Total synthesis of the salicylate enamide macrolide oximidine III: Application of relay ring-closing metathesis

Article abstract of DOI:10.1002/anie.200460042

The vacuolar-type (H⁺)-adenosine triphosphatase inhibitor oximidine III and its enamide and epoxide stereoisomers have been synthesized enantioselectively. A relay ring-closing metathesis (RCM) strategy was employed to facilitate the macrocycli

Full text of DOI:10.1002/anie.200460042

Angewandte
Chemie
Natural Product Synthesis
Total Synthesis of the Salicylate Enamide
Macrolide Oximidine III: Application of Relay
Ring-Closing Metathesis**
Xiang Wang, Emma Jean Bowman, Barry J. Bowman,
and John A. Porco, Jr.*
Vacuolar-type (H+)-adenosine triphosphatases (V-ATPases)
are important membrane-bound proteins that control the
pH level of intracellular compartments in eukaryotic cells.
These proteins affect many membrane processes, such as
intracellular membrane transport, bone resorption, and
tumor metastasis.[1] A number of structurally related salicy-
late enamide macrolides have recently been disclosed as V-
ATPase inhibitors. Unlike other known V-ATPase inhibitors
(for example, bafilomycins and concanamycins), salicylate
enamide macrolides are selective inhibitors of mammalian V-
ATPases.[2] The oximidines[3,4] (Scheme 1) are a growing
subclass of the salicylate enamides and contain both macro-
cyclic triene and diene epoxide moieties. Oximidine III (3)
from Pseudomonas sp. QN05727 was recently identified by
Hayakawa et al.[4] Compound 3 is closely related to oximi-
dines I (1) and II (2)[5] but contains an E enamide side chain
and lacks a secondary hydroxy group at C14. Interestingly, 3
exists in two conformers at ambient temperature[4] and its
activity against transformed 3Y1cells is three to eightfold
higher than that of 1. To further evaluate the structure–
activity relationships (SARs) of the oximidines, we initiated a
program to synthesize 3 and the corresponding epoxide and
enamide stereoisomers.
Our retrosynthetic analysis for oximidine III is shown in
Scheme 1. We planned to utilize late-stage copper(i)-medi-
ated cross-coupling[6] of E vinyl iodide (E)-4 with amide 5[6a]
to attach the enamide side chain. Macrocyclization would
then rely on ring-closing metathesis (RCM)[7] of precursors 6
and 7, which may be prepared by base-mediated transester-
ification[5a,8] of 4H-1,3-benzodioxin-4-one 8 and alcohol 9.
Our experience with oximidine II[5a] suggested that the ability
[*] X. Wang, Prof. J. A. Porco, Jr.
Department of Chemistry and
Center for Chemical Methodology and Library Development
Boston University, 590 Commonwealth Avenue
Boston, MA 02215 (USA)
Fax: (+1)617-353-6466
E-mail: porco@chem.bu.edu
Dr. E. J. Bowman, Prof. B. J. Bowman
Department of Molecular, Cell and Developmental Biology
Sinsheimer Labs, University of California
Santa Cruz, CA 95064 (USA)
[**] We thank Prof. Y. Hayakawa (Tokyo University of Science) for kindly
providing an authentic sample of oximidine III, Dr. J. Lee (Boston
University) for assistance with NMR spectroscopy, and Mr. R. Shen
(Boston University) for helpful suggestions. We thank the National
Institutes of Health (Grant no. GM-62842) for research support.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2004, 43, 3601 –3605
DOI: 10.1002/anie.200460042
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3601

Communications
the generation of intermediate 11 involving the introduction
of a relay moiety.[10] We hoped that the ruthenium catalyst
would react with the least sterically hindered terminal olefin
of substrate 7 to provide intermediate 15. After release of
cyclopentene as a by-product, 15 could then be converted into
the desired intermediate 11 for RCM macrocyclization.
Synthesis of vinyl epoxide fragments 9a/9b was initiated
by adding chiral, nonracemic epoxide 16[11] to an alkynyl
lithium reagent derived from 17 to afford alkynol 18
(Scheme 3). Acetylation of 18, followed by desilylation and
Lindlar semihydrogenation, afforded allylic alcohol 19 (85%,
4 steps). Sharpless asymmetric epoxidation[12] oxidized 19 to
epoxide 20 (78%, d.r. > 16:1).[13] The primary alcohol was
then oxidized by treatment with Dess–Martin periodinane[14]
to form an intermediate aldehyde, which was treated with
either methylene or 5-hexenylidenetriphenylphosphorane to
afford vinyl epoxides 9a/9b after deacetylation.
Preparation of the salicylate fragment began with Stille
coupling of triflate 21[6c] and vinyl stannane 22 (E:Z = 6:1),[15]
followed by desilylation with TBAF to form E allylic alcohol
23 (82%, 2 steps). In this transformation the Z alkene isomer
was converted into a lactone by-product that was easily
separated from the desired product by flash chromatography.
Oxidation of 23 with MnO2 and Wittig olefination gave
E,Z dienes 8a,b. Treatment of 9b with NaHMDS followed by
addition of 8a–c[5a] and in situ silylation afforded tetraenes
7a–c, respectively. Substrates 6a (R2 = cis-Me) and 6b (R2 =
trans-Me) were prepared from 8a, 8c, and 9a by employing
analogous procedures.
Scheme 1. Chemical structures of the oximidines and retrosynthetic
analysis of oximidine III (3). TBS=tert-butyldimethylsilyl, PMB=para-
methoxybenzyl, R2 =Me or iPr (see Table 1).
We next evaluated a number of substrates in RCM
macrocyclizations (Table 1). Treatment of 6b with the Grubbs
second-generation catalyst 10 (CH2Cl2, reflux) afforded a low
yield of the desired product 12, accompanied by significant
amounts of oligomers. Substrate 6a reacted more slowly with
catalyst 10 and also afforded a low yield of 12 (entries 1and 2,
Table 1). This result may be explained by the formation of the
unreactive alkylidene 14 (Scheme 2). We next examined the
relay RCM reactivity of substrates 7a–c. We found that it was
necessary to add the substrate to a solution of ruthenium
catalyst in CH2Cl2 to prevent formation of by-product 6
to perform RCM cleanly may be related to successful
initiation of the metathesis reaction. We hoped that the
ruthenium catalyst 10 (Scheme 2) would react with the epoxy
alkene function of substrate 6 to form intermediate 11, and
that 11 would undergo macrocyclization to afford the desired
12-membered macrolactone 12. Accordingly, we planned to
substitute the terminus of the conjugated diene to avoid
formation of stable ruthenium complex 14. In the event that
initiation of the metathesis at the vinyl epoxide site proved
difficult,[9] we planned to evaluate an alternative approach to
Scheme 2. Analysis of ring-closing metathesis pathways. Mes=trimethylphenyl, L=ligand.
3602
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 3601 –3605

Angewandte
Chemie
Scheme 3. Synthesis of RCM precursors. a) 17, nBuLi, then 16, BF3·OEt2, THF, À78–258C; b) 1. Ac2O, pyridine, cat. DMAP, CH2Cl2; 2. AcOH/
H2O/THF (3:1:1), 608C; 3. Lindlar catalyst, 1 atm H2, Et3N, THF, 85% for 4 steps; c) Ti(OiPr)4, d-DIPT, 4-Š molecular sieves, CH2Cl2, À208C,
+
À
+
À
=
76%; d) 1. Dess–Martin periodinane, CH2Cl2, RT; 2. Ph3P Me·Br , NaHMDS, THF, À108C, 62%; or Ph3P (CH2)4CH CH2·I , NaHMDS, THF,
À108C, 70%, Z:E>16:1; 3. K2CO3, MeOH, RT, 95%; e) 1. [Pd2(dba)3], tri(2-furyl)phosphane, LiCl, DMF, 608C; 2. TBAF, THF, RT, 82% for 2 steps;
f) 1. MnO2, CH2Cl2, RT; 2. (Ph3PEt)·Br or (Ph3PiBu)·Br, NaHMDS, À78–258C, 60–62% for 2 steps (for 8c, see ref. [5a]); g) 9, NaHMDS, 8; then
TBSCl, imidazole, THF, 0–258C, 92–96%. THF=tetrahydrofuran, DMAP=4-dimethylaminopyridine, DIPT=diisopropyl tartrate,
HMDS=1,1,1,3,3,3-hexamethyldisilazane, dba=trans,trans-dibenzylideneacetone, DMF=dimethylformamide, TBAF=tetrabutylammonium fluo-
ride.
Table 1: Evaluation of substrates and conditions for RCM macrocyclization.[a]
reactive catalyst 24 during the reac-
tion. A slower reaction rate was
observed, but the chemical yield
was not improved.
To evaluate the influence of the
vinyl epoxide olefin geometry on
the relay RCM process we prepared
epoxy alkene 26 (E:Z = 1.3:1,
Scheme 4) from 20 and phenylte-
trazole sulfone 25.[18] Compound 26
was converted into RCM precursor
27 in several steps (see Scheme 3).
Relay RCM (2.0 mm, DCE, 508C)
of substrate 27 provided only a 34%
yield of the desired product 12. The
E vinyl epoxide stereoisomer of 27
was purified and used in the further
study of this transformation. Sub-
mission of this E isomer of 27 to
relay RCM conditions gave neither
the desired macrolactone 12 nor by-
product 6a, only oligomeric by-
SM[b] R1
R2
Addition time [min] Catalyst Conversion [%][e] Yield [%][f]
1[c] 6a
2[c] 6b
H
H
cis-Me
trans-Me
–
–
10 or 24
10 or 24
10
38
75
15
12
45
30
30
58
58
71
=
=
3
4
5
6
7
7a
7b
7c
7a
7a
(CH2)3CH CH2 cis-Me
(CH2)3CH CH2 cis-iPr
(CH2)3CH CH2 trans-Me 240
(CH2)3CH CH2 cis-Me
(CH2)3CH CH2 cis-Me
(CH2)3CH CH2 cis-Me
120
150
>90
71
10
=
=
=
10
77
60
60
30
10
24
24
>90
>90
>90
8[d] 7a
=
[a] Reactions were conducted in refluxing CH2Cl2 unless otherwise noted; the final concentration of 6/7
was 2.0 mm. [b ] SM=starting material. [c] Catalyst added to the starting material in refluxing CH2Cl2
and stirred for 2 h. [d] Reaction conducted at 508C in 1,2-dichloroethane. [e] Conversions based on
recovered starting material. [f] Yields of isolated products; in all cases, only 12-membered cyclic
E,Z diene 12 was isolated.
(Scheme 2).[16] Substrate 7a was optimal for the RCM-RCM
process; significant amounts of the corresponding by-prod-
ucts 6 were obtained when substrates 7b and 7c were used,
which indicates slow reaction rates for the second RCM
(entries 3–5, Table 1). Faster addition of the substrate resulted
in reduced decomposition of the product (entry 6, Table 1).
The recyclable ruthenium catalyst 24 described by Hoveyda
and co-workers[17] showed a similar reactivity to catalyst 10
(entry 7, Table 1) and produced a colorless product. We
investigated the relationship between the addition rate and
temperature further in an attempt to minimize oligomer
formation.[7c] We found that production of oligomers was
minimized at a higher temperature than that initially used
Scheme 4. Relay RCM with substrate 27. a) 1. Dess–Martin periodi-
nane, CH2Cl2, RT; 2. 25, NaHMDS, À78–258C, 74% (E:Z=1.3:1).
b) 10 mol% 24, DCE, 2.0 mm, 508C, 30 min, 34%. DCE=dichloro-
ethane.
(entry 8, Table 1). Catalyst 24 was pretreated with 20 mol%
o-isopropoxystyrene to convert the highly reactive ruthe-
nium–alkylidene species 13 (R2 = Me, Scheme 2) into the less
Angew. Chem. Int. Ed. 2004, 43, 3601 –3605
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3603

Communications
products were formed, which indicates that the Z configura-
tion of the epoxy alkene is required for a fast reaction rate in
the first RCM step.
To advance 12 to oximidine III, the PMB protecting group
was removed by treatment with DDQ and the resulting
alcohol 28 was oxidized by PDC to produce aldehyde 29
(Scheme 5). Transformation of 29 into E vinyl iodide (E)-4
3 and a higher yield of Z enamide 31 (3 h, 21% yield, E:Z =
1:1).[6c] The synthetic product 3 was confirmed as identical to
natural oximidine III by 1H and 13C NMR spectroscopy, mass
spectrometry, [a]2D0 measurement, HPLC, and TLC Rf values
in three solvent systems.
After successful synthesis of oximidine III (3) and its
enamide stereoisomer 31, we also synthesized 32 (the C12-
C13 epimer of oximidine III) and its enamide stereoisomer 33
by employing l-DIPT in the epoxidation step. Unlike
oximidine III (3) and 31, both 32 and 33 each exist as a
1
single conformer, as shown by H NMR spectroscopy (RT).
Oximidines 3 and 31–33 were evaluated for activity against
bovine V-ATPase and their IC50 values (concentration
required for 50% inhibition) were found to be 2.2, 65, 4.3,
and 65 nm, respectively.[24] These initial SAR data indicate
that the C17-C18 olefin geometry has a more substantial
effect on V-ATPase inhibition than the C12-C13 epoxide
stereochemistry.
In summary, enantioselective total syntheses of the
natural product oximidine III (3) and stereoisomers 31–33
have been developed, which has allowed unambiguous
assignment of the relative and absolute stereochemistry of
3. A relay RCM strategy was employed to facilitate the crucial
macrocyclization reaction, and a well-defined substrate
possessing two differentially functionalized RCM alkene
partners was found to be required for the RCM-RCM
process. A novel phenyltetrazole sulfone reagent 30 was
developed for homologation of aldehydes to form vinyl
iodides under mild conditions. Further synthetic studies on
the oximidines and simplified analogues of these compounds
will be reported in due course.
Scheme 5. Syntheses of oximidine III (3) and its enamide stereoisomer
31. a) DDQ, CH2Cl2, pH 7 buffer, 0–258C, 95%; b) PDC, 4-Š molecular
sieves, CH2Cl2; c) 30, NaHMDS, THF, À78–258C, 62% for 2 steps
(E:Zꢀ1:1); d) TBAF (1.0 equiv), THF, 0–258C; concentrate, then 5,
CuTC, N,N’-dimethylethylenediamine, K2CO3, DMA, 508C, 1 h, 45%
(E:Z=7:1). DDQ=2,3-dichloro-5,6-dicyano-1,4-benzoquinone,
PDC=pyridinium dichromate, CuTC=copper(i) thiophenecarboxylate,
DMA=N,N-dimethylacetamide.
Received: March 18, 2004 [Z460042]
Keywords: antitumor agents · enamides · natural products ·
.
ring-closing metathesis · vinyl iodides
was problematic because the vinyl epoxide is sensitive to both
acidic and reductive[19] conditions. Treatment of 29 with the
Seyferth/Gilbert reagent[20] converted the aldehyde into a
terminal alkyne. Hydrozirconation of the alkyne, followed by
addition of iodine (THF, 08C) afforded vinyl iodide (E)-4 in
low yield. Direct homologation of aldehyde 29 was achieved
by KociØnski–Julia olefination with the novel phenyltetrazole
sulfone reagent 30.[21] This transformation led to an insepa-
rable mixture of vinyl iodides (E)-4 and (Z)-4 (62% yield,
E:Z = 1:1).[22]
[1] T. Nishi, M. Forgac, Nat. Rev. Mol. Cell Biol. 2002, 3, 94.
[2] a) M. R. Boyd, C. Farina, P. Belfiore, S. Gagliardi, J. W. Kim, Y.
Hayakawa, J. A. Beutler, T. C. McKee, B. J. Bowman, E. J.
Bowman, J. Pharmacol. Exp. Ther. 2001, 297, 114; b) for a recent
review of syntheses and SAR studies, see: L. Yet, Chem. Rev.
2003, 103, 4283.
[3] J. W. Kim, K. Shin-ya, K. Furihata, Y. Hayakawa, H. Seto, J. Org.
Chem. 1999, 64, 153.
[4] a) Y. Hayakawa, T. Tomikawa, K. Shin-Ya, N. Arao, K. Nagai,
K-I. Suzuki, J. Antibiot. 2003, 56, 899; b) Y. Hayakawa, T.
Tomikawa, K. Shin-Ya, N. Arao, K. Nagai, K-I. Suzuki, K.
Furihata, J. Antibiot. 2003, 56, 905.
[5] For the total synthesis of oximidine II, see: a) X. Wang, J. A.
Porco, Jr, J. Am. Chem. Soc. 2003, 125, 6040; for recent studies
on the synthesis of the oximidines, see: b) T. Haack, S. Kurtkaya,
J. P. Snyder, G. I. Georg, Org. Lett. 2003, 5, 5019; c) J. E. Harvey,
S. A. Raw, R. J. K. Taylor, Tetrahedron Lett. 2003, 44, 7209.
[6] a) R. Shen, J. A. Porco, Jr, Org. Lett. 2000, 2, 1333; b) R. Shen,
C. T. Lin, E. J. Bowman, B. J. Bowman, J. A. Porco, Jr, J. Am.
Chem. Soc. 2003, 125, 7889; for applications of this method, see:
ref. [5a] and c) A. Fürstner, T. Dierkes, O. R. Thiel, G. Blanda,
Chem. Eur. J. 2001, 7, 5286; d) R. Shen, C. T. Lin, J. A. Porco, Jr,
J. Am. Chem. Soc. 2002, 124, 5650; e) K. C. Nicolaou, D. W. Kim,
R. Baati, A. OꢀBrate, P. Giannakakou, Chem. Eur. J. 2003, 9,
We next tested the crucial late-stage CuI-mediated vinylic
amidation reaction. The TBS protecting group of 4 was first
removed by treatment with TBAF to avoid the use of excess
base.[23] The deprotected compound, presumed to be the
phenolate salt, was submitted directly to amidation by
treatment with copper(i) thiophenecarboxylate-N,N’-dimeth-
ylethylenediamine.[5a] The vinyl iodides were consumed
quickly when a stoichiometric amount of CuTC was used.
Oximidine III (3) and the corresponding Z enamide stereo-
isomer 31 were isolated in 45% yield (E:Z = 7:1) after 1 h
(508C). Extended reaction times resulted in decomposition of
3604
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2004, 43, 3601 –3605

Angewandte
Chemie
6177; f) R. S. Coleman, P.-H. Liu, Org. Lett. 2004, 6, 577; g) Q.
Su, J. S. Panek, J. Am. Chem. Soc, 2004, 126, 2425.
[7] Recent reviews: a) A. Fürstner, Angew. Chem. 2000, 112, 3140;
Angew. Chem. Int. Ed. 2000, 39, 3012; b) T. M. Trnka, R. H.
Grubbs, Acc. Chem. Res. 2001, 34, 18; c) for RCM with a vinyl
epoxide, see: K. Yamamoto, K. Biswas, C. Gaul, S. J. Danishef-
sky, Tetrahedron Lett. 2003, 44, 3297.
[8] Y. Wu, X. Liao, R. Wang, X.-S. Xie, J. K. De Brabander, J. Am.
Chem. Soc. 2002, 124, 3245.
[9] For an example of allylic substitution effects on reaction rates of
RCM, see: T. R. Hoye, H. Zhao, Org. Lett. 1999, 1, 1123.
[10] For studies regarding relay RCM, see: a) T. R. Hoye, J. Wang,
226th ACS National Meeting (New York, NY), 2003, ORGN-
670; b) T. R. Hoye, R. Thomas, H. Zhao, 218th ACS National
Meeting, (New Orleans, LA), 1999, ORGN-620; c) T. R. Hoye,
H. Zhao, 218th ACS National Meeting (New Orleans, LA), 1999,
ORGN-303; d) for another application of a “relay” entity in the
assembly of reacting sites in RCM, see: A. Fürstner, K.
Langemann, Synthesis 1997, 792.
[11] The chiral nonracemic epoxide 16 was prepared by hydrolytic
kinetic resolution of the racemic epoxide (48% yield, 95% ee):
S. E. Schaus, B. D. Brandes, J. F. Larrow, M. Tokunaga, K. B.
Hansen, A. E. Gould, M. E. Furrow, E. N. Jacobsen, J. Am.
Chem. Soc. 2002, 124, 1307.
[12] T. Katsuki, V. S. Martin, Org. React. (N.Y.) 1996, 48, 1 .
[13] Use of a sterically demanding TBS protecting group on the
secondary alcohol resulted in low diastereoselectivity of the
Sharpless asymmetric epoxidation (4:1).
[14] D. B. Dess, J. C. Martin, J. Org. Chem. 1983, 48, 4155.
[15] J. W. Labadie, D. Tueting, J. K. Stille, J. Org. Chem. 1983, 48,
4634.
[16] Use of the Grubbs first-generation catalyst led to the corre-
sponding by-products 6 (Scheme 2).
[17] S. B. Garber, J. S. Kingsbury, B. L. Gray, A. H. Hoveyda, J. Am.
Chem. Soc. 2000, 122, 8168.
[18] P. R. Blakemore, W. J. Cole, P. J. KociØnski, A. Morley, Synthesis
1996, 285. For the preparation of 25, see the Supporting
Information.
[19] K. Takai, K. Nitta, K. Utimoto, J. Am. Chem. Soc. 1986, 108,
7408.
[20] a) D. Seyferth, R. S. Marmor, P. Hilbert, J. Org. Chem. 1971, 36,
1379; b) R. Ratcliffe, B. Christensen, Tetrahedron Lett. 1973, 14,
4645; c) J. C. Gilbert, U. Weerasooriya, J. Org. Chem. 1982, 47,
1837; d) D. G. Brown, E. J. Velthuisen, J. R. Commerford, R. G.
Brisbois, T. R. Hoye, J. Org. Chem. 1996, 61, 2540.
[21] For the preparation of 30, see the Supporting Information; a) for
a related chloromethyl benzothiazole sulfone reagent, see: J. B.
Baudin, G. Hareau, S. A. Julia, R. Lorne, O. Ruel, Bull. Soc.
Chim. Fr. 1993, 130, 856; b) for use of iodoalkyl phenyl sulfone
reagents in free radical reactions, see: P. Jankowski, M. Masnyk,
J. Wicha, Synlett 1995, 866.
[22] The E:Z ratio was determined by 1H NMR analysis.
[23] In our synthesis of oximidine II (ref. [5a]), we utilized a silylated
phenol substrate treated with base and amide (2 equiv each) in a
copper-mediated coupling to effect both silyl ether deprotection
and enamide formation. However, use of excess base and amide
in the syntheses described herein led to significant decomposi-
tion of the sensitive vinyl epoxide moiety.
[24] V-ATPase activity was determined as described in the Support-
ing Information.
Angew. Chem. Int. Ed. 2004, 43, 3601 –3605
www.angewandte.org
ꢀ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3605