Potent taccalonolides, AF and AJ, inform significant structure-activity relationships and tubulin as the binding site of these microtubule stabilizers

Article abstract of DOI:10.1021/ja209045k

The taccalonolides are a class of microtubule stabilizing agents isolated from plants of the genus Tacca. In efforts to define their structure-activity relationships, we isolated five new taccalonolides, AC-AF and H2, from one fraction of an ethanol extract of Tacca plantaginea. The structures were elucidated using a combination of spectroscopic methods, including 1D and 2D NMR and HR-ESI-MS. Taccalonolide AJ, an epoxidation product of taccalonolide B, was generated by semisynthesis. Five of these taccalonolides demonstrated cellular microtubule-stabilizing activities and antiproliferative actions against cancer cells, with taccalonolide AJ exhibiting the highest potency with an IC ₅₀ value of 4.2 nM. The range of potencies of these compounds, from 4.2 nM to >50 μM, for the first time provides the opportunity to identify specific structural moieties crucial for potent biological activities as well as those that impede optimal cellular effects. In mechanistic assays, taccalonolides AF and AJ stimulated the polymerization of purified tubulin, an activity that had not previously been observed for taccalonolides A and B, providing the first evidence that this class of microtubule stabilizers can interact directly with tubulin/microtubules. Taccalonolides AF and AJ were able to enhance tubulin polymerization to the same extent as paclitaxel but exhibited a distinct kinetic profile, suggesting a distinct binding mode or the possibility of a new binding site. The potencies of taccalonolides AF and AJ and their direct interaction with tubulin, together with the previous excellent in vivo antitumor activity of this class, reveal the potential of the taccalonolides as new anticancer agents.

Full text of DOI:10.1021/ja209045k

COMMUNICATION
pubs.acs.org/JACS
Potent Taccalonolides, AF and AJ, Inform Significant
StructureÀActivity Relationships and Tubulin as the Binding Site of
These Microtubule Stabilizers
Jing Li,† April L. Risinger,†,‡ Jiangnan Peng,†,‡ Zhongliang Chen,|| Lihong Hu,|| and Susan L. Mooberry*,†,‡,§
†Department of Pharmacology, ‡Cancer Therapy & Research Center, and §Department of Medicine, University of Texas Health Science
Center at San Antonio, San Antonio, Texas 78229, United States
Shanghai Research Center for Modernization of Traditional Chinese Medicine, Shanghai Institute of Materia Medica,
Chinese Academy of Sciences, Shanghai 201203, China
S Supporting Information
b
that bind within the taxane site, cabazitaxel and the epothilone
ABSTRACT: The taccalonolides are a class of microtubule
stabilizing agents isolated from plants of the genus Tacca. In
efforts to define their structureÀactivity relationships, we
isolated five new taccalonolides, ACÀAF and H2, from one
fraction of an ethanol extract of Tacca plantaginea. The
structures were elucidated using a combination of spectro-
scopic methods, including 1D and 2D NMR and HR-ESI-
MS. Taccalonolide AJ, an epoxidation product of taccalo-
nolide B, was generated by semisynthesis. Five of these
taccalonolides demonstrated cellular microtubule-stabiliz-
ing activities and antiproliferative actions against cancer
cells, with taccalonolide AJ exhibiting the highest potency
with an IC50 value of 4.2 nM. The range of potencies of these
compounds, from 4.2 nM to >50 μM, for the first time provides
the opportunity to identify specific structural moieties
crucial for potent biological activities as well as those that
impede optimal cellular effects. In mechanistic assays, taccalo-
nolides AF and AJ stimulated the polymerization of purified
tubulin, an activity that had not previously been observed for
taccalonolides A and B, providing the first evidence that this
class of microtubule stabilizers can interact directly with
tubulin/microtubules. Taccalonolides AF and AJ were able
to enhance tubulin polymerization to the same extent as
paclitaxel but exhibited a distinct kinetic profile, suggesting a
distinct binding mode or the possibility of a new binding
site. The potencies of taccalonolides AF and AJ and their
direct interaction with tubulin, together with the previous
excellent in vivo antitumor activity of this class, reveal the
potential of the taccalonolides as new anticancer agents.
ixabepilone, have been approved for clinical use. These drugs can
circumvent some but not all of the limitations of the first- and
second-generation microtubule stabilizers.3,4
In our search for microtubule-disrupting agents, we identified
a new class of microtubule stabilizers, the taccalonolides, from
the tropical plant Tacca chantrieri.5 The taccalonolides have
highly acetylated pentacyclic steroidal skeletons and are structu-
rally distinct from other microtubule stabilizers. The cellular
effects of the most abundant taccalonolide, A (1), are almost
identical to the effects of paclitaxel. They both increase the
density of cellular microtubules, cause the formation of aberrant
mitotic spindles leading to mitotic arrest and apoptosis, and have
excellent antitumor efficacy in vivo.5,6 However, multiple experi-
mental approaches were unsuccessful in identifying a direct
interaction between 1 and tubulin/microtubules.7,8 Other stud-
ies showed that the taccalonolides A, E, B (2), and N can
circumvent clinically relevant forms of taxane resistance, includ-
ing expression of P-glycoprotein and βIII tubulin, suggesting that
the taccalonolides offer advantages over existing microtubule-
targeting agents.6 In a recent study, we isolated three additional
taccalonolides with antimitotic and microtubule-stabilizing
properties (designated as Z, AA, and AB) and began to identify
structureÀactivity relationships (SARs) for this class of com-
pounds.9 This communication focuses on the isolation and char-
acterization of five additional naturally occurring taccalonolides,
designated as AC (3), AD(4), AE(5), AF(6), andH2(8), andone
semisynthetic product, AJ (7), which allow significant refinement
of the SARs for this class of molecules. Additionally, the single-
digit nanomolar potency of 7 gave us the ability to detect a direct
interaction between a taccalonolide and tubulin.
The structures of the taccalonolides (Figure 1) were solved
using 1D and 2D NMR spectra as well as high-resolution elec-
trospray ionization mass spectrometry (HR-ESI-MS) data. All of
the compounds were obtained as white powders. The 1H spectra
of 3À6 and 8 revealed signals for five methyls, four acetyl-group
methyls, four oxygenated methines, two epoxide methines, and
one olefinic methine (Supplemental Table 1), which are char-
acteristic of the taccalonolides. The molecular formula of com-
pound 3 was determined to be C36H46O16 by HR-ESI-MS
or over 40 years, drugs that target microtubules have been
F
important in oncology. In 2010, two of the four new drugs
approved by the FDA for the treatment of cancer targeted
microtubules, suggesting that microtubules remain a valuable
target for anticancer drugs. Paclitaxel, the first microtubule
stabilizer identified, was isolated from Taxus brevifolia. Although
it has achieved significant clinical success, the limitations of
paclitaxel and the second-generation analogue docetaxel include
intrinsic and acquired multidrug resistance and dose-limiting
toxicities, prompting the development of new classes of micro-
tubule-stabilizing drugs.1,2 Recently, two microtubule stabilizers
Received: September 26, 2011
Published: November 01, 2011
r
2011 American Chemical Society
19064
dx.doi.org/10.1021/ja209045k J. Am. Chem. Soc. 2011, 133, 19064–19067
|

Journal of the American Chemical Society
COMMUNICATION
in Figure 1 was determined and the trivial name taccalonolide AD
assigned.
Compound 5 gave pseudomolecular ions at m/z 719 [M + H]+,
736 [M + NH4]+, and 741 [M + Na]+ in LCÀESI-MS. HR-ESI-
MS of 5 showed an [M + H]+ ion at m/z 719.2944 (calcd
719.2915), corresponding to a molecular formula of C36H46O15,
which differs from that of 1 by one additional O atom. The 1H
and 13C NMR spectra of 5 were very similar to those of 1. The
only difference was the occurrence of a signal for an additional
hydroxyl group at 5.01 ppm instead of H7 at ca. 4 ppm. The
location of the additional hydroxyl group at C7 to form a geminal
diol was deduced by the chemical shift of C7 at 92.4 ppm and
confirmed by the HMBCs between 7-OH at 5.01 ppm and C6/
C7/C8, between 7-OH at 3.64 ppm and C6/C7, and between
H8 and C7. Compound 5 represents a rare example of geminal
diol functionality for natural products and is designated as
taccalonolide AE. It was stable during our experiments.
The molecular formula C36H46O15 was established for com-
pound 6 by the [M + H]+ ion at m/z 719.2911 (calcd 719.2915)
in the HR-ESI-MS spectrum, indicating the presence of one
1
more O than is found in 1. The H and 13C NMR data for 6,
which were fully assigned through 2D NMR experiments, closely
resembled those of 1. The only difference was the absence of
resonances for the C22ÀC23 olefin and the appearance of
signals for one oxygenated methine (δH = 3.29 ppm and δC =
65.9 ppm) and one quaternary carbon at 92.2 ppm. In accor-
dance with the molecular formula of 6, the additional oxygen
should be present as an epoxide group at C22/C23. This was
confirmed by the HMBCs between H21 and C20/C22, between
H22 and C20/C21, and between H28 and C23/C24. The
relative configuration of this epoxide group was deduced from
Figure 1. Taccalonolide structures.
analysis (calcd m/z 735.2864, exptl 735.2893), which contains
two more O atoms than taccalonolide A (1). The proton NMR
spectrum of 3 showed a singlet resonance for the C21 methyl
group, indicating a quaternary C20, which requires an additional
substituent at C20 in comparison with 1. This was confirmed by
the heteronuclear multiple-bond correlations (HMBCs) be-
tween H22 and C17/C20/C24 and between H21 and C17/
C20/C22. The substitution was proposed to be a hydroperoxyl
group on the basis of the chemical shift of C20 at 84.5 ppm and
the two additional oxygen atoms required by the molecular
formula. Additional support for the hydroperoxyl group was
provided by the downfield shift of C20 relative to that of
taccalonolide W, which possesses a hydroxyl group at C20.10
All of the taccalonolides discovered to date possess 18β-Me and
16β-H configurations. The α-orientation of the hydroperoxyl
group was deduced by the nuclear Overhauser effect (NOE)
correlations between Me21 and H16/Me18. The other signals of
3 are similar to those of 1, so the structure of 3 depicted in Figure 1
was proposed and the trivial name taccalonolide AC given.
Liquid chromatography (LC)ÀESI-MS of 4 showed pseudo-
molecular ions at m/z 701 [M + H]+, 718 [M + NH4]+, and 723
[M + Na]+, indicating a mass 2 Da less than that of 1. The
molecular formula of 4 was determined to be C36H44O14 by HR-
ESI-MS analysis (calcd m/z 701.2807, exptl 701.2787). The 13C
NMR spectrum showed two additional olefinic carbon signals at
143.9 and 127.3 ppm, suggesting the presence of one more
double bond in 4 than in 1. The location of this double bond
between C5 and C6 was deduced by the HMBCs between C5
and H1/H3/H19. In addition, an enol hydroxyl group at the C6
position was determined by the HMBCs between the hydroxyl
proton at 6.26 ppm and C6/C7. The shift of the ketone group to
C7 was evidenced by the HMBC between C7 at 190.3 ppm and
H8/H9/6-OH. Therefore, the structure of compound 4 depicted
3
the coupling constants. The small JH20/H22 coupling constant
requires an equatorial (β-oriented) H22, so the epoxide group is
α-oriented. This epoxide-containing taccalonolide was given the
trivial name taccalonolide AF.
Compound 8 showed the same pseudomolecular ions as 4, at
m/z 701 [M + H]+, 718 [M + NH4]+, and 723 [M + Na]+. The
molecular formula was determined to be C36H44O14 by HR-ESI-
MS analysis (calcd m/z 701.2809, exptl 701.2813). Similar to 4,
compound 8 also showed carbon signals for an additional double
bond at 140 and 126 ppm in its 13C NMR spectrum. The location
of this additional double bond was determined to be C7/C8 on
the basis of the HMBCs between H9/H14 and C7 and between
H9/H14/H15 and C8. A hydroxyl group resonance at 6.27
ppm showed HMBCs with C6/C7/C8, suggesting that this
hydroxyl group is located at C6. Thus, the structure of 8 was
determined to be the enol of the previously reported taccalono-
lide H (9).11 Since 8 is stable as the enol, the trivial name
taccalonolide H2 was given to distinguish it from taccalonolide H
containing the ketone.
Compounds 1À5 and 8 were isolated in sufficient quantities
to allow accurate determination of their biological activities.
Taccalonolide AF (6) was present at low abundance in the
extract, and more material was semisynthesized by epoxidation of
the abundant compound 1 using dimethyldioxirane.12 The
reaction was carried out quantitatively under very mild condi-
tions and gave 6 as the single product. No β-epoxide isomer was
observed. After evaporation of the solvent and reagent, quantities
of 6 sufficient for biological studies were obtained. Because of the
potent activity of 6, taccalonolide AJ (7), the epoxidation
product of taccalonolide B (2), was also produced using the
same method.
19065
dx.doi.org/10.1021/ja209045k |J. Am. Chem. Soc. 2011, 133, 19064–19067

Journal of the American Chemical Society
COMMUNICATION
Table 1. Antiproliferative Potencies of the Taccalonolides
proliferation. Compound 3 appeared to be relatively stable under
the assay conditions, reducing the possibility that its inactivity
was due to rapid decomposition.
compound
IC50 (nM)a
The large number of naturally occurring taccalonolides dis-
tinguishes them as one of only a few natural product classes
where the SARs can be identified without significant synthetic
manipulation. Our previous studies with other taccalonolides
provided initial indications of SARs.6,9 A bulky isovalerate at the
C1 position, found in taccalonolide T, was associated with ex-
cellent potency relative to 1.9 Additionally, upon comparison of
1, 2, and taccalonolides E and N, the presence or absence of
acetoxy groups at C11 or C15 did not have a major effect on
potency.6 When comparing the potency of other taccalonolides,
however, the presence or absence of the C11 acetoxy group and
the hydrolysis of the C15 acetate group were contingent on
moieties present in other parts of the molecule, suggesting
interrelations across the northern and southern regions.9 In the
present study, 8 showed a 7.4-fold increase in potency relative to
1. The only difference between these two taccalonolides is the
presence of an additional double bond in 8 at C7/C8. The
location of this double bond is important, since a double bond at
C5/C6 (as in 4) does not provide the same increase in potency.
When a hydroxyl group was added to C7 of 1 to form the rare
geminal diol in 5, the potency was also unchanged (Table 1).
Surprisingly, 3, which differs from 1 only by an additional hydro-
peroxyl group at C20, showed no antiproliferative activity at con-
centrations up to 50 μM, suggesting the importance of substitu-
ents at this site. In contrast, 6 and 7, which differ from 1 and 2 by
conversion of the C22ÀC23 double bond to an epoxide group,
increased the activity 234À743-fold, giving the most potent
taccalonolides identified to date (Table 1). These results clearly
suggested that an epoxide moiety at C22/C23 provides optimal
potency, and we will test this further with additional taccalono-
lides in future studies. Taken together, these results highlight the
importance of the C20,C22ÀC23 region of the taccalonolide
molecule and suggest that this region plays a central role in its
interactions with tubulin.
Previously, all classes of microtubule-stabilizing agents that
increase the density of cellular microtubules except the taccalo-
nolides were shown to stimulate the polymerization of purified
tubulin in biochemical preparations, demonstrating a direct
interaction with tubulin.7 Multiple studies evaluating the inter-
actions of 1 and 2 with purified tubulin, microtubule protein, or
cellular lysates containing tubulin and microtubule-associated
proteins suggested that these taccalonolides are not able to bind
directly to tubulin.7,8 However, these studies were all performed
with 1 and 2, which have antiproliferative potencies in the low
micromolar range. In this study we generated 7, the first
taccalonolide to inhibit cellular proliferation at concentrations
less than 5 nM, the range of potency of microtubule stabilizers in
clinical use. In contrast to all previous studies with 1 and 2, we
found that 7 enhanced both the rate and extent of purified
tubulin polymerization (Figure 3). A direct comparison of the
effects of 7 and paclitaxel on tubulin polymerization at equimolar
concentrations showed that the two drugs are able to polymerize
tubulin to the same extent. However, a lag in the time required for
polymerization to occur was observed in the presence of 7, as
opposed to the immediate tubulin assembly that occurs with
paclitaxel. This suggests that 7 may interact with tubulin/micro-
tubules in a manner distinct from that for paclitaxel or bind to a
different binding site. Studies with the other highly potent
taccalonolide, 6, indicated that it is also able to enhance the
taccalonolide A (1)
taccalonolide B (2)
taccalonolide AC (3)
taccalonolide AD (4)
taccalonolide AE (5)
taccalonolide AF (6)
taccalonolide AJ (7)
taccalonolide H2 (8)
paclitaxel
5380 ( 230
3120 ( 180
>50000
3480 ( 230
5010 ( 210
23 ( 3
4.2 ( 0.3
730 ( 20
1.0 ( 0.1
a Concentrations of taccalonolides that caused 50% inhibition of cellular
proliferation, as measured in HeLa cells using the SRB assay (n = 3).
Figure 2. Effects of the taccalonolides on interphase microtubules.
HeLa cells were treated with vehicle or a taccalonolide for 18 h. Shown
are representative images of cells treated with (A) vehicle, (B) 1, (C) 2,
(D) 3, (E) 4, (F) 5, (G) 6, (H) 7, and (I) 8. Concentrations equal to 5
times the IC50 were used for all of the taccalonolides except 3, which was
added at 20 μM. Microtubules were visualized by indirect immuno-
fluorescence using a β-tubulin antibody.
Except for 3, all of the taccalonolides that were isolated or
synthesized inhibited the proliferation of and exhibited cytotoxi-
city toward HeLa cells, with IC50 values ranging from 4.2 nM for
7 to 5 μM for 1 and 5 (Table 1). In contrast, concentrations of 3
up to 50 μM had no effect on cellular proliferation. Significant
increases in cellular microtubule density and microtubule bund-
ling occurred at concentrations 5-fold greater than the IC50 for
each of the active compounds, consistent with the microtubule-
stabilizing activities of other taccalonolides (Figure 2). In con-
trast, concentrations of 3 up to 20 μM showed no effects on
cellular microtubules (Figure 2). In addition to increasing the
density of interphase microtubules, all of these taccalonolides
except 3 caused HeLa cells to arrest in the G2/M phase of the cell
cycle (Supplemental Figure 2) with multiple aberrant mitotic
spindles (Supplemental Figure 3), which are additional pheno-
types associated with microtubule-stabilizing agents. No change
in cell cycle distribution was observed in cells treated with a
20 μM concentration of 3 (Supplemental Figure 2). In immuno-
fluorescence studies, 3-treated cells in mitosis exhibited normal
bipolar spindles (Supplemental Figure 3) and DNA align-
ment (data not shown). These data are consistent with the
inability of 3 to disrupt interphase microtubules and inhibit cell
19066
dx.doi.org/10.1021/ja209045k |J. Am. Chem. Soc. 2011, 133, 19064–19067

Journal of the American Chemical Society
COMMUNICATION
Additionally, in intact cells, microtubule-associated proteins
modulate the effects of these agents. The current findings
demonstrate that the potent taccalonolides 6 and 7 interact
directly with tubulin. Future studies to identify the binding site
and nature of the interactions of the taccalonolides with tubulin
and/or microtubules will be performed. It is interesting to
speculate on the possibility of a third microtubule stabilizer
binding site. The potent biological activities of 6 and 7, their
interaction with tubulin, and the excellent in vivo antitumor
effects of other taccalonolides, together with the proven value of
microtubule-stabilizing drugs, clearly reveal the potential of the
taccalonolides as anticancer agents.
’ ASSOCIATED CONTENT
S
Supporting Information. Experimental details, 1D and
b
Figure 3. Effects of 6 and 7 on tubulin assembly. Purified porcine brain
tubulin was incubated at 37 °C in the presence of vehicle or 10 μM
paclitaxel, 6, or 7. Microtubule polymerization was monitored turbidi-
metrically at 340 nm.
2D NMR spectra, and additional bioassay data. This material is
available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
Mooberry@uthscsa.edu
polymerization of purified tubulin (Figure 3), demonstrating that
this is a common feature for taccalonolides with low-nanomolar
potency. Additional experiments to test this hypothesis and
define the interaction between the taccalonolides and tubulin
are ongoing and will provide important information regarding
the mechanism of action of this class of microtubule stabilizers
and whether they bind within the taxane site, the peloruside
A/laulimalide site, or a unique site on microtubules.
Taccalonolides 4 and 8 are ketone/enol tautomers of 9.
Additionally, 5 and 9 are potentially interconvertible geminal
diol and ketone moieties (Supplemental Figure 4). It is interest-
ing that we isolated all of them as stable isomers, as suggested by
the different biological potencies obtained with these com-
pounds. Additionally, significant tautomerization or interconver-
sions were not seen during purification or storage in the NMR
solvent over 3 weeks. The enol and germinal diol in 4, 5, and 8
may be stabilized by the α-carbonyl and 15-acetoxy groups
through intramolecular H-bonds. Since quantities were limited,
the conditions for tautomerization and interconversion were not
evaluated but will be investigated in future studies.
The six new taccalonolides described in this study have
expanded the SARs of this class and identified the critical
importance of the E-ring constituents at C20ÀC23. The C22/
C23 epoxide group in 6 and 7 dramatically increases the potency,
while the C20 hydroperoxyl group in 3 results in total loss of
antiproliferative and microtubule-stabilizing activities.
Most notable is the finding that taccalonolides with single-
digit nanomolar potency can be identified and that compounds 6
and 7, the most potent taccalonolides isolated to date, afforded us
the ability to detect a direct interaction between these taccalo-
nolides and purified tubulin, consistent with their actions in cells.
The rationale for why 6 and 7 interact with purified tubulin
whereas 1 and 2 cannot affect tubulin polymerization is likely
related to differences in potency. The ability of microtubule-
targeted agents to affect the polymerization of purified tubulin in
biochemical assays requires concentrations that are orders of
magnitude higher than the concentration required to observe
changes in microtubule density and antiproliferative and anti-
mitotic effects in cells. A number of factors contribute to this
discrepancy, including the ability of these compounds to be
concentrated hundreds of fold across the cell membrane.13,14
’ ACKNOWLEDGMENT
This work was supported by NCI CA121138 (S.L.M.),
COSTAR Program NIDCR DE 14318 (J.L.), DOD-CDMRP
Postdoctoral Award BC087466 (A.L.R.), and NCI P30
CA054174 (S.L.M.). Support from the Mass Spectrometry and
NMR Cores are gratefully acknowledged. We extend sincere thanks
to Prof. Doug Frantz for his critical reading of this manuscript and
Prof. Cong-Gui Zhao for providing dimethyldioxirane.
’ REFERENCES
(1) Dumontet, C.; Jordan, M. A. Nat. Rev. Drug Discovery 2010, 9, 790.
(2) Kavallaris, M. Nat. Rev. Cancer 2010, 10, 194.
(3) Galsky, M. D.; Dritselis, A.; Kirkpatrick, P.; Oh, W. K. Nat. Rev.
Drug Discovery 2010, 9, 677.
(4) Morris, P. G.; Fornier, M. N. Clin. Cancer Res. 2008, 14, 7167.
(5) Tinley, T. L.; Randall-Hlubek, D. A.; Leal, R. M.; Jackson, E. M.;
Cessac, J. W.; Quada, J. C., Jr.; Hemscheidt, T. K.; Mooberry, S. L. Cancer
Res. 2003, 63, 3211.
(6) Risinger, A. L.; Jackson, E. M.; Polin, L. A.; Helms, G. L.;
LeBoeuf, D. A.; Joe, P. A.; Hopper-Borge, E.; Luduena, R. F.; Kruh,
G. D.; Mooberry, S. L. Cancer Res. 2008, 68, 8881.
(7) Buey, R. M.; Barasoain, I.; Jackson, E.; Meyer, A.; Giannakakou,
P.; Paterson, I.; Mooberry, S. L.; Andreu, J. M.; Díaz, J. F. Chem. Biol.
2005, 12, 1269.
(8) Risinger, A. L.; Mooberry, S. L. Cell Cycle 2011, 10, 2162.
(9) Peng, J.; Risinger, A. L.; Fest, G. A.; Jackson, E. M.; Helms, G.;
Polin, L. A.; Mooberry, S. L. J. Med. Chem. 2011, 54, 6117.
(10) Yang, J. Y.; Zhao, R. H.; Chen, C. X.; Ni, W.; Teng, F.; Hao,
X. J.; Liu, H. Y. Helv. Chim. Acta 2008, 91, 1077.
(11) Chen, Z.-L.; Shen, J.-H.; Gao, Y.-S.; Wichtl, M. Planta Med.
1997, 63, 40.
(12) Adam, W.; Saha-Moller, C. R.; Zhao, C.-G. Org. React. 2002,
61, 219.
(13) Jordan, M. A.; Toso, R. J.; Thrower, D.; Wilson, L. Proc. Natl.
Acad. Sci. U.S.A. 1993, 90, 9552.
(14) Kuh, H.-J.; Jang, S. H.; Wientjes, M. G.; Au, J.L-S. J. Pharm. Exp.
Ther. 2000, 293, 761.
19067
dx.doi.org/10.1021/ja209045k |J. Am. Chem. Soc. 2011, 133, 19064–19067