Desferrioxamine Inhibits Protein Tyrosine Nitration: Mechanisms and Implications

Abstract

Tissues are exposed to exogenous and endogenous nitrogen dioxide (^•NO₂), which is the terminal agent in protein tyrosine nitration. Besides iron chelation, the hydroxamic acid (HA) desferrioxamine (DFO) shows multiple functionalities including nitration inhibition. To investigate mechanisms whereby DFO affects 3-nitrotyrosine (3-NT) formation, we utilized gas phase ^•NO₂ exposures, to limit introduction of other reactive species, and a lung surface model wherein red cell membranes (RCM) were immobilized under a defined aqueous film. When RCM were exposed to ^•NO₂ covered by +/− DFO: (i) DFO inhibited 3-NT formation more effectively than other HA and non-HA chelators; (ii) 3-NT inhibition occurred at very low [DFO] for prolonged times; and (iii) 3-NT formation was iron independent but inhibition required DFO present. DFO poorly reacted with ^•NO₂ compared to ascorbate, assessed via ^•NO₂ reactive absorption and aqueous phase oxidation rates, yet limited 3-NT formation at far lower concentrations. DFO also inhibited nitration under aqueous bulk phase conditions, and inhibited 3-NT generated by active myeloperoxidase “bound” to RCM. Per the above and kinetic analyses suggesting preferential DFO versus ^•NO₂ reaction within membranes, we conclude that DFO inhibits 3-NT formation predominantly by facile repair of the tyrosyl radical intermediate, which prevents ^•NO₂ addition, and thus nitration, and potentially influences biochemical functionalities.

Introduction

Desferrioxamine (DFO), a siderophore produced by Streptomyces pilosus, is currently the most widely employed iron chelator and is administered clinically to promote the excretion of toxic iron. It consists of a chain of three hydroxamic acids (see Fig. 1) which allows it to bind ferric (Fe³⁺) iron with a stability constant of 10³¹ [1]. Apart from its clinical use, DFO is useful in investigating free-radical induced injury [2]. DFO reacts with both superoxide and hydroxyl radicals, affects eicosanoid metabolism, acts as a substrate for peroxidases and inhibits them at high concentrations, and is capable of forming a nitroxide radical [3,4]. In animal studies, DFO has shown protective effects on ^•NO₂- and smoke-induced lung injury [5,6]. These effects have been almost entirely attributed to iron chelation and thus inhibition of iron-dependant free radical reactions. However, it has also been proposed that DFO can inhibit free radical reactions by iron chelation-independent mechanisms such as direct scavenging of reactive species including the nitrogen dioxide radical (^•NO₂) [7–9].

Figure 1. Schematic Structures of Hydroxamic Acids and Related Compounds.

Note that DFO contains three hydroxamic groups while HU, BHA, and SHA only a single group. HBED and DTPA are not hydroxamic acids although all five moieties show variable degrees of metal chelation.

It is widely believed that ^•NO₂ is pivotal in biological nitration reactions [10–13], that have been observed concomitant with a variety of pathophysiologic conditions [14–17]. In vivo exposure to ^•NO₂ occurs via direct inhalation, decomposition of higher order nitrogen oxides, peroxidase-mediated reactions, ^•NO near diffusion-controlled reaction with organic peroxyl radicals which form ^•NO₂ and the corresponding alkoxyl radicals, and, to a limited extent, autoxidation of nitric oxide (^•NO) [11–13,18–21]. The reaction of peroxynitrite with CO₂, that ultimately forms the ^•NO₂ and CO₃^•− radicals, and ^•NO₂ catalytically generated by the enzyme myeloperoxidase using nitrite as a substrate, are commonly associated with the nitration of tyrosine residues. However, these reactions may not be the sole sources of ^•NO₂, especially in the outermost tissue strata. The lung, for example, endogenously generates ^•NO and is frequently exposed to environmental ^•NO and ^•NO₂ that can be absorbed from the intrapulmonary gas phase by diffusive and reactive processes [18,20,22,23]. Within the lung, in a concentration dependent manner, ^•NO₂ exposure initiates acute epithelial injury, activation of signaling cascades, inflammation, protein and lipid oxidation, and, ultimately remodeling of lung architecture [18,24–26].

The entire respiratory tract surface is covered by a biochemically complex aqueous layer termed the epithelial lining fluid (ELF) into which inhaled gases dissolve and then must diffuse before contacting the underlying epithelia. Reactive absorption, wherein ELF reactions with solute ^•NO₂ maintain the net driving force for continued mass transfer of ^•NO₂ from the gas to aqueous phase [22,27,28], may play an important role in lung surface compartment nitration. Because ^•NO₂ undergoes facile reaction with numerous ELF constituents and because recent comprehensive analyses suggest markedly reduced hydrophobic partitioning relative to ^•NO [21,29,30], initial ^•NO₂ reactions are likely largely confined to the ELF. Thus, depending upon the respective concentrations of ^•NO₂ and ELF substrates and the ELF thickness, diffusion versus reaction properties will dictate how far ^•NO₂ penetrates into and through the ELF, which varies in thickness from 1 to 10 μm [31]. Recent estimates suggest that ^•NO₂ reaction with ELF constituents would predominate over its diffusion to the epithelial surface [21,32].

This is further supported by observations that ^•NO₂ exposure-mediated 3-nitrotyrosine (3-NT) formation occurred in cell membranes when overlain by phosphate buffer but not at appreciable antioxidant concentrations [32]. In addition, inclusion of metal chelators DFO or diethylenetriaminepenta-acetic acid (DTPA) to the model system inhibited exposure-mediated membrane oxidation but, DFO also inhibited membrane nitration while DTPA did not [32,33]. Because tyrosine nitration is generally considered to involve a two step process wherein oxidation generates a tyrosyl radical intermediate followed by ^•NO₂ addition, diverse oxidants other than ^•NO₂ (e.g., carbonate radical, lipid radicals, etc) may contribute to the first step [12,34]. Consequently, interventions that reduce the extent of tyrosine nitration may result from a variety of mechanisms.

Accordingly, we extended our previous observations of ^•NO₂ reaction/diffusion phenomena via studies to discern mechanisms by which DFO inhibits protein nitration under conditions that approximate the lung surface. A previously utilized reductionist model was employed to facilitate tight experimental controls, including the direct use of gaseous ^•NO₂ which circumvents initial generation of other reactive species that occurs during use of precursor reagents, pulse radiolysis, or flash photolysis [7,12,32–35]. Our results suggest that, in addition to modest direct ^•NO₂ scavenging, DFO reduces tyrosyl radicals and thus inhibits the radical-radical addition step that leads to 3-NT formation.

Materials and Methods

Reagents

Reagents were obtained from the following suppliers: (i)compressed ^•NO₂ (99.1 ppm in N₂) - BOC Gases (Riverton, NJ, USA); (ii) nitrotyrosine polyclonal antibody, nitrotyrosine BSA, and 3-nitrotyrosine (3-NT) - Cayman Chemical (Ann Arbor, MI, USA); (iii) acrylamide, lysine, Tris, nitrocellulose membrane paper and broad range SDS-page - Bio-Rad (Hercules, CA, USA); (iv) sodium lauryl sulfate, monobasic and dibasic sodium phosphate - Fisher Scientific (Fair Lawn, NJ, USA); (v) SuperSignal west pico chemiluminescent substrate and BCA protein assay kit - Pierce Biotechnology, Inc. (Rockford, IL, USA); (vi) L-tyrosine - ICN Biochemical (Cleveland, OH, USA); (vii) hydroxyurea (HU) - Calbiochem (La Jolla, CA, USA); (viii) N,N′-Di(2-hydroxybenzyl)ethylenediamine-N, N′-diacetic acid monohydrochloride hydrate (HBED) - Strem Chemicals (Newbury Port, MA, USA); (ix) desferrioxamine (DFO), salicylhydroxamic acid (SHA), benzohydroxamic acid (BHA), diethylenetriaminepentaacetic acid (DTPA), ascorbic acid (AH₂), ferric chloride (FeCl₃), monoclonal anti-rabbit IgG and all other materials - Sigma-Aldrich CO (St. Louis, MO, USA). Structures of the various hydroxamic acids and related reagents are presented in Figure 1. Because of the functional group structural similarities, we shall consider hydroxyurea as a hydroxamic acid.

In vitro lung surface model

Human blood was obtained by venipuncture according to the University of Alabama, Birmingham IRB protocol number X030320005. The lung surface model, as described in [24, 34] was designed to mimic the lung surface via its three compartments; air space, epithelial lining fluid (ELF), and epithelium. Herein, a monolayer of red blood cell-derived membranes (red cell membranes; RCM) bound to the bottom of Petri dishes (60 × 15 mm) represents the epithelial apical surface. Prior to application of red cells, dishes were boiled in 10% nitric acid, silanized with 3-aminopropyltriethoxysilane, and treated with glutaraldehyde to generate a surface of reactive aldehydes capable of binding to the amino termini and other amino moieties of proteins [32,33]. Erythrocytes were washed three times with 310 mOsm phosphate buffer, pH 7.4, and centrifuged at 4000 RCF. 2.3 ml of a 10% erythrocyte suspension (in 310 mOsm phosphate buffer) was added to the dishes and allowed to bind for 30 minutes. The bound cells were then lysed with 2 ml hypotonic phosphate buffer (10 mOsm, pH 7.0), and repeatedly rinsed to remove hemoglobin and other intracellular debris, resulting in a monolayer of RCM covalently bound to the glass surface of the dishes. Immediately prior to exposure, 2 ml model ELF (10 mOsm phosphate buffer + dissolved reagents, pH 7.0 [36]) were added over the RCM. A hypo-osmolar buffer was used during exposure to inhibit RCM re-annealing, which would generate a fourth (intracellular) compartment, leaving a sheet of open membranes so that both “inner” and “outer” sides were exposed to aqueous phase solutes. Reagents included AH₂, FeCl₃, iron chelators DFO, DTPA, HBED, SHA, BHA, and HU. In order to facilitate iron removal prior to exposure initiation, several experiments were conducted wherein either 50 μM DFO or 400 μM DTPA were added to the membranes for 10 minutes followed by repeated rinsing and addition of the model ELF. Ferrioxamine was formed by combining DFO with a molar excess of FeCl₃ prior to introduction into the RCM model.

RCM exposure protocol

Petri dishes with bound RCM were exposed to ^•NO₂ for 30 minutes in a small glass chamber (1500 ml). In order to mimic ELF in situ thickness, the chamber was intermittently tilted to leave an aqueous film covering the upper half of the dish, which provided a reservoir of reactive substrates pooled in the lower half of the dish. Every two minutes, the chamber was tilted to the opposite angle allowing for each half to be exposed equally [32,33]. Exposures were conducted under steady state, first order conditions with respect to ^•NO₂ [37] ([^•NO₂]_inflow ~ 4.5 ppm) at 25°C. High concentration ^•NO₂ (~ 100 ppm ^•NO₂ in N₂) was injected countercurrent into a stream of humidified air which then flowed through a diffuser located in the top of the exposure chamber [32], resulting in well-mixed, gas phase conditions (~ 1 volume change/min). To reduce background ^•NO₂ interactions with chamber components, the entire exposure system was equilibrated (wherein [^•NO₂]_inlet = [^•NO₂]_outlet) prior to Petri dish introduction. Gas phase ^•NO₂ concentrations were continuously monitored via chemiluminescence (Model 42C ^•NO_x analyzer; Thermo Fisher, Franklin, MA, USA). ^•NO concentrations were consistently well below 0.10 ppm.

Myeloperoxidase studies

To directly generate ^•NO₂ at the RCM, thus avoiding ^•NO₂ mass transfer into and reaction/diffusion within the overlying aqueous film, RCM were overlain with a 0.7 units/ml MPO solution, the Petri dishes covered, and gently but continuously rocked for 15 min. The MPO solution was then aspirated and the dishes rinsed, resulting in catalytically active MPO adhered to the RCM. MPO activity was assessed via intra-dish reaction mixtures using tetramethylbenzidine as the substrate [38]. Residual (bound) MPO activity averaged 0.56 units/ml, with little inter-dish variation. For nitration studies, RCM + MPO were overlain with various aqueous systems (+/− 100 μM NO₂⁻ and/or 100 μM H₂O₂) and continuously rocked in covered dishes for 30 min. Following treatment, membrane proteins were isolated and the presence of 3-NT determined by Western analysis.

Flask exposures

50 ml Erlenmeyer flasks were equipped with Teflon covered stoppers, Teflon inlet and exhaust tubes, and stir bars. ^•NO₂ gas phase mixtures were generated as above and the exposure system and flask conditioned until inlet (~ 150 ml/min) and exhaust [^•NO₂] were equivalent. Test solutions were directly injected into the flask, exposures conducted under first-order conditions with respect to [^•NO₂] and both gas and aqueous phase well stirred conditions, and the time-dependent change in ^•NO₂ exit concentration monitored. This approach permits mass balance computations (i.e. ([^•NO₂]_in - [^•NO₂]_out x time x flow = total uptake) of ^•NO₂ reactive absorption. Although this is a lumped measure that includes interfacial transfer, ^•NO₂ solubility, both solute ^•NO₂ and reactive substrate diffusion, and bimolecular reaction kinetics, ^•NO₂ uptake (gas phase disappearance) serves as a measure of relative reactivity since, under fixed conditions of [^•NO₂]_inflow, temperature, flow, and stirring, interfacial transfer (uptake) is determined by solute ^•NO₂ reaction which is a function of substrate concentration and reaction rate (k • [S]) [22,30]. NO₂⁻ accumulation in the model ELF was measured using the Griess reaction [39]. Larger flasks (250 ml) and volumes (20 ml) were also used to assess differential rates of AH₂ and DFO oxidation due to ^•NO₂ exposure. Oxidation was evaluated spectrophotometrically using extinction coefficients (AH₂; ε_{(265 nm)} = 14,500 M⁻¹cm⁻¹ [40]) and the loss of ability to form ferrioxamine (ε_{(425 nm)} = 2,460 M⁻¹cm⁻¹ [41]) in the presence of excess ferric ions (see supplemental data).

Bubbler exposures

Using a fine fritted bubbler equipped with a stopcock at the bottom to allow for sample removal, approximately 8 ppm ^•NO₂ in humidified air, generated as above, was bubbled through a solution of 0.8 mM tyrosine with or without 50 μM DFO (in 10 mOsm phosphate buffer, pH 7.0) for 120 minutes to assess the formation of 3-NT under aqueous bulk phase conditions. Every 30 minutes, 2.25 ml of the exposed solution were combined with 0.25 ml of 1 N NaOH and scanned from 200–800 nm (Cary 100 Bio UV-Visible Spectrophotometer, Walnut Creek, CA, USA). 3-NT displays a broad peak at approximately 430 nm [42].

Western blot analysis of 3-nitrotyrosine formation

RCM nitration was evaluated by Western blot analysis of 3-NT formation in membrane proteins relative to aqueous phase conditions (e.g., AH₂, DTPA, HBED, DFO, DFO + FeCl₃, SHA, BHA, HU), and pretreatment conditions (DFO, DTPA). Immediately after exposure, the aqueous phase was removed and RCM were rinsed with phosphate buffer and 150 μl of extraction buffer (150 mM NaCl, 50 mM Tris, 1 mM EGTA, .25% Sodium deoxycholate, 1% Igepal, 1 mM Na orthovanadate, 1 mM protease inhibitor, and 0.4% HCl) was added to each dish to solubilize membrane proteins. The dishes were scraped and the composite extract stored in sealed microfuge tubes (−80°C). Using the BCA assay to determine appropriate loading amounts, 10 μg protein from each sample and nitrated bovine serum albumin (positive control) were separated by gel electrophoresis (10% gel, 50 V, overnight) and transferred to nitrocellulose membranes (25 V, 1 hr) for Western blot analysis. After transfer, blots were stained in fast green to visualize proteins, then blocked in a 5% nonfat milk/TBST (200 mM Tris, 1.5 M NaCl, and 0.05% Tween 20) solution for 1 hr. Blots were rinsed with TBST and incubated in 1:500 nitrotyrosine polyclonal antibody for 1 hr followed by a 1:2500 monoclonal anti-rabbit IgG for an additional 1 hr incubation. The presence of 3-NT was detected with Pierce SuperSignal west pico chemiluminescent substrate.

Data analyses

Because of the high density staining of the 3-NT Western blots, densitometry was impractical for quantification [43,44] and thus we utilized a comparative system wherein each blot was run with an experiment-specific PO₄ + NO₂ positive control, and the resulting extent of 3-NT subjectively scaled. Lanes containing the experimental conditions were evaluated in comparison to the experiment- and blot-specific control lanes, denoting the amount of 3-NT as extensive (++++), significant (+++), modest (++), little (+), or none (−). Each blot was scored by blinded observation. Data represent means +/− standard deviations. Statistical analyses generally employed ANOVA with significance set at p <0.05.

Results

Previous ^•NO₂ reaction/diffusion studies, which employed the same lung surface model herein, noted that protein nitration was ablated when RCM were covered by an aqueous film containing DFO but not DTPA [32], but the mechanism remained equivocal. To elucidate mechanisms for the selective inhibition by DFO, we initially investigated the concentration profile of DFO-related 3-NT inhibition to determine a [DFO] efficacy threshold under these experimental conditions. RCM were exposed (30 min, 4.5 ppm ^•NO₂ in air) using only DFO in buffer and protein nitration evaluated. Concentrations as low as 0.5 μM DFO appreciably decreased exposure-mediated nitration (Fig. 2), suggesting that only small concentrations were required despite the fact that some DFO reaction with solute ^•NO₂ would potentially occur [7].

Figure 2. Effect of Initial DFO Concentration on ^•NO₂-Mediated RCM Nitration.

RCM, adhered in Petri dishes, were overlain with PO₄ buffer (10 mOms, pH 7.0) plus increasing initial DFO concentrations (0 – 5.0 μM) and exposed to 4.5 ppm ^•NO₂ for 30 min with intermittent cyclic tilting of the glass exposure chamber. DFO was added just prior to exposure onset. Post exposure, RCM were harvested and 3-NT formation assessed via Western analysis as described in Methods (n ≥ 4 independent RCM preparations and related exposures). Nitrated BSA was utilized as a positive control standard for 3-NT. Illustrated in this exampled gel is the fact that even 5 μM DFO extensively inhibited 3-NT formation (herein scored as “—“) while 1 μM constrained formation to little (+) and even 0.5 μM reduced generation to a significant (+++) extent. 0 μM DFO comparatively illustrates our subjective scale of extensive (++++). In the absence of ^•NO₂ exposure, note that basal levels of 3-NT within the adhered RCM were below detectable levels. Differences across the [DFO] were statistically significant.

We subsequently characterized the temporal course of DFO-mediated nitration inhibition using a [DFO] above that shown to be efficacious for a 30 minute exposure period. Figure 3 illustrates that an initial aqueous phase concentration of 25 μM DFO limited RCM 3-NT formation to near or below detectable levels for at least 90 min of continuous ^•NO₂ exposure. If one considers the mass balance between ^•NO₂ delivery and the DFO pool, approximately 20 μmoles of ^•NO₂ were delivered to the chamber over the 90 min time span (4.5 ppm = 8.46 ng/ml; (8.46 ng/ml • 1200 ml/min • 90 min)/46 ~ 20 μmole). Each RCM dish initially contained 50 nmole DFO so that even with 6 dishes concomitantly exposed (2 ml/dish • 6 dishes = 0.3 μmoles DFO), there was at least a 60-fold excess of ^•NO₂ available for mass transfer and reaction, although [^•NO₂]_exit indicated somewhat modest ^•NO₂ gas phase disappearance throughout the 90 min exposure period, consistent with limited aqueous substrate reaction [30]. Thus, the prolonged inhibition of protein nitration suggests that pathway(s) other than direct DFO scavenging of solute ^•NO₂ likely predominated.

Figure 3. Time Course of ^•NO₂-Mediated RCM Nitration in the Presence and Absence of Desferrioxamine.

RCM, adhered in Petri dishes, were overlain with PO₄ buffer (10 mOms, pH 7.0) with and without 25 μM DFO and exposed to 4.5 ppm ^•NO₂ for 30 – 90 min with intermittent cyclic tilting of the glass exposure chamber. DFO was added just prior to exposure onset. Post exposure, RCM were harvested and 3-NT formation assessed via Western analysis as described in Methods. Nitrated BSA was utilized as a positive control standard for 3-NT. As can be seen from this exampled gel (n ≥ 4 independent RCM preparations and related exposures), under these experimental conditions DFO substantially restricted 3-NT formation for at least 90 min of continuous ^•NO₂ exposure. Also note that the extent of RCM nitration in the absence of DFO increased with exposure time but was extensive (++++) even by the 30 min time point.

As a basis for direct comparison of aqueous phase reactants, 3-NT formation was assessed using a well-documented facile ^•NO₂ reactive substrate [29,30]. RCM were exposed (30 min; 4.5 ppm ^•NO₂) while covered with initial AH₂ concentrations ranging from 10 to 100 μM. As shown in Table 1, a pronounced concentration effect was observed wherein only 100 μM AH₂ substantially limited exposure-induced nitration over 30 min. At the lower initial concentrations (≤ 50 μM), no AH₂ could be detected at the end of the exposure period (data not shown). Thus, the differential extent of 3-NT formation across [AH₂] was likely related to reactive absorption-mediated depletion of AH₂, so that the temporal course of protein nitration was a function of aqueous phase conditions that modulated ^•NO₂ diffusion to the RCM. Under the specific employed conditions of initial 100 μM AH₂, ^•NO₂ delivery rate, aqueous phase mixing (tilting), total aqueous phase surface area, and exposure time, there was a sufficient initial AH₂ pool to enable continued reaction which served to scavenge ^•NO_2, thus constraining diffusion which prevented membrane protein nitration.

TABLE 1.

Influence of Initial Ascorbate Concentrations on ^•NO₂-Induced Membrane Tyrosine Nitration

Aqueous Phase	3-Nitrotyrosine
100 μM AH2	—
50 μM AH2	+ + +
25 μM AH2	+ + +
10 μM AH2	+ + + +

Red cell membranes (RCM) were covered with differing initial reduced ascorbate (AH₂) concentrations and exposed to 4.5 ppm ^•NO₂ in air for 30 min with cyclic tilting (2 min/side). Following exposure, RCM were washed, membrane proteins isolated, and the presence of 3-nitrotyrosine residues determined by Western analysis. As described in Methods, the extent of 3-NT formation is denoted as extensive (++++), significant (+++), modest (++), little (+), or none (-).

Using our well-established flask exposure approach to determine relative reactivities [30], we compared the ability of DFO, HBED, SHA, HU, and AH₂ to drive ^•NO₂ reactive absorption and estimate ^•NO₂ uptake:NO₂⁻ formation ratios. As illustrated in Figure 4, compared to AH₂, the hydroxamic acids all showed relatively modest reactivity towards ^•NO₂. Previous studies have shown a 1:1 ratio between ^•NO₂ uptake and NO₂⁻ formation for a variety of substrates [30,32], which generally held true herein. Deviations from a 1:1 ratio were likely due to ^•NO₂ mass balance measurement errors since DFO-mediated ^•NO₂ uptake rates, for example, were quite low relative to ^•NO₂ delivery but ^•NO₂ inflow rates needed to be sufficient to preclude ^•NO₂ as the limiting reagent. On the other hand, both 50 and 100 μM AH₂ displayed approximately equivalent uptake and NO₂⁻ formation, likely due to saturation of the interfacial transfer rates, which we have previously documented [37]. Furthermore, 30 min exposures of greater AH₂ volumes (with DTPA to reduce adventitious Fe-initiated autoxidation; volumes permitted multiple sample withdrawals) resulted in appreciable loss of reduced ascorbate while only marginal oxidation of DFO could be detected (Supplementary Data). Although it is difficult to completely rule out that DFO is a direct ^•NO₂ scavenger in our model system (see Fig 6), one may rank DFO against AH₂ for this capacity on the basis of ^•NO₂ flux calculated from the decrease in substrate (DFO or AH₂) during a 30 min exposure (see Supplementary Data) assuming the same reaction stoichiometry for both substrates (Substrate:^•NO₂ = 1:2). The substrate decay rates are coupled to ^•NO₂ flux and reflect reactivities for NO₂. The ^•NO₂ flux into DFO was small (57 ± 21 nmoles ^•NO₂/10 min) and can be compared to the large flux into AH₂ (374 ± 28 nmoles/10 min) (Table S1). It is interesting to note that with AH₂, the decay in concentration is nearly independent of the [AH₂] from 50 down to 20 μM, strongly suggesting interfacial saturating conditions. Thus, the ^•NO₂ flux with AH₂ ~ 7× larger than for DFO but would be even larger if the interface was not under saturation conditions. Moreover, the NO₂⁻ formation data presented in Figure 4 suggests that some DFO may be decaying by reacting with (NO₂)_solute by simple diffusion rather than by driving ^•NO₂ reactive uptake. These analyses indicate that DFO is a considerably less effective direct scavenger of ^•NO₂ than AH₂ and thus a direct scavenging mechanism is less likely to occur with DFO.

Figure 4. The Extent of ^•NO₂ Uptake and Nitrite Formation across Varying Substrates.

As described in methods, solutions at the listed concentrations were exposed under quasi-steady state (with respect to ^•NO₂), well stirred, conditions in small flasks. ^•NO₂ uptake was calculated via the mass balance of gas phase ^•NO₂ across the flasks and the nitrite concentrations determined via the Greiss reaction based on daily standard curves (containing the respective substrate) utilizing NaNO₂ as the standard. Data are presented as means +/− one standard deviation of at least 3 independent observations. This approach provides a relative measure of aqueous phase reactivity since the mixing rate, ^•NO₂ delivery rate, and exposure time were all held constant so that differences in uptake and product (nitrite) formation result from the differential driving force for ^•NO₂ interfacial transfer maintained by reaction between solute ^•NO₂ and the substrate of interest. As can be seen, all but the AH₂ showed marginal, if any, uptake rates above the buffer control, suggesting relatively limited reactivity with ^•NO₂. AH₂, which is a well documented facile reactant for ^•NO₂ produced significantly greater rates of uptake and nitrite formation that approximated the 1:1 stoichiometry observed previously. The apparent lack of dose response between 50 and 100 μM likely stemmed for saturation of the interfacial mass transfer rates, which has also been formerly characterized.

We subsequently compared the ability of the tested hydroxamic acids to inhibit ^•NO₂ exposure-mediated RCM nitration. Table 2 displays notable differences across the various moieties. Compared to 50 μM DFO, which inhibited 3-NT formation to below detectable levels, 200 μM SHA (below detection) and 200 μM BHA (largely) also inhibited membrane nitration. At a somewhat lower concentration, to more closely mimic therapeutic situations, 25 μM HU also displayed notable, but not complete, inhibitory activity. However, like the other non-hydroxamic acid DTPA, 200 μM HBED poorly inhibited exposure-mediated nitration. Thus DFO served as the most robust inhibitor of ^•NO₂ mediated tyrosine nitration

TABLE 2.

Influence of Hydroxamic Acids and Related Compounds on ^•NO₂-Induced Membrane Tyrosine Nitration

Aqueous Phase	3-Nitrotyrosine
50 μM DFO	—
200 μM SHA	—
200 μM BHA	+
25 μM Hydroxyurea	+ +
200 μM HBED	+ + +
400 μM DTPA	+ + + +

Red cell membranes (RCM), adhered in petri dishes, were covered with various aqueous phase constituents and exposed to 4.5 ppm ^•NO₂ in air for 30 min with cyclic tilting. Following exposure, RCM were washed, membrane proteins isolated, and the presence of 3-nitrotyrosine residues determined by Western analysis. As described in Methods, the extent of 3-NT formation is denoted as extensive (++++), significant (+++), modest (++), little (+), or none (-). 3-NT was undetectable during air-only exposures.

To determine whether DFO-related 3-NT inhibition occurred preferentially with membrane associated tyrosine residues, we performed aqueous bulk phase studies by directly bubbling ^•NO₂ in air through a solution of free tyrosine with and without DFO. Time dependent samples were obtained and 3-NT formation evaluated via spectrophotometry. Figure 5 shows a representative scan that illustrates 3-NT formation in the absence of DFO. However, with DFO addition we noted substantial nitration inhibition even though tyrosine was in appreciable excess. Exposures were conducted for 120 min which was the time required to generate a readily detectable 3-NT peak via this approach. The lack of complete inhibition by DFO may have been due to its direct but relatively limited reaction with ^•NO₂ and/or the stochastic probability of tyrosyl radicals interacting with continuously supplied ^•NO₂ rather than DFO interception. Under these conditions, per published spectra [42], we did not observe dityrosine accumulation again likely due to the limited probability of two tyrosyl radicals directly interacting. However, the data do illustrate that the DFO inhibitory effect occurred under purely aqueous conditions and thus membrane tyrosine residues that protrude outside the hydrophobic domain may also be affected by its presence.

Figure 5. Effect of Desferrioxamine on Aqueous Bulk Phase Tyrosine Nitration.

8 ppm ^•NO₂ (mixture of 60 ml/min ^•NO₂ and 520 ml/min air) was bubbled through a 100 ml solution of 0.8 mM tyrosine for 120 min at 25° C. The presence of 3-nitrotyrosine in the solution was measured spectrophotometrically (OD₄₃₈) at 30 min intervals (long dashed line). At each interval, 2.25 ml of exposed solution were combined with 0.25 ml 1 N NaOH and scanned from 800 – 200 nm. Baseline corrections were made with PO₄ + NaOH solution. Relative to the ^•NO₂ inflow rate, exposed solution volume, tyrosine concentration, and exposure time, the yield of 3-NT based on its alkaline extinction coefficient (~ 4000 M⁻¹ cm⁻¹) is within experimental norms. Conditions were then repeated using a 50 μM DFO + 0.8 mM tyrosine solution (short dashed line). Data shown includes t = 0 min and t = 120 min readings. At t = 0 min, spectrophotometric scans of tyrosine and tyrosine + DFO solutions overlie. The addition of DFO to the ^•NO₂ exposed tyrosine markedly reduced 3-NT accumulation.

We subsequently investigated the consequences of DFO-Fe binding with regard to nitration inhibition. Addition of FeCl₃ alone did not alter ^•NO₂ exposure-related RCM nitration (data not shown), suggesting that under these experimental conditions nitration was Fe-independent. Furthermore, preloading DFO with a Fe molar excess ablated inhibition (data not shown), substantiating that the free hydroxamic moieties were likely responsible for preventing ^•NO₂-mediated nitration. To further test these assumptions regarding Fe-independence and hydroxamic acid-dependence, we preincubated adhered, lysed RCM with either DFO or DTPA and conducted ^•NO₂ exposures with either PO₄ buffer alone or the other chelator present (Table 3). DTPA removal of adventitious iron followed by ^•NO₂ + PO₄ exposures resulted in extensive 3-NT formation. However, DTPA pretreatment followed by exposures containing DFO limited nitration below detection. Conversely, pretreatment with DFO followed by exposures with either PO₄ alone or DTPA resulted in no discernible diminution of 3-NT formation, suggesting that inhibition required the presence of DFO during ^•NO₂ exposure regardless of any pretreatments.

TABLE 3.

Differential Effects of Iron Chelator Pretreatment and Presence during Exposure on ^•NO₂-Induced Membrane Tyrosine Nitration

Pretreatment Chelator	Exposure Aqueous Phase	3-Nitrotyrosine
DTPA	PO4	+ + + +
DTPA	50 μM DFO	—
DFO	PO4	+ + + +
DFO	400 μM DTPA	+ + + +

Red cell membranes (RCM), adhered to glass petri dishes, were covered with an aqueous solution containing either 400 μM DTPA or 50 μM DFO. After two 10 min treatments with periodic mixing (gas phase = air), systems were thoroughly rinsed to remove bound Fe complexes and remaining unbound chelator. RCM were then covered with PO₄ buffer alone or buffer + chelator and exposed to 4.5 ppm ^•NO₂ in air for 30 min with cyclic tilting (2 min/side). Following exposure, RCM were washed, membrane proteins isolated, and the presence of 3-nitrotyrosine residues determined by Western analysis. As described in Methods, the extent of 3-NT formation is denoted as extensive (++++), significant (+++), modest (++), little (+), or none (-).

Lastly, in an attempt to mimic what might transpire on the lung surface wherein activated neutrophils in close approximation to epithelial apical membranes could drive cellular nitration reactions via peroxidase activity, MPO was “bound” to RCM prior to addition of NO₂⁻/H₂O₂ and DFO (Table 4). Adhered RCM incubation with MPO, followed by thorough washing, resulted in appreciable remaining MPO activity within the Petri dishes. In combination, MPO + NO₂⁻ + H₂O₂ generated extensive RCM nitration, as anticipated [19,45]. In the absence of either NO₂⁻ or H₂O₂ no detectable nitration occurred and, under these experimental conditions, NO₂⁻ + H₂O₂ in the absence of MPO also did not generate 3-NT. Addition of the MPO inhibitor 4-aminobenzoic hydrazide to MPO + NO₂⁻ + H₂O₂ also completely blocked 3-NT formation, confirming the requirement for MPO catalytic activity. Importantly, addition of 50 μM DFO to the aqueous phase also limited nitration to below detection even though the RCM-bound MPO should have been generating its nitrating species in very close proximity to the RCM surfaces. Because at these low DFO concentrations MPO catalytic activity was essentially unaffected (data not shown), the results suggest that DFO had sufficient diffusivity within and/or along the RCM to block nitration. Thus, partitioning of DFO into the membranes likely contributed to its inhibitory actions rather than exclusive interception of (^•NO₂)_solute diffusing from the gas-liquid interface and/or ^•NO₂ generated at the RCM surfaces via MPO activity.

TABLE 4.

Effect of Desferrioxamine on Myeloperoxidase-Mediated Membrane Tyrosine Nitration

30 nM MPO	100 μM NO2−	100 μM H2O2	50 μM DFO	100 μM 4-Aminobenzoic hydrazide	3-Nitrotyrosine
X	X	X	-	-	+ + + +
X	X	-	-	-	—
X	-	X	-	-	—
-	X	X	-	-	—
X	X	X	-	X	—
X	X	X	X	-	—

Adhered red cell membranes (RCM) were covered with 0.7 units/ml myeloperoxidase (MPO) solution, continuously rocked in covered dishes for 15 min, the covering solution removed, and the RCM washed leaving only RCM bound MPO (~ 0.56 units/ml MPO activity). RCM were then covered with various aqueous phase constituents and exposed to air for 30 min with continuous rocking in covered dishes. Following exposure, RCM were washed, membrane proteins isolated and the presence of 3-NT determined by Western analysis. As described in Methods, 3-NT formation is denoted as extensive (++++), significant (+++), modest (++), little (+), or none (-).

Discussion

^•NO₂ is now recognized as a ubiquitous oxidant that is derived from both exogenous and endogenous sources. The lung surface is relatively unique because of the composite exposure potentials stemming from direct ^•NO₂ inhalation and/or intrapulmonary oxidation of inhaled ^•NO, and responses to numerous stimuli, including inflammation, which endogenously generate ^•NO₂. Because ^•NO₂ is central in the overall biology of reactive nitrogen species (RNS) [11–13] and because protein tyrosine nitration, wherein ^•NO₂ serves as the terminal nitrating species, is widely considered a hallmark index of RNS biochemical interactions, it is important to elucidate the factors that might modulate the net generation of nitrated tyrosine residues.

Due to its high affinity binding, DFO is the most commonly employed Fe chelator and has been used in numerous studies to identify the importance of Fe in redox reactions (e.g., Fenton chemistry) as well as its active site contributions in protein catalysis. Although the binding affinities are somewhat less, other hydroxamic acids chelate metals as well. Previous observations, both direct and inferential [3,4,7–9,34]), have suggested that DFO may function beyond its Fe binding activities. Other related chemical species, such as HU, have broad applicability as therapeutic agents with a variety of proposed pharmacologic modalities, including, for example, inhibition of enzymes such as ribonucleotide reductase (e.g., [46,47]). Since transient radicals, including tyrosyl, are key in the catalytic actions of numerous proteins [48,49] and since tyrosine nitration has been shown to affect function in diverse proteins [50–58], elucidating whether DFO and other hydroxamic acids may modulate nitration reactions and biological functions independent of chelation is important for understanding mechanisms, and how hydroxamic acids may influence and potentially confound in vitro and in vivo experimental outcomes and interpretations.

To clarify the extent that reagent DFO might confound studies focused on protein nitration reactions and Fe contributions, a minimalist approach was utilized that incorporated both a well-defined nitration target (RCM) and gas phase ^•NO₂ exposures. Although the RCM model system involves mass transfer and reaction/diffusion complexities, delivery of gas phase ^•NO₂ (stable ^•NO₂ aqueous solutions cannot be generated/employed) limits it to the sole RNS introduced and avoids potential confounding from other reactive species generated by use of ^•NO₂ precursors (e.g., ONOOCO₂⁻), flash photolysis, or pulse radiolysis [7,12,32–35,59], thus affording a methodologic advantage.

Figures 2 and 3 illustrate that DFO effectively prevented membrane protein nitration at very low concentrations (≤ 5 μM) and over a relatively prolonged exposure period (≥ 90 min). Considering that ^•NO₂ was not the limiting reagent, due to the steady state exposure conditions, and that DFO was a relatively poor substrate for ^•NO₂ reactive absorption (Fig. 4 and Fig. S1), it does not seem likely that 3-NT inhibition resulted solely from direct (^•NO₂)_solute scavenging. In this regard, a rate constant for reaction of DFO with ^•NO₂ has been reported (k = 7.6 × 10⁶ M⁻¹s⁻¹ [7]) which is smaller than the rate constant for AH₂ with ^•NO₂ (two values in general agreement with one another have been reported for ascorbate; 1.8 × 10⁷ M⁻¹s⁻¹ and 3.5 × 10⁷ M⁻¹s⁻¹ [53,60,61]). It is interesting to note that to the best of our knowledge, the rate constant reported for DFO with ^•NO₂ is the only rate constant regarding a complex hydroxamic acid. Recently, Samuni and Goldstein reported that the rate constant for acetohydroxamic acid, the simplest member of the hydroxamic acid series, is less than 2 × 10⁴ M⁻¹s⁻¹ [62] which contrasts sharply with the larger value reported for DFO [7] and argues against direct scavenging of ^•NO₂ by DFO.

Ascorbate, which is a more facile ^•NO₂ reactant than DFO, showed some degree of 3-NT inhibition at initial concentrations as low as 25 μM. It should be noted that AH₂ will be consumed during the course of exposure so that ^•NO₂ diffusion through the overlying aqueous film is a balance among its delivery and interfacial mass transfer rates, diffusion of both solute ^•NO₂ and aqueous substrate, rate of aqueous substrate consumption, and depth of the aqueous film, among others. Thus, if the facile reactive substrate concentrations drop below a critical threshold, ^•NO₂ diffusion to directly interact with the RCM becomes more favorable. Consequently, Table 1 shows that in the presence of a sufficiently large initial substrate pool, ^•NO₂ was consumed via reaction and although secondary oxidants may have been produced thereby driving RCM oxidation, insufficient ^•NO₂ reached the membranes to add to any existing tyrosyl radicals. If aqueous antioxidant concentrations remain sufficiently elevated, both ^•NO₂ and secondary oxidants may be quenched thereby preventing or repairing tyrosine oxidation and thus limit consequent nitration.

The DFO-Fe complex can function as a reductant when in the DFO-Fe²⁺ state [63]. Because our use of lysed red blood cells, and results from previous studies, document the presence of adventitious iron, such a potential mechanism for tyrosyl radical reduction should not be ignored. However, although the DFO bound Fe may remain redox active, it would require a constant source of electrons during ^•NO₂ exposure to maintain its reductive capacity. We believe that the presence of appreciable DFO-Fe²⁺ in our system was unlikely since an applicable electron source was present only during the AH₂ studies. Even then, the continuous ^•NO₂ influx would have likely resulted in preferential ^•NO₂ reduction over tyrosyl radical repair. In the absence of aqueous phase AH₂, no obvious electron source would have been available. Despite the complexities of Fe redox activities, our studies utilizing ferrioxamine or adventitious Fe removal tend to confirm this. Iron-saturated DFO had no inhibitory capacity and adventitious Fe removal (Table 3) required DFO per se present to block ^•NO₂-mediated membrane nitration.

The ability of DFO to inhibit nitration occurred under a variety of physicochemical situations in both aqueous and hydrophobic compartments. Figure 5 illustrates that DFO constrained nitration under purely aqueous phase conditions while Table 4 demonstrates similar results with oxidants and/or nitrating species generated via MPO activity at or near membrane surfaces. We also considered DFO may directly inhibit MPO since it has been reported that certain hydroxamic acids may compete with MPO substrates (chloride, thiocyanate, and nitrite) for binding to the same active site, although inhibition was only observed at higher hydroxamic acid concentrations than we employed [64,65]. Interestingly, a published report on the inhibition of MPO by SHA used guaiacol (a phenolic compound) as a co-substrate for MPO [64]. We believe these results could have been confused by reaction of SHA with the guaiacol phenoxyl radicals because we did not observe MPO inhibition by DFO at the highest concentration levels we employed while using 3,3′,5,5′-tetramethylbenzidine as a co-substrate for MPO (data not shown). Thus, based on our empirical observations of sustained MPO catalytic activity in the presence of 50 μM DFO, we dismissed competitive inhibition of MPO as a mechanism for 3-NT inhibition during the MPO/DFO studies.

Irrespective of the ^•NO₂ source (gas phase; MPO), it is difficult to precisely discern where within the membrane structure nitration occurred but, despite the lack of a “lens” effect driving preferential ^•NO₂ partitioning into the membrane hydrophobic regions [21], one would anticipate nitration of tyrosines both within and protruding out of the membrane proper. The aqueous bulk phase studies showed a right spectral shift in the DFO+Tyr scans (Fig. 5; 325–375 nm) which may be attributable to oxidation products including adducts such as DFO-NO₂, DFO-Tyr, etc., among a variety of others. Further analyses are needed to resolve the precise genesis of the spectral shift. However, this does not detract from the observations that even under bulk phase, well mixed conditions, the presence of DFO appreciably reduced the extent of 3-NT formation. Although dityrosine formation has been observed in a number of studies [59,62,66], its apparent absence herein was reasonable given the tyrosyl radical generation constraints under these study conditions.

Regardless of the mechanism(s) by which tyrosine oxidation occurred, we propose that DFO predominantly inhibited its nitration according to the mechanism below where R^• represents a free radical such as ^•NO₂ or a lipid-derived radical:

The site of radical formation in DFO is in accordance to the structure proposed by Morehouse et. al. [67]. Per the above, DFO competes with ^•NO₂ for reaction with protein tyrosyl radicals, repairs the oxidized tyrosine residues, and inhibits nitration. DFO is only slightly lipophilic and based on its octanol-water partition coefficient, approximately 1.4% of the aqueous concentration should have partitioned into the lipid bilayer [68]. However, we believe that both aqueous and membrane-associated DFO participated in membrane protein tyrosyl radical reduction with the former reducing the exposed tyrosyl radicals while the latter reduced tyrosyl radicals embedded deeper in the membrane.

For our experimental conditions, DFO was likely to be present (both in membranes and the aqueous milieu) at higher concentration than ^•NO₂ since ^•NO₂ is sparingly soluble in aqueous media [69] and only slightly more soluble in organic media [70,71]. Moreover, recent analyses suggest that ^•NO₂ does not partition into membranes to the same extent and thus undergo accelerated reaction as has been shown for ^•NO [20,21]. The rate constants for the tyrosyl radical and DFO reactions with ^•NO₂ are 3 × 10⁹ M⁻¹s⁻¹ and 6.3 × 10⁶, respectively [72]; protein tyrosyl radicals are probably somewhat less reactive than tyrosyl radicals. Although the rate constant for nitration is some 400 times larger than the published rate constant for reaction with DFO, because [DFO] is expected to be much larger than [^•NO₂] at the reaction sites, we expect DFO to outcompete ^•NO₂ for the protein tyrosyl radicals. In this regard, one can calculate using a Henry’s law constant equal to 1.4 × 10⁻² M atm⁻¹ [70] and ideal gas behavior that for a gas phase [^•NO₂] equal to 4.5 ppm, the [^•NO₂] in the aqueous phase will approximate 4.5 × 10⁻⁶ atm × 1.4 × 10⁻² M atm = 6.3 × 10⁻⁸ M. This value should be

$$\frac{\text{rate}\text{of}\text{repair}}{\text{rate}\text{of}\text{nitration}}=\frac{k_{\mathit{TyrO}•/\mathit{DFO}}[\mathit{TyrO}•][\mathit{DFO}]}{k_{\mathit{TyrO}•/{NO}_2}[\mathit{TyrO}•][{NO}_2]}=\frac{k_{\mathit{TyrO}•/\mathit{DFO}}[\mathit{DFO}]}{k_{\mathit{TyrO}•/{NO}_2}[{NO}_2]}\ge \frac{(6.3\times {10}^6{\text{M}}^{-1}{\text{s}}^{-1})(50\times {10}^{-6}\text{M})}{(3\times {10}^9{\text{M}}^{-1}{\text{s}}^{-1})(6.3\times {10}^{-8}\text{M})}\approx 2$$

taken as an upper limit because of the high reactivity of ^•NO₂ but can be compared to the aqueous [DFO] equal to 50 μM used in our experiments, which is some 800 times larger than the calculated equilibrium [^•NO₂]. Consequently, it is reasonable to assume that in the aqueous compartment, the relatively greater DFO concentration could more than compensate for its lower reactivity with tyrosyl radicals, compared to ^•NO₂. Thus in compilation, a conservative comparison of the rate of DFO repair versus the rate of ^•NO₂ addition suggests that repair is favored in excess of two-fold over nitration. Any decrease in the rate constant that describes the DFO + ^•NO₂ reaction would increase the computed repair preferentiality.

We found that tyrosine nitration was completely inhibited by DFO, suggesting effective competition of DFO for trans-membrane protein tyrosyl radicals as well. This may be due to a steep [^•NO₂] gradient with the membrane concentration being much lower than what occurs in aqueous solution. This would be a kinetic rather than a thermodynamic effect because ^•NO₂ is slightly lipophilic [21,70,71] and may be due to faster consumption of ^•NO₂ (by chemical reactions in the membrane) than diffusion of ^•NO₂ into the membrane. Figure 6 is a schematic cartoon depicting these numerous complexities and contributory factors.

Figure 6. Schematic of the Proposed Differential Mechanisms of Tyrosine Nitration Inhibition.

Under conditions similar to the lung surface compartment, reactive uptake of ^•NO₂ may occur by solutions of AH₂ or DFO covering RCM that are chemically bound to the bottom of Petri dishes. At ELF pH (~7.0), essentially all AH₂ will exist as ascorbate (AH⁻). This oversimplified cartoon underscores critical differences between these two substrates that result in efficient inhibition of protein tyrosine nitration by DFO as: (1) a larger flux of NO₂ into the AH⁻ solution, indicated by the bold arrow for NO₂ uptake, that results from the higher reactivity of AH⁻ for NO₂ (also in bold) as compared to DFO. Notice that the higher reactivity of AH⁻ facilitates its role as a direct scavenger of NO₂ but also results in faster depletion of AH⁻, and (2) the larger log D for DFO that allows for partitioning of DFO into the cell membrane and thereby favors reaction of DFO with inter-membrane tyrosyl radicals while AH⁻ remains primarily in the aqueous phase which hinders this reaction [log D (DFO) = −3.56; log D (AH⁻) = −5.61 at pH 7 (values taken from SciFinder and calculated using Advanced Chemistry Development (ACD/Labs) Software V11.02 (© 1994–2012 ACD/Labs)].

In further support for our mechanism, hydroxamic acids and hydroxyl-amines have been reported to reduce the active site tyrosyl radical of ribonucleotide reductase, which is consistent with reports of DFO and HU inhibiting this enzyme [46,47]. Consequently, application of hydroxamic acids could affect the catalytic activities of other proteins with active site tyrosyl radicals. Non-enzymatic proteins may be important considerations as well [52,54,56–58] since tyrosine nitration of signaling molecules may also disrupt their normal function. Tyrosine nitration may affect protein function and structure, and change the rate of proteolytic degradation of nitrated proteins. Nitration often leads to loss of protein activity but recent proteomic analyses have revealed that a variety of cellular processes may be affected by protein tyrosine nitration in various ways. These processes include energy metabolism, protein, nucleic acid and ion binding, cellular structure and regulation, enzymatic activity control, and cellular signaling, including apoptosis. Thus, therapeutic intervention with hydroxamic acids when their intended use is to chelate ferric ions or inhibit histone deacetylases may have secondary effects. Further investigations are needed to determine whether these secondary effects due to inhibition of protein tyrosine nitration could be beneficiary, or detrimental, against a background of poisoning or disease wherein protein tyrosine nitration levels are found to be elevated.

Numerous investigations have utilized DFO to identify Fe as a critical effecter driving redox perturbations. However, although DFO will clearly sequester Fe, and has been shown to inhibit peroxynitrite-related interactions, if DFO is used as an investigational reagent the precise mechanisms by which altered outcomes occur may not simply involve removal of redox active Fe, limiting nutraceutical Fe availability, or scavenging of oxidant and/or nitrating species but rather, direct action on tyrosyl radical intermediates. Furthermore, DFO enters the cell mostly via endocytosis, which is the same pathway as transferrin, and thus during lysosomal maturation and acidification the Fe is liberated to form “free” redox active (Fenton) Fe. Thus, DFO may serve as an “antioxidant”, Fe chelator, and radical scavenger. Indeed a major new clinical form of DFO is “starch-DFO” which apparently confines its entry to endocytic pathways [73,74]. Such factors need to be considered, for example, if DFO is added to tissue culture media to help preserve supplemental AH₂, is included in lung lavage fluids to reduce artifactual antioxidant oxidation, or if administered in vivo if subsequent cell biology studies are to be performed. Consequently, because of the varied functions of DFO, there may be a need to reevaluate some proposed mechanisms related to Fe-mediated events/interactions via the use of non-hydroxamic acid chelators (e.g. DTPA) and a requisite need to both interpret experimental results with appropriate caution and include confirmatory studies using ancillary approaches.

Highlights.

• DFO robustly inhibits tyrosine nitration (3-NT) independent of Fe chelation.

• The mechanism likely involves tyrosyl radical repair preventing ^•NO₂ addition.

• More facile reactants scavenge ^•NO₂ but lose 3-NT inhibition due to consumption.

• DFO effectively inhibits 3-NT in both aqueous and hydrophobic compartments.

• DFO multifunctionalities should be considered during its reagent and clinical use.