Identification of a tunable site in bryostatin analogs: C20 bryologs through late stage diversification

Article abstract of DOI:10.1021/ol0501931

The C20 region of our bryostatin analogs was identified as a nonpharmacophoric site that could be varied to tune analogs for function and physical properties without significantly affecting their binding affinity for PKC. The use of this site in a late-stage diversification strategy has enabled the facile synthesis of a variety of new C20 analogs, all of which retain nanomolar affinity for PKC, in agreement with our pharmacophore hypothesis.

Full text of DOI:10.1021/ol0501931

ORGANIC
LETTERS
2005
Vol. 7, No. 6
1177-1180
Identification of a Tunable Site in
Bryostatin Analogs: C20 Bryologs
through Late Stage Diversification
Paul A. Wender*,†,‡ and Jeremy L. Baryza†
Department of Chemistry, Stanford UniVersity, Stanford, California 94305-5080, and
Department of Molecular Pharmacology, Stanford UniVersity School of Medicine,
Stanford UniVersity, Stanford, California 94305-5080
wenderp@stanford.edu
Received January 30, 2005
ABSTRACT
The C20 region of our bryostatin analogs was identified as a nonpharmacophoric site that could be varied to tune analogs for function and
physical properties without significantly affecting their binding affinity for PKC. The use of this site in a late-stage diversification strategy has
enabled the facile synthesis of a variety of new C20 analogs, all of which retain nanomolar affinity for PKC, in agreement with our pharmacophore
hypothesis.
Bryostatin 1 (Figure 1) is a highly potent compound with a
unique spectrum of biological activities. It is currently in
Phase I and II clinical trials for the treatment of cancer.
Bryostatin 1 has been shown to inhibit cancer growth in vitro
and in vivo, synergize the action of other anticancer agents,1
reverse multidrug resistance,2 and stimulate the immune
system.3 While its mode of action is not established,
bryostatin 1 is a potent activator of protein kinase C (PKC).
Unfortunately, the limited availability4 and structural com-
plexity of bryostatin have restricted synthetic and biochemical
studies that could lead to a better understanding of its
molecular mode of action and, equally importantly, the
identification of clinically superior agents. Being a marine
† Department of Chemistry.
‡ Department of Molecular Pharmacology.
(1) Ali, S.; Aranha, O.; Li, Y. W.; Pettit, G. R.; Sarkar, F. H.; Philip, P.
A. Cancer Chemother. Pharm. 2003, 52, 235-246.
(2) (a) Spitaler, M.; Utz, I.; Hilbe, W.; Hofmann, J.; Grunicke, H. H.
Biochem. Pharm. 1998, 56, 861-869. (b) AlKatib, A. M.; Smith, M. R.;
Kamanda, W. S.; Pettit, G. R.; Hamdan, M.; Mohamed, A. N.; Chelladurai,
B.; Mohammad, R. M. Clin. Cancer Res. 1998, 4, 1305-1314.
(3) Curiel, R. E.; Garcia, C. S.; Farooq, L.; Aguero, M. F.; Espinoza-
Delgado, I. J. Immunol. 2001, 167, 4828-4837.
(4) For a leading reference, see: Hildebrand, M.; Waggoner, L. E.; Liu,
H.; Sudek, S.; Allen, S.; Anderson, C.; Sherman, D. H.; Haygood, M. Chem.
Biol. 2004, 11, 1543-1552.
Figure 1. Bryostatin 1 and analogs 1 and 2.
10.1021/ol0501931 CCC: $30.25
© 2005 American Chemical Society
Published on Web 02/17/2005

natural product, bryostatin is neither produced nor optimized
for clinical use in humans.
Scheme 1. Installation of C20 Sidechain
Using pharmacophoric and docking hypotheses, we have
been engaged in the design and synthesis of structurally
simplified but functionally potent analogs of bryostatin,
whose activities, physical properties, side-effect profile,
detectability, and other properties could be tuned through
synthesis.5 Analogs 1 and 2 (Figure 1) are representative of
this effort. These analogs are readily available synthetically
and exhibit bryostatin-like activity but demonstrate greater
potency in vitro and in limited in vivo studies when compared
to the natural product.5,6
With the demonstration that comparable or superior
affinity and function can be realized in simplified, syntheti-
cally accessible bryostatin analogs, we next sought to identify
a region in our analogs that could be readily modified to
optimize in vivo performance without sacrificing potency
or function. The identification of such a site would allow
for the preparation of analogs as needed with improved
physical properties and ADME behavior and with tags for
use in biochemical and mechanistic studies. Preferably the
selected region would be amenable to modification at the
end of the current synthesis to reduce the number of steps
required to prepare each new analog and to avoid undesired
changes in other sensitive functionality.
The above considerations restrict potential sites for
modification of our analogs to the C20 ester and C21 enoate
as all other functional groups are implicated in binding and/
or preorganization. Unfortunately, our previous efforts to
modify the C21 enoate at the end of the current synthesis
have been frustrated by undesired rearrangements of the
densely functionalized C-ring. We were, however, able to
prepare ester variants at C20 (Scheme 1: 3 f 5 f 9) that
retained binding potency, but access to such analogs requires
early divergence from the current synthesis route, adding as
many as eight steps to the synthesis of each derivative and
restricting the type of derivative to those compatible with
the remaining steps.7 Attempts to modify our analogs, or their
late stage precursors, at C20 have been complicated by steric
congestion around C20 and the associated undesired involve-
ment of neighboring functionality in C20 modifications.
Notwithstanding these problems, our earlier synthesis of a
C20 ester variant and the observation that C20 ester
variations are found in active natural bryostatins suggested
that the C20 substituent could be used for tuning the
performance of our analogs.
incorporate into that group a functionality that would be
spatially removed from the steric crowding at C20 and
amenable to flexible modification at the end of the synthesis.
In our current synthesis (Scheme 1), the C20 ester group is
introduced relatively early by acylation of alcohol 3.5b,8
Unfortunately, alcohol 3 is sterically hindered and prone to
acid-promoted elimination of the C19 OMe group to produce
4 (Scheme 1), as well as base-promoted rearrangement of
the C21 enoate. Therefore, the use of less reactive acylation
agents is precluded. Electron-rich acids, such as 4-benzyl-
oxybenzoic acid, or even moderately bulky acids, such as
bromoacetic acid, failed to react at all or, under forcing
conditions, resulted in decomposition. More reactive reagents,
such as activated trifluoroacetate or chloroactetate, will
acylate alcohol 3; however, after deprotection of the C17
alcohol, these electron-deficient acyl groups migrate com-
pletely to the less hindered C17 alcohol. This “Goldilocks
effect”,9 meaning the acyl groups employed must be just
right, limits the range of C20 substituents that can be used.
The eventual choice of a 3-aminobenzoate ester as a
tunable C20 substituent was made for several reasons. The
aniline could be generated by selective reduction of a nitro
group without the need to protect other functionalities in
Given the steric problems encountered in the attachment
of a C20 group at the end of our current synthesis, we sought
to install a C20 group at an earlier synthetic point and to
(5) (a) Wender, P. A.; Baryza, J. L.; Brenner, S. E.; Clarke, M. O.; Craske,
M. L.; Horan, J. C.; Meyer, T. Curr. Drug DiscoVery Tech. 2004, 1, 1-11.
(b) Wender, P. A.; Baryza, J. L.; Bennett, C. E.; Bi, C.; Brenner, S. E.;
Clarke, M. O.; Horan, J. C.; Kan, C.; Lacoˆte, E.; Lippa, B.; Nell, P. G.;
Turner, T. M. J. Am. Chem. Soc. 2002, 124, 13648-13649.
(6) (a) Baryza, J. L.; Brenner, S. E.; Craske, M. E.; Meyer, T.; Wender,
P. A. Chem. Biol. 2004, 11, 1261-1267. (b) Stone, J. C.; Stang, S. L.;
Zheng, Y.; Dower, N. A.; Brenner, S. E.; Baryza, J. L.; Wender, P. A. J.
Med. Chem. 2004, 47, 6638-6644.
(8) Wender, P. A.; De Brabander, J.; Harran, P. G.; Jimenez, J.-M.;
Koehler, M. F. T.; Lippa, B.; park, C.-M.; Shiozaki, M. J. Am. Chem. Soc.
1998, 120, 4534-4535.
(9) (a) Hassall, J. Old Nursery Stories and Rhymes; Blackie & Son:
London, 1904. (b) Sommer, T. J.; Bertz, S. H. Chem. InnoV. 2000, 30 (8),
37-42.
(7) Wender, P. A.; Hinkle, K. W. Tetrahedron Lett. 2000, 41, 6725-
6729.
1178
Org. Lett., Vol. 7, No. 6, 2005

Scheme 2. Synthesis of Intermediate 13
Scheme 3. Synthesis of Analog 17
these complex molecules; both the C17-C18 olefin and the
C21 enoate are known to be resistant to hydrogenation in
bryostatin. The aniline moiety would provide a handle for
selective installation of a variety of groups without the need
for protection of other groups in the analog, and the phenyl
ring would provide spatial separation from the sterically
congested C20 position. The choice of a 3-amino species
was made to maximize the nucleophilicity of the nitrogen
by avoiding conjugation to the carbonyl. For pharmacokinetic
and ADME studies, the inclusion of a nitrogen atom is also
desirable due to the improved ionization in MSn analysis,
making detection easier, especially at the low doses used
for studies with these exceptionally potent compounds.
Analog 17 was prepared as shown in Schemes 2 and 3,
starting from the γ-ketoenoate 10.5 Reduction of the ketone
and acylation of the resulting alcohol 3 provided 3-nitroben-
zoate ester 11.
PyBroP/DMAP.5 Macrocyclic transacetalization and con-
comitant desilylation provided, in a single step and under
thermodynamic control, the C20 nitroarene analog 16 in 84%
yield. Gratifyingly, the nitrobenzoate could be reduced
chemoselectively by hydrogenation to provide the desired
aniline 17.
As shown in Table 1, chemoselective acylation of the
aniline moiety of 17 can be accomplished with either
symmetrical or mixed anhydrides at room temperature in
CH2Cl2, providing direct one-step access to a variety of C20
analogs, including aliphatic, aromatic, PEG, and probe
amides. The reaction is highly selective for the aniline
nitrogen, and acylation of the unprotected alcohols was not
observed.
The binding constants for analogs 18a-e to PKC (rat-
brain isozyme mixture) were determined as previously
described5 and are shown in Figure 2. It is apparent from
these data that there is a good tolerance for substituents with
different physical properties. All of the new analogs bind to
PKC with potencies sufficient for mechanistic studies. Even
the highly hydrophilic glycol chain of 18d provides a
compound that is sufficiently potent. These results suggest
that the physical properties of the molecule can be readily
tuned by modification of the C20 substituent. The mild
reaction conditions also allow moieties with other desirable
but potentially sensitive functions to be incorporated into
the analogs. An example of such a function is analog 18e,
which incorporates the NBD fluorophore that can be used
for receptor identification and imaging studies.
Removal of the C17 TBS protecting group and oxidation
of the resulting alcohol generated aldehyde 12. Dihydroxy-
lation of the terminal alkene provided an inseparable 2:1
mixture of diastereomers. Protection of the diols followed
by a four-step homologation of the aldehyde produced enal
14. Unfortunately, the use of 13 in our previously reported
one-step homologation was thwarted by decomposition of
the nitroarene moiety induced by the vinyl zincate reagent.
Deprotection of the C25-C26 diol and hydrolysis of the C19
ketal was accomplished in one step. The change to the C20
substituent appeared to retard this hydrolysis significantly,
lengthening the reaction time from 16 h to 5 days. Selective
protection of the primary alcohol and separation of the
diastereomers provided recognition domain 15, which was
coupled to the spacer domain used in the synthesis of 2 with
Org. Lett., Vol. 7, No. 6, 2005
1179

Table 1. Reagents and Yields for Synthesis of Amide
Analogsa
a All reactions were run in CH2Cl2 at room temperature using 3-4 µmol
of 17. See Supporting Information for details.
Application of a late stage divergence strategy to the
synthesis of new C20 analogs of bryostatin 1 has been
realized, generating compounds with properties tuned for in
vivo use and mechanistic studies. Starting from the aniline
analog 17, analogs with new C20 substituents can be
generated in one step and in good yield. Part of the value of
this approach arises from its step economy. Generating a
similar set of analogs by diversification at the ester formation
step that produces 11 (see Scheme 2) would require 13 steps
per analog. Any need for additional protecting groups would
only lengthen and complicate the preparation of such species.
All of the analogs generated exhibit nanomolar binding
affinity for PKC. The relatively high affinity of these analogs
suggests that changes to this portion of the molecule can be
used to improve physical properties, control selectivity, and
incorporate additional functionality, a fluorophore, to probe
for receptors and mode of action. Further biological testing
of the analogs in Figure 2 and related compounds in assays
designed to measure PKC activation and functional activity
will be reported in due course.
Figure 2. Structures and binding affinities of new analogs to
protein kinase C.
in this area and helpful discussions. We also gratefully
acknowledge Prof. Daria Mochly-Rosen for her assistance
and support.
Supporting Information Available: Experimental pro-
cedures and characterization data for compounds 10-18.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Acknowledgment. The authors thank the National Insti-
tutes of Health for their support of this work (Grant
CA31845). J.L.B. is supported by a Stanford Graduate
Fellowship and an ACS Medicinal Chemistry Predoctoral
Fellowship (Lilly). We thank Cindy Kan for her early efforts
OL0501931
1180
Org. Lett., Vol. 7, No. 6, 2005