Post-Trapping Derivatization of Radical-Derived EPR-Silent Adducts

Jun 19, 2012 - Free radicals are conventionally detected by electron paramagnetic resonance (EPR) spectroscopy after being trapped as spin adducts...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Post-Trapping Derivatization of Radical-Derived EPR-Silent Adducts: Application to Free Radical Detection by HPLC/UV in Chemical, Biochemical, and Biological Systems and Comparison with EPR Spectroscopy Karim Michail*,†,‡ and Arno G. Siraki† †

Faculty of Pharmacy and Pharmaceutical Sciences, University of Alberta, Edmonton, Alberta T6G 2E1, Canada Faculty of Pharmacy, Alexandria University, Alexandria 21521, Egypt

‡

S Supporting Information *

ABSTRACT: Free radicals are conventionally detected by electron paramagnetic resonance (EPR) spectroscopy after being trapped as spin adducts. Albeit this technique has demonstrated utmost efficacy in studying free radicals, its application to biological settings is intrinsically hampered by the inevitable bioreduction of radical-derived paramagnetic adducts. Herein, we describe a reliable technique to detect and quantify free radical metabolites, wherein reduced alkyl- and phenyl-5,5-dimethyl-1-pyrroline N-oxide (DMPO) adducts are converted into ultrastable N-naphthoate esters. To mimic the ubiquitous in vivo microenvironment, bioreductants, exogenous thiols, and sodium borohydride were studied. Nitroxyl reduction was confirmed using EPR and triphenyltetrazolium chloride. The formation of the N-naphthoyloxy derivatives was established by liquid chromatography/mass spectrometry (LC/MS). The derivatives were chromatographed using a binary eluent. HPLC and internal standards were synthesized using Grignard addition. The labeled DMPO adduct is (1) fluorescent, (2) stable as opposed to nitroxyl radical adducts, (3) biologically relevant, and (4) excellently chromatographed. Applications encompassed chemical, biochemical, and biological model systems generating C-centered radicals. Different levels of phenyl radicals produced in situ from whole blood were successfully determined. The method is readily applicable to the detection of hydroxyl radical. Analogously, DMPO, the spin trap, could be detected with extreme sensitivity suitable for in vivo applications. The developed method proved to be a viable alternative to EPR, where for the first time the reductive loss of paramagnetic signals of DMPO-trapped free radicals is transformed into fluorescence emission. We believe the proposed methodology could represent a valuable tool to probe free radical metabolites in vivo using DMPO, the least toxic spin trap.

T

transient radicals to yield relatively persistent paramagnetic spin adducts whose hyperfine splitting pattern and magnitude usually reflect the identity and nature of the initial radical trapped.8,9 Though the spin-trapping technique in conjunction with EPR spectroscopy is an effective analytical tool utilized conventionally to detect and identify free radicals,10 it is susceptible to possible pitfalls, such as (1) decay of the spin adduct (t1/2 = minutes to hours);11 (2) the EPR-active nitroxyl (aminoxyl) species is less significant to biological systems, where reduction of paramagnetic adducts into EPR-silent products by extra-/intracellular reductants and antioxidants seems inescapable;12 (3) the resolution and interpretation of composite spectra from mixtures of spin adducts is perplexing due to overlaps between adducts giving rise to barely distinguishable spectra;13,14 (4) the relative sophistication and

he inability to harness the universally generated free radicals in the course of diverse biological processes is widely accepted as a contributor to cellular damage.1 However, controlled production of free radicals and radical-derived reactive species in a biological milieu have defined physiological functions unless the redox homeostasis is disturbed.2−4 From another perspective, harmful levels of radical intermediates may arise from hepatic and peroxidase-driven extrahepatic metabolic biotoxification of some drugs and xenobiotics.5−7 Currently, simple and reliable techniques to detect and quantify low levels of free radicals, particularly in biological matrixes and/or in whole animal models, are mostly lacking. This is a leading challenge which continues to hamper our understanding of the mechanisms of production, localization, kinetics, and disposition of free radical precursors in living organisms. Electron paramagnetic resonance (EPR) spectroscopy is considered the most direct method to detect free radicals. It is typically coupled to a preceding spin-trapping step. Spin-trapping agents are diamagnetic scavenger probes which react readily with © 2012 American Chemical Society

Received: May 8, 2012 Accepted: June 19, 2012 Published: June 19, 2012 6739

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746

Analytical Chemistry

Article

as HPLC or internal standards. Both alkyl and phenyl magnesium bromide reacted with DMPO under the designated experimental conditions (Supporting Information: S3) to give predominantly the corresponding 5-substituted cyclic HAs as major products (Supporting Information: Scheme S1). Reported yields for 5-methyl-, 5-ethyl-, and 5-phenyl-DMPO HAs by Grignard synthesis are in the range of 90%.24,25 Aerial oxidation of these cyclic N,N-dialkyl HAs gave rise to trace amounts of the respective cyclic aminoxyl paramagnetic DMPO spin adducts (Supporting Information Scheme S1, traces a and b).26 As demonstrated by a nitroxide triplet (Supporting Information Scheme S1a) in EPR, a dimer aminoxyl having no β-hydrogens was also detected in case of Grignard-derived, alkyl-substituted but not phenyl-substituted DMPO HA (DMPOH). This is most probably ascribable to the presence of traces of 2-methyl-DMPO nitrone due to two-electron oxidation of alkyl HA and subsequent formation of a nitrone HA aldol-type dimer through a self-condensation reaction.27 Upon reduction using ascorbic acid (ASC), paramagnetic species turn into EPR-silent forms (Supporting Information Scheme S1, traces c and d). For a fast and convenient preparation of various standards, tetrahydrofuran proved to be a superior solvent to diethyl ether in terms of both yields and handling ease. The reaction conditions were optimized to maximize the yield of 5-substituted DMPOH on expense of other byproducts. Excessive heating (>60 °C) or prolonged reaction (>15 min) with alkyl Grignard reagents resulted in a gradual loss of HAs due to self-dimerization.28 Heating at 80 °C for 20 min resulted in a deep yellow solution containing no HAs but mostly condensation products. Effect of Different Reducing Agents on Paramagnetic C-Centered DMPO Adducts. To mimic the in vivo reducing environment, various endogenous and exogenous reducing agents were tested to accomplish the ex vivo conversion of DMPO•/CH3 and DMPO•/C6H5 into their reduced counterparts. Moreover, chemical reduction using NaBH4 was tested. The EPR spectrum was recorded before and after the addition of each reducing agent to the reaction mixture in the EPR flat cell. This confirmed the reduction of the produced radical adduct and ruled out interference with free radical generation or spin trapping. Figure 1 shows that, in procarbazine/ microperoxidase-11 (MP-11) system (Figure 1a), all reducing agents, namely, NaBH4, ASC, reduced glutathione (GSH), and N-acetyl cysteine (NAC), had a comparable reducing efficacy. In addition, ASC resulted in the formation of ascorbyl radical,

cost of EPR spectroscopy makes it generally less appealing. Alternatively, if a universal technique with high-throughput capacity, such as high-performance liquid chromatography (HPLC) combined with various detection methods, were to offer a superior alternative to EPR, it would need to (a) extend the lifetime of radical-trapped adducts, (b) detect the reduced EPR-silent adduct, (c) achieve a linear detector response, and (d) afford high-resolution separation and unambiguous identification of individual spin adducts. Previous attempts to detect free radical intermediates were focused intensively on the use of combinations of detectors and/or techniques in which EPR played a central role, such as the determination of chemically generated phenyl radicals, hydroxyl radical, and metal-catalyzed hydrazine-derived pentyl and phenyl radicals by liquid chromatography/electron paramagnetic resonance/mass spectrometry (LC/EPR/MS), tandem LC/EPR−LC/MS with dual spin trapping, and HPLC/ECD/UV/ESR/MS, respectively.15−17 In biological settings, 5,5-dimethyl-1-pyrroline Noxide (DMPO) and N-tert-butyl-α-phenylnitrone (PBN) are the two most commonly used spin traps for detecting free radicals. Only few studies not involving EPR spectroscopy employed DMPO, to detect oxygen- or sulfur-centered radicals,18−20 whereas the majority of studies utilized open chain nitrones, viz. PBN and α-(4-pyridyl-1-oxide)-N-tertbutylnitrone (4-POBN), to determine carbon-centered radicals.21−23 Nevertheless, adequate detection and quantification of spin-trapped radicals was not attainable. The objective of this work was to develop a technique which can afford robust longevity, enhanced detectability, and improved chromatographic retention of biologically relevant radical-derived adducts. Converting the radical-trapped adduct into a fluorescent, highly stable derivative appeared an interesting approach. However, this structural modification should occur exclusively after the addition of the radical to the spin trap in order not to interfere with the original spin trap’s physicochemical properties, its toxicity, or the rate constant of the trapping reaction. Ideally, the reduced EPR-silent adduct, the expected species in biological systems, should be the target of the derivatization reaction. We anticipated that for alkyl and phenyl radicals a good option would be to derivatize their DMPO spin adducts to the corresponding N-naphthoates after reduction into EPR-invisible cyclic hydroxylamines (HAs). The resulting derivative has several advantages, most significantly a linear optical response, stability, and direct pertinence to reallife matrixes, which enhance its potential for in vivo applications. To date, no specific HPLC methods in conjunction with optical detection exist for the determination of EPR-undetectable adducts. Notably, the majority if not all of the developed methodologies did not implement the use of appropriate HPLC and internal standards. To the best of our knowledge, there are no HPLC/UV methods available at the moment for the detection and quantification of drug- or xenobiotic-derived primary carbon-centered radicals. Herein, we present a viable and generic approach to detect and quantify free radicals in different systems and matrixes without the need of EPR, combinations of detectors, or laborious extraction steps.

■

RESULTS AND DISCUSSION Synthesis of HPLC and Internal Standards. Addition of the appropriate Grignard reagent to the nitrone spin trap provided a simple means to prepare different C-centered DMPO hydroxylamine adducts in sufficient yields which served

Figure 1. Effect of different reducing agents on DMPO•/CH3 derived from procarbazine/MP-11 biochemical system. 6740

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746

Analytical Chemistry

Article

nated into nitrones and HAs via a biomolecular decay mechanism.38 Attempts by Novakov et al. to determine both forms (i.e., nitroxyls and HAs of POBN) were hindered in part by alterations in the ratio of these redox states, and as a result they suggested a preceding reduction step by incubating the reaction mixture with ASC.23 Interestingly, the authors in that work reported that GSH was not able to reduce POBN•/CH3, an open chain heterocyclic nitroxide. In our hands, DMPO•/ CH3 was readily reduced using GSH with the formation of thiyl products (Figure 1d). Furthermore, in a biological milieu, the identification of radical-derived spin adducts is not a straightforward task since it is conceivable that numerous paramagnetic trapped-radical species with quite identical EPR spectra are formed, whereas only little structural information is bestowed by EPR spectroscopy. Most importantly, the flux of free radical generation, competing secondary targets for free radicals, and the lability of radical adducts remain key aspects impeding accurate detection and quantification of free radicals, especially in biomatrixes (e.g., blood or plasma) and in in vivo settings. To sum up, in biological systems, EPR-silent HAs are expected to be the predominant species; however, their determination ex vivo is difficult owing to electron shuttling with their nitroxyls counterparts. Naphthoylation of DMPO hydroxylamine adducts provided a means for their stabilization and optical detection besides considerably refining their chromatography (Supporting Information: Scheme S2). In presence of an organic base this reaction proceeds quickly and stoichiometrically at different temperatures (0−50 °C). 1Methylimidazole was the base catalyst of choice and afforded high and reproducible yields of the derivative.39 Hexane was the optimum extracting solvent and gave over 90% extraction recovery of the labeled product. This was ascertained by assaying the aqueous layer for its HA content after neutralization. The UV detector response changed quantitatively with the concentration of the labeled adducts showing excellent linearity (Supporting Information: Figure S1). By isolating the peak volumes of the analytes in question from different reaction mixtures, structures of the derivatives were unambiguously assigned to the N-naphthoate ester of different C-centered DMPO HA adducts (Supporting Information: Figure S2, parts a and b). Intriguingly, unreacted DMPO (nitrone form) undergoes a kinetically controlled hetero-Cope rearrangement yielding an O-naphthoyl imine derivative (i.e., 3naphthoyloxy-1-pyrroline) having 2 atomic mass units (m/z 268 amu) less than the N-naphthoate (m/z 270 amu) obtained after NaBH4 reduction (Supporting Information: Figure S2c). This product of DMPO has not been reported beforehand. Accordingly, DMPO can be determined fluorometrically either before or after two-electron reduction using the described derivatization procedure. Different factors that might interfere with the labeling reaction were investigated using 5-methylDMPO hydroxylamine (DMPOH/CH3) as a model HA adduct. Metal ions, such as Cu2+ and Fe3+ (ferricyanide) at 1:1 molar ratio to the adduct, did not affect naphthoyl esterification. Yet, at higher levels, they gradually suppressed the absolute peak area of the labeled compound up to 10 molar excess, where they totally abolished the response as evidenced by the disappearance of the derivative peak. Equimolar amounts of NaBH4 or ASC in presence of metal salts were able to partially restore the response. NaBH4 was more effective in this aspect than ASC because it allows a two-electron reduction of the ultimate product of the reaction, viz. the nitrone, whereas ASC reduces the nitroxyl intermediate through which the

whereas GSH and NAC produced secondary radicals, most probably thiyl intermediates (Figure 1, parts c, d, and f). NaBH4 caused two-electron reduction of nitrones beside the targeted one-electron reduction of nitroxyls (Figure 1b). Consequently, DMPO itself undergoes reduction to the corresponding HA and can react accordingly with 2-naphthoyl chloride resulting in a labeled spin trap. Therefore, if spin adducts are to be selectively labeled, NaBH4 should not be used for reduction. On the other hand, if DMPO is the analyte in question, aqueous NaBH4 is the reducing agent of choice which enables the formation of an N-naphthoate ester of this cyclic spin trap. This offers versatility to the technique, and hence it could be used as a highly sensitive tool to trace free DMPO levels in vivo. Addition of ASC followed by an equimolar amount of GSH was selected as a reducing combination for two reasons: (1) these substances are major bioreductants; (2) GSH acts as a safeguard against the back-oxidation of HAs by semidehydroascorbyl radical, and hence assists the ascorbatemediated nitroxide reduction (Figure 1e).29 Simple color reactions were devised as indicators for completeness of the Grignard addition and the reduction step. Ferric chloride oxidized 2-unsubstituted cyclic nitrones, i.e., excess unreacted DMPO, with the formation of a violet complex which absorbs at 540−545 nm.30 Triphenyltetrazolium chloride reacts with dialkyl HA bearing no substituents on the oxygen giving a characteristic purple-red color of the reduction product.31,32 Ascorbic acid, a strong reducing agent, interferes with the results of this test. These confirmatory tests were useful alternatives to spectroscopy during routine use of the technique. Labeling of Diamagnetic DMPO Adducts. In a cyclic spin trap as DMPO, there is only the nitrone functional group which could be modified post-trapping notwithstanding the type of radical added. For this purpose, we tested the feasibility of converting EPR-inactive cyclic HAs into their corresponding ester derivatives.33,34 Generally, nitrone spin traps or adducts do not show adequate chromatographic behaviors using conventional reversed-phase (RP) sorbents because of secondary interactions between the nitrogenous centers and residual silanols resulting in remarkable peak tailing. In addition, being a low molecular weight hydrophilic spin trap makes DMPO difficult to be retained satisfactorily using typical RP-chromatography. Moreover, unlike POBN and PBN, DMPO possesses neither a strong nor characteristic UV−vis absorption spectrum.35 In addition, absorption bands of nitrones are in general solvent and pH sensitive.36 Biological systems are characterized by a distinctive reducing environment, where several mediators, such as mitochondrial ubiquinone−cytochrome c reductase, NADH- and NADPHdependent reductases, thioredoxin, tissue thiols in concert with superoxide, GSH, and ASC interplay to secure an overall constant redox potential. In such an environment, an EPRactive nitroxide is very likely to undergo a facile bioreduction into the corresponding EPR-invisible hydroxylamine. Nonetheless, other aspects affect the reduction rate of cyclic nitrones, for instance the nitrone ring type, ring substituents, electron affinity of nitrones, hydrophilicity/lipophilicity balance, O2 level, subcellular compartment, etc.37 Conversely, HAs (e.g., reduced nitroxides of POBN and PBN) undergo slow autoxidation back to their nitroxyl radical form in buffer or upon solvent dilution.22 However, under the effect of mild oxidizing agents (O2, H2O2, or metal ions), HAs undergo oxidation to nitroxyl radicals which in turn are disproportio6741

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746

Analytical Chemistry

Article

Figure 2. Detection and quantification of hydroxyl radicals as DMSO-derived methyl radicals (tR = 14.0 min) at different scavenger concentrations using post-trapping labeling and HPLC/UV vs EPR spectroscopy (insets) according to the conditions stated within the text.

physicochemical properties, and it is photostable at ambient light;44 (2) it diffuses readily across cell membranes;45 (3) it has favorable tissue distribution;46 (4) it is the least toxic spin trap in vivo;47 (5) it shows adequate trapping rate constants with Ccentered free radicals in aqueous medium.48 Next, we envisaged that carbon-centered radicals would be the reactive intermediates of choice to be monitored for the following reasons: (1) they are frequent spin-trapping products in biological settings;49 (2) methyl radicals are able to act successfully as an indirect measure for •OH which is chemically trapped by an auxiliary reaction with dimethyl sulfoxide (DMSO) avoiding the production of artifacts due to the instability of the DMPO hydroxyl radical adduct (DMPO•/OH);50 (3) being weak nucleophiles, carbon-centered radicals are less vulnerable to the non-free-radical Forrester−Hepburn reaction and subsequent production of counterfeit spin adducts;51 (4) our laboratory is interested in phenyl radicals and their implications in anilinebased drug-induced agranulocytosis.52 Also, we are currently studying the mechanism of action of procarbazine, a second line antiglioma anticancer, which has been shown to generate methyl radicals.53 It is possible that these radicals might be associated with bone marrow side effects. Of note, our work is distinctly distinguished from previous work using nitroxide spin labels for the fluorescence detection of free radicals, where over a series of valuable publications, Kieber and Blough exploited the phenomenon of intramolecular quenching which limits the fluorescence of fluorescamine-linked 3-(aminomethyl)-2,2,5,5tetramethyl-1-pyrrolidinyloxy nitroxyl radical (3-AMP) for the determination of carbon-centered radicals in aqueous solutions.54 Though the work of the Blough group is highly appreciated, it has certain limitations. For instance, fluorescamine hydrolyzes quickly in aqueous solutions yielding nonfluorescent hydrolytic products and the formation of its adduct is pH-dependent with significantly decreasing yields below pH 7.5 or above pH 10.55 The stability of the hydroxylamine−fluorescamine adducts is very low unless kept under stringent anaerobic conditions.55 Another influential drawback is that the emitted light does not correlate linearly with the reduction of the spin label, a curious fact that might be attributed to a circuitous mechanism for nitroxide-mediated fluorescence quenching of fluorescamine-based fluorophores.56 Most importantly, little if any is known about the toxicity and pharmacokinetics of 3-AMP in vivo, and no studies have been conducted to explore its efficiency in biological spin labeling.

nitrone forms. No complete restoration of the response occurs probably because of the formation of other oxidation species as oxoammonium cations and other acyclic derivatives.40 In contrast, ferric chloride did not hinder the esterification reaction up to 10 molar excess but only decreased the detector response by almost 50% at 20 molar excess of the DMPO adduct (Supporting Information: Figure S3). However, ferric chloride fully abrogates the derivatization of DMPO, possibly via the formation of a hydroxamic acid derivative. Profitably, this can be maneuvered to avoid labeling of the spin trap in case a selective assay without a separation technique is required for radical-derived DMPO adducts. Under mild conditions and up to 2 molar excess, H2O2 did not negatively impact the derivatization of reduced DMPO adducts nor did it suppress the detector response. Some reports have shown that acyclic HAs were oxidized by mildly acidic solutions of peroxides to the corresponding nitrones.41−43 On the other hand, in a biochemical system such as procarbazine/horseradish peroxidase (HRP) or MP-11, H2O2 had a little effect on the final response. It is likely that these peroxidases caused a peroxidecatalyzed two-electron oxidation of the HAs. Evidently, this oxidative pathway is confirmed, at least in part, by an approximately 2-fold higher response when NaBH4, but not ASC, was used as a reducing agent (data not shown). However, the formation of higher or other oxidation products of HAs cannot be excluded. The reducing agents interacted slightly with 2-naphthoyl chloride in different ways; ASC is partly acylated by the tagging mixture, whereas NaBH4 had a weak reducing effect on the labeling reagent. At a concentration of 10 mM, ASC did not pose any problems for the labeling reaction. In the different applications, the final pH of the reaction mixture was virtually neutral (pH 6−8) and did not interfere with the naphthoylation reaction. The post-trapping labeled DMPO adduct has the following advantages: (1) it possesses strong UV and fluorescent characteristics (Supporting Information: Figure S4); (2) it is highly stable (t1/2 ≈ months); (3) it is more relevant to biological systems; (4) it imparts a superior chromatographic behavior to the reduced radical adducts, and hence their separation on conventional RPcolumns is substantially improved; (5) it has a characteristic m/ z in MS (Supporting Information: Figure S2, parts a and b); (6) it could be easily synthesized to serve as a calibration standard. We selected DMPO as a candidate spin trap because of the following reasons: (1) DMPO has unobjectionable 6742

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746

Analytical Chemistry

Article

Figure 3. Detection and quantification of methyl radicals (tR ≈ 14.0 min) produced by MP-11-catalyzed oxidation of procarbazine at two substrate concentrations (right inset) using post-trapping labeling and HPLC/UV vs EPR spectroscopy (left inset) according to the conditions stated within the text.

Figure 4. Detection and quantification of phenylhydrazine-derived phenyl radicals (tR ≈ 34.0 min) catalyzed by Cu(II) in carbonate buffer using post-trapping labeling and HPLC/UV vs EPR spectroscopy (inset) according to the conditions stated within the text.

hemoprotein of plant origin, and human plasma which were spiked with HA adduct standard of methyl-DMPO (Supporting Information: Figure S5). DMPO could be simultaneously determined using this approach after a reduction step with NaBH4 but not ASC. Determination of Hydroxyl Radicals in Terms of DMSOGenerated Methyl Radicals. A Fenton system generating •OH was incubated with varying concentrations of DMSO which effectively scavenges these radicals generating the •CH3 equivalents. Clearly, the outlined technique was superior to EPR in selectively detecting and accurately quantifying the level of •OH on the basis of HPLC and internal standards (Figure 2). At 2.8 mM, DMSO produced a fairly good signal for DMPO•/CH3 in EPR, whereas using the developed approach a striking response for DMPOH/CH3 naphthoate was obtained using HPLC/UV. On the other hand, when the concentration of DMSO was decreased to 0.35 mM, a level closer to that commonly and safely employed in cell culture applications, EPR produced a quite weak signal for •OH-derived •CH3 which was overwhelmed by the presence of a four-line spectrum due to nonscavenged •OH (Figure 2, inset). However, when the described methodology was applied in conjunction with HPLC/UV, an excellent response was readily

Another major disadvantage of using nitroxide radicals as spin labels in biological systems is the increased likelihood of instantaneous reduction and subsequent loss of their trapping capability. The prefluorescent proxyl probe has been explored in fluorescence imaging of free radicals; yet, authors were not able to achieve their goal due to response nonlinearity.56 In contrast, the use of a nontoxic nitrone spin trap, such as DMPO, is more suitable for in vivo settings. However, as discussed previously the produced radical-derived paramagnetic adducts are likely to undergo a facile extra- and intracellular reduction into the corresponding diamagnetic HAs. The current technique implements the effective spin-trapping step but circumvents the loss of EPR visibility due to the reduction of the trapped products by post-trapping derivatization. In addition, derivatized DMPO hydroxylamine adducts are resistant to back-oxidation by sample workup and dilution by various solvents or buffers. Applications. Advantages and drawbacks of the proposed technique and potential ramifications of this approach are discussed below throughout various applications encompassing chemical, biochemical, and biomatrixes, such whole blood. Initially, the applicability of the method was examined in a complex experimental matrix comprising HRP, a peroxidase 6743

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746

Analytical Chemistry

Article

Figure 5. Detection and quantification of phenylhydrazine-derived phenyl radicals (tR = 34.0 min) catalyzed by HRP/H2O2 using post-trapping labeling and HPLC/UV vs EPR spectroscopy (inset) according to the conditions stated within the text.

Figure 6. Detection and quantification of phenylhydrazine-derived phenyl radicals generated in situ from whole blood by post-trapping labeling and HPLC/UV vs EPR spectroscopy (inset) using 50:50 acetonitrile/water, v/v at 230 nm.

attainable for the labeled DMPO HA adduct. The level of •OH was estimated to be 450 μM in the studied Fenton reaction. It is noteworthy that using EPR the six-line signal of the detected radical fully decayed in approximately 10 min. Nonetheless, using the present approach signal intensity only changed by less than 2% after 72 h. Determination of Methyl Radicals Generated from Procarbazine/MP-11. Procarbazine, a hydrazine anticancer drug, is oxidized by MP-11, a low molecular weight (2 kDa) pseudoperoxidase proteolytically derived from equine cytochrome c, generating •CH3. In fact, this system is complicated and challenging because some radical species are gradually produced while others fade by time. This is especially true for HRP (Supporting Information: Figure S6). Moreover, the type and concentration of radicals vary with enzyme concentration (Supporting Information: Figures S7 and S8). Nonsurprisingly, the levels of reduced DMPO adducts were considerably low in this setup. Nevertheless, using post-trapping derivatization of the EPR-silent DMPO adducts, we were able to detect •CH3 at different levels of the substrate (Figure 3). In this system, omission of H2O2 or its replacement by glucose/glucose oxidase (5 mM/5 mU) resulted in a modest increase in the optical response. There are several reports showing that N,Ndialkyl HAs can be readily oxidized by peroxidases, such as

HRP.57,58 On the other hand, EPR in this particular model system was more successful in detecting methyl radicals (Figure 3, inset). However, this oxidative environment does not parallel the overwhelming reducing milieu in in vivo settings. Still, fluorescence detection is expected to substantially enhance the sensitivity of the method. Additionally, the current approach provides opportunities to gain insights into the mechanisms of oxidation of HAs by peroxidases. Determination of Phenyl Radicals Generated from Phenylhydrazine Chemically. Phenylhydrazine, an aromatic hydrazine xenobiotic, undergoes a one-electron metal-catalyzed oxidation in alkaline buffers to yield •C6H5 with a characteristic six-line EPR signal in 10% MeOH (Figure 4, inset). Alternatively, post-trapping derivatization of reduced phenylDMPO adducts and HPLC/UV provided a relatively clean chromatogram showing a single strong peak for 5-phenylDMPO hydroxylamine (DMPOH/C6H5) (Figure 4). A remarkable difference in sensitivity is revealed in favor of the proposed technique when compared with EPR. On basis of calibration and internal standards, the level of •C6H5 was computed to be 1.85 mM. Determination of Phenyl Radicals Generated from Phenylhydrazine/HRP. Phenylhydrazine incubated with HRP in presence of H2O2 produces •C6H5 through a peroxidase6744

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746

Analytical Chemistry driven one-electron oxidation with a typical EPR spectrum (Figure 5, inset). Unlike DMPOH/CH3, the DMPOH/C6H5 adducts appear to be more resistant to peroxidase oxidation, probably due to steric hindrance. Therefore, post-trapping naphthoylation of the phenyl-HA adduct followed by HPLC/ UV offered a valid and at least equally effective substitute to EPR for the detection of •C6H5 (Figure 5). However, the present method provides easier and more accurate quantification of those free radicals. More importantly, HPLC/UV is a popular technique available in almost every laboratory. Determination of Phenyl Radicals Generated in Situ from Phenylhydrazine in Whole Blood. Red blood cells (RBCs) destruction by phenylhydrazine has been documented for some time.59 Hemolysis has been postulated to be attributable to intracellular free radical formation, among which •C6H5 might be playing a prominent role. Previous reports detected •C6H5 originating from phenylhydrazine by the action of diluted RBCs;60 however, in whole rat blood or using undiluted RBCs, thiyl radicals were the predominant species which masked the characteristic EPR signal of phenyl radicals.61 Applying the current approach, we were able to unequivocally detect •C6H5 in whole blood at different substrate levels after spontaneous bioreduction (Figure 6). Therefore, selectivity could be considered as an added benefit to the proposed technique. Finally, the suggested method could be applied to monitor and measure the extent of spontaneous reduction of radical-derived DMPO spin adducts in various settings.

■

REFERENCES

(1) Halliwell, B.; Gutteridge, J. M. C. In Free Radicals in Biology and Medicine; Oxford University Press: Oxford and New York, 1999; p 936. (2) Dröge, W. Physiol. Rev. 2002, 82, 47−95. (3) Jackson, M. J. Free Radical Biol. Med. 2008, 44, 132−141. (4) Valko, M.; Leibfritz, D.; Moncol, J.; Cronin, M. T. D.; Mazur, M.; Telser, J. Int. J. Biochem. Cell Biol. 2007, 39, 44−84. (5) Mason, R. P. In Free-Radical Intermediates in the Metabolism of Toxic Chemicals; Pryor, W. A., Ed.; Free Radicals in Biology; Vol. 5; Academic Press: New York, 1982; p 161. (6) Mason, R. P. In Free-Radical Metabolite Formation by Mammalian Peroxidases; Fowler, B. A., Ed.; Mechanisms of Cell Injury: Implications for Human Health; John Wiley & Sons: Chichester, U.K., 1987; p 67. (7) Amacher, D. E. Curr. Drug Metab. 2006, 7, 219−229. (8) Janzen, E. G.; Blackburn, B. J. J. Am. Chem. Soc. 1968, 90, 5909− 5910. (9) Lagercrantz, C.; Forshult, S. Nature 1968, 218, 1247−1248. (10) Janzen, E. G. In Free Radicals in Biology; Pryor, W. A., Ed.; Academic Press: New York, 1980; Vol. 4, p 115. (11) Berliner, L. J.; Khramtsov, V.; Fujii, H.; Clanton, T. L. Free Radical Biol. Med. 2001, 30, 489−499. (12) Degray, J. A.; Mason, R. P. In Biological Spin Trapping; Atherton, N. M., Gilbert, B. C., Davies, M. J., Eds.; Electron Spin Resonance; Vol. 14; Royal Society of Chemistry: Cambridge, U.K., 1994; p 246. (13) Janzen, E. G. Acc. Chem. Res. 1971, 4, 31−40. (14) Janzen, E. G.; Liu, J. I.-P. J. Magn. Reson. (1969) 1973, 9, 510− 512. (15) Iwahashi, H.; Parker, C. E.; Mason, R. P.; Tomer, K. B. Anal. Chem. 1992, 64, 2244−2252. (16) Qian, S. Y.; Kadiiska, M. B.; Guo, Q.; Mason, R. P. Free Radical Biol. Med. 2005, 38, 125−135. (17) Ikeda, H.; Iwahashi, H. J. Sep. Sci. 2010, 33, 1185−1191. (18) Rao, P. S.; Weinstein, G. S.; Rujikarn, N.; Luber, J. M., Jr.; Tyras, D. H. Chromatographia 1990, 30, 19−23. (19) Towell, J.; Kalyanaraman, B. Anal. Biochem. 1991, 196, 111− 119. (20) Guo, Q.; Gao, G.; Qian, S. Y.; Mason, R. P. Chem. Res. Toxicol. 2004, 17, 1481−1490. (21) Stoyanovsky, D. A.; Cederbaum, A. I. Free Radical Biol. Med. 1998, 25, 536−545. (22) Stoyanovsky, D. A.; Melnikov, Z.; Cederbaum, A. I. Anal. Chem. 1999, 71, 715−721. (23) Novakov, C. P.; Feierman, D.; Cederbaum, A. I.; Stoyanovsky, D. A. Chem. Res. Toxicol. 2001, 14, 1239−1246. (24) Bonnett, R.; Brown, R. F. C.; Clark, V. M.; Sutherland, I. O.; Todd, A. J. Chem. Soc. 1959, 2094−2102. (25) Black, D. S. C.; Boscacci, A. B. Aust. J. Chem. 1976, 29, 2511− 2524. (26) Janzen, E. G.; Zhang, Y.; Haire, D. L. Magn. Reson. Chem. 1994, 32, 711−720. (27) Janzen, E. G.; Zhang, Y.; Arimura, M. Chem. Lett. 1993, 22, 497−500. (28) Nazarski, R. B.; Skowroński, R. J. Chem. Soc., Perkin Trans. 1 1989, 1603−1610.

CONCLUSIONS In contrast to EPR spectroscopy which detects transient DMPO radical adducts, the illustrated technique has been designed and developed to detect and quantify EPR-silent end products of spin-trapped radical adducts in the form of ultrastable, fluorescent derivatives. The proposed approach has for the first time unraveled the prospect to convert the disappearance of an EPR signal of DMPO-trapped free radicals by spontaneous reduction into a fluorescent emission. Whereas only some applications were presented in this work, potential ramifications might include probing of a wide range of free radicals in chemical and biological systems. In addition, the presented approach will likely contribute to the detection of macromolecule free radicals. Overall, our studies suggest new opportunities of post-trapping structural modification of radical-derived cyclic nitroxyl adducts in combination with a universal technique, such as HPLC/UV or fluorescence detection that would enhance the detection limits, stability, and chromatography of the free radical of interest. Clearly, the method has been demonstrated to be a viable alternative to EPR. In many instances, this novel technique proved to be superior to EPR in terms of robustness, selectivity, and sensitivity. ASSOCIATED CONTENT

S Supporting Information *

Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

■

ACKNOWLEDGMENTS

Preliminary findings from this manuscript were orally presented in the SFRBM’s 18th Annual Meeting, 2011, Atlanta, GA, U.S.A. This work was supported by a research Grant from the Canadian Institute of Health Research (CIHR reference no. 202034). The authors are thankful to Dr. Randy Whittal and Dr. Vishwa Somayaji for their assistance in MS experiments. Also, we acknowledge Dr. Richard Rowthery’s technical assistance with EPR operations and Dr. Joel Weiner for use of his EPR facility at the University of Alberta.

■

■

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: +1-780-492-1217. Notes

The authors declare no competing financial interest. 6745

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746

Analytical Chemistry

Article

(29) Bobko, A. A.; Kirilyuk, I. A.; Grigor’ev, I. A.; Zweier, J. L.; Khramtsov, V. V. Free Radical Biol. Med. 2007, 42, 404−412. (30) Elsworth, J. F.; Lamchen, M. J. Chem. Soc. C 1966, 1477−1479. (31) Rogers, M. A. T. J. Chem. Soc. 1955, 769−772. (32) Thesing, J.; Sirrenberg, W. Chem. Ber. 1959, 92, 1748−1755. (33) Huisgen, R.; Bayerlein, F. Justus Liebigs Ann. Chem. 1960, 630, 138−146. (34) Gassman, P. G.; Hartman, G. D. J. Am. Chem. Soc. 1973, 95, 349−354. (35) Kaminsky, L. S.; Lamchen, M. J. Chem. Soc. B 1968, 1085−1087. (36) Kubota, T.; Yamakawa, M.; Mori, Y. Bull. Chem. Soc. Jpn. 1963, 36, 1552−1563. (37) Swartz, H. M. Free Radical Res. Commun. 1990, 9, 399−405. (38) Grigor’ev, I. A. In Nitrones: Novel Strategies in Synthesis; Nitrile Oxides, Nitrones, and Nitronates in Organic Synthesis; Feuer, H., Ed.; Wiley-Interscience: Hoboken, NJ, 2008; p 143. (39) Connors, K. A.; Pandit, N. K. Anal. Chem. 1978, 50, 1542− 1545. (40) Bernofsky, C.; Ratnayake Bandara, B. M. Mol. Cell. Biochem. 1995, 148, 155−164. (41) Utzinger, G. E. Ann. Chem. 1944, 556, 50−64. (42) Thesing, J.; Muller, A.; Michel, G. Chem. Ber. 1955, 88, 1027− 1037. (43) Delpierre, G. R.; Lamchen, M. Q. Rev., Chem. Soc. 1965, 19, 329−348. (44) Ohkuma, T.; Kirino, Y.; Kwan, T. Chem. Pharm. Bull. 1981, 29, 25−28. (45) Anzai, K.; Aikawa, T.; Furukawa, Y.; Matsushima, Y.; Urano, S.; Ozawa, T. Arch. Biochem. Biophys. 2003, 415, 251−256. (46) Liu, K. J.; Kotake, Y.; Lee, M.; Miyake, M.; Sugden, K.; Yu, Z.; Swartz, H. M. Free Radical Biol. Med. 1999, 27, 82−89. (47) Janzen, E. G. J. Biochem. Biophys. Methods 1995, 30, 239−247. (48) Villamena, F. A.; Hadad, C. M.; Zweier, J. L. J. Phys. Chem. A 2005, 109, 1662−1674. (49) Mason, R. P.; Kadiiska, M. B. In In Vivo EPR(ESR): Theory and Applications; Berliner, L. J., Ed.; Biological Magnetic Resonance; Vol. 18; Plenum Press: New York, 2003; p 314. (50) Burkitt, M. J.; Mason, R. P. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 8440−8444. (51) Forrester, A. R.; Hepburn, S. P. J. Chem. Soc. C 1971, 701−703. (52) Siraki, A. G.; Bonini, M. G.; Jiang, J.; Ehrenshaft, M.; Mason, R. P. Chem. Res. Toxicol. 2007, 20, 1038−1045. (53) Goria-Gatti, L.; Iannone, A.; Tomasi, A.; Poli, G.; Albano, E. Carcinogenesis 1992, 13, 799−805. (54) Kieber, D. J.; Blough, N. V. Free Radical Res. Commun. 1990, 10, 109−117. (55) Kieber, D. J.; Blough, N. V. Anal. Chem. 1990, 62, 2275−2283. (56) Pou, S.; Bhan, A.; Bhadti, V. S.; Wu, S. Y.; Hosmane, R. S.; Rosen, G. M. FASEB J. 1995, 9, 1085−1090. (57) Boucher, J.-L.; Genet, A.; Vadon, S.; Delaforge, M.; Henry, Y.; Mansuy, D. Biochem. Biophys. Res. Commun. 1992, 187, 880−886. (58) Cai, T.; Xian, M.; Wang, P. G. Bioorg. Med. Chem. Lett. 2002, 12, 1507−1510. (59) Flores, G.; Frieden, E. Dev. Biol. 1972, 27, 406−418. (60) Hill, H. A. O.; Thornalley, P. J. FEBS Lett. 1981, 125, 235−238. (61) Maples, K. R.; Jordan, S. J.; Mason, R. P. Mol. Pharmacol. 1988, 33, 344−350.

6746

dx.doi.org/10.1021/ac301142c | Anal. Chem. 2012, 84, 6739−6746