Dinitrosyliron Complexes and the Mechanism(S) of Cellular Protein Nitrosothiol Formation from Nitric Oxide

Dinitrosyliron complexes and the mechanism(s) of cellular protein nitrosothiol formation from nitric oxide

Charles A. Boswortha,b, Jose´ C. Toledo, Jr.c, Jaroslaw W. Zmijewskib,d, Qian Lib,e, and Jack R. Lancaster, Jr.a,b,e,f,1

Departments of aPhysiology and Biophysics, dMedicine, eAnesthesiology, and fEnvironmental Health Sciences, and bCenter for Free Radical Biology, University of Alabama, Birmingham, AL 35205; and cCentro de Cieˆncias Naturais e Humanas, Universidade Federal do ABC, CEP 09210-170, Santo Andre´, Sao Paulo, Brazil

Edited by Louis J. Ignarro, University of California School of Medicine, Los Angeles, CA, and approved January 13, 2009 (received for review November 1, 2007) Nitrosothiols (RSNO), formed from thiols and metabolites of nitric presence of a cellular electron acceptor (6); reaction between •NO •Ϫ oxide (•NO), have been implicated in a diverse set of physiological and superoxide (O2 ) in the presence of excess •NO (7), and the and pathophysiological processes, although the exact mechanisms interaction between •NO and transition metals and/or metal- by which they are formed biologically are unknown. Several loproteins (8–14). candidate nitrosative pathways involve the reaction of •NO with In this study, we examine the contributions of potential nitrosa- O2, reactive oxygen species (ROS), and transition metals. We tive pathways in RAW 264.7 murine macrophages. We present developed a strategy using extracellular ferrocyanide to determine evidence suggesting that dinitrosyliron complexes (DNICs) formed that under our conditions intracellular protein RSNO formation at least in part from the cellular chelatable iron pool (CIP), through occurs from reaction of •NO inside the cell, as opposed to cellular an O2-independent transnitrosative process, are responsible for the entry of nitrosative reactants from the extracellular compartment. majority of RSNO formation in cells, and that reactive oxygen Using this method we found that in RAW 264.7 cells RSNO forma- species (ROS) enhance RSNO formation, possibly via increases in tion occurs only at very low (<8 ␮M) O2 concentrations and DNICs. exhibits zero-order dependence on •NO concentration. Indeed, RSNO formation is not inhibited even at O2 levels <1 ␮M. Addi- Results tionally, chelation of intracellular chelatable iron pool (CIP) reduces •NO Exposure Results in Intracellular RSNO Formation from Reactions RSNO formation by >50%. One possible metal-dependent, O2- of •NO Inside the Cell. Fig. 1A presents a scheme depicting our independent nitrosative pathway is the reaction of thiols with strategy for selectively detecting RSNO formation caused by intra- dinitrosyliron complexes (DNIC), which are formed in cells from the cellular reactions of free •NO. The presence of external ferrocya- reaction of •NO with the CIP. Under our conditions, DNIC forma- nide (FCN), which rapidly reacts with •NO2 (15), will prevent the tion, like RSNO formation, is inhibited by Ϸ50% after chelation of formation of nitrosative species in the external compartment, thus labile iron. Both DNIC and RSNO are also increased during over- preventing their entry into the cells. RAW 264.7 macrophages were production of ROS by the redox cycler 5,8-dimethoxy-1,4-naphtho- incubated 60 min with •NO donor [10 ␮M N-4–1-3-aminopropyl- quinone. Taken together, these data strongly suggest that cellular 2-hydroxy-2-nitrosohydrazinobutyl-1,3-propane-diamine (sper/ RSNO are formed from free •NO via transnitrosation from DNIC NO), t1/2 ϭ 39 min] with or without FCN (1 mM). Fig. 1B shows that derived from the CIP. We have examined in detail the kinetics and there is no statistical difference between intracellular RSNO for- mechanism of RSNO formation inside cells. mation with or without FCN. However, nitrosation of an extracellular target (BSA) was nearly completely inhibited (85.8 Ϯ 6.4%). iron ͉ nitrosation ͉ reactive nitrogen species ͉ reactive oxygen species ͉ These results indicate that under our conditions the RSNO formed chelatable iron inside the cell are primarily caused by the diffusion of •NO into the cell and reaction with cellular components. itric oxide (nitrogen monoxide, •NO) is a ubiquitous signaling Nmolecule that originally was thought to exert its effects solely Time Course of RSNO, •NO, and O2 Levels. To determine the effects through interaction with transition metal ligands of proteins, most of O2 on cellular RSNO formation, we measured RSNO and O2 in notably the heme group of soluble guanylate cyclase. However, it is cell suspensions exposed to •NO donor. Because of the dual effects now recognized that •NO is capable of affecting cell physiology by of •NO on O2 consumption (16) and O2 on •NO consumption (17) inducing oxidative and covalent modification of protein amino acid it is essential to also measure the •NO concentration because BIOCHEMISTRY residues. One such modification, S-nitrosation, the addition of a varying the concentration of one will necessarily affect the other. nitroso group to a thiol to form a nitrosothiol (RSNO), has received When cells are added in the presence of sper/NO, O2 levels rapidly ␮ Ϸ considerable attention as a potentially important posttranslational decline, reaching 8 Min 15 min, after which they decline slowly Ϸ ␮ protein modification. S-nitrosated products are found ubiquitously to 3 M at the end of the 60-min incubation (Fig. 2A). Subsequent in vivo (1), and thiol nitrosation has been found to alter the activity addition of dithionite results in immediate disappearance of O2, of a diverse set of proteins and may therefore represent an indicating that the •NO at these low O2 levels is inhibiting respi- important concept in cellular and organismal biology (2). A central issue in this paradigm is understanding the routes by Author contributions: C.A.B., J.C.T., J.W.Z., and J.R.L. designed research; C.A.B., J.C.T., which RSNO are formed inside cells. Importation of low molecular J.W.Z., and Q.L. performed research; C.A.B. and J.W.Z. contributed new reagents/analytic weight (LMW) RSNO through LAT transporters, followed by tools; C.A.B., J.C.T., J.W.Z., Q.L., and J.R.L. analyzed data; and C.A.B. and J.R.L. wrote the transnitrosation reactions, have been demonstrated to be a potent paper. route for intracellular RSNO formation (3). However, mechanisms The authors declare no conﬂict of interest. of de novo synthesis from free •NO are less clear. Proposed This article is a PNAS Direct Submission. mechanisms include the reaction of •NO with O2 (autoxidation), 1To whom correspondence should be addressed. E-mail: [email protected]. either in the aqueous phase (4) or in an accelerated manner in This article contains supporting information online at www.pnas.org/cgi/content/full/ hydrophobic phases (5); reaction of •NO with a thiol in the 0710416106/DCSupplemental.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0710416106 PNAS ͉ March 24, 2009 ͉ vol. 106 ͉ no. 12 ͉ 4671–4676 Downloaded by guest on September 30, 2021 A A 200 10 0. .7 5 M)

μ .5 0

150 ( 2 O M) 2.5 μ .0 0

( 100 3510 0 45 60 2 Time (min)

O 50

0 0 10 20 30 40 50 60 70 iT me im( n) 60 tnI ar lec l lu ar Ex rt a ralullec BC.s.n 06 dedda( SB A) B 9 00 45 54 80 ) 30 03 60 06 0 )

15 51 40

RSNO (nM) * 003 ( [NO] (nM) RSNO (pmol/mg) 0 0 ( 02 -FCN +FCN NCF- +FCN

RSNO (pmol/mg) 0 0 Fig. 1. Formation of intracellular RSNO is not caused by the autoxidation of 10 20 060504030 •NO in the extracellular space. (A) Scheme of extracellular versus intracellular Ti em )nim( nitrosative processes. External addition of FCN will prevent cellular nitrosation from entry of nitrosative species formed extracellularly from •NO autoxida- C tion, resulting in RSNO formation from only intracellular •NO. See Results for 60 details. (B) Cells were exposed to sper/NO (10 ␮M) Ϯ FCN (1 mM) for 60 min, and RSNO level in lysates was determined as described in Methods.(C) Same 45 as A except nitrosation of external BSA (0.075 mg/mL) was measured. n.s., not signiﬁcant. *, P Ͻ 0.05 compared with control. 30

15 ration. Levels of •NO remain low (Ͻ 20 nM) during the first 20 min RSNO (pmol/mg) of incubation, and then begin rising as O2 declines to Ͻ5 ␮M (Fig. 0 2 A and B). The very low levels of detectable •NO during the initial 0 10 20 30 40 50 60 )nim( emiT )nim( 20 min, despite constant •NO release from the donor, can be attributed to O2-dependent •NO consumption as shown in several Fig. 2. Time course of RSNO formation and •NO and O2 concentrations. (A) cell types (17, 18). By the end of the 60-min incubation, •NO levels Cells were added to PBSD containing sper/NO (10 ␮M) and 1 mM FCN, and O2 are 771 Ϯ 141 nM. Despite ongoing •NO formation and rapid concentration was monitored. (Inset) Scale expansion showing the slow de- cellular consumption during the first 20 min, RSNO formation is crease in O2 at longer times. At arrow, sodium dithionite was added. Repre- sentative data from 2 experiments are shown. (B) Conditions as in A except at not observed during this time (Fig. 2B). The mechanism of O2- dependent •NO consumption by cells thus does not involve nitro- 0, 12, 24, 36, 48, and 60 min cells were removed and processed for RSNO content (■). In parallel experiments, aliquots were removed at the indicated sative chemistry. In addition, nitrosation is observed only when free time points and analyzed for •NO concentration (‚). (C) Cells were incubated •NO begins to appear. Subsequent to the appearance of •NO, the with sper/NO and FCN as in A for 60 min, after which oxymyoglobin (20 ␮M) instantaneous rate of RSNO formation is constant despite a steady was added to scavenge •NO. Samples were collected 0, 20, or 60 min later and increase in •NO concentration. These surprising and unexpected processed for intracellular RSNO content. results indicate that the rate of intracellular RSNO formation is independent of the concentration of •NO. high molecular weight (HMW) proteins. To determine which pool Cellular RSNO Formation Does Not Require O2. To test whether O2 is is the primary target for nitrosative chemistry, the LMW and HMW required for RSNO formation under our conditions, cells were fractions from lysates of sper/NO-exposed cells were separated by added to buffer containing 1 mM FCN in a reaction chamber using filters (10,000-kDa cutoff). After treatment with 10 ␮M containing both an O2 and an •NO electrode. The chamber was sealed and the cells were allowed to respire until the O2 level was below the detection limit of the O2 electrode (Ͻ 100 nM), a process that took 5–10 min. Thereupon argon-purged sper/NO (5 ␮M) was AB )

) 08 added and the cells were incubated with stirring for an additional 01 00 0.1 60 min. After sper/NO addition there is an immediate increase in 57 0 8.0 06 •NO, and peak •NO levels are Ϸ700 nM (Fig. 3A), similar to the 6.0 005 experiments begun under aerobic conditions (Fig. 2B). Surpris- 0.4 04

Ϯ M) ( ingly, RSNO levels in anaerobic experiments are 65.8 7.9 052 0.2 μ 02 pmol/mg (Fig. 3B), again similar to experiments begun under 0 0.0 ( 2 NO (nM) (

aerobic conditions. There is a small increase in O2 after sper/NO O 0 0201 3 040 5 060 RSNO (pmole/mg) 0 addition, although O2 levels remain below •NO levels for the na ox duration of the experiment (Fig. 3A). These data show another T emi )nim( unexpected result, namely, that intracellular RSNO formation is Fig. 3. RSNO formation during anoxia. (A) •NO and O2 electrode traces for virtually independent of O2. anaerobic experiments. Cells (2 ϫ 107 cells/mL) were allowed to respire for 5–10 min until O2 levels reached 0 (t ϭ 0), at which point 1 mM FCN and 5 ␮M Size of Cellular RSNO. Cellular nitrosative chemistry may be target- sper/NO (argon-purged) were added. Cells were then incubated for 60 min. (B) ing either LMW thiols, primarily glutathione (GSH), or thiols on RSNO determination in anaerobically-incubated cells.

4672 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0710416106 Bosworth et al. Downloaded by guest on September 30, 2021 Table 1. Iron-dependent intracellular RSNO formation Treatment RSNO content, pmol/mg Control 52.1 Ϯ 4.3 (14) SIH (30 ␮M) 30.7 Ϯ 2.5* (3) SIH (100 ␮M) 21.2 Ϯ 3.3† (7) ‡ FeSO4 140.1 Ϯ 4.5 (3) § FeSO4 ϩ 100 ␮M SIH 17.1 Ϯ 4.1 (3) BSO 86.8 Ϯ 9.1¶ (4) BSO ϩ 100 ␮M SIH 57.1 Ϯ 7.3ʈ (4)

Cells Ϯ 100 ␮M FeSO4 (1 h) or 100 ␮M BSO (22–24 h) were treated for 1 h with 1 mM FCN and 10 ␮M sper/NO Ϯ SIH. After treatment the cells were collected and assayed for RSNO content. Number of cells is indicated by the numbers in parentheses. *P ϭ 0.044 vs. control. †P ϭ 0.00017 vs. control. ‡P ϭ 2.0 ϫ 10Ϫ7 vs. control. § Ϫ5 P ϭ 0.0027 vs. control, P ϭ 3.6 ϫ 10 vs. FeSO4. ¶P ϭ 0.002 vs. control. ʈP ϭ 0.58 vs. control, 0.044 vs. BSO.

sper/NO and 1 mM FCN, the RSNO content in the retentate is 73.4 Ϯ 17.6% of unfractionated cell extracts. Likewise, in cells treated with 2,3-dimethoxy-1,4-naphthoquinone (DMNQ) (see below), 89.6 Ϯ 8.1% of the mercury-labile signal is found in the retentate. These results demonstrate that the RSNO formed from Fig. 4. Effect of ROS on cellular RSNO formation. (A) Cells preloaded with •NO under our conditions are of large molecular weight. DCFH-DA were exposed to DMNQ (0 or 10 ␮M) for 30 min, and DCFH oxidation was determined as described in Methods. (Magniﬁcation: 20ϫ.) (B) Cells were Stability of Cellular RSNO. To assess the stability of RSNO formed exposed to sper/NO and FCN as in Fig. 1 and DMNQ at the indicated concen- from exogenously-generated •NO, cells were incubated with trations, and cells were analyzed for RSNO content. sper/NO (10 ␮M) for 60 min and then MbO2 (20 ␮M) was added Ϫ to scavenge the remaining •NO. Cells were incubated further for cell-permeable redox cycling compound that produces O• and 0, 20, or 60 min and RSNO content was determined. As shown in 2 hydrogen peroxide (H2O2) (24). Unlike other quinone-based redox Fig. 2C, the level of RSNO is relatively stable until 20 min cyclers, semiquinones produced by DMNQ reduction will not form Ϸ post-MbO2 and then RSNO level begins to decline slowly ( 40%) adducts (25) and DMNQ is therefore thought to exert its effects on during the subsequent 40-min time period. The relatively stable cells exclusively through the production of ROS. As shown in Fig. level of RSNO is consistent with other studies (1, 19, 20). These 4A confocal microscopy reveals that exposure of cells to 10 ␮M results show that the mechanism responsible for RSNO formation DMNQ results in a robust increase (Ͼ2-fold) in fluorescence of the is only slowly reversible and that the increase in RSNO levels (Fig. probe dichlorofluorescein (DCF), confirming that the redox cycler 2B) is caused primarily by the increased rate of formation as increases ROS. Fig. 4B shows that cellular RSNO are increased by opposed to decreased rate of disappearance. DMNQ in a dose-dependent fashion. To confirm that DMNQ increases RSNO because of intracellular production of ROS, we Cellular RSNO Formation Is Largely Mediated by the CIP. Exposure of performed DMNQ experiments in the presence of extracellular cells to •NO results in the formation of EPR-visible dinitrosyliron superoxide dismutase (SOD) (1,000 units/mL). There is no differ- complexes (DNICs) (21), which in acellular systems are capable of ence in RSNO in DMNQ-treated cells in the presence or absence nitrosative chemistry (8, 9). DNICs in cells exposed to •NO are of SOD (125.1 Ϯ 31.7 without SOD vs. 140.6 Ϯ 25.1 pmol/mg with formed from the CIP (22), representing a small (Ϸ5%) fraction of SOD; P ϭ 0.71). total iron thought to be in transit between the various cellular iron We attempted to verify that the DMNQ effect on cellular RSNO pools (23). To test the involvement of the CIP in cellular nitrosation, is attributable to ROS by testing the effect of manganese tetrakis sper/NO-exposed cells were incubated in the presence of the (4-benzoic acid) porphyrin (MnTBAP), a cell-permeable scavenger BIOCHEMISTRY cell-permeant CIP chelator salicylaldehyde isonicotinoyl hydrazone of reactive oxygen and nitrogen intermediates (26). As shown in (SIH) (23) and assayed for RSNO. As is also true for DNIC Table 2, MnTBAP decreases RSNO production in the presence of formation (22), the addition of SIH inhibits RSNO formation in a DMNQ to levels significantly lower than controls. Although these dose-dependent fashion (Table 1), reaching a maximum inhibition results could indicate that the effects of MnTBAP are caused by of Ϸ60% at 100 ␮M SIH. As a positive control, cells were enriched scavenging of ROS, as described in Discussion, the exact mecha- with iron by preincubation with FeSO4 (100 ␮M), and subsequently nisms are not clear. In particular, as shown in Table 2, SIH reverses treated without or with SIH (100 ␮M) during sper/NO treatment. the effect of DMNQ on cellular RSNO formation, thus indicating Iron enrichment results in a substantial increase in cellular RSNO, that the enhanced RSNO formation from DMNQ involves the CIP. which is completely reversed by SIH to the level of unenriched cells Oxidative stress can occur either via increased production of treated with chelator (Table 1). These data indicate an important ROS or decreased antioxidant defenses. To test the effect of the role for the CIP in the formation of RSNO from free •NO in cells, latter on RSNO formation, we incubated cells with the GSH possibly via DNIC. synthesis inhibitor L-buthionine sulfoximine (BSO). Treatment of cells for 20–22 h with BSO diminished the total pool of GSH [GSH Oxidative Stress Increases Cellular RSNO Formation. Oxidative stress ϩ glutathione disulfide (GSSG)] to 19.1 Ϯ 6.2% of controls as can result in increases in nitrosative chemistry as assessed by using assessed by the GSH recycling assay. LMW reduced thiol was fluorescent probes (7). To establish whether this chemistry will undetectable (Ͻ6% total control thiol). Treatment of BSO-treated result in cellular RSNO formation, we treated cells with DMNQ, a cells with sper/NO results in a significant increase in RSNO (Table

Bosworth et al. PNAS ͉ March 24, 2009 ͉ vol. 106 ͉ no. 12 ͉ 4673 Downloaded by guest on September 30, 2021 Table 2. ROS-dependent intracellular RSNO formation donor, a lag occurs in the appearance of both •NO and RSNO, Ͻ ␮ Treatment RSNO content, pmol/mg which corresponds to a decline in [O2] to very low levels ( 10 M) (Fig. 2). We attribute the lag in •NO appearance, in the face of Control 73.2 Ϯ 5.4 (7) constant •NO liberation from donor, to cellular •NO consumption, Ϯ MnTBAP 65.9 6.9* (4) which we have shown is O2 dependent (17). Eventually, as [O2] Ϯ † DMNQ 180.2 16.0 (7) decreases, •NO liberation inhibits mitochondrial O2 consumption, ϩ Ϯ ‡ MnTBAP DMNQ 10.6 3.2 (4) resulting in a plateau of [O2]. At this point both •NO and RSNO SIH ϩ DMNQ 79.9 Ϯ 21.5§ (4) appear and, importantly, the instantaneous rate of RSNO forma- Cells were treated with 1 mM FCN and 10 ␮M sper/NO, Ϯ 100 ␮M MnTBAP, tion is independent of [•NO]. This zero-order kinetic dependence 10 ␮M DMNQ, or 100 ␮M SIH as indicated. After treatment cells were collected argues against either autoxidation (4) or direct •NO/thiol reaction and assayed for RSNO content. Number of cells is indicated by the numbers in with subsequent oxidation (6) as the mechanism of RSNO forma- parentheses. tion because both processes are second order with respect to [•NO]. *P ϭ 0.22 vs. control. The zero-order kinetics suggests that the rate-limiting step is the †P ϭ 2.5 ϫ 10Ϫ5 vs. control. formation of either a short-lived or a low-abundance (and thus ‡P ϭ 3.0 ϫ 10Ϫ8 vs. control, P ϭ 2.8 ϫ 10Ϫ5 vs. DMNQ. limiting) species. §P ϭ 0.68 vs. control, P ϭ 0.0045 vs. DMNQ. One of the most surprising results from this study is the obser- vation that RSNO formation from physiological •NO concentrations (Ͻ1 ␮M) is not inhibited even under conditions of very low 1), confirming the conclusions from the DMNQ experiments that (Ͻ 300 nM) O (Fig. 3). RSNO formation under low O has been oxidative stress in cells is pronitrosative. These results also reveal 2 2 reported (27, 28); however, high concentrations of •NO donor that GSH is not involved in the CIP-dependent nitrosative pathway (50–100 ␮M) were used. Our experiments establish that under any because SIH still reduces cellular RSNO in BSO-treated cells, in net physiologically-relevant condition O will not be limiting for RSNO amount comparable to its effect on control cells (Table 1). 2 formation [although at very low levels it may become limiting for •NO generation by NO synthase (NOS) as discussed below]. These Cellular RSNO Are Formed by DNICs. If DNICs are involved in RSNO results also point out that RSNO may be readily formed in formation, it would be expected that our manipulations (iron subcellular organelles, such as mitochondria, that may experience chelation, ROS production) would have effects on cellular DNIC extremely low O tensions. pool similar to those on the RSNO pool. As shown in Fig. 5, this is 2 RSNO formation involves the participation of the CIP, as indeed the case. Changes in DNIC content with SIH and/or DMNQ demonstrated by the inhibitory effects of SIH (Table 1). More demonstrate strong correlation with the extent of RSNO formation. specifically, the species involving the CIP appears to be the DNIC, Thus, the extent of RSNO formation is predicted by DNIC content. because the extent of RSNO formation correlates with the intensity To verify that the tri-iodide chemiluminescence (CL) assay for of DNIC under a variety of conditions, including SIH-inhibitable RSNO is not measuring DNIC, samples processed as for CL (see DNIC and SIH-resistant DNIC (Fig. 5). We have shown (22) that Methods) were analyzed for DNIC by EPR. There was no decrease there appears to be mutually-exclusive quantitative conversion of in the intensity of the EPR DNIC signal upon 20-min incubation the CIP to either DNIC (upon •NO pretreatment) or conversion with HgCl /sulfanilamide as compared with sulfanilamide alone, 2 to an EPR-visible chelate complex (pretreatment with deferrox- indicating that the mercury-resistant signal in the CL assay, com- amine). These experiments were carried out under different con- monly interpreted as RSNO, is not from DNIC. ditions, in particular with much higher •NO donor doses. Overall, Room temperature EPR analysis reveals anisotropic DNIC however, similar to results presented here (Fig. 5), SIH does not spectra under both basal (sper/NO alone; also see ref. 22) and completely prevent DNIC formation from •NO and SIH-resistant ROS-stimulated (sper/NO ϩ DMNQ) conditions (Fig. S1), indi- DNIC may well be capable of transnitrosating thiols. Other studies cating that in both cases, DNICs (like RSNO) are associated with have demonstrated SIH-inhibitable and SIH-resistant DNIC for- proteins. Together, these data strongly suggest that cellular protein- mation in cells (29, 30). associated DNIC formation from •NO and the CIP is the primary We propose that the major mechanism for RSNO formation is nitrosative pathway from free •NO. via transnitrosation from DNIC to thiol. This mechanism, which has Discussion been suggested by Vanin et al. (9) based on in vitro chemical studies, would explain the surprising independence on both [•NO] and [O ]. We have used ferrocyanide to probe mechanisms of intracellular 2 Zero-order kinetics with regard to [•NO] can be explained by rapid protein RSNO formation from •NO. When concurrent [O ], 2 nitrosylation of the CIP by •NO to form DNIC (22) followed by the [•NO], and RSNO are measured after addition of cells and •NO slower transnitrosation to thiol. The O2 independence is explained by the fact that neither DNIC formation nor transnitrosation involves O . However, nitrosative chemistry requires a net 1-elec- 003 2 tron oxidation; in the case of DNIC-mediated nitrosation it has been suggested that the iron may facilitate the net transfer of an 2 00 electron from one •NO to another, producing HNO and a nitroso- nium equivalent (9). Thus, in this scheme, the oxidant would be •NO itself. Alternatively, transition metal centers other than DNIC 001 might be nitrosating species; several metalloproteins are known to % Control catalyze RSNO formation in vitro (10, 13, 14, 31), and it is unlikely 0 that SIH will be able to chelate tightly-bound metals in prosthetic C SI DM DMN groups. on H N Q Q+ One alternative possibility for the inhibition of RSNO formation SI H by SIH is that the chelator is directly reacting with •NO or a derived Fig. 5. Correlation between DNIC and RSNO. Cells were exposed to 10 ␮M nitrosating species. However, we feel that this possibility is unlikely sper/NO and 1 mM FCN for 60 min, Ϯ SIH and/or 10 ␮M DMNQ, and samples given that RSNO levels in the presence of SIH are the same in both were analyzed for RSNO content (closed bars) or DNIC (open bars) as described Fe-unenriched and Fe-enriched cells (Table 1), indicating that SIH in Methods. Values for RSNO are taken from Tables 1 and 2 and plotted in is affecting RSNO levels through an iron-dependent pathway. graphical form as percentage of control. Additionally, SIH actually decreases the rate at which cells scavenge

4674 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0710416106 Bosworth et al. Downloaded by guest on September 30, 2021 •NO (Fig. S2), opposite of what would be expected were SIH size of the nonheme iron pool. Our work, using intracellular iron directly reacting with •NO. chelators, further implicates the CIP as an important mechanistic Prooxidant treatments such as BSO (to lower GSH levels) and link between DNIC and RSNO. Yang and Loscalzo (20) reported DMNQ (a well-established redox cycler) both substantially increase evidence to suggest that mitochondria act as a source of ROS, which RSNO formation (Fig. 4 and Tables 1 and 2), but the effect is potentiate cellular RSNO formation. Our results with DMNQ and related to the CIP because it is almost completely reversed by SIH GSH depletion are consistent with this conclusion and suggest a (Tables 1 and 2). The portion that is not reversed by SIH may still specific mechanism, i.e., increasing DNIC. It is also possible that be caused by DNIC because there may be 2 pools of iron involved mitochondria regulate RSNO formation by regulating O2 levels; in DNIC formation and protein nitrosation, one chelatable and one decreased mitochondrial O2 consumption could attanuate RSNO resistant to SIH chelation. ROS and RNS have been shown to levels by increased O2-dependent •NO consumption (Fig. 2). Espey modulate intracellular iron (32). The effect of BSO also shows that et al. (38) studied intracellular nitrosation of the probe diamin- the CIP-dependent nitrosative pathway is unaffected by changes in ofluorescein (DAF) and compared the mechanism to that for GSH or other LMW thiol (Table 1), indicating that GSH-DNICs aqueous solution. They conclude that the intracellular mechanism (33) are not obligatory intermediates, nor are other products of the is different from in aqueous solution, based primarily on the interaction between GSH and the CIP. first-order rate of increase of DAF nitrosation in the presence of Another possibility for the BSO effect is that decreased GSH will cells, compared with the second-order rate in the absence of cells result in decreased S-nitrosoglutathione (GSNO) reductase (34) or with cell lysate. However, their kinetic results could reflect the activity. Assuming an equilibrium between GSNO and protein appearance and disappearance of •NO in their reaction system, RSNO as suggested by Liu et al. (35), this effect could tend to both of which will be first-order under the conditions they use. increase cellular RSNO by inhibiting the GSNO reductase- We should point out several caveats to this work. First, we mediated enzymatic breakdown pathway. However, the slow de- measure overall cellular nitrosative chemistry, rather than nitrosa- cline in RSNO upon removal of •NO (Fig. 2C) suggests that the tion reactions within specific spatially or functionally distinct mechanism of action of BSO is not on decreased breakdown via groups. Both protein structure and intracellular compartmental- transnitrosation to GSH and subsequent GSNO breakdown via ization will affect susceptibility of individual cysteine residues to GSNO reductase. nitrosation. It is also possible that a mechanism accounting for a One interesting feature of DMNQ-mediated RSNO formation is small proportion of cellular RSNO, perhaps not detectable here, that it does not exhibit a ‘‘bell shape’’ with a DMNQ dose that is will have functional consequences far beyond its contribution to •Ϫ expected for O2 -dependent nitrosative chemistry (7). The bell- overall nitrosation, because of the importance of the target proteins •Ϫ shape derives from the fact that O2 potentiates nitrosation only in in cellular processes. We also note that our work is restricted to the •Ϫ •Ϫ conditions where O2 flux is below •NO flux, because O2 will formation of RSNO from free •NO inside cells, rather than the consume •NO, and nitrosation requires free •NO (36). Assuming importation of LMW RSNO from the extracellular space. LMW a similar process inside cells, the observed plateau could be a result RSNO have been shown to be imported into the cell and cause of saturation of ROS production caused by rate-limiting DMNQ massive protein nitrosation (3), and it is possible this mechanism is reduction. Additionally, if the DMNQ effect on nitrosation is as important, if not more important, in vivo. Additionally, we have caused by mobilization of nitrosative iron (as suggested above), used •NO donor rather than •NO from endogenous formation. It •Ϫ rather than a direct reaction of RNS derived from the O2 /•NO may be, for example, that relatively long-term •NO exposure during with thiols, then a bell-shaped curve would not be predicted. up-regulation of inducible NOS results in •NO-induced modifica- We attempted to test the involvement of ROS on RSNO tions in protein expression and/or iron homeostatic responses to formation by using the scavenger MnTBAP. MnTBAP has no modifications in the CIP (39). Finally, our proposal that •NO effect on RSNO formation in the absence of DMNQ (Table 2); affects nitrosation via DNIC formation and subsequent transnitro- however, with DMNQ it dramatically reduces RSNO to below sation to thiols is based on indirect evidence (kinetics, O2 indepen- controls. However, interpretation of these results is problematic, dence, CIP dependence) and the direct correlation between RSNO because manganese porphyrinic oxidant scavengers such as MnT- and DNIC with different experimental manipulations. It may be, BAP display a wide spectrum of reactivity and are capable of •Ϫ however, that the correlation is simply reflective of some underlying reaction with several stronger oxidants than O2 (26). In addition, mechanism that produces RSNO and DNIC by separate and it has been recently demonstrated that commercial preparations of nonoverlapping processes. However, given the relative sizes of the MnTBAP are contaminated by multiple ‘‘noninnocent’’ impurities CIP, RSNO, and DNIC (22) pools and the aforementioned kinetic (26). Further work on the precise action of commercial and purified observations, the simplest (and thus most likely) explanation is that manganese porphyrinic scavengers is needed before we can make DNIC and RSNO formation are causally linked. In our previous useful speculations on the mechanistic basis of their effects on work (22), we have ruled out the possibility that cellular DNICs are

RSNO formation. formed from RSNO. Thus, if a cause and effect relationship exists, BIOCHEMISTRY We observe RSNO formation only at O2 levels where iNOS and as seems probable, it must be that DNICs form RSNO. nNOS activity will be less than maximal because of the relatively Our results may provide insight into another question of major high Km for O2 for these enzymes. However, although enzymatic importance to NO signaling biology, namely as to how protein activity will be lower than maximal below the Km, it will not be 0 specificity is achieved in nitrosation reactions. Factors that have because appreciable •NO production from NOS isozymes does been suggested to be important are thiol pKa, identity and config- indeed occur well below the Km (37). The zero-order dependence uration of neighboring amino acid residues, and accessibility (2). of RSNO formation on [•NO] implies an interesting possibility for We suggest that another factor that may contribute to specificity of the in vivo condition, that at low O2 levels, RSNO formation will nitrosation is iron-dinitrosyl coordination near the target thiol. be independent of the effects of O2 on NOS, e.g., decreasing O2 ␮ Alternatively, once the DNIC is formed on a protein separate from from 10 to 5 M will decrease the activity of the NOS, but have no the target thiol, specificity is imparted by protein–protein recogni- effect on RSNO formation. Indeed, Bryan et al. (1) have deter- tion mechanisms. mined that RSNO levels in vivo are relatively unaffected by global hypoxia, and in some tissues they are actually increased. Methods The results of this study are interesting in light of 3 other reports Materials. DMNQ was from Oxis, and sper/NO and MnTBAP were from Alexis. SIH investigating mechanisms of nitrosation in cells. Kim et al. (11) was a generous gift from Prem Ponka (Lady Davis Institute for Medical Research, showed that both DNIC and RSNO could be increased in RAW Montreal). DMEM, FBS, and penicillin/streptomycin were from Fisher. All other cells by iron loading, and that both of these species correlated to the reagents were from Sigma.

Bosworth et al. PNAS ͉ March 24, 2009 ͉ vol. 106 ͉ no. 12 ͉ 4675 Downloaded by guest on September 30, 2021 Cell Culture. Murine RAW 264.7 macrophages were cultured as described (22). Labs) against nitrite standards. For anaerobic experiments, •NO was measured by Before experiments, cells were rinsed twice and collected in HBSS supplemented using an •NO-specific electrode (ISO-NOP; World Precision Instruments). with 10 mM Hepes, pH 7.4. Cells were collected by scraping and washed in fresh HBSS/Hepes before use. For iron loading, cells were preloaded by incubation with O2 Measurements. O2 concentration of cell suspensions was determined in a 100 ␮M FeSO4 in HBSS/Hepes for 1 h. Excess iron was then removed by washing sealed, stirred chamber at 37 °C containing an O2 electrode (Oxygraph; Orobo- 3 times with HBSS/Hepes. ros). The electrode was calibrated at room air and at zero O2 by the addition of sodium dithionite. Cell •NO Exposure. Cells were added to reaction vials (to a final density of 107 cells per mL, unless otherwise indicated) containing PBS with 100 ␮M diethylenetri- GSH and Reduced Thiol Measurements. For GSH determination, cells were aminepentacetic acid (DTPA) at pH 7.4 (PBSD) with sper/NO (10 ␮M, unless rinsed twice with PBSD and lysed with PBSD containing Triton X-100 (0.1%). ϫ otherwise indicated) and (where indicated) FCN (1 mM) and concentrations of Lysates were vortexed vigorously and centrifuged 15,800 g at 4 °C. Super- ϩ SIH, DMNQ, and/or MnTBAP as indicated. Vials were capped with rubber septa natants were reserved for GSH and total thiol analysis. Total GSH (GSH GSSG) and incubated with zero headspace with rotation at 37 °C in the dark for 60 min. was measured by using the GSH recycling assay (41). Total and LMW reduced thiol levels were assessed by DTNB before and after protein precipitation by For anaerobic and time-course experiments, MbO2 (20–30 ␮M) was added at the trichloroacetic acid (42). end of the exposure to scavenge remaining •NO. Cells were processed as described by Zhang and Hogg (3), modified as follows. Cells were collected by centrifugation, washed twice with ice-cold PBSD, and resuspended in ice-cold Imaging of DCF Fluorescence. Intracellular oxidants were measured by using the probe dichlorofluoroscin diacetate and confocal microscopy as described (43). thiol blocking buffer (50 mM KPi,50mMN-ethylmaleimide, 100 ␮M DTPA, pH 7.4) for RSNO determination or in 200 ␮L of PBSD for EPR analysis. EPR samples were transferred to EPR tubes and frozen at Ϫ80 °C until analysis. Samples to be EPR Spectroscopy. EPR spectra were obtained and analyzed as described (22). analyzed for RSNO content were lysed by freeze-thaw, followed by brief sonica- tion on ice. Insoluble material was removed by centrifugation and lysates were Protein Determination: Protein content was determined by using the bicincho- stored at Ϫ80 °C until analysis for RSNO content. Cells treated with 20 ␮M nitrite ninic acid method (Pierce). had barely-detectable RSNO levels (2.6 Ϯ 0.8 pmol⅐mgϪ1 protein), indicating that Ϯ Ն RSNO formed from sper/NO were caused by the release of •NO. Statistical Analysis. All measures reflect mean SEM, with n 3. Means between groups were compared by using paired Student’s t test. P Ͻ 0.05 were considered significant. RSNO and •NO Determination. RSNO levels were determined by triiodide- dependent CL as described by Feelisch et al. (40), using an EcoMedics CLD 88 sp ACKNOWLEDGMENTS. We thank Kay Shi for excellent technical assistance, Dave detector. •NO concentration during sper/NO treatment was analyzed by remov- Kraus for use of the oxygraph, and Dr. Rakesh Patel for insightful discussions. This ing aliquots from reaction vials using a gas-tight syringe and adding them to the work was supported in part by American Heart Association (Southeastern Affil- CL reaction chamber containing 1 M NaOH to prevent further release of •NO from iate) Predoctoral Fellowship 0515098B (to C.A.B.) and National Institutes of the sper/NO. Signals were integrated by using ChromProcessor software (ACD Health Grants HL71189 and HL074391 (to J.R.L.).

1. Bryan NS, et al. (2004) Cellular targets and mechanisms of nitros(yl)ation: An insight 24. Shi MM, Kugelman A, Iwamoto T, Tian L, Forman HJ (1994) Quinone-induced oxidative into their nature and kinetics in vivo. Proc Natl Acad Sci USA 101:4308–4313. stress elevates glutathione and induces ␥-glutamylcysteine synthetase activity in rat 2. Hess DT, Matsumoto A, Kim SO, Marshall HE, Stamler JS (2005) Protein S-nitrosylation: lung epithelial L2 cells. J Biol Chem 269:26512–26517. Purview and parameters. Nat Rev Mol Cell Biol 6:150–166. 25. Gant TW, Rao DN, Mason RP, Cohen GM (1988) Redox cycling and sulphydryl arylation: 3. Zhang Y, Hogg N (2005) S-nitrosothiols: Cellular formation and transport. Free Radical Their relative importance in the mechanism of quinone cytotoxicity to isolated hepa- Biol Med 38:831–838. tocytes. Chem Biol Interact 65:157–173. 4. Kharitonov VG, Sundquist AR, Sharma VS (1995) Kinetics of nitrosation of thiols by 26. Reboucas JS, Spasojevic I, Batinic-Haberle I (2008) Pure manganese(III) 5,10,15,20- nitric oxide in the presence of oxygen. J Biol Chem 270:28158–28164. tetrakis(4-benzoic acid)porphyrin (MnTBAP) is not a superoxide dismutase mimic in 5. Moller MN, Li Q, Lancaster JR, Jr, Denicola A (2007) Acceleration of nitric oxide aqueous systems: A case of structure–activity relationship as a watchdog mechanism in autoxidation and nitrosation by membranes. IUBMB Life 59:243–248. experimental therapeutics and biology. J Biol Inorg Chem 13:289–302. 6. Gow AJ, Buerk DG, Ischiropoulos H (1997) A novel reaction mechanism for the forma- 27. Foster MW, Stamler JS (2004) New insights into protein S-nitrosylation. Mitochondria tion of S-nitrosothiol in vivo. J Biol Chem 272:2841–2845. as a model system J Biol Chem 279:25891–25897. 7. Espey MG, Thomas DD, Miranda KM, Wink DA (2002) Focusing of nitric oxide mediated 28. Jourd’heuil D, Jourd’heuil FL, Feelisch M (2003) Oxidation and nitrosation of thiols at nitrosation and oxidative nitrosylation as a consequence of reaction with superoxide. low micromolar exposure to nitric oxide. Evidence for a free radical mechanism J Biol Proc Natl Acad Sci USA 99:11127–11132. Chem 278:15720–15726. 8. Boese M, Mordvintcev PI, Vanin AF, Busse R, Mulsch A (1995) S-nitrosation of serum 29. Lewandowska H, Meczynska S, Sochanowicz B, Sadlo J, Kruszewski M (2007) Crucial albumin by dinitrosyl-iron complex. J Biol Chem 270:29244–29249. role of lysosomal iron in the formation of dinitrosyl iron complexes in vivo. J Biol Inorg 9. Vanin AF, Malenkova IV, Serezhenkov VA (1997) Iron catalyzes both decomposition Chem 12:345–352. and synthesis of S-nitrosothiols: Optical and electron paramagnetic resonance studies. 30. Meczynska S, Lewandowska H, Sochanowicz B, Sadlo J, Kruszewski M (2008) Variable Nitric Oxide 1:191–203. inhibitory effects on the formation of dinitrosyl iron complexes by deferoxamine and 10. Inoue K, et al. (1999) Nitrosothiol formation catalyzed by ceruloplasmin. Implication salicylaldehyde isonicotinoyl hydrazone in K562 cells. Hemoglobin 32:157–163. for cytoprotective mechanism in vivo. J Biol Chem 274:27069–27075. 31. Luchsinger BP et al. (2003) Routes to S-nitroso-hemoglobin formation with heme redox 11. Kim YM, Chung HT, Simmons RL, Billiar TR (2000) Cellular nonheme iron content is a and preferential reactivity in the beta subunits. Proc Natl Acad Sci USA 100:461–466. determinant of nitric oxide-mediated apoptosis, necrosis, and caspase inhibition. J Biol 32. Mladenka P, Simunek T, Hubl M, Hrdina R (2006) The role of reactive oxygen and Chem 275:10954–10961. nitrogen species in cellular iron metabolism. Free Radical Res 40:263–272. 12. English AM, Wilcox DE (2001) Effects of metal ions on the oxidation and nitrosation of 33. Cesareo E, et al. (2005) Nitrosylation of human glutathione transferase P1–1 with cysteine residues in proteins and enzymes. Met Ions Biol Syst 38:313–350. dinitrosyl diglutathionyl iron complex in vitro and in vivo. J Biol Chem 280:42172– 13. Tao L, English AM (2003) Mechanism of S-nitrosation of recombinant human brain 42180. calbindin D28K. Biochemistry 42:3326–3334. 34. Liu L, et al. (2001) A metabolic enzyme for S-nitrosothiol conserved from bacteria to 14. Weichsel A, et al. (2005) Heme-assisted S-nitrosation of a proximal thiolate in a nitric humans. Nature 410:490–494. oxide transport protein. Proc Natl Acad Sci USA 102:594–599. 35. Liu L, et al. (2004) Essential roles of S-nitrosothiols in vascular homeostasis and 15. Goldstein S, Czapski G (1995) Kinetics of nitric oxide autooxidation in aqueous solution endotoxic shock. Cell 116:617–628. in the absence and presence of various reductants. The nature of the oxidizing 36. Lancaster JR, Jr (2006) Nitroxidative, nitrosative, and nitrative stress: Kinetic predic- intermediates. J Am Chem Soc 117:12078–12084. tions of reactive nitrogen species chemistry under biological conditions. Chem Res 16. Brown GC (2007) Nitric oxide and mitochondria. Front Biosci 12:1024–1033. Toxicol 19:1160–1174. 17. Thomas DD, Liu X, Kantrow SP, Lancaster JR, Jr (2001) The biological lifetime of nitric 37. Robinson MA, Baumgardner JE, Good VP, Otto CM (2008) Physiological and hypoxic O2 oxide: Implications for the perivascular dynamics of NO and O2. Proc Natl Acad Sci USA tensions rapidly regulate NO production by stimulated macrophages. Am J Physiol 98:355–360. 294:C1079–C1087. 18. Gardner PR, Martin LA, Hall D, Gardner AM (2001) Dioxygen-dependent metabolism of 38. Espey MG, Miranda KM, Thomas DD, Wink DA (2001) Distinction between nitrosating nitric oxide in mammalian cells. Free Radical Biol Med 31:191–204. mechanisms within human cells and aqueous solution. J Biol Chem 276:30085–30091. 19. Zhang Y, Hogg N (2004) Formation and stability of S-nitrosothiols in RAW 264.7 cells. 39. Pantopoulos K, Weiss G, Hentze MW (1996) Nitric oxide and oxidative stress (H2O2) Am J Physiol 287:L467–L474. control mammalian iron metabolism by different pathways. Mol Cell Biol 16:3781– 20. Yang Y, Loscalzo J (2005) S-nitrosoprotein formation and localization in endothelial 3788. cells. Proc Natl Acad Sci USA 102:117–122. 40. Feelisch M, et al. (2002) Concomitant S-, N-, and heme-nitros(yl)ation in biological 21. Henry YA, Guissani A, Ducastel B (1997) Nitric Oxide Research from Chemistry to tissues and ﬂuids: Implications for the fate of NO in vivo. FASEB J 16:1775–1785. Biology: EPR Spectroscopy of Nitrosylated Compounds (Landes Bioscience, Austin, TX). 41. Moellering DR, et al. (2002) Induction of glutathione synthesis by oxidized low-density 22. Toledo JC, Jr, et al. (2008) Nitric oxide-induced conversion of cellular chelatable iron lipoprotein and 1-palmitoyl-2-arachidonyl phosphatidylcholine: Protection against into macromolecule-bound paramagnetic dinitrosyliron complexes J Biol Chem quinone-mediated oxidative stress. Biochem J 362:51–59. 283:28926–28933. 42. Sedlak J, Lindsay RH (1968) Estimation of total, protein-bound, and nonprotein sulf- 23. Glickstein H, El RB, Shvartsman M, Cabantchik ZI (2005) Intracellular labile iron pools hydryl groups in tissue with Ellman’s reagent. Anal Biochem 25:192–205. as direct targets of iron chelators: A ﬂuorescence study of chelator action in living cells. 43. Zmijewski JW, et al. (2005) Oxidized LDL induces mitochondrially associated reactive Blood 106:3242–3250. oxygen/nitrogen species formation in endothelial cells. Am J Physiol 289:H852–H861.

4676 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0710416106 Bosworth et al. Downloaded by guest on September 30, 2021