Isoform-selective inhibition of the human UDP-glucuronosyltransferase 2B7 by isolongifolol derivatives

Article abstract of DOI:10.1021/jm061204e

A set of 48 derivatives of the tricyclic sesquiterpenol alcohol isolongifolol was synthesized. The set comprised homochiral and diastereomeric alcohols, amines, chlorohydrins, as well as carboxylic acids, phosphonic acids, and their corresponding esters. The absolute configuration of the epimeric compounds was assigned by 2D NMR experiments [gradient heteronuclear single quantum correlation (gHSQC) and gradient nuclear Overhauser enhancement spectroscopy (gNOESY)] in agreement with crystallographic data. The tricyclic derivatives were assessed as inhibitors of the human UDP-glucuronosyltransferase (UGT) 2B7. The phenyl-substituted secondary alcohol 26b was the best inhibitor in this series and its competitive inhibition constant was 18 nM. Compound 26b was not glucuronidated by UGT2B7 and other hepatic UGT enzymes, presumably due to the high steric hindrance exerted by its bulky phenyl substituent. Its inhibitory activity toward 14 other UGT isoforms of subfamily 1A and 2B was determined, and the data indicated that the tricyclic secondary alcohol 26b was highly selective for UGT2B7 (true selectivity > 1000).

Full text of DOI:10.1021/jm061204e

J. Med. Chem. 2007, 50, 2655-2664
2655
Isoform-Selective Inhibition of the Human UDP-glucuronosyltransferase 2B7 by Isolongifolol
Derivatives
Ingo Bichlmaier,† Mika Kurkela,‡ Tanmaya Joshi,†,§ Antti Siiskonen,† Tobias Ru¨ffer,| Heinrich Lang,| Bohumila Suchanova´,†
Mikko Vahermo,† Moshe Finel,‡ and Jari Yli-Kauhaluoma*,†
DiVision of Pharmaceutical Chemistry, Faculty of Pharmacy, and Drug DiscoVery and DeVelopment Technology Center, UniVersity of Helsinki,
FI-00014 Helsinki, Finland, Department of Chemistry, Indian Institute of Technology, Kanpur 208016 (UP), India, and Lehrstuhl fu¨r
Anorganische Chemie, Institut fu¨r Chemie, Technische UniVersita¨t Chemnitz, D-09111 Chemnitz, Germany
ReceiVed October 16, 2006
A set of 48 derivatives of the tricyclic sesquiterpenol alcohol isolongifolol was synthesized. The set comprised
homochiral and diastereomeric alcohols, amines, chlorohydrins, as well as carboxylic acids, phosphonic
acids, and their corresponding esters. The absolute configuration of the epimeric compounds was assigned
by 2D NMR experiments [gradient heteronuclear single quantum correlation (gHSQC) and gradient nuclear
Overhauser enhancement spectroscopy (gNOESY)] in agreement with crystallographic data. The tricyclic
derivatives were assessed as inhibitors of the human UDP-glucuronosyltransferase (UGT) 2B7. The phenyl-
substituted secondary alcohol 26b was the best inhibitor in this series and its competitive inhibition constant
was 18 nM. Compound 26b was not glucuronidated by UGT2B7 and other hepatic UGT enzymes, presumably
due to the high steric hindrance exerted by its bulky phenyl substituent. Its inhibitory activity toward 14
other UGT isoforms of subfamily 1A and 2B was determined, and the data indicated that the tricyclic
secondary alcohol 26b was highly selective for UGT2B7 (true selectivity >1000).
Introduction
to possess broad substrate selectivities so that they can detoxify
a wide range of harmful compounds.
The UDP-glucuronosyltransferases (UGTs,a EC 2.4.1.17) are
important enzymes of the human phase II metabolic system.
These enzymes catalyze the transfer of R-D-glucuronic acid
(GlcA) from UDP-glucuronic acid (UDPGlcA) to lipophilic
substrates bearing hydroxy, thiol, amino, sulfonamide, or
carboxy groups and even enolate moieties. This conjugation
reaction renders lipophilic endobiotics and xenobiotics more
water-soluble and stimulates their urinary and biliary excretion,
thus preventing their accumulation to harmful levels.1-3 Eighteen
isoforms belonging to either subfamily UGT1 or UGT2 have
been identified, seven of which are assumed to play major roles
in drug metabolism.4,5 Perhaps the single most important UGT
in drug glucuronidation is UGT2B7, whose activity may account
for the transformation of 40% of drugs that are metabolized by
UGT isoforms.5 UGTs are promiscuous enzymes that are
commonly described to possess a high degree of flexibility to
depict their complex and often overlapping substrate selectivi-
ties.6,7 This feature is assumed to be a defining property of
metabolic enzymes such as UGTs, cytochrome P450s, and
glutathione S-transferases. These enzymes apparently evolved
To date, the crystal structure of the membrane-bound UGTs
has not been resolved and the detailed reaction mechanism of
the enzymatic glucuronidation reaction is essentially unknown.
However, the biotransformation proceeds with inversion of
configuration at the anomeric carbon atom of GlcA to afford
exclusively â-D-glucuronides. It is therefore assumed that the
transfer step is concerted and resembles the bimolecular
nucleophilic substitution reaction (SN2). Two examples for
enzyme-catalyzed SN2 reactions that display the defining
transition state with trigonal bipyramidal geometry and penta-
coordinate carbon are given by the haloalkane dehalogenase of
Xanthobacter autotrophicus and S-adenosylmethionine syn-
thetase.8,9 In contrast, members of the N-ribosyltransferase
family were shown to display an SN1-like mechanism, and
several very potent transition-state (TS) analogues for these
enzymes were developed.10,11
Some inhibitors of the UGT-catalyzed conjugation reaction
have been reported, but most of them lack desirable levels of
isoform selectivity and potency.12 Although TS analogues for
UGTs were proposed,13,14 they did not display the high potency
that is generally associated with TS mimics.15 These inhibitors
should therefore not be considered as meaningful representations
of true TS mimics and could be better described as weak
bisubstrate analogues. The lack of potency of these compounds
is probably due to the shortfall to address TS-defining features
such as bond lengths, bond angles, and electrostatic potential
surfaces. In general, true TS analogues are designed on the basis
of kinetic isotope effects, computational approaches, and
crystallographic analysis, none of which has been conducted
for UGT enzymes to date.16 Therefore and despite various
attempts, it has been proven relatively unsuccessful to pursue
the design of inhibitors for UGT enzymes by merely attaching
a UDP-like entity to a lipophilic moiety.17,18 Other attempts were
driven by random screening of commercially available com-
pounds and one screening approach led to the discovery of
* Corresponding author: phone +358 9 19159170; fax +358 9
19159556; e-mail jari.yli-kauhaluoma@helsinki.fi.
† Division of Pharmaceutical Chemistry, Faculty of Pharmacy, University
of Helsinki.
‡ Drug Discovery and Development Technology Center, University of
Helsinki.
§ Indian Institute of Technology.
| Technische Universita¨t Chemnitz.
a Abbreviations: 13C-δppm, chemical shift for 13C in parts per million;
CI95%, 95% confidence interval; gCOSY, gradient correlation spectroscopy;
gHMBC, gradient heteronuclear multiple bond coherence; gHSQC, gradient
heteronuclear single quantum correlation; GlcA, glucuronic acid; gNOESY,
gradient nuclear Overhauser enhancement spectroscopy; 1H-δppm, chemical
shift for 1H in parts per million; IBX, 2-iodoxybenzoic acid; IC50,
concentration of inhibitor that causes 50% inhibition; JHH, homonuclear
coupling constant; Kic, competitive inhibition constant; Km, Michaelis-
Menten constant; mep, mixture of epimers; PCC, pyridinium chlorochro-
mate; TS, transition state; UGT, UDP-glucuronosyltransferase.
10.1021/jm061204e CCC: $37.00 © 2007 American Chemical Society
Published on Web 05/03/2007

2656 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11
Bichlmaier et al.
hecogenin as a potentially isoform-selective inhibitor of
UGT1A4.19 Approaches to develop inhibitors for UGT enzymes
based on rational design have been rarely conducted.
Isoform-selective and potent inhibitors for UGTs and other
metabolic enzymes are valuable tools for pharmacokinetic
studies in drug design and development. They facilitate the
identification of enzymes that are responsible for drug metabo-
lism and can be applied to testing in human tissue. Furthermore,
selective inhibitors can be used to elucidate drug-drug interac-
tions and for the identification of enzyme polymorphism.7
However, the design of isoform-selective inhibitors for UGTs
is difficult because these enzymes display overlapping substrate
selectivities and only a small number of selective substrates for
some UGT isoforms have been described.7 This lack of
specificity prohibits the derivation of pharmacophores for the
design of isoform-selective lead structures. In addition, UGT
enzymes generally display low affinities toward their substrates,
which renders it difficult to identify structural features that
promote high affinity.5 Yet another major obstacle is that the
most common functional groups in organic compounds that
promote water solubility, such as hydroxy, amino, thiol, and
carboxy groups, serve as nucleophiles in the enzymatic glucu-
ronidation reaction. Hence, the incorporation of such groups
bears the risk that the designed inhibitor turns out to be a
substrate. The development of potent inhibitors is also chal-
lenging because the X-ray crystal structure of any of the UGT
isoforms is yet to be resolved. This lack of crystallographic data
prevents the design of inhibitors by computational approaches.
Figure 1. Homochiral isolongifolol derivatives.
catalyzed reaction is significantly controlled by stereochemical
and steric properties. In contrast, the affinity toward UGT2B7
was not influenced by the stereochemistry of the employed
enantiomers and diastereomers.34-36 These findings provided a
rational approach to turn high-affinity substrates into potent
inhibitors by addressing stereochemical and steric features to
prevent enzymatic glucuronidation. Especially the substrate
isolongifolol (1) displayed high affinity toward UGT2B7, and
therefore, this compound was chosen as the lead compound for
this study. In this respect, two sets of isolongifolol derivatives
were synthesized. The first set comprised homochiral homolo-
gous monofunctional alcohols, carboxylic acids, phosphonic
acids, and tertiary amines to identify structural features that are
responsible for promoting affinity toward the enzyme (Figure
1). Phosphonic acids were included because it has not been
reported to date that phosphono groups are glucuronidated by
UGT enzymes. The second set consisted of bifunctional epimeric
secondary alcohols bearing various hydrophilic and lipophilic
functional groups (Figure 2). The epimers differed from each
other only in the configuration in the side chain, designated
C(1′) (the numbering of the derivatives is given in the
Supporting Information). Therefore, the influence of the spatial
arrangement of the hydroxy group and the lipophilic or
hydrophilic substituent on the inhibition level could be assessed.
The compounds in this study were synthesized from the
sesquiterpene (+)-longifolene,20 and they possess the intricate
tricyclo[5.4.0.02,9]undecane framework, which has attracted the
attention of synthetic chemists for decades and for which
numerous total syntheses have been reported [the structure of
(+)-longifolene is shown in the Supporting Information].21-23
Furthermore, the terpene has been used for the preparation of
dilongifolylborane, an effective chiral hydroborating agent,24 and
chiral longifolylferrocenes have been described.25 Biosynthesis
of the secondary metabolite longifolene was reported and a
recent study has investigated the promiscuous enzyme γ-hu-
mulene synthase that catalyzes the formation of a prodigious
52 distinct sesquiterpenes, among them longifolene, from the
common precursor (E,E)-farnesyl pyrophosphate.26,27 The tri-
cyclic sesquiterpene occurs in abundance in the Indian turpentine
oil, which is produced commercially from the oleoresin of
Himalayan pine, Pinus longifolia Roxb. It is used in the perfume
industry owing to its woody odor and as a precursor for various
other scents.20 Despite the general interest of organic chemists
in the tricyclic terpene, its biological activity was rarely studied,
probably due to its low water solubility. In this respect,
longifolene was identified as a lipophilic pollutant discharged
from a Finnish pulp and paper mill and its potential aquatic
toxicity was assessed.28,29 Its biotransformation was reported
and one metabolite, a hydroxyaldehyde, was isolated from rabbit
urine.30 Longifolene was described to be active against Gram-
positive bacteria, various fungi, and epimastigotes of Trypano-
soma cruzi, the causative agent of Chagas disease.31 Derivatives
such as oxime ethers were found to be active against the
mosquito Culex quinquefasciatus, a vector for the West Nile
virus, and oxygenated metabolites were described as inhibitors
of butyrylcholinesterase.32,33
Results and Discussion
The syntheses of the isolongifolol derivatives are described
in detail in the Supporting Information and only selected
synthesis steps are discussed herein. Isolongifolyl aldehyde 32
was an important precursor for the syntheses of the various
compounds (compound numbers are given in Schemes 1 and 2
in the Supporting Information). However, the aldehyde 32 was
unstable even when stored under inert argon atmosphere at -19
°C, and decomposition was also observed upon storage in vacuo.
After storage overnight, decomposition into numerous com-
pounds took place as observed by gas chromatography-mass
spectrometry (GC-MS), and the impurities accounted for ap-
proximately 20% of the initial pure aldehyde. Therefore, a
convenient and fast method for the preparation of 32 was
developed by the use of 2-iodoxybenzoic acid (IBX) as the
oxidizing agent.37,38 The terpenoid aldehyde 32 was obtained
Our ongoing analysis of stereochemical events during the
enzyme-catalyzed glucuronidation reaction has recently shown
that diastereomeric tricyclic sesquiterpenoid alcohols display
high affinity toward UGT2B7. Furthermore, the rate of the UGT-

Isoform-SelectiVe Inhibition of UGT2B7
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2657
enediphosphonate and sodium hydride.42 The saturated alkyl
phosphonates 36 and 38 were synthesized by Pd/C-catalyzed
hydrogenation of the olefinic double bonds of (E)-35 and (E)-
37, respectively.43 The ester bond in alkyl phosphonates was
cleaved by use of bromotrimethylsilane followed by hydrolysis
of the resulting silyl ester.42
The diastereomeric secondary terpenoid alcohols 20a-26b
were prepared by Grignard reaction. Pronounced reduction of
the aldehyde was observed when the Grignard reagent was
employed at <5 equiv. In this case, isolongifolol (1) was
recovered in yields between 20% and 60%. In contrast, when a
solution of the aldehyde was dropped slowly into a solution
containing 5 equiv of Grignard reagent, the side reaction was
minimized and the desired secondary alcohols were obtained
in moderate to good yields. Increasing the amount of Grignard
reagent to >5 equiv did not further improve the yields. The
bifunctional â-hydroxy esters 12a and 12b were synthesized
from 32 by titanocene-promoted Reformatsky addition in
excellent yields.44 Subsequent ester hydrolysis afforded the
corresponding â-hydroxycarboxylic acids 11a and 11b, whereas
reduction of the ester group by LiAlH4 gave the corresponding
diols 18a and 18b in high yields. It should also be mentioned
that some solid compounds could be additionally purified by
sublimation at relatively low temperatures (23-90 °C; Sup-
porting Information). The chlorohydrin derivatives 19a and 19b
sublimed slowly at room temperature (23 °C), which was
observed when the compounds were dried in vacuo after
purification by flash chromatography.
The absolute configuration at C(1′) in the side chain of the
epimeric alcohols 12b, 14a, 16a, 17a, 19b, 20a, 21b, and 22a
was assigned by X-ray crystallography (Supporting Information).
However, by conducting 1D and 2D NMR experimentss1H
NMR, 13C NMR, gNOESY, gHMBC, gCOSY, and gHSQCs
it became evident that the absolute configuration at C(1′) could
be conveniently determined by 1H-13C gHSQC correlation
analysis, consistent with crystallographic data (Figure 3).45
The absolute configuration at C(1′) was assigned by compar-
ing the proton chemical shifts (1H-δppm) of the H-atom in
position 9, designated C(9)H, and that of the H-atoms of the
methyl group at position 14, C(14)H, of one secondary alcohol
to those of its respective epimer. It was observed that when the
Figure 2. Epimeric isolongifolol derivatives.
in quantitative yield and high purity after aqueous workup
without need for further purification. In contrast, when aldehyde
32 was prepared by the use of Dess-Martin reagent or PCC,
purification by flash chromatography was necessary due to
byproduct formation. Owing to the fast and convenient protocol
with IBX, aldehyde 32 was freshly prepared for subsequent
synthesis steps. Tertiary amine 10 was synthesized by reductive
alkylation from aldehyde 32 by use of triethylsilane and an
iridium complex as catalyst.39 The use of sodium cyanoboro-
hydride and sodium triacetoxyborohydride as reducing agents
did not furnish the expected tertiary amine, since the aldehyde
32 was reduced under these conditions and the corresponding
alcohol 1 was isolated as the only product.
The oxirane mep-39 was also an important precursor for the
preparation of the epimeric bifunctional â- and γ-hydroxyphos-
phonates and chlorohydrins. Its synthesis was accomplished
from aldehyde 32 by the use of diiodomethane and methyl-
lithium.40 The bifunctional â- and γ-hydroxyphosphonates were
prepared by ring opening of the monosubstituted oxirane mep-
39 by phosphorus and carbon nucleophiles derived from
dialkylphosphites and methanephosphonates, in the presence of
boron trifluoride etherate.41 The monofunctional R,â-unsaturated
alkyl phosphonates (E)-35 and (E)-37 were prepared from 32
in a Wittig-type reaction by the use of tetraisopropyl methyl-
1
hydroxy group was pointing toward C(9)H, its H-δppm value
1
was ∼2.30 ppm, whereas the H-δppm for C(14)H was ∼0.90
ppm. Vice versa, when the hydroxy group was in proximity of
C(14)H, its 1H-δppm value was shifted downfield to ∼1.10 ppm,
whereas that for C(9)H was ∼2.10 ppm (Figure 3 and Sup-
porting Information). These differences were due to the close
proximity of the hydroxy group and the substituent R to either
C(9)H or C(14)H (2.50-2.83 Å in the crystal structures)
1
resulting in different chemical environments and distinct H-
1
δppm values. The H-δppm of the endo-positioned C(10)H was
influenced accordingly, but the differences were in general not
as pronounced as those for C(9)H and C(14)H. The 13C-δppm
values were not significantly influenced by the spatial arrange-
ment of the substituents, thus facilitating identification of the
corresponding carbon atoms in the tricyclic scaffold. The
orientation of C(8)H to C(1′)H was antiperiplanar in the epimeric
alcohols (dihedral angle φ ) 168°-179°) as can be seen from
the crystal structures, and consistently, the coupling constants
determined from 1H and gHSQC NMR experiments were ∼10
Hz (Supporting Information). However, the tert-butyl-substituted
alcohol 25a displayed significantly smaller 3JHH values of 2.70
Hz in CDCl3 and 2.10 Hz in DMSO-d6 for C(1′)H and C(8)H.
This small 3JHH indicated a synclinal arrangement of C(8)H to

2658 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11
Bichlmaier et al.
Table 1. Inhibition Levels Exerted by Monofunctional Derivatives
% inhibitiona
1.0 µM 0.10 µM
Primary Alcohols
compd
100 µM
10 µM
99
92
97
IC50calc b (µM)
1c
2
3
(E)-3
(Z)-3
4
100
97
100
100
100
96
92
68*
54*
56*
60*
32*
45*
15*
11
15
14
0
0.1
0.5
0.9
0.8
0.7
3
94
92
79*
Carboxylic Acids
5
6
99
96
98
90
57*
39*
29*
23*
9
9
0
5
0
0
1
2
3
3
16
82*
78*
73*
39*
(E)-6
(Z)-6
7
98
82*
Phosphonic Acids
8
92
89
45*
11
7
0
0
0
0
0
12
8
32
41
Figure 3. Absolute configuration at C(1′), determined by 1H-13C
gHSQC correlation analysis. The graph displays 1H-δppm values
(ordinate) and 13C-δppm values (abscissa) for C(9)H (×, 9) and C(14)H
(+, 0). Clear differences in the corresponding 1H-δppm values enabled
the determination of the absolute configuration, in agreement with X-ray
crystallographic data. Small variations in the 13C-δppm values facilitated
the identification of the C-atoms. Deuterated NMR solvents had no
significant influence on either 13C-δppm or 1H-δppm (the graph displays
data from NMR experiments with CDCl3, DMSO-d6, and methanol-
d4). The smallest 1H-δppm value for both C(9)H and C(14)H was
measured for the phenyl-substituted epimeric alcohols 26a and 26b,
revealing the highly shielding character of the phenyl group. Data for
epimers 25a and 25b are not displayed. Detailed gHSQC data along
with sample spectra are given in the Supporting Information.
(E)-8
9
(E)-9
55*
21*
27*
76*
71*
0
Tertiary Amine
14*
10
91
65*
0
5
a Percent inhibition was measured at a constant estriol concentration of
25 µM. The inhibitor concentrations were 0.10, 1.0, 10, and 100 µM.
calc
b IC50
) [(100/% inhibition) - 1] × [I]. The value(s) used for the
calculation is (are) indicated by the asterisk.46 c The IC50 of isolongifolol
(1) was previously determined (97 nM).36
group from the other diastereomeric derivatives due to the high
steric demand of its tert-butyl group.
The rotation along the C(1′)-C(8) σ-bond is impaired due
to steric hindrance exerted by the C(14)-methyl group on the
one side and the endo-C(10)H on the other side of the tricyclo-
[5.4.0.02,9]undecane scaffold. Therefore, the spatial orientation
of the hydroxy group at C(1′) is rather well defined. This highly
impaired rotation might be partly responsible for the great
differences in the measured 1H-δppm values that allowed for the
assignment of the absolute configuration at C(1′) (vide supra).
It should also be mentioned that the secondary alcohols in which
the hydroxy group was pointing toward C(14) possessed higher
Rf values than their corresponding epimers and, therefore, eluted
from the chromatography column first. In the case of the tert-
butyl-substituted derivatives 25a and 25b, however, the com-
pound 25b in which the OH group is pointing toward C(14)
had a smaller Rf value than 25a and eluted later from the column.
The affinity toward UGT2B7 was assessed by measuring the
reduction of the rate of the enzyme-catalyzed estriol glucu-
ronidation (% inhibition) in the presence of each isolongifolol
derivative. The IC50 was calculated from this data and was
applied to estimate the affinity for each terpenoid derivative
(Tables 1 and 2). 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 inhibition, and the enzyme
assays displayed good reproducibility. The percent inhibition
exerted by the substrate isolongifolol (1) was applied as a
reference to indicate high affinity. The competitive inhibition
constant (Kic) of isolongifolol (26 nM) was previously deter-
mined in our laboratory.36
Figure 4. Proposed structure for compound 25a based on 1D and 2D
NMR experiments. NOE correlations for C(1′)OH and C(1′)H are
shown.
C(1′)H (φ ∼ 60°). Furthermore, gHSQC experiments revealed
1
that the H-δppm values of C(14)H for the diastereomers 25a
and 25b were 1.00 and 1.07 ppm, respectively, indicating that
the OH group was in proximity to C(14)H in both epimers. The
1H-δppm values for C(9)H were small for both stereoisomers
(2.03 and 1.90 ppm), indicating that the OH group was distant
from C(9)H in both diastereomeric alcohols. The most signifi-
cant difference in the 1H-δppm was found for the endo-positioned
C(10)H, which was significantly smaller for 25b (1.28 ppm)
compared to that for 25a (2.08 ppm). This very large ∆δppm of
0.80 ppm indicated that the OH group was close to endo-C(10)-
H in compound 25b but not in its epimer 25a. NOE correlations
confirmed the proximity of C(1′)OH to both C(14)H and endo-
C(10)H in 25a. On the basis of these results, the structure and
the absolute configuration at C(1′) for the secondary alcohol
25a are proposed as displayed in Figure 4 (cf. Supporting
Information for a full account). The data showed that the epimer
25a deviated significantly in its spatial orientation of the hydroxy
Table 1 presents the results of the screening assays for the
calc
monofunctional derivatives from Figure 1. The IC50 values
varied significantly (0.1-41 µM) and some general trends were
observed. The introduction of hydrophilic substituents such as
carboxy, phosphono, and amino groups resulted in lower
affinities toward UGT2B7. This finding indicated that these
groups did not exert significant attractive interactions toward

Isoform-SelectiVe Inhibition of UGT2B7
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2659
Table 2. Inhibition Levels Exerted by Bifunctional Derivatives
% inhibitiona
% inhibitiona
compd
100 µM
10 µM
1.0 µM
0.1 µM
IC50calc b (µM)
compd
100 µM
10 µM
1.0 µM
0.1 µM
IC50calc b (µM)
Epimeric â- and γ-Hydroxy Acids and Esters
11a
12a
13a
14a
15a
16a
17a
98
100
81*
91
87
93
53*
72*
8
12
26*
0
0
0
0.9
0.4
11b
12b
13b
14b
15b
16b
17b
97
99
90
89
83*
95
91
89
94
51
76*
10
11
7
9
23*
0
0
0
1
0.3
13
10
18
4
36*
49*
28*
75*
68*
18
43*
51*
36*
73*
70*
8
6
10
28
3
78*
99
21*
15*
4
0
20*
27*
0
0
94
5
4
Epimeric Diols and Chlorohydrins
18a
19a
96
100
96
99
82*
88*
38*
36*
0.2
0.2
18b
19b
99
97
98
98
79*
84*
24*
29*
0.3
0.2
Epimeric Alkyl, Vinyl and Aryl-Substituted Alcohols
20a
21a
22a
23a
24a
25a
26a
100
100
95
100
99
93
99
98
91
89
90
86
66*
82*
79*
74*
72*
57*
53*
13
0.5
0.2
0.3
0.4
0.4
0.8
0.9
20b
21b
22b
23b
24b
25b
26b
94
95
95
97
94
100
96
98
59*
66*
63*
81*
70*
88*
94
11
0.7
0.5
0.6
0.2
0.4
0.2
0.07
35*
27*
31*
17*
14
23*
20*
29*
19*
39*
60*
99
100
96
99
99
94
24
100
97
calc
a Percent inhibition was measured at a constant estriol concentration of 25 µM. The inhibitor concentrations were 0.10, 1.0, 10, and 100 µM. b IC50
[(100/% inhibition) - 1] × [I]. The value(s) used for the calculation is (are) indicated by the asterisk.46
)
the binding site of UGT2B7 and therefore did not promote
affinity. In this respect, the phosphonic acid derivatives exerted
the lowest inhibition levels and (E)-9 displayed the highest
IC50calc of 41 µM. Also, elongation of the side chain resulted in
lower affinity for both alcoholic and acidic derivatives. This
stemmed presumably from the higher flexibility of the longer
side chain, leading to an increased entropy cost during formation
of the enzyme-inhibitor complex. Accordingly, alcohol 4 and
chlorohydrins 19a and 19b, exerted high inhibition levels. The
alkyl, vinyl and aryl-substituted derivatives 20a-26b displayed
calc
in general high affinities as indicated by IC50 values below
1 µM. Therefore, it was concluded that lipophilic substituents
promote affinity toward UGT2B7. It is noteworthy that the
epimers of all compounds displayed highly similar IC50calc values
and it was concluded that the spatial arrangement of the hydroxy
group had no effect on the inhibition level.
calc
carboxylic acid 7 displayed the highest IC50 values within
The best inhibitor was 26b, which displayed a slightly lower
calc
the sets of homologous alcohols and carboxylic acids, respec-
tively. Introduction of olefinic double bonds and the resulting
geometric isomerism had no significant influence on the affinity
because the saturated derivatives 3, 6, 8, and 9 displayed
affinities that were comparable to those of their corresponding
unsaturated analogues (E)-3, (Z)-3, (E)-6, (Z)-6, (E)-8, and (E)-
9. On the basis of these results, it was concluded that there were
no attractive interactions between carboxy, phosphono, and
amino groups to the binding site of UGT2B7. The substrate
isolongifolol was the best inhibitor within the set of monofunc-
tional derivatives. Therefore, it was anticipated that the OH
group at C(1′) promoted affinity toward the enzyme.
To test whether the spatial orientation of the OH group at
C(1′) is crucial for high affinity, various compounds were
synthesized in which this nucleophilic group was retained
(Figure 2). The introduction of a second substituent to the
diastereotopic C(1′) in isolongifolol (1) resulted in two stere-
oisomers, which were related as epimers. The results of the
affinity assays for the epimeric derivatives are displayed in Table
2. The â-hydroxycarboxylic acids, as well as the â- and
IC50
than isolongifolol (1). This indicated that the phenyl
group did not significantly contribute to the affinity toward the
enzyme. It was therefore concluded that the high affinity
stemmed predominantly from hydrophobic interactions between
the tricyclic hydrocarbon scaffold and the binding site of the
enzyme. However, UGT2B7 was able to accommodate even
the bulky phenyl group without loss of affinity. Accordingly,
derivatives 25a and 25b bearing the sterically demanding tert-
butyl group also exerted high affinity levels. In this respect,
the deviant spatial orientation of the hydroxy group in secondary
alcohol 25a did not result in lower affinity toward the enzyme.
Inhibition of the UGT2B7-catalyzed glucuronidation reaction
exerted by compound 26b was assessed by measuring its IC50
value to confirm the apparent high affinity as indicated by the
screening assays. This analysis showed that the phenyl-
substituted derivative was a potent inhibitor of the UGT2B7-
catalyzed estriol glucuronidation as expressed by its low IC50
value of 54 nM (standard deviation ) 3.9, n ) 18; CI95%
49.2-58.7 nM), which was in good agreement with the IC50
)
calc
determined from the screening assays (70 nM). Due to its high
affinity, the reversibility of inhibition was measured to determine
whether or not the compound was a tight-binding, slowly
reversible inhibitor. The reversibility of inhibition was assessed
according to standard procedures by measuring the recovery of
enzymatic activity after a rapid and large dilution (Supporting
Information). The data showed that the sesquiterpenoid alcohol
26b was a rapidly reversible inhibitor, and standard assays using
initial rate measurements could be used for further kinetic
studies.
γ-hydroxyphosphonic acids, displayed the lowest affinities
calc
toward UGT2B7. Furthermore, the IC50
values of these
epimeric â- and γ-hydroxy acids 11a, 11b, 13a, 13b, 15a, and
15b were highly similar to those of the respective monofunc-
tional derivatives. This indicated that the hydroxy group had
no significant influence on the inhibition level. Hydroxy esters
12a, 12b, 14a, 14b, 16a, 16b, 17a, and 17b displayed 1.3-
9.3-fold higher affinities toward the enzyme than their respective
free acids. This might indicate that lipophilic interactions toward
the binding site are favored in agreement with the results for
the monofunctional derivatives. In this respect, carboxylic acid
The mode of inhibitionseither competitive, uncompetitive,
noncompetitive, or mixed-typeswas determined by measuring
the effects of the [S]/Km ratio of estriol on the apparent IC50
values of the terpenoid alcohol 26b.46 As can be seen from
calc
esters 12a and 12b displayed smaller IC50 values than the
phosphonic acid esters. Also, diols 18a and 18b, as well as

2660 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11
Bichlmaier et al.
Figure 6. Inhibition of the UGT-catalyzed glucuronidation of umbel-
liferone and 1-hydroxypyrene by compound 1 (open bars) and inhibitor
26b (solid bars). Mean values and standard deviations are shown (n )
4). The inhibitors were employed at a high concentration of 25 µM,
which corresponds to more than 1000Kic for UGT2B7. In addition to
UGT2B7, which was completely inhibited, the isoform 2B17 displayed
the highest response and was therefore chosen to assess the true
selectivity of inhibitor 26b. The phenyl group had presumably no effect
on the isoform selectivity because the percent inhibition exerted by
26b was similar to that exerted by isolongifolol (1).
Figure 5. Determination of the mode of inhibition by measuring the
effect of [S]/Km ratio on IC50. (A) Data for each IC50 determination;(B)
IC50 values for 26b as a function of the [S]/Km ratio of the reference
substrate estriol (r2 ) 0.987, left ordinate). (---) Results of the control
assays (% inhibition) employing isolongifolol (1) at a fixed concentra-
tion of 0.10 µM at each [S]/Km ratio (right ordinate). The mean values
and the standard deviations are displayed (panel A, n ) 3; panel B, n
) 15 for IC50 values, n ) 3 for control assays). [S]/Km ratios are plotted
on a logarithmic scale for clarity. Compound 26b displayed the expected
behavior for a competitive inhibitor.
Figure 5, the compound displayed competitive, active-site-
directed inhibition because the corresponding IC50 values
increased linearly with increasing concentration of the reference
substrate estriol.
The competitive inhibition constant (Kic) of the secondary
sesquiterpenoid alcohol 26b was determined in order to derive
a kinetic parameter that is independent of the substrate
concentration. Furthermore, Kic was measured to confirm the
high affinity and the mechanism of inhibition displayed by the
tricyclic alcohol as indicated by the studies described above.
The resulting data were fitted by simultaneous nonlinear
regression to the competitive, noncompetitive, mixed-type, and
uncompetitive inhibition models, which were ranked according
to Akaike’s corrected information criterion (results not shown).
On the basis of this analysis, the active-site-directed, competitive
inhibition model was chosen and the resulting Kic was in good
agreement with the expected value, which was calculated from
its IC50 (Cheng-Prusoff equation; cf. Supporting Information).
The very low Kic of 18.4 nM confirmed the high affinity of
compound 26b toward UGT2B7 (Figure 7A).
Figure 7. Determination of Kic for inhibitor 26b assayed with UGT2B7
(A) and UGT2B17 (B). Kic for the UGT2B7-catalyzed estriol glucu-
ronidation was 18.4 nM (standard deviation ) 1.51, n ) 90; CI95%
)
15.3-21.4 nM), which was assayed at inhibitor concentrations of 0.020,
0.10, 0.5, and 2.5 µM. Kic for the UGT2B17-catalyzed scopoletin
glucuronidation was 21.8 µM (standard deviation ) 1.46, n ) 54; CI95%
) 18.9-24.7 nM) and was measured at inhibitor concentrations of 10
and 25 µM (inhibitor concentrations >25 µM resulted in solubility
2B17
problems at high scopoletin concentrations). The true selectivity (Kic
/
Kic2B7) was >1000; formation of the UGT2B7/26b complex was
therefore more than 1000-fold favored over formation of the UGT2B17/
26b complex.
It was determined that the alcohol 26b displayed similar
affinity levels independent of the substrate used. The substrate
independence was assessed by measuring the IC50 of the
secondary sesquiterpenoid alcohol 26b toward five different
substrates of UGT2B7, namely, 1-naphthol, 4-nitrophenol,
scopoletin, epitestosterone, and 4-methylumbelliferone. The
measured IC50 values ranged from 47 to 79 nM (standard
deviation <8.7%, n ) 15) and confirmed that the high inhibition
exerted by 26b was independent of the substrate employed.
Glucuronidation assays using [14C]UDPGlcA revealed that
the secondary alcohol 26b was not glucuronidated by UGT2B7
at detectable rates and was therefore a true inhibitor of this
enzyme. The glucuronidation assays were conducted with the
enzyme at a high concentration of 2.0 mg of protein mL-1 and
an incubation time of 16 h in order to identify even small
amounts of [14C]â-D-glucuronide. The radioactive labeled co-
substrate was used due to the low UV absorption of compound
26b. The results suggest that steric hindrance in the vicinity of
the hydroxy group exerted by the bulky phenyl group prevented
enzyme-catalyzed conjugation of the nucleophilic group. Also,
other hepatic UGT enzymes did not conjugate the inhibitor
because assays with human liver microsomes did not produce
the corresponding [14C]â-D-glucuronide. Interestingly, molecular

Isoform-SelectiVe Inhibition of UGT2B7
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2661
33-5), and epitestosterone (CAS 481-30-1) were obtained from
Sigma (St. Louis, MO). 4-Nitrophenol (CAS 100-02-7), 1-naphthol
(CAS 90-15-3), and umbelliferone (CAS 93-35-6) were from
Aldrich (Schnelldorf, Germany). Radiolabeled [14C]UDPGlcA was
acquired from Perkin-Elmer Life and Analytical Sciences (Boston,
MA). HPLC-grade solvents were used throughout the study. The
recombinant human UGTs were expressed as His-tagged proteins
in baculovirus-infected insect cells as previously described.47 The
cDNA for the human UGT2B17 was a generous gift from Professor
Peter Mackenzie (Flinders Medical Centre, Flinders University,
Bedford Park, South Australia).
modeling showed that the phenyl-substituted derivative 26b was
not able to accommodate the GlcA moiety at its hydroxy group
due to the steric hindrance exerted by the phenyl group in
proximity to the bulky tricyclic scaffold (results not shown).
Isoform selectivity was assessed by measuring the percent
inhibition exerted by compound 26b toward the glucuronidation
reaction catalyzed by 14 different recombinant human UGT
isoforms of subfamilies 1A and 2B. Umbelliferone and 1-hy-
droxypyrene were chosen as reference substrates because these
compounds are glucuronidated by many different UGT enzymes
and can be conveniently analyzed by fluorescence spectroscopy.
The inhibitor 26b was employed at a concentration of 25 µM,
which was significantly higher than its Kic for UGT2B7 (>1000-
fold). The results indicated that inhibitor 26b was highly
isoform-selective for UGT2B7 (Figure 6). Most of the UGT1A
isoforms did not display any significant response to the inhibitor.
In contrast, the activity of the UGT2B isoforms was reduced
by approximately 30-70% by this high concentration of 26b.
In addition to UGT2B7, which was completely inhibited, the
UGT2B17-catalyzed glucuronidation reaction displayed the
highest response to the inhibitors (∼70% inhibition). Isoform
2B17 was therefore chosen to assess the true selectivity of
inhibitor 26b. The Kic value of 26b for UGT2B17 was 21.8
µM, and the true selectivity (Kic2B17/Kic2B7) of this inhibitor for
UGT2B7 was therefore >1000 (Figure 7). The large difference
in Kic values was confirmed by measuring the inhibitory activity
of 26b toward the UGT2B7- and UGT2B17-catalyzed glucu-
ronidation of their common substrates 1-naphthol, umbelliferone,
and 4-methylumbelliferone (results not shown).
Syntheses. Procedures for the syntheses and characterization of
the compounds are given in the Supporting Information.
X-ray Crystal Structures. Crystallographic data for the X-ray
measurements of 12b, 14a, 16a, 17a, 19b, 20a, 21b, and 22a are
given in the Supporting Information. Crystal data were collected
on an Oxford Gemini S diffractometer (Cu-KR radiation, λ )
1.542 48 Å, for 12b, 16a, 17a, 19b, 20a, 21b, and 22a; Mo-KR
radiation, λ ) 0.710 73 Å, for 14a) at 100 K. All structures were
solved by direct methods (SHELXS-97)49 and refined by full-matrix
least-squares methods against F2 (SHELXL-97).50 All non-hydrogen
atoms were refined anisotropically. All hydrogen atom positions,
except O-bonded hydrogen atoms, were refined by use of a riding
model. The positions of O-bonded hydrogen atoms were taken from
the difference Fourier map and refined isotropically. The absolute
structures were determined with respect to the Flack parameter.51
In 19b, the chlorine atoms of the two crystallographic indepen-
dent molecules (19ba and 19bb) are disordered and were refined
on two positions. Occupation factors were determined as follows:
19ba (Cl1 ) 0.531; Cl1A ) 0.469) and 19bb (Cl2 ) 0.484; Cl2A
) 0.516).
In 22a, the dCH2 group of one of the two crystallographic
independent molecules is disordered and was refined disordered
on two positions. Occupation factors were determined as follows:
C18 ) 0.442 and C18A ) 0.558.
Conclusion
The results of this study indicate that compound 26b is a
highly selective inhibitor for the key enzyme involved in drug
glucuronidation, UGT2B7. To our knowledge, no isoform-
selective inhibitor for this isoform has been described to date.
The preferential inhibition of UGT2B7 is remarkable when the
promiscuous character of metabolic enzymes is taken into
account. The tricyclo[5.4.0.02,9]undecane framework is presum-
ably responsible for the high affinity, whereas the phenyl group
does not significantly contribute to the inhibition level. The key
function of the phenyl group is to increase steric hindrance in
the vicinity of the hydroxy group and therefore to decrease the
accessibility to this nucleophilic functionality and to prevent
enzymatic glucuronidation. The hydroxy group has presumably
no significant effect on the formation of the enzyme-inhibitor
complex, indicating that there is no significant attractive
interaction to the binding site of UGT2B7. The OH group is
merely retained to promote the solubility of the compound.
Furthermore, the very low Kic of 18 nM is likewise noteworthy
since UGTs are commonly described as flexible, promiscuous
enzymes that in general do not display low dissociation
constants. This nanomolar Kic of the sesquiterpenoid derivative
26b might indicate specific attractive interactions toward the
binding site of UGT2B7. In this respect, it might be suggested
that the recognition of small guest molecules is predominantly
controlled by mere hydrophobic interactions during the forma-
tion of the encounter complex. Hence, the question arises to
what extent the UGT enzyme eventually recognizes those
chemical groups that serve as nucleophiles in the enzyme-
catalyzed reaction.
CCDC-623771 (12b), CCDC-623772 (14a), CCDC-623773
(16a), CCDC-623774 (17a), CCDC-623775 (19b), CCDC-623776
(20a), CCDC-623777 (21b), CCDC-623778 (22a), CCDC-623779
[(E)-33], and CCDC-623780 [(Z)-3] contain the supplementary
crystallographic data. These data can be obtained free of charge
from The Cambridge Crystallographic Data Centre via www.ccd-
c.cam.ac.uk/data_request/cif.
Inhibitor Screening. Percent inhibition was measured at four
inhibitor concentrations (0.10, 1.0, 10, and 100 µM) with estriol
as the reference substrate (25 µM). The reaction mixture consisted
of phosphate buffer (50 mM, pH 7.4), MgCl2 (10 mM), and
saccharic acid-1,4-lactone (5.0 mM). The concentration of UGT2B7
was 0.10 mg of protein mL-1. Assays in the absence of inhibitor,
blank runs in the absence of cosubstrate, and control assays
employing isolongifolol were included in each assay. The enzyme
reaction was initiated 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 (16000g, 10 min) and aliquots of
the supernatants were subjected to HPLC analysis. Results reflect
a minimum of three replicate determinations.
IC50 Values. The IC50 values of compound 26b were determined
at five concentrations bracketing its apparent IC50 (0.10, 0.25, 1.0,
4.0, and 10 × IC50). Estriol was employed at six [S]/Km ratios (0.5,
1.0, 2.0, 5.0, 10, and 20). The reaction mixture consisted of
phosphate buffer (50 mM, pH 7.4), MgCl2 (10 mM), and saccharic
acid-1,4-lactone (5.0 mM). The enzyme UGT2B7 was employed
at 0.10 mg of protein mL-1. Assays in the absence of inhibitor,
blank runs in the absence of cosubstrate, and control assays with
isolongifolol at a fixed concentration of 0.10 µM were included at
each [S]/Km ratio. The enzyme reaction was initiated after a 5 min
preincubation time at 37 °C by addition of a solution of UDPGlcA
to a final concentration of 5.0 mM. 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.
Experimental Section
Materials. UDPGlcA (trisodium salt, CAS 63700-19-6), sac-
charic acid-1,4-lactone (CAS 61278-30-6), estriol (CAS 50-27-
1), scopoletin (CAS 92-61-5), 4-methylumbelliferone (CAS 90-

2662 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11
Bichlmaier et al.
The mixtures were centrifuged (16000g, 10 min) and aliquots of
the supernatants were subjected to HPLC analysis. The data were
analyzed by nonlinear regression with the two-parameter Hill
equation. Results reflect a minimum of three replicate determina-
tions.
ice-cold perchloric acid (4.0 M) and transfer to ice. The mixtures
were centrifuged (16000 g, 10 min), and aliquots of the supernatants
were subjected to HPLC analysis. For 1-hydroxypyrene assays:
Enzyme reactions were terminated after 20 min at 37 °C by addition
of ice-cold aqueous ZnSO4 (15% by weight), followed by aceto-
nitrile, and transfer to ice. The mixtures were centrifuged (16000g,
10 min), and aliquots of the supernatants were subjected to HPLC
analysis. Results reflect a minimum of four replicate determinations.
Glucuronidation Assays. Formation of glucuronide was assayed
by use of radiolabeled [14C]UDPGlcA.48 A solution of [14C]-
UDPGlcA (400 µL, 196 mCi mmol-1, 0.02 mCi mL-1 in EtOH/
water, 7:1 v/v) was transferred to 2-mL vials and the solvent was
evaporated in vacuo at room temperature. The residue was dissolved
in reaction buffer (90 µL), which consisted of either UGT2B7 (2.0
mg of protein mL-1) or human liver microsomes (100 µg), MgCl2
(10 mM), phosphate buffer (50 mM, pH 7.4), and saccharic acid-
1,4-lactone (5.0 mM). A solution of inhibitor 26b (1.0 mM in 50
vol % DMSO/water) was added to the reaction buffer to a final
saturating concentration of 100 µM (approximately 5500Kic2B7).
After incubation for 16 h at 37 °C, the reaction mixture was
centrifuged (16000 g, 10 min), and aliquots of the supernatants
were subjected to HPLC analysis. Control assays in the presence
of the substrate isolongifolol (100 µM) and blank runs were
included. The detection limit (25 pmol, signal-to-noise ratio ) 10:
1) was determined by subjecting dilutions of reaction buffer
containing isolongifolol [14C]â-D-glucuronide to HPLC analysis.
Results reflect a minimum of two replicate determinations.
HPLC Methods. The HPLC system consisted of the Agilent
1100 series degasser, binary pump, autosampler, thermostated
column compartment, multiple wavelength detector, and fluores-
cence detector (Agilent Technologies, Palo Alto, CA). The resulting
spectra were analyzed with Agilent ChemStation software (rev.
B.01.01). Glucuronidation reaction products were separated and
detected as follows. Estriol â-D-glucuronide: Hypersil BDS-C18
(150 × 4.6 mm; Agilent Technologies, Palo Alto, CA); 35% MeOH
Testing for Reversibility. UGT2B7 at a concentration of 5.0
mg of protein mL-1 was preincubated in buffered solution
(phosphate buffer, 50 mM, pH 7.4) at 37 °C with a concentration
of 26b equivalent to 10 times 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 µM), 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 µL of the reaction buffer to a vial containing
10 µL of perchloric acid (4.0 M). The acidified mixtures were
transferred to ice and centrifuged (16000g, 10 min), and aliquots
of the supernatants were subjected to HPLC analysis. Control assays
in the absence of terpenoid alcohol were included. Results reflect
a minimum of two replicate determinations (Supporting Informa-
tion).
Inhibitory Dissociation Constants. The Kic value for UGT2B7
(0.10 mg of protein mL-1) was determined at six estriol concentra-
tions (2.5, 5.0, 10, 25, 50, and 100 µM), and the inhibitor 26b was
employed at four concentrations (0.020, 0.10, 0.50, and 2.5 µM).
The Kic value for UGT2B17 (0.15 mg protein mL-1) was measured
at six scopoletin concentrations (25, 50, 100, 250, 500, and 1000
µM), and compound 26b was used at two concentrations of 10
and 25 µM. For estriol assays: Enzyme reactions were initiated
after a 5.0 min preincubation at 37 °C by addition of a solution of
UDPGlcA to a final concentration of 5.0 mM. Enzyme reactions
were terminated after 15 min at 37 °C by addition of ice-cold
perchloric acid (4.0 M) and transfer to ice. The mixtures were
centrifuged (16000g, 10 min), and aliquots of the supernatants were
subjected to HPLC analysis. Results reflect a minimum of three
replicate determinations. For scopoletin assays: Enzyme reactions
were initiated after a 5 min preincubation at 37 °C by addition of
a solution of UDPGlcA to a final concentration of 5.0 mM. Enzyme
reactions were terminated after 10 min at 37 °C by addition of
ice-cold perchloric acid (4.0 M) and transfer to ice. The mixtures
were centrifuged (16000g, 10 min), and aliquots of the supernatants
were subjected to HPLC analysis. Results reflect a minimum of
three replicate determinations.
in phosphate buffer (50 mM, pH 3.0); flow rate 1.0 mL min-1
;
detection by fluorescence spectroscopy (excitation at λ 335 nm,
emission at λ 455 nm); retention time 4.2 min. 4-Nitrophenol â-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 min-1; detection by UV spectroscopy (λ
300 nm); retention time 1.3 min. Scopoletin â-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 min-1
(0.0-4.5 min), 0.8 mL f 2.5 mL min-1 (4.5-5.0 min), 2.5 mL
min-1 (5.0-11.0 min), 2.5 f 0.8 mL min-1 (11.0-12 min);
detection by fluorescence spectroscopy (excitation at λ 335 nm,
emission at λ 455 nm); retention time 4.2 min. Epitestosterone â-D-
glucuronide: Chromolith SpeedROD (50 × 4.6 mm); 43% MeOH
Testing for Substrate-Independent Inhibition. IC50 values of
inhibitor 26b were determined by use of five different reference
substrates. Assay conditions were balanced with the substrates at
concentrations that resembled their Km values for UGT2B7:
1-naphthol (300 µM, 0.20 mg of protein mL-1, incubation time 20
min), 4-nitrophenol (500 µM, 0.10 mg of protein mL-1, incubation
time 10 min), scopoletin (320 µM, 0.10 mg of protein mL-1
,
in phosphate buffer (50 mM, pH 3.0); flow rate 2.0 mL min-1
;
incubation time 10 min), epitestosterone (15 µM, 0.25 mg of protein
mL-1, incubation time 20 min), and 4-methylumbelliferone (600
µM, 0.20 mg of protein mL-1, incubation time 20 min). Compound
26b was employed at five concentrations bracketing its IC50 (0.10,
0.25, 1.0, 4.0, and 10 × IC50). Assays were carried out as described
above. Results reflect a minimum of two replicate determinations.
Isoform Selectivity. Umbelliferone (100 µM) was used as
substrate for UGT isoforms 1A1 (0.15 mg of protein mL-1), 1A7
(0.15 mg of protein mL-1), 1A8 (0.11 mg of protein mL-1), 1A9
(0.15 mg of protein mL-1), 1A10 (0.12 mg of protein mL-1), and
2B7 (0.10 mg of protein mL-1). UGTs 1A3 (0.40 mg of protein
mL-1), 1A4 (0.50 mg of protein mL-1), 1A5 (0.11 mg of protein
mL-1), 1A6 (0.10 mg of protein mL-1), 2B4 (0.50 mg of protein
mL-1), 2B10 (0.26 mg of protein mL-1), 2B15 (0.50 mg of protein
mL-1), 2B17 (0.20 mg of protein mL-1), and 2B28 (0.25 mg of
protein mL-1) were assayed with 1-hydroxypyrene (50 µM).
Isolongifolol (1) and inhibitor 26b were employed at concentrations
of 10, 25, and 50 µM. Enzyme reactions were initiated after 5 min
preincubation at 37 °C by addition of a solution of UDPGlcA to a
final concentration of 5.0 mM. For umbelliferone assays: Enzyme
reactions were terminated after 25 min at 37 °C by addition of
detection by UV spectroscopy (λ 246 nm); retention time 5.9 min.
1-Naphthol â-D-glucuronide: Hypersil BDS-C18 (150 × 4.6 mm);
42% MeOH in phosphate buffer (50 mM, pH 3.0); flow rate 1.0
mL min-1; detection by fluorescence spectroscopy (excitation at λ
285 nm, emission at λ 335 nm); retention time 4.9 min. Umbel-
liferone â-D-glucuronide: Chromolith SpeedROD (50 × 4.6 mm);
15% MeOH in phosphate buffer (50 mM, pH 3.0); flow rate 1.5
mL min-1; detection by fluorescence spectroscopy (excitation at λ
316 nm, emission at λ 382 nm); retention time 1.4 min. 4-Methy-
lumbelliferone â-D-glucuronide: Chromolith SpeedROD (50 × 4.6
mm); 20% MeOH in phosphate buffer (50 mM, pH 3.0); flow rate
2.0 mL min-1; detection by fluorescence spectroscopy (excitation
at λ 316 nm, emission at λ 382 nm); retention time 1.3 min.
1-Hydroxypyrene â-D-glucuronide: Hypersil BDS-C18 (150 × 4.6
mm); 40% acetic acid (0.5 vol %) in acetonitrile; flow rate 0.9 mL
min-1; detection by fluorescence spectroscopy (excitation at λ 237
nm, emission at λ 388 nm); retention time 2.3 min. [14C]â-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% f 80% MeOH (3.5-8.0 min), 80% MeOH (8.0-13 min), 80%
f 5% MeOH (13-15 min), 5% MeOH (15-20 min); flow rate

Isoform-SelectiVe Inhibition of UGT2B7
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11 2663
1.0 mL min-1; retention time for isolongifolol [14C]â-D-glucuronide
9.3 min; detection by use of a 9701 HPLC radioactivity monitor
(Reeve Analytical, Glasgow, Scotland).
(17) Said, M.; Noort, D.; Magdalou, J.; Ziegler, J. C.; van der Marel, G.
A.; van Boom, J. H.; Mulder, G. J.; Siest, G. Selective and potent
inhibition of different hepatic UDP-glucuronosyltransferase activities
by ω,ω,ω-triphenylalcohols and UDP derivatives. Biochem. Biophys.
Res. Commun. 1992, 187, 140-145.
Acknowledgment. We thank Sanna Sistonen for conducting
some of the isoform-selectivity assays, Olli Aitio for his help
with the gNOESY experiments, and Sirkku Ja¨ntti for conducting
the LC-MS analyses. This work was supported by a fellowship
from the Finnish Cultural Foundation (IB), a grant from Walter
och Lisi Wahls Stiftelse fo¨r Naturvetenskaplig Forskning, and
the Academy of Finland (Projects 207535 and 210933).
(18) Noort, D.; Coughtrie, M. W.; Burchell, B.; van der Marel, G. A.;
van Boom, J. H.; van der Gen, A.; Mulder, G. J. Inhibition of UDP-
glucuronosyltransferase activity by possible transition-state analogues
in rat-liver microsomes. Eur. J. Biochem. 1990, 188, 309-312.
(19) Uchaipichat, V.; Mackenzie, P. I.; Elliot, D. J.; Miners, J. O.
Selectivity of substrate (trifluoperazine) and inhibitor (amitriptyline,
androsterone, canrenoic acid, hecogenin, phenylbutazone, quinidine,
quinine, and sulfinpyrazone) “probes” for human UDP-glucurono-
syltransferases. Drug Metab. Dispos. 2006, 34, 449-456.
(20) Sell, C. S. A fragrant introduction to terpenoid chemistry; The Royal
Society of Chemistry: Cambridge, U.K., 2003; pp 203-214.
(21) Corey, E. J.; Ohno, M.; Vatakencherry, P. A.; Mitra, R. B. Total
synthesis of D,L-longifolene. J. Am. Chem. Soc. 1961, 83, 1251-
1253.
(22) Volkmann, R. A.; Andrews, G. C.; Johnson, W. S. Novel synthesis
of longifolene. J. Am. Chem. Soc. 1975, 97, 4777-4779.
(23) Schultz, A. G.; Puig, S. The intramolecular diene-carbene cycload-
dition equivalence and an enantioselective Birch reduction-alkylation
by the chiral auxiliary approach. Total synthesis of (()-and (-)-
longifolene. J. Org. Chem. 1985, 50, 915-916.
(24) Jadhav, P. K.; Brown, H. C. Dilongifolylborane: a new effective
chiral hydroborating agent with intermediate steric requirements. J.
Org. Chem. 1981, 46, 2988-2990.
(25) Dimitrov, V.; Linden, A.; Hesse, M. Chiral ferrocenes derived from
(+)-longifolenesdetermination of the configuration by NMR spec-
troscopy and X-ray crystallography. Tetrahedron: Asymmetry 2001,
12, 1331-1335.
(26) Arigoni, D. Stereochemical aspects of sesquiterpene biosynthesis.
Pure Appl. Chem. 1975, 41, 219-245.
(27) Yoshikuni, Y.; Ferrin, T. E.; Keasling, J. D. Designed divergent
evolution of enzyme function. Nature 2006, 440, 1078-1082.
(28) Koistinen, J.; Lehtonen, M.; Tukia, K.; Soimasuo, M.; Lahtipera¨,
M.; Oikari, A. Identification of lipophilic pollutants discharged from
a Finnish pulp and paper mill. Chemosphere 1998, 37, 219-235.
(29) Passino-Reader, D. R.; Hickey, J. P.; Ogilvie, L. M. Toxicity to
Daphnia pulex and QSAR predictions for polycyclic hydrocarbons
representative of Great Lakes contaminants. Bull. EnViron. Contam.
Toxicol. 1997, 59, 834-840.
(30) Ishida, T.; Asakawa, Y.; Takemoto, T. Hydroxyisolongifolalde-
hyde: a new metabolite of (+)-longifolene in rabbits. J. Pharm. Sci.
1982, 71, 965-966.
(31) Rubio, J.; Calderon, J. S.; Flores, A.; Castroa, C. Trypanocidal activity
of oleoresin and terpenoids isolated from Pinus oocarpa. Z. Natur-
forsch. C 2005, 60, 711-716.
(32) Sawaikar, D. D.; Sinha, B.; Hebbalkar, G. D.; Sharma, R. N.;
Patwardhan, S. A. Products active on mosquitoes. Part VII. Synthesis
and biological activity of longifolene derivatives. Indian J. Chem. B
1995, 34, 832-835.
(33) Choudhary, M. I.; Musharraf, S. G.; Nawaz, S. A.; Anjum, S.; Parvez,
M.; Fun, H.-K.; Atta-ur-Rahman. Microbial transformation of (-)-
isolongifolol and butyrylcholinesterase inhibitory activity of trans-
formed products. Bioorg. Med. Chem. Lett. 2005, 13, 1939-1944.
(34) Bichlmaier, I.; Siiskonen, A.; Finel, M.; Yli-Kauhaluoma, J. Stere-
ochemical sensitivity of the human UDP-glucuronosyltransferases
2B7 and 2B17. J. Med. Chem. 2006, 49, 1818-1827.
Supporting Information Available: Degree of purity (elemen-
tal analysis), synthesis procedures, spectroscopic data, X-ray crystal
structures, 1H-13C correlation data, gHSQC and 13C NMR spectra,
and testing for reversibility of inhibition. This material is available
free of charge via the Internet at http://pubs.acs.org.
References
(1) Radominska-Pandya, A.; Ouzzine, M.; Fournel-Gigleux, S.; Magdal-
ou, J. Structure of UDP-glucuronosyltransferases in membranes.
Methods Enzymol. 2005, 400, 116-147.
(2) Wells, P. G.; Mackenzie, P. I.; Chowdhury, J. R.; Guillemette, C.;
Gregory, P. A.; Ishii, Y.; Hansen, A. J.; Kessler, F. K.; Kim, P. M.;
Chowdhury, N. R.; Ritter, J. K. Glucuronidation and the UDP-
glucuronosyltransferases in health and disease. Drug Metab. Dispos.
2004, 32, 281-290.
(3) Fisher, M. B.; Paine, M. F.; Strelevitz, T. J.; Wrighton, S. A. The
role of hepatic and extrahepatic UDP-glucuronosyltransferases in
human drug metabolism. Drug Metab. ReV. 2001, 33, 273-297.
(4) Owens, I. S.; Basu, N. K.; Banerjee, R. UDP-glucuronosyltrans-
ferases: gene structures of UGT1 and UGT2 families. Methods
Enzymol. 2005, 400, 1-22.
(5) Williams, J. A.; Hyland, R.; Jones, B. C.; Smith, D. A.; Hurst, S.;
Goosen, T. C.; Peterkin, V.; Koup, J.; Ball, S. E. Drug-drug
interactions for UDP-glucuronosyltransferase substrates: a pharma-
cokinetic explanation for typically observed low exposure (AUCi/
AUC) ratios. Drug Metab. Dispos. 2004, 32, 1201-1208.
(6) The term catalytic promiscuity has been used to describe the ability
of an enzyme to catalyze an adventitious secondary activity at the
active site responsible for the primary activity. However, in the case
of metabolic enzymes like UGTs, the term has been used to depict
their overlapping substrate selectivities. Cf. Copley, S. D. Enzymes
with extra talents: moonlighting functions and catalytic promiscuity.
Curr. Opin. Chem. Biol. 2003, 7, 265-272.
(7) Court, M. H. Isoform-selective probe substrates for in vitro studies
of human UDP-glucuronosyltransferases. Methods Enzymol. 2005,
400, 104-116.
(8) Lightstone, F. C.; Ya-Jun, Z.; Maulitz, A. H.; Bruice, T. C. Non-
enzymatic and enzymatic hydrolysis of alkyl halides: a haloalkane
dehalogenation enzyme evolved to stabilize the gas-phase transition
state of an SN2 displacement reaction. Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 8417-8420.
(9) Markham, G. D.; Parkin, D. W.; Mentch, F.; Schramm, V. L. A
kinetic isotope effect study and transition state analysis of the
S-adenosylmethionine synthetase reaction. J. Biol. Chem. 1987, 262,
5609-5615.
(10) Schramm, V. L. Enzymatic transition states: thermodynamics,
dynamics and analogue design. Arch. Biochem. Biophys. 2005, 433,
13-26.
(35) Bichlmaier, I.; Siiskonen, A.; Kurkela, M.; Finel, M.; Yli-Kau-
haluoma, J. Chiral distinction between the enantiomers of bicyclic
alcohols by UDP-glucuronosyltransferases 2B7 and 2B17. Biol.
Chem. 2006, 387, 407-416.
(11) Taylor Ringia, E. A.; Schramm, V. L. Transition states and inhibitors
of the purine nucleoside phosphorylase family. Curr. Top. Med.
Chem. 2005, 5, 1237-1258.
(36) Bichlmaier, I.; Kurkela, M.; Siiskonen, A.; Finel, M.; Yli-Kau-
haluoma, J. Eudismic analysis of tricyclic sesquiterpenoid alcohols.
Lead structures for the design of potent inhibitors of the human UDP-
glucuronosyltransferase 2B7. Bioorg. Chem. (submitted).
(37) Ladziata, U.; Zhdankin, V. V. Hypervalent iodine(V) reagents in
organic synthesis. ARKIVOC: Arch. Org. Chem. 2006, ix, 26-58.
(38) Frigerio, M.; Santagostino, M.; Sputore, S. A user friendly entry to
2-iodoxybenzoic acid (IBX). J. Org. Chem. 1999, 64, 4537-4538.
(39) Mizuta, T.; Sakaguchi, S.; Ishii, Y. Catalytic reductive alkylation of
secondary amine with aldehyde and silane by an iridium compound.
J. Org. Chem. 2005, 70, 2195-2199.
(40) Concello´n, J. M.; Cuervo, H.; Fe´rnandez-Fano, R. An improved
preparation of epoxides from carbonyl compounds by using di-
iodomethane/methyllithium: synthetic applications. Tetrahedron
2001, 57, 8983-8987.
(41) Li, Z.; Racha, S; Dan, L.; El-Subbagh, H.; Abushanab, E. A general
and facile synthesis of â- and γ-hydroxyphosphonates from epoxides.
J. Org. Chem. 1993, 58, 5779-5783.
(12) Grancharov, K.; Naydenova, Z.; Lozeva, S.; Golovinsky, E. Natural
and synthetic inhibitors of UDP-glucuronosyltransferase. Pharmacol.
Ther. 2001, 89, 171-186.
(13) Timmers, C. M.; Dekker, M.; Buijsman, R. C.; van der Marel, G.
A.; Ethell, B.; Anderson, G.; Burchell, B.; Mulder, G. J.; van Boom,
J. H. Synthesis and inhibitory effect of a trisubstrate transition state
analog for UDP glucuronosyltransferases. Bioorg. Med. Chem. Lett.
1997, 7, 1501-1506.
(14) Noort, D.; Meijer, E. A.; Visser, T. J.; Meerman, J. H. N.; van der
Marel, G. A.; van Boom, J. H.; Mulder, G. J. Selective inhibition of
glucuronidation by 2,2,2-triphenylethyl-UDP in isolated rat hepato-
cytes: conjugation of harmol, 3,3′,5-triiodothyronine, and N-hydroxy-
2-acetylaminofluorene. Mol. Pharmacol. 1991, 40, 316-320.
(15) Schramm, V. L. Enzymatic transition states and transition state
analogues. Curr. Opin. Struct. Biol. 2005, 15, 604-613.
(16) Schramm, V. L. Enzymatic transition states and transition state analog
design. Annu. ReV. Biochem. 1998, 67, 693-720.

2664 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 11
Bichlmaier et al.
(42) Samadder, P.; Bittman, R.; Byun, H.-S.; Arthur, G. Synthesis and
use of novel ether phospholipid enantiomers to probe the molecular
basis of the antitumor effects of alkyllysophospholipids: Correlation
of differential activation of c-Jun NH2-terminal protein kinase with
antiproliferative effects in neuronal tumor cells. J. Med. Chem. 2004,
47, 2710-2713.
(43) Chandran, S. S.; Frost, J. W. Aromatic inhibitors of dehydroquinate
synthase: Synthesis, evaluation and implications for gallic acid
biosynthesis. Bioorg. Med. Chem. Lett. 2001, 11, 1493-1496.
(44) Parrish, J. D.; Shelton, D. R.; Little, R. D. Titanocene(III)-promoted
Reformatsky additions. Org. Lett. 2003, 5, 3615-3617.
(45) Alternatively, gNOESY experiments could be used for the assignment
of the absolute configuration at C(1′). The NOE correlation between
C(1′)OH and endo-C(10)H was present when the OH group was
pointing toward endo-C(10)H. In contrast, this NOE correlation was
not found in its corresponding epimer in which the OH group was
pointing away from endo-C(10)H. However, these correlations were
often difficult to assign due to signal overlap, and, therefore, gHSQC
analysis was superior because all corresponding signals were resolved.
(46) Copeland, R. A. EValuation of enzyme inhibitors in drug discoVery;
John Wiley & Sons: Hoboken, NJ, 2005.
(47) Kurkela, M.; Garcia-Horsman, A.; Luukkanen, L.; Mo¨rsky, S.;
Taskinen, J.; Baumann, M.; Kostiainen, R.; Hirvonen, J.; Finel, M.
Expression and characterization of recombinant human UDP-glucu-
ronosyltransferases (UGTs). UGT1A9 is more resistant to detergent
inhibition than other UGTs and was purified as an active dimeric
enzyme. J. Biol. Chem. 2003, 278, 3536-3544.
(48) Kaivosaari, S.; Salonen, J. S.; Mortensen, J.; Taskinen, J. High-
performance liquid chromatographic method combining radiochemi-
cal and ultraviolet detection for determination of low activities of
uridine 5′-diphosphate-glucuronosyltransferase. Anal. Biochem. 2001,
292, 178-187.
(49) Sheldrick, G. M. Phase annealing in SHELX-90: direct methods for
larger structures. Acta Crystallogr., Sect. A 1990, 46, 467-473.
(50) Sheldrick, G. M. SHELXL-97, A program for crystal structure
refinement; Go¨ttingen, Germany, 1996.
(51) Flack, H. D. On enantiomorph-polarity estimation. Acta Crystallogr.,
Sect. A 1983, 39, 876-881.
JM061204E