Enzyme-assisted synthesis and structure characterization of glucuronide conjugates of eleven anabolic steroid metabolites

Article abstract of DOI:10.1016/j.steroids.2007.10.008

Enzyme-assisted in vitro synthesis of eleven glucuronide-conjugated anabolic androgenic steroid (AAS) metabolites was performed using biphenyl-induced rat liver microsomal enzymes. The substrates within the study were the main compounds and metabolites detected in human urine after dosing of, e.g. metandienone, metenolone, methyltestosterone, nandrolone, and testosterone. Yields of glucuronidation reactions were 13-28% for most compounds, but significantly higher (77-78%) for the substrates with 4-ene-3-one double bond system of the steroid A-ring. Characterization of glucuronide-conjugated AAS structures was based on nuclear magnetic resonance spectroscopy (¹H NMR) and on liquid chromatographic-mass spectrometric (LC-MS) and tandem mass spectrometric (LC-MS/MS) analyses in positive and negative ion mode electrospray ionization (ESI). Only minor differences were observed in optimal synthesis conditions between various substrates, which offer a potential to apply this in vitro assay as a default method for glucuronidation of new AAS substrates. The method allowed for a rapid production pathway of stereochemically pure AAS glucuronides in milligram amount, such as needed, e.g. in the development of analytical methods in forensic or pharmaceutical sciences, as well as in doping control.

Full text of DOI:10.1016/j.steroids.2007.10.008

steroids 7 3 ( 2 0 0 8 ) 257–265
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/steroids
Enzyme-assisted synthesis and structure characterization
of glucuronide conjugates of eleven anabolic
steroid metabolites
Laura Hintikkaa, Tiia Kuuranneb,∗, Olli Aitioa, Mario Thevisc,
Wilhelm Scha¨nzerc, Risto Kostiainena
a
Division of Pharmaceutical Chemistry, Faculty of Pharmacy, P.O. Box 56, Viikinkaari 5E, FIN-00014 University of Helsinki, Finland
b
¨
¨ ¨
¨
United Laboratories Ltd., P.O. Box 222, Hoylaamotie 14, FIN-00380 Helsinki, Finland
c
Center for Preventive Doping Research – Institute of Biochemistry, German Sport University, Carl-Diem Weg 6, 50933 Cologne, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Enzyme-assisted in vitro synthesis of eleven glucuronide-conjugated anabolic androgenic
steroid (AAS) metabolites was performed using biphenyl-induced rat liver microsomal
enzymes. The substrates within the study were the main compounds and metabolites
detected in human urine after dosing of, e.g. metandienone, metenolone, methyltestos-
terone, nandrolone, and testosterone. Yields of glucuronidation reactions were 13–28% for
most compounds, but significantly higher (77–78%) for the substrates with 4-ene-3-one
double bond system of the steroid A-ring. Characterization of glucuronide-conjugated AAS
structures was based on nuclear magnetic resonance spectroscopy (1H NMR) and on liquid
chromatographic–mass spectrometric (LC–MS) and tandem mass spectrometric (LC–MS/MS)
analyses in positive and negative ion mode electrospray ionization (ESI). Only minor differ-
ences were observed in optimal synthesis conditions between various substrates, which
offer a potential to apply this in vitro assay as a default method for glucuronidation of new
AAS substrates. The method allowed for a rapid production pathway of stereochemically
pure AAS glucuronides in milligram amount, such as needed, e.g. in the development of
analytical methods in forensic or pharmaceutical sciences, as well as in doping control.
© 2007 Elsevier Inc. All rights reserved.
Received 3 January 2007
Received in revised form
12 October 2007
Accepted 21 October 2007
Published on line 26 October 2007
Keywords:
Anabolic steroids
Glucuronidation
Enzyme-assisted synthesis
LC–MS
1.
Introduction
potency also to cause serious health problems as side effects,
e.g. cardiovascular or liver diseases [2].
Anabolic-androgenic steroids (AAS) represent an important
class of abused drugs in sport, and are classified as pro-
hibited substances according to World Anti-Doping Agency
(WADA). The main desired effects of these testosterone-
derived compounds are their potential ability to improve
physical performance of skeletal muscle and to balance
catabolic condition in body after stress [1–2]. However, medi-
ated mainly by their androgenic activity, the AAS have the
Within the human body, the non-polar AAS are trans-
formed extensively by phase-I and phase-II metabolic reac-
tions. Enzymatically catalyzed phase-I reactions (oxidation,
reduction or hydroxylation) introduce more functionalities
into steroid structure, which increase the polarity and serve
as sites for phase-II reactions. Phase-II reactions (glucuronida-
tion, sulfation, methylation, acetylation and conjugation with
amino acids or glutathione) are conjugation reactions, which
∗
Corresponding author. Tel.: +358 9 5060 5441; fax: +358 9 5060 5420.
E-mail address: tiia.kuuranne@yhtyneetlaboratoriot.fi (T. Kuuranne).
0039-128X/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2007.10.008

258
steroids 7 3 ( 2 0 0 8 ) 257–265
Fig. 1 – Formation of 5␤-MTG in glucuronidation reaction.
terminate the activity of the compound and enhance elim-
ination. The most common pathways of AAS conjugation
in human body are glucuronidation and sulfation. Steroids
possessing a 3␤-hydroxyl structure are mostly excreted as sul-
fates, whereas steroids with 3␣- or 17␤-hydroxyl group favor
glucuronide-conjugation [3].
At present, the analytical methods for the detection of
glucuronide-conjugated AAS are based on gas chromato-
graphic (GC) separation and mass spectrometric (MS) detec-
tion of hydrolyzed and derivatized compounds. Although the
methods are robust and sensitive, sample preparation of the
indirect method is time-consuming and the GC–MS sample
throughput is relatively low [4]. Thus, the development of
more straightforward methods based on the direct analysis
of steroid conjugates is of great interest. Application of liquid
chromatographic (LC) separation interfaced by soft ionization
techniques, such as electrospray ionization (ESI), with tandem
mass spectrometry (MS/MS) offers an effective analytical tool
for the direct monitoring of AAS glucuronides. One significant
drawback of the conjugate analysis, however, is still the lack
of reference material, which is a necessity in the development
and application of analytical methods. In addition to doping
Table 1A – Nomenclature, precursors, and structures of the AAS glucuronides, as well as sources of the steroid aglycones
used as starting material
Abbreviation
Compound
Structure
Precursor
Source
5␣-NG
5␣-Estran-17-one-3␣-O-
Nandrolone
S
glucuronide
5␤-NG
5␤-Estran-17-one-3␣-O-
Nandrolone
S
glucuronide
5␤-EPIMG
5␣-MTG
17␤-Methyl-5␤-androst-1-
ene-17␣-ol-3␣-O-
glucuronide
Metandienone
DSHS
S
17␣-Methyl-5␣-androstane-
17␤-ol-3␣-O-glucuronide
Mestanolone
Methyltestosterone
Oxymetholone
S
5␤-MTG
17␣-Methyl-5␤-androstane-
17␤-ol-3␣-O-glucuronide
Metandienone
Metandriol
Methyltestosterone
5␣-MEG
1-Methylen-5␣-androstan-
17-one-3␣-O-glucuronide
Metenolone
DSHS
DSHS = Institute of Biochemistry, Cologne, Germany; S = Steraloids, Wilmington, USA.

steroids 7 3 ( 2 0 0 8 ) 257–265
259
control analytical purposes, the conjugated reference stan-
dards are also needed within steroid metabolism research in
pharmaceutical industry and research institutes.
misused by sportsmen [20]. The group of compounds repre-
sents also steroids with small differences in the substitution
of carbons 1, 3, 5, 10 and 17. The glucuronides were synthe-
sized by using enzyme-assisted synthesis method, which was
optimized in respective to substrate, UDPGA and protein con-
centration. The synthesized products were characterized by
nuclear magnetic resonance (NMR) spectroscopy and ESI tan-
dem mass spectrometry, and used as reference material for
the development of an LC–MS-based method suitable for UGT
enzyme kinetic studies.
In vivo production of reference substances in human runs
often into ethical as well as practical problems associated
with the isolation of pure metabolites from urine. Chemi-
cal synthesis methods for several metabolites of AAS have
been developed, both phase-I [5,6] and recently also phase-II
metabolites [7–9]. Some issues related to chemical synthesis,
e.g. formation of racemic reaction mixtures has been reported
earlier [10]. An alternative approach to chemical in vitro glu-
curonide synthesis is enzyme-assisted synthesis catalyzed
with microsomal uridine diphosphoglucuronosyl-transferase
(UGT) enzymes (Fig. 1). One of the main advantages of the
enzyme-assisted synthesis over the chemical synthesis is the
stereospecificity of the enzymes, which allows synthesis of
stereospecifically pure conjugates. Secondly, enzyme-assisted
synthesis offers a practical pathway for the rapid produc-
tion of small amounts of glucuronides, such as needed in the
build-up of an analytical method. Rat liver preparations as a
source of conjugating enzymes have been used for several
groups of substrates, e.g. p-nitrophenol [11], nitrogen con-
taining antidepressants [12], catechols [13,14] nitrocatechols
[15,16], androsterone, androstanediol, dihydrotestosterone
[17], epitestosterone [18], and testosterone [19].
2.
Methods
2.1.
Substrates and reagents
Nomenclature, precursors, and structures of the AAS glu-
curonides, as well as sources of the steroid aglycones of
this study are presented in Tables 1A and 1B. Uridine-
5ꢀ-diphosphoglucuronic acid (UDPGA) and d-saccharic acid
1,4-lactone were purchased from Sigma (MO, USA) and Aro-
clor 1254 (RCS-088/Analabs, Lot No K040) was from the
Foxboro Company (CT, USA). Acetonitrile (Rathburn, Scotland),
dichloromethane (Baker, The Netherlands), and methanol
(Baker) were of HPLC-grade. For the NMR studies methanol-
d4 (99.50%, Acros Organics, Belgium) and acetone-d6 (99.9%,
Aldrich, Germany) were used.
Liver microsomes were prepared from Aroclor 1254
induced (a single dose of 500 mg/4.5 mL olive oil/body
weight male Wistar rats (n = 5) at the department of
Industrial Hygiene and Toxicology (Finnish Institute of
Occupational Health, Helsinki, Finland) according to ear-
The aim of this study was to synthesize glucuronide conju-
gate standards of eleven AAS. The selection of the compounds
in this work represent the parent compounds and metabo-
lites which can be detected in human urine after dosing of
exogenous anabolic steroids, e.g. metandienone, metenolone,
methyltestosterone and nandrolone, as well as endogenous
testosterone (Tables 1A and 1B), which all are reported to be
Table 1B – Nomenclature, precursors, and structures of the AAS glucuronides, as well as sources of the steroid aglycones
used as starting material
Abbreviation
Compound
Structure
Precursor
Source
AG
5␣-Androstane-3␣-ol-17␤-O-glucuronide
Testosterone
DSHS
d3-TG
NG
[16,16,17-2H3]4-androsten-3-one-17␤-O-glucuronide
Estr-4-en-3-one-17␤-O-glucuronide
–
DSHS
Nandrolone
Metenolone
Methyltestosterone
Diosynth
DSHS
5␣-1-MEG
MTG
1-Methyl-5␣-androst-1-en-3-one-17␤-O-glucuronide
17␣-Methyl-5␤-androstane-3␣-ol-17␤-O-glucuronide
DSHS
DSHS = Institute of Biochemistry, Cologne, Germany; S = Steraloids, Wilmington, USA.

260
steroids 7 3 ( 2 0 0 8 ) 257–265
Table 2 – Optimized conditions of the synthesis of AAS glucuronides
Abbreviation
Substrate (␮M)
UDPGA (mM)
Protein (mg/mL)
5␣-NG
5␤-NG
5␤-EPIMG
5␣-MTG
5␤-MTG
5␣-MEG
250
500
500
1000
1000
500–1000
5
5
5
2.5
2.5
2.5
1
0.75
1
0.5
0.5
0.75
AG
d3-TG
NG
500
500
500
2.5
5
5
1
1
1
5␣-1-MEG
MTG
500
1000
5
7.5
1
0.5
See Tables 1A and 1B for nomenclature.
lier described procedure [21]. The treatment of the animals
was approved by the local ethical committee for animal
studies. BCA protein assay kit (Pierce, IL, USA) was used
for the determination of protein concentration (18.5 mg/mL),
which was used to standardize the amount of microsomal
enzymes.
were 50 ␮M (substrate), 5 mM (co-substrate), 0.5 mg/mL (pro-
tein) and incubation time of 2 h. d-Saccharic acid 1,4-lactone
(5 mM) was added to the reaction mixture as ␤-glucuronidase
inhibitor to prevent the effect of bond-breaking enzymes
of the tissue preparation. Incubation reaction was car-
ried out in 50 mM phosphate buffer (pH 7.4) with 5 mM
MgCl2. After centrifugation the internal standard (d3-5␤-
MTG) was added to the supernatant of the small-scale
samples and purified in SPE prior to LC–MS/MS analy-
sis.
For the synthesis of AAS glucuronide steroid aglycone was
diluted in 2 mL of methanol and filled up to the total volume
of 20 mL of earlier described phosphate buffer. The reaction
was initiated with the addition of UDPGA, carried out in
water bath of 37 ◦C for 12–15 h with continuous magnet stir-
ring, and terminated with transferring the incubation mixture
2.2.
Enzyme-assisted synthesis and purification
Synthesis conditions were optimized in small-scale stud-
ies in total volume of 100 ␮L with respect to concentration
of the substrate 1–1000 ␮M (AAS or its phase-I metabolite),
UDPGA 0.5–10 mM, and enzymatic protein 0.1–1.85 mg/mL.
Optimization of each parameter was carried out in turn
while keeping the other parameters constant in an overnight
incubation. The constant concentrations of each component
Fig. 2 – Optimization of synthesis conditions for 5␣-NG. Formation of the conjugate as function of (A) substrate
concentration, (B) UDPGA concentration, (C) microsomal protein concentration, and (D) incubation time. Constant
conditions: 50 ␮M substrate, 5 mM UDPGA, 0.5 mg/mL microsomal protein, and 2 h incubation time.

steroids 7 3 ( 2 0 0 8 ) 257–265
261
with nitrogen and reconstituted in LC–MS/MS solvents A:B
(75:25).
3.
Analytical methods
3.1.
Liquid chromatography
The LC separation was done by a Hewlett-Packard model
1100 binary pump and autosampler (Hewlett-Packard, Palo
Alto, CA, USA), using an endcapped Purospher RP-18 col-
umn (125 mm × 3 mm i.d., particle size 5 ␮m) (Merck). Eluent A
was 15 mM ammonium acetate–formic acid buffer (pH 4.2) in
water–acetonitrile (9:1, v/v) and eluent B contained the same
buffer, with the water–acetonitrile ratio 1:9 (v/v). The injection
volume was 5 ␮L. In gradient run eluent B was increased lin-
early from 25 to 40% in 11 min, followed by a rapid raise to
90% for 2 min and equilibration of 3 min. The LC flow rate of
0.5 mL/min was post-column split (1:10) (Acurate; LC-Packings,
The Netherlands) prior to introduction to ESI source.
3.2.
Mass spectrometry
The instrument used in MS and MS/MS measurements was
Applied Biosystems API3000 (Perkin-Elmer SCIEX, Concord,
ON, Canada) triple quadrupole instrument with a turbo ion-
spray source. The ionspray voltages were optimized to 5500
and 4500 V, and the orifice (declustering) voltages to 20 V, in
positive and negative ion mode ESI, respectively. Purified air
(custom-made air purifier; Atlas Copco, Belgium) was used
as the nebulizing gas (0.82 L/min) and curtain gas. Nitro-
gen (Generator Systems 75–72; Whatman, USA) was used as
the collision gas and turbo gas (6 L/min) and the ion source
temperature was adjusted to 350 ◦C. The MS and MS/MS spec-
tra were measured by using direct infusion of the purified
compounds with flow rate of 10 ␮L/min. The samples were
dissolved in the LC eluent consisting of 7.5 mM ammonium
acetate in water–methanol (50–50, v/v). In negative ion mode
the eluent was neutral, whereas in positive ion mode the
pH was adjusted to 4.2 with formic acid. The recorded mass
ranges were m/z 100–600 and m/z 50–550 in MS and MS/MS
experiments, respectively, with the scan speed of 5 s. Colli-
sion offset voltages 20–30 and 40 V were applied in positive
and negative ion mode, respectively, adjusted to produce
the spectrum of representative fragmentation for each glu-
curonide.
Fig. 3 – Elution of AAS glucuronide and steroid aglycone
from the reaction mixture of 5␣-1-MEG (A) and NG (B) as
function of methanol concentration in the eluent of
solid-phase extraction.
to ice bath, and precipitating the enzymatic protein with
dichloromethane, which also extracted the excess of start-
ing material. The aqueous phase of the synthesis, containing
the AAS glucuronide, was removed and further purified
by solid-phase extraction (SPE) using C18 cartridge (Isolute,
100 mg, International Sorbent Technology, UK) according to
the method described earlier by Borts and Bowers [22], except
to an additional SPE step, which was added to the proce-
dure to ensure the purity of steroid glucuronide fraction
with the final elution of 60% methanol in water. AAS glu-
curonide fraction was then evaporated to dryness in 60 ◦C
Table 3 – Synthesized steroid glucuronides and their yields
Compound
Yield (mg)
Yield (%)
Compound
Yield (mg)
Yield (%)
5␣-NG
5␤-NG
5␤-EPIMG
5␣-MTG
5␤-MTG
5␣-MEG
1.3
1.1
1.4
2.1
2.6
2.5
16
13
25
22
26
28
AG
d3-TG
NG
5␣-1-MEG
MTG
1.9
6.5
3.5
1.7
1.2
16
77
77
24
14
See Tables 1A and 1B for nomenclature.

262
steroids 7 3 ( 2 0 0 8 ) 257–265
3.3.
NMR spectroscopy
The synthesis method was suitable for producing steroid
glucuronides in milligram amounts (1.1–6.5 mg), with yields of
13–78% (Table 3). Steroid substrates with 4-ene-3-one structure
and 17␤-OH position solely available for O-glucuronidation
(NG, d3-NG and d3-TG) showed superior yields of 77–78% in
comparison to the substrates with completely saturated or
1-ene A-ring structure (13–28%). Glucuronidation of methyl-
testosterone was exceptional (14%), as the only possible site
for O-glucuronidation in methyltestosterone is the 17␤-OH-
group, sterically hindered by the neighboring 17␣-CH3, which
is likely to inhibit the catalysis of the bulky UGT enzymes.
Based on the detected formation of MTG, although in low
yield, it may be proposed that the structure still allows con-
jugation to some extent in the absence of a more favorable
functionality. Within the group of selected compounds 5␣-
androstane-3␣,17␤-diol and its d5-labeled analog were the
only ones with sterically non-hindered 3␣- and 17␤-hydroxyl
groups.
The 1H NMR experiments were carried out on a Varian
Unity 500 spectrometer at 23 ◦C using a gHX nano-NMR
probe. Acetone-d6 was used as solvent for MTG, 5␤-EPIM
and 5␤-EPIMG. The rest of the compounds were dissolved
in methanol-d4. The 1H chemical shifts were referenced to
residual acetone or methanol, 2.05 and 3.31 ppm, respec-
tively.
4.
Results and discussion
4.1.
Enzyme-assisted synthesis
Glucuronidation synthesis was optimized using sample vol-
umes of 100 ␮L with respect to substrate (steroid aglycone),
co-substrate (UDPGA) and UGT-enzymes (microsomal pro-
tein) concentration, in order to rationalize the consumption
of expensive starting materials, especially that of UDPGA.
The optimization experiments were carried out with 10%
methanol, which was a compromise between dissolution of
non-polar steroids and easily denaturizing enzymal protein.
All the tested steroid substrates were possible to conjugate
via enzyme-assisted pathway using rat liver microsomes, and
only slight differences were detected in the optimal condi-
tions for the various substrates (Table 2). Substrate inhibition
was not observed for any of the steroids within the concen-
tration range of 1–1000 ␮M, but when examining the effect
of substrate concentration, a plateau of glucuronidation was
observed at 500–1000 ␮M, except for 5␣-NG, for which it was
250 ␮M (Fig. 2A). The first point of the plateau was selected as
the concentration level for the bulk synthesis. The increase
of UDPGA concentration increased the yield of the glucu-
curonides until the plateau was achieved (Fig. 2B). Optimal
concentration of UDPGA was 2.5–5 mM for all the other com-
pounds except for methyltestosterone, which required 7.5 mM
concentration of co-substrate for maximal yield of MTG. The
formation of the steroid glucuronides was linearly propor-
tional to the concentration of protein up to 0.5–1 mg/mL
(Fig. 2C). The effect of incubation time was tested up to 24 h,
and observed to reach the plateau after 6 h of incubation at the
latest (Fig. 2D). However, there was no evidence of breakdown
of the glucuronide conjugate during that time, which allowed
an overnight incubation (12–15 h).
5.
Structure characterization
5.1.
Mass spectrometry
The abundant ammonium adduct ion [M+NH4]+ was detected
in positive ion ESI-MS for the glucuronides with saturated
A-ring system, whereas for d3-TG, NG and d3-NG with 4-ene-
3-one structure protonated molecule [M+H]+ was detected as
base peak. The ratio of protonated molecule and ammonium
adduct ion was found to strongly depend on the proton affinity
(PA) of the compound, higher proton affinity resulting into for-
mation of stronger protonated molecule. Proton affinity of the
4-ene-3-one structure in the steroid ring results higher pro-
ton affinity than hydroxyl group or carbonyl group without
conjugated double bond [23]. For 5␣-1-MEG and MTG, the rel-
ative abundance of [M+NH4]+ and [M+H]+ was approximately
equal. Formation of sodium adduct [M+Na]+ was also detected
for most of the compounds, although with variable relative
abundances. In negative ion ESI an intensive deprotonated
molecule [M−H]− was observed as the base peak for all AAS
glucuronides (Table 4). In negative ion ESI the most potential
Table 4 – Positive and negative ion ESI-MS of AAS
glucuronides
The larger scale enzymatic synthesis of the steroid glu-
curonides were carried out under optimized conditions
(Table 2) using 20 mL reaction volume. The excess of start-
ing material was removed in liquid–liquid extraction (LLE)
with dichloromethane, and recycled in prospective (non-
quantitative) synthesis. The more polar steroid glucuronide
remained in aqueous layer, but also some residue of start-
ing material was found to remain in that fraction. To increase
purity the sample was further purified with SPE. Retention of
steroid substrate and glucuronide conjugate in SPE was exam-
ined as a function of methanol percentage (1–100%) (Fig. 3).
The difference between the polarities of the two compounds
was sufficient for the complete isolation of the AAS glu-
curonide when methanol–water (60:40, v/v) was applied in the
final elution.
Compound
POS ESI
NEG ESI
[M+H]+
[M+NH4]+
[M+Na]+
[M−H]−
5␣-NG
5␤-NG
5␤-EPIMG
5␣-MTG
5␤-MTG
5␣-MEG
AG
d3-TG
NG
5␣-1-MEG
MTG
n.d.
n.d.
n.d.
n.d.
n.d.
479 (2)
n.d.
468 (100)
451 (100)
479 (89)
479 (94)
470 (100)
470 (100)
498 (100)
500 (100)
500 (100)
496 (57)
486 (100)
485 (41)
n.d.
n.d.
475 (43)
n.d.
451 (100)
451 (100)
479 (100)
481 (100)
481 (100)
477 (100)
467 (100)
466 (100)
449 (100)
477 (100)
477 (100)
505 (36)
505 (38)
501 (100)
491 (6)
490 (66)
473 (50)
501 (11)
501 (15)
496 (100)
496 (100)
See Tables 1A and 1B for nomenclature.

steroids 7 3 ( 2 0 0 8 ) 257–265
263
site for deprotonation [M−H]− is the carboxylic acid moiety of
the glucuronic acid, as the steroid aglycones are not ionized
in negative ion ESI at all. These results are in good agreement
with earlier studies by Bowers and Sanaullah [24], Williams
et al. [25] and Kuuranne et al. [26,27], and gave evidence on
the corrected molecule weights of the synthesized steroid glu-
curonides.
The positive ion MS/MS spectra (Table 3) showed abundant
product ions formed by loss of glucuronide moiety without
or with loss of one or two water molecules. The positive
charge stayed predominantly on the aglycone side, and only
weak fragments were detected from glucuronic acid side,
e.g. [Glu+H]+ = 177, [Glu+H−H2O]+ = 159, [Glu+H−2H2O]+ = 141
(Table 5). The negative ion MS/MS spectra of [M−H]− showed
intense product ions derived from the deprotonated glu-
curonic acid [Glu−H]− (Table 6). The spectra showed also weak
product ions formed by the cleavage of glucuronic acid moi-
ety [M−H−Glu]−, followed by the abstraction of two hydrogens
[M−H−Glu−2H]−, and losses of water [M−H−Glu−nH2O]−. The
positive and negative ion MS and MS/MS spectra indicate
clearly the presence of glucuronide moiety. However, the MS
data does not provide unambiguous information on the site of
glucuronidation.
5.2.
NMR spectroscopy
One-dimensional 1H NMR was used to determine the position
of the glycosidic linkage and the anomericity of the glucuronic
acid moiety in the steroid glucuronides. Each glucuronide
spectrum contained the five methine signals arising from the
glucuronic acid. The measured vicinal 3JGH1,GH2 coupling con-
stants of approximately 7.8 Hz confirmed the ␤-configuration
for the glucuronic acid for each glucuronide. Integration of the
signals gave a 1:1 glucuronic acid/aglycone ratio indicating the
correct stoichiometry.
As the steroid substrates resulting d4-5␣-NG, 5␣-MEG, d3-
TG, NG, d3-NG, 5␣1-MEG and MTG contain only one possible
site for glucuronidation there was no ambiguity about the
position of the glycosidic linkage in the corresponding glu-
curonides. In the structures of AG and 5␤-EPIMG there are
two possible glucuronidation sites, namely hydroxyl groups in
positions 3 and 17. For these compounds 1H spectra of both 5␣-
androstane-3␣,17␤-diol (A, steroid aglycone) and the steroid
glucuronide (AG) were measured. In 5␣-androstane-3␣,17␤-
diol H3 and H17 resonate downfield to the steroid bulk region
and can be distinguished from each other due to their different
coupling patterns. H17 of steroid aglycone is a triplet [28] with
a coupling constant of 8.5 Hz while H3 is typically a multiplet
resulting from several small couplings [29]. Comparison of the
steroid aglycone spectrum with its glucuronide-conjugated
product (AG) spectrum showed that the H17 signal shifted by
0.25 ppm downfield upon glucuronidation and on the other
hand that H3 signal did not shift, which suggests that the
glucuronide indeed is a 17-glucuronide.
For 5␤-EPIMG none of the methyl signals, including 17␤-
methyl, shifted upon glucuronidation while H1, H2 and H3
shifted downfield by 0.09, 0.12 and 0.16 ppm, respectively.
This suggests that 5␤-EPIMG is a 3-glucuronide. 5␣-NG, 5␤-NG,
5␣-MTG and 5␤-MTG have been fully characterized by NMR
previously [30].

264
steroids 7 3 ( 2 0 0 8 ) 257–265
6.
Conclusions
Glucuronide-conjugated metabolites of anabolic androgenic
steroids were produced by in vitro enzyme-assisted synthesis
using induced rat liver microsomes as the source of the conju-
gating uridine diphosphoglucuronosyl-transferase enzymes.
The method offers a practical pathway for the rapid produc-
tion of stereochemically pure AAS glucuronides in milligram
amount, such as needed in the development of analytical
methods. The yields of syntheses were mostly 13–28%, though
for substrates with conjugated 4-ene-3-one double bond sys-
tem significantly higher (77–78%). Due to relatively simple
reaction mixture and only minor differences between the opti-
mal conditions for the various AAS substrates, the addition of
a new structural analogue to the test set should be straight-
forward and thus also easily applied.
The intact AAS glucuronides can be ionized in both pos-
itive and negative ion ESI. The positive ion ESI mass the
base peak was either [M+H]+ or [M+NH4]+, depending on
the A-ring structure of the steroid. In negative ion mode an
intense deprotonated molecule [M−H]− was observed simi-
larly with each analyte, completing the molecular weight data
of the synthesized conjugates. The most representative mass
spectrometric information on the AAS glucuronide structure
was obtained from MS/MS fragmentation studies in positive
ion ESI, which is also the method of choice for the routine
LC–MS/MS analysis of AAS glucuronides.
Acknowledgments
Dr. Eivor Elovaara from National Institute of Health is grate-
fully acknowledged for preparation of microsomes, as well as
financial support from European Union (G6RD-CT-2001-00513)
and Finnish Funding Agency for Technology and Innovation
(TEKES).
r e f e r e n c e s
[1] Celotti F, Negri Cesi P. Anabolic steroids: a review of their
effects on the muscles, of their possible mechanisms of
action and of their use in athletics. J Steroid Biochem Mol
Biol 1992;43:469–77.
[2] Maravelias C, Dona A, Stefanidou M, Spiliopoulou C. Adverse
effects of anabolic steroids in athletes—a constant threat.
Toxicol Lett 2005;158:167–75.
[3] Scha¨nzer W. Metabolism of anabolic androgenic steroids.
Clin Chem 1996;42:1001–20.
[4] Thevis M, Scha¨nzer W. Mass spectrometry in sports drug
testing: structure characterization and analytical assays.
Mass Spectrom Rev 2007;26:79–107.
[5] Sekera MH, Ahrens BD, Chang YC, Starcevic B,
Georgakopoulos C, Catlin DH. Another designer steroid:
discovery, synthesis, and detection of ‘madol’ in urine.
Rapid Commun Mass Spectrom 2005;19:781–4.
[6] Scha¨nzer W, Donike M. Metabolism of anabolic steroids in
man: synthesis and use of reference substances for
identification of anabolic steroid metabolites. Anal Chim
Acta 1993;275:23–48.
[7] Thevis M, Opfermann G, Schmickler H, Scha¨nzer W. Mass
spectrometry of steroid glucuronide conjugates. I. Electron

steroids 7 3 ( 2 0 0 8 ) 257–265
265
impact fragmentation of 5␣-/5␤-androstan-3␣-ol-17-one
[20] WADA statistics of adverse analytical findings at
http://www.wada-ama.org/rtecontent/document/
LABSTATS 2005.pdf. Accessed on 7th December
2006.
glucuronides, 5␣-estran-3␣-ol-17-one glucuronide and
deuterium-labelled analogues. J Mass Spectrom
2001;36:159–68.
[8] 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
[21] Luukkanen L, Elovaara E, Lautala P, Taskinen J, Vainio H.
Characterization of 1-hydroxypyrene as a novel marker
substrate of 3-methylcholantrene-inducible phenol
UDP-glucuronosyltransferase(s). Pharmacol Toxicol
1997;80:152–8.
[22] Borts DJ, Bowers LD. Direct measurement of urinary
testosterone and epitestosterone conjugates using
high-performance liquid chromatography/mass
spectrometry. J Mass Spectrom 2000;35:
50–61.
3-keto-5␣-steroid-17-O-␤ glucuronides and
5␣-steroid-3␣,17␤-diol 3- and 17-glucuronides. J Mass
Spectrom 2001;36:998–1012.
[9] Sanaullah, Bowers LD. Facile synthesis of
[16,16,17-2H3]-testosterone, -epitestosterone and their
glucuronides and sulfates. J Steroid Biochem Mol Biol
1996;58:225–34.
[10] Stachulski AV, Jenkins GN. The synthesis of O-glucuronides.
Nat Prod Rep 1998;15:173–86.
[23] Harrison AG. Chemical ionization mass spectrometry. FL:
CRC: Boca Raton; 1982. p. 117.
[11] Lehman JP, Ferrin L, Fenselau C, Yost GS. Simultaneous
immobilization of cytochrome P-450 and
[24] Bowers LD, Sanaullah. Direct measurement of steroid
sulfate and glucuronide conjugates with high-performance
liquid chromatography–mass spectrometry. J Chromatogr B
1996;687:61–8.
glucuronyltransferase for synthesis of drug metabolites.
Drug Metab Dispos 1981;9:15–8.
[12] Lehman JP, Fenselau C, Depaulo JR. Quaternary
ammonium-linked glucuronides of amitriptyline,
imipramine, and chlorpromazine. Drug Metab Dispos
1983;11:221–5.
[25] Williams TM, Kind AJ, Houghton E, Hill DW. Electrospray
collision-induced dissociation of testosterone and
testosterone hydroxy analogs. J Mass Spectrom
1999;34:206–16.
[13] Antonio L, Grillasca JP, Taskinen J, Elovaara E, Burchell B, Piet
MH, et al. Characterization of catechol glucuronidation in
rat liver. Drug Metab Dispos 2002;30:199–207.
[14] Elovaara E, Mikkola J, Luukkanen L, Antonio L,
Fournel-Gigleux S, Burchell B, et al. Assessment of catechol
induction and glucuronidation in rat liver microsomes. Drug
Metab Dispos 2004;32:1426–33.
[15] Luukkanen L, Kilpela¨inen I, Kangas H, Ottoila P, Elovaara E,
Taskinen J. Enzyme-assisted synthesis and structural
characterization of nitrocatechol glucuronides. Bioconjugate
Chem 1999;10:150–4.
[16] Lautala P, Kivimaa M, Salomies H, Elovaara E, Taskinen J.
Glucuronidation of entacapone, nitecapone, tolcapone, and
some other nitrocatechols by rat liver microsomes. Pharm
Res 1997;14:1444–8.
[17] Rittmaster RS, Leopold CA, Thompson DL. Androgen
glucuronyl transferase activity in rat liver, evidence for the
importance of hepatic tissue in 5 alpha-reduced androgen
metabolism. J Steroid Biochem 1989;33:1207–12.
[18] Falany CN, Tephly TR. Separation, purification and
characterization of three isoenzymes of
[26] Kuuranne T, Vahermo M, Leinonen A, Kostiainen R.
Electrospray and atmospheric pressure chemical ionization
tandem mass spectrometric behavior of eight anabolic
steroid glucuronides. J Am Soc Mass Spectrom
2000;11:722–30.
[27] Kuuranne T, Kotiaho T, Pedersen-Bjergaard S, Rasmussen
KE, Leinonen A, Westwood S, et al. Feasibility of a
liquid-liquid-phase microextraction sample clean-up and
liquid chromatographic/mass spectrometric screening
method for selected anabolic steroid glucuronides in
biological samples. J Mass Spectrom 2003;38:
16–26.
[28] Hayamizu K, Kamo O. Complete assignments of the 1H and
13C NMR spectra of testosterone and
17␣-methyltestosterone and the 1H parameters obtained
from 600 MHz spectra. Magn Res Chem 1990;28:250–6.
[29] Kirk DN, Gaskell SJ, Marples BA. Spectroscopic methods of
steroid analysis. In: Makin HLJ, Gover DB, Kirk DN, editors.
Steroid analysis. Great Britain: Chapman & Hall; 1995.
p. 72.
[30] Kuuranne T, Aitio O, Vahermo M, Elovaara E, Kostiainen R.
Enzyme-assisted synthesis and structure characterization of
glucuronide conjugates of methyltestosterone (17
alpha-methylandrost-4-en-17-beta-ol-3-one) and
nandrolone (estr-4-en-17-beta-ol-3-one) metabolites.
Bioconjugate Chem 2002;13:194–9.
UDP-glucuronyltransferase from rat liver microsomes. Arch
Biochem Biophys 1983;227:248–58.
[19] Rao GS, Haueter G, Rao ML, Breuer H. An improved assay for
steroid glucuronyltransferase in rat liver microsomes. Anal
Biochem 1976;74:35–40.