A Concise Synthesis of the Octalactins

Article abstract of DOI:10.1021/ja038353w

The total synthesis of octalactins A and B has been achieved in 15 steps (longest linear sequence) and 10% overall yield from commercially available materials. Key steps include the Paterson-Aldol reaction for the rapid assembly of the carbonate 46, methylenation of 46 and subsequent Claisen rearrangement of the corresponding alkenyl-substituted cyclic ketene acetal to provide the core unsaturated medium-ring lactone 47, and the use of enzyme-mediated acetate deprotection in the presence of a medium-ring lactone.

Full text of DOI:10.1021/ja038353w

Published on Web 02/03/2004
A Concise Synthesis of the Octalactins
Paul T. O’Sullivan, Wilm Buhr, Mary Ann M. Fuhry, Justin R. Harrison,
John E. Davies, Neil Feeder, David R. Marshall,† Jonathan W. Burton,* and
Andrew B. Holmes*
Contribution from the Department of Chemistry, UniVersity of Cambridge, Lensfield Road,
Cambridge, CB2 1EW, U.K.
Received September 5, 2003; E-mail: abh1@cam.ac.uk; jwb1004@cam.ac.uk
Abstract: The total synthesis of octalactins A and B has been achieved in 15 steps (longest linear sequence)
and 10% overall yield from commercially available materials. Key steps include the Paterson-Aldol reaction
for the rapid assembly of the carbonate 46, methylenation of 46 and subsequent Claisen rearrangement
of the corresponding alkenyl-substituted cyclic ketene acetal to provide the core unsaturated medium-ring
lactone 47, and the use of enzyme-mediated acetate deprotection in the presence of a medium-ring lactone.
tion7,8 to form the central eight-membered lactone of 1 and 2.
Introduction
This synthesis demonstrated for the first time that it was feasible
In 1991, Fenical and Clardy reported the isolation of the
saturated eight-membered lactones octalactin A (1) and B (2)
from a marine-derived actinomycete of the genus Streptomyces
collected from a gorgonian octocoral.1,2 The structure and
relative stereochemistry of 1 and 2 were assigned by a
combination of spectroscopic analysis and X-ray crystal-
lography; the absolute stereochemistry of the metabolites was
established by total synthesis.3,4 Biological evaluation of these
natural products demonstrated that octalactin A (1) was sig-
nificantly cytotoxic in tests with B-16-F10 murine melanoma
and HCT-116 human colon tumor cell lines, whereas octalactin
B (2) was completely inactive.1,5 The therapeutic potential of
octalactin A (1) coupled with the unusual structural features of
both metabolites and the challenges associated with the con-
struction of such systems has rendered the octalactins attractive
targets for total synthesis.
to construct an eight-membered lactone from a saturated seco-
acid precursor using a direct lactonization. This methodology
has further been extended by Buszek.9 Latterly, the Buszek
group have demonstrated the utility of ring-closing metathesis
for the construction of the key eight-membered lactone core of
1 and 2 as well as documenting a large-scale route for the
synthesis of 1.10,11 A completely different approach was adopted
by McWilliams and Clardy for the synthesis of ent-1 and ent-
2.4 Their elegant synthesis involved a Baeyer-Villiger ring
expansion for the conversion of a cycloheptanone into the
corresponding eight-membered lactone core of 1 and 2. Many
other groups have reported formal total syntheses and ap-
proaches toward the octalactins.12-17 Our own approach to
medium-ring oxacycles has been based on ring expansion
reactions of cyclic ketones and ketene acetals to produce both
saturated18-23 and unsaturated23-29 medium-ring lactones. In
(7) Corey, E. J.; Brunelle, D. J.; Stork, P. J. Tetrahedron Lett. 1976, 38, 3405-
3408.
(8) Corey, E. J.; Nicolaou, K. C. J. Am. Chem. Soc. 1974, 96, 5614-5616.
(9) Buszek, K. R.; Jeong, Y.; Sato, N.; Still, P. C.; Muin˜o, P. L.; Ghosh, I.
Synth. Commun. 2001, 31, 1781-1791.
(10) Buszek, K. R.; Jeong, Y. Tetrahedron Lett. 1995, 36, 7189-7192.
(11) Buszek, K. R.; Sato, N.; Jeong, Y. Tetrahedron Lett. 2002, 43, 181-184.
(12) Bach, J.; Garcia, J. Tetrahedron Lett. 1998, 39, 6761-6764.
(13) Inoue, S.; Iwabuchi, Y.; Irie, H.; Hatakeyama, S. Synlett 1998, 735-736.
(14) Kodama, M.; Matsushita, M.; Terada, Y.; Takeuchi, A.; Yoshio, S.;
Fukuyama, Y. Chem. Lett. 1997, 117-118.
In the first total synthesis of the octalactins,3 Buszek used
(15) Andrus, M. B.; Argade, A. B. Tetrahedron Lett. 1996, 37, 5049-5052.
(16) Hulme, A. N.; Howells, G. E. Tetrahedron Lett. 1997, 38, 8245-8248.
(17) Bluet, G.; Campagne, J.-M. Synlett 2000, 221-222.
(18) Carling, R. W.; Holmes, A. B. Chem. Commun. 1986, 565-567.
(19) Clark, J. S.; Holmes, A. B. Tetrahedron Lett. 1988, 29, 4333-4336.
(20) Carling, R. W.; Curtis, N. R.; Holmes, A. B. Tetrahedron Lett. 1989, 30,
6081-6084.
the Gerlach modification6 of the Corey-Nicolaou lactoniza-
† Formerly at GlaxoSmithKline, Gunnels Wood Road, Stevenage, SG1
2NY, U.K.
(1) Tapiolas, D. M.; Roman, M.; Fenical, W.; Stout, T. J.; Clardy, J. J. Am.
Chem. Soc. 1991, 113, 4682-4683.
(2) Throughout the paper, all eight-membered lactones are named as oxocan-
2-ones in accord with Chemical Abstracts nomenclature.
(3) Buszek, K. R.; Sato, N.; Jeong, Y. J. Am. Chem. Soc. 1994, 116, 5511-
5512.
(4) McWilliams, J. C.; Clardy, J. J. Am. Chem. Soc. 1994, 116, 8378-8379.
(5) Perchellet, J.-P.; Perchellet, E. M.; Newell, S. W.; Freeman, J. A.; Ladesich,
J. B.; Jeong, Y.; Sato, N.; Buszek, K. Anticancer Res. 1998, 18, 97-106.
(6) Gerlach, H.; Thalmann, A. HelV. Chim. Acta 1974, 57, 2661-2663.
(21) Carling, R. W.; Clark, J. S.; Holmes, A. B. J. Chem. Soc., Perkin Trans.
1 1992, 83-94.
(22) Carling, R. W.; Clark, J. S.; Holmes, A. B.; Sartor, D. J. Chem. Soc., Perkin
Trans. 1 1992, 95-101.
(23) Burton, J. W.; Clark, J. S.; Derrer, S.; Stork, T. C.; Bendall, J. G.; Holmes,
A. B. J. Am. Chem. Soc. 1997, 119, 7483-7498.
(24) Carling, R. W.; Holmes, A. B. Chem. Commun. 1986, 325-326.
(25) Curtis, N. R.; Holmes, A. B.; Looney, M. G. Tetrahedron 1991, 47, 7171-
7178.
9
2194
J. AM. CHEM. SOC. 2004, 126, 2194-2207
10.1021/ja038353w CCC: $27.50 © 2004 American Chemical Society

A Concise Synthesis of the Octalactins
A R T I C L E S
Figure 1. Transfer of stereochemical information during the Claisen rearrangement.
Scheme 1. Synthesis of the Lactone 7
in the synthesis.31 Claisen rearrangement of the alkenyl-
substituted ketene acetal 8 derived from the corresponding anti-
1,3,-diol would likely proceed via a chair-chair transition state
(Scheme 2) to deliver the desired cis-disubstituted lactone 9;
previous work had indicated that the synthesis of medium-ring
lactams proceeded more efficiently from alkenyl-substituted
ketene aminals derived from anti-1,3-amino alcohols than from
the corresponding syn diastereomers.32
The unsaturated lactone 9 was prepared by the route outlined
in Scheme 3. Aldol reaction of the boron enolate derived from
acetophenone 10 with TMS-propynal 1133 provided the â-keto
alcohol 12 in good yield (74%). Hydroxy-directed anti reduction
of 12 was effected using the Evans protocol34 to provide the
diol 13 as a single diastereomer (mp 92-94 °C from EtOAc/
hexane) as judged by 200 MHz 1H NMR. The TMS protecting
group was removed (NaOH, MeOH), and the acetonide 14 was
formed under standard conditions. Silylcupration of 14 was
conducted according to the Fleming protocol35 and the resultant
alkenylcuprate was quenched with methyl iodide furnishing the
(E)-alkene 15a in respectable yield.36 While this reaction is
known to be very efficient for simpler acetylenes, the yields
for this system were variable even with strict monitoring of the
reaction conditions. At this stage, it was possible to confirm
the relative stereochemistry of the substituents on the acetonide
15a and hence the diol 13. In line with the results of
Rychnovsky,37,38 the 13C NMR chemical shifts of the methyl
groups of the acetonide 15a were at δ 25.0 and 25.1 indicating
that 15a adopts a twist-boat conformation and is therefore
derived from an anti-diol.39 Acetonide removal transformed 15a
into the corresponding diol which was converted into the
selenoacetal 17 in the presence of phenylselanylacetaldehyde
diethylacetal 1630 and PPTS. The selenides 17 were isolated as
a mixture of diastereomers, and no attempt was made to separate
them, although they both appeared to undergo selenoxide
particular, we have exploited the Claisen rearrangement of
alkenyl-substituted ketene acetals to prepare unsaturated medium-
ring lactones, a reaction first reported by Petrzilka,30 although
we have demonstrated that a direct lactonization approach to
unsaturated medium-ring lactones is also feasible.23 Herein we
report a short total synthesis of the octalactins based on the
two-carbon ring expansion reaction of an alkenyl-substituted
ketene acetal to form an unsaturated eight-membered ring
lactone which is stereoselectively hydrogenated to produce the
core saturated lactone of 1 and 2. A notable feature of our
approach is the faithful transfer of stereochemical information
from the sp3 center at C6 and the E-alkene in 3 to the sp3 center
at C4 and the Z-alkene in the lactone 4, presumably through
the chairlike transition state of the Claisen rearrangement (Figure
1).27-29
Results and Discussion
Synthesis of a Model System. At the outset of this work,
we had demonstrated that it was possible to synthesize diaster-
eomerically pure disubstituted seven- and eight-membered
lactones by Claisen rearrangement of the corresponding vinyl-
substituted ketene acetals.27 For example, heating a solution of
the selenoxides 5 and DBU in xylene in a sealed tube at 180
°C afforded the diastereomerically pure lactone 7 with complete
transfer of stereochemical information (Scheme 1).
Before we embarked upon the synthesis of the octalactins, it
was necessary to demonstrate that it would be possible to
synthesize an eight-membered lactone such as 9 containing a
5,6-trisubstituted double bond with concomitant introduction of
the C4 substituent by transfer of stereochemical information
from the sp3 center (C6) of the alkene-substituted ketene acetal
8 (Scheme 2).28,29
For synthetic simplicity, we selected the racemic lactone 9
as our initial target with the knowledge that the phenyldi-
methylsilyl group could potentially be transformed into the
desired hydroxy substituent by Fleming-Tamao oxidation later
(31) Fleming, I. Chemtracts 1996, 9, 1-64.
(32) Evans, P. A.; Holmes, A. B.; McGeary, R. P.; Nadin, A.; Russell, K.;
O’Hanlon, P. J.; Pearson, N. D. J. Chem. Soc., Perkin Trans. 1 1996, 123-
138.
(33) Kruithof, K. J. H.; Schmitz, R. F.; Klumpp, G. W. Tetrahedron 1983, 39,
3073-3081.
(34) Evans, D. A.; Chapman, K. T.; Carreira, E. M. J. Am. Chem. Soc. 1988,
110, 3560-3578.
(35) Fleming, I.; Newton, T. W.; Roessler, F. J. Chem. Soc., Perkin Trans. 1
1981, 2527-2532.
(36) If the silylcupration reaction was prematurely quenched then the corre-
sponding (E)-disubstituted olefin 15b was isolated (Jalkene ) 18.9 Hz). The
alkene 15b arises via protonation of the vinyl copper intermediate. By
analogy, the stereochemistry of vinyl silane 15a was assigend as (E).
(37) Rychnovsky, S. D.; Yang, G.; Powers, J. P. J. Org. Chem. 1993, 58, 5251-
5255.
(38) Rychnovsky, S. D.; Rogers, B.; Yang, G. J. Org. Chem. 1993, 58, 3511-
3515.
(26) Congreve, M. S.; Holmes, A. B.; Hughes, A. B.; Looney, M. G. J. Am.
Chem. Soc. 1993, 115, 5815-5816.
(39) Due to the sterically undemanding nature of a terminal acetylene (A-value
) 0.41-0.52 kcal mol-1) alkynyl-substituted acetonides of anti-1,3-diols
generally exist as an equilibrium mixture of chair and twist-boat conform-
ers.37 For these systems, Rychnovsky has shown that the difference in the
13C NMR chemical shifts for the methyl groups of the acetonide (∆δ) are
usually <5 ppm. For the alkynyl-substituted acetonide 14, ∆δ ) 5.8 ppm
indicating that 14 was most likely derived from an anti-1,3-diol. Confirma-
tion of this stereochemical assignment was therefore made on the vinyl
silane 15a.
(27) Fuhry, M. A. M.; Holmes, A. B.; Marshall, D. R. J. Chem. Soc., Perkin
Trans. 1 1993, 2743-2746.
(28) Anderson, E. A.; Davidson, J. E. P.; Harrison, J. R.; O’Sullivan, P. T.;
Burton, J. W.; Collins, I.; Holmes, A. B. Tetrahedron 2002, 58, 1943-
1971.
(29) Harrison, J. R.; Holmes, A. B.; Collins, I. Synlett 1999, 972-974.
(30) Petrzilka, M. HelV. Chim. Acta 1978, 61, 3075-3078.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2195

A R T I C L E S
O’Sullivan et al.
Scheme 2. Proposed Claisen Rearrangement of the Ketene Acetal 8
Scheme 3. Synthesis of the Lactone 9
elimination and Claisen rearrangement with equal facility. The
selenides were oxidized to the corresponding selenoxides which,
on being heated as a dilute solution in xylene in a sealed tube
at 180 °C in the presence of DBU and the silyl ketene acetal
18, provided the ketene acetal 8 which underwent in situ Claisen
rearrangement to provide the desired eight-membered lactone
9 in 31% yield. The use of the ketene acetal 18 as a trap for
“PhSe+” had previously been shown to enhance related Claisen
ring expansion reactions which were sensitive to side reactions
involving disproportion promoted by “PhSe+”.32 The lactone 9
was isolated as a single diastereomer indicating complete transfer
of stereochemical information between C6 in 8 and C4 in 9.
Due to overlapping 1H NMR signals from several key protons
on the ring, assignment of the relative stereochemistry of 9 on
Hiyama-Kishi reaction40 (NHK reaction), a process which has
precedent from a previous synthesis of 2.3 The vinyl iodide 23
was to be generated by hydrozirconation of the corresponding
acetylene which would be available through elaboration of the
boronic acid 24 according to the procedure of Yamamoto;41
Andrus has also used this methodology for the synthesis of the
side chain of the octalactins.15 The saturated eight-membered
lactone 19 was to be prepared by substrate controlled hydro-
genation of the trisubstituted ring-alkene 4.42 A key test of the
synthetic strategy would be the stereoselective reduction of the
double bond in 4, and we anticipated that the reduction would
occur preferentially from the face opposite the C4 and C8
substituents. Success in the model Claisen rearrangement study
implied that we would be able to prepare the unsaturated lactone
4 by [3,3]-sigmatropic rearrangement of the alkenyl-substituted
ketene acetal 3 which was to be generated either by in situ
selenoxide elimination from 20 or by methylenation of the
carbonate 21 using dimethyltitanocene.28 The synthesis therefore
relied upon the preparation of the triol derivative 20 or 21. The
anti-1,3-diol precursor to 20 and 21 would be necessary in order
to set the desired stereochemistry at C4 in the lactone 4, and
based on previous experience, we were confident that Claisen
rearrangement from the derived alkenyl-substituted ketene acetal
3 would occur in good yield.32 The elegant aldol methodology
of Paterson,43-45 based on Ipc boron enolates, was to be used
to prepare 22 which would undergo anti-selective reduction34
1
the basis of H NMR NOE experiments proved inconclusive.
The stereochemistry of 9 was therefore tentatively assigned as
cis on the basis of the likely chair-chair transition state of the
Claisen rearrangement which places the phenyl group in a
pseudoequatorial position (see Scheme 2).23
The success of this strategy toward a disubstituted eight-
membered lactone bearing a trisubstituted double bond, which
was the most stereochemically demanding Claisen rearrange-
ment which we had realized at that time, encouraged us to press
on with the total synthesis of the octalactins. However, due to
the capricious nature of the silylcupration reaction used to
generate 15a, we decided to investigate the possibility of
carrying a protected hydroxy group through the synthesis from
the outset.
Retrosynthetic Analysis and Synthetic Planning. As had
been previously demonstrated, octalactin A (1) may be synthe-
sized from octalactin B (2) by hydroxy directed epoxidation.4
Our initial target was therefore the enone 2. The retrosynthetic
analysis of 2 (Scheme 4) depends on the convergent union of
the aldehyde 19 with the vinyl iodide 23 using the Nozaki-
(40) Fu¨rstner, A. Chem. ReV. 1999, 99, 991-1045.
(41) Ikeda, N.; Arai, I.; Yamamoto, H. J. Am. Chem. Soc. 1986, 108, 483-
486.
(42) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV. 1993, 93, 1307-
1370.
(43) Paterson, I.; Goodman, J. M.; Isaka, M. Tetrahedron Lett. 1989, 30, 7121-
7124.
(44) Paterson, I.; Goodman, J. M.; Lister, M. A.; Schumann, R. C.; McClure,
C. K.; Norcross, R. D. Tetrahedron 1990, 46, 4663-4684.
(45) Paterson, I.; Oballa, R. M. Tetrahedron Lett. 1997, 38, 8241-8244.
9
2196 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

A Concise Synthesis of the Octalactins
A R T I C L E S
Scheme 4. Retrosynthetic Analysis of the Octalactinsa
a P,P′ ) protecting groups.
Scheme 5. Synthesis of the Aldehyde 25
Scheme 6. Synthesis of the Diol 29
First Generation Synthesis of the Core Saturated Eight-
Membered Lactone. The construction of the required core
eight-membered lactone began with the convergent aldol union
of the silyloxy aldehyde 25 with the methyl ketone 26. The
aldehyde 2546 was prepared from the enolate 27 by adaptation
of a literature procedure (Scheme 5).47 The salt 27 is readily
obtained from either tetraethoxymethylpropane or 3-ethoxymeth-
acrolein via acid-catalyzed hydrolysis followed by addition of
potassium hydroxide.46,48,49 The stereochemistry of 25 was
to provide the desired diol. Note that this strategy combines
the requirement of an anti-1,3-diol with the use of substrate-
induced stereocontrol to generate the required absolute stereo-
chemistry at C4 and C8 in the lactone 4. Not only does the
Claisen rearrangement induce the C4 stereocenter in 4, but it
also produces the desired (Z)-trisubstituted double bond at the
expense of the C6 stereocenter in the ketene acetal 3. As in the
synthesis of many complex natural products, the choice of
protecting groups would prove crucial to the success of our plan.
At the outset of the synthesis, we chose to prepare the
vinylogous silyl ester 25 (Scheme 5) and the known benzyl-
protected methyl ketone 26 (Scheme 6).43 This choice was
guided, in part, by Clardy’s synthesis of the octalactins where
a silyl group had been used to protect the oxygen functionality
at C4 of 1 and 2.4
1
assigned by H NMR NOE measurements; a strong reciprocal
NOE was observed between H1 and H3 with no NOE being
observed between H3 and the methyl group. The known methyl
ketone 26 was prepared from (R)-methyl-3-hydroxy butyrate.43
The aldol reaction of the Ipc-boron enolate derived from the
methyl ketone 26, and the aldehyde 25 afforded the 1,4-syn-
â-hydroxy ketone 28 in good yield and reasonable diastereo-
selectivity (71%) (Scheme 6).43-45,50 Attempted formation of
(46) Ullrich, F.-W.; Rotscheidt, K.; Breitmaier, E. Chem. Ber. 1986, 119, 1737-
1744.
(47) The corresponding trimethylsilyl-protected aldehyde has been prepared.46
(48) Klimo, V. T.; Skoldinov, A. P. Chem. Abstr. 1960, 54, 20870.
(49) Isler, O.; Zeller, P. (Hoffmann-La Roche) Ger., 1,011,881, 11 Jul 1957.
Chem. Abstr. 1959, 53, P15116e.
(50) Cowden, C. J.; Paterson, I. Org. React. 1997, 51, 1-200.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2197

A R T I C L E S
O’Sullivan et al.
basic ethanol.30 A plausible strategy for the formation of the
required 1,3-dioxane could involve generation of the mixed
acetals 32 and 33 under mild basic conditions followed by mild
intramolecular transacetalization. Addition of a mixture of ethyl
vinyl ether and phenylselenyl chloride to a solution of the diol
29 and diisopropylethylamine in ethanol provided the two mixed
acetals 32 and 33 as a 2:1 mixture in a combined yield of 55%.
The structures of the regioisomeric acetals were assigned on
Chart 1
1
the basis of H NMR COSY experiments. Treatment of the
major acetal 33 with a range of mild protic and Lewis acids
resulted in formation of either the diene 30 or the R,â-
unsaturated aldehyde 31. Indeed heating a solution of the acetal
33 in toluene at reflux in the presence of 4 Å molecular sieves
resulted in formation of the diene 30 indicating the acute acid
sensitivity of the mixed acetal. In a further series of experiments,
formation of the desired acetal was attempted using Noyori’s
protocol54 (TMS-triflate and the bis-TMS ether derived from
29), but these conditions resulted in decomposition of the
substrate.
A further method which would potentially allow access to
the desired ketene acetal would involve reaction of a titanium
methylidene carbene with the cyclic carbonate derived from the
1,3-diol 29.28 However, the difficulty in forming the Mosher
esters of the aldol 28 was most probably a result of the very
facile elimination that occurs when the hydroxy group in 28 is
made into a good leaving group. Therefore formation of
corresponding cyclic carbonate derived from 29 was likely to
prove challenging. Indeed exposure of the diol 29 to either
triphosgene55 or carbonyldiimidazole56 resulted in decomposi-
tion. By this stage, it was apparent that the silylenol ether was
too electron rich and was promoting â-elimination reactions
under both acidic and basic conditions. We therefore switched
our attention to the use of an electron-withdrawing enol-acetate
substituent to be derived from the aldehyde 34 (Scheme 7). At
this stage, we were also mindful of the impending challenge of
selective deacetylation of the C4 acetoxy substituent in the
presence of a medium-ring lactone at a late stage in the
synthesis.
the Mosher esters from 28 was unsuccessful and resulted in
decomposition of the substrate. The stereochemistry of the major
diastereomer 28 was therefore tentatively assigned as 1,4-syn
on the basis of precedent.43,44 The decomposition of 28 under
the conditions for Mosher ester formation indicated the delicate
nature of the compound and hinted at the problems that we were
soon to face. Anti-reduction34 of the aldol 28 provided the
corresponding anti-diol 29. The diol 29 was relatively unstable
and was therefore stored as a dilute solution in ethyl acetate at
-20 °C and was not purified by column chromatography.
Attempted formation of the acetonide derived from 29 to allow
stereochemical assignment resulted in decomposition. However,
the stereochemistry of the diol could be assigned by comparison
1
1
of the H NMR with the H NMR of 15a (see Supporting
Information).
We have gained extensive expertise in the formation of
medium-ring lactones (and lactams32,51-53) by the Claisen
rearrangement of alkenyl-substituted ketene acetals (and ami-
nals). In these studies, we have generally formed the ketene
acetal according to the original procedure of Petrzilka,30 namely
by pyrolytic selenoxide elimination from the corresponding
acetal. These acetals are formed by acid-catalyzed acetal
exchange between phenylselanylacetaldehyde diethyl acetal 16
and the 1,n-diol. More recently, we have reported a comple-
mentary method for the formation of ketene acetals via
methylenation of the corresponding carbonates using a titanium-
methylidene reagent.28 The carbonates (e.g., 21) are formed at
low temperature under mildly basic conditions thus comple-
menting the acid-catalyzed formation of the selenoacetals.
Exposure of the diol 29 to our standard conditions for
selenoacetal formation (acetal 16, PPTS, toluene at reflux)
yielded the diene 30 (Chart 1), the structure of which was
assigned as the 2(E),4(E) isomer (J4,5 ) 15.0 Hz). The diene is
assumed to arise via acid-catalyzed elimination of the hydroxy
groups at C3 and C5 of 29. A wide variety of protic and Lewis
acids were screened in an effort to form the desired cyclic
selenoacetal; however, all efforts met with failure. Either starting
material (29) was recovered, or more usually, the diol 29
decomposed or was transformed into the aldehyde 30. The
synthesis of the selenoacetal 16 involves the addition of
phenylselenyl chloride to ethyl vinyl ether in the presence of
The enol acetate 34 was synthesized in an analogous manner
to that used for the enol silane 25 by reaction of the salt 2746
with acetyl chloride (Scheme 7); the stereochemistry of the
1
product 34 was assigned by H NMR NOE experiments. The
Paterson-Aldol reaction43-45,50 between 26 and 34 proceeded
without incident to provide the 1,4-syn-aldol 35 with good
1
diastereocontrol (88% de as judged by 400 MHz H NMR).
The absolute configuration at the newly formed stereocenter
was assigned using Kakisawa’s extension of Mosher’s method
(see Supporting Information).57 Anti-selective reduction of the
keto-alcohol 35 with tetramethylammonium triacetoxyborohy-
dride34 provided the anti-diol 36 as a single diastereomer; the
relative stereochemistry of 36 was assigned as anti on the basis
of much precedent. Exposure of 36 to phenylselanylacetaldehyde
diethyl actetal30 16 and PPTS in toluene at reflux in the presence
of 4 Å molecular sieves provided the corresponding selenoac-
etals 37 in 56% optimized yield as a mixture of acetal
diastereomers. These reactions have been shown to be highly
(51) Evans, P. A.; Holmes, A. B.; Collins, I.; Raithby, P. R.; Russell, K. Chem.
Commun. 1995, 2325-2326.
(54) Tsunoda, T.; Suzuki, M.; Noyori, R. Tetrahedron Lett. 1980, 21, 1357.
(55) Burk, R. M.; Roof, M. B. Tetrahedron Lett. 1993, 34, 395-398.
(56) Kutney, J. P.; Ratcliffe, A. H. Synth. Commun. 1975, 5, 47-52.
(57) Ohtani, I.; Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem. Soc.
1991, 113, 4092-4096.
(52) Evans, P. A.; Holmes, A. B.; Russell, K. J. Chem. Soc., Perkin Trans. 1
1994, 3397-3409.
(53) Evans, P. A.; Holmes, A. B.; Russell, K. Tetrahedron Lett. 1992, 33, 6857-
6858.
9
2198 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

A Concise Synthesis of the Octalactins
A R T I C L E S
Scheme 7. Synthesis of the Lactone 38
Scheme 8. Conversion of the Lactone 38 into the Aldehyde 41
susceptible to the presence of residual acetic acid from the
triacetoxyborohydride reduction. If the final product is not
carefully washed with saturated sodium hydrogencarbonate
solution, it is impossible to form the acetals 37; the only products
observed are the aldehydes 30 and 31. Even with these
precautions, a more polar highly UV active component (the
hydroxy-aldehyde 31) was always observed. It was therefore
not possible to completely suppress the formation of the
elimination product 31 even when the olefin carried an electron-
withdrawing group. The oxidation of the selenides 37 to the
corresponding selenoxides proceeded without event. Heating the
selenoxides using our standard conditions (xylene at reflux in
the presence of DBU)23 provided the desired eight-membered
lactone 38 in poor yield (22%) along with recovered selenides
37 (25%). The low yield in this case was attributed to the known
thermal cleavage of acetate esters by DBU58 and to dispropor-
tionation reactions mediated by “PhSe+”.
As previously noted, the selenides are most likely formed
by reduction of the selenoxides with diphenyl diselenide.28 The
diphenyl diselenide is produced by the disproportionation of
the benzeneseleninic acid which is a product of the selenoxide
elimination. The disproportionation may be suppressed by the
use of non-nucleophilic bases and the addition of a nucleophilic
silyl ketene acetal (e.g. 18) as a selenium scavenger. After much
experimentation, it was discovered that heating the selenoxides
in a sealed tube in xylene at 180 °C in the presence of potassium
carbonate and 18 furnished the desired lactone 38 in an
optimized yield of 40% as well as 24% of recovered selenides.
Use of silyl-ketene aminals as selenium scavengers27 or the
addition of mild bases gave inferior yields of the lactone. In all
cases, 38 was isolated as a single diastereomer.
transition state with the R-methylbenzyloxyethyl side chain
adopting a pseudoequatorial position (see Figure 1).
Transformation of the lactone 38 into the aldehyde 41 required
for the side chain coupling necessitated the removal of the
benzyl protecting group and the hydrogenation of the trisub-
stituted double bond from the top face as drawn. In initial
catalyst screens, it was shown that palladium hydroxide effected
the hydrogenolysis of the benzyl group with concomitant
hydrogenation of the olefin to provide 40. However, the product
was isolated as a 1:1 mixture of C5 diastereomers. The use of
Adams catalyst (PtO2) provided the saturated lactones 39 as a
mixture of inseparable C5 diastereomers (39a:39b, 4:1) (Scheme
8); careful monitoring of the reaction was required to minimize
subsequent reduction of the benzyl group to a cyclohexylmethyl
group, giving 42.
Many of the hydrogen atoms adorning the ring of the lactone
38 showed vicinal 1H NMR coupling constants of approximately
7 Hz, indicating that the ring was conformationally mobile and
that the conformers were in “fast exchange” on the NMR time
scale. Nevertheless, the stereochemistry of the lactone 38 could
1
be assigned on the basis of H NMR NOE data; a reciprocal
NOE was observed between H4 (δ 5.65, t, J ) 7.2 Hz, 1H)
and H8 (δ 4.66, q, J ) 7.5 Hz, 1H) in the lactone 38, indicating
that these two substituents were on the same face of the ring.
The formation of the (4R)-lactone 38 may be explained by the
preference for the Claisen rearrangement to adopt a chairlike
The stereochemistry of the major lactone 39a was assigned
on the basis of a H NMR NOESY experiment. An NOE was
(58) Baptistella, L. H. B.; dos Santos, J. F.; Ballabio, K. C.; Marsaioli, A. J.
1
Synthesis 1989, 436-438.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2199

A R T I C L E S
O’Sullivan et al.
Scheme 9. Synthesis of the TBS-Protected Lactone 47
observed between H5 (δ 1.77-1.72, m, 1H) and H8 (4.45-
4.40, m, 1H) and between H5 and H4 (δ 5.00, brd, J ) 6.2 Hz,
1H). Further NOEs supported the assignment of the structure
as shown and indicated that hydrogenation of 38 had occurred
preferentially on the face of the olefin opposite to both the C4
and C8 side chains. Although we had not experienced any
difficulty in hydrogenating the trisubstituted ring-double bond
in 38 to provide 39, Buszek and co-workers,3 in the first
synthesis of the octalactins, had found it impossible to hydro-
genate a disubstituted double bond between C6 and C7 in a
similar medium-ring lactone. However, if the reduction of the
disubstituted double bond was delayed until installation of the
complete octalactin side chain, then the hydrogenation could
be accomplished with ease.11 The benzyl group in 39 was
removed by hydrogenolysis in the presence of palladium
hydroxide, providing the primary alcohol 40 (9:1 mixture of
C5 diastereomers after purification by flash chromatography)
which was oxidized using the Dess-Martin periodinane59-61
to provide the aldehyde 41 in 49% yield in readiness for side
chain introduction. We had now managed to construct the core
of the octalactins utilizing our Claisen rearrangement methodol-
ogy. However, the modest yields encountered in the later stages
of the synthesis meant that this route was unattractive for
delivery of sufficient material for the completion of the
synthesis.
Second Generation Synthesis of the Core Saturated Eight-
Membered Lactone. Some of the difficulties associated with
the first generation synthesis of the core lactone 41 resulted
from our use of a benzyl protecting group. This group had
interfered with the hydrogenation of the trisubstituted double
bond in 38 and its removal by hydrogenolysis had not been
high yielding. We therefore elected to use a more labile silyl
ether protecting group. Furthermore, the formation of the
selenides 37 under acidic conditions occurred in moderate yield
and also resulted in the formation of some elimination products.
We therefore decided to use our more recently developed
methodology for the formation of ketene acetals from carbonates
via methylenation with dimethyltitanocene.28 Although we had
already demonstrated that it was possible to form medium-ring
lactones from alkenyl-substituted cyclic carbonates by exposure
to dimethyltitanocene in toluene at reflux,28 the use of a substrate
bearing an acetate protecting group would provide a further test
of this methodology; would it be possible to selectively
methylenate a carbonate carbonyl group in the presence of an
acetate carbonyl group?
The synthesis of the anti-diol 45 followed the same course
as the synthesis of 36 (Scheme 9). Thus, 1,4-syn-selective aldol
reaction between the Ipc boron enolate of the ketone 4362 and
the aldehyde 34 proceeded without incident to provide 4443-45,50
(the stereochemistry at C3 was assigned on the basis of
precedent).63 Evans34 anti-reduction then delivered the diol 45
(94% de) in good yield. Exposure of the diol 45 to our optimized
conditions for carbonate formation (triphosgene, pyridine,
triethylamine, 4 Å molecular sieves, dichloromethane, -78 f
0 °C)28 provided the unstable cyclic carbonate 46 in good yield
(71%). Treatment of the carbonate 46 with dimethyltitanocene
in toluene at reflux28,64 for 30 min resulted in methylenation
and subsequent Claisen rearrangement to provide the desired
eight-membered lactone 47 in reasonable yield (42%) after
purification by flash chromatography. The stereochemistry of
47 [δ 5.70 (t, J ) 7.4 Hz, 1H, H4), 5.43 (t, J ) 6.2 Hz, 1H,
H6), 4.67 (dt, J ) 5.0, 9.3 Hz, 1H, H8)] was initially assigned
by comparison of the 1H NMR spectrum with the benzyl-
protected lactone 38 [δ 5.65 (t, J ) 7.2 Hz, 1H, H4), 5.42 (t, J
) 6.7 Hz, 1H, H6), 4.66 (q, J ) 7.5 Hz, 1H, H8)] and was
proven by crystal structure analysis of a later translactonization
product 48 (vide infra). Although the yield for the Claisen
rearrangement to form the lactone 47 was similar to the
selenium-based route to form 38, the formation of the carbonate
46 occurred in significantly higher yield (71%) compared with
the formation of the selenoacetals (56%). Furthermore, the
conversion of the carbonate 46 into the lactone 47 occurred in
toluene at reflux under relatively concentrated conditions (0.053
M), whereas the formation of the lactone 38 from the selenox-
ides derived from 37 required heating at 180 °C in a sealed
tube at lower concentration (0.017 M). We28 and others65-67
had previously shown that it is possible to perform chemose-
lective methylenations of carbonyl groups using titanium based
reagents; however, we believe that this is the first report of the
selective methylenation of a carbonate carbonyl group in the
presence of an ester carbonyl group. A plausible explanation
for this chemoselectivity is that the lone pair electrons of the
carbonate carbonyl group interact more readily with the bulky
electrophilic titanium carbene generated from dimethylti-
(62) The enantiomer of the ketone 43 is known. Denmark, S. E.; Fujimori, S.
Synlett 2001, 1024-1029.
(63) Based on the isolated yield of diastereomerically pure 44, the lower limit
of the de of the aldol reaction of 43 and 34 is 76%.
(64) Petasis, N. A.; Bzowej, E. I. J. Am. Chem. Soc. 1990, 112, 6392-6394.
(65) Pine, S. H.; Pettit, R. J.; Geib, G. D.; Cruz, S. G.; Gallego, C. H.; Tijerina,
T.; Pine, R. D. J. Org. Chem. 1985, 50, 1212-1216.
(66) Vauzeilles, B.; Sinay¨, P. Tetrahedron Lett. 2001, 42, 7269-7272.
(67) Bartlett, P. A.; Nakagawa, Y.; Johnson, C. R.; Reich, S. H.; Luis, A. J.
Org. Chem. 1988, 53, 3195-3210.
(59) Meyer, S. D.; Schreiber, S. L. J. Org. Chem. 1994, 59, 7549-7552.
(60) Ireland, R. E.; Liu, L. B. J. Org. Chem. 1993, 58, 2899.
(61) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4156.
9
2200 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

A Concise Synthesis of the Octalactins
A R T I C L E S
Scheme 10. Deprotection with Concommitant Translactonization
Scheme 11. Synthesis of the Aldehyde 41
Figure 2. Chem-3D representation of the X-ray crystal structure of the
lactone 48. The acetate group carbonyl oxygen atom is disordered.
tanocene68 than the lone pairs of the enol acetate carbonyl group.
This may be mainly a steric effect as the alkyl groups of the
carbonate 46 are “tied-back” into a six-membered ring rendering
the carbonate carbonyl group lone pairs more available (and
more basic) than those of the enol acetate of 46.69 Furthermore,
the lone pairs of the enol acetate carbonyl group are likely to
be less basic than those of a standard acetate ester.
With the lactone 47 in hand, we had the choice of either
introducing the side chain and then hydrogenating the trisub-
stituted double bond or vice versa. Exposure of 47 to HF‚
pyridine (Scheme 10) resulted in cleavage of the silyl group
and translactonization to provide the crystalline ten-membered
lactone 48 in excellent yield. The lactone 48 formed crystals
suitable for X-ray structure determination which allowed
assignment of the relative (and hence absolute) stereochemistry
of C4 in 47. The structure is shown in Figure 2.70 We have
noted this type of ring-expansion reaction previously,26 and the
reaction is likely driven by the thermodynamic stability of the
ten-membered lactone 48 compared with the desilylated eight-
membered lactone derived from 47.71
Due to the rearrangement encountered during desilylation of
47, we elected to hydrogenate the trisubstituted double bond of
47 before desilylation. We switched our attention to the use of
rhodium on alumina as the use of Adams catalyst had effected
the transformation of 38 into 39 with moderate diastereoselec-
tivity. In an initial encouraging experiment, the desired product,
the saturated lactones 49, was isolated as an inseparable 5:1
mixture of diastereomers (by 1H NMR analysis), although this
result was not reproducible in subsequent experiments (Scheme
11). The stereochemistry at C5 of the lactones 49 was assigned
by comparison of the 1H NMR spectra of 49a [δH 5.03 (d, J )
6.1 Hz, 1H, H4)] and 49b [δH 4.77 (ddd, J ) 8.6, 5.0, 3.6 Hz,
1H, H4)] with 39a [δH 5.00 (d, J ) 6.0 Hz, 1H, H4)] and 39b
[δH 4.76 (ddd, J ) 8.8, 5.1, 3.7 Hz, 1H, H4)] and ultimately
rests with the conversion of the major diastereomer 49a into
the octalactins.
As discussed above, the benzyl-protected lactone 38 exists
as a mixture of medium-ring conformers at room temperature.
1
Due to similarities in the H NMR spectrum it is most likely
that the lactone 47 also exists as a mixture of medium-ring
conformers. Therefore we postulated that the selectivity of the
hydrogenation might be increased by conducting the reaction
at lower temperature. In the event, a stepwise lowering of the
temperature at which the hydrogenation was conducted im-
proved the selectivity. At -22 °C, the saturated lactones were
formed in excellent yield and with good selectivity (8.2:1) in
favor of the desired diastereomer 49a; further lowering of the
reaction temperature led to poor conversion and no improvement
in selectivity. Removal of the silyl ether protecting group from
a mixture of the lactones 49 with HF‚pyridine provided the
alcohol 40 and the corresponding C5 epimer which were readily
separable by flash chromatography.72 Oxidation of the alcohol
40 with freshly prepared Dess-Martin periodinane59-61 provided
the acid-sensitive unstable aldehyde 41 which was purified by
filtration through Florisil and used immediately. Use of older
batches of the periodinane resulted in significant epimerization
of the methyl group adjacent to the aldehyde in 41, and the
product was isolated in much poorer yield (see Scheme 8).
By the judicious choice of protecting groups, and the use of
a cyclic carbonate (46) as a Claisen rearrangement precursor,
we had secured an efficient, reproducible, and scalable synthesis
(68) The Petasis olefination has been shown to proceed by way of a titanium
carbene. Hughes, D. L.; Payack, J. F.; Cai, D.; Verhoeven, T. R.; Reider,
P. J. Organometallics 1996, 15, 663-667. Meurer, E. C.; Santos, L. S.;
Pilli, R. A.; Eberlin, M. N. Org. Lett. 2003, 5, 1391-1394.
(69) The hydrogen-bond basicity pKHB of carbonates and esters has been
measured. Besseau, F.; Laurence, C.; Berthelot, M. J. Chem Soc., Perkin
Trans. 2 1994, 485-489.
(70) Compound 48, C13H20O5, MW 256.30, was obtained as clear colorless
crystals, space group P31, a ) 12.843(2) Å, c ) 7.107(3) Å, V ) 1015.2(5)
Å3, Z ) 3, Dcalcd ) 1.258 g cm-3, F(000) ) 414. More detailed crystal
structure data can be viewed in the Supporting Information.
(72) The eight-membered lactone 40 did not undergo translactonization under
(71) Wiberg, K. B.; Waldron, R. F. J. Am. Chem. Soc. 1991, 113, 7697-7705.
the reaction conditions.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2201

A R T I C L E S
O’Sullivan et al.
Scheme 12. Synthesis of the TPS-Protected Vinyl Iodide 50
Scheme 13. Side Chain Coupling
of the core eight-membered lactone of the octalactins, thus
allowing us access to significant quantities of the aldehyde 41.
All that remained for the completion of the natural products
was to elaborate the C8 side chain.
Side Chain Synthesis. The vinyl iodide 50 was synthesized
in a manner similar to the route developed by Andrus.15 Thus
the known boronic acid 24 was synthesized from propargyl
bromide and trimethylborate according to the procedure of
Yamamoto.41 Addition of (+)-diethyl tartrate allowed formation
of the corresponding enantiomerically pure boronic ester deriva-
tive (Scheme 12). Subsequent reaction with isobutyraldehyde
provided the homopropargylic alcohol 51 in 80% ee (Mosher
ester analysis; see Supporting Information). The volatile alcohol
51 was stored as a solution in toluene and was converted to the
corresponding silyl ether as required (TPSCl, imidazole, DMF).
Methylation of the terminal acetylene occurred in quantitative
yield (n-BuLi, MeI) to provide 52. Finally, following the route
of Buszek,3 hydrozirconation of the internal acetylene 52 using
Schwartz’s reagent (4 equiv) in THF, equilibration at room
temperature for 4 h, followed by the addition of a solution of
iodine in THF provided the vinyl iodides 50a (δ 5.98, tq, J )
7.5, 1.3 Hz, 1H, H3) and 50b (δ 6.20, q, J ) 7.0 Hz, 1H, H2)
as a 9:1 inseparable mixture of regioisomers in reasonable yield
(60%).
All that remained for the synthesis of octalactin B (2) was
oxidation of the allylic alcohols to a ketone followed by
protecting group removal. The formation of the ketone 54 from
a mixture of alcohols 53 occurred without event. However,
removal of the silyl protecting group with aqueous HF was
inefficient (41% yield). It appeared that the long reaction time
(40 h) required for removal of the bulky silyl protecting group
was causing decomposition of the substrate 54, the product 55,
or both.73 Our attention therefore turned to the use of the TBS-
protected vinyl iodide 56a in the coupling reaction, as Buszek
had demonstrated that removal of this protecting group occurred
with good efficiency.3 The vinyl iodide 56a was synthesized in
an analogous manner to the TPS-protected vinyl iodide 50a but
with three important modifications. First, benzene was used as
the solvent for the hydrozirconation; second, the vinyl zirconium
intermediates were allowed to equilibrate for 4 h at 40 °C before
addition of iodine; and third, an aqueous workup procedure was
employed.74 The vinyl iodides 56a and 56b were isolated as a
10:1 mixture of regioisomers in a combined yield of 79%. The
previously optimized procedure for side chain coupling provided
the desired products 57a and 57b. However, 1H NMR indicated
the presence of two inseparable impurities, assigned as 57c and
57d, that had resulted from coupling of the regioisomeric iodide
56b (Chart 2). Surprisingly the difference in steric bulk of a
TBS-ether compared with a TPS-ether in the vinyl iodides 50
and 56 is sufficient to influence the outcome of the NHK
coupling reaction.
Fragment Coupling and Completion of the Syntheses. The
final stages of the synthesis involved coupling of the side chain
50a with the aldehyde 41. Initial studies focused on the use of
the Nozaki-Hiyama-Kishi (NHK) reaction to effect this
transformation.40 Thus treatment of a solution of the aldehyde
41 and the mixture of vinyl iodides 50 in DMSO with 11 equiv
of CrCl2/0.1% NiCl2 provided the desired coupled products 53a
and 53b as a mixture of diastereomers in good overall yield
(69%) (Scheme 13). Pleasingly, only the less-hindered vinyl
iodide 50a underwent the coupling reaction. The absolute
stereochemistry of the secondary alcohols 53 was proved using
Kakisawa’s extension of Mosher’s method57 (see Supporting
Information), thus demonstrating that the coupling reaction had
occurred with modest Felkin-Anh selectivity as expected.3,40
The reaction with the pure vinyl iodide 56a (obtained by
careful flash chromatography in hexane) provided the desired
(73) Clardy reported a similar low yield for removal of the corresponding side
chain TPS group.4
(74) Nakatsuka, M.; Ragan, J. A.; Sammakia, T.; Smith, D. B.; Uehling, D. E.;
Schreiber, S. L. J. Am. Chem. Soc. 1990, 112, 5583-5601.
9
2202 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

A Concise Synthesis of the Octalactins
A R T I C L E S
Chart 2
Scheme 14. Synthesis of Acetyl-Protected Octalactin B
Chart 3
products 57a and 57b in excellent yield (91%) as a 1.8:1 mixture
of separable diastereomers. Surprisingly the H NMR spectra
Scheme 15. Synthesis of Octalactin B
1
of 57a and 57b differed significantly from the 1H NMR spectra
of 53a and 53b. Nevertheless, we presumed that the reaction
had proceeded under Felkin-Anh control (vide supra), and thus
the major diastereomer was assigned the (2′R)-configuration 57a
and the minor diastereomer was assigned the (2′S)-configuration
57b. Oxidation of the minor diastereomer 57b with the Dess-
Martin periodinane59-61 provided the ketone 58 (Scheme 14).
Pleasingly, deprotection of the silyl group with HF‚pyridine
Our attention therefore turned to the use of an enzymatic
method for the final deprotection.75,78 After much experimenta-
tion, it was discovered that lipase type VII from Candida
cylindracea (an enzyme known to accommodate bulky esters
in its active site78) could be used for the deprotection. Thus
exposure of 39a to the enzyme in diethyl ether containing 3%
pH 7 buffer provided a small quantity of the desired hydroxy-
lactone 61 after stirring for 22 days (Chart 3). Complete
conversion could be achieved after 7 days if the reaction was
conducted in 10% DMF/pH 7 buffer.79 In the real system,
exposure of 55 to the lipase in 10% DMF/pH 7 buffer provided
1
provided the alcohol 55 in good yield (80%). The H NMR
spectrum of 55 indicated the presence of a minor impurity
(<10%) [δ 6.81 (t, J ) 6.7 Hz, 1H)]. The impurity was
identified as being the C6′ side chain epimer. The epimer arises
from the moderate enantiomeric excess (80%) obtained in the
synthesis of the (S)-alkynol 51 from the allenyl boronic acid
24. The minor impurity was removed by preparative thin-layer
chromatography after deprotection of the acetate ester (vide
infra).
All that remained for the synthesis of octalactin B (2) was
the deprotection of the acetate group. The lactones 39a and 42
were used as model substrates to test this transformation.
Unfortunately, attempted deprotection of the acetate group in
these systems under a wide variety of conditions (NH3, MeOH,
0 °C; KCN, EtOH, 5 °C; DBU, toluene; DBU, MeOH; K2CO3,
MeOH; Et3N, THF, water, MeOH; NH2-NH2, water, pyridine,
AcOH; BF3‚OEt2, MeCN, water)75,76 resulted in decomposition
of the substrate, formation of the R,â-unsaturated lactone 59,
or formation of ring-opened products 60 which retained the
acetate group (Chart 3).77
octalactin B (2) {[R]D -123 (c 0.04 in CDCl3), lit.3 [R]D
24
-126} in 96% yield as a clear and colorless oil (Scheme 15).
The spectral data (1H NMR, 13C NMR, IR, MS) for our
synthetic sample of 2 were in complete agreement with the data
reported by Buszek for synthetic (-)-octalactin B3 and that
(77) Due to problems with the deprotection of the acetate ester in the presence
of a medium-ring lactone, we explored the possiblity of synthesizing the
corresponding lactone bearing a chloroacetate protecting group. Clardy had
previously demonstrated that this group is more labile than a medium-ring
lactone to treatment with ammonia in THF.4 Unfortunately it has not so
far proved possible to synthesize the chloroacetate-protected aldehyde
corresponding to 34.
(78) Wong, C.-H.; Whitesides, G. M. Enzymes in Synthetic Organic Chemistry;
Pergamon: Oxford, 1994; Vol. 12.
(79) We are much indebted to Prof. C. -H. Wong for advice in using DMF as
a cosolvent for this reaction and for screening a range of other lipases.
(75) Greene, T. W.; Wuts, P. G. M. ProtectiVe Groups in Organic Synthesis,
3rd ed.; John Wiley & Sons: New York, 1999.
(76) Ishido, Y.; Nakazki, N.; Sakairi, N. Chem. Commun. 1976, 832-833.
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2203

A R T I C L E S
O’Sullivan et al.
Scheme 16. Synthesis of Octalactin A
Figure 3. Chem-3D representation of the X-ray crystal structure of synthetic
octalactin A (1).
The optical rotation of our synthetic sample of 1 was ap-
proximately an order of magnitude greater than that published
for the natural sample1 (vide supra). Synthetic octalactin A (1)
formed crystals suitable for X-ray structure determination; the
structure is shown in Figure 3 and confirms our previous
stereochemical assignments.82
Conclusion
In summary, we have reported above an efficient synthesis
of the octalactins. The synthesis of octalactin A (1) proceeds in
15 steps (longest linear sequence) and 10% overall yield (86%
average yield per step) from commercially available materials.
We have demonstrated the utility of the Paterson-Aldol reaction
for the rapid assembly of the Claisen rearrangement precursor
46, transfer of stereochemical information during the Claisen
rearrangement of alkenyl-substituted cyclic ketene acetals in the
formation of ∆5-oxocenes, as well as documenting the use of
our recently developed carbonate methylenation strategy for the
formation of ketene acetals28 and the use of enzyme-mediated
acetate deprotection in the presence of a medium-ring lactone.
Further examples of the use of the Claisen rearrangement for
the construction of medium-ring natural products and natural
product fragments will be reported in due course.
reported for the isolated natural product by Fenical and Clardy.1
The optical rotation of the synthetic natural product was
approximately 10 times larger than that reported for the natural
product, as noted by Buszek3 and Clardy;4 this is likely due to
a calculational error in the original isolation paper.1
In the original synthetic approaches to octalactin A (1), a
vanadium-directed allylic epoxidation had been used to install
the required epoxide. The proof of the stereochemical outcome
of this reaction had relied on application of the Sharpless
mnemonic80 and comparison of spectral data from the synthetic
and natural samples. We adopted this approach to complete the
synthesis of octalactin A (1). Thus, in a manner similar to that
reported by Buszek,3 we subjected the major allylic alcohol 57a
to hydroxy directed epoxidation80 [VO(acac)2, tert-butyl hy-
droperoxide] to provide 62 as a single diastereomer (Scheme
16); however, we had no method of assigning the relative
stereochemistry at the epoxide stereocenters. The stereochem-
istry of the epoxide moiety was ultimately proven by conversion
of 62 into octalactin A 1 and the use of X-ray crystal structure
analysis, which indicated that the reaction had proceeded
according to the model proposed by Sharpless.80 Attempts to
epoxidize the minor diastereomer 57b from the same face as
the hydroxy group failed.3
Experimental Section
For general experimental techniques, see Supporting Information.
(E)-3-Acetoxy-2-methylpropenal 34. To a suspension of 2746(see
Supporting Information) (4.0 g, 32.3 mmol) in Et2O (30 mL) and
triethylamine (6.5 mL, 47.1 mmol) at -10 °C was added a solution of
acetyl chloride (3.36 mL, 47.1 mmol) in Et2O (10 mL) dropwise via
syringe at such a rate as to maintain the temperature below -5 °C.
The reaction mixture was warmed to ambient temperature over 1.5 h
and was stirred for an additional 6 h during which time a brown
precipitate formed. The reaction was quenched with water (50 mL)
and was extracted with Et2O (3 × 30 mL). The combined organic layers
were dried (MgSO4) and concentrated in vacuo. Purification by flash
chromatography (ethyl acetate/hexane, 9:1) provided the title compound
34 as a clear colorless oil (3.2 g, 25 mmol, 79%); Rf 0.5 (ethyl acetate/
hexane, 2:3); 1H NMR (250 MHz, CDCl3) δ 9.46 (s, 1H), 8.01 (q, J )
1.3 Hz, 1H), 2.29 (s, 3H), 1.77 (d, J ) 1.3 Hz, 3H); 13C NMR (62.5
MHz, CDCl3) δ 166.5, 152.9, 124.4, 20.6, 7.0; IR (CHCl3) 1743 (CO)
cm-1; HRMS (EI) m/z 128.0462 (128.0473 calcd for C6H8O3, M).
(E),3(S),6(R)-1-Acetoxy-7-(tert-butyldimethylsiloxy)-3-hydroxy-
2,6-dimethyl-hept-1-ene-5-one 44. To a solution of (+)-chlorodiiso-
pinocampheylborane (2.29 g, 7.10 mmol) in Et2O (150 mL) cooled to
0 °C was added Et3N (1.04 mL, 7.40 mmol) followed by a solution of
the ketone 43 (1.05 g, 4.80 mmol) in Et2O (15 mL). The resultant white
Oxidation of 6259-61 provided the corresponding ketone 63
which was desilylated on exposure to HF‚pyridine giving 64.81
Finally, enzymatic hydrolysis of the acetate protecting group78
provided octalactin A 1 {[R]D -153 (c 0.14 in CHCl3), lit.3
20
[R]D -156 (c 0.7 in CHCl3)} in quantitative yield as a colorless
crystalline solid mp 154-157 °C (ether) {lit.1 mp 155-157 °C
(CHCl3/EtOAc)}. Octalactin A had spectral data (1H NMR, 13
C
NMR, IR, MS) in complete agreement with the published data.1,3
(80) Sharpless, K. B.; Verhoeven, T. R. Aldrichimica Acta 1979, 12, 63-74.
(81) The 1H NMR of 64 indicated the presence of a slight impurity arising from
coupling of the minor enantiomer of the vinyl iodide present in the scaelmic
mixture of iodides derived from 51. As for 55, this minor component was
removed by preparative thin-layer chromatography after deprotection of
the acetate ester.
(82) Compound 1, (-)-octalactin A, C19H32O6, MW 356.46, was obtained as a
clear colorless crystal, space group P212121, a ) 9.961(4) Å, b ) 20.869(5)
Å, c ) 9.596(4) Å, R ) 90°, â ) 90°, γ ) 90°, V ) 1994.8(12) Å3, Z )
4, Dcalcd ) 1.187 g cm-3, F(000) ) 776. More detailed crystal data can be
viewed in the Supporting Information.
9
2204 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

A Concise Synthesis of the Octalactins
A R T I C L E S
suspension was stirred for 2.5 h. The mixture was then cooled to -78
°C, and a solution of aldehyde 34 (1.27 g, 10.0 mmol) in Et2O (60
mL) was added dropwise via syringe. After the mixture was stirred for
2 h at -78 °C, it was left to warm slowly to -20 °C in a freezer. After
15 h, the mixture was warmed to 0 °C and stirred for 40 min. The
reaction was quenched with a mixture of methanol (150 mL), pH 7
buffer (75 mL), and H2O2 (100 v/v, 19 mL) and stirred vigorously at
0 °C for 10 min followed by warming to ambient temperature and
stirring for 1 h. The mixture was poured into water (150 mL) and was
extracted with EtOAc/hexane, 1:1 (2 L). The aqueous layer was
saturated with solid sodium chloride and was re-extracted with the same
solvent mixture (2 × 250 mL). The combined organic extracts were
washed with brine (200 mL), dried (MgSO4), and concentrated in vacuo.
Purification by flash chromatography (EtOAc/hexane, 10:1 f 5:1)
afforded the title compound 44 (1.44 g, 4.20 mmol, 88%) as a clear
formed from the 1,3-syn-diol. For characterization purposes, a small
sample was purified by flash chromatography (ethyl acetate/hexane,
2:3); Rf 0.3 (ethyl acetate/hexane, 2:3); [R]D20 +24.1 (c 0.61, CH2Cl2);
1H NMR (400 MHz, CDCl3) δ 7.24 (qu, J ) 1.3 Hz, 1H), 4.95 (t, J )
6.6 Hz, 1H), 4.50 (dt, J ) 7.6, 6.6 Hz, 1H), 3.72 (dd, J ) 10.1, 4.8
Hz, 1H), 3.60 (dd, J ) 10.1, 4.8 Hz, 1H), 2.16 (s, 3H), 2.15-1.99 (m,
3H), 1.72 (d, J ) 1.2 Hz, 3H), 0.95 (d, J ) 7.0 Hz, 3H), 0.87 (s, 9H),
0.04 (d, J ) 2.8 Hz, 6H); 13C NMR (100 MHz, CDCl3) δ 167.4, 149.0,
133.8, 117.9, 63.5, 45.8, 38.9, 26.6, 25.9, 20.6, 18.2, 12.3, 10.2, 8.6,
-5.6, -5.8; IR (CDCl3) 1752 (CO) cm-1; HRMS (Electrospray,
Q-TOF) m/z 395.1872 (395.1866 calcd for C18H32O6SiNa, MNa).
(Z),4(R),8(S)-4-Acetoxy-8-[(R)-2-(tert-butyldimethylsiloxy)-1-methyl-
ethyl]-5-methyl-3,4,7,8-tetrahydro-oxocin-2-one 47. To the carbonate
46 (0.98 g, 2.64 mmol) in toluene (500 mL) was added dimethyl
titanocene (603 mg, as a 0.24 M solution in toluene, 2.90 mmol) via
syringe. The mixture was excluded from light and heated at 110 °C
for 30 min. After allowing the mixture to cool, it was concentrated in
vacuo to 30 mL and filtered through a plug of silica (hexane/ethyl
acetate, 1:1). The concentrated residue was subjected to flash chroma-
tography (hexane/ethyl acetate, 19:1) to give the title compound 47
(392 mg, 1.10 mmol, 42%) as a light yellow oil; Rf (hexane/ethyl
20
colorless oil; Rf 0.16 (hexane/ethyl acetate, 5:1); [R]D -62.2 (c 3.7,
1
CHCl3); H NMR (250 MHz, CDCl3) δ 7.20 (s, 1H), 4.55 (m, 1H),
3.78-3.60 (m, 2H), 3.14 (s, 1H), 2.85-2.70 (m, 3H), 2.14 (s, 3H),
1.70 (s, 3H), 1.02 (d, J ) 6.9 Hz, 3H), 0.86 (s, 9H), 0.04 (s, 3H), 0.02
(s, 3H); 13C NMR (62.5 MHz, CDCl3) δ 214.3, 167.9, 132.3, 122.0,
68.8, 65.6, 49.1, 47.8, 25.8, 20.7, 18.2, 12.6, 9.9, -5.6; IR (CHCl3)
3491 (OH), 1758 (CO), 1710 (CO) cm-1; MS (CI, NH3) m/z (rel
intensity) 362 [8, (M + NH4)+]; HRMS (CI, NH3) m/z 362.2363
(362.2363 calcd for C17H36NO5Si, MNH4).
20
1
acetate, 3:1) 0.61; [R]D -17.0 (c 2.4, CHCl3); H NMR (400 MHz,
CDCl3) δ 5.70 (t, J ) 7.4 Hz, 1H), 5.43 (t, J ) 6.2 Hz, 1H), 4.67 (dt,
J ) 9.3, 5.0 Hz, 1H), 3.67 (dd, J ) 9.7, 4.9 Hz, 1H), 3.54 (dd, J )
9.7, 3.3 Hz, 1H), 3.18 (dd, J ) 14.2, 7.4 Hz, 1H), 2.89 (dd, J ) 14.2,
7.4 Hz, 1H), 2.52-2.35 (m, 2H), 2.11 (s, 3H), 1.87 (m, 1H), 1.74 (s,
3H), 0.98 (d, J ) 6.8 Hz, 3H), 0.90 (s, 9H), 0.04 (s, 3H), 0.04 (s, 3H);
13C NMR (100 MHz, CDCl3) δ 172.2, 170.0, 137.6, 121.7, 79.0, 69.9,
64.2, 42.5, 40.3, 31.2, 25.9, 20.9, 19.7, 18.3, 13.1, -5.5, -5.6; IR (film)
1739 (CO), 1693 (CO) cm-1; MS (CI, NH3) m/z (rel intensity) 371
[42, (M + H)+]; HRMS (CI, NH3) m/z 371.2254 (371.2254 calcd for
C19H35O5Si, MH).
(E),3(S),5(S),6(R)-1-Acetoxy-7-(tert-butyldimethylsilyloxy)-2,6-
dimethyl-hept-1-ene-3,5-diol 45. To a solution of tetramethylammo-
nium triacetoxyborohydride (3.76 g, 14.3 mmol) in acetonitrile (40 mL)
and acetic acid (10 mL) at -45 °C was added via cannula a cooled
solution of the aldol 44 (1.40 g, 4.08 mmol) in acetonitrile (10 mL, 5
mL rinse). The reaction mixture was stirred at -35 °C for 24 h and
then quenched with saturated aqueous sodium potassium tartrate (40
mL). The mixture was allowed to warm to ambient temperature, stirred
for 50 min, and then poured onto ice, neutralized with saturated aqueous
sodium hydrogen carbonate (200 mL), and extracted with CH2Cl2 (200
mL). The aqueous layer was saturated with solid sodium chloride and
re-extracted with CH2Cl2 (2 × 200 mL). The combined organic layers
were washed with saturated aqueous sodium hydrogencarbonate (2 ×
150 mL), dried (Na2SO4), and concentrated to give the title compound
45 as a clear colorless oil (1.31 g, 3.80 mmol, 93%). The relatively
unstable diol 45 was used without purification. For characterization
purposes, flash chromatography (ethyl acetate/hexane, 2:3) of the
residue yielded the pure compound; Rf 0.31 (ethyl acetate/hexane, 2:3);
[R]D22 -17.0 (c 0.87, CDCl3); 1H NMR (250 MHz, CDCl3) δ 7.19 (s,
1H), 4.38 (br s, 1H), 3.81-3.74 (m, 2H), 3.82 (dd, J ) 10.1, 7.1 Hz,
1H), 3.81-3.74 (m, 1H), 3.60-3.54 (m, 2H), 2.10 (s, 3 H), 1.83-
1.77 (m, 2H), 1.65 (s, 3H), 0.86 (s, 9H), 0.78 (d, J ) 6.9 Hz, 3H),
0.05 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 168.0, 131.7, 123.9, 74.8,
70.1, 68.8, 39.0, 38.6, 25.8, 20.7, 18.0, 13.3, 10.0, -5.6, -5.7; IR
(CDCl3) 3423 (OH), 1748 (CO) cm-1; HRMS (electrospray, Q-TOF)
m/z 369.2103 (369.2073 calcd for C17H34O5SiNa, MNa).
4(R),5(S),8(S)-4-Acetoxy-8-[(R)-2-(tert-butyldimethylsiloxy)-1-
methyl-ethyl]-5-methyl-oxocan-2-one 49a. To a solution of the lactone
47 (5 mg, 13.4 µmmol) in EtOAc (5 mL) was added 5% rhodium on
alumina (6 mg, 2.6 µmol Rh). This mixture was stirred vigorously under
an atmosphere of H2 at -22 °C for 20 h. Filtration of the mixture
through a plug of silica (eluting with EtOAc) followed by concentration
in vacuo and purification by flash chromatography (hexane/ethyl acetate,
9:1) afforded the reduced lactones 49 as an 8.2:1 mixture of C5 epimers
(4.8 mg, 12.8 µmol, 96%) as a clear colorless oil. The diastereoisomers
were inseparable at this stage and were characterized as a mixture; Rf
1
(hexane/ethyl acetate, 9:1) 0.10; H NMR (400 MHz, CDCl3 major
diastereomer) δ 5.03 (d, J ) 5.7 Hz, 1H), 4.45 (m, 1H), 3.78 (dd, J )
9.8, 3.9 Hz, 1H), 3.44 (dd, J ) 9.8, 3.6 Hz, 1H), 2.95 (dd, J ) 13.4,
6.2 Hz, 1H), 2.78 (dd, J ) 13.4, 1.7 Hz, 1H), 2.10 (s, 3H), 1.95-1.65
(m, 5H), 1.34-1.22 (m, 1H), 1.03 (d, J ) 6.6 Hz, 3H), 1.00 (d, J )
7.0 Hz, 3H), 0.88 (s, 9H), 0.03 (s, 3H), 0.01 (s, 3H); 13C NMR (62.5
MHz, CDCl3 major diastereomer) δ 170.9, 78.0, 73.2, 64.1, 40.0, 36.8,
36.0, 30.5, 25.8, 21.8, 20.9, 18.8, 13.4; IR (film) 1736 (CO) cm-1; MS
(CI, NH3) m/z (rel intensity) 390 [25, (M + NH4)+], 373 [100, (M +
H)+]; HRMS (CI, NH3) m/z 373.2410 (373.2410 calcd for C19H37O5-
Si, MH).
(E),4(S),6(S)-6-(2-Acetoxy-1-methyl-vinyl)-4-[(R)-2-(tert-butyl-
dimethylsilyloxy)-1-methyl-ethyl]-[1,3]dioxan-2-one 46. The diol 45
(118.0 mg, 0.20 mmol) was dissolved in CH2Cl2 (6 mL), pyridine (166
µL, 2.05 mmol), and triethylamine (569 µL, 4.10 mmol). Crushed 4 Å
molecular sieves (250 mg) were added, and the reaction mixture was
cooled to -78 °C. A solution of triphosgene (80.8 mg, 0.273 mmol)
in CH2Cl2 (3 mL) was added dropwise via syringe. The suspension
was stirred for 25 min after which it was warmed to 0 °C and quenched
with saturated aqueous ammonium chloride solution (10 mL). The layers
were separated, and the aqueous phase was extracted with CH2Cl2 (2
× 10 mL). The combined organic layers were washed with brine (10
mL), dried (Na2SO4), and concentrated in vacuo. The mixture was
filtered through a plug of silica to give the title compound 46 (90.0
mg, 0.24 mmol, 71%) which was used without further purification.
The 1H NMR spectrum indicated the presence of an inseparable impurity
(judged from integration to be < 5%), corresponding to the carbonate
4(R),5(S),8(S)-4-Acetoxy-8-[(R)-2-hydroxy-1-methyl-ethyl]-5-methyl-
oxocan-2-one 40. To a solution of the lactones 49 (38 mg, 0.102 mmol)
in THF (2.0 mL) at 0 °C was added crushed glass followed by pyridine
(150 µL) and 70% HF‚pyridine (100 µL). The mixture was warmed to
ambient temperature and was stirred for 20 h. Additional HF‚pyridine
(70 µL) was added, and the mixture was stirred for a further 3 h. The
mixture was then filtered through a short wide plug of silica, eluting
with EtOAc. The eluent was evaporated, and the residue was purified
by flash column chromatography (ethyl acetate/hexane, 2:1) yielding
the separated diastereomeric products as clear oils (total yield 26.5 mg,
0.102 mmol, 100%); Rf (major diastereomer) 0.16 (ethyl acetate/hexane,
2:1); [R]D20 -88 (c 0.57, CHCl3); 1H NMR (400 MHz, CDCl3) δ 5.03
(d, J ) 6.3 Hz, 1H), 4.46-4.39 (m, 1H), 3.77 (dd, J ) 10.5, 4.3 Hz,
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2205

A R T I C L E S
O’Sullivan et al.
1H), 3.57 (dd, J ) 10.5, 4.2 Hz, 1H), 2.94 (dd, J ) 13.5, 6.4 Hz, 1H),
2.81 (dd, J ) 13.5, 2.3 Hz, 1H), 2.09 (s, 3H), 1.94-1.65 (m, 5H),
1.03 (d, J ) 7.0 Hz, 3H), 1.00 (d, J ) 7.1 Hz, 3H); 13C NMR (100
MHz, CDCl3) δ 171.3, 170.8, 78.9, 73.3, 64.1, 39.8, 36.3, 35.5, 31.7,
21.0, 13.4, 1.0; IR (film) 3454 (OH), 1720 (CO) cm-1; MS (CI, NH3)
m/z (rel intensity) 259 [20, (M + H)+]; HRMS (CI, NH3) m/z 259.1545
(259.1546 calcd for C13H23O5 MH). Anal. Calcd for C13H22O5: C, 60.5;
H, 8.6. Found: C, 60.6; H, 8.6.
10.9, -4.2, -4.6; IR (film) 3467 (OH), 1742 (CO) cm-1; MS (CI,
NH3) m/z (rel intensity) 516 [15, (M + NH4)+], 499 [18, (M + H)+];
HRMS (CI, NH3) m/z 499.3445 (499.3455 calcd for C27H51O6Si, MH).
4(R),5(S),8(S)-4-Acetoxy-8-[(E),1(S),6(S)-6-(tert-butyl-dimethyl-
siloxy)-1,3,7-trimethyl-2-oxo-oct-3-enyl]-5-methyl-oxocan-2-one 58.
To a solution of 57b (11.2 mg, 22.5 µmol) in CH2Cl2 (2 mL) was
added the Dess Martin periodinane (38 mg, 90 µmol). The mixture
was stirred at ambient temperature for 3 h. The solvent was exchanged
with EtOAc, and the reagent was removed by passing the mixture
through a short plug of silica, eluting with EtOAc. The crude product
was purified by column chromatography (hexane/ethyl acetate, 3:1) to
furnish the title compound 58 (9.5 mg, 77%) as a colorless oil; Rf 0.25
4(R),5(S),8(S)-4-Acetoxy-5-methyl-8-[(S)-1-methyl-2-oxoethyl]-
oxocan-2-one, 41. To a solution of 40 (12.0 mg, 46 µmol) in CH2Cl2
(3 mL) was added the Dess Martin periodinane (67 mg, 160 µmol).
After 2 h at ambient temperature, wet CH2Cl2 (1 mL) was added and
the mixture was stirred another 30 min. The solvent was exchanged
with EtOAc by careful evaporation, and the insoluble Dess Martin
reagent was filtered off by quickly passing the mixture through a short
wide plug of Florisil (EtOAc/hexane, 3:1) giving the crude title
compound 41 (12 mg, 46 µmol, 100%). The unstable aldehyde 41 was
used in the next step without further purification; Rf (EtOAc/hexane,
7:3) 0.4; 1H NMR (400 MHz, CDCl3) δ 9.72 (d, J ) 1.2 Hz, 1H), 5.06
(d, J ) 6.1 Hz, 1H), 4.70 (ddd, J ) 13.5, 7.5, 3.4 Hz, 1H), 3.00 (dd,
J ) 13.5, 6.6 Hz, 1H), 2.87 (dd, J ) 13.5, 1.9 Hz, 1H), 2.78 (dqn, J
) 1.2, 7.5 Hz, 1H), 2.10 (s, 3H), 1.94-1.52 (m, 5H), 1.15 (d, J ) 7.5
Hz, 3H), 1.00 (d, J ) 6.6 Hz, 3H).
22
1
(hexane/ethyl acetate, 3:1); [R]D -95.8 (c 0.36, CDCl3); H NMR
(400 MHz, CDCl3) δ 6.76 (t, J ) 6.6 Hz, 1H), 5.07 (d, J ) 5.7 Hz,
1H), 4.79 (ddd, J ) 12.0, 9.1, 2.9 Hz, 1H), 3.63 (td, J ) 6.0, 5.0 Hz,
1H), 3.48 (m, 1H), 2.98 (dd, J ) 13.6, 2.2 Hz, 1H), 2.92 (dd, J )
13.7, 6.0 Hz, 1H), 2.45-2.31 (m, 2H), 2.09 (s, 3H), 1.94-1.64 (m,
5H), 1.77 (s, 3H), 1.05 (d, J ) 2.1 Hz, 3H), 1.03 (d, J ) 1.8 Hz, 3H),
0.90 (m, 12H), 0.89 (s, 3H), 0.06 (s, 6H); 13C NMR (100 MHz, CDCl3)
δ 203.0, 170.7, 170.5, 141.2, 136.8, 78.6, 77.2, 75.8, 73.1, 44.5, 36.4,
35.5, 33.6, 33.3, 31.9, 25.8, 24.0, 21.7, 20.9, 18.1, 17.9, 15.1, 11.9,
-4.3, -4.5; IR (film) 1732 (CO), 1665 (CO) cm-1; MS (CI, NH3) m/z
(rel intensity) 514 [4, (M + NH4)+], 497 [6, (M + H)+]; HRMS (CI,
NH3) m/z 497.3297 (497.3298 calcd for C27H49O6Si, MH).
4(R),5(S),8(S)-4-Acetoxy-8-[(E),1(R),2(R),6(S)-6-(tert-butyldimeth-
ylsilyloxy)-2-hydroxy-1,3,7-trimethyl-oct-3-enyl]-5-methyl-oxocan-
2-one 57a and 4(R),5(S),8(S)-4-Acetoxy-8-[(E),1(R),2(S),6(S)-6-(tert-
butyldimethylsilyloxy)-2-hydroxy-1,3,7-trimethyl-oct-3-enyl]-5-methyl-
oxocan-2-one 57b. The aldehyde 41 (8.0 mg, 0.031 mmol) was
azeotropically evaporated twice with toluene (2 × 2 mL) and dried
under high vacuum in a Schlenk flask prior to use. DMSO (1 mL) was
added via syringe followed by the vinyl iodide 56a (44.9 mg, 0.122
mmol) as a solution in DMSO via cannula (0.5 mL, 0.5 mL rinse). A
mixture of 1% NiCl2 in CrCl2 (38.1 mg, 0.31 mmol) was preweighed
in a glovebox and added to the reaction mixture under a stream of
argon. The dark green solution was stirred at ambient temperature with
the exclusion of light for 24 h, after which time the reaction mixture
was quenched with saturated aqueous ammonium chloride solution (8
mL) and extracted with EtOAc (5 × 10 mL). The combined organic
layers were washed with brine (10 mL) and dried (Na2SO4). Concentra-
tion in vacuo was followed by flash chromatography (ethyl acetate/
hexane, 1:4) to provide the diastereomeric allylic alcohols 57a (9.0
mg, 0.018 mmol, 58%) and 57b (5.1 mg, 0.010 mmol, 33%) as clear
4(R),5(S),8(S)-4-Acetoxy-8-[(E),1(S),6(S)-6-hydroxy-1,3,7-trimethyl-
2-oxo-oct-3-enyl]-5-methyl-oxocan-2-one 55. Crushed glass (150 mg)
was added to a solution of 58 (6.3 mg, 12.6 µmol) in THF (2.0 mL) at
0 °C followed by anhydrous pyridine (300 µL) and 70% HF‚pyridine
(150 µL). The cooling bath was removed, and the mixture was stirred
for 48 h at ambient temperature with additional HF‚pyridine (100 µL)
being added after 24 h. The mixture was filtered through a short wide
plug of silica (EtOAc). Evaporation was followed by flash chroma-
tography (hexane/ethyl acetate, 1:1) to give the title compound 55 (3.8
mg, 9.9 µmol, 80%) as a clear colorless oil; Rf 0.21 (hexane/ethyl
acetate, 1:1); [R]D -126.6 (c 0.29, CH2Cl2); H NMR (400 MHz,
CDCl3) δ 6.87 (t, J ) 7.1 Hz, 1H), 5.05 (brs, 1H), 4.72 (ddd, J 12.2,
9.3, 3.0 Hz, 1H), 3.59-3.50 (m, 2H), 2.95-2.89 (m, 2H), 2.48-2.32
(m, 2H), 2.09 (s, 3H), 1.90-1.67 (m, 7H), 1.35-1.27 (m, 3H), 1.05
(d, J ) 5.5 Hz, 3H), 1.04 (d, J ) 5.2 Hz, 3H), 0.97 (d, J ) 3.6 Hz,
3H), 0.95 (d, J ) 3.7 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ 203.3,
170.7, 141.5, 137.5, 79.1, 75.9, 73.0, 44.4, 36.4, 35.5, 34.0, 33.9, 32.0,
30.9, 29.7, 24.0, 21.7, 21.0, 18.7, 17.4, 15.0, 11.8; IR (film) 3468 (OH),
1731 (CO), 1661 (CO) cm-1; MS (FAB) m/z (rel intensity) 383 [50,
(M + H)+]; HRMS (CI, NH3) m/z 383.2435 (383.2434 calcd for
C21H35O6, MH).
22
1
20
colorless oils. Data for 57a: Rf 0.3 (hexane/ethyl acetate, 3:1); [R]D
-27.6 (c 0.25, CHCl3); 1H NMR (400 MHz, CDCl3) δ 5.47 (t, J ) 6.9
Hz, 1H), 5.05 (d, J ) 5.4 Hz, 1H), 4.48-4.44 (m, 1H), 4.28 (s, 1H),
3.49 (m, 1H), 2.95 (dd, J ) 13.5, 6.2 Hz, 1H), 2.87 (d, J ) 13.5 Hz,
1H), 2.20 (m, 2H), 2.11 (s, 3H), 1.95-1.86 (m, 2H), 1.84-1.62 (m,
5H), 1.38 (s, 1H), 1.32-1.22 (m, 3H), 1.04 (d, J ) 7.0 Hz, 3H), 0.89
(s, 9H), 0.87 (d, J ) 7.0 Hz, 3H), 0.85 (d, J ) 7.0 Hz, 3H), 0.76 (d,
J ) 7.0 Hz, 3H), 0.04 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 171.3,
170.8, 136.8, 120.8, 79.2, 77.0, 73.9, 73.3, 40.0, 36.5, 35.6, 32.8, 32.5,
32.3, 25.9, 24.5, 21.3, 21.0, 18.9, 18.1, 17.1, 14.1, 8.8, -4.2, -4.6; IR
(film) 3477 (OH), 1731 (CO) cm-1; MS (CI, NH3) m/z (rel intensity)
516 [40, (M + NH4)+], 499 [8, (M + H)+]; HRMS (CI, NH3) m/z
516.3712 (516.3720 calcd for C27H54O6NSi, MNH4). Anal. Calcd for
C27H50O6Si: C, 65.0; H, 10.1. Found: C, 64.9; H, 10.0.
4(R),5(S),8(S)-4-Hydroxy-8-[(E),1(S),6(S)-6-hydroxy-1,3,7-tri-
methyl-2-oxo-oct-3-enyl]-5-methyl-oxocan-2-one, (-)-Octalactin B,
2. To a solution of 55 (2.8 mg, 7.3 µmol) in 10% DMF/pH 7 buffer
(2.0 mL) was added type VII lipase from Candida cylindracea (Fluka;
35.3 mg). The flask was sealed, and the mixture was excluded from
light and stirred at ambient temperature for a period of 6 d, adding 13
mg, 9.0 mg, 13 mg, 5 mg of enzyme on days 2, 3, 4, and 5, respectively.
Water was added to the mixture, and the aqueous layer was extracted
with Et2O (5 × 10 mL). The combined organic layers were washed
with brine (20 mL), dried (MgSO4), and filtered. The eluent was passed
through a short wide plug of silica (Et2O). Concentration of the eluent
was followed by flash chromatography (ethyl acetate/hexane, 3:1) and
PTLC to give the title natural product (-)-octalactin B, 2 (2.4 mg, 7.0
µmol, 96%), as an oil; Rf 0.15 (hexane/ethyl acetate, 1:2); [R]D24 -123
(c 0.04, CDCl3) {lit.3 [R]D -126}; 1H NMR (400 MHz, CDCl3) δ 6.86
(t, J ) 7.0 Hz, 1H), 4.75 (t, J ) 10.7 Hz, 1H), 4.03 (brs, 1H), 3.56-
3.51 (m, 2H), 3.05 (dd, J ) 13.3, 1.5 Hz, 1H), 2.72 (dd, J ) 13.3, 6.5
Hz, 1H), 2.45 (m, 1H), 2.35 (m, 1H), 1.96 (d, J ) 4.8 Hz, 1H), 1.81-
1.74 (m, 2H), 1.78 (s, 3H), 1.73-1.68 (m, 3H), 1.24-1.18 (m, 1H),
1.14 (d, J ) 7.1 Hz, 3H), 1.05 (d, J ) 7.2 Hz, 3H), 0.98 (d, J ) 6.6
Hz, 3H), 0.96 (d, J ) 6.8 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ
22
Data for 57b: Rf 0.13 (ethyl acetate/hexane, 1:3); [R]D -49.2 (c
1
0.6, CH2Cl2); H NMR (400 MHz, CDCl3) δ 5.42 (dd, J ) 7.0, 0.5
Hz, 1H), 5.04 (d, J ) 6.0 Hz, 1H), 4.80 (dt, J ) 12.3, 3.9 Hz, 1H),
3.76 (dd, J ) 8.7, 2.1 Hz, 1H), 3.48 (dt, J ) 4.4, 5.9 Hz, 1H), 2.98
(dd, J ) 13.5, 6.2 Hz, 1H), 2.85 (dd, J ) 13.3, 1.9 Hz, 1H), 2.21-
2.15 (m, 3H), 2.10 (s, 3H), 1.68-1.63 (m, 1H), 1.61 (s, 3H), 1.03 (d,
J ) 6.8 Hz, 3H), 0.89 (s, 9H), 0.87 (d, J ) 7.0 Hz, 3H), 0.85 (d, J )
6.8 Hz, 3H), 0.80 (d, J ) 7.0 Hz, 3H), 0.03 (s, 6H); 13C NMR (100
MHz, CDCl3) δ 171.1, 170.8, 136.4, 126.4, 80.4, 77.9, 76.5, 73.3, 39.8,
36.3, 35.9, 32.7, 32.2, 28.3, 25.9, 24.2, 21.5, 21.0, 18.7, 17.2, 11.1,
9
2206 J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004

A Concise Synthesis of the Octalactins
A R T I C L E S
203.3, 172.5, 141.4, 137.5, 79.4, 75.9, 71.4, 44.2, 39.2, 38.0, 34.1, 33.9,
32.2, 22.6, 22.0, 18.7, 17.4, 15.0, 11.7; IR (film) 3444 (OH), 1714
(CO), 1660 (CO) cm-1; HRMS (FAB) m/z 364.2225 (364.2226 calcd
for C19H33O5Na, MH + Na).
5.7 Hz, 1H), 4.60 (m, 1H), 3.57 (t, J ) 6.2 Hz, 1H), 3.47 (m, 1H),
3.02-2.83 (m, 4H), 2.07 (s, 3H), 1.90-1.58 (m, 6H), 1.43 (s, 3H),
1.32-1.22 (m, 2H), 1.04 (d, J ) 7.0 Hz, 3H), 1.00 (d, J ) 7.0 Hz,
3H), 0.93 (d, J ) 6.8 Hz, 3H), 0.92 (d, J ) 6.8 Hz, 3H); 13C NMR
(100 MHz, CDCl3) δ 212.6, 170.8, 170.6, 79.2, 77.2, 74.3, 72.8, 62.5,
58.8, 42.4, 36.5, 35.5, 34.0, 32.1, 31.8, 23.8, 21.9, 20.9, 18.3, 17.6,
13.5, 12.5; IR (CHCl3) 3470 (OH), 1734 (CO), 1709 (CO) cm-1; MS
(CI, NH3) m/z (rel intensity) 416 [12, (M + NH4)+], 399 [18, (M +
H)+]; HRMS m/z 399.2386 (399.2383 calcd for C21H35O7, MH).
4(R),5(S),8(S)-4-Hydroxy-8-{2-[3-((S)-2-hydroxy-3-methyl-butyl)-
2(S),3(R)-2-methyl-oxiranyl]-1(S)-methyl-2-oxo-ethyl}-5-methyl-
oxocan-2-one, (-)-Octalactin A, 1. To a vigorously stirred solution
of the acetate ester 64 (2.0 mg, 50 µmol) in 10% DMF/pH 7 buffer (1
mL) was added type VII lipase from Candida cylindracea (Fluka, 20
mg). This mixture was stirred with the exclusion of light for 7 days,
additional enzyme (20 mg) being added on days 2 and 4. The suspension
was extracted with ether (3 × 5 mL), and the combined organic layers
were filtered through a short wide plug of silica (Et2O), washed with
brine (5 mL), and dried (Na2SO4). The solvent was removed in vacuo,
and purification by flash chromatography (ethyl acetate/hexane, 1:1)
gave the synthetic natural product (-)-octalactin A, 1 (2.0 mg, 50 µmol,
100%), as a colorless solid which was crystallized from ether; mp 154-
4(R),5(S),8(S)-4-Acetoxy-8-{2-[3-((S)-2-(tert-butyldimethylsilyloxy)-
3-methyl-butyl)-2(R),3(R)-2-methyl-oxiranyl]-1(R),2(R)-2-hydroxy-
1-methyl-ethyl}-5-methyl-oxocan-2-one 62. To a stirred solution of
57a (24.0 mg, 48 µmol) in benzene (3 mL) was added t-BuOOH (27
µL, 144 µmol) and VO(acac)2 (3 mg, 11.3 µmol). The mixture was
stirred at ambient temperature for 1 h. Concentration in vacuo was
followed by flash chromatography (hexane/ethyl acetate, 3:1), which
gave the title compound 62 (23.8 mg, 46 µmol, 98%) as a clear colorless
23
oil; Rf 0.08 (hexane/ethyl acetate, 4:1); [R]D -14.3 (c 0.91, CHCl3);
1H NMR (400 MHz, CDCl3) δ 5.04 (d, J ) 5.5 Hz, 1H), 4.47 (m, 1H),
4.02 (s, 1H), 3.60 (q, J ) 5.2 Hz, 1H), 3.29 (dd, J ) 7.8, 3.1 Hz, 1H),
2.94 (dd, J ) 13.7, 5.8 Hz, 1H), 2.88 (dd, J ) 13.7, 2.2 Hz, 1H), 2.10
(s, 3H), 1.97-1.76 (m, 6H), 1.69-1.59 (m, 2H), 1.33-1.28 (m, 1H),
1.28-1.25 (m, 1H), 1.21 (s, 3H), 1.07 (d, J ) 7.0 Hz, 3H), 0.91 (d, J
) 2.2 Hz, 3H), 0.89 (s, 12H), 0.83 (d, J ) 7.0 Hz, 3H), 0.06 (s, 3H),
0.04 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 171.2, 170.8, 78.5, 75.3,
73.2, 70.2, 60.8, 55.5, 39.3 36.4, 35.5, 32.8, 32.8, 32.5, 25.9, 24.4,
21.6, 21.0, 18.6, 18.2, 18.1, 14.9, 9.5, -4.3, -4.6; IR (CHCl3) 3476
(OH), 1738 (CO) cm-1; MS (CI, NH3) m/z (rel intensity) 532 [20, (M
+ NH4)+], 515 [9, (M + H)+]; HRMS m/z 515.3404 (515.3404 calcd
for C27H51O7Si, MH).
157 °C (ether){lit.1 155-157 °C (CHCl3/EtOAc}; [R]D -153.1 (c
20
0.14, CHCl3), {lit.3 [R]D -156 (c 0.7 in CHCl3)}; Rf 0.09 (ethyl acetate/
1
hexane, 1:1); H NMR (500 MHz, CDCl3) δ 4.60 (t, J ) 10.3 Hz,
4(R),5(S),8(S)-4-Acetoxy-8-{2-[3-((S)-2-(tert-butyldimethylsilyloxy)-
3-methyl-butyl)-2(S),3(R)-2-methyl-oxiranyl]-1(S)-methyl-2-oxo-
ethyl}-5-methyl-oxocan-2-one 63. To a solution of the epoxide 62
(12.0 mg, 23 µmol) in CH2Cl2 (2 mL) was added the Dess Martin
periodinane (51.0 mg, 120 µmol) under a stream of argon. The
suspension was stirred for 2 h at 25 °C. The solvent was exchanged
for EtOAc on a rotary evaporator while ensuring the mixture did not
go dry. The resultant white suspension was filtered through a short
wide plug of silica to provide the title compound 63 (11.9 mg, 23 µmol,
1H), 4.03 (brs, 1H), 3.55 (t, J ) 6.2 Hz, 1H), 3.58-3.48 (m, 1H),
3.00-2.93 (m, 2H), 2.79 (br d, J ) 3.1 Hz, 0.5H, OH), 2.72 (dd, J )
13.4, 6.3 Hz, 1H), 1.89 (br d, J ) 3.1 Hz, 0.5H, OH), 1.79-1.57 (m,
6H), 1.44 (s, 3H), 1.23-1.18 (m, 1H), 1.13 (d, J ) 7.1 Hz, 3H), 1.00
(d, J ) 7.0 Hz, 3H), 0.94 (d, J ) 6.8 Hz, 3H), 0.92 (d, J ) 6.8 Hz,
3H); 13C NMR (100 MHz, CDCl3) δ 212.5, 172.5, 79.3, 74.5, 71.3,
62.4, 58.9, 42.4, 39.2, 38.0, 34.0, 32.2, 32.0, 22.4, 22.1, 18.5, 17.6,
13.5, 12.6; IR (CHCl3) 3458 (OH), 1708 (CO), 1462 cm-1; HRMS
(Q-TOF) m/z 379.2122 (379.2097 calcd for C19H32O6Na, MNa).
22
100%) as a clear colorless oil; Rf 0.23 (hexane/ethyl acetate, 4:1); [R]D
-124.0 (c 0.59, CHCl3); 1H NMR (400 MHz, CDCl3) δ 5.05 (brs, 1H),
4.67 (ddd, J ) 12.4, 9.4, 3.0 Hz, 1H), 3.68 (dd, J ) 10.4, 5.9 Hz, 1H),
3.16 (dd, J ) 7.1, 4.0 Hz, 1H), 2.96-2.83 (m, 3H), 2.08 (s, 3H), 1.92-
1.68 (m, 5H), 1.65-1.55 (m, 2H), 1.43 (s, 3H), 1.32-1.25 (m, 1H),
1.03 (d, J ) 6.9 Hz, 3H), 1.01 (d, J ) 7.2 Hz, 3H), 0.92-0.88 (m,
12H), 0.86 (d, J ) 6.8 Hz, 3H), 0.06 (s, 3H), 0.05 (s, 3H); 13C NMR
(100 MHz, CDCl3) δ 211.8, 170.7, 170.1, 77.3, 74.9, 73.0, 62.4, 57.9,
43.5, 36.4, 35.5, 33.1, 32.3, 31.7, 25.9, 25.9, 23.9, 21.6, 20.9, 18.2,
18.1, 17.4, 13.4, 13.3, -4.4, -4.6; IR (CHCl3) 1738 (CO), 1708 (CO)
cm-1; MS (CI, NH3) m/z (rel intensity) 530 [4, (M + NH4)+]; HRMS
m/z 530.3512 (530.3513 calcd for C27H52NO7Si, MNH4).
Acknowledgment. We thank the Engineering and Physical
Sciences Research Council (UK) for financial support and
provision of the Swansea Mass Spectrometry service, and the
Royal Society (University Research Fellowship to J.W.B.), the
Deutsche Forschungsgemeinschaft (W.B.), Cambridge European
Trust and the Robert Gardiner Memorial Fund (P.T.O.’S.),
GlaxoSmithKline and Universities UK (ORS and studentship
to M.A.M.F.), and Merck Sharp and Dohme (CASE studentship
to J.R.H.) for financial support. Mr. R. A. Robinson provided
helpful initial suggestions.
4(R),5(S),8(S)-4-Acetoxy-8-{2-[3-((S)-2-hydroxy-3-methyl-butyl)-
2(S),3(R)-2-methyl-oxiranyl]-1(S)-methyl-2-oxo-ethyl}-5-methyl-
oxocan-2-one 64. To a solution of the silyl ether 63 (5.0 mg, 98 µmol)
in THF (2 mL) stirred at 0 °C was added crushed glass (150 mg)
followed by pyridine (300 µL) and 70% HF‚pyridine (150 µL). The
mixture was warmed to ambient temperature and stirred for 24 h. The
resulting white suspension was diluted with EtOAc (3 mL) and filtered
through a small wide plug of silica (EtOAc). The eluent was
concentrated in vacuo and purified by flash chromatography (hexane/
ethyl acetate, 2:1) to give the title compound 64 (3.5 mg, 88 µmol,
Supporting Information Available: Experimental procedures
for the synthesis of 9, 12-15, the diol derived from 15a, 17,
25, 27, 28-39a, 42, 43, 48, 50-54, and 56; 1H NMR analysis
of 29, 35, 51, and 53a; CIF files and ORTEP plots of 48 and 1
(the data has been deposited with the Cambridge Crystal-
1
lographic Database); H and 13C NMR spectra of synthetic
octalactins A (1) and B (2) and tabulated comparisons with the
natural octalactins. This material is available free of charge via
the Internet at http://pubs.acs.org.
22
90%) as a clear colorless oil; Rf 0.19 (hexane/ethyl acetate, 2:1); [R]D
1
-126.3 (c 0.24, CHCl3); H NMR (400 MHz, CDCl3) δ 5.03 (d, J )
JA038353W
9
J. AM. CHEM. SOC. VOL. 126, NO. 7, 2004 2207