Syntheses and biological evaluation of iriomoteolide 3a and analogues

Article abstract of DOI:10.1002/anie.200903379

Metathesis on stage: A combined crossmetathesis (CM)/ring-closing metathesis (RCM) approach has led to the stereocontrolled synthesis of iriomoteolide 3a, a smaller but equally cytotoxic congener of amphidinolides. Chemical editing of the molecule has pro

Full text of DOI:10.1002/anie.200903379

Communications
DOI: 10.1002/anie.200903379
Total Synthesis
Syntheses and Biological Evaluation of Iriomoteolide 3a and
Analogues**
Riccardo Cribiffl, Corinna Jäger, and Cristina Nevado*
Amphidinium species are an extremely prolific source of
marine secondary metabolites.[1] Structurally unique poly-
ketides such as amphidinolides, caribenolide I, and amphidi-
nolactones have fostered the interest of chemists not only as
challenging targets for total synthesis but also because of their
potent anticancer activity.[2] Recently, the Amphidinium
strain HYA024 was found to produce cytotoxic compounds
such as iriomoteolides 1a–c,[3] and a rare 15-membered
macrolide, iriomoteolide 3a (1).[4] With a novel carbon frame-
work comprising eight stereogenic centers, four of them in
allylic positions, compound 1 represents the first member of a
unique and unprecedented 15-membered macrolide class.
Compound 1 represents the first member of a unique and
unprecedented 15-membered macrolide class. In addition, the
preliminary physiological properties disclosed for 1 and its
Scheme 1. Retrosynthetic analysis for iriomoteolide 3a (1).
7,8-O-isopropylidene derivative 2 are very promising, show-
ing potent cytotoxicity against lymphoma cell lines in the low
nanomolar range.[4]
that the formation of a medium-sized ring would also be E,E
stereoselective (Scheme 1).
To confirm the assigned structure, further evaluate its
biological activity, and determine whether its cellular targets
are related to those of larger congeners such as amphidino-
lides,[5] substantial quantities of these compounds are
required. Our retrosynthetic approach to 1 involved four
major disconnections, which revealed key fragments 3–6 as
summarized in Scheme 1. Fragment 6 was planned to be
incorporated at the end of our synthetic sequence by a Julia–
Kocienski olefination because of its widely recognized
performance in the elaboration of such sensitive settings
and also to ensure a flexible late-stage diversification of the
parent compound. An intermolecular esterification was
envisioned to assemble fragments 3 and 4. Finally, we
hypothesized that the C2-symmetry of the diol precursor of
fragment 5 could be advantageously used to construct the 1,5-
diene upon ring closure by a cross-metathesis (CM)/ring-
closing metathesis (RCM) approach. We were relying on the
excellent results achieved by the Grubbs-type carbene com-
plexes in both CM and RCM processes with the expectation
The required building block 3 (Scheme 2) was prepared
by alkylation of the Evans oxazolidinone 7[6] with iodide 8.[7]
Scheme 2. a) Na[N(SiMe3)2], THF, ꢀ788C, 85%; b) ADmix-a,
MeSO2NH2, tBuOH/H2O, 08C, 83% (94% de); c) PMBNHCCl3, CSA,
CH2Cl2, RT, 89%; d) TBAF, THF, 89%; e) TBSCl, imidazole, DMF, 08C,
94%; f) LiBH4, Et2O, 91%; g) TBDPSCl, imidazole, CH2Cl2, 08C, 90%;
h) TBSOTf, 2,6-lutidine, CH2Cl2, 08C, 98%; i) PPTS, EtOH, RT, 82%;
j) DMSO, (COCl)2, Et3N, CH2Cl2, ꢀ788C!RT, then Ph3PCHCO2Me,
CH2Cl2, RT, 91%; k) DIBAL-H, CH2Cl2, ꢀ788C, 88%; l) tBuOOH,
Ti(OiPr)4, (+)-DIPT, CH2Cl2, MS (4 ꢀ), 94%, (92% de); m) DMSO,
(COCl)2, Et3N, CH2Cl2, ꢀ788C!RT, then [Ph3PCH3]Br, Na[N(SiMe3)2],
THF, 73%; n) DDQ, CH2Cl2, pH 7 buffer, RT, 85%. CSA=camphorsul-
fonic acid, DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone, DIBAL-
H=diisobutylaluminium hydride, DIPT=diisopropyl tartrate, DMF=
N,N-dimethylformamide, DMSO=dimethylsulfoxide, MS=molecular
sieves, PMB=para-methoxybenzyl, PPTS=pyridinium para-toluenesul-
fonate, TBAF=tetra-n-butylammonium fluoride, TBDPS=tert-butyldi-
phenylsilyl, TBS=tert-butyldimethylsilyl.
[*] Dr. R. Cribiffl, C. Jäger, Prof. Dr. C. Nevado
Organic Chemistry Institute, Universität Zürich
Winterthurerstrasse 190, 8057 Zürich (Switzerland)
Fax: (+41)446-356-888
E-mail: nevado@oci.uzh.ch
[**] R.C. and C.J. thank the Legerlotz Stiftung and OCI for financial
support. We are indebted to Prof. N. Luedtke, B. Vummidi, and
C. Hemmerle for their invaluable help with the cell-based assays.
Noam Prywes, Karine Lafleur, and Teresa de Haro are also
acknowledged for their contributions at the initial stage of this
project.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200903379.
8780
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8780 –8783

Angewandte
Chemie
Asymmetric dihydroxylation of the double bond and subse-
quent spontaneous lactonization to release the chiral auxiliary
afforded 9 with excellent selectivity (d.r. > 20:1). The free
hydroxy group in 9 was protected as p-methoxybenzyl ether,
and the TBDPS protecting group on the primary alcohol was
exchanged for a TBS group.[8] Reductive opening of the
lactone afforded diol 10. A series of protecting group
manipulations delivered alcohol 11 in excellent yield. Inter-
mediate 12 was obtained after a reaction sequence including
an oxidation, a Wittig reaction with (carbamethoxymethyl-
ene)triphenylphosphorane, and a DIBAL-H reduction of the
methyl ester. Sharpless asymmetric epoxidation[9] allowed the
installation of the required oxirane with high stereochemical
control (> 92% de). Straightforward oxidation-state and
protecting group manipulation led to the alkene moiety in
fragment 3 which is necessary for the envisioned metathesis
reaction.
Scheme 3. a) Me2AlCl, AcBr, iPr2EtN, CH2Cl2, ꢀ788C, 94%; b) CuBr,
Me2S, MeMgBr, THF, ꢀ508C, 68%; c) Boc2O, DMAP, tBuOH, RT,
82%; d) Pd/C (10 mol%), H2, MeOH, RT; e) DMSO, (COCl)2, Et3N,
CH2Cl2, ꢀ788C!RT, then [Ph3PCH3]Br, nBuLi, ꢀ788C!RT, 80% over
3 steps; f) trifluoroacetic acid, CH2Cl2, RT, 72%; g) BTSH, NaH, DMF,
RT, 80%; h) Na2WO4 (10 mol%), H2O2, RT, 75%. Ac=acetyl, Boc=
tert-butoxycarbonyl, BTSH=1-phenyl-1H-benzothiazol-5-thiol.
five equivalents of 5c afforded an inseparable mixture of
regioisomers 22a and 22b in 49% yield (Scheme 4).
The route to fragment 4 commenced with the catalytic
asymmetric cyclocondensation of 4-(benzyloxy)-1-butanal
(13)[10] with acetyl bromide using triamine ligand 14[11] by a
modification of the procedure previously reported by Nelson
et al. (Scheme 3).[12] Dimethyl cuprate, generated in situ, was
reacted with 15 to afford the corresponding acid, which was
readily transformed into tert-butyl ester 16. Hydrogenolysis of
the benzyl protecting group and subsequent oxidation of the
primary alcohol gave the corresponding aldehyde, which was
then subjected to a Wittig olefination to provide compound 17
in good yield.[13] Hydrolysis of the tert-
Surprisingly, when the mixture of 22a and 22b was
submitted to classical RCM conditions (21, 0.005m in toluene,
RTor 608C), the desired products were obtained in low yields,
in addition to the RCM product of 20 and dimeric 5c. In our
view, the ruthenium–carbene complex might react first with
the terminal double bonds present in 22a and 22b (shown in
red in Scheme 4), but the steric hindrance of the double bond
close to the silyl group prevents the desired RCM event.
Instead, the ruthenium–carbene complex reacted with the
internal double bond close to the free hydroxy group (shown
butyl ester under acidic conditions deliv-
ered fragment 4 in 40% overall yield after
six steps. Fragment 6 was prepared from
(E)-5-bromo-2-pentene (18)[14] in two steps
by halogen displacement to give sulfide 19
and then chemoselective oxidation, using
H2O2/Na2WO4, to give sulfone
6
(Scheme 3).[15] Analogues of fragment 5
(Scheme 1) were synthesized from l-tarta-
ric acid according to known procedures.[16]
Esterification of acid 4 with alcohol 3
was cleanly effected using EDC as activat-
ing agent in the presence of 4-pyrrolidino-
pyridine (Scheme 4). Completion of the
macrolide required the assembly of com-
pound 20 with fragment 5 by a CM/RCM
sequence using ruthenium complex 21.[17]
When 5a was used, a complex mixture of
products was obtained, which was prob-
ably a result of the high reactivity of the
diol. In contrast, under the same reaction
conditions, olefin 5b was recovered
unreacted, and 20 underwent both RCM
to give a 10-membered ring lactone and
also self-immolative CM.[18] We hypothe-
sized that the steric hindrance imposed by
the TBS group in 5c could prevent the
coordination of the ruthenium catalyst 21
to the substrate and thereby prevent the
CM reaction on its neighboring double
bond.[19] Remarkably, CM between 20 and
Scheme 4. a) EDC·HCl, 4-pyrrolidinopyridine, CH2Cl2, RT, 72%; b) 5c (5 equiv), 21
(5 mol%), toluene, 508C, 49%; c) TBSOTf, 2,6-lutidine, THF, 08C, 80%; d) 21 (12 mol%),
toluene, RT, 76%. EDC=3-(3-dimethylaminopropyl)-1-ethylcarbodiimide.
Angew. Chem. Int. Ed. 2009, 48, 8780 –8783
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8781

Communications
Table 1: Syntheses of iriomoteolide 3a analogues.
in blue) to produce a ring-contracted homologue of 20 and
released 5c, which then dimerized. To circumvent this
problem, we decided to silylate the mixture of 22a and 22b
with the hope that upon the first cycloaddition of the
ruthenium catalyst to the less hindered double bond (red),
the remaining available double bonds of the molecule would
have similar steric constraints.[20] To our delight, the mixture
of 23a and 23b underwent clean RCM under the above-
mentioned reaction conditions to give compound 24 as a
single isomer in 72% overall yield.
The access to compound 24 enabled us to test the critical
chemoselective removal of the primary OTBDPS group in the
presence of three secondary OTBS groups.[21] After extensive
experimentation, we found that an excess of ammonium
fluoride in methanol[22] resulted in a slow but clean conversion
of the starting material into the desired product 25. Oxidation
of the primary alcohol with Dess–Martin periodinane and
subsequent Julia–Kocienski olefination[23] with sulfone 6
afforded the immediate precursor of iriomoteolide 3a in
76% yield over two steps (Scheme 5). Notably, this last
transformation was highly E stereoselective (> 93:7) and the
Entry
Reaction
R1–R4
Product
(yield %)
conditions[18]
1
2
3
4
5
1
Ac2O, Pyridine
R1 =R2 =R3 =Ac,
26
R4 =trans-CH2CHCHCH3 (quant.)
25[a] Na[N(SiMe3)2],
C7H15SO2PT[b,c]
R1 =R2 =R3 =TBS,
R4 =C6H13
27 (72)
28 (71)
27 TBAF (4 equiv)[d]
R1 =R2 =R3 =H,
R4 =C6H13
25[a] Na[N(SiMe3)2],
[Ph3PCH3]Br[b]
R1 =R2 =R3 =TBS, R4 =H 29 (75)
29 TBAF (4 equiv)[d]
R1 =R2 =R3 =R4 =H
30 (88)
[a] DMP, CH2Cl2, RT (quant.). [b] Reaction performed in THF at ꢀ788C.
[c] PT=1-phenyl-1H-tetrazol-5-thiol. [d] Reaction performed in THF at
258C.
Table 2: Antiproliferative activity of 1 and analogues (2, 26, 28, 30) in the
alamarBlue fluorimetric assay.[a]
Cell line
1
2
26
28
30
DAUDI
HL-60
0.080
2.6
0.048
2.0
0.737
n.d.
0.083
2.8
n.d.
n.d.
[a] GI50 values in mm. n.d.=not determined; no activity was observed up
to a concentration of 10 mm.
Scheme 5. a) NH4F, MeOH, RT, 58%; b) DMP, CH2Cl2, RT; c) 6, K[N-
(SiMe3)2], THF, 08C, 93:7 E/Z, 76% (over 2 steps); d) TBAF, THF, RT,
86%; e) 2,2-dimethoxypropane, PPTS, CH2Cl2, 20%. DMP=Dess–
Martin periodinane.
737 nm). The introduction of a truncated side chain (30)
also compromised the antiproliferative activity even at 10 mm
concentration. Noticeably, a more lipophilic pendant chain
(28) afforded similar levels of potency as 1 and 2, thus
highlighting the importance of the lateral chain for the
cytotoxicity of these molecules. A similar pattern was
observed for HL-60 with activities in the low mm range. In
light of these results, we were able to deduce the following
trends: first, the enhanced activity of compound 28 compared
to that of 30 could be explained by simple increase in
lipophilicity, which might facilitate the cell penetration of the
molecule. This fact is partially confirmed by the higher
activity of acetonide 2 compared to that of parent compound
1. Second, as peracetylated 26 showed very low cytotoxicity
compared to 1 and 2, we conclude that other factors might
also influence the activity of these compounds and the
presence of the free OH group on C15 is important for the
interaction with their biological targets.
In summary, we report the first total synthesis of
iriomoteolide 3a (1), which confirmed the absolute config-
uration of this potent cytotoxic macrolide and provided
sufficient quantities for additional biological evaluation. The
key ring-closure to construct the 15-membered ring macro-
cycle relies on a highly E,E stereoselective cross-metathesis/
ring-closing metathesis sequence. Through a modular syn-
thetic approach, we have synthesized a small collection of
non-natural derivatives of 1, and tested their antiproliferative
activity, revealing the suitable sites for structural modifica-
tions in the original core. Additional chemical editing of this
mild conditions preserved the stereochemical integrity of the
intermediate a-branched aldehyde.[24] Final removal of the
three silyl groups was achieved with TBAF to afford
iriomoteolide 3a (1) (Scheme 5), whose analytical and spec-
troscopic properties were in good accordance with the
published data.[4] 7,8-O-isopropylidene derivative 2 was
obtained by treatment of 1 with 2,2-dimethoxypropane in
the presence of pyridinium para-toluenesulfonate.[4]
A systematic structural editing of the natural product
became our next immediate goal. First, the hydroxy groups in
1 were fully acetylated to afford compound 26 (Table 1,
entry 1). As originally planned, the side chain was used for
structural diversification. Starting from alcohol 25, and after
Dess-Martin periodinane oxidation, longer (27) and shorter
(29) side chains were assembled through Julia and Wittig
olefination reactions, respectively (Table 1, entries 2 and 4).
The macrolides 27 and 29 were deprotected using TBAF in
THF to afford triols 28 and 30, respectively (Table 1, entries 3
and 5).
The growth inhibitory activities of compounds 1, 2, 26, 28,
and 30 were investigated on two different human cancer cell
lines: DAUDI (lymphoma) and HL-60 (leukemia) using the
alamarBlue fluorometric assay (Table 2).[18,25] Synthetic 1 and
2 showed high potency against lymphoma cell lines (GI50 = 80
and 48 nm, respectively) confirming the preliminary results
reported in the isolation paper.[4] However, the activity of
peracetylated derivative 26 dramatically decreased (GI50 =
8782
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 8780 –8783

Angewandte
Chemie
[7] a) X.-T. Zhou, R. G. Carter, Chem. Commun. 2004, 2138 – 2140;
promising structure, and studies to elucidate both, its mode of
action and cellular targets, are currently underway.
b) K. Ohtsuki, K. Matsuo, T. Yoshikawa, C. Moriya, K. Tomita-
Yokotani, K. Shishido, M. Shindo, Org. Lett. 2008, 10, 1247 –
1250.
Received: June 22, 2009
Revised: August 18, 2009
Published online: October 7, 2009
[8] Alternative synthesis of compound 9 starting with a TBS-
protected alkyl iodide (8’) gave lower yields and poor diaste-
reomeric ratios. The TBDPS to TBS swap was necessary to allow
further functional group manipulations at a later stage of the
synthesis.
[9] Y. Gao, J. M. Klunder, R. M. Hanson, H. Masamune, S. Y. Ko,
K. B. Sharpless, J. Am. Chem. Soc. 1987, 109, 5765 – 5780.
[10] P. Perlmutter, W. Selajerern, F. Vounatsos, Org. Biomol. Chem.
2004, 2, 2220 – 2228.
Keywords: anticancer agents · macrocycles · metathesis ·
.
natural products · total synthesis
[1] a) J. Kobayashi, M. Tsuda, Nat. Prod. Rep. 2004, 21, 77 – 93; b) J.
Kobayashi, T. Kubota, J. Nat. Prod. 2007, 70, 451 – 460.
[2] Leading references for the total synthesis of amphidinolides:
Amphidinolide A: a) B. M. Trost, P. E. Harrington, J. D. Chis-
holm, S. T. Wrobleski, J. Am. Chem. Soc. 2005, 127, 13598 –
13610; b) R. E. Maleczka, Jr., L. R. Terrell, F. Geng, J. S.
Ward, Org. Lett. 2002, 4, 2841 – 2844; c) H. W. Lam, G.
Pattenden, Angew. Chem. 2002, 114, 526 – 529; Angew. Chem.
Int. Ed. 2002, 41, 508 – 511; Amphidinolide E: d) P. Va, W. R.
Roush, J. Am. Chem. Soc. 2006, 128, 15960 – 15961; Amphidi-
nolide H–G: e) A. Fꢀrstner, L. C. Bouchez, J.-A. Funel, V.
Liepins, F.-H. Porꢁe, R. Gilmour, F. Beaufils, D. Laurich, M.
Tamiya, Angew. Chem. 2007, 119, 9425 – 9430; Angew. Chem. Int.
Ed. 2007, 46, 9265 – 9270; Amphidinolide K: f) D. R. Williams,
K. G. Meyer, J. Am. Chem. Soc. 2001, 123, 765 – 766; Amphidi-
nolide T1: g) A. K. Ghosh, C. Liu, J. Am. Chem. Soc. 2003, 125,
2374 – 2375; Amphidinolide T4: h) A. Fꢀrstner, C. Aꢂssa, R.
Riveiros, J. Ragot, Angew. Chem. 2002, 114, 4958 – 4960; Angew.
Chem. Int. Ed. 2002, 41, 4763 – 4766; Amphidinolide V: i) A.
Fꢀrstner, O. Larionov, S. Flꢀgge, Angew. Chem. 2007, 119, 5641 –
5644; Angew. Chem. Int. Ed. 2007, 46, 5545 – 5548.
[3] a) M. Tsuda, K. Oguchi, R. Iwamoto, Y. Okamoto, J. Kobayashi,
E. Fukushi, J. Kawabata, T. Ozawa, A. Masuda, Y. Kitaya, K.
Omasa, J. Org. Chem. 2007, 72, 4469 – 4474; Progress towards
the total synthesis of iriomoteolide-1a: b) L. Fang, H. Xue, J.
Yang, Org. Lett. 2008, 10, 4645 – 4648; c) A. K. Ghosh, H. Yuan,
Tetrahedron Lett. 2009, 50, 1416 – 1418; d) J. Xie, D. Horne,
Tetrahedron Lett. 2009, 50, 4485 – 4487.
[4] K. Oguchi, M. Tsuda, R. Iwamoto, Y. Okamoto, J. Kobayashi, E.
Fukushi, J. Kawabata, T. Ozawa, A. Masuda, Y. Kitaya, K.
Omasa, J. Org. Chem. 2008, 73, 1567 – 1570.
[11] M. Cernerud, A. Skrinning, I. Bꢁrgꢃre, C. Moberg, Tetrahedron:
Asymmetry 1997, 8, 3437 – 3441.
[12] a) S. G. Nelson, T. J. Peelen, Z. Wan, J. Am. Chem. Soc. 1999,
121, 9742 – 9743; b) S. G. Nelson, Z. Wan, M. A. Stan, J. Org.
Chem. 2002, 67, 4680 – 4683.
[13] Benzyl ether removal in presence of the free acid, even under
neutral conditions, led to partial lactonization of the resulting
alcohol.
[14] Y. C. Hwang, F. W. Fowler, J. Org. Chem. 1985, 50, 2719 – 2726.
[15] A. B. Charette, C. Bethelette, D. St. -Martin, Tetrahedron Lett.
2001, 42, 5149 – 5153.
[16] For the formation of the diol 5a, see a) M. A. Robbins, P. N.
Devine, T. Oh, Org. Synth. 1999, 76, 101 – 109; b) C. C. Marvin,
A. J. L. Clemens, S. D. Burke, Org. Lett. 2007, 9, 5353 – 5356; for
the preparation of 5b and 5c, see: c) K. J. Quinn, A. K. Isaacs,
B. A. DeChristopher, S. C. Szklarz, R. A. Arvary, Org. Lett.
2005, 7, 1243 – 1245.
[17] M. Scholl, S. Ding, C. W. Lee, R. H. Grubbs, Org. Lett. 1999, 1,
953 – 956.
[18] For further details, see the Supporting Information.
[19] See reference [16c] and also: S. Michaelis, M. Blechert, Org.
Lett. 2005, 7, 5513 – 5516, and references therein.
[20] Compounds 22a and 22b could be separated at this stage to give
a 2:1 ratio, confirming the E configuration of the olefin produced
in the CM event.
[21] A single successful precedent for such transformation is reported
using TASF: R. Lira, W. R. Roush, Org. Lett. 2007, 9, 533 – 536.
[22] W. Zhang, M. J. Robins, Tetrahedron Lett. 1992, 33, 1177 – 1180.
[23] a) P. J. Kocienski, A. Bell, P. R. Blakemore, Synlett 2000, 365 –
366; b) P. R. Blakemore, J. Chem. Soc. Perkin Trans. 1 2002,
2563 – 2585.
[5] a) T. Usui, S. Kazami, N. Dohmae, Y. Mashimo, H. Kondo, M.
Tsuda, A. G. Terasaki, K. Ohashi, J. Kobayashi, H. Osada,
Chem. Biol. 2004, 11, 1269 – 1277.
[6] D. A. Evans, J. Bartroli, T. L. Shih, J. Am. Chem. Soc. 1981, 103,
2127 – 2129.
[24] C. Aꢂssa, Eur. J. Org. Chem. 2009, 1831 – 1844.
[25] M. M. Nociari, A. Shalev, P. Benias, C. Russo, J. Immunol.
Methods 1998, 213, 157 – 167.
Angew. Chem. Int. Ed. 2009, 48, 8780 –8783
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8783