Protein-degrading enediynes: Library screening of Bergman cycloaromatization products

Article abstract of DOI:10.1021/ol005926q

(Formula presented) A screening method based on Bergman cycloaromatization product was applied to a compact library of estrogenic-enediyne hybrids. An enediyne candidate identified from the screen was subsequently synthesized, and it induced temperature-

Full text of DOI:10.1021/ol005926q

ORGANIC
LETTERS
2000
Vol. 2, No. 13
1863-1866
Protein-Degrading Enediynes: Library
Screening of Bergman
Cycloaromatization Products
,†
Graham B. Jones,* Justin M. Wright,† George Hynd,† Justin K. Wyatt,†
Molly Yancisin,‡ and Myles A. Brown‡
Bioorganic and Medicinal Chemistry Laboratories, Department of Chemistry,
Northeastern UniVersity, 360 Huntington AVenue, 102HT, Boston, Massachusetts
02115, and Department of Adult Oncology, Dana-Farber Cancer Institute,
HarVard Medical School, 44 Binney Street, Boston, Massachusetts 02115
grjones@lynx.neu.edu
Received April 11, 2000
ABSTRACT
A screening method based on Bergman cycloaromatization products was applied to a compact library of estrogenic-enediyne hybrids. An
enediyne candidate identified from the screen was subsequently synthesized, and it induced temperature- and concentration-dependent
degradation of human estrogen receptor r upon cycloaromatization.
The enediyne antitumor antibiotics are capable of inducing
a wide range of biological events, including DNA strand
scission, RNA cleavage, protein agglomeration, and apop-
tosis.1 Derivatives of both calicheamicin and neocarzinostatin
are currently undergoing clinical evaluation, and the chal-
lenge of designing synthetic enediyne hybrids continues to
attract interest.2 While the in vitro and in vivo effectiveness
of enediynes against certain cancers is unquestioned, the
exact mechanism(s) of biological activity remains to be
clarified. On the basis of the reported protein-modulating
ability of synthetic enediynes,3 and the recent confirmation
that amino acid radicals can be generated from enediynes,4
we became interested in the possibility of designing enediyne
hybrids 1 capable of interacting with specific receptor targets
to form 2 (Scheme 1). Such systems could have vast
Scheme 1. General Strategy for Enediyne “Affinity Cleavage
Agents”
† Northeastern University.
‡ Harvard Medical School.
(1) Xi, Z.; Goldberg, I. ComprehensiVe Natural Products Chemistry;
Barton, D. H. R., Nakanishi, K., Eds.; Pergamon: Oxford, 1999; Vol. 7, p
553.
(2) Smith, A. L.; Nicolaou, K. C. J. Med. Chem. 1996, 39, 2103. Maier,
M. E. Synlett 1995, 13.
(3) Zein, N.; Solomon, W.; Casazza, A. M.; Kadow, J. F.; Krishnan, B.
S.; Tun, M. M.; Vyas, D. M.; Doyle, T. W. Bioorg. Med. Chem. Lett. 1993,
3, 1351. Zein, N.; Schroeder, D. R. AdVances in DNA Sequence Specific
Agents; Jones, G. B., Ed.; JAI Press Inc., Greenwich, CT, 1998; Vol. 3, p
201.
10.1021/ol005926q CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/24/2000

Figure 1. Library of estrogenic probes synthesized 7, R ) 9-24.
potential, interaction of the cycloaromatized diyls with the
protein target (3) serving either as dynamic probes of protein
architecture or as irreversible inhibitors of protein function.
Should proteolysis result (4), the enediyne would constitute
an affinity cleavage system, complementing established
methods.5 An enediyne-based system could offer several
advantages in this regard, most importantly that the entire
probe is hydrophobic, ideal for receptors whose endogenous
ligands are lipophilic, including the nuclear receptor super-
family.6
Due to the thermal instability and often sensitive chemistry
involved, the synthesis and manipulation of enediynes
requires special attention. Accordingly, rather than as-
sembling an enediyne library, we wished to outline a rapid
screening method for the identification of promising candi-
date compounds, based on their presumed affinity for a target
receptor. Since the cycloaromatization of C-10 carbocyclic
enediynes 5 (Scheme 2) takes place via a late-stage transition
state, the active diyl radicals 6 bear a close structural
resemblance to the arene products 7. A viable strategy could
therefore be to couple readily available alcohol 87 with
structures having affinity to the receptor of interest and screen
for receptor affinity, thus identifying the most promising
enediyne candidates for subsequent synthesis. Due to its
importance in endocrinological pathways related to cancer,
we focused our attention on the human estrogen receptor
R(hERR).8 High affinity ligands identified for this receptor
include a wide variety of phenols which mimic the endog-
enous agonist, â-estradiol.9 Accordingly, 8 was coupled to
a library of known ligands, using Mitsonobu coupling
methods with the appropriate phenols and alcohols (DEAD,
PPh3, DMF, 12 h/0 °C). Mimics prepared (Figure 1) include
the tetrahydronaphthyl ethers of hexestrols and stilbestrols
9-12, â-estradiol 13, zeranols 14-16, hydroxylated tamox-
ifen derivatives 17 and 18, biochanin A 19, and the alkyl
phenols 20-22. Analysis of affinity for the ligand-binding
(4) Jones, G. B.; Plourde, G. W.; Wright, J. M. Org. Lett. 2000, 2, 811.
(5) Rana, T. M.; Meares, C. F. J. Am. Chem. Soc. 1991, 113, 1859. Hoyer,
D.; Cho, H.; Schultz, P. G. J. Am. Chem. Soc. 1990, 112, 3249. Cuenoud,
B.; Tarasow, T. M.; Schepartz, A. Tetrahedron Lett. 1992, 33, 895. Miyake,
R.; Owens, J. T.; Xu, D.; Jackson, W. M.; Meares, C. F. J. Am. Chem. Soc.
1999, 121, 7453. Kumar, C. V.; Buranaprapuk, A. J. Am. Chem. Soc. 1999,
121, 4262.
Scheme 2. Post-Bergman Cycloaromatization Products as Diyl
Isosteres
(6) O’Malley, B. W.; Mol. Endocrinol. 1990, 4, 363. Evans, R. M.
Science 1988, 240, 889.
(7) Beckwith, A. L. J.; Gerba, S. Aust. J. Chem. 1992, 45, 289.
(8) Brzozowski, A. M.; Pike, A. C. W.; Dauter, Z.; Hubbard, R. E.; Bonn,
T.; Engstrom, O.; Ohman, L.; Greene, G. L.; Gustafsson, J.-A.; Carlquist,
M.; Nature 1997, 389, 753. Kumar, V.; Green, S.; Stack, G.; Berry, M.;
Jin, J.-R.; Chambon, P. Cell 1987, 51, 941.
(9) Gao, H.; Katzenellenbogen, J. A.; Garg, R.; Hansch, C. Chem. ReV.
1999, 99, 723.
1864
Org. Lett., Vol. 2, No. 13, 2000

domain of hERR was conducted using a competitive
displacement assay,10 in most cases indicating relatively poor
affinity for the receptor. This is presumably due to the
reduction in hydrogen bonding capacity relative to the native
ligands, two such interactions between â-estradiol and the
receptor having been identified.8 Our attention therefore
turned to accessible ligands which possess an enediyne-
tethering site in addition to diol functionality. It is known
that certain 17R-alkynyl steroid derivatives possess compa-
rable binding to the parent steroidal nucleus;9 thus analogues
23 and 24 were prepared by coupling 8, with either the
alkynyl carboxylate or propargyl alcohol, both of which are
readily prepared from commercially available 17R-ethynyl
estradiol. Ester 23 and ether 24 both showed sub-micromolar
affinity to hERR;10 however, in the case of 24, the compound
proved unstable, decomposing to a mixture of products on
standing for extended periods. Subsequent investigations
revealed that 23, like â-estradiol, is capable of recruiting the
AIB1, GRIP1, and RAC3 estrogen receptor coactivator
proteins, which are important for formation of estrogenic
complexes competent for transcription.11 For these reasons,
the enediyne-estrogen derived from 23 became the candidate
for synthesis.12 The thermally labile enediyne core 25 was
prepared in eight steps from commercially available methyl
hexynoate, the key enediyne closure utilizing an intramo-
lecular carbenoid coupling reaction.13 Alkyne 26, prepared
from ethynyl estradiol, was converted to alkynyl carboxylate
27 and immediately coupled with freshly prepared 25
(Scheme 3). Subsequent desilylation gave key enediyne-
cycloaromatization to yield adduct 23 (half-life approxi-
mately 15 h at 37 °C), identical in all respects to the authentic
material, and underscoring our design philosophy (Scheme
2). We were now able to study the key interaction of diyl
29 with the receptor target. Accordingly, a freshly prepared
sample of 28 was incubated with 35S-labeled full length
hERR at various concentrations for two half-lives (36 h),
and then the protein was separated using SDS-PAGE and
visualized using fluorography. The results indicate that the
enediyne induces degradation of the receptor (Figure 2, lanes
Figure 2. ERR degradation mediated by enediyne 28. 35S-
Methionine-labeled full length hERR incubated with either ethanol
alone (lanes 1 and 8), estradiol (lanes 2 and 9, 10 µM), 4-OH-
tamoxifen (lanes 3 and 10, 10 µM), ICI182,780 (lanes 4 and 11,
10 µM), and 28 (lanes 5-7 and 12-14 at concentrations indicated)
at either 37 °C or 4 °C for 36 h. The samples were resolved (10%
SDS-PAGE), fixed, enhanced, dried, and visualized using fluorog-
raphy. Incubation with 23 or 25 (1 mM) produced no change relative
to control (data not shown).
Scheme 3. Preparation of Chemically Reactive
Estrogen-Enediyne Probe
13 and14), and that the process has concentration and
temperature-dependent components (lanes 6 and 7). This
finding constitutes the first example of targeted protein
degradation using a designed enediyne. Control reactions
either with estradiol, the antiestrogens 4-hydroxytamoxifen
or ICI 182,780,14 arene 23, or enediyne 25 indicate the
enediyne-estrogen conjugate is responsible for the degrada-
tion, which implies a proteolytic mechanism involving diyl
29.3,4 The obserVation of receptor degradation at micromolar
concentration is especially encouraging giVen the fact that
the affinity of 23 for hERR was only in the low micromolar
range. It is thus possible that improved analogues can be
found which approach the nanomolar affinity levels observed
for natural ligands, including â-estradiol, which in turn may
improve both the specificity and selectivity of the degradation
event.
estrogen 28 in good yield. As expected, in the presence of a
hydrogen donor, this enediyne underwent Bergman type
(12) Wang, J.; DeClercq, P. J. Angew. Chem., Int. Ed. Engl. 1995, 34,
1749. Py, S.; Harwig, C. W.; Banerjee, S.; Brown, D. L.; Fallis, A. G.
Tetrahedron Lett. 1998, 39, 6139.
(13) Jones, G. B.; Huber, R. S.; Mathews, J. E. J. Chem. Soc., Chem.
Commun. 1995, 1791. Jones, G. B.; Wright, J. M.; Plourde, G. W., II; Hynd,
G.; Huber, R. S.; Mathews, J. E. J. Am. Chem. Soc. 2000, 122, 1937.
(14) Wakeling, A. E.; Bowler, J. J. Steroid Biochem. Mol. Biol. 1992,
43, 173. Osborne, C. K.; Coronado, E.; Allred, D. C.; Wiebe, V.;
DeGregorio, M. J. Natl. Cancer Inst. 1991, 83, 1477.
(10) Relative binding affinities (RBA’s) determined by displacement of
3H estradiol from ligand binding domain of hERR using increasing
concentrations (nM through mM) of candidate compounds at 4 °C/37 °C.
RBA’s of compounds 9-22 were all >1 µM, 23 (0.5 µM), and 24 (0.1
µM) relative to estradiol (1 nM). Specific details of the entire screen will
be published in a full account of this work.
(11) Chen, H.; Lin, R. J.; Xie, W.; Wilpitz, D.; Evans, R. M. Cell 1999,
98, 675.
Org. Lett., Vol. 2, No. 13, 2000
1865

subjected to photolytic ligand exchange with polymer-
supported triphenylphosphine to yield adduct 31 (Scheme
4). Esterification with 27 followed by deprotection were
successful on the polymer-supported analogue, decomplex-
ation giving a sample of 23, identical with that prepared using
the solution-phase method. We envisage this technology will
now permit the rapid assembly of diverse libraries of
enediyne mimics. Due to the promising activity of the
estrogen enediyne hybrid 28 (Figure 2), we anticipate many
applications will be forthcoming.
Scheme 4. Polymer-Supported Route to Library
In summary, a convenient library screening method
identified an estrogenic enediyne candidate. Following
synthesis, the enediyne-induced temperature- and concentra-
tion-dependent degradation of the human estrogen receptor
R occurred at micromolar levels,16 providing the first proof-
of-concept for designed protein-targeted enediynes.17
Having demonstrated a screening method based on post-
Bergman activated enediynes, it will now be of interest to
perform refined screens to find optimized candidates, targeted
toward this and other nuclear receptors. In pursuit of this
goal, it may be desirable to employ solid-phase synthesis
methods, and we elected to investigate the utility of the newly
discovered “traceless linker” method.15 Accordingly, alcohol
8 was converted to tricarbonyl chromium complex 30 and
OL005926Q
(16) At micromolar levels, degradation of other proteins might also be
expected to compete. In a preliminary survey, trace degradation of lysozyme
and bovine serum albumin is observed, whereas moderate degradation of
the transcription factor FKHR is induced. The scope and selectivity of
degradation induced by 28 and refined analogues will be reported in a full
account of this work.
(15) Gibson, S. E.; Hales, N. J.; Peplow, M. A. Tetrahedron Lett. 1999,
40, 1417.
(17) This work was generously supported by the National Institutes of
Health [R01GM57123].
1866
Org. Lett., Vol. 2, No. 13, 2000