A series of 23,24-dihydrodiscodermolide analogues with simplified lactone regions

Article abstract of DOI:10.1016/j.bmcl.2005.12.066

A collection of seven new 23,24-dihydrodiscodermolide analogues have been synthesized with modifications to the lactone ring, some of which show antiproliferative activities similar to discodermolide.

Full text of DOI:10.1016/j.bmcl.2005.12.066

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1961–1964
A series of 23,24-dihydrodiscodermolide analogues
with simplified lactone regions
Simon J. Shaw,* Kurt F. Sundermann, Mark A. Burlingame, Dan Zhang,
Joseph Petryka and David C. Myles
Kosan Biosciences, Inc., 3832 Bay Center Place, Hayward, CA 94545, USA
Received 2 December 2005; revised 16 December 2005; accepted 20 December 2005
Available online 18 January 2006
Abstract—A collection of seven new 23,24-dihydrodiscodermolide analogues have been synthesized with modifications to the
lactone ring, some of which show antiproliferative activities similar to discodermolide.
Ó 2005 Elsevier Ltd. All rights reserved.
24
The marine natural product (+)-discodermolide 1
(Fig. 1), isolated from the sponge Discodermia dissoluta,
is potently cytotoxic to human cancer cell lines.1 It has
been well established that it has a similar mechanism
of action to paclitaxel, namely the binding and stabilisa-
tion of microtubules, which leads to mitotic arrest and
ultimately apoptosis.2 Discodermolide also shows some
activity against cell lines that express the P-glycoprotein
efflux pump and are thus resistant to paclitaxel.
13
O
O
OH
OH
7
NH2
HO
5
HO
1
O
1
O
The organism responsible for the biosynthesis of 1, pre-
sumably a bacterial symbiont, has not been identified.
Thus, total chemical synthesis has been the only viable
option for obtaining useful quantities of this material.3
Indeed, a combination of two of these syntheses has
been used to generate the material required to support
a phase I clinical trial.4 However, the 13 stereocentres
and three cis double bonds make the manufacture of this
material a daunting proposition.
Figure 1. (+)-Discodermolide 1.
which the terminal double bond has been reduced to
yield a monoene.
Modification of the diene has a number of attractions; it
has been reported by the Novartis group on the parent
discodermolide where it was shown to result in similar
potency7 and it allows for a more facile construction
of the C-24 to C-19 section of the molecule through a
straightforward Wittig reaction resulting in improved
material throughput.
There have been a number of efforts to generate simpli-
fied analogues that maintain the biological activity of
discodermolide itself.5 Of these the most promising have
been modifications to the lactone region, in which a
number of stereocentres can be removed without
impacting the cytotoxicity.6 In this communication, we
report a group of new lactone analogues that have been
further simplified by replacing the diene with a system in
When considering a synthetic strategy, we choose to use
a similar approach to the Novartis group in which the
molecule was divided into three sections, all available
from the Smith common precursor 2 (Fig. 2).4 This
essentially combines the Smith coupling of the A and
B sections with the Paterson approach to the AB and
C coupling.8 However, because we were intending to
Keywords: Discodermolide; Analogues; Antitumor.
*
Corresponding author. Tel.: +1 510 731 5269; fax: +1 510 732
8401; e-mail: shaw@kosan.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.12.066

1962
S. J. Shaw et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1961–1964
the methyl ketone derived from the common precursor4
PMBO
N
OMe
using (+)-diisopinocampheylboron chloride generated
the complete carbon skeleton. Subsequent directed anti-
reduction of the resulting ketone and acidic deprotection
yielded 7 (Scheme 2).
OH
O
Smith common precursor, 2
With the monoene version of discodermolide 7 in hand
we began to look at modifications of the lactone region.
The Smith group has reported the 2,3-anhydro com-
pound to be particularly potent.9 Subsequently in col-
laboration with the Smith group we demonstrated that
the methyl substituents could be removed from the ring
while still maintaining strong potency.6a However, there
has been only one report in which the 3-hydroxyl group
was investigated in isolation. In this report, the 3-O-
acetyl compound showed activity superior to that of
discodermolide itself.10 The excellent activity of the
3-acetate justified the further investigation of this posi-
tion and the 3-methoxy- and 3-deoxy-23,24-dihydrodis-
codermolide derivatives, 8 and 9, were targeted. Both of
these compounds are accessed through the common pre-
cursor 2 by O-methylation using the Meerwein reagent
and Barton–McCombie deoxygenation via the pentaflu-
orophenylthiocarbonate,11 respectively, followed by
conversion to the corresponding methyl ketones 10
and 11 (Scheme 3) using standard procedures.4
I
PMBO
OTBDMS
Modified A Fragment, 3
O
O
OH
OH
I
NH2
OTBDMS
HO
OPMB
Analogue
B Fragment, 4
OH
Analogue
O
Figure 2. Approach to discodermolide analogues.
use a monoene instead of the diene it was considered
that this could be introduced prior to the coupling reaction
with the Smith B piece. In this way, late stage modifica-
tions were avoided making the route more convergent.
Both of these substrates underwent aldol reaction in
good yield with excellent selectivity and were taken
through to obtain the desired discodermolide analogues.
In the case of the 3-deoxy compound, it was found
that in addition to the expected 3-deoxy-23,24-dihydro-
To obtain the modified A fragment 3, we used chemistry
similar to that recently reported by the Smith group.6b
This route could be easily carried out on a multigram
scale in good yield.
As expected, coupling to the B piece 4 was not affected
by the monoene and yielded the complete C-24 to
C-12 carbon backbone 5. Using the standard conditions
reported by the Paterson group this material was con-
verted to the enal 6 in six steps (Scheme 1, see Support-
ing information for further details).8c
N
MeO
O
O
O
O
O
OH
OH
TBDMS
i. (+)-Ipc2BCl
Et3N, Et2O
NH2
6
HO
O
ii. NaB(OAc)3H
iii. HCl, MeOH
We were now in a position to begin generating lactone
analogues, however, we first generated the previously
described 23,24-dihydrodiscodermolide 7 to validate this
compound in our own assays. Thus, aldol reaction with
HO
7
O
Scheme 2. Completing the synthesis of 23,24-dihydrodiscodermolide 7.
4
3
Me3OBF4,
proton sponge
PMBO
OTBDMS
OTBDMS
OPMB
PMBO
N
t-BuLi, ZnCl2
Et2O
2
OMe
R
O
or i. C6F5OCSCl
ii. Bu3SnH, AIBN
5
R = OMe
R = H
i. DDQ
i. H2, Pearlmans cat
ii. (COCl)2, Me2SO, Et3N
ii. TEMPO Oxidation
iii. Still-Gennari
O
O
NH2
OTBDMS
OTBDMS
N
OMe
iv. Carbamate Formation
v. DiBAL-H
vi. Dess-Martin Oxidation
iii. MeMgBr
iv. (COCl)2, Me2SO, Et3N
O
R
O
6
R = OMe 10
R = H 11
For full details see supplementary material
O
Scheme 1. Coupling of the modified A fragment and conversion to the
enal.
Scheme 3. Generation of the 3-methoxy and 3-deoxy methyl ketones,
10 and 11, respectively.

S. J. Shaw et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1961–1964
1963
discodermolide 9, the open chain methyl ester 12 was
also obtained (Fig. 3). Presumably the lack of a substi-
tuent at the 3-position makes the acid catalysed cycliza-
tion less favourable.
lin mutation making it resistant to paclitaxel. It was
encouraging to see that the 23,24-dihydrodiscodermo-
lide
discodermolide 1 in line with what has been reported
7 is slightly more potent than the parent
(Table 1).7
Next, we examined the nature of the lactone itself by
first converting it to a carbamate 13. In this case, the
necessary azido methyl ketone 14 could be easily ob-
tained from Roche ester 15 (Scheme 4). The carbamate
was installed prior to removal of the silyl groups by
Staudinger reduction and treatment with carbonyldiim-
idazole. When the same approach was attempted to
form the corresponding carbonate from methyl ketone
16, the final deprotection removed the carbonate result-
ing in the triol 17. Attempted reinstallation of the
carbonate at this stage was not successful.
The two 3-position analogues show excellent activity.
However, the open chain form of the 3-deoxy com-
pound 12 is not active. This contrasts with the methyl es-
ter of discodermolide, which was reported to have
nanomolar activity.12,13 The activity of 3-methoxy com-
pound 8 is consistent both with the reports from the
Curran and Day groups in which a simplified disco-
dermolide analogue with a methoxymethyl ether at the
3-position showed a slightly improved activity over the
parent molecule5c and the 3-O-acetyldiscodermolide
reported by Harbor Branch.10
To complete the series, the simple butyrolactone 18 sim-
ilar to that reported in the diene series was also synthe-
sized along with its 7,5-epimer 19.6a
Both the carbamate and the triol compounds 13 and 17,
respectively, do not show significant cytotoxicity. This
lack of activity suggests that the nature of the carbonyl
of the lactone ring is important. As expected the butyro-
lactone 18 with the stereochemistry found in disco-
dermolide is an active compound.
The compounds were tested for their antiproliferative
activity in the standard cell lines, including the MDR
cell line NCI/ADR, which expresses the P-glycoprotein
efflux pump and the A549/t12 cell line which has a tubu-
The three most active compounds were further investi-
gated in a range of resistant leukaemia cell lines. All cell
lines express the P-glycoprotein efflux pump but to vary-
ing degrees. All compounds are active with each com-
pound showing improved activity over discodermolide
in at least one cell line (Table 2).
O
O
OH
OH
NH2
In conclusion, we have generated a series of new ana-
logues of discodermolide in the 23,24-dihydro series,
of which three are as good or better than discodermo-
lide. This scaffold does not appear to affect the activity
of the compounds and allows for material to be moved
HO
O
HO
O
HO
MeO
MeO
O
OH
Table 1. Antiproliferative activity for compounds 9–16 and disco-
dermolide 1
O
O
9
8
12
Compound
Antiproliferative activity, GI50 (nM)
MCF-7
NCI/ADR
A549
A549t12
HO
O
HO
HO
O
HO
1
7
28
6.2
240
190
22
7
30
nd
HN
OH OH
O
8
3.8
5.8
250
140
6.5
11
12
50
9
O
O
O
12
13
17
18
19
1000
110
1000
6.4
10,000
10,000
10,000
730
4000
520
3000
46
4000
1000
5000
63
13
17
18
19
Figure 3. Compounds in the study.
2000
10,000
4000
10,000
Tf2O,
pyridine
MeMgBr
HO
HO
OMe
OMe
N3
OMe
N3
Table 2. Antiproliferative activity for compounds 10, 11, 15 and
discodermolide 1 in a range of resistant cell lines
NaN3
O
O
O
O
15
15
14
Compound
Antiproliferative activity, GI50 (nM)
TBDMSOTf,
lutidine
CCRF- CCRF-
CEM
CCRF-
CEM/PTX CEM/VBL CEM/VP16
CCRF-
TBDMSO
MeMgBr
O
1
8
16
92
79
158
416
52
8.4
12
7.1
16
4.4
8.6
6.5
9
18
65
285
209
3807
Scheme 4. Generation of the methyl ketones 14 and 16, for forming
the carbamate 13 and triol 17, respectively.

1964
S. J. Shaw et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1961–1964
W.-C.; Wang, R.-M.; Waykole, L.; Xu, D. D.; Xue, S. Org.
Process Res. Dev. 2004, 8, 92, and following papers.
5. (a) Burlingame, M. A.; Shaw, S. J.; Sundermann, K. F.;
Zhang, D.; Petryka, J.; Mendoza, E.; Lui, F.; Myles, D.
C.; LaMarche, M. J.; Hirose, T.; Freeze, B. S.; Smith, A.
B., III Bioorg. Med. Chem. Lett. 2004, 14, 2335; (b)
Curran, D. P.; Furukawa, T. Org. Lett. 2002, 4, 2233; (c)
Choy, N.; Shin, Y.; Nguyen, P. Q.; Curran, D. P.;
Balachandran, R.; Madiraju, C.; Day, B. W. J. Med.
Chem. 2003, 46, 2846.
through the synthetic sequence more effectively. From
these compounds, it appears the lactone ring is impor-
tant for activity. Changing to a carbamate or opening
the ring lowers the activity. The 3-methoxy and 3-deoxy
compounds are both more potent than discodermolide
and have a better profile in the resistant cell lines. It is
noteworthy that the most simplified analogue in this ser-
ies, the butyrolactone 18, is also one of the most active
compounds.
6. (a) Shaw, S. J.; Sundermann, K. F.; Burlingame, M. A.;
Myles, D. C.; Freeze, B. S.; Xian, M.; Brouard, I.; Smith,
A. B., III J. Am. Chem. Soc. 2005, 127, 6532; (b) Smith, A.
B., III; Xian, M. Org. Lett. 2005, 6, 5229.
Acknowledgments
7. Palmero, M. G.; Bair, K. W.; Chen, W.; Guo, Q.; Lassota,
P. T.; Sorrensen, E.; Wang, R. M.; Kinder, Jr., F. R.,
Abstracts of Papers, 224th Meeting of the American
Chemical Society, Boston, MA; American Chemical
Society: Washington, DC, 2002; Abstract 791.
8. (a) Smith, A. B., III; Kaufman, M. D.; Beauchamp, T. J.;
LaMarche, M. J.; Arimoto, H. Org. Lett. 1999, 1, 1823;
(b) Smith, A. B., III; Beauchamp, T. J.; LaMarche, M. J.;
Kaufman, M. D.; Qui, Y.; Arimoto, H.; Jones, D. R.;
Kobayashi, K. J. Am. Chem. Soc. 2000, 122, 8654; (c)
Paterson, I.; Florence, G. J.; Gerlach, K.; Scott, J. P.;
Sereining, N. J. Am. Chem. Soc. 2001, 123, 9535.
9. Martello, L. A.; LaMarche, M. J.; He, L.; Beauchamp, T. J.;
Smith, A. B., III; Horwitz, S. B. Chem. Biol. 2001, 8, 843.
10. Gunasekera, S. P.; Longley, R. E.; Insbrucker, R. A. J.
Nat. Prod. 2001, 64, 171.
We thank Fenghua Liu and Wei Ma for antiprolifera-
tive assays.
Supporting information
A full scheme outlining the synthesis of 8 from the cou-
pling of 3 with 4 and full experimental details are avail-
able in the supporting information. Supplementary data
associated with this article can be found, in the online
version, at doi:10.1016/j.bmcl.2005.12.066.
References and notes
11. Ding, Y.; Wang, J.; Abboud, K. A.; Xu, Y.; Dolbier,
W. R., Jr.; Richards, N. G. J. J. Org. Chem. 2001, 66,
6381.
12. Kinder, F. R.; Bair, K., W.; Chen, W.; Florence, G. L.;
Francavilla, C.; Geng, P.; Gunerasekera, S.; Lassota, P.
T.; Longley, R.; Palermo, M.; Paterson, I.; Pomponi, S.;
Ramsey, T. M.; Rogers, L.; Sabio, M.; Sereinig, N.;
Sorensen, E.; Wang, R.; Wright, A. 93rd Annual Meeting,
San Francisco, CA; American Association for Cancer
Research, 2002; Poster 3650.
1. Gunasekera, S. P.; Gunasekera, M.; Longley, R. E.;
Schulte, G. K. J. Org. Chem. 1990, 55, 4912; Correction .
J. Org. Chem. 1991, 56, 1346.
2. ter Haar, E.; Kowalski, R. J.; Hamel, E.; Lin, C. M.;
Longley, R. E.; Gunasekera, S. P.; Rosenkranz, H. S.;
Day, B. W. Biochemistry 1996, 35, 243.
3. For a review, see: Paterson, I.; Florence, G. J. Eur. J. Org.
Chem. 2003, 2193.
4. Mickel, S. J.; Sedelmeier, G. H.; Niederer, D.; Daeffler, R.;
Osmani, A.; Schreiner, K.; Seeger-Weibel, M.; Berod, B.;
Schaer, K.; Gamboni, R.; Chen, S.; Chen, W.; Jagoe, C. T.;
Kinder, F. R., Jr.; Loo, M.; Prasad, K.; Repic, O.; Shieh,
13. We speculate that the discodermolide methyl ester cyclises
to discodermolide in the assay but the 3-deoxy compound
does not.