Combination of Electrochemistry and Nuclear Magnetic Resonance

Sep 12, 2012 - Safety & Environmental Assurance Centre, Unilever U.K., Colworth Science Park, Sharnbrook, Bedfordshire MK44 1LQ, United. Kingdom...
1 downloads 0 Views 243KB Size
Article pubs.acs.org/ac

Combination of Electrochemistry and Nuclear Magnetic Resonance Spectroscopy for Metabolism Studies Hannah Simon,† Daniel Melles,† Sandrine Jacquoilleot,‡ Paul Sanderson,‡ Raniero Zazzeroni,‡ and Uwe Karst*,† †

Institut für Anorganische und Analytische Chemie, Westfälische Wilhelms-Universität Münster, Corrensstrasse 30, 48149 Münster, Germany ‡ Safety & Environmental Assurance Centre, Unilever U.K., Colworth Science Park, Sharnbrook, Bedfordshire MK44 1LQ, United Kingdom S Supporting Information *

ABSTRACT: During the development of new materials demonstrating biological activity, prediction and identification of reactive intermediates generated in the course of drug metabolism in the human liver is of great importance. We present a rapid and purely instrumental method for the structure elucidation of possible phase I metabolites. With electrochemical (EC) conversion adopting the oxidative function of liver-inherent enzymes and nuclear magnetic resonance (NMR) spectroscopy enabling structure elucidation, comprehensive knowledge on potential metabolites can be gained. Paracetamol (APAP) has been known to induce hepatotoxicity when exceeding therapeutic doses and was therefore selected as the test compound. The reactive metabolite N-acetyl-p-benzoquinone imine has long been proven to be responsible for the toxic side effects of APAP and can easily be generated by EC. EC coupled online to NMR is a straightforward technique for structure elucidation of reactive drug intermediates at an early stage in drug discovery.

A

detection of molecular sites labile toward oxidation offers the opportunity to investigate the formation of phase I metabolites at an early stage in drug discovery. Various oxidation products generated by EC via a single electron transfer show a good correlation to metabolites found in vivo. However, the method is limited to reactions initiated by a single electron transfer. Thus, CYP-catalyzed reactions that proceed via hydrogen abstraction cannot be mimicked.6 EC/ESI-MS offers the possibility to characterize reactive metabolites due to the absence of a biological matrix and hence potential reaction partners. By extending the instrumental setup by high-pressure liquid chromatography (HPLC), phase II adduct formation of reactive intermediates can be achieved.7 Lohmann et al. were able to simulate the detoxifying pathway of APAP by trapping the electrochemically generated NAPQI with GSH and the antidote N-acetylcysteine. Furthermore, a chromatographic separation and a structural confirmation of the formed adducts by means of HPLC and tandem MS were achieved. The oxidation pathway of APAP as well as the reaction of NAPQI with GSH proposed by Lohmann et al. is presented in Scheme 1.8 EC/LC/MS was also used to reveal the covalent binding of NAPQI to the proteins β-lactoglobulin and human serum albumin.9

crucial step in the development of new materials demonstrating biological activity is the prediction of their metabolic fate in the human body. Metabolic transformation of xenobiotics such as the analgesic and antipyretic paracetamol (APAP) into reactive intermediates has been shown to induce toxic side effects. An APAP overdose has been known to induce acute liver necrosis since the 1960s.1 The metabolite responsible for the hepatotoxicity could be identified as N-acetyl-p-benzoquinone imine (NAPQI), which is a highly reactive electrophile.2 Detoxification of NAPQI generally proceeds via conjugation with reduced glutathione (GSH).3 At higher doses, the hepatic GSH will finally be depleted. Due to its highly electrophilic nature, NAPQI will then bind covalently to other thiol groups of cellular macromolecules such as liver proteins, thus inducing toxic side effects.4 The biotransformation reactions leading to the formation of the reactive metabolite NAPQI are initiated by oxidation reactions catalyzed by enzymes of the cytochrome P450 superfamily (CYP).5 Direct identification of short-lived intermediates such as NAPQI in the course of drug development is therefore of great importance to prevent toxic side effects. However, these highly reactive species tend to rearrange themselves or react very quickly with available sites on proteins during in vitro studies and are subsequently removed during cleanup and protein precipitation procedures before analysis. The potential of electrochemistry (EC) coupled online to ESI-MS as a complementary tool in metabolism studies has been investigated extensively in the past decade. Rapid © XXXX American Chemical Society

Received: July 27, 2012 Accepted: September 12, 2012

A

dx.doi.org/10.1021/ac302152a | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

decrease in resolution due to magnetic field inhomogeneity during electrolysis.18 NMR spectroscopy has been shown to be a valuable tool for structure elucidation of potential in vivo metabolites in electrochemical metabolism studies.19,20 A major drawback is the need for sufficient quantities of the desired metabolite. This often results in a very long electrolysis time due to low conversion rates and small electrode surfaces.21,22 Time is an important factor, especially for the characterization of reactive metabolites, which is only possible after adduct formation with soft nucleophiles.23−25 In our studies, we describe the direct coupling of a commercially available EC cell with an NMR spectrometer equipped with a flow probe for the online monitoring of electrochemical reactions. Homogeneity of the magnetic field is maintained due to the utilization of an external reactor cell. Our approach allows a very rapid detection of electrochemically generated oxidation products by means of 1H NMR and is therefore suitable for high-throughput screening of pharmaceuticals. Structure elucidation even of oxidation products that are not amenable to MS is possible. Our aim was to prove the applicability of our setup by monitoring the electrochemical formation of NAPQI and BQ without the need for trapping agents.

Scheme 1. Structures and Reactions for the Electrochemical Oxidation of APAP and the Adduct Formation of APAP with GSH



EXPERIMENTAL SECTION Chemicals. Paracetamol, p-benzoquinone, glutathione, ammonium acetate, sodium sulfate (Na2SO4), and acetonitrile-d3 (CD3CN) were obtained from Sigma-Aldrich Chemie GmbH (Steinheim, Germany). Acetonitrile (AcN) was purchased from Merck (Darmstadt, Germany). Deuterated water (D2O) was ordered from Euriso-Top (Gif sur Yvette Cedex, France). All chemicals were used in the highest quality available. Water was purified before utilization using an Aquatron A4000D still system (Barloworld Scientific, Nemours Cedex, France). EC/MS. The oxidation behavior of APAP was studied by using an online setup consisting of an electrochemical cell (Flex Cell, Antec Leyden, Zoeterwoude, The Netherlands) coupled directly to the ESI source of a high-resolution mass spectrometer (Exactive Orbitrap, Thermo Fisher Scientific, Bremen, Germany). A 10 μM solution of APAP in 1 mM ammonium acetate (pH 7.4) and AcN (50/50, v/v) was passed through the cell and into the ESI interface of the mass spectrometer at a flow rate of 10 μL/min by a Fusion 100 classic syringe pump (Chemyx, Stafford, TX). The electrochemical cell was equipped with a glassy carbon (GC) working electrode. The experimental setup is presented in Figure 1. A homemade potentiostat was used to apply a potential ramp between 0 and 2000 mV at a scan rate of 10 mV/s against Pd/ H2 to generate three-dimensional mass voltammograms, which provide a precise overview of all oxidation products of APAP. The mass spectrometer was tuned and externally calibrated according to the instructions of the manufacturer. Operation was carried out in the negative ion mode with the following parameters: ESI(−); resolution, high, AGC target, high dynamic range, sheath gas flow (N2), 15 au; spray voltage, 3.0 kV; capillary temperature, 275 °C; capillary voltage, −42.5 V; tube lens voltage, −100 V; skimmer voltage, −34.0 V; maximum injection time, 100 ms; mass-to-charge ratio (m/z) range, 50−1000. Each mass voltammogram was recorded at least three times to ensure reproducibility of the measurements.

Coupling of EC to high-resolution MSn is able to provide important structural information about the generated oxidation products. However, to achieve a comprehensive structural identification, complementary spectroscopic methods such as nuclear magnetic resonance (NMR) spectroscopy are needed. Its ability to distinguish between structurally similar species, e.g., isomers, makes it particularly useful for monitoring of electrochemical reactions. EC in combination with NMR spectroscopy can be realized by two different approaches. An external flow cell outside the NMR probe can be used for monitoring comparably stable oxidation products. In situ electrolysis within the NMR probe offers the data acquisition of very unstable oxidation products at the electrode surface, as first reported by Richards and Evans. Nonetheless, it should be mentioned that the species-generating electrode was not placed in the actual detection region.10 Placement of an electrolysis cell within the NMR probe gives rise to several obstacles. The presence of asymmetric surfaces could cause a degradation of the homogeneity of the magnetic field. Metallic electrodes could furthermore alter the radio frequency field due to the presence of an electric current. This generally results in a loss of sensitivity and resolution.11 Various approaches to facilitate the application of in situ spectroelectrochemistry have been reported since then.12,13 Dedicated symmetrical EC units with working electrodes made of NMR tubes internally coated with conducting materials were designed to prevent a loss of resolution during in situ spectroelectrochemistry.11,14−16 A new spectroelectrochemical cell design consisting of almost metal free, symmetrically arranged carbon fiber electrodes was used to study the electrochemical behavior of p-benzoquinone (BQ)17 A comparison of the two instrumental setups emphasized the possibility to maintain the resolution for the combination of electrochemical and NMR flow cells. Insertion of the electrode assembly into a spinning NMR tube, however, showed a B

dx.doi.org/10.1021/ac302152a | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

the metabolites formed in vivo. Electrochemical formation of the reactive metabolite NAPQI is thus expected under the conditions applied. NAPQI can further be hydrolyzed, leading to the formation of BQ after reaction with one molecule of water.26 A mass voltammogram showing the EC oxidation of APAP is provided in the Supporting Information (Figure S-1). A signal at m/z 150 could successfully be assigned to the deprotonated parent ion with a mass deviation of 1.7 ppm. In agreement with the previously reported electrochemical studies of APAP, a decrease in the signal intensity was observed starting at 500 mV, which indicates that an oxidation of APAP took place. Nevertheless, no oxidation products were monitored with the parameters used. This is most likely due to the lack of acidic protons of the oxidation products. Thus, ionization in negative ion mode is hindered. Mass voltammograms of APAP were also recorded in the positive ion mode. Likewise, no oxidation products could be monitored. In further experiments, the soft nucleophile GSH was added to the effluent of the reactor cell to trap the highly electrophilic NAPQI and therefore visualize its electrochemical generation. Starting at an oxidation potential of 500 mV, the formation of a signal with m/z 455 was observed (Figure S-2, Supporting Information). With a deviation of 0.9 ppm between calculated and determined accurate masses, this signal could be assigned to C18H23N4O8S resulting from the adduct formation between NAPQI and GSH. This reaction is known to be the common detoxification pathway of NAPQI in the human liver.3 Further oxidation of NAPQI to the known oxidation product BQ was not observed. The reason for this is most likely the lack of acidic protons needed for successful ionization of BQ. EC/MS has proven its feasibility in metabolism simulation during the past decade, but is often limited by factors such as varying ionization efficiencies and exact structure elucidation of oxidation products. The addition of trapping agents such as GSH is one way to compensate for the limitations of the ionization technique and may allow the detection of oxidation products that are usually not accessible to ESI. However, the determination of the oxidation site in a molecule is of great importance for the elucidation and the prediction of the toxicity of potential reactive metabolites. To overcome these obstacles, online EC/NMR was integrated in the investigation of APAP as a novel complementary technique for in situ structure elucidation of oxidation products and to demonstrate its application potential in this field. EC/NMR. APAP provides a model reaction system that is readily amenable to characterization by 1H NMR spectroscopy, as APAP only forms two products (NAPQI and BQ) upon electrochemical oxidation.26 Therefore, APAP was chosen to demonstrate that oxidation products can be generated in the electrochemical cell and then detected and identified by onflow NMR spectroscopy analysis. Due to its symmetrically substituted aromatic moiety, two ortho coupled doublets are expected for APAP in the one-dimensional 1H NMR experiment. For the performance of online EC/NMR experiments of APAP, an electrochemical setup identical to the coupling with the MS instrument was employed. The effluent from the cell was directly introduced into the flow-through probe head of a 600 MHz NMR spectrometer by means of a transfer capillary. To obtain 1H NMR spectra in a reasonable time, the initial concentration of APAP was increased to 2 mM due to the concentration dependency of the NMR signal. A further change was undertaken in the buffer system, which was switched to Na2SO4 to avoid intense 1H NMR signals that

Figure 1. Experimental setup consisting of an electrochemical thinlayer cell coupled online to HR-ESI-MS (a) and NMR (b) instruments, respectively.

Mass voltammograms of trapping experiments with GSH were generated similarly to the procedures described above. The reaction solution was mixed with a 50 μM solution of GSH in doubly distilled water after electrochemical oxidation via a Tpiece. GSH was added with a constant flow of 10 μL/min. The reaction time after mixing was set to approximately 6 min. Operation was carried out in the negative ion mode with the following parameters: ESI(−); resolution, high; AGC target, balanced; sheath gas flow (N2), 15 au; spray voltage, 3.0 kV; capillary temperature, 250 °C; capillary voltage, −30.0 V; tube lens voltage, −175 V; skimmer voltage, −28.0 V; maximum injection time, 50 ms; mass-to-charge ratio (m/z) range, 50− 1000. Each mass voltammogram was recorded at least three times to ensure reproducibility of the measurements. EC/NMR. The same electrochemical setup as applied for the MS studies was used for oxidation. NMR experiments were performed at 300 K on a Bruker Avance (II) 600 NMR spectrometer operating at 600.13 MHz using standard Bruker pulse sequences. For direct online monitoring of reaction products, a 2 mM solution of APAP in CD3CN and 1 mM Na2SO4 dissolved in D2O (50/50, v/v) were transferred directly from the electrochemical thin-layer cell into the flow cell (60 μL volume) of an LC/NMR probe head at a rate of 5 μL/min (Figure 1). Transfer of the EC effluent to the NMR spectrometer took approximately 8 min. 1H NMR spectra were recorded from these steady-state reaction product mixtures in “on-flow” mode. Chemical shifts are referenced to the residual deuteroacetonitrile-d3 signal at 1.94 ppm. NMR. A spiking experiment was performed to confirm the electrochemical formation of BQ. The experiment was carried out at 300 K on a Bruker Avance (II) 600 NMR spectrometer operating at 600.13 MHz using standard Bruker pulse sequences. For sample preparation, 0.7 mL of a 2 mM solution of APAP in CD3CN/D2O (50/50, v/v) was spiked with 0.01 mL of a 10 mM solution of BQ in CD3CN/D2O (50/50, v/v). The final concentration of APAP was 1.97 mM, and that of BQ was 0.14 mM.



RESULTS AND DISCUSSION EC/MS. The electrochemical reaction pathway of APAP and also the adduct formation with GSH have previously been described by Getek et al. and our group. Therefore, the compound should be a well-suited model system for first online EC/NMR experiments.8,26 The electrochemically generated oxidation products are known to show a good correlation with C

dx.doi.org/10.1021/ac302152a | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 2. Aromatic proton region of the 1H NMR spectra of APAP at (a) 0 mV, (b) 600 mV, (c) 1200 mV, and (d) 1700 mV.

Figure 3. 1H NMR spectrum of (a) APAP spiked with BQ and (b) electrochemically oxidized APAP to assign the signal formed at 1700 mV.

would arise from the use of ammonium acetate. The aromatic proton region of the results is displayed in Figure 2 and demonstrates that significant levels of oxidation products are generated in the electrochemical cell under the reaction conditions used. Figure 2a shows the 1H NMR spectrum of

APAP without oxidation. As expected, two doublets can be observed in the aromatic region, due to the symmetrical substitution of the ring system (6.74 and 7.23 ppm). The signals for the remaining protons located at the methyl group, the nitrogen atom, and the hydroxyl function are not included D

dx.doi.org/10.1021/ac302152a | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

concentrations due to its inherently low sensitivity. Especially for substances that show a low electrochemical conversion rate, this could result in more lengthy infusion time to achieve a reasonable signal-to-noise ratio. This could lead to the decomposition of highly reactive oxidation products and hinder their detection by NMR. EC coupled online to MS will thus remain an important tool in the electrochemical mimicry of oxidative drug metabolism. Its high sensitivity makes it irreplaceable for the detection of oxidation products generated in low yields and complements the data obtained by EC/NMR. Fast screening by EC/MS and presentation of the obtained results in three-dimensional mass voltammograms deliver a precise overview on the generated oxidation products. Therefore, EC/NMR and EC/MS can ideally be used as complementary techniques in the simulation of oxidative metabolism.

in Figure 2a, since they are not needed for monitoring the electrochemical conversion of APAP. The complete spectrum of the electrochemical oxidation of APAP is supplied in the Supporting Information (Figure S-3). An oxidation potential of 600 mV was selected to obtain a 1H NMR spectrum of the reactive metabolite NAPQI in the onflow mode (Figure 2b). Under these conditions, very little oxidation has occurred as 1H NMR signals can be observed for the parent compound with only very small additional signals for a second species (6.64 and 7.02 ppm). The upfield chemical shift changes for the aromatic protons of the second species and the downfield chemical shift change for the methyl signal (not shown) are consistent with the electrochemical formation of NAPQI. An increase of the oxidation potential to 1200 mV led to a higher yield of oxidation product (Figure 2c). Integration of the signals for NAPQI showed that approximately 5% of the parent compound was converted at 600 mV, while doubling the potential leads to approximately 15% conversion of APAP. At 1700 mV (Figure 2d), the level of this product remains at approximately 15% conversion of parent compound. However, at this higher potential, the formation of an additional signal, which was putatively identified as BQ, was observed at 6.78 ppm. NMR signal integration indicates that approximately 4% of the parent has been oxidized to BQ. To confirm the formation of BQ at 1700 mV, a spiking experiment was performed in a standard NMR tube in the same solvent system as used in the flow cell. A 10 mM solution of BQ was added to a 2 mM solution of APAP, resulting in final concentrations of 1.97 mM for APAP and 0.14 mM for BQ. Figure 3 shows a comparison of APAP spiked with BQ (a) and APAP electrochemically oxidized at 1700 mV (b). The signal resulting from the spiking with BQ (Figure 3a) showed the same chemical shift as the singlet observed in Figure 3b and thus confirms the oxidation of NAPQI to BQ. The oxidation pathway of APAP, which was summarized in Scheme 1, was therefore confirmed by the results gained from online EC/MS and EC/NMR experiments. Upon electrochemical oxidation of APAP (151.1 Da), one electron and one proton are abstracted in the first step, resulting in the formation of an APAP radical. Further loss of one electron and one proton leads to the generation of NAPQI (149.1 Da). Finally, after hydrolysis with water from the buffer solution and the loss of the acetyl moiety, the p-benzoquinone (108.1 Da) is formed. However, if GSH is added to the oxidized solution as a trapping agent, adduct formation with NAPQI can be observed (456.1 Da).



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.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +49 251 83-33141, Fax: +49 251 83-36013. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Davidson, D. G.; Eastham, W. N. Br. Med. J. 1966, 2, 497−499. (2) Dahlin, D. C.; Miwa, G. T.; Lu, A. Y. H.; Nelson, S. D. Proc. Natl. Acad. Sci. U.S.A. 1984, 81, 1327−1331. (3) Mitchell, J. R.; Jollow, D. J.; Potter, W. Z.; Gillette, J. R.; Brodie, B. B. J. Pharmacol. Exp. Ther. 1973, 187, 211−217. (4) Jollow, D. J.; Mitchell, J. R.; Potter, W. Z.; Davis, D. C.; Gillette, J. R.; Brodie, B. B. J. Pharmacol. Exp. Ther. 1973, 187, 195−202. (5) Thummel, K. E.; Lee, C. A.; Kunze, K. L.; Nelson, S. D.; Slattery, J. T. Biochem. Pharmacol. 1993, 45, 1563−1569. (6) Jurva, U.; Wikström, H. V.; Weidolf, L.; Bruins, A. P. Rapid Commun. Mass Spectrom. 2003, 17, 800−810. (7) Faber, H.; Melles, D.; Brauckmann, C.; Wehe, C.; Wentker, K.; Karst, U. Anal. Bioanal. Chem. 2012, 403, 345−354. (8) Lohmann, W.; Karst, U. Anal. Bioanal. Chem. 2006, 386, 1701− 1708. (9) Lohmann, W.; Hayen, H.; Karst, U. Anal. Chem. 2008, 80, 9714− 9719. (10) Richards, J. A.; Evans, D. H. Anal. Chem. 1975, 47, 964−966. (11) Mincey, D. W.; Popovich, M. J.; Faustino, P. J.; Hurst, M. M.; Caruso, J. A. Anal. Chem. 1990, 62, 1197−1200. (12) Sandifer, M. E.; Zhao, M.; Kim, S. H.; Scherson, D. A. Anal. Chem. 1993, 65, 2093−2095. (13) Mairanovsky, V. G.; Yusefovich, L. Y.; Filippova, T. M. J. Magn. Reson. 1983, 54, 19−35. (14) Prenzler, P. D.; Bramley, R.; Downing, S. R.; Heath, G. A. Electrochem. Commun. 2000, 2, 516−521. (15) Webster, R. D. Anal. Chem. 2004, 76, 1603−1610. (16) Zhang, X.; Zwanziger, J. W. J. Magn. Reson. 2011, 208, 136−147. (17) Klod, S.; Ziegs, F.; Dunsch, L. Anal. Chem. 2009, 81, 10262− 10267. (18) Albert, K.; Dreher, E. L.; Straub, H.; Rieker, A. Magn. Reson. Chem. 1987, 25, 919−922. (19) Madsen, K. G.; Skonberg, C.; Jurva, U.; Cornett, C.; Hansen, S. H.; Johansen, T. N.; Olsen, J. Chem. Res. Toxicol. 2008, 21, 1107−19. (20) Thevis, M.; Lohmann, W.; Schrader, Y.; Kohler, M.; Bornatsch, W.; Karst, U.; Schänzer, W. Eur. J. Mass Spectrom. 2008, 14, 163−170.



CONCLUSIONS Online EC/NMR is presented here as a simple and elegant way for the generation, detection, and structure elucidation of potential reactive intermediates that might be formed in vivo. NMR analysis provides not only detailed structure elucidation but also quantification of generated oxidation products. It offers precise knowledge about the oxidation sites, which could help to predict toxic side reactions of potential active molecules. Extension of the instrumental setup with HPLC could make the presented method available to more complex systems, where various products are generated upon oxidation. The use of a flow probe allows the flow to be stopped and single peaks to be “parked” in the probe head of the NMR. This would ensure enough analysis time to obtain spectra of very small amounts of oxidation products with a sufficient signal-to-noise ratio. However, it has to be noted that NMR is limited to higher E

dx.doi.org/10.1021/ac302152a | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(21) Johansson, T.; Jurva, U.; Grö nberg, G.; Weidolf, L.; Masimirembwa, C. Drug Metab. Dispos. 2009, 37, 571−9. (22) Tahara, K.; Yano, Y.; Kanagawa, K.; Abe, Y.; Yamada, J.; Iijima, S.; Mochizuki, M.; Nishikawa, T. Chem. Pharm. Bull. 2007, 55, 1207− 1212. (23) Jurva, U.; Holmen, A.; Grönberg, G.; Masimirembwa, C.; Weidolf, L. Chem. Res. Toxicol. 2008, 21, 928−35. (24) Madsen, K. G.; Olsen, J.; Skonberg, C.; Hansen, S. H.; Jurva, U. Chem. Res. Toxicol. 2007, 20, 821−831. (25) Madsen, K. G.; Grönberg, G.; Skonberg, C.; Jurva, U.; Hansen, S. H.; Olsen, J. Chem. Res. Toxicol. 2008, 21, 2035−41. (26) Getek, T. A.; Korfmacher, W. A.; McRae, T. A.; Hinson, J. A. J. Chromatogr. 1989, 474, 245−256.

F

dx.doi.org/10.1021/ac302152a | Anal. Chem. XXXX, XXX, XXX−XXX