A Cell-Permeable Biscyclooctyne As a Novel Probe for the Identi fi cation of Protein Sulfenic Acids

Reactive oxygen species act as important second messengers in cell signaling and homeostasis through the oxidation of protein thiols. However, the dynamic nature of protein oxidation and the lack of sensitivity of existing molecular probes have hindered our understanding of such reactions; therefore, new tools are required to address these challenges. We designed a bifunctional variant of the strained bicyclo[6.1.0]nonyne (BCN-E-BCN) that enables the tagging of intracellular protein sulfenic acids for biorthogonal copper-free click chemistry. In validation studies, BCN-E-BCN binds the sulfenylated form of the actin-severing protein cofilin, while mutation of the cognate cysteine residues abrogates its binding. BCN-E-BCN is cell permeable and reacts rapidly with cysteine sulfenic acids in cultured cells. Using different azide-tagged conjugates, we demonstrate that BCN-E-BCN can be used in various applications for the detection of sulfenylated proteins. Remarkably, cycloaddition of an azide-tagged fluorophore to BCN-E-BCN labeled proteins produced in vivo can be visualized by fluorescence microscopy to reveal their localization. These findings demonstrate a novel and multifaceted approach to the detection and trapping of sulfenic acids. P thiol (sulfhydryl) groups are important mediators of redox signaling. In response to increases in cellular oxidation, thiols act as nucleophiles to reduce reactive oxygen species (ROS) in order to maintain redox homeostasis. In general, thiol oxidation leads to the formation of disulfides and oxidative intermediates, leading to structural changes of the protein thiols. In response to ROS, free thiols become oxidized to form sulfenic acids (sulfenylation), unstable moieties that can contribute to the formation of post-translational modifications that may vary depending on microenvironmental conditions and the duration of the ROS generated. Sulfenic acids may undergo further oxidation to sulfinic and sulfonic acids, or react with other cysteines or glutathione to form disulfides or thiosulfinates, respectively. These structural changes are diverse in nature and may impact protein function, activity, and localization. As such, ROS has significant influence on cell functions, acting as a second messenger to promote a diverse number of signaling events and subsequent responses including proliferation and differentiation, hypoxia, ferroptosis, and autophagy. Furthermore, there are clear ROS signatures in the etiology of several diseases; therefore, identifying the proteins subject to ROS regulation is critical to understanding how oxidative stress can influence these pathways and biological processes. Protein thiol sulfenylation is central to ROS physiology and pathophysiology as it is the first intermediate step in protein oxidation. As such, a number of methods have been developed to exploit the reactivity of the protein thiol group in order to “trap” proteins in their sulfenylated state to characterize the proteins susceptible to ROS modification. Typically, the use of 1,3-dicarbonyl-based probes, such as dimedone, has been employed to label sulfenylated proteins, with over 700 proteins having been identified to date. Although the use of dimedone and dimedone-like compounds is generally accepted, there are a number of limitations associated with these probes. Dimedone derivatives have slow rates of reaction toward sulfenic acids; in fact, there have been several recent efforts to overcome this problem by balancing the pKa and nucleophilicity of the enolates of beta-functionalized carbonyl compounds. There is also evidence suggesting that they may react with electrophilic sulfenamides. Furthermore, due to the limited range of applications with such probes, there is a lack of spatial and temporal data revealing the subcellular distribution of sulfenylated proteins. As a result, there is a need to develop better tools to overcome the challenges to examining this dynamic and labile redox modification. Strained cycloalkynes, such as bicyclo[6.1.0]nonyne (BCN), show great promise as sulfenic acid traps as they react with sulfenic acids with rate constants that are several-hundred-fold greater than other sulfenic acid probes. Rather than acting similarly to dimedone-based probes that exploit the electrophilicity of sulfenic acids (Supporting Information Figure S1A), or to probes that use the weak nucleophilicity of this group, the Received: August 25, 2016 Accepted: October 28, 2016 Published: October 28, 2016 Letters pubs.acs.org/acschemicalbiology © 2016 American Chemical Society 3300 DOI: 10.1021/acschembio.6b00742 ACS Chem. Biol. 2016, 11, 3300−3304 This is an open access article published under a Creative Commons Attribution (CC-BY) License, which permits unrestricted use, distribution and reproduction in any medium, provided the author and source are cited.

P rotein thiol (sulfhydryl) groups are important mediators of redox signaling. 1 In response to increases in cellular oxidation, thiols act as nucleophiles to reduce reactive oxygen species (ROS) in order to maintain redox homeostasis. In general, thiol oxidation leads to the formation of disulfides and oxidative intermediates, leading to structural changes of the protein thiols. 2 In response to ROS, free thiols become oxidized to form sulfenic acids (sulfenylation), unstable moieties that can contribute to the formation of post-translational modifications that may vary depending on microenvironmental conditions and the duration of the ROS generated. 3 Sulfenic acids may undergo further oxidation to sulfinic and sulfonic acids, or react with other cysteines or glutathione to form disulfides or thiosulfinates, respectively. 4 These structural changes are diverse in nature and may impact protein function, activity, and localization. As such, ROS has significant influence on cell functions, acting as a second messenger to promote a diverse number of signaling events and subsequent responses including proliferation and differentiation, 1,5 hypoxia, 6 ferroptosis, 7 and autophagy. 8 Furthermore, there are clear ROS signatures in the etiology of several diseases; 9 therefore, identifying the proteins subject to ROS regulation is critical to understanding how oxidative stress can influence these pathways and biological processes.
Protein thiol sulfenylation is central to ROS physiology and pathophysiology as it is the first intermediate step in protein oxidation. As such, a number of methods have been developed to exploit the reactivity of the protein thiol group in order to "trap" proteins in their sulfenylated state to characterize the proteins susceptible to ROS modification. Typically, the use of 1,3-dicarbonyl-based probes, such as dimedone, has been employed to label sulfenylated proteins, with over 700 proteins having been identified to date. 10,11 Although the use of dimedone and dimedone-like compounds is generally accepted, there are a number of limitations associated with these probes. Dimedone derivatives have slow rates of reaction toward sulfenic acids; in fact, there have been several recent efforts to overcome this problem by balancing the pK a and nucleophilicity of the enolates of beta-functionalized carbonyl compounds. 12,13 There is also evidence suggesting that they may react with electrophilic sulfenamides. 14 Furthermore, due to the limited range of applications with such probes, there is a lack of spatial and temporal data revealing the subcellular distribution of sulfenylated proteins. As a result, there is a need to develop better tools to overcome the challenges to examining this dynamic and labile redox modification.
Strained cycloalkynes, such as bicyclo[6.1.0]nonyne (BCN), show great promise as sulfenic acid traps as they react with sulfenic acids with rate constants that are several-hundred-fold greater than other sulfenic acid probes. 15 Rather than acting similarly to dimedone-based probes that exploit the electrophilicity of sulfenic acids (Supporting Information Figure S1A), 12,13 or to probes that use the weak nucleophilicity of this group, 16 the reaction involves a specific concerted syn addition to the alkyne (Supporting Information Figure S1B). In this study, we describe the development and validation of a novel cell-permeable bifunctional reagent, consisting of two linked BCN moieties (BCN-E-BCN), which makes use of the BCN reactivity for sulfenic acids to enable the detection of sulfenylated proteins.
BCN-E-BCN is comprised of two symmetrical strained cyclooctynes connected by a short ethylenediamine-derived linker, which allows protein sulfenic acids to be selectively tagged with a BCN group for copper-free click chemistry with azide-tagged reagents ( Figure 1). The symmetrical design of the bifunctional probe allows for straightforward two-step synthesis from commercially available compounds 17 (Supporting Information Figure S2) and exploits the most reactive and practical functionalities for sulfenic acid detection and click chemistry.
To validate the specificity of this probe, we incubated BCN-E-BCN with cofilin, a filamentous actin (F-actin) severing protein that has been shown to be oxidized by hydrogen peroxide (H 2 O 2 ). 18 Recombinant human cofilin was first incubated with 100 μM BCN-E-BCN, followed by a copper-free click reaction with azide-PEG3-biotin. BCN-E-BCN tagging was characterized by Western blotting using fluorescently labeled streptavidin to detect consequent protein-biotin conjugates formed by this reaction. BCN-E-BCN tagged cofilin in vitro, which could be increased above spontaneously oxidized basal levels with increasing H 2 O 2 concentrations ( Figure 2A). We previously determined that cofilin oxidation by H 2 O 2 occurs at cysteines 139 and 147, resulting in reduced binding and severing of F-actin by cofilin. 18 Mutation of these cysteines to alanine residues (C139/ 147A) decreased H 2 O 2 -induced BCN-E-BCN labeling in vitro ( Figure 2B), indicating that BCN-E-BCN tags cofilin at these cysteine sulfenic acids. Mass spectrometry analysis of recombinant cofilin incubated with BCN-E-BCN verified that cysteine 139 was a BCN-modified residue ( Figure 2C). In addition, we did not detect that BCN-E-BCN labeled the free thiol, sulfinic, or sulfonic forms of cofilin by mass spectrometry.
MDA MB 231 cells are triple negative human breast cancer cells that have been shown to have high basal levels of sulfenic Biotin can be attached using a copper-free click reaction between the BCN-E-BCN-protein conjugate and an azide-tagged biotin, which could be used to assist protein or peptide enrichment and subsequent analysis. (C) In the same way, the BCN-E-BCN-protein conjugates can be labeled using an azide-tagged fluorophore to enable visualization of the subcellular distribution of protein sulfenic acids.

ACS Chemical Biology
Letters acid formation. 19 To test the reactivity of BCN-E-BCN in a whole cell system, we incubated cultured cells with varying concentrations of BCN-E-BCN for 30 min at 37°C, followed by cell lysis and conjugation with azide-PEG3-biotin for detection. BCN-E-BCN was determined to be cell permeable and labeled sulfenic acids in a concentration-dependent manner ( Figure 3A), with efficient labeling at concentrations lower than typically used for dimedone based probes. 11 When compared to a commercially available biotinylated form of BCN (BCN-biotin), BCN-E-BCN demonstrated significantly greater efficacy in labeling sulfenylated proteins ( Figure 3B). Treatment with 50 μM H 2 O 2 induced a modest 50% increase in BCN-E-BCN labeling, as calculated by quantifying fluorophore-conjugated band intensities using an Odyssey CLx Imaging System ( Figure 3C). To attenuate this increase and demonstrate that the labeling responds to changes in redox state, we incubated the reaction in the presence of the antioxidant N-acetylcysteine (NAC; Figure 3C), which resulted in decreased BCN-E-BCN labeling. Furthermore, treatment with high H 2 O 2 concentrations decreased BCN-E-BCN labeling and correlated with increased peroxiredoxin trioxide (Prx-O 3 ) formation ( Figure 3D), suggesting that proteins were oxidized beyond the single oxygen-containing sulfenic acid and were forming BCN nonreactive sulfinic and sulfonic acids at these concentrations. Collectively, these data are consistent with BCN-E-BCN being specific for the sulfenic acid form of thiol oxidation.
We next examined BCN-E-BCN tagging of a previously characterized sulfenylated candidate protein by immunoprecipitation from whole cell lysates. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) is a key regulator of the glycolytic pathway and

ACS Chemical Biology
Letters is susceptible to thiol oxidation. 20−22 HEK 293 human embryonic kidney cell lysates were oxidized with 0.5 mM H 2 O 2 before labeling with 250 μM BCN-E-BCN. GAPDH was immunoprecipitated and BCN-E-BCN labeling determined by Western blotting. As shown in Figure 3E, BCN-E-BCN labeled GAPDH in HEK293 cell lysates, demonstrating the utility of this compound in identifying sulfenylated proteins from complex lysates.
To utilize the flexibility of the cyclooctyne ring in click chemistry reactions, we employed an azide-tagged fluorophore to label BCN-E-BCN-protein conjugates in fixed cells in order to visualize their subcellular localization. MDA MB 231 cells were fixed with 4% paraformaldehyde, tagged with 100 μM BCN-E-BCN followed by Alexa Fluor 488 azide and analyzed by fluorescence microscopy. We were able to visualize the spatial distribution of BCN-E-BCN tagged proteins, with the majority being predominantly cytoplasmic, with some mitochondrial, plasma membrane and weak nuclear signals detected (Figure 4).
It has been suggested that free thiols may undergo thiol-yne addition with cyclooctyne groups, such that pretreating proteins with iodoacetamide (IAM) would be expected to reduce nonspecific binding. 23 However, we found that IAM pretreatment of MDA MB 231 cells did not markedly affect BCN-E-BCN labeling (Figure 4), consistent with the sulfenic acid specificity of the reaction and in line with previous findings. 15 It should be noted that BCN should not be regarded as an electrophilic reagent for sulfenic acids but reacts via a concerted syn addition of the sulfenic acid to the alkyne: the C−S bond and the C−H bond are formed in the same step via a cyclic transition state, analogous

ACS Chemical Biology
Letters to a pericyclic reaction (Supporting Information Figure S1B). This type of reactivity is not possible for protein thiols or sulfenamides, and our results indicate that electrophilic reaction with thiolates or thiyl radicals is unlikely to contribute to the observed BCN-tagging of proteins under physiological conditions. Protein sulfenylation has received much recent interest because of its dynamic and complex role in ROS mediated cell signaling. However, a major limitation is a lack of convenient tools to measure this post-translational modification. Existing reagents are compromised by low reaction rates for sulfenic acids or have limited utility in biochemical and cell biological applications. Here, we now report the development and characterization of a novel bifunctional cyclooctane compound, BCN-E-BCN, which can rapidly label sulfenylated proteins in cultured cells and be used in a number of diverse applications. We demonstrate that BCN-E-BCN is cell permeable, has high reactivity compared to other BCN derivatives, and can facilitate a range of experimental approaches via copper-free click chemistry reactions that allow for labeling of BCN-tagged proteins with a variety of readily detectable moieties.