S-Nitrosylation Is Emerging as a Specific and Fundamental Posttranslational Protein Modification: Head-to-Head Comparison with O-Phosphorylation

See allHide authors and affiliations

Science's STKE  12 Jun 2001:
Vol. 2001, Issue 86, pp. re1
DOI: 10.1126/stke.2001.86.re1


Nitric oxide (NO) is a free-radical product of mammalian cell metabolism that plays diverse and important roles in the regulation of cell function. Biological actions of NO arise as a direct consequence of chemical reactions between NO or NO-derived species and protein targets. Reactions of NO with transition metals in target proteins have garnered the most attention to date as the principal mechanism of NO signaling; nonetheless, S-nitrosylation of protein Cys residues is rapidly moving to center stage in importance. In general, however, there has been a delay in adequate appreciation of the role of S-nitrosylation in biological signaling by NO. This lag is attributed to a poor understanding of the basis for selective targeting of NO to particular thiols, and methodological limitations in accurately quantifying this modification--recent breakthroughs in concepts and methods diminish these barriers. Here, we consider the wheres and whys of protein S-nitrosylation and its basis for specificity. Protein S-nitrosylation potentially represents a ubiquitous and fundamental mechanism for posttranslational control of protein activity on a par with that of O-phosphorylation.


The field of nitric oxide (NO) biology was born in 1987, when it was first revealed that this reactive free radical gas is a product of mammalian cell metabolism. Since that time, NO has been shown to function as a ubiquitous cell-signaling molecule with diverse physiological and pathophysiological roles (1, 2). In contrast to previously characterized signaling molecules that act by noncovalent binding to a receptor, NO and derived species exert their biological actions by chemical modification of targets--preferentially interacting with free radicals, transition metals, and thiol groups. The initial discovery of NO-mediated cell signaling focused attention on the reaction with metals; endothelium-derived NO was thought to elicit its physiological vasorelaxant activity by reaction with heme-iron in soluble guanylate cyclase of blood vessels, leading to increased 3′-5′ guanosine monophosphate (cGMP) accumulation. Recognition of the enormous importance of the NO-heme reaction in guanylate cyclase came in 1998 with an award of the Nobel Prize in Physiology or Medicine to pioneering discoverers Robert Furchgott, Louis Ignarro, and Ferid Murad.

Another reaction of NO (or more correctly, any of several NO-derived species) is S-nitrosylation of Cys residues in target proteins. This posttranslational protein modification was already recognized 10 years ago for albumin, by authors who portended its potential generality and importance in cell-signaling (3). S-nitrosylation has since been implicated in the control of a wide array of protein functions and cell activities, but research in this area has been hindered by methodological difficulties. A recent breakthrough by Jaffrey and Snyder (4) enables facile detection of proteins that are S-nitrosylated in vivo, providing an important new tool for identifying S-NO-modified proteins in biological systems.

In the mid-1950s, Krebs and Fischer made the seminal discovery that glycogen phosphorylase a and b are in fact the phosphorylated and native forms, respectively, of the same enzyme. Although the full importance of this discovery was not appreciated at the time, it served as the impetus for nearly a half-century of research that established reversible O-phosphorylation of protein Ser, Thr, and Tyr residues as having a "role in the regulation of essentially all cellular functions" (5). Tens of thousands of published articles have since revealed the intricacies of O-phosphorylation in the control of protein activity and attest to an evolutionarily ancient appearance of this mode of posttranslational protein modification. This knowledge and recent discoveries in the NO field provide insights as to how specificity for S-nitrosylation of mammalian cell proteins is achieved. Indeed, head-to-head comparison of the systems that evolved to mediate protein S-nitrosylation and O-phosphorylation reveal remarkable similarities as well as intriguing differences. Like the seeds planted by Krebs and Fisher in the field of protein phosphorylation, S-nitrosylation is anticipated to blossom and reveal itself to be a ubiquitous and fundamental posttranslational protein modification that influences nearly all key cellular functions.

Identification of Protein Targets and Cys Residues that Undergo S-Nitrosylation

Benefits of the Biotin Switch method

Until recently, only a handful of proteins had been identified as targets for in vivo S-nitrosylation; these targets include serum albumin (3), hemoglobin β-subunits (6), ryanodine-sensitive calcium release channels (7), N-methyl-D-aspartate (NMDA) receptors (8), methionine adenosyl transferase (9) and caspase-3 (10). A physiological role has been ascribed to S-nitrosylation in each case. More than 80 additional proteins have been asserted to undergo reversible regulation by S-nitrosylation based on results of in vitro studies. Nonetheless, proving that S-nitrosylation plays a biological role in regulating the activity of any given protein has been hindered by technical limitations: (i) there is no radioisotopic form of nitrogen or oxygen suitable for direct assessment of low-level NO incorporation into proteins, analogous to 32P for tracking phosphate incorporation, and (ii) protein S-NO groups are labile and subject to loss and rearrangement during sample preparation and analysis, causing identification of S-nitrosylated proteins to be tenuous and S-NO quantitation problematic. Protein S-NO can be lost by the action of intracellular reducing agents such as ascorbate (11), and reduced metal ions, especially Cu(I) (12). It can also be transferred to NO-accepting groups such as heme, glutathione, and other Cys-containing proteins. The inherent lability of protein S-NO groups makes any direct measurement prone to both false-negative and false-positive signals, arising from intra- and intermolecular NO group transfer reactions. These limitations in the analysis of protein S-nitrosylation contrast with O-phosphorylation, a chemical modification that is chemically resistant and readily quantifiable. Nonetheless, the recent methodological advance described by Jaffrey and Snyder (4) and applied in recent studies for the first time (13, 14) permits sites of S-NO modification in proteins to be effectively preserved and identified with high sensitivity in tissue extracts.

The Jaffrey and Snyder Biotin Switch method (4) for detection of S-NO in cell proteins is indirect, involving the substitution of a biotin group at every Cys-sulfur that had been modified by nitrosylation. In the first step, proteins in a tissue extract are denatured and all free SH groups are chemically blocked by treatment with methanethiosulfonate (MMTS), under conditions in which nitrosothiol or disulfide groups are unmodified. Subsequently, nitrosothiols, but not disulfides, are selectively reduced to free SH groups by treatment with ascorbate. The resulting newly formed SH groups are then biotinylated with a sulfhydryl-specific reagent, N-[6-(biotinamido)hexyl]-3′-(2′-pyridyldithio)propionamide (biotin-HPDP). The biotin thus introduced serves both as a placeholder that identifies thiol residues that had been S-nitrosylated in proteins at the time of cell lysis, and an efficient means of concentrating these proteins by affinity chromatography with immobilized streptavidin. This procedure minimizes the possibility of S-NO loss or rearrangement during tissue sample preparation. The biotinylated proteins obtained presumably represent most of the in vivo S-nitrosylated proteins, which may then be resolved by sodium dodecyl sulfate-poylacrylamide gel electrophoresis (SDS-PAGE) and identified either by mass spectrometry-based fingerprint analysis or immunoblotting with protein-specific antibodies.

Mechanisms of targeting S-nitrosylation

Application of this approach for S-NO detection has already borne fruit. The method has confirmed the previously held belief that intracellular proximity of an NO generator and acceptor provide one mechanism for selective targeting of NO to specific protein thiols in vivo (14). Although H-Ras and other members of the Ras family of small guanosine triphosphatases (GTPases) are present in neurons and can be activated by S-nitrosylation in vitro (15, 16), only the Ras family member with dexamethasone-inducible expression Dexras1 was found to be selectively activated and S-nitrosylated upon NMDA receptor-activated NO synthesis in cortical neurons (14). Investigations into the basis for the selective activation of Dexras1 revealed that this protein is indirectly tethered to the neuronal isoform of NO synthase (nNOS) by a linker protein termed carboxy-terminal PDZ ligand of nNOS (CAPON). The entire complex is anchored to the neuronal cell membrane by protein-protein interactions involving PDZ domains in nNOS and an integral membrane protein, PSD93/95 (17). The results confirm that NMDA-induced activation of Dexras1 in cells requires assembly with nNOS through the CAPON link, demonstrating that the spatial arrangement of the NO donor and NO recipient is a major determinant of S-nitrosylation specificity. This macromolecular assembly also contains the NMDA receptor (a PDZ domain-containing protein), which, like Dexras1, is S-nitrosylated in vivo (8). S-nitrosylation of the NMDA receptor leads to receptor desensitization. It is likely, but as yet unproven, that proximity to nNOS is key for S-nitrosylation of NMDA receptors and perhaps multiple other proteins included in PSD93/95 complexes in neurons, as well as of the analogous dystrophin complexes of skeletal muscle (18). The basis for specific NO modification of particular protein Cys residues awaits full definition. Although proximity to a source of NO may be a major factor for the specificity of S-nitrosylation, as discussed below, it is clear that some protein Cys residues are more susceptible to modification than others and that protein structural context does play a determining role.

Functional consequences of S-nitrosylation

The Biotin Switch method has also enabled the identification of a select set of sixteen proteins that are specifically S-NO modified in mouse brain by nNOS (13). This finding is remarkable in two regards: (i) it confirms that protein S-nitrosylation occurs physiologically, and (ii) it shows that, despite the existence of Cys residues in virtually all proteins, S-nitrosylation is restricted to only a small subset of mouse brain proteins and is absolutely dependent on nNOS gene expression. It will likely be discovered that, with physiological or pathophysiological alterations in neuronal activity (such as may occur during development and learning, or as a consequence of stroke, neurodegeneration, or oncogenesis), regional NO synthesis is altered and the distribution pattern of S-nitrosylated proteins is modified. Evaluation of the functional consequences of such anticipated changes in protein S-nitrosylation will be a difficult task, but could provide valuable insights into normal and pathophysiology.

Several S-nitrosylated brain proteins identified by Jaffrey et al. (13) had previously been shown to be S-NO-modified in vivo (NMDA receptors, ryanodine receptors, and glyceraldehyde-3-phosphate dehydrogenase), and approximately half had been suggested to be S-NO-modified based on in vitro findings. The broad spectrum of functions ascribed to proteins found to be S-nitrosylated (13) affects essentially all major cellular activities and is reminiscent of the diversity of protein activities controlled by O-phosphorylation (Fig. 1).

Fig. 1.

Diversity of function is apparent in the array of proteins discovered to be regulated by reversible S-nitrosylation. Reversible protein phosphorylation regulates all facets of cellular life. For each of the depicted functional classes of target proteins known to be regulated by phosphorylation, at least one example of a target protein has been revealed to be regulated by S-nitrosylation. This list of proteins shown here is far from complete; it represents only a subset of those reported to be modified by the indicated posttranslational modificatons.

The identification of novel S-nitrosylated proteins brings with it new challenges. In each case, molecular biologists will need to define physiological relevance, to specify the particular Cys residue(s) modified, and to confirm functionality of that Cys as an NO-sensor by mutagenesis, protein phenotype analysis, or other approaches. Structural biologists will need to define the molecular basis by which S-nitrosylation modulates protein activity. Recurring patterns of S-NO-induced conformational control of protein structure are anticipated to emerge from these studies. Presumably, the Biotin Switch method (4) can be extended to identify the relevant S-nitrosylated Cys residue(s) by SDS-PAGE followed by in-gel proteolysis and mass spectrometric identification of biotinylated peptides.

A tandem mass spectrometric method for identification of S-nitrosylation sites in a pure protein

An alternative method for identifiying the site(s) of S-NO-modified proteins has been developed that relies on an electrospray ionization tandem mass spectrometer (MS/MS) for identification of specific S-nitrosylated Cys residues in a pure protein digest (19). This method takes advantage of the lability of the S-N bond. Two mass analyzers are used sequentially (designated M1 and M2, respectively), separated by an argon gas collision cell. The collision cell is set to impart sufficient energy to rupture the S-N bond, but not enough to fragment other more stable chemical bonds. The two mass analyzers are set to monitor "neutral loss," that is, a mode in which only parental peptide ions are identified at M1 that lose precisely the uncharged mass of NO at M2 (30 dalton, 15 dalton, and 7.5 dalton for peptide ions with charges of +1, +2, and +3, respectively). The utility of this technique is that highly complex and poorly resolved spectra of peptide ions can often be simplified to reveal a single mass ion representing a lone S-nitrosylated peptide. An example of the application of this technique to detect the in vitro nitrosylation of argininosuccinate synthetase by S-nitrosoglutathione (GSNO) is shown (Fig. 2).

Fig. 2.

Electrospray ionization mass spectrometric (MS/MS) neutral loss spectrum of GSNO-treated human argininosuccinate synthase (huAS), revealing the identity of a single peptide that contains an S-nitrosylated Cys residue. A function is served by huAS in regenerating the NOS substrate arginine from the NOS product citrulline. Although huAS contains five Cys residues, only a single residue (Cys132) is revealed by this MS/MS analysis to be amenable to stable transnitrosation. Other studies have shown that S-nitrosylation of Cys132 causes dithiothreitol-reversible inhibition of huAS activity, thereby limiting access of NOS to its substrate. Notably, S-nitrosylation of AS was confirmed to occur in tissues from immunostimulant-treated rats in vivo (35) using the Biotin Switch method of Jaffrey and Snyder (4). The neutral loss MS/MS procedure is as follows: recombinant huAS protein was incubated with 100 µM GSNO at 25°C for 30 min. After buffer exchange with an ammonium bicarbonate-based buffer, trypsin was added (1:100) and the samples were incubated at 37°C for 4 hours, followed by MS/MS analysis. A control sample of AS was processed in an identical manner, but not exposed to GSNO (untreated). Tryptic peptides were injected into a Micromass QuatroII electrospray ionization tandem mass spectrometer, and neutral loss spectrum of 15 (the mass of NO, where charge = 2) and 30 (the mass of NO, where charge = 1) were acquired. (A) Neutral loss spectrum of 15 dalton GSNO-treated AS reveals a single detectable peak at m/z 648.3, which corresponds to the doubly-protonated peptide containing Cys132. (B) Neutral loss spectrum of 30 dalton again reveals a single peak, occurring at m/z 1527.8 and corresponding to the same Cys132-containing peptide as above, but in its singly protonated charge-state.

Determinants of Protein S-Nitrosylation

A comparison to O-phosphorylation

Delivery and removal of protein phosphate groups is engendered by the antagonistic actions of an array of kinase and phosphatase enzymes that require specific protein-protein interactions for recognition of protein targets. Phosphorylation is a highly ordered process, both spatially and temporally, providing integrated effector pathways for the amplification of specific receptor-mediated cell signals. Although mechanisms differ, specificity and order are also features of the S-nitrosylation system. A side-by-side alignment of steps that mediate reversible S-nitrosylation and O-phosphorylation, considering 3′-5′ adenosine monophosphate (cAMP)-dependent protein kinase (PKA) as a prototype to exemplify the latter, highlights the similarities and differences of these two signaling cascades (Fig. 3).

Fig. 3.

Side-by-side comparison of events that sequentially activate and deactivate both S-nitrosylation and O-phosphorylation in response to ligand binding to a GPCR. For O-phosphorylation, an adenylyl cyclase (AC)-coupled receptor is provided as an example. NO synthesis and presumably S-nitrosylation can also be triggered by some growth factors [for example, insulin and vascular endothelial growth factor (VEGF); not shown], which act through Akt-mediated phophorylation of eNOS at a defined Ser residue (Ser1179 in bovine eNOS) (36, 37). S-nitrosylation has been reported to inhibit Ca2+-induced activation of AC1 (38). The latter are only two examples of potential cross talk between phosphorylation and nitrosylation signaling cascades.

Like O-phosphorylation, S-nitrosylation is understood to contribute to a signal transduction system, involving signal recognition, amplification, and control of effector protein activities, that has ubiquitous impact on cell functions. Regulation of these systems can occur at multiple levels. Both phosphorylation and nitrosylation can be triggered by signaling molecules that bind to G protein-coupled receptors (GPCRs), causing activation of an effector enzyme which, in turn, produces a second messenger molecule. If the second messenger is intracellular Ca2+ [elevated by phospholipase C-β (PLCβ) activation], then Ca2+ will bind calmodulin and the complex can transiently bind to and activate any constitutively expressed nNOS or endothelial NOS (eNOS) present in the cell. If the second messenger is elevated intracellular cAMP (mediated by adenylyl cyclase activation), then cAMP will bind PKA regulatory subunits, releasing PKA catalytic subunits to selectively phosphorylate Ser and Thr residues in accessible proteins that contain the requisite consensus sequence.

Despite superficial similarities, a fundamental difference between these systems is that O-phosphorylation and dephosphorylation is conducted exclusively by enzyme-catalyzed reactions that require direct binding of enzyme and substrate, whereas S-nitrosylation and denitrosylation is accomplished in solution phase through chemical reactions involving reactive diffusible molecules. This distinction becomes partially blurred with the discovery that NO groups can be transferred between protein thiols by direct protein-protein interactions, exemplified by the discovery that S-NO-hemoglobin selectively transnitrosates anion exchange protein 1 (AE1) in red blood cell membranes (20). Nonetheless, the principal reliance on solution-phase chemistry endows S-nitrosylation with a sensitivity to changing conditions in the intracellular milieu that does not extend to the control of O-phosphorylation. For S-nitrosylation, the identity of protein targets and their degree of modification will be influenced in large measure by competing reactions with other species.

It is important to recognize that NO itself does not appear to be a biologically relevant nitrosylating agent; it is instead a precursor to several species that can effectively mediate protein S-nitrosylation. S-nitrosylation of any given protein will be influenced by four key factors: (i) rate of formation of the biologically relevant nitrosylating species, (ii) proximity of protein thiol targets to the nitrosylating species, (iii) inherent susceptibility of protein thiol targets to nitrosylation, and (iv) rate of S-NO-protein denitrosylation. Each of these four factors is considered below.

Formation and localization of biologically relevant S-nitrosylating species

The preeminent molecules that are thought to mediate protein S-nitrosylation in biological systems are the higher oxides of nitrogen, NO2 and N2O3 (21, 22), OONO- (peroxynitrite) (23), and transition metal-NO adducts (24); transnitrosation may be mediated by low molecular weight nitrosothiols such as GSNO (21), or by S-NO-proteins (3). The rate of formation of each of these species will predictably depend on concentrations of NO and companion precursor molecules, in accord with the law of mass action. Each of the above mentioned nitrosylating species will be formed preferentially under differing conditions and in distinct subcellular loci.

Consider, for example, NO2 and N2O3, species that are formed by reaction of NO and O2. These higher oxides of nitrogen are likely to be the major nitrosylating species in cells that produce moderate to high levels of NO. Whereas NO2 arises from reaction of two moles of NO with one mole of O2, N2O3 is formed by reaction of NO2 and NO. Accordingly, these nitrosylating agents are produced at very low rates when the NO concentration is low, and rates of their synthesis increase as a function of the square (for NO2) or cube (for N2O3) of rising intracellular NO concentrations. Because both O2 and NO are more soluble in lipid than water [partition coefficient of NO ≈ 20 (25)], these precursors concentrate in lipid membranes. Thus, formation of NO2 and N2O3 should preferentially occur in cell membranes, predictably a major site for protein S-nitrosylation in cells. This possibility is supported by results from Liu et al. (25), showing that the reaction of NO with O2 is accelerated 300-fold in membranes, compared with the surrounding medium. Not surprisingly, therefore, protein S-nitrosylation can be "catalyzed" by cell membranes. In addition to intracellular S-nitrosylation, the membrane locale may mediate nitrosylation of extracellular thiols -- this could provide a major route for the efflux of NO equivalents from cells. Further transnitrosylation reactions involving small molecules and proteins would serve to ferry NO equivalents from membranes to more distal sites, leading to possible S-NO gradients within and between cells.

A fascinating extension of the idea that hydrophobic sites catalyze protein S-nitrosylation is that a hydrophobic core within a protein can concentrate O2 and NO. This concentration promotes the formation and accumulation of a reactive species that causes S-nitrosylation of one or more Cys residues within the hydrophobic protein interior (Fig. 4) (26). The possibility that a protein could effectively catalyze its own S-nitrosylation (26) challenges the view that a protein Cys residue must be solvent-exposed to be a biological target for modification by NO. S-NO-Cys residues in a hydrophobic protein core may also accumulate as a result of a more limited access to species that mediate denitrosylation. These mechanisms may explain the accumulation and persistence of S-NO groups in the hydrophobic pockets of circulating serum albumin in vivo (27).

Fig. 4.

Schematic model to depict how hydrophobic amino acid residues can confer selectivity for S-nitrosylation of a Cys residue that lies within the hydrophobic core of a protein. The basis for specificity of S-nitrosylation in this case is that concentrations of the precursors, NO and O2, become enriched in the hydrophobic compartment (shaded area), where they can react to produce relatively high levels of the potent nitrosating species, N2O3. S-nitrosylation of a Cys in this hydrophobic compartment may be further enhanced due to limited access of hydrophilic reductants and low molecular weight thiols that would otherwise promote denitrosylation.

Under conditions of low NO production, N2O3 production would be minimal; accordingly, a greater fraction of NO would be expected to undergo addition reactions with transition metals. Resulting metal-NO complexes formed in this circumstance may become a dominant S-nitrosylating species and operate efficiently in the aqueous phase, rather than the membrane phase. These differences in the nitrosylating species and environment for nitrosylation may result in an altered array of S-nitrosylated proteins, compared to those targeted by N2O3 in membranes. In sites of enhanced superoxide production (such as mitochondria) or under conditions of oxidative stress, peroxynitrite could become a major fate for NO and serve as a biologically important nitrosylating species (23).

Proximity of a protein thiol target to an S-nitrosylating species

A well-recognized evolutionary strategy for the efficient control of protein phosphorylation involves tethering protein substrates to their cognate kinases and phosphatases (for review, see 28, 29). Targeting of PKA, and additional enzymes including kinase substrates, other kinases, and phosphatases, to the cytoskeleton is achieved by PKA-anchoring proteins (AKAPs). By effectively enhancing the concentrations of all the interacting species, these macromolecular assemblies increase the rapidity and specificity of phosphoregulation of protein activities in response to activation of cell surface receptors. As is the case for Dexras1 (14), and as will undoubtedly prove to be the case for many other proteins, molecular assemblies will have also evolved to compartmentalize a source of NO in a complex with key protein targets of S-nitrosylation. Directed S-nitrosylation may be achieved in this setting by presenting high local concentrations of a nitrosylating species in either the membrane phase or solution phase. It would not be surprising if NOSs are themselves targets for S-nitrosylation in vivo--this possibility warrants exploration inasmuch as it could represent an overlooked mechanism for modulating NOS activity.

Once a protein within a large protein assembly becomes S-nitrosylated, subsequent transnitrosation reactions may serve to sequentially deliver NO to neighboring proteins within the complex, potentially creating a cascade of spatially and temporally organized movements of NO through protein Cys residues. It is tempting to speculate that NO can thus modulate the function of an entire signaling complex in a highly organized manner, as may be the case for PSD93/95-containing complexes in neurons and dystrophin complexes in skeletal muscle, both of which contain nNOS. In endothelial cell membranes, eNOS is also found assembled into a complex with a large number of signaling molecules (30). Here, eNOS is bound to caveolin-1, the defining protein component of caveolae, which are specialized membrane patches enriched in proteins that serve in cell signaling [such as receptors, heterotrimeric GTP-binding proteins (G proteins), and effector proteins]. This localization of eNOS within caveolae may enable NO to function as a master controller that orchestrates cell signaling. In this regard, NO has been reported to depolymerize the caveolin scaffold (31) and generally disturb signaling through caveolae. The possible role of protein S-nitrosylation in NO-mediated disassembly of caveolae has not been explored.

Inherent susceptibility of protein thiol targets to nitrosylation

Proximity contributes to the specificity of a kinase-protein interaction, but is by no means sufficient. Amino acids flanking a given Ser, Thr, or Tyr define a consensus sequence that engenders selective recognition of sites for phosphorylation and dephosphorylation by a panoply of kinases and phosphatases. The question therefore arises as to whether there is a consensus motif for S-nitrosylation of protein Cys residues. Inasmuch as S-nitrosylation occurs nonenzymatically, a "consensus motif" for this activity would denote a sequence that facilitates the solution-phase transfer of NO to a protein thiol, rather than one that promotes the interaction of two proteins (as in the case of a protein substrate for recognition by its cognate kinase). Because NO group transfer to protein Cys residues involves reaction with thiolate anion (protein-S-), rather than protein-SH, the pKa of a Cys will affect its potential for S-nitrosylation. Notably, intramolecular hydrogen bonding to protein thiols will increase the nucleophilicity of sulfur and tend to activate it for nitrosylation. Accordingly, protein folding and allostery will influence the chemical susceptibility of Cys residues to be targeted for nitrosylation.

A model has been proposed based on analysis of the nitrosylation and denitrosylation of hemoglobin as it transitions from the R-state to the T-state (6, 32). β-Cys93 in hemoglobin is a site that can undergo alternate nitrosylation and denitrosylation in association with the allosteric transitions of hemoglobin from R-state to T-state, respectively. β-Cys93 has a basic residue on its NH2-terminal side (His) and an acidic residue on its COOH-terminal side (Glu); thus, the model suggests that the flanking acidic and basic residues play a key role in mediating NO addition and release (6, 32). Structural studies have revealed (6) that when hemoglobin is in R-state, the sulfur of β-Cys93 is opposed to the acidic residue (promoting nitrosylation), whereas transition to T-state moves the sulfur toward the basic residue (promoting denitrosylation). The chemical basis for this acid-base catalyzed NO group transfer, as described by Stamler et al. (32), is presented schematically (Fig. 5). An acid/base consensus sequence for S-nitrosylation is observed in many proteins where the modified Cys residue has been defined (32). The most important characteristic of this motif, XYCZ (where X = G, S, T, C, Y, or N; Y = K, R, H, D, or E; and Z = D or E), is thought to be the Asp or Glu residue following the target Cys. The absence of this motif in the primary sequence of a protein that is S-nitrosylated in vivo has been explained in the case of methionine adenosyl transferase by the appearance of a "consensus" acid-base motif for nitrosylation in the 3-dimensional structure created as a consequence of protein folding (33).

Fig. 5.

Acid-base catalysis of nitrosylation and denitrosylation of β-Cys93 in hemoglobin, as described (6). Transnitrosylation by an intermediary nitrosothiol (R-S-NO) is catalytically assisted by neighboring His and Asp residues, acting as base and acid, respectively. This mode of catalysis may serve a generic role in facilitating S-NO traffic to and from protein thiols. Other residues can fulfill the acid-base roles of His and Asp as served in hemoglobin (for example, Glu, Tyr, Lys, Arg). Catalytic residues may not be apparent upon consideration of the linear amino acid sequences flanking a target Cys in a given protein, but nonetheless may materialize in the 3-dimensional protein structure as a result of folding.

Rate of S-NO-protein denitrosylation

In the case of phosphorylation, phosphatase enzymes play a critical role in determining the phosphorylation status of kinase substrates; importantly, the phosphatases are subject to diverse regulatory control mechanisms. Similarly, the extent of S-nitrosylation in any protein will depend on the rate of denitrosylation as well as nitrosylation.

Denitrosylation of S-NO-proteins in cells can be accomplished by simple chemistry, wherein intracellular glutathione (GSH) or other intracellular thiols, including other protein Cys residues, act as acceptors and effectively remove nitrosyl groups via transnitrosation reactions. Additionally, light, ascorbate, and metal ions [Cu(I)] can promote S-NO-protein decomposition (11, 12). In this system, the rate of S-NO-protein decomposition would be modulated by changing levels of intracellular thiols; in other words, conditions that promote glutathione oxidation in cells would enhance steady-state levels of protein S-nitrosylation. This mechanism would put protein S-nitrosylation under the control of environmental changes that affect the intracellular redox milieu.

Formaldehyde dehydrogenase, an enzyme that is evolutionarily conserved in bacteria and humans, was recently shown to catalyze the selective reduction of GSNO at the expense of reduced nicotinamide adenine dinucleotide (NADH), forming glutathione disulfide and ammonia (34). Importantly, deletion of the gene encoding formaldehyde reductase in both mice and yeast resulted in increased levels of both intracellular GSNO and S-NO-proteins. This finding identifies the first biologically relevant mammalian denitrosylase and confirms that levels of GSNO determine intracellular levels of S-NO-proteins. Although this denitrosylase does not act directly on S-NO-proteins, its indirect actions may modulate the longevity of S-NO-proteins and thereby significantly influence NO bioactivities.


The signaling molecule NO is far more complex in its biological activities than "conventional" cell signaling molecules that adhere to the paradigm of selective receptor activation by lock-and-key fit. Because of the possibility for chemical reaction with almost any thiol-containing molecule in cells, in the past it was not evident that NO could modify protein thiols with sufficient specificity to function as an effective posttranslational regulator of protein function. Nonetheless, in recent years it has become clear that mechanisms have evolved to impose spatial and temporal constraints on the access of NO-derived nitrosylating species to appropriate protein targets in cells. On a broad level, numerous intracellular and extracellular factors can affect nitrosylation chemistry (including levels of thiols, ascorbate, metal ions, and light). Additionally, it is becoming apparent that the molecular environment of each Cys residue in a protein influences the efficiency of S-nitrosylation and denitrosylation; thus, allostery and protein-protein interactions can impose a further level of control.

It is quite possible that protein S-nitrosylation will ultimately come to occupy a position of importance akin to that now held by O-phosphorylation. In fact, the evident ubiquity of protein S-nitrosylation attests to its importance in a wide variety of tissues, including those central to disease processes. Until now, research on nitrosylation has been hindered by methodological difficulties that have not obstructed work on O-phosphorylation. The Jaffrey Biotin Switch method (4) offers a powerful new tool that largely overcomes these obstacles. Extension of the method to studies to an array of cell types, tissues, and organisms, in the contexts of pharmacological and genetic manipulations, promises a substantial surge in our understanding of when, where, and how protein S-nitrosylation contributes to the moment-to-moment control of protein activity and cell function.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
View Abstract

Stay Connected to Science Signaling

Navigate This Article