Unbiased High-Throughput Screening of Reactive Metabolites on the

Anal. Chem. 2008, 80, 6410–6422

Unbiased High-Throughput Screening of Reactive Metabolites on the Linear Ion Trap Mass Spectrometer Using Polarity Switch and Mass Tag Triggered Data-Dependent Acquisition Zhengyin Yan,* Gary W. Caldwell, and Noureddine Maher Division of Drug Discovery, Johnson & Johnson Pharmaceutical Research & Development, LLC, Spring House, Pennsylvania 19477 Constant neutral loss (CNL) and precursor ion (PI) scan have been widely used for the in vitro screening of glutathione conjugates derived from reactive metabolites, but these two methods are only applicable to triple quadrupole or hybrid triple quadrupole mass spectrometers. Additionally, the success of CNL and PI scanning largely depends on structure and CID fragmentation pathways of GSH conjugates. In the present study, a highly efficient methodology has been developed as an alternative approach for high-throughput screening and structural characterization of reactive metabolites using the linear ion trap mass spectrometer. In microsomal incubations, a mixture of glutathione [GSH, γ-glutamyl-cystein-glycin] and the stable-isotope labeled compound [GSX, γ-glutamylcystein-glycin-13C2-15N] was used to trap reactive metabolites, resulting in formation of both labeled and unlabeled conjugates at a given isotopic ratio. A mass difference of 3.0 Da between the natural and labeled GSH conjugate (mass tag) at a fixed isotopic ratio constitutes a unique mass pattern that can selectively trigger the datadependent MS2 scan of both isotopic partner ions, respectively. In order to eliminate the response bias of GSH adducts in the positive and negative mode, a polarity switch is executed between the mass tag-triggered data dependent MS2 scan, and thus ESI- and ESI+ MS2 spectra of both labeled and nonlabeled GSH conjugates are obtained in a single LC-MS run. Unambiguous identification of glutathione adducts was readily achieved with great confidence by MS2 spectra of both labeled and unlabeled conjugates. Reliability of this method was vigorously validated using several model compounds that are known to form reactive metabolites. This approach is not based on the appearance of a particular product ion such as MH+ - 129 and anion at m/z 272, whose formation can be structure-dependent and sensitive to the collision energy level; therefore, the present method can be suitable for unbiased screening of any reactive metabolites, regardless of their CID fragmentation pathways. Additionally, this methodology can potentially be applied * To whom correspondence should be addressed. Zhengyin Yan, Ph.D., Drug Discovery, R2013 Johnson & Johnson Pharmaceutical Research & Development, LLC Spring House, PA 19477-0779. Phone: (215)-628-5036. Fax: (215)-540-4878. E-mail: [email protected].

6410

Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

to triple quadrupole or hybrid triple quadrupole mass spectrometers. Cytochrome P450 enzymes (CYPs) play an essential role in metabolism and detoxicification of xenobiotics such as drugs and chemical toxicants, by which foreign compounds are converted to stable and more polar metabolites that can be readily eliminated from the human body. However, it is well recognized that, for some drugs and other xenobiotics, CYPs-mediated metabolism can also lead to the formation of reactive intermediates that can modify cellular proteins and nucleic acids via covalent binding. Such covalent modifications of critical cellular proteins by reactive metabolites can potentially lead to idiosyncratic drug toxicity, which is often characterized by a high degree of individual susceptibility and lacking simple dose dependence.1 Idiosyncratic toxicity has been a serious safety concern in medicine, since such drug reactions are sometimes very severe and even lethal, and can result in restricted drug use and even withdrawal from the market. Although biochemical mechanisms of drug-induced idiosyncratic toxicity have not been clearly defined, there is a substantial amount of evidence that chemically reactive metabolites are a critical cellular mediator, especially for liver toxicity.2,3 Presently, there has been a consensus in the field of drug discovery and development that, because of the potential safety concern, formation of reactive metabolites is not a desirable characteristic for a drug candidate. Therefore, in order to reduce attrition rate in drug development, screening and structural characterization of reactive metabolites has been integrated in the ADMET-guided lead optimization process in drug discovery.4,5 As a result, there has been a great demand in drug discovery for an efficient and unbiased method that can be used to screen a large volume of compounds for formation of reactive metabolites, since such information can be very helpful for medicinal chemists to optimize structural scaffolds at an early stage. (1) Lazarou, J.; Pomeranz, B. H.; Corey, P. N. J. Am. Med. Assoc. 1998, 279, 1200–1205. (2) Park, B. K.; Kitteringham, N. R.; Maggs, J. L.; Pirmohamed, M.; Williams, D. P. Annu. Rev. Pharmacol. Toxicol. 2005, 45, 177–202. (3) Walgren, J. L.; Mitchell, M. D.; Thompson, D. C. Crit. Rev. Toxicol. 2005, 35, 325–361. (4) Evans, D. C.; Watt, A. P.; Nicoll-Griffith, D. A.; Baillie, T. A. Chem. Res. Toxicol. 2004, 17, 3–16. (5) Caldwell, G. W.; Yan, Z. Curr. Opin. Drug Discovery Dev. 2006, 9, 47–60. 10.1021/ac800887h CCC: $40.75  2008 American Chemical Society Published on Web 07/19/2008

Because of instability of reactive metabolites, the current analytical methodology is to use glutathione (GSH) trapping, which results in formation of stable GSH adducts that can be subsequently characterized by tandem mass spectrometry. To meet the throughput of reactive metabolite screening in drug discovery, several MS methods have been developed during the past several years. Traditionally, constant neutral loss scanning (CNL) in the ESI positive mode has been employed due to its high efficiency and sensitivity.6 This method is based on the finding that molecular ions of GSH conjugates undergo a neutral loss of the pyroglutamic acid moiety (129 Da) under collisioninduced dissociation (CID). A major drawback of the conventional CNL is false positives resulting from nonspecific responses of endogenous components. This problem has been overcome using stable isotope trapping,7 in which an equal mixture of labeled [GSX, γ-glutamyl-cystein-glycin-13C2-15N] and unlabeled glutathione is used to trap reactive metabolites in microsomal incubations; as a result, GSH conjugates display unique doublet ions with a mass difference of 3.0 Da in the CNL scan mode, while the same MS feature is not seen for interfering components (false positives). Another concern is that the neutral loss of 129 Da is both structure and collision energy dependent, and some GSH adducts such as aliphatic and benzylic thioether conjugates may not afford the neutral loss of 129 Da.8,9 In order to address this concern, the precursor ion (PI) scan in the negative electrospray ionization mode was developed as an alternative for screening GSH conjugates.10 This method is based on the observation that, in the negative mode, glutathione conjugates form a predominated anion at m/z 272, which corresponds to deprotonated γ-glutamyldehydroalanyl-glycine moiety. However, few product ions are derived from CID fragmentation of GSH conjugates in the negative mode, which can significantly hinder structure elucidation. Therefore, a separate product ion MS full scan is required in the positive mode in order to obtain sufficient structural information. To overcome this drawback, a polarity switching approach has recently been developed,11 in which a negative precursor ion scan (m/z 272) is used as the survey scan to trigger the product ion MS full scan in the positive mode to obtain a more informative MS2 spectrum. This strategy has been demonstrated on the hybrid triple quadrupole linear ion trap mass spectrometer as a sensitive and effective method for screening GSH conjugates. However, it should be noted that responses of GSH conjugates can be polarity dependent. Therefore, if a particular conjugate is not sensitive in the negative mode, a false negative can still be reported. On the other hand, similar to the CNL scan, the precursor ion scan may still be biased, because a certain class of GSH conjugates may not form these characteristic product ions such as m/z 272. Additionally, application of both CNL and PI scan in screening

reactive metabolites is limited to triple quadrupole or hybrid triple quadrupole mass spectrometers. There is a great need to develop a MS screening method that is unbiased and has broad utility on different mass spectrometers. Presently, linear ion trap mass spectrometers have become a very powerful MS platform widely used for the identification and structural characterization of small molecules, peptides, and proteins, largely due to their superior sensitivity and versatile multiple-staged CID capabilities. So far, ion trap mass spectrometers have not been used for high-throughput screening of reactive metabolites because these types of instruments, unlike triple quadrupole mass spectrometers, perform ion isolation, CIDinduced fragmentation, ion storage, and mass analysis in the same ion device in a sequential manner, and these generic methods such as CNL and PI scan can not be physically executed. The present study describes a novel approach that can be routinely used as a generic method on ion trap mass spectrometers for high-throughput screening of reactive metabolites. In the current method, reactive metabolites were captured by stable isotope trapping, resulting in formation of both natural and stableisotope labeled GSH adducts that appear as a discernible isotope doublet with a mass difference of 3.0 Da in the survey MS full scan. This isotopic mass pattern is highly unique; therefore, it can be used in both positive and negative mode to trigger a datadependent MS2 scan of both isotopic conjugates. As a result, identification and structural characterization can be readily accomplished by comparing tandem MS spectra of labeled and nonlabeled conjugates. In contrast to CNL and PI scan, this approach is not associated with the CID fragmentation pathway and thus is unbiased in the screening of reactive metabolites.

(6) Chen, W. G.; Zhang, C.; Avery, M. J.; Fouda, H. G. Biological Reactive Intermediates VI. Chemical and Biological Mechanisms in Susceptibility to and Preventation of Environmental Diseases; Dansetter, P. M., Snyder, R., Delaforge, M., Gibson, G. G., Greim, H., Jollow, D. J., Monks, T. J., Sipes, I. G., Eds.; Kluwer Academic/Plenum Press: New York, 2001; 521-524. (7) Yan, Z.; Caldwell, G. W. Anal. Chem. 2004, 76, 6835–6847. (8) Baillie, T. A.; Davis, M. R. Biol. Mass Spectrom. 1993, 22, 319–325. (9) Grillo, M. P.; Hua, F.; Knutson, C. G.; Ware, J. A.; Li, C. Chem. Res. Toxicol. 2003, 16, 1410–1417. (10) Dieckhaus, C. M.; Fernandez-Metzler, C. L.; King, R.; Krolikowski, P. H.; Baillie, T. A. Chem. Res. Toxicol. 2005, 18, 630–638. (11) Wen, B.; Ma, L.; Nelson, S. D.; Zhu, M. Anal. Chem. 2008, 80, 1788– 1799.

EXPERIMENTAL SECTION Materials. Reagents and solvents used in the current study were of the highest possible grade available. The following chemicals were purchased from Sigma-Aldrich (St. Louis, MO), including acetaminophen, 4-methylphenol, diclofenac, β-estradiol, clozapine, 4-hydroxyestrone, ticlopidine, 2,3-dimethylindole, 3-methyloxindole, glutathione, glucose-6-phosphate, 4-hydroxylestrone, β-nicotinamide adenine dinucleotide phosphate (NADP+), and glucose-6-phosphate dehydrogenase. Stable-isotope labeled glutathione [GSX, γ-glutamyl-cystein-glycin-13C2-15N] was obtained

Figure 1. MS methodology of the mass tag-triggered data dependent acquisition. R denotes the drug moiety.


6411

Scheme 1. Bioactivation of 4-Methylphenol in Human Liver Microsomes

from Cambridge Isotope Laboratories (Andover, MA), and isotopic purity was 90% estimated by the supplier using NMR. Pooled human liver microsomes were purchased from Gentest Corp. (Woburn, MA). Methanol was from EMD Chemicals (Gibbstown, NJ). Formic acid. Stable Isotope Trapping. All incubations were performed at 37 °C in a water bath, as previously described.7 Test compounds were mixed with human microsomal proteins in 50 mM potassium phosphate buffer (pH 7.4) supplemented with GSH and GSX at an approximate molar ratio of 1:0.85. Reaction mixtures were prewarmed at 37 °C for 5 min. Reactions were initiated by the addition of a NADPH generating system to give a final volume of 1 mL. Final reaction mixtures contained 10 µM test compounds, 1 mg/mL microsomal proteins, 1 mM GSH and GSX, respectively, 1.3 mM NADP+, 3.3 mM glucose-6-phosphate, 0.4 U/mL glucose6-phosphate dehydrogenase, and 3.3 mM magnesium chloride. After a 60 min incubation, reactions were terminated by the addition of an equal volume of acetonitrile. Samples were centrifuged at 10 000g for 15 min at 4 °C to pellet the precipitated protein, and supernatants were concentrated on a Speed-Vac drier to a final volume of 300 µL approximately. LC-MS/MS Analyses. LC-MS/MS analyses were performed on a LTQ XL linear ion trap mass spectrometer interfaced with an Accela HPLC system (Thermo-Fisher, San Jose, CA). The ESI source was operated under the following conditions: capillary temperature 275 °C, sheath gas flow 19, auxiliary gas flow 42, sweep gas flow 1, source voltage 5.0 kV, and capillary voltage 13 V. Instrument CID settings are below: ion isolation width 2 Da, default charge state 1, normalized collision energy 15 eV, activation Q 0.25, and activation time 30 ms. The minimal signal required to trigger MS2 acquisition was 25 000 counts in the positive mode and 10 000 counts in the negative mode, respectively. An Agilent Zorbax SB C18 column (2.1 mm × 150 mm, 3.5 µm particle size) was used for chromatographic separations. The mobile phases are methanol and 0.1% formic acid in deionized water, and the flow rate was 0.3 mL/min. A generic LC gradient profile was below: 0.0-1.5 min, 95% aqueous and 5% methanol,; 1.5-20 min, 6412


from 5% to 85% methanol; 20-22 min, 85% methanol; 22-24 min, 85% methanol to 95% methanol; 24-26 min, from 95% to 5% methanol. LC-MS/MS analyses were carried out on 10 µL aliquots of samples. RESULTS AND DISCUSSION The Methodology. The methodology is described in Figure 1. In microsomal incubations, the stable isotope trapping results in formation of two isotopic adducts. Because two isotopic partners differ in a mass of 3.0 Da and appear at a fixed intensity ratio, they can be recognized by the computer as a unique MS pattern to selectively trigger data dependent MS2 scans of both labeled and unlabeled GSH conjugates. Considering the possibility that, in the ESI source, the ionization efficiency of a given GSH conjugate can change significantly between the positive and the negative mode, a polarity switch is executed to avoid the bias in ionization polarity. As shown in Figure 1, scan event one is the survey MS full scan in the positive mode, which is followed by mass pattern-triggered MS2 full scans of both isotopic isomer ions (event two). After a polarity switch, the second survey MS full scan is executed in the negative mode followed by the mass pattern triggered MS2 full scans of both isotopic partners in the negative mode (event four). Therefore, MS2 spectra of two isotopic isomers can be acquired in both negative and positive mode in a single run. Positive identification of a particular GSH conjugate can be readily accomplished by examining MS2 spectra of both isotopic isomers. Proof of Concept. The feasibility of the current methodology was demonstrated by analyzing the microsomal sample of 4-methylphenol. In human liver micrsomes, bioactivation of 4-methylphenol produces two reactive metabolites: a quinone methide and 4-methyl-ortho-benzoquinone via oxidation mediated by CYPs.12 These two reactive metabolites were trapped by GSH to form 4-(glutathione-S-yl-methyl)-hydroquinone (MP1, m/z 414) and 3-(glutathione-S-yl)-5-methyl-ortho-hydroquinone (12) Yan, Z.; Easterwood, L. M.; Maher, N.; Torres, R.; Huebert, N.; Yost, G. S. Chem. Res. Toxicol. 2007, 20, 140–148.

Figure 2 Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6413

Figure 2. Mass tag-triggered data dependent MS/MS analysis of reactive metabolites of 4-methylphenol. (A) ESI+ TIC trace of the MS2 scan; (B) ESI- TIC trace of the MS2 scan; (C) ESI+ MS/MS spectrum of MP1; (D) ESI+ MS/MS spectrum of MP2; (E) ESI- MS/MS spectrum of MP1; (F) ESI- MS/MS spectrum of MP2. The minor peak marked by / at RT 8.1 min was a minor isomer of MP2, m/z 430 (ref 14). 6414


Figure 3. Mass tag-triggered data dependent MS/MS analysis of reactive metabolites of diclofenac. (A) ESI+ TIC trace of MS2 scan; (B) ESITIC trace of MS2 scan; (C) ESI+ MS/MS spectrum of DC1 (m/z 583); (D) ESI+ MS/MS spectrum of DC2 (m/z 617). (/, indicates a positive response from an impurity in the incubation).

(MP2, m/z 430), respectively (Scheme 1). As shown in Figure 2, MP1 and MP2 were detected at retention time (RT) 8.4 and 7.9 min, respectively, by the mass tag triggered data-dependent scan in both the positive (Figure 2A) and negative mode (Figure 2B). In the positive mode, MS2 spectra of MP1 showed a precursor ion at m/z 414 and product ions at m/z 179, 285,

308, and 339 (Figure 2C). Ions at m/z 285 and 339 were derived from the characteristic neutral losses of 129 and 75 Da, respectively. For isotope labeled MP1, the precursor ion appeared at m/z 417, and product ions were detected at m/z 182, 288, 311, and 339 (spectrum not shown). The results are Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6415

Figure 4. Mass tag-triggered data dependent MS/MS analysis of reactive metabolites of 3-methyloindole. (A) ESI+ TIC trace of MS2 scan; (B) ESI- TIC trace of MS2 scan; (C) ESI+ MS/MS spectrum of MO1 (m/z 469); (D) ESI+ MS/MS spectrum of MO2 (m/z 453).

consistent with the structure of the conjugate (MP1). Similarly, MP2 exhibited a precursor ion at m/z 430 and product ions at m/z 130, 198, 284, 301, and 355 (Figure 2D). The isotope labeled MP2 displayed a precursor ion at m/z 433 and product ions at 133, 198, 287, 304, and 355 (not shown). The data is also in agreement with the structure of MP2. 6416


In the negative mode, MP1 and MP2 exhibited a quite different CID fragmentation pattern, even two conjugates are highly structure-related. As shown in Figure 2E, MP1 showed a precursor ion at m/z 412, and the GSH anion at m/z 306 as the most abundant product ion, whereas the characteristic

Figure 5. ESI+ MS2 spectra of two “false isotopic partners” at m/z 585 (A) and m/z 588 (B).

Scheme 2. Putative Bioactivation Pathway of 3-Methyloindole in Human Liver Microsomes

product ion at m/z 272 was seen at a very low level. In contrast, MP2 (m/z 428) gave rise to the most prominent ion at m/z

272 (Figure 2F), and other anions at m/z 254 and 179 appeared at very low levels. All these anions (m/z 272, 254, and 179) were derived from the GSH moiety.10 These results clearly indicate that the PI scan at m/z 272 in the negative mode as a generic MS method11 is also biased and sometimes can potentially lead to false negatives. Impacts of Halogenated Compounds. Compounds containing halogens such as Cl and Br can have a high isotopic ratio. The impact of a high isotope ratio on the effectiveness of the current method was evaluated using diclofenac, a dichlorosubstituted compound that is known to form two reactive intermediates. In the presence of GSH, these two reactive intermediates can be captured to form stable conjugates with a molecular weight of 582 and 616, respectively.13,14 In the microsomal incubation sample of diclofenac, two positives (DC1, 14.9 min; DC2, 15.8 min) were successfully detected in both positive (Figure 3A) and negative mode (Figure 3B). As expected, DC1 and DC2 exhibited precursor ions at m/z 583 and 617, respectively, in the positive mode and m/z 581 and 615 in the negative mode (data not shown). As shown in Figure (13) Tang, W.; Stearns, R. A.; Bandiera, S. M.; Zhang, Y.; Raab, C.; Braun, M. P.; Dean, D.; Pang, J.; Leung, K.; Doss, G.; Strauss, J. R.; Kwei, G. Y.; Rushmore, T. H.; Chiu, S.-H. L.; Baillie, T. A. Drug Metab. Dispos. 1999, 27, 365–372. (14) Yan, Z.; Li, J.; Huebert, N.; Caldwell, G. W.; Du, Y.; Zhong, H. Drug Metab. Dispos. 2005, 33, 706–713.


6417

Table 1. Major GSH Conjugates Identified by the Mass Tag-Triggered Data Dependent Data Acquisition

a

ESI+ MS/MS data of isotope labeled conjugates were not listed. b ESI- MS/MS data were not reported.

3C (DC1) and Figure 3D (DC2), MS2 spectra are consistent with the conjugates previously identified.13,14 This example demonstrated that the mass pattern-triggered data dependent acquisition is suitable for detecting reactive metabolites derived from halogen-containing compounds. It should be noted that, for halogen-containing compounds, MS2 scans of less abundant isotopic isomers could also be triggered by the mass pattern-dependent scan. For example, in the case of diclofenac, MS2 scan of precursor ions at m/z 585 and 619 (MH+ + 2) were also triggered by the mass patterndependent scan, in addition to the most abundant isotopic isomers at m/z 583 and 617. Therefore, caution should be taken in data interpretation when halogen-containing compounds are analyzed. It is also important to point out that the molar ratio of two trapping 6418


agents (GSH/GSX) should be under 1:0.9 in the stable isotope trapping incubation. Otherwise, the intensity of the labeled conjugate would be higher than its unlabeled partner for a halogen-containing compound, and such a mass tag can not be recognized by computer to trigger data dependent MS2 scan due to the default setting of the software. As a result, only the minor isotopic partners (m/z 585 and 619 for diclofenac) are detected, and the most abundant isotopic isomers (m/z 583 and 617 for diclofenac) are missing. Polarity Bias in Detection of GSH Conjugates. Traditionally, screening of GSH conjugates is performed in the ESI positive mode. To address the bias in CNL in the positive mode, precursor ion scan at m/z 272 in the negative mode has recently proposed as the survey scan for screening of reactive metabolites10,11 simply

Figure 6. Screening and structural characterization of reactive metabolites derived from 2,3-dimethylindole. (A) ESI+ TIC trace of MS2 scan; (B) ESI- TIC trace of MS2 scan; (C) ESI+ MS/MS spectrum of DM1 (m/z 451); (D) ESI- MS/MS spectrum of DM1 (m/z 449). Responses of DM2-DM6 in both (A) and (B) were enlarged 10-fold in order to be visualized.

by assuming that all GSH conjugates can be ionized comparably well in both polarities. However, the polarity bias in two ESI modes was found during our method development. Figure 4 shows TIC traces of an incubation sample of 3-methyloxindole obtained in both positive and negative mode. In the positive mode (Figure 4A), two positives were detected at 8.0 (MO1) and 9.0 min (MO2), respectively, whereas only one positive (MO1) was

seen in the negative mode (Figure 5B). In the positive mode, MO1 and MO2 showed a precursor ion at m/z 469 and m/z 453 respectively. As shown in Figure 4C, MO1 displayed product ions at m/z 194, 237, and two characteristic ions at m/z 394 and 340 resulting from neutral losses of 75 and 129 Da, respectively. MO2 exhibited product ions at m/z 179, 308, 324, and 378. Presumably, both MO1 and MO2 were GSH conjugates derived from bioactivation of Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6419

Scheme 3. Putative Bioactivation Pathway of 2,3-Dimethylindole in Human Liver Microsomes

3-methyloxindole (Scheme 2). Further confirmation was obtained by MS2 spectra of their isotope labeled partners at m/z 472 and 456 (data not shown). This example clearly suggested that the polarity switch is necessary to ensure a full coverage of different reactive metabolites. Additionally, this case has demonstrated that, similar to CNL in the positive mode, screening of reactive metabolites in the negative mode could also result in false negatives, due to the polarity bias of GSH conjugates. The Validation Study. In order to thoroughly assess the reliability of the method, more compounds were included for further evaluation, which include acetaminophen, clozapine, β-estradiol, 4-hydroxyestrone, clozapine and ticlopidine. As summarized in Table 1, reactive metabolites were detected for all test compounds, and results are inconsistent with the literature.7,10 No false positives were detected for negative controls such as testosterone, ketoconazole, midazolam, and terfenadine. Because of the polarity switch, four MS2 spectra can be acquired in a single run except for those with the polarity bias in ESI. For those uncommon conjugates, tandem MS spectra of two isotopic partners can be obtained in either the positive or negative mode. Therefore, identification can be readily achieved with great confidence by the appearance of some of those characteristic product ions. In the positive mode, in addition to the neutral losses of 129 and 75 Da, the protonated GSH moiety ion (m/z 308 for unlabeled and m/z 311 for labeled) can be used for the positive identification. In the negative mode, anions at m/z 306, 272, 256, and 179 are commonly seen for unlabeled conjugates; for labeled conjugates, these characteristic anions appear at m/z 308, 275, 257, and 179 Da. Although the mass pattern is selective, it is not completely specific to isotopic labeled and unlabeled GSH adducts. In theory, any pair of coeluted precursor ions can potentially trigger a MS2 scan (false positives) if they meet two criteria: a mass difference of 3.0 Da and an intensity ratio in the defined range. Apparently, background ions from mobile phases can 6420


significantly increase false positives, because they often appear over a long period of time and have a greater possibility of pairing with another ion to constitute a “false mass pattern”. Such false responses can be minimized by “dynamic background subtraction” by which some ions detected in a previous scan can be eliminated as background ions from the subsequent scan; as a result, the possibility of two ions pairing to constitute a “false mass pattern” is significantly reduced. In MS2 TIC traces, false responses usually do not show a classic chromatographic peak. Instead, they appear as individual “spikes”. False positives are seen more frequently in the positive mode, but they can be easily eliminated by tandem spectra. For example, analysis of an incubation sample of acetaminophen detected a minor peak that exhibited two partner ions at m/z 585 and m/z 588. As shown in Figure 5, MS2 spectra of two “false isotopic partners” look very different. Also, they do not exhibit those characteristic product ions. A Case Study: Bioactivation of 2,3-Dimethylindole. Metabolic bioactivation of 2-dimethylindole has not been investigated, and therefore, this compound was chosen in the present study to demonstrate the feasibility of the current method for detecting reactive metabolites derived from unknown compounds. Figure 6 show the MS2 TIC traces obtained with the incubation sample of 2-dimethylindole. The most predominant peak (DM1) was observed at 12.5 min in both the positive (Figure 6A) and negative mode (Figure 6B). In the positive mode, DM1 showed a precursor ion at m/z 451 and a very predominant product ion at m/z 308 (Figure 6C). In the negative mode, DM1 exhibited a precursor ion at m/z 449 and a predominant product ion at m/z 306 (Figure 6D). The results appear to suggest that DM1 was derived from direct conjugation of GSH to 2,3-dimethylenenindolenine followed by dehydrogenation (Scheme 3), which is highly similar to bioactivation of 3-methylindole.15 Since DM1 did not show either a neutral loss of 129 Da in the positive mode or a product ion at m/z 272 in the negative mode, this major reactive metabolite could

Figure 7. MS/MS spectra of six minor GSH adducts. (A) a representative ESI+ MS2 spectrum of group I (DM2, DM4); (B) a representative ESI- MS2 spectrum of group I (DM2 and DM4); (C) a representative ESI+ MS2 spectrum of group II (m/z 467, DM3, DM5, DM6); (D) a representative ESI- MS2 spectrum of group II (m/z 465, DM3, DM5, DM6). Analytical Chemistry, Vol. 80, No. 16, August 15, 2008

6421

be missing if either CNL or PI scan was utilized for screening of reactive metabolites. In addition, five other conjugates were detected at 10.2 (DM2), 10.8 (DM3), 11.2 (DM4), 11.8 (DM5), and 12.0 min (DM6) in the same incubation sample by comparing two MS2 TIC traces obtained in positive and negative mode. Because they appeared at a very low level, these conjugates could be readily ignored as noises if the polarity switch was not utilized. The five conjugates (DM2-DM6) are isomers, and they all showed a precursor ion at m/z 469 (m/z 472 for the labeled ones) in the positive mode and m/z 467 (m/z 470 for the labeled ones) in the negative mode. The results appear to suggest that those five adducts were derived from oxidation and subsequent GSH conjugation (Scheme 3), which is also similar to what has been found for 3-methylindole.15 Those six minor conjugates can be divided into two groups (I, DM2 and DM4; II, DM3, DM5,and DM6), based on their tandem MS spectra. In group I, DM2 and DM4 showed identical MS2 spectra. Both conjugates gave rise of the most prominent product ion as the protonated GSH moiety at m/z 308 in the positive mode (Figure 7A) and the GSH anion at m/z 306 in the negative mode (Figure 7B). For group II, ESI+ MS2 spectra of DM3, DM5, and DM6 are highly similar and exhibited three key diagnostic ions at m/z 308 (the protonated GSH), 338 and 392 resulting from the neutral losses of 129 and 75 Da, respectively. In the negative mode, DM3, DM5, and DM6, like many common GSH conjugates, showed a serial of characteristic anions derived from fragmentation of the GSH moiety, which include m/z 306, 272, 254, and 179. From these CID fragmentation pathways, it can be expected that only three minor conjugates, DM3, DM4, and DM5, will be detected by both the CNL and PI scan, while DM2, DM4, and the most abundant DM1 will be missing since they do not show either the neutral loss of 129 Da in the positive mode or the product ion at m/z 272 in the negative mode. This example has clearly demonstrated that misleading results can be obtained for some unusual compounds when a biased MS method such as CNL and PI scan is deployed for screening of reactive metabolites.

proach is sensitive, rugged, and reliable, which is achieved by using the stable isotope trapping in combination with mass tag triggered data dependent acquisition. This study also indicated that the MS behavior of GSH conjugates, such as polarity dependence and CID fragmentation, are more complex than what has originally been recognized. In comparison to other methods such as CNL and PI scan, the current approach offers several advantages. First, because the MS trigger is mass pattern-based, neither conjugate structures nor CID fragmentation pathways are likely to cause false negatives. Second, this approach provides broader coverage of GSH conjugates, regardless of their polarity preference. Additionally, structural identification by the current method is of great confidence, because MS2 spectra of both labeled and nonlabeled conjugates in both positive and negative mode can be obtained for most compounds in a single run. This feature is more useful for positive identification of GSH conjugates with few diagnostic ions. Although high-resolution mass spectrometry in combination of mass defect filtering has been proven as a very effective strategy for the detection and structural characterization of GSH conjugates without concerns of CID pathways, its success is largely limited to those metabolically predictable conjugates.16 As an alternative, the present method does not need to rely on either metabolism prediction or highresolution mass spectrometers. As the application of parallel synthesis and high-throughput screening in drug discovery has resulted in relatively large numbers of drug candidates, the need for a more efficient method to screen for reactive metabolites has become very obvious. Since ion trap mass spectrometers have been widely available, it is reasonable to anticipate that the present method will be rapidly adopted in the mass spectrometry community. Additionally, the same methodology can be potentially applied to triple quadrupole or hybrid triple quadrupole mass spectrometers.

Received for review April 30, 2008. Accepted June 16, 2008. AC800887H

CONCLUSIONS The current study has demonstrated that ion trap mass spectrometers are another powerful MS platform for highthroughput screening of reactive metabolites. The present ap-

6422


(15) Yan, Z.; Easterwood, L. M.; Maher, N.; Torres, R.; Huebert, N.; Yost, G. S. Chem. Res. Toxicol. 2007, 20, 140–148. (16) Zhu, M.; Ma, L.; Zhang, H.; Humphreys, W. G. Anal. Chem. 2008, 79, 8333–8341.

Unbiased High-Throughput Screening of Reactive Metabolites on the

Recommend Documents