Synthesis of a simplified bryostatin C-ring analogue that binds to the CRD2 of human PKC-α and construction of a novel BC-analogue by an unusual julia olefination process

Article abstract of DOI:10.1021/ol027392u

(Matrix presented) The synthesis of two truncated bryostatin analogues 2 and 3 is described. High-field NMR measurements on the C-ring analogue 3 in C²H₃CN containing 25% ²H₂O have shown that it binds to the CRD

Full text of DOI:10.1021/ol027392u

ORGANIC
LETTERS
2003
Vol. 5, No. 4
499-502
Synthesis of a Simplified Bryostatin
C-Ring Analogue That Binds to the
CRD2 of Human PKC-r and
Construction of a Novel BC-Analogue
by an Unusual Julia Olefination Process
Karl J. Hale,* Mark Frigerio, Soraya Manaviazar, Marc G. Hummersone,
Ian J. Fillingham, Igor G. Barsukov, Christian F. Damblon, Andreas Gescher, and
Gordon C. K. Roberts*
The Christopher Ingold Laboratories, UniVersity College London, 20 Gordon Street,
London WC1H 0AJ, England, and The Departments of Biochemistry and Oncology,
The UniVersity of Leicester, PO Box 138, UniVersity Road, Leicester LE1 9HN, U.K.
k.j.hale@ucl.ac.uk; gcr@leicester.ac.uk
Received December 3, 2002
ABSTRACT
2
The synthesis of two truncated bryostatin analogues 2 and 3 is described. High-field NMR measurements on the C-ring analogue 3 in C H CN
3
containing 25% 2H O have shown that it binds to the CRD2 of human PKC-r at virtually the same position as phorbol-13-acetate (PA) and
2
bryostatin 1 (1). NMR titration studies have also revealed that 3 binds to the CRD2 with a potency similar in magnitude to PA but much less
potently than 1.
Protein kinase C (PKC) isozymes are serine and threonine
kinases that play a pivotal role in the signal transduction
pathways that control mammalian cell division.1 They
phosphorylate a range of intracellular protein targets1,2 that
drive the processes of mitogenesis,3 tumorigenesis,3 growth
arrest,3 and differentiation,4 and they are activated by various
extracellular signals. Not surprisingly, their over- or under-
expression in many cells has been correlated with the
transition into malignancy3 and with increased tumorigenic-
ity.3 For example, in A549 human lung cancer cells, PKC-R
levels are considerably higher than in other nearby healthy
alveolar tissue.5 PKC-R levels are also significantly elevated
(1) Reviews: (a) Newton, A. C. Curr. Opin. Cell. Biol. 1997, 9, 161.
(b) Nishizuka, Y. FASEB J. 1995, 9, 484. (c) Nishizuka, Y. Science 1992,
258, 607. (d) Mahoney, C. W.; Huang, K.-P. In Molecular and catalytic
properties of protein kinase c in Protein Kinase C; Kuo, J. F., Ed.; Oxford
University Press: New York, 1994; p 16.
(2) Fields, A. P.; Pettit, G. R.; Stratford May, W. J. Biol. Chem. 1988,
263, 8253.
(3) (a) Borner, C.; Guadagno, S. N.; Fabbro, D.; Weinstein, I. B. J. Biol.
Chem. 1992, 267, 12892. (b) Mischak, H.; Goodnight, J.; Kolch, W.;
Martiny-Baron, G.; Schaechtle, C.; Kazanietz, M. G.; Blumberg, P. M.;
Pierce, J.; Mushinski, J. F. J. Biol. Chem. 1993, 268, 6090.
10.1021/ol027392u CCC: $25.00 © 2003 American Chemical Society
Published on Web 01/23/2003

in MCF-7 human breast cancer cells compared with normal
breast tissue.5 In NIH-3T3 cells, PKC-ꢀ overexpression
causes them to become highly neoplastic and tumorigenic
toward nude mice,4 while PKC-δ overexpression halts
proliferation.4 In mouse keratinocytes, PKC-δ overexpression
results in apoptosis,6 while in rat vascular smooth muscle
cells, it causes a dramatic decrease in the levels of cyclins
D1 and E. The latter enzymes are critically involved in cell
cycle progression.7 PKC-δ overexpression in this system also
up-regulates p27, which is a cyclin-dependent kinase inhibi-
tor.7 PKC-δ has additionally been implicated in the anti-
proliferative signaling pathway that is mediated by cell-
cell contacts.8 It will thus be appreciated that PKCs can not
only serve as oncogenes in certain circumstances but also
function as tumor-suppressors, depending upon their exact
levels of expression and their precise distribution within cells.
One naturally occurring antitumor agent that binds selec-
tively to a range of different PKC isoforms, to thereafter
modulate their levels of expression, is bryostatin 1 (1).5,9,10
It potently inhibits the growth of several cancer cell lines in
a way that correlates closely with its selective down-
regulation of PKC-R.5,9b It can also halt the growth of some
cell systems by apparently preventing certain PKC-δs from
undergoing down-regulation.11
So far, only Wender’s group has made substantial progress
on the bryostatin analogue/probe front,14 their team having
already identified several simplified structures that combine
good PKC inhibiting activity with powerful in vitro antitumor
effects (see Figure 1). Their cumulative data on this class
Figure 1. Some of Wender’s bryostatin analogues.
suggest (1) that a 20-membered macrolactone ring, contain-
ing an intact “Southern Hemisphere”, is needed for good
PKC binding activity; (2) that a C(3)-hydroxyl with (R)-
stereochemistry is important for a high enzyme affinity; and
(3) that a free hydroxyl at C(26) is essential for a good
interaction with PKC isozymes. Of further relevance is
Wender’s observation that the B-ring exocyclic olefin and
the A-ring can both be deleted from 20-membered macro-
lactone analogue structures without having serious conse-
quences for PKC-binding affinity.
Bryostatin 1 (1) binds to the diacylglycerol/phorbol ester
binding sites of phosphorylated PKC isozymes at two highly
conserved regions known as the cysteine-rich domains 1 and
2 (PKC CRDs 1 and 2),12 and in so doing, it creates ligated
proteins that have greatly differing intracellular translocation
properties,5 as well as differing susceptibilities toward
undergoing degradation13 (and down-regulation) at their final
intracellular destination(s).
One of the main goals of our PKC probe/analogue program
in the bryostatin area is to characterize the binding of our
simplified probe molecules to the CRDs of various human
PKCs by NMR methods. The primary advantage of using
this technique to identify novel probe structures is that one
can readily establish whether a particular analogue/probe is
binding to the relevant CRD of a human PKC in exactly the
same way as the lead structure (viz. bryostatin 1). The NMR
approach also allows small and often quite subtle changes
in ligand binding to be easily detected; these can subse-
quently be correlated with downstream effects on PKC
signaling and with the antitumor profile of a particular
analogue. The NMR method also gives precise information
on which amino acid residues must be targeted in a PKC to
produce a particular biological effect, enabling the rational
design of new analogues with tailored biological profiles.
At the outset of this project, we thought it important to
identify the minimum pharmacophoric elements of bryostatin
1 that would be needed for successful binding to the CRD2
of human PKC-R. Given that Wender and co-workers14 had
already demonstrated that substantial changes could be
Much has still to be learned about how the different PKCs
regulate cellular proliferation, differentiation, and morphol-
ogy, and it is clear that our understanding will only improve
if increased access is gained to novel PKC-interactive probe
molecules. The fact that bryostatin 1 can selectively modulate
PKC isozymes at very low drug concentrations, and simul-
taneously prevent tumor cell growth, makes it an especially
attractive lead for the design of new PKC probes, for there
is the added possibility that some of these structures will
themselves function as novel anticancer drugs through their
effects on deregulated PKC-signaling pathways.
(4) Mischak, H.; Pierce, J. H.; Goodnight, J.; Kazanietz, M. G.; Blumberg,
P. M.; Mushinski, J. F. J. Biol. Chem. 1993, 268, 20110.
(5) Stanwell, C.; Gescher, A.; Bradshaw, T. D.; Pettit, G. R. Int. J. Cancer
1994, 56, 585.
(6) Deacon, E. M.; Pongracz, J.; Griffiths, G.; Lord, J. M. Mol. Pathol.
1997, 50, 124.
(7) Fukumoto, S.; Nishizawa, Y.; Hosoi, M.; Koyama, H.; Yamakawa,
K.; Ohno, S.; Morii, H. J. Biol. Chem. 1997, 272, 13816.
(8) Heit, I.; Weiser, R. J.; Herget, T.; Faust, D.; Borchert,-Stuhltrager,
M.; Oesch, F.; Dietrich, C. Oncogene 2001, 20, 5143.
(9) (a) Kraft, A. S.; Smith, J. B.; Berkow, R. L. Proc. Natl. Acad. Sci.,
U.S.A. 1986, 83, 1334. (b) Jalava, A.; Lintunen, M.; Heikkila, J. Biochem.
Biophys. Res. Commun. 1993, 191, 472.
(10) For reviews on bryostatin chemistry and biology, see: (a) Norcross,
R.; Paterson I. Chem. ReV. 1995, 95, 2041. (b) Mutter, R.; Wills, M. Biorg.
Med. Chem. 2000, 8, 1841. (c) Hale, K. J.; Hummersone, M. G.; Manaviazar,
S.; Frigerio, M. Nat. Prod. Rep. 2002, 413.
(11) Szallasi, Z.; Denning, M. F.; Smith, C. B.; Diugosz, A. A.; Yuspa,
S. H.; Pettit, G. R.; Blumberg, P. M. Mol. Pharmacol. 1994, 46, 840.
(12) Zhang, G.; Kazanietz, M. G.; Blumberg, P. M.; Hurley, J. H. Cell
1995, 81, 917.
(13) For the mechanism of PKC-R and -ꢀ degradation in human renal
epithelial cells and fibroblasts, see: Lee, H.-W.; Smith, L.; Pettit, G. R.;
Bingham Smith. Mol. Pharmacol. 1997, 51, 439.
(14) (a) Wender, P. A.; Hinkle, K. W.; Koehler, M. F. T.; Lippa, B.
Med. Res. ReV. 1999, 19, 388. (b) Wender, P. A.; Baryza, J. L.; Bennett,
C. E.; Bi, F. C.; Brenner, S. E.; Clarke, M. O.; Horan, J. C.; Kan, C.; Lacote,
E.; Lippa, B.; Nell, P. G.; Turner, T. M. J. Am. Chem. Soc. 2002, 124,
13648.
500
Org. Lett., Vol. 5, No. 4, 2003

tolerated within the A- and B-regions of their PKC-binding
bryostatin analogues, we decided to investigate whether
C-ring analogues that lacked both the 20-membered mac-
rolactone ring and the A- and B-ring systems could still bind
effectively to the CRDs of human PKCs.
of Ichikawa17a and Hommel.17b On addition of PA or phorbol-
12,13-dibutyrate (PB) to our CRD2 in H2O, the ligand-
protein complexes immediately precipitated. Following a
2
Scheme 2. Synthesis of BC-Analogue 2a
Although Wender’s results on the C-ring enal 6 (Figure
1) had not been especially encouraging in this respect,14 we
remained undaunted, for the X-ray crystal structure of
tetradecanoyl phorbol-13-acetate 7 (TPA) complexed to
the CRD2 of murine PKC-δ had very clearly shown that
only two of its hydroxyls, OH(20) and OH(4), and its
C(3)-carbonyl were involved in binding of the ligand to the
protein.12 Moreover, the fact that all three of these TPA
recognition sites were positioned within a short five-carbon
chain,15 suggested to us that the bryostatin macrolactone
system might not actually be necessary for effective binding
to PKC and that a truncated C-ring analogue which lacked
this feature might still bind to the CRD2, provided its
H-bonding capabilities were suitably magnified. However,
such a system would still need to project an appropriately
oriented lipophilic group from its core to create the long,
continuous, hydrophobic protein domain needed for PKC
membrane insertion.15 Herein, we now report on how such
planning recently allowed a C-ring analogue 3 to be iden-
tified, which binds to the human PKC-R CRD2 with an
affinity of the same order of magnitude as phorbol-13-acetate
(PA).
Our pathway to 3 set off from the known phenyl sulfone
8, whose synthesis had previously been described by us from
(E)-1,4-hexadiene.16 Compound 8 (Scheme 1) underwent
Scheme 1. Pathway to C-Ring Analogue 3a
a Reagents and conditions: (a) Ac2O (5 equiv), Py (5 equiv),
DMAP (0.5 equiv), CH2Cl2 ([8] ) 0.05 M); (b) 40% aq HF (4
equiv), MeCN/THF (1:1, 0.05 M), 2 h.
smooth acetylation with acetic anhydride, pyridine, and
DMAP to give the corresponding O-acetate, which readily
underwent O-desilylation with 40% aq HF in 53% overall
yield. Importantly, these conditions also caused glycoside
hydrolysis to provide 3 as a 3.6:1 mixture of ring-closed/
ring-opened hemiketals. The interaction of 3 with the human
a
Reagents and conditions: (a) Et3SiOTf (1.5 equiv), 2,6-lutidine
(5 equiv), CH2Cl2 ([9] ) 0.08 M), -20 °C, 7 min; (b) DDQ (1.5
equiv), CH2Cl2/H2O (18.1), rt, 1.5 h; (c) TPAP (0.1 equiv), NMO
(2 equiv), MeCN (0.05M), 4A MS, rt, 0.5 h; (d) 12, PhLi (1.8 M
solution in hexane, 1.26 equiv), THF (0.1 M), -78 °C; warm to
-50 °C for 20 min, re-cool to -78 °C, add 11 in THF (0.1 M)
dropwise, stir 35 min, warm to rt for 2 h, quench with saturated
aqueous NH4Cl at -78 °C; (e) to 13 in MeOH/EtOAc (2:1) ([13]
) 0.01 M) at -20 °C was added 6% Na/Hg (2.5 g per 80 mg of
13) and Na2HPO4 (1.25 g per 80 mg of 13) over 4.5 h, then warmed
to 0 °C for 2 h; (f) BzCl (1 equiv), DMAP (0.1 equiv), Py (0.23M),
CH2Cl2 ([14] ) 0.015 M); (g) 40% aq HF (6 equiv), THF/MeCN
(2:1, 0.02 M), 0 °C, 0.75 h, rt, 12 h.
1
PKC-R CRD2 was subsequently studied by 600 MHz H
NMR spectroscopy.
With the kind assistance of Dr. Peter Parker of the Cancer
Research UK Laboratories at Lincoln’s Inn Fields, we
prepared a CRD2 construct of human PKC-R and verified
1
its sequence and folding by comparison of its H spectrum
in 2H2O with the previously published resonance assignments
Org. Lett., Vol. 5, No. 4, 2003
501

suggestion in the literature,18 we added 25% C2H3CN to these
complexes, which markedly improved their solubility. A
detailed comparison of the 1D and 2D spectra of the CRD2
in this solvent system confirmed that the structure of the
protein had not been significantly affected,19 indicating that
it will be a generally useful solvent mixture for studying the
binding of very hydrophobic ligands to this protein.
sulfone 12. n-BuLi, t-BuLi, MeLi, and LDA all proved
unsatisfactory in this regard. However, by adopting the PhLi/
THF metalation conditions of Masamune,21 we were able to
eventually prepare the desired anion and add it onto aldehyde
11 to obtain 13 in 26% isolated yield. However, the actual
yield of 13 was closer to 63% when it was calculated upon
the quantity of recovered 12.
Unfortunately, difficulties were also encountered when we
attempted to O-acylate the intermediary â-lithioalkoxy- and
â-hydroxyphenyl sulfone adducts of this union with various
acylating agents (e.g., BzCl, Ac2O, MsCl, etc.). Desulfo-
nylative elimination was therefore attempted directly on 13
using 6% Na/Hg amalgam. To our delight, the partially
deprotected alkene 14 was obtained as a single geometrical
isomer, but only in 32% yield. To complete the synthesis of
2, alcohol 14 was O-benzoylated and the product globally
deprotected using 40% aq HF in MeCN.
In summary, we have shown that it is possible to design
a bryostatin C-ring analogue that can bind to the CRD2 of
human PKC-R with an affinity of the same order of mag-
nitude as phorbol-13-acetate. Importantly, we have shown
that a 20-membered macrolactone structure is not a prereq-
uisite for effective analogue binding to the PKC-R CRD2.
Future work will focus on elucidating the structure of the
complex formed between 3 and the 13C/15N-labeled PKC-R
CRD2 by high-field NMR. NMR binding studies will also
be conducted with 2 and the CRD2, and analogues 2 and 3
will have their antitumor and PKC-modulating properties
fully evaluated.
1
A detailed description of the H resonance assignments
for the PKC-R CRD2 in 25% aq C2H3CN, and of the effects
of ligand binding, will be presented elsewhere.19 Residues
specifically affected by binding of PA and PB correspond
to those observed in the binding pocket of the TPA-PKCδ
CRD2 X-ray crystal structure.12 Bryostatin 1 also affects
several of these residues but clearly binds much more tightly
than PA or PB. Our observations conclusively demonstrate
for the first time eVer that the binding site of bryostatin 1
does indeed overlap with that of PA and PB. The binding of
our simplified analogue 3 also affected some, but not all, of
the residues that bind TPA, indicating that 3 occupies a part
of the same binding pocket. Chemical shift changes in the
amide region were observed over similar concentration
ranges for 3 and for PA and PB, showing that their binding
constants are very similar under our conditions. Our com-
bined data thus indicates that it is possible to design
bryostatin C-ring analogues that bind specifically to the
human PKC-R CRD2, in a novel manner, and with an affinity
of the same order of magnitude as those of the phorbol esters.
As the prelude to future protein NMR binding studies with
structurally more elaborate BC-fragments, we have recently
completed a synthesis of the novel BC-analogue 2. Impor-
tantly, our synthesis has highlighted a number of potential
problems that can be faced when using the Julia olefination
reaction to connect B- and C-ring bryostatin fragments
together.
Acknowledgment. We thank the BBSRC for support
through Project Grant No. 31/B09691 and for a studentship
to M.F. We also thank Merck Sharp & Dohme (Harlow)
and Pfizer (Sandwich) for additional financial support and
the London School of Pharmacy for HRMS measurements.
We thank Professor G. R. Pettit (Arizona State University)
for a sample of natural bryostatin 1.
Our route to 2 (Scheme 2) converted our previously
reported alcohol 920 into the aldehyde 11 in three steps.
Considerable experimentation was required before conditions
were identified that could successfully metalate the phenyl
1
Supporting Information Available: 500 MHz H and
(15) The bulk of the TPA molecule apparently serves as a hydrophobic
cap for the polar inside of the CRD2. Its binding to a phosphorylated PKC
favours its membrane translocation and insertion.
(16) Hale, K. J., Frigerio, M.; Manaviazar, S. Org. Lett. 2001, 3, 3791.
(17) (a) Ichikawa, S.; Hatanaka, H.; Takeuchi, Y. Ohno, S.; Inagaki, F.
J. Biochem. 1995, 117, 566. (b) Hommel, U.; Zurini, M.; Luytenm, M.
Nature Struct. Biol. 1994, 1, 383.
125 MHz 13C NMR and mass spectra of all new compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL027392U
(18) Groves, P.; Searle, M. S.; Waltho, J. P.; Williams, D. H. J. Am.
Chem. Soc. 1995, 117, 7958.
(19) Fillingham, I. J.; Barsukov, I. G.; Damblon, C. F.; Hale, K. J.;
Frigerio, M.; Manaviazar, S.; Hummersone, M. G.; Gescher, A.; Roberts,
G. C. K. Manuscript in preparation.
(20) Hale, K. J.; Hummersone, M. G.; Bhatia, G. S. Org. Lett. 2000, 2,
2189.
(21) Kageyama, M.; Tamura, T.; Nantz, M. H.; Roberts, J. C.; Somfai,
P.; Whritenour, D. C.; Masamune, S. J. Am. Chem. Soc. 1990, 112, 7407.
502
Org. Lett., Vol. 5, No. 4, 2003