells exposed to acute hypoxia⁹⁶. Taken together, the intermittent reperfusion can increase generation of NO/SNO, and the formed SNO will be stabilized by the low oxygen during the sustained ischemia. On reperfusion oxygen is restored and this would accelerate the breakdown of the SNO. This concept is supported by a study in which an increase in SNO occurred during PC while a decline in SNO occurred upon reperfusion¹¹. Thus, PC would provide an environment that might favor S-nitrosylation, including NO/SNO formation, ion content, and redox equilibrium.

B. Reversibility of SNO: transnitrosylation and denitrosylation

If SNO has a regulatory role in cell biology, it is expected that there are regulated mechanisms for removal of SNO, similar to the role of phosphatases in proteins phosphorylation. In fact, a number of mechanisms have been described leading to the removal of protein S-nitrosylation (see⁹⁷). The nitrosyl moiety could be removed by transnitrosylation (i.e., transfer of NO moiety between proteins) with low-molecular-mass thiols, such as GSH to form GSNO, or cysteine thiols to form CysNO⁸⁴. However, unless the GSNO is denitrosylated by GSNO reductase, the low-molecular-mass thiol formed SNO could also transnitrosylate some other proteins, in which case the overall protein SNO will not change.

Compared to transnitrosylation, protein denitrosylation plays an important role in eliminating SNO proteins therefore regulating cellular SNO levels. Some enzymatic systems have been shown to function as denitrosylase, such as GSNO reductase (GSNOR)^{97, 98}, the thioredoxin system⁹⁹, carbonyl reductase¹⁰⁰, Cu,Zn-superoxide dismutase¹⁰¹, and xanthine oxidoreductase¹⁰². Their physiological role as a critical regulator of SNO biology has been discussed in detail elsewhere^{103, 104}. Given that these denitrosylases are indeed coupled to cellular antioxidant redox enzymatic systems, the level of protein S-nitrosylation will be regulated by cellular redox state. Cellular redox changes during ischemia and reperfusion and this is likely to have important effects on protein SNO.

V. Detection of S-nitrosylated protein

Most methods used to detect S-nitrosylated proteins are based on the biotin switch method¹⁰⁵. Briefly in the biotin switch method, free thiol groups are blocked with a methyl methylating agent, the SNO groups are then reduced with ascorbate to free thiol groups that are then labeled with a sulfhydryl specific biotinylating agent. In this way the proteins that were originally S-nitrosylated are now labeled with a biotin and can be detected with an anti-biotin antibody. As is the case with any protein identification strategy, the set of SNO proteins that are detected depends to some extent on the details of the method employed. For example, if one were attempting to determine what proteins were phosphorylated by cardioprotection, one might employ different methods. Typically one will test for a specific pathway using a specific antibody such as an antibody that recognizes phospho-AKT for example. Therefore to use a similar approach to detect SNO proteins, following the biotin switch one can immunoprecipitate with antibodies for specific proteins and then immunoblot with anti-biotin to determine if the protein is S-nitrosylated. Another approach used to detect phosphorylated proteins in cardioprotection is to run a gel and then blot with an anti-phosphoserine/threonine or anti-phosphotyrosine antibody. Analogous to this, after the biotin switch method, one can run the entire extract on a gel and probe with an anti-biotin antibody. Anti-SNO specific antibodies (analogous to anti-phospho antibodies) are available and are currently being compared to results obtained with the biotin switch. Recently the use of two dimensional (2D) gel electrophoresis methods and mass spectrometry (MS) has been applied to identify phosphorylated proteins in cardioprotection¹⁰⁶. A similar approach can also be used to measure protein SNO^{31, 47, 107}. As is the case with phosphorylation, these different methods can result in the identification of many proteins. Although 2D gel methods provide unbiased information on a large number of proteins, high molecular weight and membrane proteins do not readily enter the 2D gel and are therefore often undetected with 2D methods¹⁰⁸. Also because of dynamic range issues, most of the proteomic methods (2D gels and MS) are biased towards detection of high abundance proteins. Thus, lack of detection of a low abundance protein in a 2D gel or MS analysis does not mean that the protein is not subjected to SNO modification. To date, many of the studies identifying SNO proteins have used 2D gels and therefore there are likely to be many low abundance signaling molecules, which are S-nitrosylated, that have not been currently identified. It should also be mentioned that in assessing the effect of a treatment on SNO, it is important to normalize the change in SNO to total protein, similar to what is routinely done in the evaluation of phosphorylation. SNO content can be detected and quantified by colorimetric (HgCl₂-coupled Griess reagent), chemiluminescent (Ozone-coupled photolysis), fluorescent methods (4,5-diaminofluorescein, DAN), and anti-CysNO-based methods.

VI. SNO and Cardioprotection

A. Cardioprotection results in SNO of many common proteins

Although NO is clearly established as a mediator of cardioprotection, there have been only a few of studies that have examined the role of protein S-nitrosylation in cardioprotection. PC has been shown to result in S-nitrosylation of a number of proteins³¹, as might be expected given the increase of NO formation in PC. Interestingly, although postconditioning and delayed PC both require NO, protein S-nitrosylation has not been examined under these conditions. NO donors such as GSNO have been shown to also increase S-nitrosylation of a number of proteins, many of which overlap with proteins undergoing S-nitrosylation during PC³¹. Estrogen has also been shown to be cardioprotective and it results in the S-nitrosylation of many proteins in common with PC and GSNO⁴⁷. A recent study has shown that mice lacking GSNOR, an enzyme involved in removing SNO from proteins, results in increased protein SNO and cardioprotection⁹⁷. Thus, considerable evidence is accumulating suggesting that an increase in SNO is associated with cardioprotection.

Table 1 lists a number of proteins that have been reported to show a change in SNO with different cardioprotective treatments. In this Table we have focused on acute protection in cardiomyocytes. An interesting general observation is that many of the proteins showing increased S-nitrosylation with cardioprotection are mitochondrial proteins. Possible reasons for this include the increased stability of N₂O₃ in the hydrophobic milieu of the mitochondria, which would favor S-nitrosylation, and the high level of reactive cysteines in mitochondrial proteins^{109, 110}. As NO is labile, the high level of SNO of mitochondrial proteins suggests a nearby source of NO, such as the endoplasmic/sarcoplasmic reticulum as discussed previously. In addition, protein S-nitrosylation in mitochondria can occur as a result of transnitrosylation from low-molecular-mass SNO such as GSNO⁹⁴.

As mentioned earlier, low abundance, signaling proteins are not well represented in the 2D proteomic screens that have been largely used to identify S-nitrosylated protein by cardioprotective treatments. Thus, it is very likely that there are other SNO proteins, some of which have not yet been identified, that play a role in cardioprotection. For example, phosphatase and tensin homolog deleted on chromosome 10 (PTEN) has been reported to be inhibited by S-nitrosylation^{111, 112} and inhibition of PTEN would be expected to increase signaling through the PI3K pathway, a pathway that has been shown to initiate cardioprotection^113-115. S-nitrosylation of PTEN has not been observed in heart in cardioprotection, but this might just reflect the low abundance of PTEN, which might be difficult to detect with 2D screening methods. Much additional work will be needed to fully elucidate the mechanism by which SNO modulates cardioprotection.

B. Is SNO an epiphenomena or involved in protection?

Consistent with a role for NO/SNO in protection, L-NAME blocks both the protection and SNO generation that are mediated by preconditioning or treatment with an estrogen agonist⁴⁷. Furthermore the finding that loss of GSNOR is cardioprotective supports a protective role for SNO⁹⁷. But to clearly establish the importance of S-nitrosylation of any protein or class of proteins in protection, it will be necessary to mutate the thiols that are S-nitrosylated and determine if this blocks protection. One of the challenges is that many proteins are S-nitrosylated in PC and other models of cardioprotection. Are one or a few of these proteins causally involved or are many or all important? If S-nitrosylation of multiple proteins and/or pathways is necessary, it may be difficult to sort out which proteins are important using a mutation strategy for a single protein. Also, it is likely that SNO can have beneficial and detrimental effects and that cardiorprotection depends on the balance.

C. Possible mechanism(s) for SNO mediated cardioprotection

General Mechanisms

Table 1 shows a large and growing list of proteins that are S-nitrosylated in the setting of cardioprotection. How might S-nitrosylation of these proteins lead to reduced cell death? Analogous to phosphorylation, S-nitrosylation can alter the activity of proteins. In addition to modulating the activity of proteins, it has been suggested that S-nitrosylation, either directly or by enhancing S-glutathionylation, can also protect reactive thiol groups from more irreversible oxidation, thereby allowing the activity of these enzyme to be restored more quickly following oxidative stress as might occur during I/R^{116, 117}. S-nitrosylation not only modifies proteins, leading to altered activity and protection against further oxidation, but it can also facilitate additional posttranslational modifications such as S-glutathionylation, which can alter protein activity^{117, 118}. As discussed by Baba et al, the S-nitrosylated form of aldose reductase, but not the reduced form can react with GSH to form S-glutathionylated aldose reductase and increase its activity¹¹⁸.

S-nitrosylation alters protein activity

S-nitrosylation of the mitochondrial complex I^{31, 64, 119}, cytochrome c oxidase¹²⁰, the F1F0ATPase (complex V)³¹, and creatine kinase¹²¹ inhibit their activities, whereas S-nitrosylation of α-ketoglutarate dehydrogenase (α-KGDH) increases its activity³¹. S-nitrosylation of the L-type Ca²⁺ channel has been reported to reduce the channel activity in the hearts under adrenergic stimulation¹¹, under oxidative stress during PC³¹, and with atrial fibrillation¹²². S-nitrosylation of the cardiac SR calcium release channel/ryanodine receptor (RyR2) has been shown to increase its open channel probability. Also the decreased S-nitrosylation of RyR2 in nNOS-KO hearts has been suggested to cause cysteine residues of RyR2 to be more oxidized under oxidative stress, leading to SR Ca²⁺ leaking and arrhythmogenesis¹²³. Cardiac SR Ca²⁺-ATPase (SERCA2a) has also been shown to be activated by an NO dependent modification. Adachi et al showed that NO dependent S-glutathionylation (occurring via peroxynitrite and therefore dependent on superoxide generation as well as NO and GSH) of SERCA2a resulted in an increase in SERCA2a activity¹²⁴. Sun et al also found S-nitrosylation of SERCA2a was associated with increased SR Ca²⁺-ATPase activity in PC hearts³¹. However, whether S-nitrosylation of SERCA2a per se increases activity or whether S-nitrosylation increases SERCA2a activity by enhancing S-glutathionylation remains to be sorted out.

Protein S-nitrosylation has also shown to play an important cardioprotective role by modulating the activity of proteins involved in apoptosis and oxidative stress such as caspase 3¹²⁵, cyclooxygenase-2 (COX-2)¹²⁶, hypoxia inducible factor 1α (HIF1α)⁹⁷, NADPH oxidase¹²⁷, and thioredoxin¹²⁸. Furthermore, recent studies suggest that NOS generates NO locally and regulates compartmentalized S-nitrosylation and protein trafficking in the cardiovascular system^{129, 130}. Several important proteins that regulate protein trafficking, such as dynamin^{131, 132}, N-ethylmaleimide-sensitive factor (NSF)¹³³, G protein-coupled receptor kinase 2 (GRK-2)¹³⁴, and β-arrestin 2¹³⁵, have been identified to be S-nitrosylated and S-nitrosylation of these proteins play an important role in regulating protein trafficking and signaling transduction.

Figure 1 provides a potential mechanism(s) by which protein S-nitrosylation might lead to acute protection as occurs in PC. PC results in S-nitrosylation and inhibition of the L-type Ca²⁺ channel, which would reduce Ca²⁺ entry into the myocytes during ischemia and early reperfusion. S-nitrosylation (by itself or by promoting S-glutathionylation) also results in activation of SERCA2a, which would further reduce cytosolic Ca²⁺ during ischemia and early reperfusion. In fact, enhanced expression of SERCA2a has been shown to be cardioprotective ¹³⁶. PC also results in S-nitrosylation and inhibition of the F1F0ATPase which would reduce ATP consumption by reverse mode of the F1F0 ATPase³¹. This would preserve ATP and would also reduce the mitochondrial membrane potential (Δψ) thereby reducing the driving force for Ca²⁺ uptake into the mitochondrial matrix. PC has also been shown to lead to S-nitrosylation and inhibition of complex I, which has been suggested to reduce ROS generation^{64, 119}. Taken together, the increase of protein S-nitrosylation during PC would be expected to lead to reduced mitochondrial Ca²⁺ and reduced ROS generation, which together would be expected to prevent the MPTP opening and reduce cell death, as MPTP opening has been shown to be an important determinant of myocardial I/R death^{137, 138}.

Figure 1. Potential mechanism(s) by which protein S-nitrosylation might lead to acute protection as occurs in PC.

PC results in S-nitrosylation and inhibition of the L-type Ca²⁺ channel, which would reduce Ca²⁺ entry into the myocyte during ischemia and early reperfusion. S-nitrosylation also results in activation of SERCA2a, thus further reducing cytosolic Ca²⁺ during ischemia and early reperfusion. PC also results in S-nitrosylation and inhibition of the F1F0ATPase which would reduce ATP consumption by reverse mode of the F1F0 ATPase. PC has also been shown to lead to S-nitrosylation and inhibition of complex I, which has been suggested to reduce ROS generation. Taken together, the increase of protein S-nitrosylation during PC would be expected to lead to reduced Ca²⁺ overload and reduced ROS generation, therefore preventing cell death during I/R injury.

SNO enhances S-glutathionylation

S-nitrosylation can also enhance the addition of glutathione to proteins through a modification known as S-glutathionylation^{117, 118}. Similar to S-nitrosylation, S-glutathionylation can alter protein activity and protect cysteines from irreversible oxidation^{117, 139}. For example, aldose reductase activity is increased by S-glutathionylation¹¹⁸, and S-glutathionylation of complex II has also been shown to enhance electron transfer¹⁴⁰. Interestingly, S-glutathionylation of complex II is decreased during reperfusion following ischemia¹⁴⁰; it might be interesting to determine if cardioprotection reverses this loss of S-glutathionylation. S-glutathionylation can be reversed, following the restoration of a reducing GSH/oxidized glutathione (GSSG) ratio. However, S-glutathionylation can also lead to irreversible protein posttranslational modification under excessive oxidative and nitrosative stress¹³⁹.

SNO protects against irreversible oxidation

The modification of thiols can be regulated by redox-related signaling in the cell including ROS/RNS³. In addition to S-nitrosylation and S-glutationylation, free thiols can be oxidized to sulfenic acid (SOH), sulfinic acid (SO₂H), disulfide bonds, or sulfonic acid (SO₃H). With increasing oxidation states, the modifications become irreversible and this typically leads to irreversible modification of the activity of the protein^{24, 141}. It has been suggested that S-nitrosylation and S-glutathionylation of thiol groups in proteins can protect these proteins against irreversible oxidative modifications^{116, 117, 142}. Furthermore, irreversible oxidation of thiols can block the physiological modification by S-nitrosylation or S-glutathionylation and thereby interfere with normal physiological signaling¹²⁴.

An example of how S-nitrosylation can protect against irreversible oxidation of a protein is provided by studies on protein-tyrosine phosphatase (PTP)¹⁴³. The activity of PTP can be regulated by thiol redox modification at an active cysteine residue, which is also susceptible to S-nitrosylation. A recent study revealed that the active site cysteine 215 was the primary cysteine residue susceptible to S-nitrosylation, and the S-nitrosylation at cysteine 215 protects PTP from subsequent irreversible oxidation. Thus, S-nitrosylation of cysteine 215 might prevent PTP from permanent inactivation caused by oxidative stress¹⁴³. It is suggested that S-nitrosylation might form a reversible “molecular shield” that prevents further thiol oxidation during excessive oxidative/nitrosative stress¹¹⁶.

S-nitrosylation can also control cellular redox through the regulation of some important redox-active enzymes, such as GSNOR^{97, 98}, thioredoxin^144-146, glutaredoxin¹⁴⁷, peroxiredoxin¹⁴⁸ and their corresponding reductase system. It has been suggested that NO can protect cells from oxidative stress, while loss or inhibition of NOS enhances oxidative stress^{123, 149}.

Possible role of S-nitrosylation in explaining discrepant effects of NO

As mentioned earlier, with high levels of NO or an increase in the ratio of ROS to NO the beneficial effects of NO are lost and NO can have detrimental effects. The variability in the amount of ROS might explain some of the discrepant findings in the literature. As just discussed S-nitrosylation of several proteins might reduce production of ROS during ischemia and early reperfusion, which in turn would reduce peroxynitrite generation and might contribute to the beneficial effects of NO. Thus it is possible that the depending on the levels of NO generation and ROS generation different proteins might be S-nitrosylated and this might alter the balance between cell protection and cell death.

D. Timing of S-nitrosylation and denitroyslation

Protein S-nitrosylation is a very labile modification and it appears that SNO-mediated cardioprotection induced by PC has a very short duration. This would be an ideal property for a cardioprotective signal. For example, long-term inhibition of the F1F0ATPase or complex I would likely have detrimental consequences. However, it appears that SNO is reversed early during reperfusion. Thus, SNO can modify activity of these proteins during ischemia and the start of reperfusion, but then with the loss of SNO during early reperfusion, the activity of these proteins returns to normal. The lability of SNO might also account for the loss of PC when the period between the PC and the sustained ischemia is extended.

E. Does protection depend on the level of SNO?

The scenario proposed in Figure 1 and data in several studies^{31, 47, 97, 119} suggest that cardioprotection is associated with an increase in SNO. Although a general increase in SNO appears to be associated with cardioprotection, it is likely an over-simplification to assume that S-nitrosylation of all proteins is cardioprotective. If one considers phosphorylation, depending on the target, phoshorylation can stimulate cell death as well as protection and therefore it is likely that S-nitrosylation can mediate cell death as well as protection. For example, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), a protein involved in glycolysis, which has recently been shown to have an additional function in regulating cell death, has been reported to initiate a cell death pathway in neurons¹⁵⁰. S-nitrosylation of GAPDH enhances its binding to Siah1 (seven in absentia homolog 1, an E3 ubiquitin ligase) and mediates its translocation to the nucleus and the initiation of cell death. Interestingly Sen et al have reported that GOSPEL, a protein which competes with GAPDH for binding to Siah,is also S-nitrosylated¹⁵¹. S-nitrosylation of GOSPEL promotes the binding of GOSPEL to GAPDH which reduces GAPDH binding to Siah1 thus preventing neuronal cell injury and death¹⁵¹. Thus it is proposed that an NO/SNO signaling will lead to S-nitrosylation of different proteins depending on the level of SNO generation or the activation of other signaling pathways. It will be important in future studies to identify (as in the case of GAPDH and GOSPEL) the details of the pathways by which SNO regulates protection as well as death.

Therefore, it is likely that S-nitrosylation of some proteins will promote rather than inhibit cell death, and the outcome depends on the balance. We are just beginning to understand the complex relationship between SNO, cell death and cardioprotection. Similar to phosphorylation, SNO signaling is compartmentalized. There is also likely to be cross-talk with other signaling pathways such as redox and phosphorylation, and ROS/RNS have been suggested to alter the balance between phosphorylation and S-nitrosylation of key signaling molecules^{1, 3, 152}. The association between increased SNO and cardioprotection^{31, 47, 97, 119} might suggest that under these cardioprotective conditions the beneficial effects of SNO predominate over the detrimental effects.

VII. Conclusion and Future Perspective

Protein S-nitrosylation, a reversible, redox-sensitive, thiol-based posttranslational modification, has been found to play an important role in a wide range of NO-mediated cardiovascular effects, including but not limited to mitochondrial metabolic regulation, intracellular Ca²⁺ handling, protein trafficking, and regulation of cellular defense against apoptosis and oxidative stress. As shown in Figure 2, S-nitrosylated proteins could elicit their regulatory effects and protect cells by (1) changing the structure and function of proteins due to SNO modification on the active thiol(s) and (2) shielding the modified cysteine residues (by S-nitrosylation or S-glutathionylation) from further irreversible modification under oxidative/nitrosative stress. Furthermore, depending on the localization of NO/SNO signaling, the level of protein S-nitrosylation, and/or interaction with other signaling pathways, the overall effect of protein S-nitrosylation can be protective or detrimental (e.g., GOSPEL/GAPDH). The feasibility of pharmacological preconditioning with SNO has provided an intriguing therapeutic strategy for protecting against myocardial I/R injury^{69, 97, 153-155}. Taken together, protein S-nitrosylation and S-nitrosylated proteins have emerged as an important contributor to cardioprotection.

Figure 2. Possible mechanisms of S-nitrosylated protein in cardioprotection.

S-nitrosylated proteins could elicit their regulatory effects and protect cells by (A) changing the structure and function of protein due to SNO modification on the active thiol(s); (B) shielding the modified cysteine residues (by S-nitrosylation or S-glutathionylation) from further irreversible modification (indicated as “X”) under oxidative/nitrosative stress; (C) signaling transduction via protein interaction (e.g., GOSPEL/GAPDH).

There are a number of important areas for future studies. It has been proposed that S-nitrosylation of cysteine residues might protect them from oxidation. It will be important to determine exactly which proteins and cysteine residues are S-nitrosylated under cardioprotection which would otherwise be oxidized. It will also be important to better understand the relationship between redox, ROS levels, NO levels and protection. As discussed, S-nitrosylation of some proteins will be beneficial whereas S-nitrosylation of others will be detrimental. It will be important to understand how S-nitrosylation of different proteins is regulated. Is the S-nitrosylation of different proteins controls soley by varying the amount of NO (probably not this simple)? We know there is localization of NOS and this likely contributes to the differences in protein S-nitrosylation. There are also subcellular differences in enzymes that denitrosylate proteins and this will also contribute to the regulation of compartmentalized subcellular SNO signaling. The past decade has provided a wealth of new information on how SNO modifies cell biology, but we have only just begun!

Non-standard Abbreviations and Acronyms

α-KGDH

α-ketoglutarate dehydrogenase

Δψ

mitochondrial membrane potential

cGMP

cyclic guanosine monophosphate

COX-2

cyclooxygenase-2

CysNO

S-nitrosocysteine

eNOS

endothelial NOS

GAPDH

glyceraldehyde-3-phosphate dehydrogenase

GOSPEL

GAPDH's competitor of Siah protein enhances life

GRK-2

G protein-coupled receptor kinase 2

GSH

reducing glutathione

GSNO

S-nitrosoglutathione

GSNOR

GSNO reductase

GSSG

oxidized glutathione

HIF1α

hypoxia inducible factor 1α

iNOS

inducible NOS

I/R

ischemia/reperfusion

KO

knockout

L-NAME

N-nitro-L-arginine methyl ester

MPTP

mitochondrial permeability transition pore

MS

mass spectrometry

N₂O₃

dinitrogen trioxide

NADPH

nicotinamide adenine dinucleotide phosphate

nNOS

neuronal NOS

NO

nitric oxide

NOS

NO synthase

NSF

N-ethylmaleimide-sensitive factor (NSF)

ODQ

1H-[1,2,4]-oxadiazole-[4,3-a]-quinoxalin-1-one

PC

preconditioning

PI3K

phosphatidylinositol 3-kinase

PKG

protein kinase G

PTEN

phosphatase and tensin homolog deleted on chromosome 10

PTP

protein tyrosine phosphatase

RNS

reactive nitrogen species

ROS

reactive oxygen species

RyR2

cardiac SR calcium release channel/ryanodine receptor

SERCA2a

cardiac SR Ca²⁺ ATPase

Siah1

seven in absentia homolog 1, an ubiquitin E3 ligase

sGC

soluble guanylyl cyclase

SNO

S-nitrosothiol

SR

sarcoplasmic reticulum

WT

wild type