From Sulfenylation to Sulfhydration: What a Thiolate Needs to Tolerate

See allHide authors and affiliations

Science Signaling  13 Mar 2012:
Vol. 5, Issue 215, pp. pe10
DOI: 10.1126/scisignal.2002943


There is a growing appreciation that oxidants such as hydrogen peroxide (H2O2) and gases such as nitric oxide (NO) and hydrogen sulfide (H2S) can act as modulators of various signaling pathways. Much of this signaling occurs through the modifications of specific, critical cysteine residues in target proteins. How such small, diffusible molecules (H2O2, NO, H2S) can achieve the required specificity is incompletely understood. Now, new findings provide considerable insight into these and related questions.

In 1903, Edwin Binney and C. Harold Smith introduced the first set of Crayola crayons (Crayola, Easton, Pennsylvania). That initial box contained only eight varieties, including the basic coloring necessities of blue, green, and yellow. Although at first ecstatic, children soon realized that the magic of the world could not be completely captured using a palette of eight colors. In response to this growing frustration, Binney and Smith decided in the late 1950s to create the magical 64-count Crayola set, complete with its own built-in sharpener, which is etched in many of our memories. As such, children could now reject a world composed of simple primary colors and instead express themselves using sepia, lavender, and raw umber. With that backdrop, it is perhaps no accident that this first cohort of children, exposed to the freedom of 64 exotic choices, would grow up to create acid rock, Haight-Ashbury, and the “summer of love.” Like popular culture, science too can often find itself severely restricted by its available tool kit. These limitations are particularly true for newer, emerging fields. Fortunately, the addition of two recent manuscripts substantially expands the palette of strategies available to one such emerging field, namely, reactive oxygen species (ROS)–mediated signaling. The first of these studies describes a new chemoselective probe that allows for the in situ detection of sulfenylated proteins (1). The authors demonstrated that this technique provides unrivaled spatial and temporal resolution to analyze intracellular redox signaling. In the second manuscript, the biological activity of the gas hydrogen sulfide is explored. The authors of this study demonstrated that endoplasmic reticulum (ER) stress triggers a large increase in hydrogen sulfide production leading to the sulfhydration of the tyrosine phosphatase PTP1B (2). Together, these two studies illustrate the increasing complexity and the rapidly expanding array of modifications, including both sulfenylation and sulfhydration, that can take place on critical cysteine residues.

Although initially viewed as a toxic by-product, ROS such as hydrogen peroxide are increasingly viewed as important regulators of intracellular signaling pathways (36). This reassessment began after some initial studies demonstrated that peptide growth factors such as platelet-derived growth factor (PDGF) (7) and epidermal growth factor (EGF) (8) produce a burst of ROS immediately after ligand binding and that this oxidant burst was not toxic, but rather is actually essential for downstream signaling. How does an increase in hydrogen peroxide contribute to a specific downstream intracellular response? For peptide growth factors, it was initially hypothesized that oxidants might target cysteine-dependent tyrosine and dual specificity phosphatases whose redox-dependent inactivation would in turn promote tyrosine phosphorylation (7). This conjecture was subsequently established by elegant studies that convincingly demonstrated ROS-dependent inactivation of phosphatase activity after growth-factor addition (911). Together, these observations established the general principle that the production of ROS was a regulated process and that specific protein targets existed in the cell whose biological activity was regulated by reversible oxidation and reduction.

The basis for much of redox-dependent signaling lies in the unique properties of certain reactive cysteine residues. Depending on the local protein environment, particular cysteine residues can have a pKa (where Ka is the dissociation constant) of 4 or lower, thereby existing predominantly as a thiolate anion (S) at physiological pH conditions. In this form, the cysteine residue acts as a strong nucleophile and, for the case of tyrosine phosphatases, the nucleophilic property of the active site cysteine is essential for its ability to attack phosphotyrosine substrates (12). When the thiolate anion encounters a ROS such as hydrogen peroxide, the residue can be easily oxidized to the sulfenic form (SOH). For the case of tyrosine or dual-specificity phosphatases, this oxidation results in enzymatic inactivation. Once the sulfenic form is generated, further oxidation can lead to the generally irreversible sulfinic (SO2H) or sulfonic (SO3H) species. Alternatively, the sulfenic form can be reversed, either through condensing with a second intramolecular or intermolecular cysteine residue or by reacting to form a mixed disulfide often with the abundant intracellular tripeptide (Glu-Cys-Gly) glutathione. These latter pathways eventually restore the thiolate anion and thereby provide a fully reversible pathway whereby protein activity is regulated by cysteine redox biochemistry.

Unfortunately, there have been problems in the methods used to detect the broad spectrum of proteins that undergo reversible sulfenylation, wherein a thiolate anion is converted to a sulfenic acid. Past methodologies have, in general, relied on the use of chemical agents such as N-ethylmalemide (NEM) or iodoacetamide (IAM) that act as thiol-specific alkylating agents. Although these agents readily react with the reduced (SH) form of a given thiol, they cannot react with the sulfenic, sulfinic, or sulfonic forms. Both NEM and IAM can be potentially modified by the addition of fluorophores, biotin, or radioactive isotopes (13, 14). Such derivations allow for the subsequent detection, purification, and sometimes quantification of proteins that have undergone cysteine oxidation. Although useful, such strategies suffer from various methodological shortcomings. First, these approaches generally rely on complete blocking of all free thiols with an alkylating agent (NEM or IAM). This constraint requires these procedures to be performed, in most cases, on cell lysates rather than intact cells. Thus, these techniques provide limited temporal resolution and no spatial resolution, and raise concerns as to whether a given cysteine residue was oxidized before or after cell lysis.

In contrast, the recently described new technology allows for the detection and visualization of sulfenylated proteins within intact cells (1). The authors described a new chemoselective probe they called DYn-2. This reagent builds on early work that demonstrated a selective reaction between 5,5-dimethyl-1,3-cyclohexanedione (dimedone) with protein sulfenic acids (15). In the new study, a membrane-permeable analog of dimedone was coupled to an alkyne-based reporter group. Binding of the dimedone moiety to a sulfenic acid residue could then be detected by an azide-bearing detection tag (Fig. 1A).

Fig. 1.

Signal transduction through the modification of reactive cysteine residues. (A) The basis for the chemoselective probe DYn-2 is shown. Target proteins that contain sulfenic acid (SOH) react with the dimedone moiety of the cell-permeable molecule DYn-2. As synthesized, DYn-2 has an alkyne-based reporter group that can be detected by an azide-bearing biotin of fluorophore detection tag. (B) The increased sensitivity of DYn-2 allows for the detection of the EGFR as a direct target of ROS signaling. Growth-factor binding stimulates ROS generation from the membrane-bound NADPH oxidase, Nox2. These oxidants modify cysteine residues in protein tyrosine phosphatases (PTPs) and the EGFR (Cys797) that both contribute to an increase in downstream signaling. (C) The reactive cysteine in the active site of protein tyrosine phosphatase PTP1B can be modified by hydrogen peroxide, hydrogen sulfide, and nitric oxide, leading to three separate chemical modifications.


The initial studies reported suggest that this strategy provides substantially more insight than previous approaches. First, as the authors demonstrate, cysteine oxidation can now be visualized in the intact cell by techniques such as confocal imaging. For instance, after EGF stimulation, the distribution of sulfenylated proteins and the nonphagocytic reduced form of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (Nox2) colocalize. As such, the source of oxidant production (Nox2) and the targets of ROS actions exhibit spatial overlap. Moreover, this strategy appears to be considerably more sensitive than previous approaches. As an example of this increased sensitivity, the authors demonstrate that the EGF receptor (EGFR) is directly oxidized by growth-factor addition. A specific cysteine residue (Cys797) was subsequently identified by mass spectroscopy. This particular cysteine residue is in close proximity to the adenosine triphosphate binding site and is actually the target of several drugs in development that seek to target tumors with increased EGFR activity (16). In this study, oxidation of Cys797 also resulted in increased EGFR activity, at least for low concentrations of hydrogen peroxide. Thus, it would appear that as our technology has improved, the link between growth factors, ROS production, and downstream tyrosine kinase signaling has gotten a bit more complex. Indeed, with the superior sensitivity of DYn-2, ROS act not only by inactivating tyrosine or dual-specific phosphatases, but also apparently by directly increasing tyrosine kinase signaling by activating upstream kinase activity (Fig. 1B).

The preceding results using DYn-2 involve the process of sulfenylation, wherein a reactive thiolate anion is converted to the sulfenic acid (-SOH) form. However, reactive cysteines in proteins can also undergo other important modifications. For instance, reaction with nitric oxide can result in the cysteine being modified to a nitrosothiol (SNO). It is now generally appreciated that the gas nitric oxide belongs to a family of signaling molecules often termed “gasotransmitters.” This family of diffusible mediators includes NO, CO, and H2S. The latter molecule is synthesized through at least four separate biochemical pathways, although the predominant pathway in most cells appears to involve the enzyme cystathionine γ lyase (CSE). To date, H2S appears to have pleiotropic physiological effects, although most studies have concentrated on its vasodilatory properties and its role as a modulator of inflammation (17). Indeed, CSE knockout mice exhibit marked hypertension (18) and increased cell death after treatment with inflammatory cytokines such as tumor necrosis factor–α (19).

Although the physiological effects of H2S have been established, only a relatively small number of direct molecular targets have been identified, which include mitochondrial proteins such as complex IV of the electron-transport chain (20), the KATP ion channel (21), and the p65 subunit of the transcription factor nuclear factor κB (19). In tissues with abundant H2S production, such as the liver, mass spectroscopic analysis suggests that as much as 10 to 25% of certain proteins exhibit sulfhydrated residues (22). For the case of glyceraldehyde phosphate dehydrogenase (GAPDH) and actin, two abundant liver proteins, sulfhydration leads to an increase in GAPDH activity as well as an increase in actin polymerization. Again, in these examples, reactive cysteines, often in the active site, have been sulfhydrated (-SSH), leading to the observed change in biological activity.

In this context, the new report by the Tonks laboratory suggests an expanding role for H2S-mediated signaling. The tyrosine phosphatase PTP1B was identified as a new target of H2S sulfhydration (2). When cells were exposed to the agent tunicamycin, an inhibitor of protein N-glycosylation and an inducer of ER stress, a CSE-dependent increase in H2S production was observed, which was accompanied by sulfhydration of PTP1B, with more than 40% of the phosphatase demonstrating covalent modification. Sulfhydration of PTP1B occurred on the active site cysteine (Cys215) and led to a corresponding reduction of PTP1B activity. This inactivation of PTP1B was subsequently linked to an increase in phosphorylation and activation of PERK, a major arm of the ER stress response.

These results raise a number of interesting points. First, it now appears that the active-site cysteine of PTP1B can react with hydrogen peroxide (9, 10), nitric oxide (23), and H2S (2). These modifications appear to lead to inactivation of the phosphatase, and it remains to be seen whether there are important biological differences, hierarchies, and physiological contexts for each of these different chemical modifications (Fig. 1C). Although previous examples have demonstrated that sulfhydration increases GAPDH activity and the rates of actin polymerization, the new study showed that sulfhydration of PTP1B led to phosphatase inactivation. Thus, like phosphorylation, sulfhydration can, depending on the context, lead to either an increase or decrease of target-enzyme activity. Finally, it will be interesting to see if H2S plays a major physiological role in the overall ER stress response. Given the links between ER stress and inflammation (24) and the immune-modulatory role of H2S (17), it is tempting to think there might be important in vivo connections. It is currently unclear why the concentration of H2S increased so dramatically when agents like tunicamycin are added to cells. The induction of the unfolded protein response leads to a disruption in protein translation, and given that the substrates for enzymes such as CSE are amino acids (cysteine and methionine derivatives), it is tempting to speculate that induction of the unfolded protein response and the subsequent inhibition of translation augments the intracellular pool of amino acids that, in turn, might regulate CSE activity.

The expanding array of covalently modulations that target reactive cysteine residues suggests that we are just beginning to understand the basis for the presumed intrinsic specificity of redox signaling. In many ways, this array of cysteine modifications mirrors the posttranslational diversity involving other amino acids such as specific lysine residues that can undergo competing reactions involving acetylation, sumoylation, or ubiquitination (25). One element that is clearly essential for continued progress is the ability to develop an expanding array of specific reagents and methodologies that allows for the detection and purification of targets of redox-dependent signaling and gasotransmitters. In that sense, strategies such as those discussed here, coupled with improving methods to detect various oxidant species (2628), should increase our overall understanding. Indeed, it has been only a decade and a half since it was observed that growth factors such as PDGF and EGF signal in part through ROS-dependent pathways (7, 8). In many ways, due to increasingly sophisticated techniques, the science has rapidly progressed beyond the stage of thinking in primary colors. It is as if a field, initially confined to a world of blue and green, has finally discovered the freedom that accompanies periwinkle and aquamarine.

References and Notes

Acknowledgments: I am grateful to members of my laboratory for helpful suggestions and to I. Rovira for assistance with the manuscript. Funding: This work was supported by NIH Intramural funds.

Stay Connected to Science Signaling

Navigate This Article