Superoxide Anion Radical Modulates the Activity of Ras and Ras-related GTPases by a Radical-based Mechanism Similar to That of Nitric Oxide*

Ras GTPases cycle between inactive GDP-bound and active GTP-bound states to modulate a diverse array of processes involved in cellular growth control. The activity of Ras is up-regulated by cellular agents, including both protein (guanine nucleotide exchange factors) and redox-active agents (nitric oxide (NO) and superoxide anion radical (\batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document})). We have recently elucidated the mechanism by which NO promotes guanine nucleotide dissociation of redox-active NKCD motif-containing Ras and Ras-related GTPases. In this study, we show that guanine nucleotide dissociation is enhanced upon exposure of the redox-active GTPases, Ras and Rap1A, to \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document} and provide evidence for the efficient guanine nucleotide reassociation in the presence of the radical quenching agent ascorbate to complete guanine nucleotide exchange. In vivo, guanine nucleotide reassociation is necessary to populate Ras in its biologically active GTP-bound form after the dissociation of GDP. We further show that treatment of the redox-active GTPases with \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document} releases GDP in form of an unstable the oxygenated GDP adduct, putatively assigned as 5-oxo-GDP. 5-Oxo-GDP was not produced from either the C118S or the F28L Ras variants upon the treatment of \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\) \end{document}, supporting the involvement of residues Cys118 and Phe28 in \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\mathrm{-mediated}\) \end{document} Ras guanine nucleotide dissociation. These results indicate that the mechanism of \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(\mathrm{O}_{2}^{{\bar{{\cdot}}}}\mathrm{-mediated}\) \end{document} Ras guanine nucleotide dissociation is similar to that of NO/O2-mediated Ras guanine nucleotide dissociation.

The p21 Ras proto-oncoprotein (Ras) is a founding member of the Ras superfamily of GTPases and plays an important role in a number of biological processes involved in cellular growth control, including cell growth, differentiation, and apoptosis (1,2). Ras cycles between an inactive GDP-bound and active GTPbound form to modulate association with regulators and downstream targets. High affinity interactions with downstream effectors are achieved via interaction with GTP-bound Ras rather than the GDP-bound form of Ras, which in turn leads to effector activation, stimulation of downstream signaling pathways, and a plethora of biological responses (3)(4)(5). Given the importance of Ras in various growth control processes, the population of Ras in its biologically active GTP-bound or inactive GDP-bound form is critically regulated by cellular factors. For example, the intrinsically slow rate of Ras guanine nucleotide dissociation and GTP hydrolysis is modulated by regulatory proteins and free radicals. GTPase-activating proteins down-regulate Ras activity by stimulating the intrinsically slow rate of GTP hydrolysis to populate Ras in its inactive GDP-bound form, whereas guanine nucleotide exchange factors, reactive nitrogen species, and reactive oxygen species (ROS) 1 (6 -18) up-regulate Ras function by promoting guanine nucleotide exchange (GNE) to generate the active GTP-bound state of Ras in vivo. A number of reactive species may contribute to the regulation of Ras activity, including the reactive nitrogen species nitric oxide (NO), nitrogen dioxide ( ⅐ NO 2 ), and dinitrogen trioxide (N 2 O 3 ) as well as ROS such as the superoxide anion radical (O 2 . ) and hydrogen peroxide (H 2 O 2 ).
NO is the best characterized redox modulator of Ras activity as NO has been shown to promote Ras GDP dissociation in vitro, GTP binding to Ras in vivo, and stimulation of pathways downstream of Ras (9 -14, 19, 20). The target site of NO modification and NO-mediated guanine nucleotide dissociation on Ras is Cys 118 , which is located in the nucleotidebinding NKCD motif (11,12,20,21). Our recent studies indicated that ⅐ NO 2 , a reaction product of NO with O 2 , reacts with the Ras Cys 118 thiol to produce a Ras Cys 118 -thiyl radical (Ras-S 118 ⅐ ) intermediate (22). Once generated, Ras-S 118 ⅐ induces a series of radical reactions involving the Phe 28 side chain and the Ras-bound GDP base to cause the conversion of Ras-bound GDP into a Ras-bound GDP neutral radical (G ⅐ -DP) (23). Ras-bound G ⅐ -DP further reacts with an additional ⅐ NO 2 to produce a GDP-NO 2 adduct, 5-nitro-GDP. 5-Nitro-GDP can then be degraded into NIm-DP by decarboxylation of the 5-nitro-GDP C 6 atom. This NO/O 2 -mediated radicalbased process perturbs Ras GDP binding interactions, resulting in the dissociation of GDP from Ras in the form of 5-nitro-GDP (23).
In addition to NO, ROS have also been shown to activate GNE on Ras and Ras-related GTPases (8,15). Early evidence for ROS-mediated activation of Ras was obtained in studies in which various oxidative agents (i.e. H 2 O 2 and hemin) could enhance GNE on Ras in vitro and enhance Ras-mediated mitogen-activated protein (MAP) kinase activity in Jurkat cells (15). Moreover, activation of NF-B activity by these oxidative agents was blocked by expression of a dominant negative Ras mutant, suggesting that direct activation of Ras may be a central mechanism by which a variety of oxidative redox-stress stimuli transmit their signals to the nucleus (15). Recently, Adachi et al. (8) showed that H 2 O 2 -induced modification of Ras at Cys 118 facilitates activation of p38 and Akt but not extracellular signal-regulated kinase (ERK) activation in rat vascular smooth muscle cells.
The enzymes NADPH oxidase (24), xanthine oxidase (25)(26)(27), and nitric oxide synthase (28 -33)  . production from NADPH oxidase (18). Intriguingly, the redox-active NKCD motif-containing Rap1A, a member of the Ras subclass of GTPases, has been observed to colocalize with NADPH oxidase (40). Although Ras and Rap proteins commonly regulate activation of the MAP kinase cascade in some cell types (41), they regulate distinct cellular effectors. In fact, Rap GTPases are best characterized as critical regulators of integrin-mediated cell adhesion; however, their mechanism of action is poorly understood (42,43).
Since the redox-active Rap1A GTPase colocalizes with NADPH oxidase (40) and is sensitive to NO (19,20), it is possible that Rap1A activity may also be regulated  (44,45). The final proteins were Ͼ95% pure as determined by SDS-PAGE. The protein concentration was determined by the Bradford method (46). Xanthine oxidase was purchased from Sigma. Bovine liver catalase (40,000 units/mg) and bovine liver superoxide dismutase (50,000 units/mg, Cu,Zn-superoxide dismutase) were purchased from Sigma and further purified using a size-exclusion column (Sephacryl S-100, 2.5 ϫ 60 cm) prior to use.

Kinetic Measurements of Guanine Nucleotide Exchange on Ras and
Rap1A-Excess amounts of xanthine oxidase, its substrate xanthine, or catalase do not interfere with either the fluorescence intensity of the mant-labeled 2Ј-(or-3Ј)-O-(N-methylanthraniloyl)guanosine 5Ј-diphosphate (mant-GDP) or the GDP binding properties of Ras and Rap1A. Therefore, we employed a mant-GDP fluorescence assay (50,51) to conduct kinetics studies of O 2 . -mediated Ras and Rap1A guanine nucleotide dissociation. Wild-type (WT) Ras, the Ras variants C118S and F28L, and Rap1A were preloaded with mant-GDP as described previously (50,51). A standard assay mixture for the fluorescence measurements consisted of 10 mM xanthine, 20 mM GDP, 50 mM NaCl, 5 mM MgCl 2 , and 0.5 mM EDTA in ammonium acetate buffer (10 mM, pH 7.5).
Once the GTPases was loaded with mant-GDP, xanthine oxidase was added, and the dissociation of mant-GDP from the GTPase was measured as a change in the fluorescence intensity over time using a LS50B PerkinElmer Life Sciences fluorimeter. Apparent second-order rates of O 2 . -mediated Ras and Rap1A GDP dissociation ( app O2 . k off ) were determined by fitting the data to a simple exponential decay.
We also employed 3 H-radiolabeled GDP to assess the rates of Ras GDP . The apparent WT Ras GDP association rates in the absence and presence of ascorbate were determined to be 0.14 ϫ 10 3 and 1.61 ϫ 10 3 s Ϫ1 , respectively, by fitting the data to a simple exponential association. Since the association process observed in these assays could include the sequential process of Ras GDP dissociation followed by Ras GDP association, the conversion of the apparent WT Ras GDP association rates into the corresponding true values could not be easily conducted. Hence, the estimated values for the apparent WT Ras GDP association rates were not converted into true kinetic values. The data presented in this figure are mean values of triplicate measurements. Standard errors of each data point are Ͻ48%, and the regression values associated with the fit were r 2 Ͼ 0.7595. The ␤-errors associated with entire assay process were not included to account for the standard errors and r 2 values. K D and true O2 . k off (Fig. 1)  Preparation of Ras and Rap1A-released Nucleotide Nitration Product-An aliquot of Ras or Rap1A protein was introduced into the xanthine oxidase-containing assay solution, and the mixture was incubated for 300 s. The Ras or Rap1A-released nucleotide oxygenation product(s) was separated from Ras or Rap1A protein using a centricon (10-kDa cut off, Millipore).
Mass Spectrometry Analysis-Ras or Rap1A-released nucleotide oxygenation product(s) was acidified with formic acid prior to sample application since MS analysis requires the samples to be positively charged. Thus, only charged ions (i.e. [MϩH] ϩ ) were observed so that the molecular masses determined by MS are 1 Da higher than the molecular masses for the same chemicals at neutral pH (i.e. pH 7.5). Briefly, an aliquot of the sample was dissolved in 50% methanol:0.1% formic acid and analyzed within 30 min by nanospray mass spectrometry on an ABI QSTAR-Pulsar QTOF MS spectrometer with a nanoelectrospray source (Applied Biosystems Division, PerkinElmer Life Sciences). The sample (ϳ1 l) was loaded into a Protana distal coated nanospray needle (Protana, Odense, Denmark), and spectra were acquired over a mass range of 400 -500 in MS mode.

RESULTS
In addition to NO (8 -14) (22). In addition to the fluorescence-based measurements, a radioactive-based true O2 . k off measurement was performed (Fig. 2). The value of true O2 . k off (ϳ3.9 ϫ 10 Ϫ3 s Ϫ1 ) for Ras GDP, determined using the radioactive assay, is similar to that determined from the fluorescence-based measurement (Fig. 1). The fluorescence method employing mant-GDP was also applied to determine the value of O2 .
K D (11.3 mM, Fig. 3 Fig. 2). A slow increase in radioactivity, due to the rebinding of [ 3 H]GDP to Ras, was observed (Fig. 2), suggesting that, after a prolonged time period (i.e. 500 s), true O2 . k on slowly increases. A possible explanation for these results is that the fast true O2 . k off and slow true O2 . k on gradually decay over time, and thus, the intrinsic rates of Ras guanine GDP dissociation ( int k off ) and GDP association ( int k on ) slowly recover (51). However, the rate of Ras GDP association is greatly enhanced by the addition of a radical scavenger, 1 mM ascorbate (Fig. 2) (58). If H 2 O 2 facilitates Ras guanine nucleotide dissociation, the addition of catalase to the xanthine oxidase assay should inhibit the rate of Ras guanine nucleotide dissociation. The addition of an excess amount of catalase (ϳ1,000 units) did not impede but rather slightly facilitated the rate of the xanthine oxidase-mediated Ras guanine nucleotide dissociation (Fig. 1). It is difficult to rationalize the minor stimulation of the catalase-mediated xanthine oxidase-coupled Ras guanine nucleotide dissociation rate as multiple factors may be involved. For example, removal of H 2 O 2 using catalase may enhance the release of H 2 O 2 (xanthine oxidase end product) from xanthine oxidase, which in turn may facilitate the turnover rate of  It has previously been shown that a protein thiol (i.e. bovine serum albumin Cys 34 ; BSA-S 34 ) can be oxidized by H 2 O 2 to produce a protein sulfenic acid adduct (i.e. BSA-S 34 -OH) (59). If the Ras Cys 118 thiol oxidizes to sulfenic acid (Ras-S 118 -OH) due to the presence of H 2 O 2 , the formation of a Ras sulfenic acid adduct does not facilitate GDP dissociation from Ras (Fig. 1). Consistent with these observations, we have previously shown that modification of the Ras Cys 118 thiol by NO to produce S-nitrosylated Ras (Ras-SNO) does not promote NO/O 2 -mediated Ras guanine nucleotide dissociation (22,23).
Characterization . , and ⅐ NO 2 can commonly react with Ras Cys 118 thiol to produce Ras-S 118 ⅐ . A detailed mechanism for Ras-S 118 ⅐ -initiated Ras guanine nucleotide dissociation has been proposed (23). In this mechanism, electron transfer from the Ras-bound GDP guanine base to Ras-S 118 ⅐ via Phe 28 side chain produces a Ras-bound G ⅐ ϩ -DP, which is subsequently converted into Ras-bound G ⅐ -DP. This radical reaction process is likely to perturb the aromatic-aromatic interaction between the Phe 28 side chain and the Rasbound guanine base as well as interactions between Ras and the guanine base of bound GDP, resulting in the release of GDP from Ras. Consistent with this premise, the use of Ras variants containing mutations in the redox-active Ras residues Cys 118 and Phe 28 (23) abolish O 2 . -mediated guanine nucleotide dissociation from Ras ( Fig. 1 (23). The original study (23) that led to our proposed radical-based mechanism was conducted in the presence of NO/O 2 to generate ⅐ NO 2 , where we observed that Ras-bound G ⅐ -DP can react with ⅐ NO 2 to produce 5-nitro-GDP, which is subsequently decarboxylated to produce NIm-DP as an end product. The end products of O 2 . -mediated guanine nucleotide dissociation of Ras and Ras-related GTPases, however, are likely to differ from those produced by ⅐ NO 2 -mediated guanine nucleotide dissociation of Ras and Ras-related GTPases (5-nitro-GDP) since reaction of the proposed G ⅐ -DP with O 2 . will generate nucleotide oxygenation products. However, as suggested in our previous study (23), since the C 5 site of Ras-bound G ⅐ -DP is exposed to solvent (hence O 2 . is accessible to the C 5 site of Ras-bound G ⅐ -DP), whereas the C 8 site of Ras-bound G ⅐ -DP is not solventaccessible (63-68), 5-oxo-GDP is likely to be the dominant product formed over 8-oxo-GDP (Fig. 4). The proposed reaction process is shown in Fig. 4, where reaction of the G ⅐ -DP C 5 with O 2 . produces 5-oxo-GDP (step i). 5-Diamino-4O-imidazolone ri-bose diphosphate (DIm-DP) is formed after decarboxylation of 5-oxo-GDP (step iv), which can be further degraded into 5-amino-oxazolone ribose diphosphate (AOn-DP) (69) (steps v1 and vi), 5-amino-4O-imidazolone ribose diphosphate (AIm-DP) (step v2), and/or 5-imino-4O-imidazolone ribose diphosphate (IIm-DP) (step v3). Adducts released from Ras upon exposure to xanthine oxidase were analyzed by MS. Two main peaks with molecular masses of 415.26 and 417.26 Da were observed that correspond to end products of O 2 . -treated Ras (Fig. 5A). Given that the predicted molecular masses of AIm-DP and IIm-DP are 415. 26 and 417.26 Da, respectively, we putatively assigned the two Ras-released GDP-O 2 derivatives as AIm-DP and IIm-DP. As anticipated, we were unable to detect the MS peak corresponding to AOn-DP since our MS sample analysis was performed within 30 min after the sample was prepared, and the formation of AOn-DP from DIm-DP requires a prolonged incubation time (Ͼ24 h) in the dark (69). We were also unable to detect MS peaks corresponding to the molecular weight of 8-oxo-GDP as well as possible degradation products of 8-oxo-GDP. The inability to detect these adducts is consistent with our premise that the C 8 site of G ⅐ ϩ -DP is not exposed to solvent so that the reaction product of the G ⅐ -DP C 8  . to produce GTPase-S 118 ⅐ . (b) The GTPase-S 118 ⅐ then withdraws an electron from the guanine nucleotide base to produce G ⅐ ϩ -DP. Similar to the mechanism described for ⅐ NO 2 -mediated Ras guanine nucleotide dissociation (23), the Phe 28 side chain may serve as an electron conduit for this process. G ⅐ ϩ -DP is thus expected to be formed and converted to G ⅐ -DP by elimination of H ϩ from the N 1 atom of G ⅐ ϩ -DP. The process disrupts key hydrogen bond interactions as well as the n-interaction between the GTPase and its ligand nucleotide. (c) The Ras-bound G ⅐ -DP can then react with O 2 . to produce 5-oxo-GDP, releasing 5-oxo-GDP from the GTPase. Depending on the experimental conditions, the Ras-released 5-oxo-GDP is further degraded into oxygenated-nucleotide derivatives, such as AIm-DP or/and IIm-DP. We also show that O 2 . -mediated Ras signaling is unidirectional (i.e. guanine nucleotide dissociation but not association is enhanced) in the absence of a radical scavenging agent. O 2 .
effectively dissociates GDP from Ras in the form of 5-oxo-GDP but does not enhance association of GDP. Like NO/O 2 -mediated Ras guanine nucleotide dissociation (23), treatment of Ras-GDP with O 2 . promotes GDP dissociation to produce a nucleotide-deficient radical form of Ras (apo-Ras O2 . ) that cannot efficiently associate with guanine nucleotide ligands. Moreover, results from our kinetic studies indicate that apo-Ras O2 .
can be converted to an active guanine nucleotide binding form of Ras by the addition of a radical scavenger (i.e. ascorbate or GSH) in the presence of GDP. On the basis of our results and analyses, we propose that both O 2 . and a radical quenching agent are required to complete Ras GNE (both Ras guanine nucleotide dissociation and association). Therefore, given the cellular abundance of GTP over GDP, the combined action of O 2 . and a radical scavenger could lead to Ras activation in situ. In cells, GSH is present at high levels and may function as a radical scavenger. Natural quenching agents ascorbate (vitamin C) and tocopherol (vitamin E) may also serve as additional radical scavenging agents in cells (34)