Synthesis and biological activity of 7,8,9-trideoxy- and 7R DesTHP-peloruside A

Article abstract of DOI:10.1002/chem.201301796

The stereoselective syntheses of 7,8,9-trideoxypeloruside A (4) and a monocyclic peloruside A analogue lacking the entire tetrahydropyran moiety (3) are described. The syntheses proceeded through the PMB-ether of an ω-hydroxy β-keto aldehyde as a common intermediate which was elaborated into a pair of diastereomeric 1,3-syn and -anti diols by stereoselective Duthaler-Hafner allylations and subsequent 1,3-syn or anti reduction. One of these isomers was further converted into a tetrahydropyran derivative in a high-yielding Prins reaction, to provide the precursor for bicyclic analogue 4. Downstream steps for both syntheses included the substrate-controlled addition of a vinyl lithium intermediate to an aldehyde, thus connecting the peloruside side chain to C15 (C13) of the macrocyclic core structure in a fully stereoselective fashion. In the case of monocyclic 3 macrocyclization was based on ring-closing olefin metathesis (RCM), while bicyclic 4 was cyclized through Yamaguchi-type macrolactonization. The macrolactonization step was surprisingly difficult and was accompanied by extensive cyclic dimer formation. Peloruside A analogues 3 and 4 inhibited the proliferation of human cancer cell lines in vitro with micromolar and sub-micromolar IC₅₀ values, respectively. The higher potency of 4 highlights the importance of the bicyclic core structure of peloruside A for nM biological activity. I need the THP ring: Convergent stereoselective syntheses of mono- and bicyclic analogues 2 and 3 of the natural product mitosis inhibitor peloruside A (1) have been developed based on macrocyclization by ring-closing metathesis (2) or macrolactonization (3). Both analogues are less potent than natural 1, but the activity difference is distinctly less pronounced for bicyclic analogue 3. The bicyclic core structure of peloruside A thus appears to be a necessary, but not sufficient provision for peloruside A-like activity (IC₅₀ = 2: 1-5 μM, 3: 100-300 nM). Copyright

Full text of DOI:10.1002/chem.201301796

FULL PAPER
DOI: 10.1002/chem.201301796
Synthesis and Biological Activity of 7,8,9-Trideoxy-
and 7R DesTHP-Peloruside A
Christoph W. Wullschleger,[a] Jꢀrg Gertsch,[b] and Karl-Heinz Altmann*[a]
Abstract: The stereoselective syntheses
of 7,8,9-trideoxypeloruside A (4) and a
monocyclic peloruside analogue
provide the precursor for bicyclic ana-
logue 4. Downstream steps for both
syntheses included the substrate-con-
trolled addition of a vinyl lithium inter-
mediate to an aldehyde, thus connect-
ing the peloruside side chain to C15
(C13) of the macrocyclic core structure
in a fully stereoselective fashion. In the
case of monocyclic 3 macrocyclization
was based on ring-closing olefin meta-
thesis (RCM), while bicyclic 4 was cy-
clized through Yamaguchi-type macro-
lactonization. The macrolactonization
step was surprisingly difficult and was
accompanied by extensive cyclic dimer
formation. Peloruside A analogues 3
and 4 inhibited the proliferation of
human cancer cell lines in vitro with
micromolar and sub-micromolar IC50
values, respectively. The higher potency
of 4 highlights the importance of the
bicyclic core structure of peloruside A
for nm biological activity.
A
lacking the entire tetrahydropyran
moiety (3) are described. The syntheses
proceeded through the PMB-ether of
an w-hydroxy b-keto aldehyde as a
common intermediate which was ela-
borated into a pair of diastereomeric
1,3-syn and -anti diols by stereoselec-
tive Duthaler–Hafner allylations and
subsequent 1,3-syn or anti reduction.
One of these isomers was further con-
verted into a tetrahydropyran deriva-
tive in a high-yielding Prins reaction, to
Keywords: natural products · pe-
loruside A · structure–activity rela-
tionships · total synthesis · tubulin
Introduction
formed on a limited number of natural congeners of 1,[9,10b]
but the overall SAR picture for peloruside-type structures is
still spotty at this point.
Peloruside A (1) is a polyketide-derived 16-membered mac-
rolide that was isolated from the marine sponge Mycale
hentscheli by Northcote and co-workers and shown to exhib-
it potent in vitro antitumor activity.[1] The compound was
subsequently found to be a microtubule-stabilizing agent
(MSA)[2] with a tubulin binding site distinct from that of
most other natural product MSAs,[3] which almost uniformly
bind to the taxol site on b-tubulin.[4,5] As a result, peloruside
A (1) exhibits synergistic effects with a number of taxol site
agents on tubulin assembly[6] and cancer cell growth.[7]
Peloruside A (1) has been a frequent target for total syn-
thesis;[8,9] in addition, the synthesis or semisynthesis of a
number of (simplified) analogues of 1 has been reported for
SAR studies.[2,10,11] Biological studies have also been per-
We have recently reported the stereoselective synthesis of
the 14-membered monocyclic peloruside A analogue 2, as
part of a program directed at the exploration of the impor-
tance of the pyranose ring in the bicyclic core structure of
peloruside A (1) and its particular substitution pattern.[11]
[a] Dr. C. W. Wullschleger, Prof. Dr. K.-H. Altmann
Department of Chemistry and Applied Biosciences
Institute of Pharmaceutical Sciences
Swiss Federal Institute of Technology (ETH) Zꢀrich
HCI H405, Wolfgang-Pauli-Str. 10
While
2 retained measurable antiproliferative activity
8093 Zꢀrich (Switzerland)
Fax : (+41)44-6331369
E-mail: karl-heinz.altmann@pharma.ethz.ch
against A549 cells, it was generally >1000-fold less potent
than 1; it remained unclear, however, to what extent the bio-
logical activity of 2 was affected by the configuration of the
C7 chiral center (corresponding to C9 in natural 1), al-
though calculations had indicated that in order for the C7-
hydroxyl group to mimic the anomeric hydroxyl group in
peloruside A (1) the preferred configuration would have to
[b] Prof. Dr. J. Gertsch
University of Bern, Institute of Biochemistry
and Molecular Medicine
Bꢀhlstrasse 28, 3012 Bern (Switzerland)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201301796.
Chem. Eur. J. 2013, 19, 13105 – 13111
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
13105

be 7S (as in 2) rather than 7R.[12] In order to clarify this
question experimentally, we have now also prepared 7R ana-
logue 3 and we have assessed its effects on cancer cell prolif-
eration. At the same time we felt that the lack of significant
antiproliferative activity of 2 might point to the need for a
conformationally more restricted and/or larger sized macro-
cycle as a prerequisite for potent biological activity (see also
ref. [2]). As a consequence, we have started to investigate
the stepwise reconstruction of natural 1 from monocyclic
analogues 2/3 and to assess the changes in activity that
would be associated with the restoration of individual mod-
ules or substituents of the peloruside structure that are not
present in 2/3. In a first step in this process we have investi-
gated the greatly simplified bicyclic peloruside A analogue
4, which served to assess the
importance of the bicyclic core
structure of 1 for biological ac-
tivity, independent of any sub-
stituents on the tetrahydropyr-
Scheme 1. Retrosynthesis of 7R desTHP-peloruside A (3).
an (THP) ring. In an independ-
ent parallel study Zimmermann
et al. have investigated the re-
ether with DDQ[14] followed by Dess–Martin oxidation of
the ensuing primary alcohol gave the required aldehyde 10
in 58% overall yield for the 4-step sequence from homoal-
lylic alcohol 13.
lated analogue 5;[10d] this com-
pound was found to exhibit only moderate antiproliferative
activity (IC50 values between 10 and 20 mm).
In order to ascertain the predicted 1,3-syn relationship of
the diol obtained by DIBALH reduction of 13, the former
Results and Discussion
Synthesis of 7R DesTHP-peloruside A (3): In analogy to
our previous work on analogue 2,[11] the core element of our
synthetic strategy towards 7R desTHP-peloruside A (3) was
the anticipated RCM-based macrocyclization of diene 6,
which would be followed by TBS removal, double-bond re-
duction and (if possible simultaneous) cleavage of the
benzyl protecting groups (Scheme 1). Diene 6 would be ob-
tained by esterification of acid 8 with alcohol 7; the latter
was to be produced through addition of a vinyl metal inter-
mediate derived from vinyl iodide 9 to aldehyde 10, which
in turn would be derived from b-keto aldehyde 11 by asym-
metric allylation and subsequent 1,3-syn reduction as the de-
fining steps. The synthesis of aldehyde 11, vinyl iodide 9,
and acid 8 has been described.[11]
The implementation of these concepts in the first step en-
tailed the asymmetric allylation of aldehyde 11 under
Duthaler–Hafner conditions (Scheme 2).[13] When the allylti-
tanation of 11 was performed with a 1.4-fold excess of 12
the desired 7R homoallylic alcohol 13 was obtained in 93%
yield and with excellent diastereoselectivity (d.r. >25:1); an
excess of 12 was crucial for complete conversion of the start-
ing aldehyde 11.
After treatment of 13 with DIBALH in THF the desired
syn-diol could be isolated in excellent yield (84%). (The re-
action proceeded with a d.r. of 15:1, the minor isomer could
be removed by FC.) The reduction product was then con-
verted into the corresponding bis-TBS-ether 14 by reaction
with TBSOTf; subsequent oxidative cleavage of the PMB-
Scheme 2. a) 12, Et2O, ꢀ788C, 2.5 h, 93% (d.r. >25:1); b) slow addition
of DIBALH (1.0 equiv over 80 min), THF, ꢀ788C, then 4.0 equiv
DIBALH, ꢀ788C, 2 h, 84%; c) TBSOTf, 2,6-lutidine, CH2Cl2, ꢀ788C !
RT, 3.5 h, 95%; d) DDQ, CH2Cl2/H2O 12:1, 08C, 3 h, 93%; e) DMP,
CH2Cl2, RT, 90 min, 78%; f) 2.1 equiv 9, 4.0 equiv tBuLi, Et2O, ꢀ788C,
30 min, then 1.0 equiv 10, ꢀ788C, 18 h; 7: 58%, 15: 18% (based on 9); g)
8, 2,4,6-trichlorobenzoyl chloride, Et3N, toluene, RT, 1 h, then 7, DMAP,
RT, 90 min, 83 %.
13106
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13105 – 13111

Peloruside A Derivatives
FULL PAPER
was converted into the corresponding acetonide 16 by treat-
ment with CSA in 2,2-dimethoxypropane. The 13C chemical
shifts for the ketal methyl groups of 16 (19.47 ppm, axial
methyl group; 30.21 ppm, equatorial methyl group) and for
with HF·pyridine in THF gave partially deprotected macro-
lactone 18 in 89% yield. When submitting 18 to catalytic hy-
drogenation conditions (Pd/C, H2 at ambient pressure in
AcOEt), complete decomposition was observed within 2 h.
In contrast, the endocyclic double bond could be selectively
reduced with diimide (PADA, AcOH, CH2Cl2, reflux tem-
perature, 30 h)[21] to deliver the saturated macrolactone 19
in excellent yield (83%); interestingly, however, the reaction
appeared to be significantly more sluggish than the same
transformation with the 7S isomer of 18, that is, longer reac-
tion times (30 h vs. 17 h) and more equivalents of PADA/
AcOH (260 equiv vs 80 equiv) were required to drive the re-
action to completion. The final benzyl ether cleavage from
19 was first attempted by catalytic hydrogenation in AcOEt
(Pd/C, H2, room temperature, ambient pressure); while
these conditions had been successfully employed for the de-
benzylation of the 7S isomer of 19, for 19 they only led to
decomposition. Likewise, treatment of 19 with BCl3 did not
produce any of the desired des-THP-peloruside 3. These
problems could be finally overcome by changing the solvent
for the hydrogenation from AcOEt to ethanol, which led to
a significant enhancement in reaction rate. At the same time
decomposition pathways could be suppressed by limiting the
amounts of Pd/C to 2–4.5%. Under these conditions the de-
benzylation of 19 went to completion within ca. 7 h at room
temperature and ambient pressure to give 3 in 73% yield.
Only minor amounts of polar side products were formed
(TLC: <5–10%) and the trisubstituted side chain double
bond remained unaffected (according to MS-based reaction
monitoring).
its quaternary carbon (98.40 ppm) are in perfect agreement
with the predicted shifts for acetonides of syn-1,3-diols.[15]
Reaction of aldehyde 10 with the vinyllithium reagent de-
rived from vinyl iodide 9 by treatment with tBuLi at
ꢀ788C[11] gave the desired 13S alcohol 7 (numbering for 3)
as the sole isomer (in 58% yield), but the product was con-
taminated with methyl ketone 15 (18%, based on 9) which
could not be removed chromatographically.[16] However, the
stereoselectivity of this transformation was not immediately
recognized, as the signals originating from 15 in the NMR
spectra of the 7/15 mixture were initially believed to arise
from the 13R diastereoisomer of 7. The mixture was thus
oxidized to enone 17, which was then re-reduced with cate-
cholborane in the presence of the (R)-Corey–Bakshi–Shiba-
ta (CBS) catalyst,[17] to provide 7 as a single isomer and free
of ketone 15 (Scheme 3); the latter could be readily re-
moved by FC at the enone stage.
Scheme 3. a) DMP, NaHCO3, CH2Cl2, RT, 90 min, 89% (based on the
amount of 7 in the starting mixture); b) catecholborane, (R)-B-Me-CBS
oxazaborolidine, toluene, ꢀ78 ! 08C, 7 h, 87%.
Scheme 4. a) Grubbs II catalyst, 1,2-dichloroethane, 808C, 5 h, 60%; b)
HF·py, THF, RT, 19 h, 89%; c) PADA, AcOH, CH2Cl2, reflux, 30 h,
83%; d) Pd/C, H2 (atmospheric pressure), EtOH, RT, 6.5 h, 73%.
The S configuration of the chiral center formed in the
CBS-reduction of enone 17 was established by Mosher ester
analysis;[18] this stereochemical outcome is in perfect agree-
ment with a large body of information on the stereodirect-
ing properties of CBS catalysts.[17] Importantly, the product
of the CBS-reduction was spectroscopically indistinguishable
from the allylic alcohol obtained in the reaction between al-
dehyde 10 and the vinyllithium reagent derived from vinyl
iodide 9. The esterification of 7 and carboxylic acid 8 under
Yamaguchi conditions[19] then gave the desired diene 6 in
83% yield (Scheme 2).
As shown in Scheme 4, treatment of 6 with Grubbs
second-generation catalyst[20] in 1,2-dichloroethane at 808C
provided the E-configured macrolactone as the only isolable
cyclization product in yields around 60%, thus reproducing
the outcome of the corresponding step in the synthesis of
analogue 2.[11] Subsequent cleavage of the two TBS-ethers
Synthesis of 7,8,9-trideoxy-peloruside A (4): Due to the
presence of the tetrahydropyran ring in 4 and its particular
positioning, RCM-mediated macrocyclization at the site ex-
ploited in the synthesis of 3 (and also 2[11]) was no longer
feasible for 4. As a consequence, the synthesis of 4 was to
be based on ring-closure by macrolactonization of an appro-
priately protected seco-acid, such as 20 (Scheme 5).[22] Seco-
acid 20 was planned to be accessed through a stereoselective
glycolate aldol reaction with aldehyde 21; the latter was to
be obtained from aldehyde 22 and vinyl iodide 9, in analogy
to the formation of 7 from 9 and 10 (Scheme 2). Lastly, the
THP-containing fragment 22 was to be accessed from homo-
allylic alcohol 23 and aldehyde 24[23] via Prins reaction and
subsequent functional group manipulations. The synthesis of
Chem. Eur. J. 2013, 19, 13105 – 13111
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13107

K.-H. Altmann et al.
Scheme 6. a) 24 (2.4 equiv), SmI2 (16 mol%), THF, ꢀ108C, 1 h, 88 %
(d.r. 15:1); b) 26, CH2Cl2, 08C, 4 h 45 min, 88% (8:1 mixture of 27/28);
c) DIBALH, CH2Cl2, ꢀ1008C, 2.5 h, then pyridine, DMAP, Ac2O,
ꢀ1008C, 5 h, ꢀ788C, 14 h, 82% (d.r. 5:3); d) CeCl3, LiI, CH2Cl2, RT, 7 h
15 min, quant. (30/31 11:2); e) Pd/C, H2 (atmospheric pressure),
NaHCO3, MeOH/AcOEt 4:1, RT, 2 h, 90%; f) DDQ, CH2Cl2/H2O 10:1,
08C, 50 min, 80%; g) DMP, CH2Cl2, 08C, 4 h, 83%.
Scheme 5. Retrosynthesis of 7,8,9-trideoxy-peloruside A (4).
building block 23 had already been established as part of
our work on analogue 2 (and involved the stereoselective al-
lylation of aldehyde 11 with ent-12).[11]
excellent results were obtained for the construction of the
desired 2,6-cis-disubstituted-4-halotetrahydropyrans 30/31 by
treatment of 29 with CeCl3 and LiI as Lewis acids.[27] The re-
action could be conducted at room temperature in CH2Cl2
and led to a 11:2 mixture of diastereoisomers 30 and 31 in
quantitative yield (>95%; based on the amount of 29 in the
29/28 mixture). At this stage, 28 could be readily removed
by FC.
The expected cis-arrangement of the substituents at posi-
tions 2 and 6 of the THP ring in 30 and 31 was confirmed by
NOE measurements.[28] A clear NOE was detectable be-
tween H5 and H9 for both isomers; in contrast, NOEs be-
tween H7 and H5/H9 were only observed for 31, thus point-
ing to an equatorial orientation of the iodo substituent on
the THP ring in this isomer. The axial selectivity of the
Prins cyclization reaction, with 30 as the major isomer, is
further supported by chemical shifts of d = 4.87 and
4.21 ppm for the H7 proton in 30 and 31, respectively. These
numbers are clearly indicative of an equatorial orientation
of the H7 proton in 30, while its orientation must be axial in
31.[26]
The attempted reductive removal of the C7 iodo group by
radical dehalogenation (Bu3SnH/AIBN)[29] was associated
with complete decomposition of starting material. In con-
trast, hydrodehalogenation[30] of the 30/31 mixture with Pd/C
and H2 led to the desired des-iodo-30/31 as a single isomer
in high yield (90%). For the reaction to proceed smoothly,
ca. 5 equiv of NaHCO3 had to be added to the reaction mix-
ture, in order to trap hydroiodic acid and prevent TBS-ether
cleavage. Finally, PMB-removal with DDQ followed by
DMP oxidation of the resulting primary alcohol furnished
As illustrated in Scheme 6, 1,3-reduction of 23 with alde-
hyde 24 under Evans–Tishchenko conditions[24] led to the re-
quired 1,3-anti product 25 in high yield (88%) and with ex-
cellent diastereoselectivity (d.r. 15:1). The predicted 1,3-anti
relationship at C9/C11 was ascertained by the conversion of
25 into the corresponding diol and the derived acetonide;
the analytical data for both compounds were in perfect
agreement with those reported previously.[11] While TBS
protection of 25 with TBSOTf and 2,6-lutidine resulted in
extensive acyl migration (2:3 mixture of 27 and 28; 75%
yield), reaction with N-tert-butyldimethylsilylpyridinium tri-
flate (26) led to a 8:1 mixture of the desired C11-TBS ether
27 and the undesired C9-TBS regioisomer 28 in a total yield
of 88%; 27 and 28 proved to be chromatographically insepa-
rable.
With ester 27 in hand we then turned our attention to the
Rychnovsky segment coupling Prins cyclization.[25] To this
end the mixture of 27/28 was treated with a slight excess of
DIBALH in CH2Cl2 at ꢀ1008C to produce, after in situ ace-
tylation of the ensuing aluminated hemiacetal with acetic
anhydride in the presence of pyridine and DMAP,[25b] a-ace-
toxy ether 29 (Scheme 6). The latter was obtained in 82%
yield as a 5:3 mixture of diastereoisomers at C5. Interesting-
ly, the undesired regioisomer 28 did not react under these
conditions and was recovered from the reaction mixture un-
changed (although it was not separated from 29). Initial at-
tempts to perform the Lewis acid-promoted cyclization of
29 under standard conditions (e.g., with TMSI or SnBr4)[26]
resulted in decomposition (with TMSI) or the formation of
only traces of the desired product (with SnBr4). However,
13108
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13105 – 13111

Peloruside A Derivatives
FULL PAPER
aldehyde 22 in 60% overall yield for the 3-step sequence
from 30/31.
Metalation of vinyl iodide 9 and subsequent addition of
the resulting vinyllithium species to aldehyde 22 according
to the protocol that had been employed for the preparation
of 7 (Scheme 2) provided the desired 15S-allylic alcohol 32
as
a single isomer in 65–73% yield (based on 22)
(Scheme 7). As for 7, however, the product was again conta-
minated with methyl ketone 15, which was obtained in 13%
yield (based on 9); 32 and 15 could not be separated and
were carried into the next step as a mixture. Treatment of
the 32/15 mixture with 3,4-dihydro-2H-pyran (DHP) and
20 mol% PPTS produced acetal 33 as a mixture of two dia-
stereoisomers in 87% yield (based on the amount of 32 in
the starting material) without affecting the secondary TBS-
ether. The two THP isomers could be separated by FC; they
were processed into seco-acid 20 either separately, in order
to facilitate the interpretation of NMR spectra, or as the 1:1
mixture. No significant differences in yields were observed
between reactions with single isomers and the 1:1 mixtures.
The TBDPS group in 33 was then selectively removed with
refluxing methanolic NaOH (10%) in 71–88% yield.[31] Sub-
sequent oxidation of the liberated hydroxyl group with
DMP gave aldehyde 21, which was submitted to an Evans
aldol reaction[32] with a 10-fold excess of the dibutylboron
enolate 34 (derived from the corresponding glycolyl imide
and Bu2BOTf). This reaction afforded the desired aldol
product as a single isomer in 88% yield. The use of a large
excess of enolate 34 was found to be crucial for the effec-
tiveness of the reaction, as lower amounts of 34 resulted in
incomplete conversion of aldehyde 21. Due to the large
excess of 34, the aldol product could only be obtained as a
1:9 mixture with the starting glycolyl imide; however, the
latter could be readily removed by FC after methylation of
the free hydroxyl group with Meerwein salt (Me3OBF4).[33]
Selective cleavage of the THP-acetal with MgBr2[34] followed
by removal of the chiral auxiliary with LiOH/H2O2 then fur-
nished seco-acid 20 in 51% overall yield for the 4-step se-
quence from aldehyde 21.
At this stage we were confident that the completion of
the synthesis of 4 would be reasonably straightforward.
However, very surprisingly, the macrolactonization of seco-
acid 20 turned out to be highly challenging. Among the vari-
ous cyclization methods investigated the Yamaguchi proto-
col[19] proved to be the most effective, even if the desired
macrolactone 36 was obtained in only moderate yield (25%,
3.1 mmol) and the reaction produced an inseparable ca. 2:1
mixture (on a molar basis) of 36 and diolide 37 (25%,
1.5 mmol) (Scheme 8). Based on TLC, the starting material
was completely consumed and, in addition to 36 and 37, was
converted into unidentified polar side products; the latter
may be speculated to be non-cyclized oligomers.[35] Subse-
quent TBS-deprotection of the 36/37 mixture with HF·pyri-
dine afforded partially deprotected macrolactones 38 and 39
as single compounds in 80 and 85% yield, respectively.
Compound 38 and 39 showed dramatically different polari-
ties and thus could be easily separated by FC on silica. Final
Scheme 7. a) 2.5 equiv 9, 4.2 equiv tBuLi, Et2O, ꢀ788C, 40 min, then
1.0 equiv 22, ꢀ788C, 14 h, 65–73%; plus 15, 13% (based on 9); b) DHP,
PPTS, CH2Cl2, RT, 6 h, 87%; c) NaOH (10%), MeOH/THF 10:1, reflux,
14 h, 71–88%; d) DMP, CH2Cl2, 08C, 90 min, 86–92%; e) 34, CH2Cl2,
ꢀ788C, 2.5 h, 88%; f) Me3OBF4, proton sponge, CH2Cl2, 08C, 2 h, 91 %;
g) MgBr2·OEt2, Et2O, RT, 1 h, 77%; h) LiOH, H2O2, THF/H2O 4:1, 08C,
1 h 15 min, 83%.
deprotection of 38 and 39 with H2 and Pd/C in ethanol pro-
vided the targeted bicyclic peloruside A analogue 4 and dio-
lide 40 in 77 and 50% yield, respectively. No complications
arose in this step from the presence of the olefinic double
bond in the side chain.
Biological evaluation: The in vitro antiproliferative activity
of desTHP-peloruside A (3) and 7,8,9-trideoxy-peloruside A
(4) was evaluated in three human cancer cell lines and the
results of these studies are summarized in Table 1. Diolide
40 did not show any appreciable antiproliferative activity
(IC50 >10 mm). Monocyclic peloruside A analogue 3 dis-
played single digit mm IC50 values against all three cell lines
investigated. The compound, thus, is more potent in this cell
line panel than its previously reported 7S isomer 2,[11] al-
though it still remains several-hundred fold less active than
natural peloruside A (1). (IC50 values of 27 nm[7a] and 7-
32 nm[2,7a,9] have been reported for 1 against 1A9 human
ovarian carcinoma cells and HL-60 leukemia cells, respec-
tively; the IC50 against the MCF-7 breast carcinoma cell line
was 4.9 nm[36]). Importantly, bicyclic analogue 4 inhibits
cancer cell proliferation with distinctly sub-mm activity,
which makes it a significantly more potent cell growth inhib-
itor than either of the monocyclic analogues 2 or 3. These
findings clearly suggest a defining character of the bicyclic
core structure of peloruside A (1) for potent cellular activi-
ty. At the same time it is clear, however, that the bicyclic
system by itself is not sufficient for peloruside A-like anti-
proliferative activity, as 4 (with a “naked” THP ring) is still
Chem. Eur. J. 2013, 19, 13105 – 13111
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13109

K.-H. Altmann et al.
monocyclic analogue 3, or macrolactonization, in the case of
bicyclic analogue 4. RCM-mediated formation of the 14-
membered ring in 3 was unproblematic; in contrast, macro-
lactone formation was surprisingly inefficient for the bicyclic
system in analogue 4 and was accompanied by extensive
cyclic dimer formation. Bicyclic analogue 4 inhibits cancer
cell proliferation with distinctly higher potency than 3 (ca.
one order of magnitude), which points to the bicyclic nature
of the peloruside macrolactone core structure as a crucial
determinant for sub-mm growth inhibitory activity. At the
same time, trideoxy analogue 4 is still less potent than the
parent natural product. It remains to be investigated if any
single one of the oxygen substituents attached to the THP
ring in peloruside A (1) may be omitted individually without
an erosion of potency.
Experimental Section
For further details see Supporting Information.
Acknowledgements
This work was supported by the ETH Zꢀrich (Grant TH-25 06-3). The
project was also conducted within the framework of COST Action
CM0804. We are indebted to Bernhard Pfeiffer, Fabienne Gaugaz, and
Leo Betschart for NMR support, and to Kurt Hauenstein (all ETHZ) for
practical advice, and to the ETHZ-LOC MS-Service for HRMS spectra.
Scheme 8. a) 2,4,6-Trichlorobenzoyl chloride, NEt3, THF, 08C, 3 h; then
dilution with toluene and addition of the mixed anhydride solution to a
solution of DMAP in toluene (via syringe pump, over 28 h), RT, 15 h,
608C, 18 h, 50%, 2:1 mixture of 36/37; b) HF·py, THF, RT, 36 h, 38:
80%, 39: 85% (separable); c) Pd/C, H2, EtOH, RT, 3 h 15 min, 77%; d)
Pd/C, H2, EtOH, RT, 5 h, 50%.
[1] L. M. West, P. T. Northcote, C. N. Battershill, J. Org. Chem. 2000,
65, 445.
[2] K. A. Hood, L. M. West, B. Rouwꢂ, P. T. Northcote, M. V. Berridge,
S. J. Wakefield, J. H. Miller, Cancer Res. 2002, 62, 3356.
ca. 10-fold less potent than natural peloruside A (1) (based
on the comparison with literature values). The relative im-
portance of the individual oxygen substituents on the THP
ring remains to be established; given the activity of 4, how-
ever, it is well conceivable that not all of these groups are
required for full activity.
[3] T. N. Gaitanos, R. M. Buey, J. F. Dꢃaz, P. T. Northcote, P. Teesdale-
Spittle, J. M. Andreu, J. H. Miller, Cancer Res. 2004, 64, 5063.
[4] K.-H. Altmann, J. Gertsch, Nat. Prod. Rep. 2007, 24, 327.
[5] Peloruside A (1) binds to the same site as laulimalide, but the loca-
tion of this site on tubulin is still unknown. For an investigation of
this question see: a) J. T. Huzil, J. K. Chik, G. W. Slysz, H. Freed-
man, J. Tuszynski, R. E. Taylor, D. L. Sackett, D. C. Schriemer, J.
Mol. Biol. 2008, 378, 1016; b) A. Begaye, S. Trostel, Z. Zhao, R. E.
Taylor, D. C. Schriemer, D. L. Sackett, Cell Cycle 2011, 10, 3387.
[6] E. Hamel, B. W. Day, J. H. Miller, M. K. Jung, P. T. Northcote, A. K.
Ghosh, D. P. Curran, M. Cushman, K. C. Nicolaou, I. Paterson, E. J.
Sorensen, Mol. Pharmacol. 2006, 70, 1555.
Table 1. Antiproliferative activity of 7S desTHP-peloruside A (2), 7R
desTHP-peloruside A (3), and 7,8,9-trideoxy-peloruside A (4) in three
human cancer cell lines (IC50 values [nm]).[a]
Compound
A549 (lung)
MCF-7[b] (breast)
HCT116 (colon)
[7] a) A. Wilmes, K. Bargh, C. Kelly, P. T. Northcote, J. H. Miller, Mol.
Pharm. 2007, 4, 269; b) A. Wilmes, D. OꢄSullivan, A. Chan, C.
Chandrahasen, I. Paterson, P. T. Northcote, A. C. La Flamme, J. H.
Miller, Cancer Chemother. Pharmacol. 2011, 68, 117.
2[c]
3
4
16400
1390ꢁ180
124ꢁ12
>20000
2050ꢁ270
247ꢁ8
>20000
1170ꢁ100
163ꢁ15
[8] a) X. Liao, Y. Wu, J. K. De Brabander, Angew. Chem. 2003, 115,
1686; Angew. Chem. Int. Ed. 2003, 42, 1648 (ent-1); b) M. Jin, R. E.
Taylor, Org. Lett. 2005, 7, 1303; c) A. K. Ghosh, X. Xu, J. H. Kim,
C. X. Xu, Org. Lett. 2008, 10, 1001; d) A. B. Smith III, J. M. Cox, N.
Furuichi, C. S. Kenesky, J. Zheng, O. Atasoylu, W. M. Wuest, Org.
Lett. 2008, 10, 5501 (2-epi-1); e) D. A. Evans, D. S. Welch, A. W.
Speed, G. A. Moniz, A. Reichelt, S. Ho, J. Am. Chem. Soc. 2009,
131, 3840; f) M. A. McGowan, C. P. Stevenson, M. A. Schiffler, E. N.
Jacobsen, Angew. Chem. 2010, 122, 6283; Angew. Chem. Int. Ed.
2010, 49, 6147; g) T. R. Hoye, J. Jeon, L. C. Kopel, T. D. Ryba, M. A.
Tennakoon, Y. Wang, Angew. Chem. 2010, 122, 6287; Angew. Chem.
Int. Ed. 2010, 49, 6151.
[a] Cells were exposed to compounds for 72 h. [b] An IC50 value of 4.9ꢁ
1.2 nm against the MCF-7 cell line has been reported for peloruside A
(1).[36] [c] Data from ref. [11].
Conclusion
The stereoselective synthesis of two new analogues of the
potent mitosis inhibitor peloruside A (1) has been accom-
plished based on macrocyclization by RCM, in the case of
13110
www.chemeurj.org
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2013, 19, 13105 – 13111

Peloruside A Derivatives
FULL PAPER
[9] For a synthesis of the related peloruside B see: A. J. Singh, C. X.
Xu, X. Xu, L. M. West, A. Wilmes, A. Chan, E. Hamel, J. H. Miller,
P. T. Northcote, A. K. Ghosh, J. Org. Chem. 2010, 75, 2.
[10] a) B. Pera, M. Razzak, C. Trigili, O. Pineda, A. Canales, R. M. Buey,
J. Jimenez-Barbero, P. T. Northcote, I. Paterson, I. Barasoain, J. F.
Dꢃaz, ChemBioChem 2010, 11, 1669; b) A. J. Singh, M. Razzak, P.
Teesdale-Spittle, T. N. Gaitanos, A. Wilmes, I. Paterson, J. M. Good-
man, J. H. Miller, P. T. Northcote, Org. Biomol. Chem. 2011, 9, 4456;
c) Z. Zhao, R. E. Taylor, Org. Lett. 2012, 14, 669; d) N. Zimmer-
mann, P. Pinard, B. Carboni, P. Gosselin, C. Gaulon-Nourry, G. Du-
jardin, S. Collet, J. Lebreton, M. Mathꢂ-Allainmat, Eur. J. Org.
Chem. 2013, 2303.
[21] K. Biswas, H. Lin, J. T. Njardarson, M. D. Chappell, T.-C. Chou, Y.
Guan, W. P. Tong, L. He, S. B. Horwitz, S. J. Danishefsky, J. Am.
Chem. Soc. 2002, 124, 9825.
[22] For a review on macrolactonizations see: A. Parenty, X. Moreau, J.-
M. Campagne, Chem. Rev. 2006, 106, 911.
[23] L. V. Heumann, G. E. Keck, Org. Lett. 2007, 9, 1951.
[24] D. A. Evans, A. H. Hoveyda, J. Am. Chem. Soc. 1990, 112, 6447.
[25] a) S. D. Rychnovsky, Y. Hu, B. Ellsworth, Tetrahedron Lett. 1998, 39,
7271; b) D. J. Kopecky, S. D. Rychnovsky, J. Org. Chem. 2000, 65,
191.
[26] R. Jasti, J. Vitale, S. D. Rychnovsky, J. Am. Chem. Soc. 2004, 126,
9904.
[11] C. W. Wullschleger, J. Gertsch, K.-H. Altmann, Org. Lett. 2010, 12,
1120.
[12] B. Pfeiffer, unpublished results.
[13] A. Hafner, R. O. Duthaler, R. Marti, G. Rihs, P. Rothe-Streit, F.
Schwarzenbach, J. Am. Chem. Soc. 1992, 114, 2321.
[14] The alcohol obtained after PMB-cleavage was contaminated with
anisaldehyde, which could be readily removed at the stage of alde-
hyde 10.
[27] J. S. Yadav, B. V. Subba Reddy, G. G. K. S. Narayana Kumar, G.
Madhusudhan Reddy, Chem. Lett. 2007, 36, 426.
[28] While the subsequent hydrodehalogenation was conducted with the
mixture of 30/31, the isomers were separated for analytical purposes.
[29] S. Marumoto, J. J. Jaber, J. P. Vitale, S. D. Rychnovsky, Org. Lett.
2002, 4, 3919.
[30] a) K.-P. Chan, Y. Hui Ling, J. Li-Ting Chan, T.-P. Loh, J. Org. Chem.
2007, 72, 2127; b) H.-W. Hagedorn, R. Brossmer, Helv. Chim. Acta
1986, 69, 2127.
[15] a) S. D. Rychnovsky, D. J. Skalitzky, Tetrahedron Lett. 1990, 31, 945;
b) C. F. Tormena, L. C. Dias, R. Rittner, J. Phys. Chem. A 2005, 109,
6077.
[31] S. Hatakeyama, H. Irie, T. Shintani, Y. Noguchi, H. Yamada, M.
Nishizawa, Tetrahedron 1994, 50, 13369.
[16] In ref. [11] we reported the addition of the vinyllithium species de-
rived from 9 to an aldehyde related to 10 (or 22) to proceed with a
d.r. of ca. 2:1. However, we have now established that this previous
addition reaction was also completely selective and that what we
had assumed to be the undesired diastereoisomer was in fact ketone
15. Ketone 15 has been described: A. P. Duncan, J. L. Leighton,
Org. Lett. 2004, 6, 4117.
[32] D. A. Evans, J. Bartroli, T. L. Shih, J. Am. Chem. Soc. 1981, 103,
2127.
[33] D. A. Evans, A. M. Ratz, B. E. Huff, G. S. Sheppard, Tetrahedron
Lett. 1994, 35, 7171.
[34] J. D. White, R. G. Carter, K. F. Sundermann, M. Wartmann, J. Am.
Chem. Soc. 2001, 123, 5407.
[35] Dimer or oligomer formation has not been reported for any of the
seco-acid intermediates employed in the various total syntheses of 1
(cf. refs. [8,9]).
[17] E. J. Corey, C. J. Helal, Angew. Chem. 1998, 110, 2092; Angew.
Chem. Int. Ed. 1998, 37, 1986.
[18] a) J. A. Dale, H. S. Mosher, J. Am. Chem. Soc. 1973, 95, 512;
b) T. R. Hoye, C. S. Jeffrey, F. Shao, Nat. Protoc. 2007, 2, 2451.
[19] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, Bull.
Chem. Soc. Jpn. 1979, 52, 1989.
[36] A. Chan, P. M. Andreae, P. T. Northcote, J. H. Miller, Invest. New
Drugs 2011, 29, 615.
[20] a) M. Scholl, T. M. Trnka, J. P. Morgan, R. H. Grubbs, Tetrahedron
Lett. 1999, 40, 2247; b) C. N. Kuzniewski, J. Gertsch, M. Wartmann,
K.-H. Altmann, Org. Lett. 2008, 10, 1183.
Received: May 10, 2013
Published online: August 12, 2013
Chem. Eur. J. 2013, 19, 13105 – 13111
ꢁ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13111