Eudismic analysis of tricyclic sesquiterpenoid alcohols: Lead structures for the design of potent inhibitors of the human UDP-glucuronosyltransferase 2B7

Article abstract of DOI:10.1016/j.bioorg.2007.07.002

The epimeric tricyclic sesquiterpenoid alcohols globulol, epiglobulol, cedrol, epicedrol, longifolol, and isolongifolol were investigated in their ability to inhibit the recombinant human UDP-glucuronosyltransferase (UGT) 2B7. The stereoisomers displayed rapidly reversible competitive inhibition, which was substrate-independent. Longifolol and its stereoisomer isolongifolol displayed the lowest competitive inhibition constants (K_ic) of 23 and 26 nM, respectively. The K_ic values of cedrol and its epimer epicedrol were 0.15 and 0.21 μM, those of globulol and epiglobulol were 5.4 and 4.0 μM, respectively. The diastereomeric alcohols exhibited nearly identical affinities toward UGT2B7 indicating that the spatial arrangement of the hydroxy group had no influence on the dissociation of the enzyme-terpenoid complex. The high affinities stemmed presumably from mere hydrophobic interactions between the hydrocarbon scaffold of the terpenoid alcohol and the binding site of the enzyme. Glucuronidation assays revealed that there were large differences in the rates at which the epimeric alcohols were conjugated. Therefore, the spatial arrangement of the hydroxy group controlled the rate of the UGT2B7-catalyzed reaction. The introduction of a methyl group into the side chain of isolongifolol and longifolol increased the steric hindrance. As a result, the rate of the UGT2B7-catalyzed reaction was decreased by more than 88%. The findings indicated that the rate of the UGT2B7-catalyzed glucuronidation is significantly controlled by stereochemical and steric factors. Considering the high inhibition levels exerted by the tricyclic sesquiterpenoid alcohols, these compounds might serve as valuable lead structures for the design of potent inhibitors for UGT2B7.

Full text of DOI:10.1016/j.bioorg.2007.07.002

Bioorganic Chemistry 35 (2007) 386–400
www.elsevier.com/locate/bioorg
Eudismic analysis of tricyclic sesquiterpenoid
alcohols: Lead structures for the design
of potent inhibitors of the human
UDP-glucuronosyltransferase 2B7
a
a
Ingo Bichlmaier , Mika Kurkela a,b, Antti Siiskonen ,
b
a,
*
Moshe Finel , Jari Yli-Kauhaluoma
a
Faculty of Pharmacy, Division of Pharmaceutical Chemistry, University of Helsinki, P.O. Box 56,
FI-00014 Helsinki, Finland
b
Faculty of Pharmacy, Drug Discovery and Development Technology Center, University of Helsinki,
P.O. Box 56, FI-00014 Helsinki, Finland
Received 25 April 2007
Available online 15 August 2007
Abstract
The epimeric tricyclic sesquiterpenoid alcohols globulol, epiglobulol, cedrol, epicedrol, longifolol,
and isolongifolol were investigated in their ability to inhibit the recombinant human UDP-glucuron-
osyltransferase (UGT) 2B7. The stereoisomers displayed rapidly reversible competitive inhibition,
which was substrate-independent. Longifolol and its stereoisomer isolongifolol displayed the lowest
competitive inhibition constants (Kic) of 23 and 26 nM, respectively. The Kic values of cedrol and its
epimer epicedrol were 0.15 and 0.21 lM, those of globulol and epiglobulol were 5.4 and 4.0 lM,
respectively. The diastereomeric alcohols exhibited nearly identical affinities toward UGT2B7 indi-
cating that the spatial arrangement of the hydroxy group had no influence on the dissociation of the
enzyme–terpenoid complex. The high affinities stemmed presumably from mere hydrophobic inter-
actions between the hydrocarbon scaffold of the terpenoid alcohol and the binding site of the
enzyme. Glucuronidation assays revealed that there were large differences in the rates at which
the epimeric alcohols were conjugated. Therefore, the spatial arrangement of the hydroxy group con-
trolled the rate of the UGT2B7-catalyzed reaction. The introduction of a methyl group into the side
chain of isolongifolol and longifolol increased the steric hindrance. As a result, the rate of the
UGT2B7-catalyzed reaction was decreased by more than 88%. The findings indicated that the rate
*
Corresponding author. Fax: +358 9 19159556.
E-mail address: jari.yli-kauhaluoma@helsinki.fi (J. Yli-Kauhaluoma).
0045-2068/$ - see front matter ꢀ 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.bioorg.2007.07.002

I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
387
of the UGT2B7-catalyzed glucuronidation is significantly controlled by stereochemical and steric
factors. Considering the high inhibition levels exerted by the tricyclic sesquiterpenoid alcohols, these
compounds might serve as valuable lead structures for the design of potent inhibitors for UGT2B7.
ꢀ 2007 Elsevier Inc. All rights reserved.
Keywords: Metabolism; Enzyme; Inhibition; Stereochemistry; UGT; Longifolene
1. Introduction
The UDP-glucuronosyltransferases (UGTs) are membrane-bound enzymes of the
endoplasmic reticulum that catalyze the transfer of glucuronic acid (GlcA) from UDP-
a-D-glucuronic acid (UDPGlcA) to many lipophilic endogenous and xenobiotic com-
pounds [1–4]. This conjugation to GlcA increases the water solubility of the aglycone
and aids in its urinary and biliary excretion. The human genome contains many UGT-
encoding genes, which have been divided into two subfamilies, UGT1 and UGT2. The
UGT1A isoforms are encoded by a single large gene cluster on chromosome 2 and the
expression of individual UGTs of this subfamily is governed by exon sharing. In contrast,
UGT2B isoforms are encoded by separate genes clustered on human chromosome 4 [5–7].
The UGTs play important roles in the metabolism of prominent endobiotics and xeno-
biotics such as morphine, codeine, androgens, oestrogens, cholic acid derivatives, biliru-
bin, nonsteroidal anti-inflammatory drugs such as ibuprofen, and terpenoid alcohols
[8–11]. One of the major challenges in the study of UGTs is understanding their complex
substrate selectivity. Most, if not all human UGTs exhibit promiscuity1 with respect to
their substrates, meaning that each isoform can glucuronidate more than one aglycone
of rather variable chemical structure. This leads to partial overlap in the substrate selec-
tivities of different UGTs. On the other hand, even isoforms that can glucuronidate the
same compound often do it at different efficiencies. Furthermore, there are some aglycones
that are highly selective for a single UGT, like morphine that is glucuronidated at the
6-hydroxy position by UGT2B7, although other isoforms can catalyze the glucuronidation
of this opioid at the 3-hydroxy position [12]. Little is currently known, however, about the
structural elements both in the aglycone and the UGTs that are directly involved in the
determination of this complex substrate selectivity, also because the X-ray crystal structure
of any of the UGT isoforms has not been resolved to date. However, studies employing
computational chemistry tools led to predictive models for some UGT1A isoforms [13,14].
Despite their promiscuous behavior, we have recently shown that the UGT isoforms
2B7 and 2B17 displayed high selectivities toward stereoisomers that are related as mirror
images, or enantiomers [15,16]. The enantiomers used in these studies displayed nearly
identical affinities toward the enzymes, but they were glucuronidated at significantly
different rates, meaning that chiral distinction2 was limited to the actual UGT-catalyzed
1
The term catalytic promiscuity has been originally used to describe the ability of an enzyme to catalyze an
adventitious secondary activity at the active site responsible for the primary activity (see Copley SD. Enzymes
with extra talents: moonlighting functions and catalytic promiscuity. Curr. Opin. Chem. Biol. 2003, 7, 265–272).
However, in the case of metabolic enzymes like UGTs the term has been used to depict their overlapping
substrate selectivities.
2
Chiral discrimination is used synonymously for chiral distinction. It describes the stereoselective recognition
of chiral guests (e.g., chiral drugs) by chiral molecular hosts (e.g., enzymes, receptors, transporters, channel
proteins, etc.).

388
I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
substitution reaction. It was concluded that the spatial arrangement of the hydroxy group
had no influence on the affinity but merely on the glucuronidation reaction per se.
This study investigates stereochemical events during the formation of the enzyme–ter-
penoid complex and the UGT2B7-catalyzed GlcA-transfer reaction. The UGT isoform
2B7 was chosen for this study because this enzyme was identified as the key enzyme
responsible for drug glucuronidation. In this respect, UGT2B7 accounts for the transfor-
mation of 40% of drugs that are metabolized by UGT isoforms [17]. Since many endobi-
otics and xenobiotics occur as chiral entities [18–20], and UGT2B7 has displayed
stereoselective glucuronidation toward enantiomeric probes (vide supra) we decided to
study the behavior of this enzyme towards diastereomeric compounds. For this purpose,
the results of the eudismic analysis3 of the epimeric sesquiterpenoid alcohols globulol, epi-
globulol, cedrol, epicedrol, longifolol, and isolongifolol are reported (Scheme 1) [21].
These sesquiterpenoid alcohols are related as diastereomers and display distinct physico-
chemical properties, in contrast to enantiomers. However, the only configurational differ-
ence between the epimers is the spatial orientation of the hydroxy group in the case of
globulol, epiglobulol, cedrol, and epicedrol, and the hydroxymethyl group in the case of
longifolol and isolongifolol, the very site of the UGT-catalyzed glucuronidation reaction.
This structural feature enables the investigation of chiral distinction during the formation
of the enzyme–substrate complex as well as during the GlcA-transfer reaction. Further-
more, this study demonstrates that stereochemical analysis could provide the key to the
design of potent inhibitors of UGT2B7.
The affinity of the different stereoisomers toward UGT2B7 was determined by mea-
suring their IC50 values and dissociation constants. Estriol was chosen as the reference
substrate because this steroid was conveniently detected by fluorescence spectroscopy.
In addition, estriol showed simple Michaelis–Menten kinetics without substrate inhibi-
tion, and the enzyme assays displayed good reproducibility. Moreover, estriol dis-
played high affinity toward UGT2B7 (Km 4.6 lM), hence the total concentration of
substrate and inhibitor in the enzyme assays was low. This became important when
compounds that display lower solubilities such as the terpenoid alcohols in this study
were assayed.
2. Materials and methods
2.1. Materials
(+)-Longifolol and (À)-isolongifolol were synthesized from (+)-longifolene following
published procedures [28]. (+)-Longifolene was a generous gift from Camphor & Allied
Products Ltd. (Mumbai, India) and was purified before use as previously described [29].
(À)-Globulol (CAS 489-48-1), (À)-epiglobulol (CAS 88728-58-9), (+)-cedrol (CAS
77-53-2), and (À)-epicedrol (CAS 19903-73-2) were obtained from Fluka (Buchs, Switzer-
land). All the terpenoid alcohols were purified by flash chromatography and, in addition,
3
Eudismic analysis investigates the interaction between chiral compounds and their biological targets. This
analytical method is used in drug design and development to increase the selectivity and potency of a chiral lead
compound toward the pharmacological target [19,20]. The term eudismic analysis was derived from the
expressions eutomer and distomer. The more biologically active stereoisomer is termed eutomer, the less active
one is called distomer.

I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
389
Scheme 1. The structures of the epimeric tricyclic alcohols.
solid terpenoid alcohols were recrystallized prior to use. The identity of the terpenoid
1
alcohols was confirmed by H NMR, 13C NMR, and GC–MS. The purity of the com-
pounds was determined by measuring the optical rotation and by combustion analysis.
The longifolol and isolongifolol analogues 1 and 2 were synthesized from longifolyl alde-
hyde and isolongifolyl aldehyde4 according to published procedures [30]. All terpenoid
1
alcohols were identical in their H NMR, 13C NMR, GC–MS and optical rotation. The
combustion analysis showed that all the applied terpenoid alcohols were of high purity
(deviation from theoretical values <0.3%). UDPGlcA (trisodium salt, CAS 63700-19-6),
saccharic acid-1,4-lactone (CAS 61278-30-6), estriol (CAS 50-27-1), estriol-17b-(b-D-glu-
curonide) (CAS 7219-89-8), scopoletin (CAS 92-61-5), and epitestosterone (CAS 481-30-
1) were obtained from Sigma (St. Louis, MO, USA). 4-Nitrophenol (CAS 100-02-7)
was from Aldrich (Schnelldorf, Germany). Radiolabeled [14C]UDPGlcA was acquired
from PerkinElmer Life and Analytical Sciences (Boston, MA, USA). HPLC grade solvents
4
Longifolyl and isolongifolyl aldehyde were prepared as follows: to a solution of (+)-longifolol or
(À)-isolongifolol (0.23 g, 1.0 mmol) in anhydrous DMSO (15 mL) was added IBX (0.70 g, 2.5 mmol) and the
resulting suspension was stirred under an atmosphere of dry argon at room temperature (23 ꢁC) for 4 h. The
reaction was slightly exothermic and the initial slurry became a colorless solution, which finally turned into a
finely dispersed suspension. After completion of the reaction, water (5.0 mL) was added and the suspension was
extracted with Et2O (3 · 75 mL). The combined organic layers were washed with saturated NaHCO3
(1 · 100 mL), water (3 · 100 mL), and brine (100 mL), dried over anhydrous Na2SO4, and filtered through a
small pad of silica. The silica pad was washed with Et2O (2 · 30 mL). The organic solvent was removed under
reduced pressure and the resulting colorless oil was dried in vacuo affording the aldehydes in quantitative yields
and high purities (>99%, GC–MS).

390
I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
were used throughout the study. The recombinant human UGT2B7 was expressed as a
His-tagged protein in baculovirus-infected insect cells as previously described [31].
2.2. Analytics
The compounds were analyzed by NMR on a Varian Mercury 300 MHz spectrometer
(Varian, Palo Alto, CA, USA) and by GC–MS. The GC–MS system consisted of a
5890A gas chromatograph and a 5970 mass selective detector (Hewlett Packard, Palo
Alto, CA, USA). NMR spectra were processed by use of Varian VNMR software
(Varian, Palo Alto, CA, USA) on a UNIX Solaris workstation. GC–MS spectra were
analyzed with the HP ChemStation program on a Windows 3.1 workstation. Optical
rotation was determined by use of a Polartronic E polarimeter with a micro tube for
small sample volumes (100 mm, 1.10 mL; Schmidt + Haensch, Berlin, Germany). The
elemental analyses were conducted by Robertson-Microlit Laboratories (Madison, NJ,
USA).
2.3. IC50 values
The IC50 values of the terpenoid alcohols were determined at five concentrations brac-
keting their apparent IC50 (0.10, 0.25, 1.0, 4.0, and 10 · IC50). Estriol was employed at
25 lM. The Km value of estriol for UGT2B7 (4.6 lM) was determined frequently in our
laboratory and the value represents the average value of experiments conducted during
the last 7 months using the same enzyme batch. The reaction mixture consisted of phos-
phate buffer (50 mM, pH 7.4), 10 mM MgCl2, and 5.0 mM saccharic acid-1,4-lactone.
The enzyme was employed at 0.1 mg protein mLÀ1. Control assays in the absence of inhib-
itor and blank runs in the absence of cosubstrate and estriol were performed for each inhi-
bition assay. The enzyme reaction was initiated after a 5 min pre-incubation time at 37 ꢁC
by the addition of a solution of UDPGlcA to a final concentration of 5.0 mM. The enzyme
reactions were terminated after an incubation time of 15 min at 37 ꢁC by the addition of
ice-cold perchloric acid (4.0 M) and transfer to ice. The mixtures were centrifuged
(16,000g, 10 min) and aliquots of the supernatants were subjected to HPLC analysis.
The data were analyzed by nonlinear regression applying the 2-parameter Hill equation.
The results reflect a minimum of three replicate determinations.
2.4. Testing for reversibility
UGT2B7 at a concentration of 5.0 mg protein mLÀ1 was pre-incubated in buffered
solution (phosphate buffer, 50 mM, pH 7.4) at 37 ꢁC with a concentration of terpenoid
alcohol equivalent to 10-fold its IC50. After an equilibration time of 45 min, this mixture
was diluted 50-fold into reaction buffer (37 ꢁC) containing the reference substrate estriol
(25 lM), phosphate buffer (50 mM, pH 7.4), UDPGlcA (5.0 mM), MgCl2 (10 mM), and
saccharic acid-1,4-lactone (5.0 mM) to initiate reaction. The formation of glucuronide
was monitored every 2.5 min over 60 min by pipetting 100 lL of the reaction buffer to a
vial containing 10 lL perchloric acid (4.0 M). The acidified mixtures were transferred to
ice, centrifuged (16,000g, 10 min), and aliquots of the supernatants were subjected to
HPLC analysis. Control assays in the absence of terpenoid alcohol were included. The
results reflect a minimum of two replicate determinations.

I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
391
2.5. Mode of inhibition
The mode of inhibition was determined by measuring the IC50 values of the terpenoid
alcohols at six [S]/Km ratios of estriol (0.5, 1.0, 2.0, 5.0, 10, and 15). The terpenoid alcohol
was employed at five concentrations bracketing its estimated IC50 (0.10, 0.25, 1.0, 4.0, and
10 · IC50). One control assay to determine the % inhibition at each [S]/Km ratio of estriol
employing longifolol at a fixed concentration of 0.10 lM was included. The assays were
carried out as described above. The results reflect a minimum of two replicate
determinations.
2.6. Inhibition constants
The Kic values were determined at five estriol concentrations (2.5, 5.0, 10, 25, and
50 lM). The terpenoid alcohols were employed at three concentrations bracketing their
estimated Kic values (0.25, 1.0, and 4.0 · Kic), which were calculated from the respective
IC50 values by the Cheng–Prusoff equation [11]. The assays were carried out as described
above. The results reflect a minimum of three replicate determinations.
2.7. Testing for substrate-independent inhibition
The IC50 values of the terpenoid alcohols longifolol and isolongifolol for UGT2B7 were
determined by the use of four structurally different reference substrates. The assay condi-
tions were balanced using the substrates at concentrations resembling their Km values,
which have been previously determined in our laboratory: 4-nitrophenol (500 lM,
0.10 mg protein mLÀ1, incubation time 10 min), scopoletin (320 lM, 0.10 mg protein
mLÀ1, incubation time 10 min), morphine (200 lM, 0.20 mg protein mLÀ1, incubation
time 35 min), and epitestosterone (15 lM, 0.25 mg protein mLÀ1, incubation time
15 min). Longifolol and isolongifolol were employed at five concentrations bracketing
their IC50 values (0.10, 0.25, 1.0, 4.0, and 10 · IC50). The assays were carried out as
described above. The results reflect a minimum of three replicate determinations.
2.8. Glucuronidation assays
The formation of terpenoid glucuronide was measured by the use of radiolabeled
[14C]UDPGlcA [32].
A
solution of [14C]UDPGlcA (50 lL, 196 mCi mmolÀ1
,
0.02 mCi mLÀ1 in EtOH/water 7:1 v/v) was transferred to 1-mL vials and the solvent
was evaporated in vacuo. The residue was dissolved in reaction buffer (90 lL), which con-
sisted of UGT2B7 (2.0 mg protein mLÀ1), cold UDPGlcA (450 lM), MgCl2 (10 mM),
phosphate buffer (50 mM, pH 7.4) and saccharic acid-1,4-lactone (5.0 mM). The reaction
was initiated by the addition of a solution of terpenoid alcohol (10 lL, 1.0 mM in 50-vol %
DMSO/water) to a final saturating concentration of 100 lM. After incubating for 14 h at
37 ꢁC, the reactions were terminated by the addition of perchloric acid (10 lL, 4.0 M) and
transfer to ice. The mixtures were centrifuged (16,000g, 10 min) and aliquots of the super-
natants were subjected to HPLC analysis. Control assays in the presence of estriol (25 lM)
and blank runs were included. The detection limit was determined by subjecting dilutions
of reaction buffer containing estriol [14C]b-D-glucuronide to HPLC analysis. The results
reflect a minimum of six replicate determinations.

392
I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
2.9. HPLC methods
The HPLC system consisted of the Agilent 1100 series degasser, binary pump, autosam-
pler, thermostated column compartment, multiple wavelength detector, and fluorescence
detector (Agilent Technologies, Palo Alto, CA, USA). The resulting spectra were analyzed
with Agilent ChemStation software (Rev B.01.01). Glucuronidation reaction products
were separated and detected as follows. Estriol b-D-glucuronide: Hypersil BDS-C18
(150 · 4.6 mm; Agilent Technologies, Palo Alto, CA, USA); 35% MeOH in phosphate
buffer (50 mM, pH 3.0); flow rate 1.0 mL minÀ1; detection by fluorescence spectroscopy
(excitation at k 335 nm, emission at k 455 nm); retention time 4.2 min. 4-Nitrophenol
b-D-glucuronide: Chromolith SpeedROD (50 · 4.6 mm; Merck, Darmstadt, Germany);
15% acetonitrile in phosphate buffer (50 mM, pH 3.0); flow rate 1.0 mL minÀ1; detection
by UV spectroscopy (k 300 nm); retention time 1.3 min. Scopoletin b-D-glucuronide:
Chromolith SpeedROD (50 · 4.6 mm); 10% MeOH in phosphate buffer (50 mM, pH
3.0); flow rate (gradient run) 0.8 mL minÀ1 (0.0–4.5 min), 0.8 mL fi 2.5 mL minÀ1 (4.5–
5.0 min), 2.5 mL minÀ1 (5.0–11.0 min), 2.5 fi 0.8 mL minÀ1 (11.0–12 min); detection by
fluorescence spectroscopy (excitation at k 335 nm, emission at k 455 nm); retention time
4.2 min. Epitestosterone b-D-glucuronide: Chromolith SpeedROD (50 · 4.6 mm); 43%
MeOH in phosphate buffer (50 mM, pH 3.0); flow rate 2.0 mL minÀ1; detection by UV
spectroscopy (k 246 nm); retention time 5.9 min. Morphine 6-b-D-glucuronide: Hypersil
BDS-C18 (250 · 4.0 mm); 1% acetonitrile in phosphate buffer (50 mM, pH 3.0); flow rate
0.80 mL minÀ1; detection by fluorescence spectroscopy (excitation at k 285 nm, emission
at k 335 nm); retention time 14.7 min. Terpenoid [14C]b-D-glucuronides: Chromolith
SpeedROD (50 · 4.6 mm); gradient run 5% MeOH in phosphate buffer (50 mM, pH
3.0; 0.0–3.5 min), 5% fi 80% MeOH (3.5–10 min), 80% MeOH (10–15 min), 80% fi 5%
MeOH (15–20 min), 5% MeOH (20–25 min); flow rate 1.0 mL minÀ1; retention times
9.0–9.4 min for terpenoid [14C]b-D-glucuronides; retention times 1.5–1.7 min for
[14C]UDPGlcA; detection by use of a 9701 HPLC Radioactivity Monitor (Reeve Analy-
tical, Glasgow, Scotland).
3. Results
The IC50 values of the sesquiterpenoid alcohols was measured using the reference sub-
strate estriol at a concentration of 25 lM. This concentration, which is higher than the Km
of estriol for UGT2B7 (4.6 lM), was chosen to allow for a reliable integration of the
estriol glucuronide signal also when high-affinity substrates such as longifolol and isol-
ongifolol were assayed at higher concentrations. It should be mentioned that estriol is
selectively glucuronidated by UGT2B7 at the hydroxy group in position 17b. The IC50 val-
ues along with the eudismic ratios are displayed in Table 1. The eudismic ratio was calcu-
lated by dividing the IC50 values of the epimeric alcohols. The eudismic ratio therefore
quantifies the preferential dissociation of one enzyme–epimer complex over the other.
As can be seen from Table 1, the IC50 values of the terpenoid alcohols ranged from
74 nM to 29 lM, representing an approximately 400-fold difference. Longifolol and
isolongifolol displayed the highest affinity toward UGT2B7 with IC50 values of 74 and
97 nM, respectively. The epimeric alcohols displayed nearly identical affinities as expressed
by eudismic ratios that ranged merely between 1.1 and 1.3. However, these eudismic ratios
were meaningful only when the stereoisomeric compounds displayed identical modes of

I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
393
Table 1
IC50 and Kic values of the epimeric alcohols
a
b
c
b
Compound
IC50 (lM)
CI95%
Eudismic ratiod Kic (lM)
CI95%
Eudismic ratioe
1.4
Globulol
Epiglobulol
Cedrol
Epicedrol
Longifolol
29.2 1.5
24.5 1.6
0.63 0.042
0.60 0.045
0.074 0.0055 0.063–0.086 1.3
26.0–32.3
21.0–28.0
0.54–0.72
0.50–0.69
1.2
5.4 0.79
3.9–7.0
3.4–4.7
0.13–0.18
0.18–0.25
4.0 0.32
0.15 0.012
0.21 0.016
0.023 0.0011 0.021–0.025 1.1
0.026 0.0020 0.022–0.030
1.1
1.4
Isolongifolol 0.097 0.0070 0.082–0.11
a
Mean value standard deviation (n = 15).
95% Confidence interval.
Mean value standard deviation (n = 60).
Calculated by dividing the higher IC50 value of one epimer by the lower value of its respective stereoisomer.
b
c
d
e
Calculated from the respective Kic values.
inhibition. These findings could also have been a result of unspecific, noncompetitive and
uncompetitive binding to the enzyme, blurring the chiral distinction of the active-site
directed, competitive interaction. In this respect, the compounds were tested for reversibil-
ity and their mode of inhibition. The reversibility of inhibition was assessed according to
standard procedures by measuring the recovery of enzymatic activity after a rapid and
large dilution. It was found that the sesquiterpenoid alcohols were identical in their beha-
vior and they were rapidly reversible inhibitors (Fig. 1).
The mode of inhibition—either competitive, uncompetitive, noncompetitive, or mixed-
type—was determined by measuring the effects of the [S]/Km ratio on the apparent IC50
values of the terpenoid alcohols [22]. As can be seen from Fig. 2, the stereoisomeric alco-
hols displayed competitive inhibition toward UGT2B7 because the corresponding IC50
values increased linearly with increasing concentration of the reference substrate estriol.
Taken together, the six tricyclic sesquiterpenoid alcohols displayed an identical mode of
inhibition and the calculated eudismic ratios were therefore indeed meaningful. The eud-
ismic ratios close to unity expressed that both epimers had similar affinities toward the
binding site of UGT2B7. Therefore, there was no chiral distinction during the formation
Fig. 1. The reversibility of inhibition was determined by measuring the recovery of enzyme activity after a rapid
and large dilution. The data for longifolol are displayed (h, n = 2). One control assay in the absence of inhibitor
was included (n, n = 2). The curve represents the expected behavior for a rapidly reversible inhibitor. The other
epimeric alcohols displayed similar curves. The ordinate displays the amount of formed estriol b-D-glucuronide
per mg protein.

394
I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
Fig. 2. (a) Typical semilogarithmic concentration-response plot for the inhibition assays. Globulol was used as
the inhibitor. Estriol was used as the reference substrate at a concentration of 10 lM. The IC50 for globulol was
12.7 lM 0.88 (CI95% 10.8–14.7). (b) IC50 values for globulol (s) at different [S]/Km ratios of estriol. The
abscissa is plotted on a logarithmic scale for clarity. The mean values and standard deviations are displayed
(n = 10). Control assays to determine the % inhibition exerted by longifolol at a fixed concentration of 0.10 lM
were included at each [S]/Km ratio (n, n = 2). The curves represent the expected behavior for a competitive
inhibitor. The other terpenoid alcohols displayed similar curves.
of the encounter complex, even though the epimers possessed different physicochemical
properties.
The competitive inhibition constants (Kic) of the sesquiterpenoid alcohols were deter-
mined to derive a kinetic parameter that was independent of the substrate concentration.
Furthermore, the Kic values were measured to confirm the high affinities, low eudismic
ratios and the mechanism of inhibition displayed by the tricyclic alcohols as indicated
by the studies described above (Table 1). The data of this analysis were fitted to the com-
petitive, noncompetitive, mixed-type, and uncompetitive inhibition models. The different
models were ranked according to the corrected Akaike’s information criterion (results
not shown). Based on these results, the active-site directed, competitive inhibition model
was chosen and the resulting Kic values were in good agreement with those calculated from
the respective IC50 values by the Cheng–Prusoff equation (not shown) [23]. Accordingly,
longifolol and isolongifolol displayed very low Kic values of 23 and 26 nM, respectively.
The Kic values of cedrol and its diastereomer epicedrol were approximately 10-fold higher
(0.15 and 0.21 lM, respectively), and those of globulol and epiglobulol were 5.4 and
4.0 lM, respectively (Table 1). The Kic values of the epimers were nearly identical, in
agreement with the corresponding IC50 values (vide supra). This study substantiated the

I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
395
finding that there was no significant level of chiral distinction during the formation of the
enzyme–epimer complex, meaning that the configuration at the chiral center bearing the
hydroxy or hydroxymethyl group had no significant influence on the affinity toward
UGT2B7.
Furthermore, it was determined that the compounds displayed similar affinity levels
independently of a second substrate. The substrate-independency was assayed by measur-
ing the IC50 values of the sesquiterpenoid alcohols longifolol and isolongifolol by the use
of four structurally different substrates of UGT2B7. In this respect, the IC50 values of the
two epimeric alcohols were determined using the reference substrates 4-nitrophenol, mor-
phine, scopoletin, and epitestosterone. It should be mentioned here that morphine was
included as a substrate because morphine glucuronidation at the 6-hydroxy position is
selectively catalyzed by UGT2B7 [24]. The measured IC50 values that were determined
with the four reference substrates ranged from 45 to 69 nM (standard deviation <9.6%;
n = 10) and from 42 to 78 nM (standard deviation <8.7%; n = 10) for longifolol and isol-
ongifolol, respectively. The corresponding eudismic ratios were between 1.1 and 1.4. These
results confirmed that the inhibition exerted by longifolol and isolongifolol was indepen-
dent of the second substrate applied. Also the eudismic ratio as a measure of the chiral
distinction during the formation of the encounter complex was independent of the
employed second substrate.
In the preceding sections, the influence of the spatial orientation of the hydroxy group
on the affinity was investigated. Furthermore, the mode of inhibition and substrate-inde-
pendency were assessed. The subsequent sections deal with the influence of structural fea-
tures, i.e., stereochemistry and steric hindrance, on the rate of the UGT2B7-catalyzed
glucuronidation reaction.
The sesquiterpenoid alcohols and their corresponding b-D-glucuronides could not be
detected either by UV or fluorescence spectroscopy due to the lack of appropriate chro-
mophores in the saturated tricyclic hydrocarbon scaffold. Therefore, the formation of
the terpenoid b-D-glucuronides was assayed by the use of radiolabeled [14C]UDPGlcA
after incubating the reaction buffer for 14 h at a high UGT2B7 concentration of 2.0 mg
protein mLÀ1. The detection limit was 8.0 pmol terpenoid b-D-glucuronide at a signal-
to-noise ratio of 10. The high concentration of enzyme and the long incubation time were
necessary to unambiguously identify and detect the formed terpenoid [14C]b-D-glucuro-
nides. The concentration of radiolabeled cosubstrate was limited due to the high costs
and to minimize the radioactive waste output. However, the employed cosubstrate concen-
tration of 500 lM was sufficient for the purpose of this study. The tricyclic alcohols were
employed at a saturating concentration of 100 lM to ensure that depletion of terpenoid
alcohol did not occur. The best conversion was 8.1% (longifolol), therefore, no depletion
of terpenoid alcohol due to glucuronide formation took place and the results of this study
were consistent and comparable. Hence, the amount of formed b-D-glucuronide could be
applied as a measure of the glucuronidation rate. The radiochromatograms of this analysis
are displayed in Fig. 3 and the quantification results are summarized in Table 2. It was
found that UGT2B7 glucuronidated globulol at a much higher rate than its epimer epi-
globulol. The amount of globulol glucuronide was 789 pmol, that of its stereoisomer epi-
globulol was below the detection limit of 8.0 pmol. A similar pattern was observed for the
UGT2B7-catalyzed glucuronidation of longifolol and its epimer isolongifolol. A large
amount of longifolol glucuronide was formed (810 pmol), in contrast to 65 pmol of isol-
ongifolol glucuronide. Cedrol and epicedrol were significantly poorer substrates of

396
I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
Fig. 3. Radiochromatograms of the formed terpenoid [14C]b-D-glucuronides of longifolol (a), isolongifolol (b),
globulol (c), and epiglobulol (d). The curves depict clear differences in the amount of formed glucuronide
(retention time approx. 9.2 min). Longifolol and globulol were much better substrates of UGT2B7 compared to
their epimers isolongifolol and epiglobulol. The glucuronide of epiglobulol was not detected. The peak at approx.
1.6 min is the not consumed radiolabeled cosubstrate.
Table 2
Glucuronidation of the sesquiterpenoid alcohols
Compound
Glucuronidea (pmol)
% Conversion
Eudismic ratio
>98
Globulol
Epiglobulol
Cedrol
Epicedrol
Longifolol
Isolongifolol
a
789 55
<8b
31 1.6
<8b
810 70
65 4.2
7.9
<0.08
0.31
<0.08
<8.1
0.65
>3.9
12
Mean value standard deviation (n = 6).
Signal below detection limit.
b
UGT2B7 compared to the other sesquiterpenoid alcohols. The response of epicedrol glu-
curonide was below the detection limit and the amount of cedrol glucuronide was 31 pmol.
This analysis showed that the eudismic ratios were significantly higher than those of the
IC50 and Kic values, indicating significant levels of chiral distinction during the actual glu-
curonidation reaction. Especially the UGT2B7-catalyzed glucuronidation of globulol was
highly favored over the enzymatic conjugation of its stereoisomer epiglobulol (eudismic
ratio > 98). Furthermore, the glucuronidation of longifolol and cedrol were preferred over
the conjugation of their respective stereoisomers by factors of 12 and more than 3.9,
respectively.

I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
397
Fig. 4. Graphical representation of steric hindrance. The introduction of one additional methyl group to the side
chain in longifolol (compound 1, top) and isolongifolol (compound 2, bottom) increased steric hindrance
(depicted by the solid arcs). This resulted in decreased accessibility to the hydroxy group, which resulted in lower
glucuronidation rates.
Two derivatives of longifolol and its epimer isolongifolol were synthesized by the for-
mal replacement of one H-atom by a methyl group to investigate the influence of steric
repulsion on the glucuronidation rate (Fig. 4). The accessibility of the hydroxy group
for nucleophilic substitution in the derivatives 1 and 2 was decreased due to the introduc-
tion of an additional alkyl substituent in the vicinity of the reaction center. Accordingly, a
decrease in the glucuronidation rate was anticipated. As was expected, the amount of
formed glucuronide of the derivatives 1 and 2 (31 and < 8 pmol, respectively) decreased
significantly compared to those of longifolol and isolongifolol. The concentration of
formed glucuronide was reduced by 96% (compound 1) and by more than 88% (compound 2)
compared to longifolol and isolongifolol, respectively. On the other hand, both compounds
displayed high affinity toward the enzyme. The IC50 of compound 1 was 77 nM (standard
deviation = 5.3, n = 15) and that of compound 2 was 0.40 lM (standard deviation = 0.039,
n = 15). This result showed that the steric hindrance exerted by the additional methyl group
had a significant influence on the glucuronidation rate, whereas the high affinity toward
UGT2B7 was retained.
4. Discussion
The IC50 and Kic values of the epimeric alcohols were nearly identical (Table 1). This
finding indicated that the spatial orientation of the hydroxy and hydroxymethyl group
had no effect on the affinity toward the binding site of UGT2B7. Since the IC50 and Kic
relate to the dissociation (formation) of the enzyme–epimer complex [22], it was concluded
that the formation of the encounter complex was not influenced by the configuration at the

398
I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
asymmetric carbon atom bearing the hydroxy or hydroxymethyl group. Therefore, the
enzyme exhibited no significant chiral distinction during this recognition process towards
the epimeric terpenoid alcohols, and UGT2B7 accepted the different stereoisomers at
nearly identical affinities. This might suggest that there was no significant interaction
between the hydroxy group of the epimeric alcohols and the binding site of UGT2B7 that
could contribute to the affinity level, and therefore did not result in different IC50 and Kic
values between the two stereoisomers. Taking this into account, it might be concluded that
the high affinities of especially longifolol and isolongifolol stemmed merely from hydro-
phobic interactions between the tricyclic hydrocarbon scaffold of the terpenoid alcohols
and the binding site of UGT2B7. This indicated that the enzyme did not recognize the
hydroxy group within the tricyclic substrates during the formation of the enzyme–terpe-
noid complex.
In contrast to the formation of the enzyme–terpenoid complex, the actual UGT2B7-cat-
alyzed transfer of GlcA was stereoselective, because globulol, longifolol, and cedrol were
conjugated at significantly higher rates than their respective epimers (Table 2). This finding
indicated chiral distinction during the GlcA-transfer reaction. Especially the conjugation of
globulol was highly stereoselective as indicated by its eudismic ratio of >98 and the glucu-
ronidation of longifolol was favored over the conjugation of its epimer by a factor of 12.
The smaller eudismic ratio of the latter compound might be explained by the higher flexibility
of the hydroxymethyl group, rendering also isolongifolol a better substrate. In the cases of
globulol and cedrol the hydroxy group is directly attached to the rigid tricyclic hydrocarbon
scaffold, which results in decreased flexibility of the hydroxy group (Scheme 1). The eudismic
ratio of cedrol and epicedrol glucuronidation could not be assessed sufficiently due to very
low conjugation rates. However, the eudismic ratio was at least 3.9 indicating chiral distinc-
tion that favors the glucuronidation of cedrol over its epimer.
Cedrol and epicedrol were poor substrates of UGT2B7 despite their high affinities
toward this enzyme (Tables 1 and 2). The common characteristic of these two epimers
was the high steric hindrance in the vicinity of the nucleophilic hydroxy group, resulting
in decreased rates of the UGT2B7-catalyzed conjugation reaction. Furthermore, also epi-
globulol was found to be a very poor substrate since its glucuronide could not be detected.
The common feature between cedrol, epicedrol, and epiglobulol was the methyl group at
the asymmetric carbon atom that bears the nucleophilic hydroxy group. It was therefore
concluded that the steric hindrance exerted by this methyl group could be partly respon-
sible for the low conjugation rates. Since the methyl group is also encountered in globulol,
which was a very good substrate of UGT2B7, we concluded that the combination of spa-
tial orientation and steric hindrance resulted in low glucuronidation rates. To test this
hypothesis, two derivatives of longifolol and isolongifolol were synthesized (Fig. 4). It
was anticipated that the additional methyl group would decrease the accessibility to the
hydroxy group during the UGT2B7-catalyzed reaction, resulting in a decreased glucuron-
idation rate. The glucuronidation assays showed that the introduction of the methyl group
into the longifolol and isolongifolol scaffold indeed decreased the glucuronidation by at
least 88%. Therefore, it was concluded that the glucuronidation rate was controlled by
the orientation of the hydroxy group as well as by steric hindrance.
In this study, it was demonstrated that the spatial arrangement of the hydroxy and
hydroxymethyl group and the steric demand in the vicinity of this nucleophilic group gov-
erned the glucuronidation of epimeric terpenoid alcohols. However, the affinity level was
not affected by the configuration at the asymmetric carbon atom bearing the hydroxy

I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
399
group. The latter finding might indicate that there was no significant interaction between
the hydroxy group and the binding site of UGT2B7.
5. Conclusions
The enzymatic glucuronidation of high-affinity substrates such as longifolol and iso-
longifolol might be prevented by addressing steric factors and stereochemical features.
We conclude that the glucuronidation of substrates can be prevented by introducing bulky
substituents in the vicinity of the reaction center to turn high-affinity substrates into potent
true inhibitors. This is especially important for the design of inhibitors of UGT isoforms,
since functional groups that usually promote water solubility such as hydroxy, thiol,
amino, and carboxy groups are commonly glucuronidated by these enzymes. This study
suggests that it might be possible to overcome this major obstacle by considering stereo-
chemical and steric properties to prevent glucuronidation of these functional groups. Espe-
cially longifolol and its epimer isolongifolol represent suitable lead compounds for the
design of inhibitors, since both epimers displayed very low Kic values—a feature that is
rarely seen with UGT enzymes [17]—and they bear a primary hydroxy group that is read-
ily accessible to synthetic modifications. This could be exploited to synthesize derivatives
of these epimers to obtain detailed structure–activity relationships that provide more
insight in the features that control affinity and glucuronidation rate and that could finally
lead to potent true inhibitors of this UGT isoform. In this context it should be mentioned
that to date no sufficiently UGT isoform-selective and potent inhibitor has been discov-
ered [25]. Such inhibitors are important in the study of drug glucuronidation because they
could be used to identify UGT isoforms that mediate a particular glucuronidation in
human tissues and employed to verify drug–drug interactions [26]. Furthermore, selective
probes can be used to measure the phenotype of genetical polymorphism [24,27]. One
advantage of inhibitors based on a hydrocarbon scaffold such as the terpenoid alcohols
investigated in this study would be that they are not detectable by UV or fluorescence
spectroscopy and therefore would not interfere with the detection and analysis of UV
and fluorescence active substrates. Furthermore, detailed eudismic analysis could provide
novel insights in the UGT-catalyzed glucuronidation reaction at the molecular level. Pres-
ently, we are conducting advanced studies to investigate the intriguing behavior of UGT
isoforms toward stereoisomers and to demonstrate that the principles outlined in this
study can be used for the design of potent true inhibitors of UGT2B7.
References
[1] A. Radominska-Pandya, M. Ouzzine, S. Fournel-Gigleux, J. Magdalou, Methods Enzymol. 400 (2005) 116–147.
[2] P.G. Wells, P.I. Mackenzie, J.R. Chowdhury, C. Guillemette, P.A. Gregory, Y. Ishii, A.J. Hansen, F.K.
Kessler, P.M. Kim, N.R. Chowdhury, J.K. Ritter, Drug Metab. Dispos. 32 (2004) 281–290.
[3] M.B. Fisher, M.F. Paine, T.J. Strelevitz, S.A. Wrighton, Drug Metab. Rev. 33 (2001) 273–297.
[4] C.D. King, G.R. Rios, M.D. Green, T.R. Tephly, Curr. Drug Metab. 1 (2000) 143–161.
[5] I.S. Owens, N.K. Basu, R. Banerjee, Methods Enzymol. 400 (2005) 1–22.
[6] J.K. Ritter, F. Chen, Y.Y. Sheen, H.M. Tran, S. Kimura, M.T. Yeatman, I.S. Owens, J. Biol. Chem. 267
(1992) 3257–3261.
[7] G. Monaghan, D.J. Clarke, S. Povey, C.G. See, M. Boxer, B. Burchell, Genomics 23 (1994) 496–499.
[8] R.H. Tukey, C.P. Strassburg, Annu. Rev. Pharmacol. Toxicol. 40 (2000) 581–616.
[9] D. Turgeon, J.S. Carrier, S. Chouinard, A. Belanger, Drug Metab. Dispos. 31 (2003) 670–676.

400
I. Bichlmaier et al. / Bioorganic Chemistry 35 (2007) 386–400
[10] G.E. Kuehl, J.W. Lampe, J.D. Potter, J. Bigler, Drug Metab. Dispos. 33 (2005) 1027–1035.
[11] G.J. Kilpatrick, T.W. Smith, Med. Res. Rev. 25 (2005) 521–544.
[12] A.N. Stone, P.I. Mackenzie, A. Galetin A, J.B. Houston, J.O. Miners, Drug Metab. Dispos. 31 (2003)
1086–1089.
[13] M.J. Sorich, J.O. Miners, R.A. McKinnon, P.A. Smith, Mol. Pharmacol. 65 (2004) 301–308.
[14] M.J. Sorich, R.A. McKinnon, J.O. Miners, D.A. Winkler, P.A. Smith, J. Med. Chem. 47 (2004) 5311–5317.
[15] I. Bichlmaier, A. Siiskonen, M. Finel, J. Yli-Kauhaluoma, J. Med. Chem. 49 (2006) 1818–1827.
[16] I. Bichlmaier, A. Siiskonen, M. Kurkela, M. Finel, J. Yli-Kauhaluoma, Biol. Chem. 387 (2006) 407–416.
[17] J.A. Williams, R. Hyland, B.C. Jones, D.A. Smith, S. Hurst, T.C. Goosen, V. Peterkin, J.R. Koup, S.E. Ball,
Drug Metab. Dispos. 32 (2004) 1201–1208.
[18] I. Agranat, H. Caner, A. Caldwell, Nat. Rev. Drug Discov. 1 (2002) 753–768.
[19] R. Crossley, Chirality and the Biological Activity of Drugs, CRC Press, Boca Raton, 1995.
[20] M. Eichelbaum, B. Testa, A. Somogyi, Stereochemical Aspects of Drug Action and Disposition, Springer-
Verlag, Berlin, 2003.
[21] C.S. Sell, A Fragrant Introduction to Terpenoid Chemistry, The Royal Society of Chemistry, Cambridge,
2003.
[22] R.A. Copeland, Evaluation of Enzyme Inhibitors in Drug Discovery, John Wiley & Sons, Hoboken, 2005.
[23] Y. Cheng, W.H. Prusoff, Biochem. Pharmacol. 22 (1973) 3099–3108.
[24] M.H. Court MH, Methods Enzymol. 400 (2005) 104–116.
[25] K. Grancharov, Z. Naydenova, S. Lozeva, E. Golovinsky, Pharmacol. Ther. 89 (2001) 171–186.
[26] U. Ebert, N.Q. Thong, R. Oertel, W. Kirch, Eur. J. Clin. Pharmacol. 56 (2000) 299–304.
[27] M.H. Court, Q. Hao, S. Krishnaswamy, T. Bekaii-Saab, A. Al-Rohaimi, L.L. von Moltke, D.J. Greenblatt,
J. Pharmacol. Exp. Ther. 310 (2004) 656–665.
[28] P. Vyas, U.R. Nayak, Indian J. Chem. B 20 (1981) 337–339.
[29] P.K. Jadhav, J.V.N. Vara Prasad, H.C. Brown, J. Org. Chem. 50 (1985) 3203–3206.
[30] V. Dimitrov, G. Hopp-Rentsch, A. Linden, M. Hesse, Helv. Chim. Acta 86 (2003) 106–121.
[31] M. Kurkela, J.A. Garcia-Horsman, L. Luukkanen, S. Mo¨rsky, J. Taskinen, M. Baumann, R. Kostiainen, J.
Hirvonen, M. Finel, J. Biol. Chem. 278 (2003) 3536–3544.
[32] S. Kaivosaari, J.S. Salonen, J. Mortensen, J. Taskinen, Anal. Biochem. 292 (2001) 178–187.