Characterization of chemically modified steroids for doping control purposes by electrospray ionization tandem mass spectrometry

Article abstract of DOI:10.1002/jms.820

The discovery of the designer steroid tetrahydrogestrinone (THG) in elite athletes' doping control samples in 2003 demonstrated the availability of steroid derivatives prepared solely for doping purposes. Modern mass spectrometers utilizing electrospray ionization and collisionally activated dissociation (CAD) of analytes allow the structural characterization of steroids and their derivatization sites by the elucidation of fragmentation behaviors. A total of 21 steroids comprising either a 4,9,11-triene, a 3-keto-4-ene or a 3-keto-1-ene nucleus were investigated regarding their dissociation pathways, deuterated analogues were synthesized and fragmentation routes were postulated, permitting the identification of steroidal structures and modifications. Compounds based on a 4,9,11-triene steroid with an ethyl residue at C-13 (gestrinone analogues) generate abundant fragment ions at m/z 241 and 199, whereas the substitution of the C-13 ethyl group by a methyl residue (trenbolone analogues) results in a shift of m/z 241 to 227. Substances related to testosterone with a 3-keto-4-ene structure give rise to abundant fragment ions at m/z 109 and 97 whereas steroids with a 3-keto-1-ene nucleus eliminate the C-ring including the carbons C-1-C-4, in addition to C-19 that is proposed to migrate from C-10 to C-1 under CAD conditions. Copyright

Full text of DOI:10.1002/jms.820

JOURNAL OF MASS SPECTROMETRY
J. Mass Spectrom. 2005; 40: 494–502
Published online 15 February 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.820
Characterization of chemically modified steroids for
doping control purposes by electrospray ionization
tandem mass spectrometry
2
1
1
Mario Thevis,1∗ Ute Bommerich, Georg Opfermann and Wilhelm Schanzer
¨
1
Institute of Biochemistry, German Sport University Cologne, Carl-Diem Weg 6, 50933 Cologne, Germany
Institute for Physical and Theoretical Chemistry, University of Bonn, Wegelerstrasse 12, 53115 Bonn, Germany
2
Received 19 October 2004; Accepted 8 December 2004
The discovery of the designer steroid tetrahydrogestrinone (THG) in elite athletes’ doping control samples
in 2003 demonstrated the availability of steroid derivatives prepared solely for doping purposes. Modern
mass spectrometers utilizing electrospray ionization and collisionally activated dissociation (CAD) of
analytes allow the structural characterization of steroids and their derivatization sites by the elucidation
of fragmentation behaviors. A total of 21 steroids comprising either a 4,9,11-triene, a 3-keto-4-ene or a
3-keto-1-ene nucleus were investigated regarding their dissociation pathways, deuterated analogues were
synthesized and fragmentation routes were postulated, permitting the identification of steroidal structures
and modifications. Compounds based on a 4,9,11-triene steroid with an ethyl residue at C-13 (gestrinone
analogues) generate abundant fragment ions at m/z 241 and 199, whereas the substitution of the C-13 ethyl
group by a methyl residue (trenbolone analogues) results in a shift of m/z 241 to 227. Substances related to
testosterone with a 3-keto-4-ene structure give rise to abundant fragment ions at m/z 109 and 97 whereas
steroids with a 3-keto-1-ene nucleus eliminate the A-ring including the carbons C-1–C-4, in addition to
C-19 that is proposed to migrate from C-10 to C-1 under CAD conditions. Copyright  2005 John Wiley &
Sons, Ltd.
KEYWORDS: designer steroids; doping; electrospray ionization; mass spectrometry; fragmentation
2003, the discovery of another steroid named tetrahy-
drogestrinone (THG, 13-ethyl-17-hydroxy-18,19-dinor-17˛-
pregn-4,9,11-trien-3-one, Scheme 1, 3) triggered an even
greater avalanche in the world of sport4–7 as several elite ath-
letes were convicted of its abuse and sanctioned by WADA.
THG was presumably prepared from gestrinone (Scheme 1,
1), a drug utilized for the treatment of endometriosis,2 by
catalytic hydrogenation of the 17-ethinyl function, and its
effects on human or animal physiology have never been clin-
ically tested. As a consequence of gestrinone hydrogenation,
conventional screening procedures of doping control labora-
tories did not register the presence of the resulting prohibited
compound, because most analytical assays are based on tar-
get analysis, i.e. the determination of known compounds or
their metabolites.8–12 Mass spectrometry has proven mani-
fold to be a powerful tool for clinical, forensic and doping
control purposes, and it has been employed routinely for
decades for the detection and identification of drugs in var-
ious biological matrices. In order to understand common
and individual mass spectrometric properties of analytes of
interest, numerous studies on many classes of substances
have been performed in the past, in particular on steroids
after electron ionization by Djerassi and co-workers.13–17
To improve the comprehensiveness of doping controls, the
enormous potential of modern mass spectrometric instru-
ments have to be exploited, in particular of those analyzers
INTRODUCTION
Ever since anabolic androgenic steroids were prohibited by
international sports federations such as the International
Olympic Committee (IOC)1 and the World Anti-Doping
Agency (WADA), representatives of this class of drugs
have frequently been detected in urine specimens from
elite athletes. Doping offenses with remedies developed to
treat emaciation or debilitating diseases2 were revealed by
the determination of the administered drug, its metabolites
and/or their carbon isotope ratios, in particular by means
of mass spectrometry (MS) following chromatographic sepa-
ration by gas chromatography (GC) or liquid chromatog-
raphy (LC). In 2002, an anabolic agent termed norbo-
lethone (13-ethyl-17-hydroxy-18,19-dinor-17˛-pregn-4-en-3-
one, Scheme 1, 18) was detected in a doping control urine
sample3 although this compound was never marketed by
pharmaceutical companies, indicating the willingness of
some athletes to employ chemicals without medical approval
in order to promote muscle growth and strength. In October
ŁCorrespondence to: Mario Thevis, Institute of Biochemistry,
German Sport University Cologne, Carl-Diem Weg 6, 50933
Cologne, Germany. E-mail: m.thevis@biochem.dshs-koeln.de
Contract/grant sponsor: Manfred-Donike Society.
Copyright  2005 John Wiley & Sons, Ltd.

Characterization of chemically modified steroids by ESI-MS/MS 495
Scheme 1. Structure formulae of steroids investigated by means of ESI and CAD: gestrinone (1), dihydrogestrinone (DHG,
2), tetrahydrogestrinone (THG, 3), d4-THG (4), altrenogest (5), propyltrenbolone (6), trenbolone (7), 17-ethyltestosterone
(8), d4-17-ethyltestosterone (9), ethisterone (10), methyltestosterone (11), testosterone (12), 1˛,2˛-dideuterotestosterone (13),
19-trideuterotestosterone (14), 19-nortestosterone (15), norethisterone (16), norgestrel (17), norbolethone (18), 5˛-androst-1-en-17ˇ-
ol-3-one (1-testosterone, 19), metenolone (20) and 5˛-androst-1-en-3,17-dione (1-androstenedione, 21).
equipped with atmospheric pressure interfaces allowing the
detection and characterization of modified or derivatized
steroidal agents. Hence a variety of known steroids, stable
labeled counterparts and synthesized analogues (Scheme 1)
were investigated in the present study. Common and indi-
vidual fragmentation routes resulting from collisionally
activated dissociation (CAD) after electrospray ionization
(ESI) were elucidated for steroids with common nucleic struc-
tures, providing information necessary for the identification
of alterations and their influences on the mass spectrometric
behaviors of analytes relevant for doping control analyses.
Merck (Darmstadt, Germany). Benzene (dried) was obtained
from Aldrich (Deisendorf, Germany). tert-Butyl methyl ether
(TBME), (distilled before use) and ethyl acetate were bought
from Kraemer & Martin (St. Augustin, Germany).
Steroids
5˛-Androst-1-en-17ˇ-ol-3-one, gestrinone and altrenogest
were obtained from Thinker Chemical (Hangzhou, China).
Trenbolone, norgestrel, ethisterone, norethisterone, testos-
terone, nortestosterone, metenolone and methyltestosterone
were purchased from Sigma-Aldrich. 19-d3-Testosterone
was
a generous gift from Dr R. Kazlauskas of the
Australian Drug and Sport Testing Laboratory (ADSTL).
5˛-Androst-1-en-3,17-dione was obtained from Steraloids
(Newport, RI, USA). Ethyltestosterone, d4-ethyltestosterone,
norbolethone, dihydrogestrinone, tetrahydrogestrinone, d4-
tetrahydrogestrinone, propyltrenbolone and 1˛,2˛-dideute-
rotestosterone were prepared in our laboratory.
EXPERIMENTAL
Chemicals
Platinum(IV) oxide hydrate was obtained from Fluka (Buchs,
Switzerland). Palladium on activated carbon (10% Pd)
and ethylene glycol (99%) were purchased from Sigma-
Aldrich (Steinheim, Germany). Hydrogen and deuterium
were bought from Linde (Cologne, Germany). Toluene-
4-sulfonic acid (p.a.), glacial acetic acid (p.a.), n-pentane
(p.a.), and silica gel 60 (70–230 mesh) were purchased from
Selective hydrogenation and H/D exchange of
steroids
Various derivatives of commercially available compounds
were prepared by commonly employed derivatization and
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502

496
M. Thevis et al.
hydrogenation procedures,18 which are described in brief in
the following.
flash chromatography on silica gel (n-pentane–ethyl acetate
(4 : 1, v/v)), yielding 24 mg (12% of theory) of 8 (purity >95%)
and 30 mg (15% of theory) of 9 (purity >95%).
Hydrogenation of gestrinone (1)—preparation of DHG
(2), THG (3) and d4-THG (4)
Hydrogenation of norgestrel (17)—preparation of
norbolethone (18)
In order to prepare THG (3) or d4-THG (4), 500 mg of 1
(1.6 mmol) were dissolved in 250 ml of methanol in a round-
bottomed flask, 10 mg of PtO2 were added and hydrogen
or deuterium, respectively, was flushed into the reaction
mixture by means of a glass capillary. After 25 min, the
reaction mixture was allowed to stand at room temperature
for 30 min for sedimentation of the catalyst, the methanolic
layer was transferred to a new flask, evaporated to dryness,
and the resulting product was purified from residues of
PtO2 by the addition of 100 ml of deionized water and
liquid–liquid extraction into TBME. After evaporation of
TBME, pure THG and its d4-labeled analogues were obtained
at purities of >97% and yields of 78 and 81% of theory,
respectively.
Compound 2 was prepared from 1 in analogous manner,
except that PtO2 was replaced with palladium on charcoal
and 300 µl of pyridine were added to the reaction mixture
prior to hydrogenation in order to limit reduction potency.18
Here, the synthesis yielded 75% of 2 that was further puri-
fied by HPLC fractionation. By means of a Merck (Darmstadt,
Germany) semi-preparative reversed-phase column (LiChro-
spher, 250 ð 10 mm i.d., particle size 10 µm), 2 was separated
from 3 utilizing a gradient composed of (A) deionized water
and (B) acetonitrile (97% A to 0% A in 11 min, 0% A until
18 min, re-equilibration at 97% A for 5 min), and the respec-
tive fractions were collected and evaporated to dryness,
yielding 2 with a purity of >97%.
Norbolethone (18) was obtained by hydrogenation of
200 mg (0.65 mmol) of 17 (dissolved in methanol) uti-
lizing 20 mg of (1,4-bis(diphenylphosphino)butane)(1,5-
cyclooctadiene)rhodium(I)BF4 as catalyst. Here, PtO2 or Pd
on activated carbon were not an option as their selectivity was
not sufficient, resulting in additional hydrogenation of the
steroidal A-ring. After complete reduction of the 17-ethinyl
residue of 17, the methanolic solution was evaporated to
dryness, 100 ml of deionized water were added and the tar-
get compound was extracted into 300 ml of TBME, yielding
173 mg (83% of theory) of 18 (purity >97%).
H/D exchange of 5˛-androst-1-en-17ˇ-ol-3-one (19) and
metenolone (20)
The exchange of labile hydrogens by deuterium atoms was
accomplished by incubation of 5 mg of each steroid in 5 ml
of a mixture of CH3OD, D2O and 40% NaOD in D2O
°
(4 : 0.95 : 0.05, v/v/v) for 12 h at 60 C. The hydrogens located
at C-1 and C-4 or C-2, C-4 and C-20 were substituted in 19 or
20, respectively. The efficiency of deuterium exchange was
tested by ESI-MS, and the results were as follows: 19 d1 D 4%,
d2 D 100%, d3 D 88%; 20 d5 D 18%, d6 D 100%, d7 D 64%.
Purity
The purity of the synthesized material was determined by
HPLC/UV analysis performed on a Hewlett-Packard (Wald-
bronn, Germany) 1090 Series II HPLC system, equipped with
a Nucleosil 120-5 C18 column (Macherey–Nagel, Du¨ren, Ger-
many) of 120 ð 4 mm i.d. The solvents used were (A) distilled
water and (B) acetonitrile. A gradient was employed starting
with 65% A decreasing to 30% A in 14 min at a flow-rate
of 1 ml minꢀ1. Wavelengths of 208, 246 and 350 nm and
also full UV spectra were utilized for the detection of target
compounds and contaminants.
Hydrogenation of altrenogest (5)—preparation of
propyltrenbolone (6)
Propyltrenbolone was obtained by hydrogenation of 5 in
accordance to the protocol described above at a purity of
>95% and yield of 81%.
Synthesis of ethyltestosterone (8) and d4-ethyltestosterone
(9)
NMR spectroscopy
The testosterone analogues 8 and 9 were synthesized from
ethisterone (10). An amount of 400 mg (1.28 mmol) of 10 was
dissolved in 60 ml of benzene containing 30 mg of toluene-
4-sulfonic acid and 5 ml of ethylene glycol. The reaction
mixture was refluxed with water separation for 16 h, 100 ml
of deionized water were added and ethisterone-3-ethylene
ketal was recovered by liquid–liquid extraction into 300 ml
of TBME. The organic layer was evaporated to dryness and
the crude product was stored in a desiccator over phos-
phorus pentoxide for 12 h. Hydrogenation or deuteration of
ethisterone-3-ethylene ketal was performed according to the
procedure described above, and subsequently the ketal func-
tion was removed by acidic hydrolysis. For this purpose, the
intermediate products were dissolved in 50 ml of acetic acid
(2 M) and refluxed for 1 h. A volume of 50 ml of deionized
water was added, and the desired compounds were extracted
into 300 ml of TBME. The organic layer was washed twice
with 100 ml of aqueous potassium hydroxide (1 M) and evap-
orated to dryness. The resulting products were purified by
Structure confirmation of the synthesized products was
1
obtained by H NMR spectroscopy performed on a Bruker
(Bremen, Germany) Avance DPX 200 instrument. Amounts
of 5 mg of each compound were dissolved in d6-acetone and
spectra were recorded at room temperature.
Mass spectrometry
Commercially obtained and synthesized steroids were dis-
solved in a mixture of 0.1% acetic acid and acetonitrile
(1 : 1, v/v) at concentrations of 5 µg mlꢀ1 and analyzed by
ESI and CAD on an Applied Biosystems (Darmstadt, Ger-
many) QTrap instrument at room temperature. Nitrogen
was employed as curtain and collision gas (5.33 ð 10ꢀ3 Pa)
delivered from a Whatman K75-72 nitrogen generator, and
samples were introduced into the mass spectrometer by
means of a syringe pump at a flow-rate of 5 µl minꢀ1. Declus-
tering potentials were optimized for respective protonated
molecules, and collision energies (CE) were adjusted to allow
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502

Characterization of chemically modified steroids by ESI-MS/MS 497
241.2
efficient fragmentation of analytes retaining the precursor ion
at a relative abundance of ¾10% in product ion spectra. Unit
resolution was employed for mass selection in Q1, and the
100%
OH
CH
90%
CH
2
3
80%
70%
60%
50%
40%
30%
20%
10%
linear ion trap was operated at a scan speed of 1000 u sꢀ1
.
With MS3 experiments, the first precursor ion was
isolated in Q1, dissociated in Q2 and the resulting fragment
ion serving as second precursor ion was isolated in the
linear ion trap of the mass spectrometer and dissociated with
different excitation energies providing respective product
ion spectra.
O
199.2
237.1
159.1
197.1
213.1
266.1
183.1
169.2
157.1
211.1
225.1
313.2
295.2
284.1
251.2
255.1
135.1
131.1
119.1
105.1
141.1
265.2
91.1
Aliquots from H/D exchange experiments were dis-
solved only in acetonitrile and introduced into the mass
spectrometer according to the conditions described above.
60 80 100 120 140 160 180 200 220 240 260 280 300 320
m/z
(a)
109.1
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
RESULTS AND DISCUSSION
OH
CH
CH
2
3
NMR spectroscopy
The structural modifications of all synthesized compounds
97.1
1
(2–4, 6, 8, 9, 18) were confirmed by H NMR spectroscopy
O
and correspond to data for the structurally related reference
compound 1 and also published data.7,19
123.1
299.3
227.3
281.3
173.2
159.1
203.3
121.1
133.1
185.2
215.3
225.2
239.3
252.2
241.2
161.2
157.2
317.3
Mass spectrometry
93.1
79.1
284.2
67.1
ESI product ion experiments were performed on all analytes,
and in Fig. 1 the resulting tandem mass spectra of 3, 8
and 19 are shown exemplarily. All major common and
individual fragment ions of the compounds analyzed are
listed in Table 1.
60 80 100 120 140 160 180 200 220 240 260 280 300 320
(b)
m/z
187.2
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
OH
Steroids with 4,9,11-triene nucleus
This class of steroids includes compounds 1–7 in Scheme 1.
The principal fragmentation behavior of representatives will
be discussed by means of the product ion spectrum of
substance 3 (Fig. 1(a)). The protonated molecule [M C H]C at
m/z 313 gives rise to a variety of abundant fragment ions
upon CAD. The neutral loss of water (ꢀ18 u), presumably
originating from C-17, generates the product ion at m/z
295. Here, no hydrogen from the ethyl residue linked
to C-17 is involved in the elimination process as the
deuterated analogue 4 also releases 18 u, proving the
origin of the hydrogen necessary for the neutral loss of
water from a position remote of the ethyl group at C-
17. Several mass spectrometric studies demonstrated that
eliminations of water upon electron ionization (EI) from
both steroidal and other cyclic organic molecules include
hydrogens originating from various positions with particular
regard to stereochemical properties. Primarily 1,3- and
1,4-eliminations were observed, and losses of hydrogens
adjacent to the hydroxyl function were excluded.16,20–22
The fragment at m/z 295 (Fig. 1(a)) subsequently releases
an ethyl radical, generating the ion at m/z 266. This is an
exception to the commonly accepted ‘even-electron rule’,23
several of which have been reported in the literature
recently.24–27 All gestrinone analogues (compounds 1–4)
expel 29 u from the fragment ion obtained from the respective
protonated molecules after elimination of water (Table 1).
Hence the primary origin of the ethyl radical is proposed
to be C-13 that is located in an allylic position with respect
to a large conjugated 8-ꢀ-electron system, explaining the
O
131.1
145.2
159.2
105.2
117.2
133.2
289.3
161.3
91.1
205.2
157.2
213.2
227.3
253.3
245.3
173.3
271.3
69.1
300
60 80 100 120 140 160 180 200 220 240 260 280
m/z
(c)
Figure 1. ESI product ion spectra of (a) tetrahydrogestrinone
(Mr D 313 u, CE D 40 eV), (b) 17-ethyltestosterone
(Mr D 317 u, CE D 40 eV) and
(c) 5˛-androst-1-en-17ˇ-ol-3-one (Mr D 289, CE D 40 eV).
stable character of the generated cation. Compounds 3 and
4 contain additional ethyl side-chains at C-17 amenable for
elimination, and the product ion spectrum of 4 (Table 1)
provides information that proves the generation of m/z 266
by release of an ethyl residue originating either from C-13 or
C-17. The introduction of four deuterium atoms into the C-17
side-chain allows their distinction as both 29 and 33 u are
expelled from the ion at m/z 299 giving rise to the product
ions at m/z 270 and 266, respectively, in a ratio of ¾2.5 : 1.
In contrast to the fragments generated by the losses
of water and ethyl residues, ions at m/z 241 and 199 are
present in all gestrinone analogues (compounds 1–4, Table 1)
requiring a structure lacking the C-17 substituents. Hence a
neutral loss of C-16 and C-17 including their substituents
is postulated, giving rise to m/z 241 as demonstrated in
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502

498
M. Thevis et al.
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502

Characterization of chemically modified steroids by ESI-MS/MS 499
in the product ion spectra of the 5–7 (Table 1). The ion at
m/z 227 eliminates 28 u (ethene) upon CAD in a comparable
way as suggested for the propene elimination from m/z
241. Hence a distinction between substances related to either
trenbolone or gestrinone (i.e. with a methyl or ethyl residue
at C-13) can be accomplished by the determination of m/z
227 or 241, a valuable parameter for the identification of
unknown designer steroids or their metabolites. An example
is given, for instance, with THG (3) and propyltrenbolone (6),
both of which consist of C21H28O2 and generate a protonated
molecule at m/z 313. Product ion spectra of both compounds
contain signals at [M C H]C ꢀ 18 and m/z 199, but 3 gives
rise to m/z 241 whereas 6 generates a base peak at m/z
227.
Steroids with 3-keto-4-ene nucleus
Testosterone and 19-nortestosterone analogues comprise 3-
keto-4-ene nuclei and demonstrate a distinct difference in
dissociation behavior compared with the class of 4,9,11-
triene steroids. Product ion spectra of compounds related
to testosterone (8–12) contain abundant fragments at m/z
97 and 109 (Table 1), the proposed generation of which was
described for testosterone by Williams et al.28 This partic-
ular dissociation pathway is valid also for 17-substituted
testosterone analogues such as 8–11, which yield abundant
fragments originating from A- and B-ring cleavages. How-
ever, to a smaller extent, also fragment ions generated by
D-ring dissociation are found in product ion spectra of 8–11.
Ethisterone (10) bears an ethinyl residue at C-17, and its
protonated molecule eliminates 68 u to give m/z 245 in
accordance with the dissociation of gestrinone (1, Table 1
and Scheme 2(a)). Owing to four additional hydrogens in
the B- and C-rings, the fragment ion at m/z 241, as found
in the product ion spectra of 1–4, is shifted to m/z 245. The
hydrogenation of the 17-ethinyl residue of 10 gives rise to 8,
which demonstrates comparable fragmentation routes upon
CAD. Here, the protonated molecule expels 72 u by means of
D-ring dissociation, giving rise to m/z 245, and an additional
fragment ion is detected at m/z 241, which must comprise the
17-alkyl residue as with the quadruply deuterated counter-
part 9 this particular fragment is incremented by 4 u, yielding
an ion at m/z 245 obtained from two different fragmentation
pathways. Moreover, 10 produces a fragment at m/z 237
substantiating the proposal of a fragmentation remote from
the D-ring generating m/z 241 in the case of 8. In order to
determine the part of molecule released to generate m/z 241
of 8, testosterone (12), 1˛,2˛-dideuterotestosterone (13) and
19-trideuterotestosterone (14) were investigated regarding
analogous fragment ions. Compound 12 generates m/z 213
corresponding to m/z 241 of 8. The ion at m/z 213 of 12 is
incremented to m/z 214 in the case of 13, and m/z 216 is
observed in the product ion spectrum of 14, demonstrating
the presence of one and three deuterium atoms, respectively,
in the fragment ion of interest. Hence an elimination of water
from C-17 followed by the release of carbons C-2, C-3 and
C-4 as acetone (ꢀ58 u) is postulated, giving rise to m/z 213
and 241 in the case of 12 and 8, respectively. The origins of
the hydrogens that migrate to the leaving group allowing
the loss of acetone from [M C H]C ꢀ 18 were not assigned,
Scheme 2. (a) Proposed fragmentation pathway of tetrahy-
drogestrinone (3 in Scheme 1) generating fragment ions at
m/z 241 and 199; (b) proposed fragmentation pathway of
5˛-androst-1-en-17ˇ-ol-3-one generating the fragment ion at
m/z 187.
Scheme 2(a). The ion at m/z 241 is generated directly from
the protonated molecules as MS3 experiments, e.g. from
[M C H]C ꢀ H2O (i.e. m/z 291), did not yield the respective
fragment. The ion at m/z 199 results from a neutral loss
of 42 u from m/z 241, also confirmed by MS3 experiments,
where the protonated molecule (m/z 309) was dissociated,
and the resulting fragment ion at m/z 241 was isolated
and fragmented in the linear ion trap. Here, an elimination
of propene is suggested, including C-15, C-18 and C-19,
which requires a migration of the ethyl residue as shown in
Scheme 2(a). The linkage between C-13 and C-18 is proposed
to generate a double bond between C-13 and C-14, while C-
18 establishes a new bond to C-15. A neutral loss of propene
(ꢀ42 u) accompanied by a hydrogen shift from the leaving
group (C-15, C-18 and C-19) is postulated, giving rise to m/z
199. This fragment in particular is observed in all product ion
spectra of steroids with a 4,9,11-triene nucleus investigated
in this study, whereas compounds bearing a methyl residue
at C-13 instead of an ethyl group give rise to a fragment
at m/z 227 instead of m/z 241. These phenomena are in
accordance with the fragmentation pathway proposed for
the CAD of gestrinone analogues, where m/z 241 includes
the C-13-linked alkyl residue, and m/z 199 is generated by
removal of this particular part of the molecule. Exchanging
the ethyl function of gestrinone by a methyl group results in
a decrement of m/z 241 by 14 u, yielding m/z 227, as detected
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502

500
M. Thevis et al.
as various hydrogen transfer reactions during fragmentation
processes are reported in the literature, which are to be
elucidated by specific deuterium labeling.
molecule at m/z 287 was dissociated to a fragment at m/z
203 that subsequently expelled 18 u to give m/z 185. The
migration of methyl residues of steroids in solution under
various conditions is a well-known phenomenon,29–32 in
particular for 17-methyl steroids such as methyltestosterone
and methandienone,33 but it has never been reported for
C-19 of 5˛-androst-1-en-17ˇ-ol-3-one and analogues in the
gas phase. This migration can be explained according to a
Wagner–Meerwein rearrangement starting with a positive
charge at C-1 that triggers the shift of C-19 from C-10 to
C-1 followed by the release of the A-ring. Complementary
observations in acidified solutions have been described in the
literature,30 where the migration of C-19 to C-5 was reported,
initiated by the elimination of water including a hydroxyl
function at C-5, causing a positive charge at this particular
carbon. NMR studies confirmed the shift of C-19 to C-5 and
the corresponding final structure of the steroid.
Product ion spectra of compounds structurally related
to 19-nortestosterone (15) such as 16, 17 and 18 demon-
strate dissociation behaviors that are in accordance with
both gestrinone and testosterone analogues. The protonated
molecule of 17, which can also be considered tetrahydro-
genated gestrinone (in the B- and C-rings), eliminates 18 u
(H2O) and subsequently 29 u (ethyl radical), giving rise to
m/z 266 as detected also in the product ion spectrum of 3
(Table 1). In contrast to 3, 17 does not generate a base peak at
m/z 241 but at m/z 245 owing to four additional hydrogens
in the B- and C-rings and an expulsion of C-16 and C-17 of
the D-ring as described for 1–7. Hence the resulting product
ion composed of intact A-, B- and C-rings is incremented
by 4 u, allowing enabling the differentiation of the principle
steroidal structures, i.e. the 3-keto-4-ene and 4,9,11-triene
nucleus, of these closely related substances. The product
ion spectrum of 16 contains fragment ions corresponding to
those observed with 17. Here, the protonated molecule at
m/z 299 is decremented by 14 u compared with 17 owing to
a methyl residue at C-13 instead of an ethyl group. Hence
fragments resulting from neutral losses of water or D-ring
dissociation such as m/z 231 are also reduced by 14 u as the
steroid nucleus including the C-ring and its substituent at
C-13 are still present.
Doping control analysis
The knowledge of dissociation pathways of selected steroid
nuclei upon ESI and CAD allows the creation of screening
procedures complementary to established methods in dop-
ing control analysis. The possibility of utilizing precursor
ion scanning or neutral loss scanning allows the determina-
tion of classes of steroids and does not limit the analytical
assay to selected drugs, hence permitting also the detection
of artificially modified drugs such as THG or norbolethone.
Precursor ion scans of m/z 97 and 109 indicate steroids with
a 3-keto-4-ene structure (e.g. testosterone (12), ethyltestos-
terone (8)), while the presence of a signal at m/z 109 in
combination with the absence of m/z 97 can represent 19-
norsteroids with a 3-keto function such as nortestosterone
(15) and analogues. Precursor ion scans of m/z 241 indicate
structures comparable to gestrinone (1), signals generated by
m/z 227 represent trenbolone (7) nuclei and fragment ions
at m/z 187 represent the A/B-ring structure of an androst-3-
keto-1-ene steroid (19). With sufficient sensitivities of modern
triple-quadrupole mass spectrometers, doping control urine
samples can be analyzed in addition to commonly accepted
procedures by means of LC/ESI-MS/MS in precursor ion
scanning modes. In the case of suspicious signals in the
resulting urinary steroid profiles, subsequent experiments
are required either to substantiate or to negate the suspicion
regarding drugs of abuse and their putative metabolites.
A characteristic A- and B-ring dissociation of 19-
nortestosterone analogues is observed with the fragment ion
at m/z 109. This ion corresponds to m/z 123 of testosterone
analogues generated from A- and B-ring cleavages28 and
is detected in all spectra of the investigated compounds
related to 19-nortestosterone (16–18) regardless of C-13 or
C-17 substitution.
Steroids with 3-keto-1-ene nucleus
Steroids such as 5˛-androst-1-en-17ˇ-ol-3-one (19), metenol-
one (20) and 5˛-androst-1-en-3,17-dione (21) demonstrate a
different preference regarding their fragmentation behavior
upon CAD compared with those comprising a 3-keto-4-ene
structure. In product ion spectra of the respective protonated
molecules, fragment ions at m/z 187 (in the case of 19 and
20) and m/z 185 (in the case of 21) are generated (Table 1),
which are proposed to result from an elimination of the
carbons C-1–C-4 and C-19, followed by a loss of water,
as depicted in Scheme 2(b). Indications for the suggested
fragmentation route were obtained by the detection of m/z
187 in the product ion spectra of 19 and 20, proving the
release of C-1 (and its substituent) from the steroids. In
addition, H/D exchange of 20 at C-2, C-4 and C-20 did
not influence the fragment at m/z 187 (data not shown),
demonstrating the elimination of the entire steroidal A-ring
during the fragmentation process. The postulated generation
of m/z 187 requires the migration of C-19 from C-10 to C-1 of
the leaving group (Scheme 2(b)). The basis of this proposal
was MS3 experiments proving the generation of m/z 205
from 289 of 19 (ꢀ84 u) and the subsequent elimination of 18
u giving rise to m/z 187. In accordance with this, m/z 185 was
observed in MS3 experiments with 21, where the protonated
CONCLUSION
ESI and CAD-MS/MS of steroids with a 4,9,11-triene or a
3-keto-4-ene nucleus provide distinct information about the
principle structure of the analyte. The detection of abundant
product ions at m/z 241 and 199 or m/z 227 and 199
indicate the presence of a 4,9,11-triene steroid with a C-
13-linked ethyl or methyl group, respectively. In addition,
the neutral loss of 68 u from the protonated molecule is
generally observed in the case of a D-ring bearing a 17-
ethinyl and 17-hydroxyl function. In contrast, in product ion
spectra of testosterone analogues comprising a 3-keto-4-ene
structure, fragments generated by A- and B-ring cleavages
are predominant, in particular the ions at m/z 97 and
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502

Characterization of chemically modified steroids by ESI-MS/MS 501
liquid chromatography–electrospray ionization tandem mass
spectrometry. Steroids 2004; 69: 101.
109. With removal of the angular 19-methyl group giving
rise to molecules related to 19-nortestosterone, product ion
spectra are obtained that are not dominated by A- and B-ring
fragment ions but contain an ion at m/z 109 (corresponding to
m/z 123 of testosterone analogues) and fragments generated
by commonly observed D-ring fragmentation releasing C-16
and C-17 including their substituents. 5˛-Androst-1-en-17ˇ-
ol-3-one and its analogue metenolone eliminate primarily the
carbons C-1–C-4 including all substituents, in addition to C-
19 that migrates to the leaving group under CAD conditions
generating an abundant fragment at m/z 187. The removal
of the A-ring is charge-driven, presumably according to a
Wagner–Meerwein rearrangement, and was only observed
with testosterone analogues bearing a double bond between
C-1 and C-2. Based on this information, unknown substances
with steroidal structure can be classified and elucidated
regarding the principal structure and possible modifications.
Moreover, more comprehensive screening procedures can be
established focusing on common fragmentation pathways.
Here, features of modern triple-stage mass spectrometers
such as neutral loss or precursor ion scanning as employed
routinely in metabolism studies can provide a broader view
in urine analysis, supporting an improved analytical assay
at required sensitivities. Compounds with fragmentation
routes comparable to known steroids but with different
molecular masses or different chromatographic behavior
can be detected and probably identified by MSn analysis,
chemical derivatization, synthesis of the putative compound,
etc. Hence substances that do not belong to regular urinary
steroid profiles may be discovered and give indications for
the abuse of unknown drugs.
11. Deventer K, Delbeke FT. Validation of
a
screening
method for corticosteroids in doping analysis by liquid
chromatography/tandem mass spectrometry. Rapid Commun.
Mass Spectrom. 2003; 17: 2107.
12. Thevis M, Opfermann G, Scha¨nzer W. Liquid chromatogra-
phy/electrospray ionization tandem mass spectrometric screen-
ing and confirmation methods for ˇ2-agonists in human or
equine urine. J. Mass Spectrome. 2003; 38: 1197.
13. Powell H, Williams DH, Budzikiewicz H, Djerassi C. Mass
spectrometry in structural and stereochemical problems. XLVIII.
A study of the hydrogen transfer reactions accompanying
fragmentation processes of 1-keto steroids. Synthesis of
deuterated 5˛-androstan-1-ones. J. Am. Chem. Soc. 1964; 86:
2623.
14. Beugelmans R, Shapiro RH, Durham LJ, Williams DH,
Budzikiewicz H, Djerassi C. Mass spectrometry in structural
and stereochemical problems. LI. Mass spectral end enolization
studies on 7-keto-5˛-androstanes. J. Am. Chem. Soc. 1964; 86:
2832.
15. Shapiro RH, Williams DH, Budzikiewicz H, Djerassi C. Mass
spectrometry in structural and stereochemical problems. LIII.
Fragmentation and hydrogen transfer reactions of a typical 3-
keto steroid, 5˛-androstan-3-one. J. Am. Chem. Soc. 1964; 86:
2837.
16. Karliner J, Budzikiewicz H, Djerassi C. Mass spectrometry in
structural and stereochemical problems. XCI. The electron
impact induced elimination of water from 3-hydroxy steroids. J.
Org. Chem. 1966; 31: 710.
17. Budzikiewicz H, Ockels W. Massenspektrometrische Fragmen-
tierungsreaktionen—VIII. Charakteristische Bruchstu¨cke von
υ5-3-Hydroxysteroiden. Tetrahedron 1976; 32: 143.
18. Fried J, Edwards JA (eds). Organic Reactions in Steroid Chemistry.
Van Nostrand Reinhold: New York, 1972.
19. Thevis M, Opfermann G, Schmickler H, Scha¨nzer W. Mass
spectrometry of steroid glucuronide conjugates. II—Electron
impact fragmentation of 3-keto-4-en- and 3-keto-5alpha-steroid-
17-o-beta glucuronides and 5alpha-steroid-3alpha,17beta-diol 3-
and 17-glucuronides. J. Mass Spectrom. 2001; 36: 998.
20. Egger H, Spiteller G. Massenspektren und Stereochemie von
Hydroxyverbindungen, 1. Mitt.: Hydroxysteroide. Monatsh.
Chem. 1966; 97: 579.
21. Macdonald CG, Shannon JS, Sugowdz G. Studies in mass
spectrometry stereochemistry and elimination reactions of
hydroxy and acetoxy compounds. Tetrahedron Lett. 1963; 4:
807.
22. Meyerson S, Leitch LC. Organic ions in the gas phase. XIV. Loss
of water from primary alcohols under electron impact. J. Am.
Chem. Soc. 1964; 86: 2555.
23. McLafferty FW, Turecek F. Interpretation of Mass Spectra.
University Science Books: Mill Valley, CA, 1993.
24. Thevis M, Schmickler H, Scha¨nzer W. Mass spectrometric
behavior of thiazide-based diuretics after electrospray ionization
and collision-induced dissociation. Anal. Chem. 2002; 74:
3802.
25. Karni M, Mandelbaum A. The ‘even-electron-rule’. Org. Mass
Spectrom. 1980; 15: 53.
26. Justesen U. Collision-induced fragmentation of deprotonated
methoxylated flavonoids, obtained by electrospray ionization
mass spectrometry. J. Mass Spectrom. 2001; 36: 169.
Acknowledgement
The authors thank the Manfred-Donike Society, Cologne, for
financial support
REFERENCES
1. List of Prohibited Classes of Substances and Prohibited Methods of
Doping. International Olympic Committee: Lausanne, 2003.
2. Reynolds JEF (ed). Martindale—the Extra Pharmacopoeia (33rd
edn). Pharmaceutical Press: London, 2002.
3. Catlin DH, Ahrens BD, Kucherova Y. Detection of norbolethone,
an anabolic steroid never marketed, in athletes’ urine. Rapid
Commun. Mass Spectrom. 2002; 16: 1273.
4. Kondro W. Athlete’s ‘designer steroid’ leads to widening
scandal. Lancet 2003; 362: 1466.
5. Knight J. Drugs bust reveals athletes’ secret steroid. Nature 2003;
425: 752.
6. Knight J. Drugs in sport: no dope. Nature 2003; 426: 114.
7. Catlin DH, Sekera MH, Ahrens BD, Starcevic B, Chang Y-C,
Hatton CK. Tetrahydrogestrinone: discovery, synthesis, and
detection in urine. Rapid Commun. Mass Spectrom. 2004; 18:
1245.
8. Thieme D, Grosse J, Lang R, Mu¨ller RK, Wahl A. Screening,
confirmation and quantification of diuretics in urine
for doping control analysis by high-performance liquid
chromatography–atmospheric pressure ionisation tandem mass
spectrometry. J. Chromatogr. B 2001; 757: 49.
27. Qin X-Z. Tandem mass spectrum of
a farnesyltransferase
inhibitor—gas phase rearrangements involving imidazole. J.
Mass Spectrom. 2001; 36: 911.
28. Williams TM, Kind AJ, Houghton E, Hill DW. Electrospray
collision-induced dissociation of testosterone and testosterone
hydroxy analogs. J. Mass Spectrom. 1999; 34: 206.
9. Thevis M, Opfermann G, Scha¨nzer W. High speed determi-
nation of beta-receptor blocking agents in human urine by
liquid chromatography/tandem mass spectrometry. Biomed.
Chromatogr. 2001; 15: 393.
29. Wendler NL. In Molecular Rearrangements. Part 1, de Mayo P (ed).
Wiley–Interscience: New York, 1963; 1020.
10. Leinonen A, Kuuranne T, Kotiaho T, Kostiainen R. Screening of
free 17-alkyl-substituted anabolic steroids in human urine by
30. Segaloff A, Gabbard RB. Anti-androgenic activity of 17,17-
dimethyl-18-norandrost-13-enes. Steroids 1964; 4: 433.
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502

502
M. Thevis et al.
31. Cohen A, Cook JW, Hewett CL. The synthesis of compounds
related to sterols, bile acids, and oestrus-producing hormones.
Part VI. Experimental evidence of the complete structure of
oestrin, equilin, and equilenin. J. Chem. Soc. 1935; 445.
32. Johns WF. Retropinacol rearrangement of estradiol 3-methyl
ether. J. Org. Chem. 1961; 26: 4583.
33. Scha¨nzer W, Opfermann G, Donike M. 17-Epimerization of
17 alpha-methyl anabolic steroids in humans: metabolism and
synthesis of 17 alpha-hydroxy-17ˇ-methyl steroids. Steroids
1992; 57: 537.
Copyright  2005 John Wiley & Sons, Ltd.
J. Mass Spectrom. 2005; 40: 494–502