Studies on the metabolism of five model drugs by fungi colonizing cadavers using LC-ESI-MS/MS and GC-MS analysis

ORIGINAL PAPER

Studies on the metabolism of five model drugs by fungi colonizingcadavers using LC-ESI-MS/MS and GC-MS analysis

Jorge A. Martínez-Ramírez & Kerstin Voigt &Frank T. Peters

Received: 10 May 2012 /Revised: 18 June 2012 /Accepted: 18 June 2012 /Published online: 24 July 2012# Springer-Verlag 2012

Abstract It is well-known that cadavers may be colonizedby microorganisms, but there is limited information if or towhat extent these microbes are capable of metabolizingdrugs or poisons, changing the concentrations and metabolicpattern of such compounds in postmortem samples. The aimof the present study was to develop a fungal biotransforma-tion system as an in vitro model to investigate potentialpostmortem metabolism by fungi. Five model drugs (ami-triptyline, metoprolol, mirtazapine, promethazine, and zol-pidem) were each incubated with five model fungi known tocolonize cadavers (Absidia repens, Aspergillus repens, As-pergillus terreus, Gliocladium viride, and Mortierella poly-cephala) and with Cunninghamella elegans (positivecontrol). Incubations were performed in Sabouraud mediumat 25 °C for 5 days. After centrifugation, a part of thesupernatants was analyzed by liquid chromatography-tandem mass spectrometry with product ion scanning. An-other part was analyzed by full scan gas chromatography-

mass spectrometry after extraction and derivatization. Allmodel drugs were metabolized by the control fungus result-ing in two (metoprolol) to ten (amitriptyline) metabolites. Ofthe model fungi, only Abs. repens and M. polycephalametabolized the model drugs: amitriptyline was metabolizedto six and five, metoprolol to two and two, mirtazapine tofive and three, promethazine to six and nine, and zolpidemto three and four metabolites, respectively. The main meta-bolic reactions were demethylation, oxidation, and hydrox-ylation. The presented in vitro model is applicable tostudying drug metabolism by fungi colonizing cadavers.

Keywords Metabolism . Fungi . Cadavers .

LC-ESI-MS/MS . GC-MS

Introduction

It is well-known that cadavers are colonized by microorgan-isms like bacteria and fungi which are present either asnatural colonization or as contaminants from the placewhere the dead body was found [1–7]. This colonizationmay become important because the microbes can migratefrom the mucosal surface to different body tissues andfluids, and potentially cause microbiological degradationof drugs and poisons affecting their concentrations and/ormetabolic profiles in the body.

In vitro studies with bacteria such as Escherichia coli andClostridium sp. isolated from post-mortem material [8] andwith clinical urine samples containing confirmed bacterialgrowth [9] have shown that such bacteria are capable ofhydrolyzing the ethanol metabolite ethyl glucuronide. Inanother study, Moriya and Hashimoto [10] reported degra-dation of morphine glucuronides in postmortem material.These phase II metabolites are relevant as alcohol and opiateconsumption markers and their metabolic degradation may

Parts of these results were presented at the 17th GTFCh Symposium,April 14–16, 2011, Mosbach, Germany, and at the 2011 SOFT-TIAFTmeeting, September 25–30, San Francisco, USA.

Electronic supplementary material The online version of this article(doi:10.1007/s00216-012-6212-3) contains supplementary material,which is available to authorized users.

J. A. Martínez-Ramírez : F. T. Peters (*)Institute of Forensic Medicine, Jena University Hospital,Fürstengraben 23,07743 Jena, Germanye-mail: [email protected]

J. A. Martínez-RamírezDepartment of Pharmacy, National University,A.A. 14490 Bogotá, D.C, Colombia

K. VoigtInstitute of Microbiology, Friedrich Schiller University of Jena,Neugasse 24/25,07743 Jena, Germany

Anal Bioanal Chem (2012) 404:1339–1359DOI 10.1007/s00216-012-6212-3

http://dx.doi.org/10.1007/s00216-012-6212-3

therefore cause false-negative analytical findings or lead tomisinterpretation. Likewise, Yajima [11] showed that somespecies of bacteria and yeast, including Candida albicansisolated from antemortem blood diluted by intravenousinfusions of persons involved in traffic accidents, increasedthe postmortem alcohol concentration. Another example formicrobial postmortem degradation is the decrease of theconcentrations of the nitrobenzodiazepines flunitrazepam,clonazepam, and nitrazepam in postmortem blood due totheir reduction to the respective 7-amino metabolites by gutbacteria such as Streptococcus faecalis and Clostridiumperfringens [12]. Except for the above-mentioned referen-ces, postmortem drug metabolism by microbes has not beensystematically studied. Especially the potential involvementof fungi in postmortem drug metabolism has not beenaddressed so far, even though the basic cell functions ofeukaryotic fungi are more closely related to those of humansthan those of prokaryotic bacteria.

The aim of the present study was therefore to develop anin vitro fungal biotransformation model for studying themetabolic capability of fungi colonizing cadavers usinggas chromatography-mass spectrometry (GC-MS) and liq-uid chromatography-electrospray ionization-tandem mass

spectrometry (LC-ESI-MS/MS). Five drugs from differentpharmacological groups, which undergo extensive metabo-lism in humans and are among the most frequently pre-scribed drugs in Germany [13] were selected as modelcompounds. Five fungal species previously reported to bepresent in or on corpses [14–18] and classified as risk Iorganisms by “Technical Rules for Biological Agents” inGermany, were selected as model organisms for the devel-opment of the in vitro model. Additionally, the fungusCunninghamella elegans was used as an incubation control,as it has been widely employed in different models formammalian drug metabolism [19–25].

Materials and methods

Chemicals and reagents

Hydrochlorides of amitriptyline (AT), promethazine (PMZ),and the internal standard cyproheptadine (IS1), tartrate ofmetoprolol (MET), free bases of mirtazapine (MRT), andzolpidem (ZOL) as well as a solutions (0.1 mg/mL) of theinternal standard 7-aminoflunitrazepam-d7 (IS2) in

Table 1 Experiments used inMS/MS with their precursor ionsand the correspondingmetabolites

Drug model Experiment Precursorion

Metabolites / parent drug

Amitriptyline 1 264 Amitriptyline-M (nor-)

2 278 Parent drug

3 280 Amitriptyline-M (nor-HO-) (isomer 1,2,3,4)

Amitriptyline-M (nor-) (N-oxide)

4 294 Amitriptyline-M (HO-) (isomer 1,2,3,4)

Amitriptyline-M (N-oxide)

Metoprolol 1 254 Metoprolol-M (O-demethyl-)

2 268 Parent drug

Metoprolol-M (HCOOH-)

3 284 Metoprolol -M (HO-)

Mirtazapine 1 252 Mirtazapine-M (nor-)

2 266 Parent drug

3 268 Mirtazapine-M (nor-HO-) (isomer 1,2,3,4)

4 282 Mirtazapine-M (HO-) (isomer 1,2,3)

Promethazine 1 271 Promethazine-M (nor-)

2 285 Parent drug

3 287 Promethazine-M (nor-HO-)

Promethazine-M (nor-sulfoxide)

4 301 Promethazine-M (HO-)

Promethazine-M (sulfoxide), Promethazine-M (N-oxide)

5 317 Promethazine-M (sulfoxide-HO-) Promethazine-M (sulfoxide) (N-oxide)

Zolpidem 1 308 Parent drug

2 324 Zolpidem-M (HO-) (isomer 1,2), Zolpidem-M (N-oxide)

3 338 Zolpidem-M (HOOC-) (isomer 1,2)

1340 J.A. Martínez-Ramírez et al.

acetonitrile were purchased from Merck (Darmstadt,Germany). Acetonitrile (mass spec grade), dimethyl sulfox-ide (DMSO), methanol, ethyl acetate, acetic acid, iso-prop-anol, tetramethylammonium hydroxide, dichloromethane,acetic anhydride, and pyridine were obtained from Merck

(Darmstadt, Germany). A stabilized aqueous solution(100,000 Fishmann units per mL) of β-glucuronidase (GRD;EC 3.2.1.31) and arylsulfatase (ARS; EC 3.1.6.1) from Helixpomatia L. was obtained from Roche (Mannheim, Germany).Glucose, protease-peptone, yeast extract, and agar were

Table 2 Metabolites detected by in LC-MS/MS and GC-MS after a biotransformation period of five days with C. elegans (positive control), Abs.repens andM. polycephala and the main formed metabolites (*) in comparison with the main (+) metabolites detected in authentic human urine (√)

Drug/metabolitesdetected by LC-MS/MS

Drug/metabolites/artifactsdetected by GC-MS (MPW library)

Fungi Metabolites detectedin authentic human urine

C. elegans Abs. repens M. polycephala

Fig

2a Amitriptyline (AT)

2b Amitriptyline-M (nor-HO-) (isomer 1) Amitriptyline-M (nor-OH-)-H2O AC √* √ √ √

2c Amitriptyline-M (HO-) (isomer 1) Amitriptyline-M (OH-)-H2O AC √* √ √ √+

2d Amitriptyline-M (nor-HO-) (isomer 2) Amitriptyline-M (nor-OH-) AC √* √

2e Amitriptyline-M (HO-) (isomer 2) Amitriptyline-M (OH-) AC √ √

2f Amitriptyline-M (nor-HO-) (isomer 3) Amitriptyline-M (nor-OH-) AC √ √ √

2g Amitriptyline-M (HO-) (isomer 3) Amitriptyline-M (OH-) AC √ √ √ √

2h Amitriptyline-M (nor-HO-) (isomer 4) Amitriptyline-M (OH-)-H2O √ √

2i Amitriptyline-M (HO-) (isomer 4) Amitriptyline-M (OH-) AC √ √

2j Amitriptyline-M (nor-) Amitriptyline-M (nor-) AC √* √* √* √

2k Amitriptyline-M (N-oxide) Amitriptyline-M (N-oxide)-(CH3)2NOH √* √ √* √

3a Metoprolol (MET)

3b Metoprolol-M (HO-) Metoprolol-M (HO-) 3AC √

3c Metoprolol-M (O-demethyl-) Metoprolol-M (O-demethyl-) 3AC √* √* √ √

3d Metoprolol-M (HOOC-) – √* √* √ √+

4a Mirtazapine (MRT)

4b Mirtazapine-M (HO-) (isomer 1) Mirtazapine-M (OH-) AC √

4c Mirtazapine-M (nor-HO-) (isomer 1) Mirtazapine-M (nor-OH-) AC √ √

4d Mirtazapine-M (HO-) (isomer 2) Mirtazapine-M (OH-) AC √ √ √+

4e Mirtazapine-M (nor-HO-) (isomer 2) Mirtazapine-M (nor-OH-) AC √

4f Mirtazapine-M (nor-) Mirtazapine-M (nor-) AC √* √* √ √

4g Mirtazapine-M (HO-) (isomer 3) Mirtazapine-M (OH-) AC √* √* √ √

4h Mirtazapine-M (nor-HO-) (isomer 3) Mirtazapine-M (nor-OH-) AC √ √ √

4i Mirtazapine-M (nor-HO-) (isomer 4) Mirtazapine-M (nor-OH-) AC √ √ √

5a Promethazine (PMZ)

5b Promethazine-M (sulfoxide-HO-) – √ √ √

5c Promethazine-M (nor-sulfoxide) Promethazine-M (nor-sulfoxide) AC √* √* √* √+

5d Promethazine-M (sulfoxide) Promethazine-M/Artifact (sulfoxide) √* √* √* √+

5e Promethazine-M (sulfoxide) (N-oxide) – √ √ √

5f Promethazine-M (nor-HO-) Promethazine-M (nor-OH-) AC √ √ √ √

5g Promethazine-M (HO-) Promethazine-M (OH-) AC √ √ √

5h Promethazine-M (HO-) (N-oxide) – √ √ √ √

5i Promethazine-M (nor-) Promethazine-M (nor-) AC √ √ √ √

5j Promethazine-M (N-oxide) – √ √ √

6a Zolpidem (ZOL)

6b Zolpidem-M (4′-HO-) Zolpidem-M (4′-HO-) AC √* √ √ √

6c Zolpidem-M (4′-HOOC-) – √ √ √+

6d Zolpidem-M (6-HO-) Zolpidem-M (6-HO-) AC √ √ √ √

6e Zolpidem-M (6-HOOC-) – √ √

6f Zolpidem-M (N-oxide) – √ √ √ √

Metabolism of five model drugs by fungi colonizing cadavers 1341

obtained from Carl Roth (Karlsruhe, Germany). Potassiumdihydrogen phosphate, dipotassium hydrogen phosphate, andmagnesium sulfate were obtained from neoLab (Heidelberg,Germany). All chemicals were analytical grade or higher.

Solutions and biological samples

Working solutions (1 mmol/L, free base) of the model drugswere prepared in sterile water with exception of the MRT

Fig. 1 TIC profile chromatograms of AT (a), MET (b), MRT (c), PMZ (d), and ZOL (e) and their metabolites obtained after an incubation periodfor 5 days at 25 °C with C. elegans in comparison with the fungi Abs. repens and M. polycephala


solution which was prepared in sterile water–DMSO (9:1 v/v).Working solutions of IS1 (20 mg/L) and IS2 (1 mg/L) wereprepared in methanol. All solutions were stored at 4 °C for upto 1 month. In order to compare human and fungal metabo-lism, 90 authentic human urine samples from routine clinicaltoxicology casework containing the model drugs AT (n017),MET (n018), MRT (n019), PMZ (n021), and ZOL (n015)and/or their metabolites were analyzed in the same way asincubation supernatants (see below).

Fungal strains and incubations

The six fungal strains, C. elegans FSU 10230 (C. elegans),Absidia repens FSU 4726, Aspergillus repens FSU 1443,Aspergillus terreus FSU 2938, Gliocladium viride FSU5326, and Mortierella polycephala FSU 696 were providedby the Fungal Reference Center of Jena. Each of the modelfungi was incubated during three days in 9 mL Sabour-aud medium (SM) at 25 °C and a shaking velocity of90 rpm. After incubation, 1 mL of AT, MET, MRT,PMZ, or ZOL solution (1 mmol/L) was added to each

flask and incubation was continued for another fivedays. This procedure was carried out three times onthree different days in order to assess the repeatabilityduring the incubation process. C. elegans was used as apositive control (PC). Additionally, blank incubations(BI) containing the model fungi but none of the modeldrugs and negative control (NC) incubations of themodel drugs with SM only were performed (NC-AT,NC-MET, NC-MRT, NC-PMZ, and NC-ZOL). From allincubation mixtures 800 μL samples were drawn every24 h and immediately centrifuged for 3 min at 9,600×g.

Sample preparation for liquid chromatography-tandem massspectrometry-based analysis

One part (50 μL) of the supernatants of the fungal incuba-tions was transferred to a 2-mL autosampler vial, 10 μL ofIS2 was added and the mixture was diluted with 940 μL ofaqueous ammonium formate solution (50 mmol/L, pH 3.0).Urine samples were first subjected to enzymatic hydrolysis.For this purpose, 50 μL of β-glucuronidase/arylsulfatase

Fig. 1 (continued)


solution and 100 μL of sodium acetate buffer (2 mol/L,pH 5.5) were added to 1 mL of urine and the resultingmixture was incubated for 90 min at 55 °C. After centrifu-gation, a 50-μL aliquot of the hydrolyzed urine was further

treated in the same way as the supernatants of the fungalincubations. Aliquots of 50 μL of prepared incubationsupernatants and urine samples were analyzed by LC-MS/MS.

Fig. 2 ESI-MS/MS enhanced product ion (EPI) mass spectra of AT metabolites with liquid chromatographic retention times (RT), postulatedstructures, and predominant fragmentation patterns (a–k). The numbers of the respective peaks in Fig. 1a are given below each spectrum


Sample preparation for gas chromatography-massspectrometry-based analysis

The rest (750 μL) of the supernatants of the fungalincubations was transferred to a 2-mL Eppendorf tube,

5 μL of IS1 was added, and the mixture was adjustedto pH 8–9 by adding 200 μL of 2.5 mol/L phosphatebuffer. After adding 500 μL of ethyl acetate/iso-propa-nol/dichloromethane (60:20:20, v/v/v), the vials werecapped, left on a rotary shaker for 2 min, and finally

Fig. 2 (continued)


centrifuged for 3 min at 9,600×g. Thereafter, 500 μL ofthe upper organic phase were transferred to a cleanautosampler vial and evaporated under a stream ofnitrogen at 40 °C. The residue was acetylated with50 μL of an acetic anhydride/pyridine mixture (3:2; v/v)under microwave irradiation (5 min, 450 W). After

evaporation of the derivatization reagent under a streamof nitrogen at 40 °C, the derivatized extract was dis-solved in 50 μL of methanol and 2 μL were analyzedby GC-MS. The rest of the hydrolyzed urine samplesdescribed in the previous section was treated in thesame way.

Fig. 2 (continued)


LC-MS/MS analysis

The LC-MS/MS system consisted of LC-20AD HPLCsystem (Shimadzu, Jena, Germany) interfaced with a4000 Q Trap® mass spectrometer (AB Sciex, Darmstadt,Germany) equipped with a TurboIonSpray ESI sourceoperated in the positive mode. Nitrogen was used ascurtain, source, and collision gas. The system was con-trolled by Analyst 1.5.1. software which was also usedfor data analysis.

For the chromatographic separation, an Eclipse XDB C18

5 μm (4.1×150 mm) column was used with the followinglinear gradient of the mobile phases A (aqueous ammoniumformate, 50 mmol/L, pH 3.0) and B (0.1 % formic acid inacetonitrile): 10 % B to 60 % B within 10 min. A 2 minequilibration to 10 % B was used between injections. Theflow rate was 1.4 mL/min and the injection volume was50 μL.

For targeted metabolite search, data acquisition wasperformed in the product ion scan mode with the

following interface settings: source temperature, 550 °C;ion spray voltage, 5500 V; ion source gas 1 and 2,75; curtain gas, 20; collision gas, high; declusteringpotential, 50 V. The MS/MS experiments performedduring the run were drug specific with the pseudomo-lecular ions of the parent drug and major metabolitesdescribed in the literature being used as precursor ions.The collision energy (CE) was 25 V in all experimentsand the scan range for the product ion spectra was fromm/z 50 to the respective pseudomolecular ion +1. Themonitored precursor ions for each model drug and itsmetabolites are listed in Table 1.

For untargeted metabolite search, data acquisition andmass spectrometric evaluation was performed as proposedby Sauvage et al [26] with modifications. Information-dependent acquisition (IDA) using the enhanced MS(EMS) mode for the survey scan and the enhanced production (EPI) mode for the dependent scan. For EMS, thedynamic fill-time option was used, with a mass range ofm/z 50–360 at a rate of 4,000 amu/s. For EPI, the CE was

Fig. 2 (continued)


set at 40 V and the collision energy spread (CES) at 25 V.Mass range and scan rate were the same as for the surveyscan. The complete cycle lasted 0.38 s. The interface set-tings were the same as described in the previous paragraph.

Matrix effect experiments

For evaluation of matrix effects, i.e., ion suppression orenhancement, post-column infusion was performed as

Fig. 3 ESI-MS/MS enhanced product ion (EPI) mass spectra of MET metabolites with liquid chromatographic retention times (RT), postulatedstructures, and predominant fragmentation patterns (a–d). The numbers of the respective peaks in Fig. 1b are given below each spectrum


proposed by Bonfiglio et al. [27]. A solution containing amixture of all model drugs at a concentration of 0.01 mmol/Leach in a mixture of mobile phases A and B (50:50 v/v) wasinfused into the source at a constant rate of 50μL/min via a post-column tee connection. At the same time prepared supernatantsof negative control incubations with each model fungus (day 5)were injected on the column while the MS was operated in theproduct ion scan mode as described in the previous section.

GC-MS analysis

GC-MS analyses were performed on an Agilent 6890 GCconnected to a 5973N MS (Agilent, Darmstadt, Germany).The GC conditions were as follows: injection volume, 2 μL;pulsed splitless injection; column, V-5MS (Varian, Darmstadt,Germany) 30 m×0.25 mm, 0.25 μm film thickness; injectionport temperature, 300 °C; carrier gas, helium; flow rate,0.6 mL/min; column temperature, programmed from 100 °Cto 310 °C at 30 °C/min; final time, 5 min; The MS conditionswere as follows: full-scan mode (m/z 50–600); electron ioni-zation (EI) mode; ionization energy, 70 eV; ion source tem-perature, 230 °C; interface temperature, 280 °C. Dataevaluation was performed using AMDIS software with aMaurer/Pfleger/Weber (MPW) 2007 target library and decon-volution parameters according to Meyer et al [28].

Results and discussion

GC-MS and LC-MS/MS analysis

The metabolites of the five model drugs which were detectedin incubation samples with the five model organisms or inhuman urine are listed in Table 2. The vast majority of themetabolites were detected both by GC-MS and LC-MS/MS.

However, the N-oxide and aliphatic hydroxy metabolites ofATwere only detectable as artifacts by GC-MS after elimina-tion of hydroxylamine or water, respectively. In addition, thecarboxy metabolites of metoprolol and zolpidem, and the N-oxide metabolites of promethazine and zolpidem were onlyfound by LC-MS/MS. While it can be expected that thementioned carboxy metabolites would also be detectable byGC-MS after alternative workup with methylation or silyla-tion, N-oxides are generally thermolabile compounds notamenable to GC-MS analysis. This shows that LC-MS/MShas distinct advantages over GC-MS due to the possibility todirectly inject diluted incubation supernatants avoiding extrac-tion losses and the need for derivatization.

Total ion chromatograms of supernatants of incubations ofthe five model drugs with the model fungi Abs. repens andM.polycephala as well as the PC C. elegans are shown in Fig. 1.With exception of the co-eluting hydroxy and carboxy metab-olites of metoprolol (peaks 1 and 2 in Fig. 1b) and combinedN-demethyl, hydroxyl and sulfoxide metabolites of prometha-zine (peaks 2–4 in Fig. 1d), which could however bedifferentiated via their different molecular masses and frag-mentation patterns, all metabolites were sufficiently separatedunder the specified chromatographic conditions. Post-columninfusion experiments showed only a minor suppression zonearound the retention time 1 min, whereas even the morehydrophilic metabolites of the model compounds alleluted at 3 min or later (see Fig. S1 in the ElectronicSupplementary Material that accompanies the onlineversion of this article). Using the given EPI settings,fragment-rich EPI mass spectra were obtained allowinginterpretation of fragmentation patters for elucidation ofthe metabolite structures. The EPI mass spectra of themetabolites detected in the present study are shown inFigs. 2, 3, 4, 5, and 6 along with the proposed struc-tures and fragmentation patterns.

Fig. 3 (continued)


Identification of metabolites

AT and all of its metabolites show three characteristic frag-mentation patterns: neutral loss of dimethylamine (Fig. 2a,c, e, g, i), methylamine (Fig. 2b, d, f, h, j), or

dimethylhydroxylamine (Fig. 2k), α-cleavage of the sidechain (Fig. 2a–k), and loss of the side chain. The hydroxymetabolites carrying an aliphatic hydroxyl moiety are fur-ther characterized by a neutral loss of water (Fig. 2b, c, f, g)which is not present in the EPI spectra of hydroxy

Fig. 4 ESI-MS/MS enhanced product ion (EPI) mass spectra of MIR metabolites with liquid chromatographic retention times (RT), proposedstructures, and predominant fragmentation patterns (a–i). The numbers of the respective peaks in Fig. 1c are given below each spectrum


metabolites carrying a phenolic hydroxyl moiety (Fig. 2e, h, i).A further differentiation of the formed isomeric hydroxymetabolites regarding the exact position and stereochemistryof hydroxylation is not possible based on the EPI spectra alone.

According to Ma et al. [22], the common fragmentationpathways of MET and its metabolites can be interpreted as

follows: neutral loss of the propyl moiety, and common neu-tral losses of isopropylamine and water (Fig. 3a–d). The latterfragmentation reaction leads to fragments with m/z 191 forMETand its carboxy metabolite (Fig. 3a, d),m/z 177 for itsO-demethyl metabolite (Fig. 3c) and m/z 207 for its hydroxymetabolite (Fig. 3b). A further loss of methanol and water

Fig. 4 (continued)


from MET and O-demethyl MET, respectively, leads to thecommon fragment m/z 159 (Fig. 3a, c). Further loss of meth-anol from the hydroxy metabolite leads to a fragment at m/z175 (Fig. 3b). In contrast, the respective fragment of thecarboxy metabolite further reacts by loss of carbon dioxideleading to a fragment of m/z 145 (Fig. 3d).

A common fragmentation reaction of MRT and all ofits metabolites is a cleavage through the piperazinemoiety leaving a fragment with m/z 195 [29] or m/z211 representing the unchanged (Fig. 4a, e, f, g, h, i) orhydroxylated (Fig. 4b, c, d) tricyclic ring system, re-spectively. The sites of ring hydroxylation cannot be

Fig. 4 (continued)


further determined by interpretation of mass spectra.Moody et al. [23] described hydroxylation of MRT inpositions 8, 12, and 13 after incubation with C. elegansso it can be presumed that the EPI spectra shown inFig. 4b and d represent metabolites hydroxylated at oneof these three positions. Considering the mass shift of +16in comparison to the parent drug, the metabolite with a

pseudo-molecular ion mass of 282 and an unchanged tricyclicring system (Fig. 4g) could either be hydroxylated at thepiperazine moiety or an N-oxide. Considering the relativelyprominent neutral loss of water in its EPI spectrum hydroxyl-ation of the piperazine ring seems much more likely. Accord-ingly, the metabolites with pseudomolecular ions at m/z 268and unchanged tricyclic ring systems can be interpreted as

Fig. 5 ESI-MS/MS enhanced product ion (EPI) mass spectra of PMZ metabolites with liquid chromatographic retention times (RT), proposedstructures, and predominant fragmentation patterns (a–j). The numbers of the respective peaks in Fig. 1d are given below each spectrum


combinations of N-demethylation and hydroxylation at thepiperazine moiety.

The common fragmentation reactions of PMZ and all ofits metabolites is cleavage between the tricyclic ring systemand the side chain. For the parent drug (Fig. 5a) and allmetabolites with an unchanged side chain (Fig. 5b, d, g) thiscleavage leads to a fragment at m/z 86 representing theunchanged side chain. A fragment at m/z 86, albeit less

abundant, can also be observed in the N-oxide metabolites(Fig. 5e, h, j). However, unchanged and N-oxidized sidechains can be differentiated via further fragments. If the sidechain is unchanged, a neutral loss of dimethylamine (45mass units) can also be observed (Fig. 5a, b, d, g), whereasmetabolites with N-oxidated side chains show a loss of N,N-dimethylhydroxylamine (61 mass units; Fig. 5e, h, j). Accord-ingly, the N-demethyl metabolites show a fragment at m/z 72

Fig. 5 (continued)


Fig. 5 (continued)


which is accompanied by a neutral loss of methylaminecorresponding to 31 mass units (Fig. 5f, i) with exception ofthe N-demethylated sulfoxide metabolite (Fig. 5c).Promethazine itself and its metabolites with an un-changed ring system are characterized by a fragment at

m/z 198. In metabolites with hydroxylated ring systems,the corresponding fragment is observed at m/z 214(Fig. 5f, g, h). Again the position of hydroxylationcannot be determined based on the fragmentation pat-tern. The peak with the EPI spectrum shown in Fig. 5d

Fig. 6 ESI-MS/MS enhanced product ion (EPI) mass spectra of ZOL metabolites with liquid chromatographic retention times (RT), proposedstructures, and predominant fragmentation patterns (a–f). The numbers of the respective peaks in Fig. 1e are given below each spectrum


was also detected in the NC incubations and hence mostlikely belongs to promethazine sulfoxide which is notonly a metabolite but can only be formed by non-metabolic oxidation of promethazine. Due to similaritiesin fragmentation pattern of this spectrum with thoseshown in Fig. 5c, e, the latter two can be assigned to

the sulfoxides of N-demethyl-promethazine and prome-thazine N-oxide, respectively.

ZOL and all of the metabolites detected in the present studyshare three major fragmentation reactions: cleavage of theheterocyclic ring system, and cleavages on both sides of thecarbonyl group in the side chain. The first cleavage leaves a

Fig. 6 (continued)


methylpyridinium ion with m/z 92 from unchanged hetero-cycles (Fig. 6a, b, c, f), m/z 108 from methyl hydroxylatedheterocycles (Fig. 6d), and m/z 122 for ring systems in whichthe methyl group of the heterocyclic ring system has beenoxidized to a carboxyl moiety (Fig. 6e). Both of these meta-bolic reactions have been described in the literature [30,31]. Afurther hydroxy metabolite carrying the hydroxyl function atthe tolyl moiety has also been described in the literature [30].Considering the mass shifts of +16 mass units in comparisonto both side chain fragments of the parent drug zolpidem, theEPI spectra shown in Fig. 6b and f would both be in line withthis structure. However, as shown in Fig. 1e, only the peakcorresponding to the spectrum in Fig. 6b elutes before theparent drug as expected on a reversed phase column. Hence,this spectrum can be assigned to the tolyl hydroxyl metabolite.The isobaric metabolite (Fig. 6f) elutes after the parent drugwhich would be in line with its postulated N-oxide structure.Finally, the mass shifts of +30 mass units in both side chainfragments in comparison to the parent can be assigned to thetolyl carboxy metabolite which has also been described in theliterature [30,31].

Formation of metabolites by the model fungi

Of the model fungi used in the present study, Asp. terreus,Asp. repens, and G. viride were not able to form any detect-able amounts of metabolites under the employed conditions.Although the supernatants of incubations of these fungi withpromethazine contained minor peaks of promethazine sulf-oxide it could not be attributed to metabolism of this drugbecause peaks of similar abundance were also present in theNC with PMZ (see Fig. S2 in the Electronic SupplementaryMaterial). Apart from that, no relevant peaks were detectedin any of the BI (without fungi) or NC (without drug).

The metabolites detected in incubation supernatants ofthe other two model fungi Abs. repens and M. polycephala,the positive control fungus C. elegans, and in authentichuman urine samples are listed in Table 2. As can be seenfrom this table, C. elegans was capable of producing allmetabolites detected in human urine with exception of theα-hydroxy metabolite of metoprolol. In addition, it pro-duced a number of metabolites not detected in the authentichuman urine samples which partly were also produced bythe model fungi Abs. repens and M. polycephala. Thiscapability of C. elegans to produce a very wide range ofhuman and/or fungal drug metabolites is in accordance withprevious findings of other authors [22–24] and confirms theapplicability of this C. elegans as a positive control organ-ism in studies on fungal drug metabolism.

While the model fungi Abs. repens and M. polycephaladid not produce as many metabolites as C. elegans they stillformed multiple metabolites also formed in humans and twometabolites of mirtazapine (Fig. 4h) and promethazine

(Fig. 5j) which were not detected in human urine samplesor not described in previously published metabolism studiesof these drugs. Such metabolites could therefore be potentialmarkers for differentiating postmortem fungal metabolismfrom antemortem human metabolism.

Generally the peak areas of the metabolites increased in amore or less linear fashion with incubation times. The onlyexceptions were the two hydroxmethyl metabolites of zol-pidem and the O-demethyl metabolite of metoprolol thepeaks of which reached a maximum after approximatelytwo days of incubation and decreased after that. This de-crease can be explained by the fact that these metaboliteswere further oxidized to the respective carboxy metabolites.Replicate incubations showed that metabolite formation wasfairly reproducible with coefficients of variations for thepeak areas ranging from 7.6 % to 33.5 % (n03; see TableS1 in the Electronic Supplementary Material)

Conclusions

The presented data indicate that the described system of fungalincubations followed by GC-MS and LC-MS/MS analysis ofdiluted incubation supernatants is applicable for studies onfungal drug metabolism. It provides sufficient separation andselectivity for detection of known and at least tentative iden-tification of previously unknown metabolites. Major interfer-ences or ion suppression/enhancement by the supernatantmatrix are not to be expected. C. elegans proved to be usefulas positive control organism for studies on fungal drug me-tabolism. Altogether, the described approach should proveuseful for the first systematic studies on postmortem metabo-lism by fungi colonizing cadavers and for identifying potentialmarker metabolites to detect such effects.

Acknowledgments Jorge Martinez thanks DAAD for the support ofthis project and the Gesellschaft für Toxikologische und ForensicheChemie (Society of Toxicological and Forensic Chemistry, GTFCh) forproviding the travel fund to present part of this work at the 2011 SOFT-TIAFT Meeting in San Francisco-California. The authors further thankDr Daniela Remane and Dr Christoph Sauer for their assistance withthe preparation of the manuscript.

References

1. Carter DO, Yellowlees D, Tibbett M (2007) Cadaver decomposi-tion in terrestrial ecosystems. Naturwissenschaften 94:12–24

2. Guarner F, Malagelada JR (2003) Gut flora in health and disease.Lancet 361:512–519

3. Thorn JL, Gilchrist KB, Sobonya RE, Gaur NK, Lipke PN, KlotzSA (2010) Postmortem candidaemia: marker of disseminated dis-ease. J Clin Pathol 63:337–340

4. Daldrup T, Huckenbeck W (1984) Significance of the putrefactivebacterium Clostridium sordellii for the determination of age of thecadaver. Z Rechtsmed 92:121–125


5. Morris JA, Harrison LM, Partridge SM (2006) Postmortem bac-teriology: a re-evaluation. J Clin Pathol 59:1–9

6. Kellerman GD, Waterman NG, Scharefenberger LF (1976) Dem-onstration in vitro of postmortem bacterial transmigration. Am JClin Pathol 66:911–915

7. Paakko P, Nurmi T, Sarkioja T, Hirvonen J, Sutinen S (1986) Post-mortem bacterial culture of bronchial mucus and heart blood inhospital and non-hospital autopsies: effect of morgue time andlength of hospitalization. Zentralbl Bakteriol Mikrobiol Hyg B182:360–371

8. Baranowski S, Serr A, Thierauf A, Weinmann W, Grosse Perde-kamp M, Wurst FM, Halter CC (2008) In vitro study of bacterialdegradation of ethyl glucuronide and ethyl sulphate. Int J LegalMed 122:389–393

9. Helander A, Dahl H (2005) Urinary tract infection: a risk factor forfalse-negative urinary ethyl glucuronide but not ethyl sulfate in thedetection of recent alcohol consumption. Clin Chem 51:1728–1730

10. Moriya F, Hashimoto Y (1997) Distribution of free and conjugatedmorphine in body fluids and tissues in a fatal heroin overdose: isconjugated morphine stable in postmortem specimens? J ForensicSci 42:736–740

11. Yajima D, Motani H, Kamei K, Sato Y, Hayakawa M, Iwase H(2006) Ethanol production by Candida albicans in postmortemhuman blood samples: effects of blood glucose level and dilution.Forensic Sci Int 164:116–121

12. Robertson MD, Drummer OH (1995) Postmortem drug metabo-lism by bacteria. J Forensic Sci 40:382–386

13. Schwabe U, Paffrath D (2009) Arzneiverordnungsreport. Springer,Berlin

14. Ishii K, Hitosugi M, Kido M, Yaguchi T, Nishimura K, Hosoya T,Tokudome S (2006) Analysis of fungi detected in human cadavers.Leg Med (Tokyo) 8:188–190

15. Hitosugi M, Ishii K, Yaguchi T, Chigusa Y, Kurosu A, Kido M,Nagai T, Tokudome S (2006) Fungi can be a useful forensic tool.Leg Med (Tokyo) 8:240–242

16. Hawksworth DL, Wiltshire PE (2011) Forensic mycology: the useof fungi in criminal investigations. Forensic Sci Int 206:1–11

17. Tibbett M, Carter DO (2003) Mushrooms and taphonomy: thefungi that mark woodland graves. Mycologist 17:20–24

18. Carter DO, Tibbett M (2003) Taphonomic mycota: fungi withforensic potential. J Forensic Sci 48:168–171

19. Wackett LP, Gibson DT (1982) Metabolism of xenobiotic com-pounds by enzymes in cell extracts of the fungus Cunninghamellaelegans. Biochem J 205:117–122

20. Asha S, Vidyavathi M (2009) Cunninghamella—a microbial modelfor drug metabolism studies—a review. Biotechnol Adv 27:16–29

21. Zhang D, Yang Y, Leakey JE, Cerniglia CE (1996) Phase Iand phase II enzymes produced by Cunninghamella elegansfor the metabolism of xenobiotics. FEMS Microbiol Lett138:221–226

22. Ma B, Huang HH, Chen XY, Sun YM, Lin LH, Zhong DF (2007)Biotransformation of metoprolol by the fungus Cunninghamellablakesleeana. Acta Pharmacol Sin 28:1067–1074

23. Moody JD, Freeman JP, Fu PP, Cerniglia CE (2002) Biotransfor-mation of mirtazapine by Cunninghamella elegans. Drug MetabDispos 30:1274–1279

24. Zhang D, Evans FE, Freeman JP, Duhart B Jr, Cerniglia CE (1995)Biotransformation of amitriptyline by Cunninghamella elegans.Drug Metab Dispos 23:1417–1425

25. Bjornstad K, Helander A, Hulten P, Beck O (2009) Bioanalyticalinvestigation of asarone in connection with Acorus calamus oilintoxications. J Anal Toxicol 33:604–609

26. Sauvage FL, Saint-Marcoux F, Duretz B, Deporte D, Lachatre G,Marquet P (2006) Screening of drugs and toxic compounds withliquid chromatography-linear ion trap tandem mass spectrometry.Clin Chem 52:1735–1742

27. Bonfiglio R, King RC, Olah TV, Merkle K (1999) The effects ofsample preparation methods on the variability of the electrosprayionization response for model drug compounds. Rapid CommunMass Spectrom 13:1175–1185

28. Meyer MR, Peters FT, Maurer HH (2010) Automated mass spec-tral deconvolution and identification system for GC-MS screeningfor drugs, poisons, and metabolites in urine. Clin Chem 56:575–584

29. Doherty B, Rodriguez V, Leslie JC, McClean S, Smyth WF (2007)An electrospray ionisation tandem mass spectrometric investigationof selected psychoactive pharmaceuticals and its application in drugand metabolite profiling by liquid chromatography/electrospray ion-isation tandem mass spectrometry. Rapid Commun Mass Spectrom21:2031–2038

30. Vajta S, Thenot JP, de Maack F, Devant G, Lesieur M (1988)Thermospray liquid chromatography tandem mass spectrometry:application to the elucidation of zolpidem metabolism. BiomedEnviron Mass Spectrom 15:223–228

31. Klupsch F, Houssin R, Humbert L, Imbenotte M, Henichart JP,Lhermitte M (2006) Major metabolites of zolpidem: expeditioussynthesis and mass spectra. Chem Pharm Bull (Tokyo) 54:1318–1321


Documents

Studies on the metabolism of five model drugs by fungi colonizing cadavers using LC-ESI-MS/MS and GC-MS analysis