Oxygen-18 exchange reactions in steroidal ketones. Determination of relative rates of incorporation by gas chromatography-mass spectrometry

In-depth gas chromatography/tandem mass spectrometry fragmentation analysis of formestane and evaluation of mass spectral discrimination of isomeric 3-keto-4-ene hydroxy steroids

RATIONALE

The aromatase inhibitor formestane (4-hydroxyandrost-4-ene-3,17-dione) is included in the World Anti-Doping Agency's List of Prohibited Substances in Sport. However, it also occurs endogenously as do its 2-, 6- and 11-hydroxy isomers. The aim of this study is to distinguish the different isomers using gas chromatography/electron ionization mass spectrometry (GC/EI-MS) for enhanced confidence in detection and selectivity for determination.

METHODS

Established derivatization protocols to introduce [² H₉ ]TMS were followed to generate perdeuterotrimethylsilylated and mixed deuterated derivatives for nine different hydroxy steroids, all with 3-keto-4-ene structure. Formestane was additionally labelled with H₂ ¹⁸ O to obtain derivatives doubly labelled with [² H₉ ]TMS and ¹⁸ O. GC/EI-MS spectra of labelled and unlabelled TMS derivatives were compared. Proposals for the generation of fragment ions were substantiated by high-resolution MS (GC/QTOFMS) and tandem mass spectrometry (MS/MS) experiments.

RESULTS

Subclass-specific fragment ions include m/z 319 for the 6-hydroxy and m/z 219 for the 11-hydroxy compounds. Ions at m/z 415, 356, 341, 313, 269 and 267 were indicative for the 2- and 4-hydroxy compounds. For their discrimination the transition m/z 503 → 269 was selective for formestane. In 2-, 4- and 6-hydroxy steroids loss of a TMSO radical takes place as cleavage of a TMS-derived methyl radical and a neutral loss of (CH₃ )₂ SiO. Further common fragments were also elucidated.

CONCLUSIONS

With the help of stable isotope labelling, the structures of postulated diagnostic fragment ions for the different steroidal subclasses were elucidated. ¹⁸ O-labelling of the other compounds will be addressed in future studies to substantiate the obtained findings. To increase method sensitivity MS³ may be suitable in future bioanalytical applications requiring discrimination of the 2- and 4-hydroxy compounds.

Mass spectral fragmentation analyses of isotopically labelled hydroxy steroids using gas chromatography/electron ionization low-resolution mass spectrometry: A practical approach

RATIONALE

Gas chromatography coupled to electron ionization mass spectrometry (GC/EI-MS) is used for routine screening of anabolic steroids in many laboratories after the conversion of polar groups into trimethylsilyl (TMS) derivatives. The aim of this work is to elucidate the origin and formation of common and subclass-specific fragments in the mass spectra of TMS-derivatized steroids. Especially in the context of metabolite identification or analysis of designer drugs, isotopic labelling is helpful to better understand fragment ion generation, identify unknown compounds and update established screening methods.

METHODS

Stable isotope labelling procedures for the introduction of [² H₉ ]-TMS or ¹⁸ O were established to generate perdeuterotrimethylsilylated, mixed deuterated and ¹⁸ O-labelled derivatives for 13 different hydroxy steroids. Fragmentation proposals were substantiated by comparison of the abundances of isotopically labelled and unlabelled fragment ions in unit mass resolution GC/MS. Specific fragmentations were also investigated by high-resolution MS (GC/quadrupole time-of-flight MS, GC/QTOFMS).

RESULTS

Methyl radical cleavage occurs primarily from the TMS groups in saturated androstanes and from the steroid nucleus in the case of enol-TMS of oxo or α,β-unsaturated steroid ketones. Loss of trimethylsilanol (TMSOH) is dependent on steric factors, degree of saturation of the steroid backbone and the availability of a hydrogen atom and TMSO group in the 1,3-diaxial position. For the formation of the [M - 105]⁺ fragment ion, methyl radical cleavage predominates from the angular methyl groups in position C-18 or C-19 and is independent of the site of TMSOH loss. The common [M - 15 - 76]⁺ fragment ion was found in low abundance and identified as [M - CH₃ - (CH₃ )₂ SiH - OH]⁺ . For the different steroid subclasses further diagnostic fragment ions were discussed and structure proposals postulated.

CONCLUSIONS

Stable isotope labelling of oxo groups as well as derivatization with deuterated TMS groups enables the detection of structure-related fragment ion generation in unit mass resolution GC/EI-MS. This may in turn allow us to propose isomeric assignments that are otherwise almost impossible using MS only.

Reconsidering mass spectrometric fragmentation in electron ionization mass spectrometry - new insights from recent instrumentation and isotopic labelling exemplified by ketoprofen and related compounds

RATIONALE

In various fields of chemical analyses, structurally unknown analytes are considered. Proper structure confirmation may be challenged by the low amounts of analytes that are available, e.g. in early stage drug development, in metabolism studies, in toxicology or in environmental analyses. In these cases, mass spectrometric techniques are often used to build up structure proposals for these unknowns. Fragmentation reactions in mass spectrometry are known to follow definite pathways that may help to assign structural elements by fragment ion recognition. This work illustrates an investigation of fragmentation reactions for gas chromatography/electron ionization mass spectrometric characterization of benzophenone derivatives using the analgesic drug ketoprofen and seven of its related compounds as model compounds.

METHODS

Deuteration and ¹⁸ O-labelling experiments along with high-resolution accurate mass and tandem mass spectrometry (MS/MS) were used to further elucidate fragmentation pathways and to substantiate rationales for structure assignments. Low-energy ionization was investigated to increase confidence in the identity of the molecular ion.

RESULTS

The high-resolution mass analyses yielded unexpected differences that led to reconsideration of the proposals. Site-specific isotopic labelling helped to directly trace back fragment ions to their respective structural elements. The proposed fragmentation pathways were substantiated by MS/MS experiments.

CONCLUSIONS

The described method may offer a perspective to increase the level of confidence in unknown analyses, where reference material is not (yet) available.

Asymmetric synthesis, structure, and reactivity of unexpectedly stable spiroepoxy-beta-lactones including facile conversion to tetronic acids: application to (+)-maculalactone A

A novel class of small spirocyclic heterocycles, spiroepoxy-beta-lactones (1,4-dioxaspiro[2.3]-hexan-5-ones), is described that exhibit a number of interesting reactivity patterns. These spiroheterocycles, including an optically active series, are readily synthesized by epoxidation of ketene dimers (4-alkylidene-2-oxetanones) available from homo- or heteroketene dimerization. An analysis of bond lengths in these systems by X-ray crystallography and comparison to data for known spirocycles and those determined computationally suggest that anomeric effects in these systems may be more pronounced due to their rigidity and may contribute to their surprising stability. The synthetic utility of spiroepoxy-beta-lactones was explored, and one facile rearrangement identified under several conditions provides a three-step route from acid chlorides to optically active tetronic acids, ubiquitous heterocycles in bioactive natural products. The addition of various nucleophiles to these spirocycles leads primarily to addition at C5 and C2. The utility of an optically active spiroepoxy-beta-lactone was demonstrated in the concise, enantioselective synthesis of the antifouling agent, (+)-maculalactone A, which proceeds in five steps from hydrocinnamoyl chloride by way of a tetronic acid intermediate.

Novel pathways of bile acid metabolism involving CYP3A4

The hepatic predominating cytochrome P450, CYP3A4, plays an essential role in the detoxification of bile acids and is important in pathological conditions such as cholestasis where CYP3A4 is adaptively up-regulated. However, the mechanism that triggers the up-regulation of CYP3A4 is still not clear. In this study, using recombinant CYP3A4 and human liver microsomes, we demonstrate that CYP3A4 can metabolise lithocholic acid into 3-dehydrolithocholic acid, a potent activator of the nuclear receptors, pregnane X receptor and 1,25-dihydroxy vitamin D3 receptor, which are known to regulate the expression of CYP3A4. This process thus provides a feed-forward metabolism of toxic bile acid that may be of importance in maintaining bile acid homeostasis. We also provide evidence for a novel CYP3A4-mediated metabolic pathway of the secondary bile acid deoxycholic acid. Patients treated with the antiepileptic drug carbamazepine, a CYP3A4 inducer, had markedly elevated urinary excretion of 1beta-hydroxydeoxycholic acid compared to healthy controls. The importance of CYP3A4 in this process was verified by incubations with recombinant CYP3A4 and human liver microsomes, both of which efficiently converted deoxycholic acid into 1beta-hydroxydeoxycholic acid. Interestingly, CYP3A4 was also found to be active against the secondary bile acid ursodeoxycholic acid.

17O nuclear magnetic resonance spectra of steroids

The 17O NMR spectra of cholesterol and 31 other steroid alcohols, esters, ketones, and acids enriched by synthesis with 17O from H2(17)O have been observed under ordinary operating conditions, and correlations between 17O chemical shift and structure have been adduced. Spectra-structure correlations for these steroids are in conformity with those previously adduced with simpler compounds by others.

Peroxidation in position C-6 of progesterone-3-ethanolimine is increased by the presence of enzymes generating oxygen radicals

The generation of 6-oxygenated (6 beta-hydroxy, 6 beta-hydroperoxy, and 6-oxo) progesterone derivatives during the hydrolysis of progesterone-3-ethanolimine has been shown to be increased in the presence of xanthine/xanthine oxidase. The combination of xanthine/xanthine oxidase with other enzymes and/or reagents that catalyze transformation (or formation) of oxygen radicals suggested that the most likely oxygen species participating in the 6-oxygenation was the protonated acid of the superoxide anion, i.e., the hydroperoxy radical. The suggestion was further supported by experiments with oxygen scavengers. However, the data presented do not rule out a radical propagation reaction since the steroid compound used may be more reactive than the scavengers tested. A stimulation of 6-oxygenation of progesterone-3-ethanolimine by NADPH-supplemented rat liver microsomes was found. This reaction was inhibited by the only oxygen scavenger (reduced glutathione) found to be effective in the xanthine/xanthine oxidase experiments. The similarities between the two oxygenation systems may implicate a mechanism for 6 beta-hydroperoxidation of 3-oxo-4-ene steroids in rat liver microsomes.

Limits for the use of (18O)cholesterol and (18O)sitosterol in studies of cholesterol metabolism in humans

The present study was undertaken in order to determine whether 18O-labeled sterols could be used in place of 14C-sterols in clinical studies of cholesterol metabolism. (3 beta-18OH)Cholesterol and (3 beta-18OH)sitosterol were simply and inexpensively synthesized and precisely and accurately quantified by gas chromatography/mass spectrometry. 18O-Sterols added to fecal homogenate and saponified were completely recovered. However, in a series of validation studies in humans, the fecal recoveries of orally administered (18O)cholesterol and (18O)sitosterol were significantly lower than the recoveries of 14C-sterols given simultaneously. We found that the losses were largely limited to the coprostanol and ethylcoprostanol fecal metabolites. In vitro fecal incubations of 18O-sterols and unlabeled water or of unlabeled sterols with H2(18)O indicated that the losses occurred during fecal bacterial metabolism and were likely due to 3 beta-oxygen exchange with the oxygen of water, possibly via a 3-ketosteroid intermediate. These data indicate that (18O)cholesterol and (18O)sitosterol are invalid tracers for the measurement of human cholesterol metabolism by methods based on fecal sterol recovery.

Studies on rat liver microsomal steroid metabolism using 18O-labelled testosterone and progesterone

In order to investigate the possible involvement of oxygen functions in the rat liver microsomal metabolism of progesterone and testosterone these steroids were specifically labelled with 18O in their oxo-functions and incubated with NADPH supplemented 105,000 g sediments. Gas chromatography-mass spectrometry was used to identify the metabolites formed as well as to quantitate the losses of 18O-label. With 18O-labelled testosterone as substrate two of the major monohydroxylated metabolites, i.e. 2 beta- and 6 beta-hydroxytestosterone were shown to have lost about 25 and 50% of their 18O respectively. A complete retention of label was found in 7 alpha- and 16 alpha-hydroxytestosterone. None of the monohydroxylated progesterone metabolites, i.e. the 2 alpha-, 6 beta- and 16 alpha-hydroxyprogesterone had lost any 18O following incubation with 3,20-18O-labelled progesterone. Control incubation (30', 37 degrees C) with buffer and 18O-labelled progesterone and testosterone revealed no exchange of 18O. Thus the partial loss of 3-18O-label during 2 beta- and 6 beta-hydroxylation of testosterone may indicate a covalent interaction between the steroid 3-oxo-group and one or more cytochrome P-450 species in the rat liver microsomes. In view of the potentiating effect of a 3-imine group in spontaneous 6 beta-hydroxylation the present in vitro data suggest that a steroid protein-interaction may occur via a 3-imine group during 6 beta-hydroxylation of testosterone in rat liver microsomes. Analysis of 5 alpha-reduced metabolites of both progesterone and testosterone showed significant losses of 3-18O, but due to the ease with which 3-oxo-5 alpha-steroids exchange their 3-18O with aqueous media an enzymatically induced loss of 3-18O could not be safely established. The 20-oxido-reductase which converted progesterone did not induce a loss of 20- or 3-18O thus indicating that the oxofunctions were not covalently engaged in the enzymatic binding of the steroid.

Mechanism of inactivation of 3-oxosteroid delta 5-isomerase by 17 beta-oxiranes

The affinity label (17S)-spiro[estra-1,3,5(10),6,8-pentaene-17,2'-oxiran]-3-ol (5 beta) inactivates 3-oxosteroid delta 5-isomerase from Pseudomonas testosteroni by formation of a covalent bond between Asp-38 of the enzyme and the steroid. High-performance liquid chromatography (HPLC) analysis of tryptic digests of inactivated enzyme shows that two isomeric steroid-containing peptides are formed in a ratio of 9:1 at pH 7 (TPS1 and TPS2). Hydrolysis of each of these peptides produces a different steroid: TPS1 releases 17 alpha-(hydroxymethyl)estra-1,3,5(10),6,8-pentaene-3,17 beta-diol (S1) whereas TPS2 yields 17 beta-(hydroxymethyl)estra-1,3,5(10),6,8-pentaene-3,17 alpha-diol (S2). Inactivation of the enzyme by (17S)-spiro[estra-1,3,5(10),6,8-pentaene-17,2'-oxiran-18O]-3-ol, followed by mass spectral analysis of the diacetate of the steroid released upon hydrolysis of the enzyme-inhibitor bond, reveals that TPS1 is formed by attack of Asp-38 at the methylene carbon of the oxirane. In contrast, TPS2 is produced by Asp-38 attack at the tertiary carbon. These results imply that inactivation occurs through concurrent SN1 and SN2 reactions of Asp-38 with the protonated inhibitor and that Asp-38 is located on the alpha face of the steroid when it is bound to the active site in the correct manner to react for both the SN1 and SN2 processes.

Formation of imine derivatives between tris(hydroxymethyl)-aminomethane and saturated 3-oxo steroids during conventional work up of biological samples dissolved in Tris-HCl buffers

Extraction with organic solvents of Tris-HCl buffer solutions containing microgram amounts of saturated 3-oxo-steroids, followed by evaporation of the organic solvent, produce a condensation product identified as the Tris-3-imino-derivative of the steroid. The yields range between 20-40%. The conditions governing the formation of the derivatives have been explored.

The electron impact induced cleavage of the C-17--C-20 bond and D-ring in trimethylsilyl derivatives of C21 steroids. Reciprocal exchange of trimethylsilyl groups

The cleavage of the C-17--C-20 bond in the trimethylsilyl derivatives of C21 steroids has been investigated with the aid of isotopic labelling. A significant amount of reciprocal exchange was observed between trimethylsilyl groups or hydrogen atoms on the 20-O- and 17-O-positions. The loss of positional identity of these groups was found to influence the specificity of fragmentation reactions associated with the 17 beta-sidechain and the D-ring. It is supposed that the interchange of these silyl groups is facilitated by donation of the non-bonding electrons on oxygen to the empty d orbitals of silicon.

Trimethylsilyl group migration during electron impact and chemical ionization mass spectrometry of the trimethylsilyl ethers of 20-hydroxy-5alpha-pregnan-3-ones and 20-hydroxy-4-pregnen-3-ones

The electron impact and chemical ionization (isobutane) mass spectra of the TMS derivatives of 20-hydroxy-5alpha-pregnan-3-ones and 20-hydroxy-4-pregnen-3-ones include ions [M--44]+-attributable to loss of a CH3CHO fragment from C-17 with migration of the TMS group to the charge-retaining fragment. Mass spectra of isotopically labelled ([3-18O], [20-18O] and [2H9-TMS]) analogues are consistent with this mechanism; [2H9-TMS] labelling further indicates that subsequent loss of a methyl group from [M--44]+-ions does not involve the TMS group. Corresponding ions are not observed at significant abundance in the spectra of 20beta-trimethylsilyloxy-5alpha-pregnane and 20beta-trimethylsilyloxy-4-pregnen-3-one 3-O-methyloxime. In the electron impact mass spectrum of the t-butyldimethylsilyl ether of 20beta-hydroxy-4-pregnen-3-one, an ion is observed corresponding to loss of 44 atomic mass units from the intense [M--C4H9]+ ion.

Is a Schiff base involved in the mechanism of the delta4-3-oxo steroid 5alpha- or 5beta-reductases from mammalian liver?

The saturation of the delta4-double bond of delta4-3-oxo steroids by the mammalian 5alpha- or 5beta-reductases is by an enzyme mechanism that does not involve a Schiff-base intermediate.