Total synthesis of (+)-amphidinolide A. Structure elucidation and completion of the synthesis

Article abstract of DOI:10.1021/ja053365y

The structure elucidation of (+)-amphidinolide A, a cytotoxic macrolide, has been accomplished by employing a combination of NMR chemical shift analysis and total synthesis. The 20-membered ring of amphidinolide A was formed by a ruthenium-catalyzed alken

Full text of DOI:10.1021/ja053365y

Published on Web 09/07/2005
Total Synthesis of (+)-Amphidinolide A. Structure Elucidation
and Completion of the Synthesis
Barry M. Trost,* Paul E. Harrington, John D. Chisholm, and Stephen T. Wrobleski
Contribution from the Department of Chemistry, Stanford UniVersity,
Stanford, California 94305-5080
Received May 23, 2005; E-mail: bmtrost@stanford.edu
Abstract: The structure elucidation of (+)-amphidinolide A, a cytotoxic macrolide, has been accomplished
by employing a combination of NMR chemical shift analysis and total synthesis. The 20-membered ring of
amphidinolide A was formed by a ruthenium-catalyzed alkene-alkyne coupling to forge the C15-C16
bond. Using the reported structure 1 as a starting point, a number of diastereomers of amphidinolide A
were prepared. Deviations of the chemical shift of key protons in each isomer relative to the natural material
were used as a guide to determine the locations of the errors in the relative stereochemistry. The
spectroscopic data for the synthetic and natural material are in excellent agreement.
Introduction
give the fully conjugated 1,3-diene. Attempts to suppress this
isomerization by using HOBt as an additive2 or by converting
In the previous paper in this issue,1 three strategies were
outlined for the synthesis of 1 (Figure 1). Either 2, 3, or 4 could
potentially serve as a precursor to 1. Once a reliable route to 1
is developed, the door is open to allow access to diastereomers
of 1, with the ultimate goal being the determination of the
structure of amphidinolide A and confirmation by synthesis.
However, before this can begin, one or more of the routes shown
in Figure 1 must be established.
18 to the corresponding acid chloride were also unsuccessful.
The use of tributylphosphine as a nonbasic acylation catalyst3
in place of DMAP gave similar results. To circumvent this
problem, alcohol 13 was esterified with 2-butynoic acid to yield
14 in 57% yield. Conjugate addition of allyl copper at -78 °C
to 14 afforded the desired macrocyclization substrate 15 in 55%
yield.
Results and Discussion
Macrocyclization at C6-C7. The first approach to the
proposed structure of amphidinolide A (1) was via an intermo-
lecular ruthenium-catalyzed alkene-alkyne coupling to form
the C15-C16 bond followed by the intramolecular coupling to
form the C6-C7 bond. As shown in Scheme 1, [CpRu(MeCN)3]-
PF6-catalyzed alkene-alkyne coupling of 10 with (trimethyl-
silyl)alkyne 11 effected coupling. In DMF, the reaction pro-
ceeded slowly, giving only a 34% yield [96% based upon
recovered starting material (brsm)] but in a gratifying 9:1
branched (12) to linear (17) regioselectivity with respect to
the alkyne. In acetone, the conversion improved (76%
yield, 91% brsm), but the regioselectivity decreased to 4:1.
Adding 5 vol % DMF to acetone restored the regioselectivity
but at some expense of conversion (48-66% yield, ∼96%
brsm).
The initial attempts to form the 20-membered ring employed
[CpRu(MeCN)3]PF6. However, exposure of a 0.05 M solution
of 15 in DMF at room temperature to 10 mol % [CpRu(MeCN)3]-
PF6 afforded trace amounts of unidentified byproducts along
with significant amounts of recovered 15. Although the byprod-
Diene 12 was converted to 13 with barium hydroxide in
methanol. Under these conditions, the alkyne in 12 was also
deprotected. The direct esterification of 13 with 18 proved to
be problematic. Esterification with DCC/DMAP resulted in
complete isomerization of the base-sensitive 1,4-diene in 18 to
1
ucts were not rigorously characterized, their H NMR spectra
(2) Trost, B. M.; Belletire, J. L.; Godleski, S.; McDougal, P. G.; Balkovec, J.
M. J. Org. Chem. 1986, 51, 2370.
(3) Vedejs, E.; Bennett, N. S.; Conn, L. M.; Diver, S. T.; Gingras, M.; Lin, S.;
Oliver, P. A.; Peterson, M. J. J. Org. Chem. 1993, 58, 7286.
(1) Trost, B. M.; Wrobleski, S. T.; Chisholm, J. D.; Harrington, P. E.; Jung,
M. J. Am. Chem. Soc. 2005, 127, 13589.
9
13598
J. AM. CHEM. SOC. 2005, 127, 13598-13610
10.1021/ja053365y CCC: $30.25 © 2005 American Chemical Society

Completion of the Synthesis of (
+
)-Amphidinolide A
A R T I C L E S
Figure 1. Retrosynthetic analysis of amphidinolide A.
were not consistent with 16. Further examination suggested that
the byproducts may have been derived from linear isomer 19
rather than the desired branched isomer 16. Variation of the
catalyst, temperature, and solvent yielded similar results. Forma-
tion of the linear isomer over the desired branched isomer was
not completely unexpected at this point since previous results
from our laboratories have shown that propargylic ethers tend
to favor the linear products in intermolecular addition reactions
under similar conditions.4a We were hoping that reaction through
a conformationally restricted macrocyclization cycloisomeriz-
ation might offset the propensity to form the linear products,
thereby giving the desired branched isomer.
In another attempt at macrocyclization at C6-C7, the
cyclopentylidene protecting groups in 15 were removed by
treatment with catalytic CSA in aqueous methanol (Scheme
2). Exposure of 20 to [CpRu(MeCN)3]PF6 in anhydrous
DMF at room temperature gave none of the desired macro-
cycle 21, and only unreacted 20 was recovered from the
reaction.
[CpRu(MeCN)3]PF6 in DMF provided branched isomer 25 with
excellent selectivity over linear isomer 27 (Table 1, entry 1).
However, the yield of 25 was low possibly due to complexation
of the diene product with the active catalyst. To circumvent
this issue, terminal alkyne 23 was examined. Alkyne 23 should
be more reactive than 22, but may favor the undesired linear
isomer 28. With [CpRu(MeCN)3]PF6 as the catalyst, 23 coupled
with 24 to provide an acceptable combined yield of 26 and 28;
however, the product ratio favored linear isomer 28 in either
acetone or DMF (entries 3 and 4). A literature report suggested
that the [Cp*Ru(MeCN)3]PF65 catalyst may improve turnovers
in reactions where dienes are the products due to the reduced
likelihood of the more sterically demanding catalyst to undergo
product inhibition.6 In the event, an excellent combined yield
of 26 and 28 was isolated employing 10 mol % [Cp*Ru-
(MeCN)3]PF6 in DMF (entry 5). In less polar solvents, the
product ratio favored branched isomer 26 (entries 6-8), thus
overriding the preference usually observed for propargylic
ethers.
Macrolactonization. With the inability to form the macro-
cycle at the C6-C7 bond, a macrolactonization approach was
undertaken. Since this approach required an intermolecular
addition reaction to form the C6-C7 bond, the model substrates
22-24 were utilized to identify the optimal conditions for this
key coupling step (Table 1). This optimization was especially
relevant after the failure to convert 15 to 16.
Coupling of the TMS-protected alkyne 22 was initially
investigated since prior work from our laboratories have shown
that TMS-protected propargylic ethers tend to favor branched
products over linear products in intermolecular additions.4b As
predicted, exposure of a solution of 22 and 24 to 10 mol %
Application of the optimal conditions shown in entry 6 to 29
yielded branched isomer 30 in 53% yield as a 3:1 mixture of
branched and linear isomers along with 22% recovered alkyne
29 and 63% recovered alkene 24 (Scheme 3). Hydrolysis of
the ethyl ester and acetate in 30 was the next challenge. The
sensitivity of the C1-C5 diene to basic conditions has already
been described and was a concern here. In the event, exposure
of 30 to NaOH in methanol resulted in considerable iso-
merization of the C1-C5 diene. However, after an extensive
screen of conditions, a 47% yield of 31 was realized by
treatment of 30 with Ba(OH)2 in dioxane/water.
(4) (a) Trost, B. M.; Indolese, A. F.; Mueller, T. J. J.; Treptow, B. J. Am.
Chem. Soc. 1995, 117, 615. (b) Trost, B. M.; Machacek, M.; Schnaderbeck,
M. J. Org. Lett. 2000, 2, 1761.
(5) Steinmutz, B.; Schenk, W. A. Organometallics 1999, 18, 943.
(6) Yamamoto, Y.; Kitahara, H.; Ogawa, R.; Itoh, K. J. Org. Chem. 1998, 63,
9610.
9
J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005 13599

A R T I C L E S
Trost et al.
Scheme 1. Route to Macrocyclization Substrate 15
Scheme 2. Attempted Macrocyclization of Free Tetrol 20
Cyclization of 31 under the Yamaguchi conditions7 also
proved problematic due to the sensitive 1,4-diene in 31. Under
the standard Yamaguchi conditions, extensive isomerization of
the unsaturated C1-C5 portion occurred, providing a number
of byproducts in addition to a 24% yield of 16. Therefore, an
improved method for formation of the lactone, without iso-
merization of the base-sensitive diene, was sought. A number
of lactonization methods are available; however, each requires
either high temperatures and/or the presence of a base. In an
effort to circumvent these limitations, a new macrolactonization
protocol was developed based upon the work of Kita.8 Under
these conditions, a carboxylic acid is converted to the corre-
(7) Inanaga, J.; Hirata, H.; Saeki, H.; Katsuki, T.; Yamaguchi, M. Bull. Chem.
Soc. Jpn. 1979, 52, 1989.
9
13600 J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005

Completion of the Synthesis of (
+
)-Amphidinolide A
A R T I C L E S
Table 1. Optimization of the Alkene-Alkyne Coupling of 22 or 23 with 24
yield of
25:27 or
26:28
25
+
27 or
entry
R
catalyst (10 mol %)
solvent
26
+ 28 (%)
1
2
3
4
5
6
7
8
TMS
TMS
H
H
H
H
H
H
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
DMF
acetone
DMF
acetone
DMF
acetone
DCE
>20:1
>20:1
1:1.7
1:1.4
1:1.2
2.3:1
2.5:1
2.8:1
24
20
48
40
84
72
55
33
THF
Scheme 3. Route to Yamaguchi Macrolactonization Substrate 31
sponding ethoxyvinyl ester with ethyl ethynyl ether and catalytic
[RuCl2(p-cymene)]2.9 The intermediate ethoxyvinyl esters par-
ticipate in intermolecular couplings with amines and alcohols
to form amides and esters, respectively, under mildly acidic
conditions. Application of these conditions to 31 provided a
40% yield of macrolactone 16 as shown in Scheme 4. More
significantly, 16 was contaminated with only trace amounts of
byproducts resulting from diene isomerization.
Scheme 4. Macrolactonization of 31
Macrocyclization at C15-C16. To complete the synthesis
of the proposed structure of amphidinolide A from 16, instal-
lation of the C21-C25 epoxide side chain remained. Although
this appeared possible starting from 16, a more convergent route
was desired whereby incorporation of this side chain would be
performed prior to macrocyclization. Keeping this in mind, we
(8) Kita, Y.; Maeda, H.; Omori, K.; Okuno, T.; Tamura, Y. Synlett 1993, 273.
(9) Trost, B. M.; Chisholm, J. D. Org. Lett. 2002, 4, 3743.
9
J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005 13601

A R T I C L E S
Trost et al.
Scheme 5. Completion of the Synthesis of 1
also realized that one of our original endgame strategies (shown
in Figure 1) utilizing a macrocyclization of 3 at C15-C16 had
been unexplored up to this point and appeared tantalizing. One
initial concern with this approach was the envisioned coupling
of alkyne 6 with alkene 5 due to the possibility of undesirable
self-condensation of 6, which also contained a reactive terminal
alkene. To address this concern, it was anticipated that an excess
of the C1-C6 partner 24 would be required. The alkene partner
ethyl ester 24 was replaced with Fmoc ester 5, which was
anticipated to cleave to the corresponding acid under more mild
conditions, thereby avoiding a potentially low yielding hydrol-
ysis step similar to the hydrolysis of 30.10 Application of the
conditions used for the conversion of 29 to 30swith an increase
in the equivalents of alkene partner 5sgave 32 along with its
linear isomer as a minor product in 46% yield (76% brsm,
Scheme 5). Deprotection of the Fmoc ester with piperidine in
CH2Cl2 provided acid 33 in 90% yield with no detectable alkene
reported by Kobayashi.13,14 As shown in Table 2, there are
significant deviations in the 1H NMR between 1 and the natural
material.
At this juncture, we were intrigued with the opportunity to
elucidate the correct structure of amphidinolide A. There are
numerous examples in the literature where the total synthesis
of a complex natural product reveals the structure was deter-
mined incorrectly.15 Unfortunately, only an extremely small
sample of natural amphidinolide A remains;16 thus, additional
NMR experiments are not possible. Therefore, total synthesis
represents the only practical method by which the correct
structure can be determined unambiguously.
Since the relative stereochemistry of 1 was assigned with
NOE data measured on a macrolide possessing considerable
flexibility, an error in the relative stereochemistry and not gross
structure seemed most likely. This conclusion is also supported
by a comparison of the spectral data for the natural material
and the synthetic compound 1. The differences in chemical shifts
and coupling constants are not as large as would be expected
for an error in connectivity.
1
isomerization by H NMR. Esterification of 33 with 7 under
the Kita conditions gave ester 3 in 66% yield. Despite the failure
of 15 to undergo macrocyclization, the reaction of 3 proceeded
smoothly with 10 mol % [CpRu(MeCN)3]PF6 in DCE, to
provide macrolide 34 in 58% yield. Given the high degree of
unsaturation present in this substrate, the chemoselectivity of
this process is remarkable. The reason for the stark difference
in reactivity between 3 and 15 is not clear; however, differences
in conformational mobility may be a factor. Deprotection of
34 under acidic conditions provided 1, which was identical by
1H NMR to the material prepared by Maleczka11 and Pat-
tenden,12 and as anticipated did not match the spectroscopic data
Initially, it was assumed that the error in relative stereochem-
istry was in the epoxide region, as opposed to the tetrol, where
the acyclic nature of the side chain would result in the least
reliable NOE data. The trans stereochemistry of the epoxide
was assumed to be correct on the basis of a good correlation
between the reported JH20/ H21 value and other trans-epoxides.
An error in the relative stereochemistry of the epoxide was
(13) (a) Kobayashi, J.; Ishibashi, M.; Nakamura, H.; Ohizumi, Y.; Yamasu, T.;
Sasaki, T.; Hirata, Y. Tetrahedron Lett. 1986, 27, 5755. (b) Kobayashi, J.;
Ishibashi, M.; Hirota, H. J. Nat. Prod. 1991, 54, 1435.
(14) Trost, B. M.; Chisholm, J. D.; Wrobleski, S. T.; Jung, M. J. Am. Chem.
Soc. 2002, 124, 12420.
(15) For a comprehensive review, see: Nicolaou, K. C.; Snyder, S. A. Angew.
Chem., Int. Ed. 2005, 44, 1012.
(16) Professor J. Kobayashi, personal communication.
(10) (a) Kessler, H.; Siegmeier, R. Tetrahedron Lett. 1983, 24, 281. (b) Bednarek,
M. A.; Bodansky, M. Int. J. Pept. Protein Res. 1983, 21, 196.
(11) Maleczka, R. E.; Terrell, L. R.; Geng, F.; Ward, J. S. Org. Lett. 2002, 4,
2841.
(12) Lam, H. W.; Pattenden, G. Angew. Chem., Int. Ed. 2002, 41, 508.
9
13602 J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005

Completion of the Synthesis of (
+
)-Amphidinolide A
A R T I C L E S
Table 2. Deviation of the 1H NMR Chemical Shifts of Isomers 1,
36, 38-41, 54, 56, 57, and 59 Relative to the Values Reported for
the Isolated Materiala
natural product (Table 2), the JH18/H19 value for 39 was
10.3 Hz, whereas the values for 1, 36, 38, and the natural product
were 3.3-3.8 Hz, thus suggesting the requirement for trans (as
drawn) C18-C19 stereochemistry.
a Spectra were measured in CDCl3 at 500 MHz. Differences are reported
in parts per million. Values in black represent deviations of <0.04, blue
0.04-0.10, green 0.11-0.20, and red >0.20
.
addressed by Pattenden, who reported the synthesis of epoxide
epimer 35, but it failed to match the spectroscopic data of the
natural product.12
An error in correlating the tetrol and epoxide portions, which
was determined stepwise by NOE from H11 to H13, from H13
to H15, and finally from H15 to H18, was considered by
Maleczka.11 However, as shown in Table 2, isomer 36 did not
match the natural product. Finally, the possibility of the
trisubstituted (E)-alkene being the source of the error was ruled
out by Maleczka, who prepared (Z)-alkene 37.11
Synthesis of Isomers Derived from 6. Although the most
obvious structures had been ruled out by Pattenden and
Maleczka, our attention remained focused on the epoxide
portion.17 The most logical targets were C22 epimer 38 and the
isomer 39 from inversion of the C19-C21 triad. Using the
endgame established for the synthesis of 1, 38 and 39 were
prepared. Although neither matched the data reported for the
Proceeding on the belief that the correlation of the tetrol to
the epoxide was tenuous at best, isomers 40, 41, and 54 were
targeted, combining changes in the epoxide with a change in
tetrol configuration. Although 40 and 41 were prepared un-
eventfully using the endgame established for 1, the hydrolysis
of 42 proved to be very problematic. As shown in Scheme 6,
(17) Trost, B. M.; Harrington, P. E. J. Am. Chem. Soc. 2004, 126, 5028.
9
J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005 13603

A R T I C L E S
Trost et al.
Scheme 6. Attempted Conversion of 42 to 54
Scheme 7. Hydrolysis of 44
Scheme 8. Alkene-Alkyne Coupling of 46 with 5
exposure of 42 to aqueous acetic acid failed to provide 54. A
number of byproducts were produced including tetrol 43, which
was isolated in 11% yield as a single diastereomer.18,19 Other
byproducts were isolated that appeared to be derived from
hydrolysis of the epoxide in 42 before complete cleavage of
the ketals. A wide variety of conditions were screened for the
conversion of 42 to 54; however, all resulted in hydrolysis of
the epoxide before complete deprotection of the tetrol.
As shown in Scheme 10, switching the order of the final two
steps gave a modest yield of 54; however, it should be noted
that this substrate is quite challenging given all the unprotected
and potentially reactive functionality presentsfour additional
double bonds, four free OH groups, and a sensitive epoxide
and ester. Although 54 did not match, it did provide a very
good fit in the epoxide region, thus suggesting that the relative
stereochemistry of the C18-C22 portion may be correct.
To this point, comparisons to the natural product were focused
on coupling constants, and apart from the requirement for trans
C18-C19 stereochemistry, little information was gleaned by
this approach. As shown in Figure 2, there were no other obvious
trends in the vicinal JH/H values that could be exploited to predict
the relative stereochemistry in the epoxide region. Clearly, a
new method for comparison was required.
It was suspected that the ketals in seco-alkyne 44 might be
more easily hydrolyzed than was the case for 42. In the event,
exposure of 44 to aqueous acetic acid gave hexol 45 as the major
product in 17% yield as a single diastereomer (Scheme 7).19
Products analogous to 43 were also isolated from the reaction.
In light of the difficulty encountered for the conversion of
42 to 54, a new route to 54 was pursued. The TES group was
chosen to replace the problematic cyclopentylidene ketals. Silyl
ether 46 was prepared from 6 by hydrolysis with aqueous acetic
acid followed by silylation with (TES)OTf. Somewhat surpris-
ingly, alkene-alkyne coupling of 46 with 5 catalyzed by either
[CpRu(MeCN)3]PF6 or [Cp*Ru(MeCN)3]PF6 gave only trace
amounts of 47 (Scheme 8). Alternatively, acid 49 was prepared
in good yield as shown in Scheme 9. Hydrolysis of 33 followed
by TES protection gave silyl ether 49 in 75% yield from 33.
Ester formation was lower yielding than was the case for ketal-
protected acid 33. The formation of ester 51 was accompanied
Initially, little effort was expended in comparing differences
in the proton chemical shifts of each isomer from those of the
natural product. However, upon careful examination of the
1
by a number of byproducts as judged by H NMR and TLC
analysis of the crude reaction mixture. However, alkene-alkyne
coupling of 51 catalyzed by either [CpRu(MeCN)3]PF6 or
[Cp*Ru(MeCN)3]PF6 rapidly cleaved the TES groups in 51. No
products that appeared to be derived from macrocycle 52 were
detected by 1H NMR, and only partially desilylated byproducts
were isolated.
(18) The location of the acetonide in 43 was not determined and is depicted
arbitrarily for 43.
(19) The relative stereochemistry was not determined.
Figure 2. Comparison of the vicinal JH/H values (Hz).
9
13604 J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005

Completion of the Synthesis of (
+
)-Amphidinolide A
A R T I C L E S
Scheme 9. Conversion of 33 to 51
Scheme 10. Macrocyclization of 53
Scheme 11. Route to 56 via Alkene-Alkyne Coupling of 55 with
chemical shifts of various protons, a rather significant departure
in the values relative to those of the natural product was
observed. As shown in Table 2, this departure was most dramatic
for the H9 and H11 protons. In addition, these differences were
not random in sign or magnitude, but were consistently upfield
relative to the values of the natural product. This was surprising
since the variations were made in the epoxide region, where
only small differences in the chemical shifts were observed in
comparison to the large deviations in the tetrol region. In light
of this realization, an error in the relative stereochemistry of
the tetrol appeared likely.
5
the epoxide was assumed to be correct. Therefore, 56, with the
C8-C9 diol inverted, became the primary target.
Synthesis of Isomers Derived from 55. Applying the same
sequence that was used for the synthesis of 1 and 38-41, isomer
56 was prepared uneventfully from 55 (Scheme 11). However,
isomer 56, with the C8-C9 diol inverted and epoxide stereo-
chemistry of 1, failed to match. Although the chemical shift of
H11 was closer than was the case for all the previous isomers,
H9 remained 0.30 ppm upfield from the natural product.
Although we were concerned that errors within the epoxide
combined with the tetrol would make the possibilities so
numerous that we would be faced with nearly an impossible
task, key spectral data provided direction. Because the JH8/H9
and JH11/H12 values for 1, 36, 38-41, and 54 were consistent
with the natural product, it appeared likely that the relative
stereochemistry of the diols was correct, but the stereochemistry
of the C8-C9 diol was incorrect relative to that of the C11-
C12 diol. Since H11 was correlated stepwise to H18 by NOE,
whereas the C8-C9 diol was not correlated to the epoxide
portion, the stereochemistry of the C11-C12 diol relative to
With these data, the combination of an error in the relative
stereochemistry of the tetrol and epoxide appeared likely.
Therefore, right-hand-side enantiomer 57 was targeted. Again,
an identical sequence was used for the conversion of 55 to 57.
Although 57 did not match the natural product, the chemical
9
J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005 13605

A R T I C L E S
Trost et al.
the epoxide in 61 before complete cleavage of the ketals. A
variety of conditions were examined for the cleavage of the
ketals; however, all failed to provide 59. If the hydrolysis
reactions were halted prior to the complete consumption of 61,
products analogous to 43 were detected in the crude reaction
1
mixture by H NMR.
The second route to 59 that was pursued utilized an approach
analogous to the one shown in Schemes 9 and 10 that was used
to prepare 54. Ketal 62 was hydrolyzed and converted to silyl
ether 64 in 83% yield over two steps from 62 (Scheme 13).
Ester formation gave 66 in 51% yield, and silyl cleavage with
buffered TBAF provided 67 in 79% yield. However, treatment
of 67 with catalytic [CpRu(MeCN)3]PF6 in DCEsthe same
conditions that were used to convert 53 to 54sgave a multitude
of byproducts. No products that appeared to be derived from
shifts in the tetrol region were extremely close to those of the
natural product as shown in Table 2. Only in the epoxide region
of 57 were the chemical shifts significantly different from those
of the natural product.
Since the chemical shifts deviated significantly from those
of the natural material only in the epoxide region, we assumed
that the relative stereochemistry of the 20-membered ring of
57 matched that of the natural product. We continued to assume
a trans-epoxide and trans H18/H19 stereochemistry. Therefore,
only diastereomers 58-60 remained as possibilities. Isomer 58,
1
59 were detected in the crude reaction mixture by H NMR.
With the failure of 67 to provide even trace amounts of 59,
a model system for the conversion of 67 to 59 was examined.
Alkyne 68 was chosen as a suitable model for 67. As shown in
Table 3, exposure of 68 to 10 mol % [CpRu(MeCN)3]PF6 in
DCE gave a number of byproducts, but no 69 was isolated
(Table 3, entry 1). Increasing the [CpRu(MeCN)3]PF6 loading
to 1 equiv gave the desired macrocycle 69 in 8% yield (entry
2). Reaction of 68 in acetone gave a mixture of two products
that appeared to be the two monoacetonides of 68. Switching
to the [CpRu(COD)]Cl catalyst (entries 9-14) provided only
trace amounts of 69 in MeOH and none in a mixed solvent of
DMF/H2O (3:1).20 Additives such as NH4PF6 (entry 10) were
not beneficial. Examination of the [Cp*Ru(MeCN)3]PF6 catalyst
proved to be worthwhile. Macrocycle 69 was isolated in modest
yield from reactions conducted in DCE (entries 15 and 18) with
[Cp*Ru(MeCN)3]PF6. Chloride ion was examined as an additive
for coordination to the active catalyst, but no product was
detected in the reaction (entry 19).
Application of the conditions shown in entry 18 to 67
provided 59 (Scheme 14). Given the extraordinary chemo-
selectivity required in this fully unprotected highly funtionalized
substrate, the modest yield is still gratifying. The spectral data
for 59 provided an excellent fit to the natural product as shown
in Table 2. One proton deviated by 0.03 ppm, one by 0.02 ppm,
the C22 epimer of 57, did not seem a likely candidate given
the differences in chemical shift that exist between 1 and 38.
Therefore, only 59 and 60 remained; however, 59 was the
primary target given the analysis that follows. Isomer 54
provided the best overall fit in terms of chemical shift in the
epoxide region of any isomer. This is the same relative
stereochemistry in the epoxide region as found in 59. The more
quantitative analysis of the data in Table 2 that follows also
pointed to 59. The relationship between 57 and 59 is analogous
to that between 36 and 54, inversion of the C20-C22 triad. If
the changes in chemical shift that occur when the C20-C22
triad of 36 is inverted, thus yielding 54, are applied to 57, a
nearly perfect match to the natural material is obtained. For
example, the chemical shift of H19 in 36 and 54 is 4.58 and
4.67 ppm, respectively. This represents a downfield shift of
0.09 ppm. The shift of H19 in 57 is 4.65 ppm. A 0.09 ppm
downfield shift yields a predicted shift of 4.74 ppm for H19 of
59. This value compares well with the shift of 4.72 ppm for
H19 of the natural product. Analysis of the other protons yields
similar results.
1
and the remainder by 0.01 ppm or less. The H NMR spectra
in C6D6 and CD3OD deviated by 0.01 ppm or less from those
of the natural product in those solvents. The J values in all three
solvents were also in agreement. The 13C NMR spectrum
deviated by 0.1 ppm or less in CDCl3. These results are well
within experimental error.21 The optical rotation of [R]24D +56
(c 0.05, CHCl3) was identical in sign to but slightly higher than
the reported value of [R]24 +46 (c 1.0, CHCl3), therefore
D
establishing the absolute stereochemistry as shown for 59.
Finally, subsequent comparison of 59 and a sample of natural
amphidinolide A by HPLC and NMR further confirmed the
structural assignment.
Conclusion
We have used a combination of synthesis and NMR spec-
troscopy to determine the correct relative and absolute stereo-
chemistries of amphidinolide A. Overall, 23 linear steps and
The first route to 59 is shown in Scheme 12. Ketal 61 was
prepared from 55 via the standard route. As was the case for
42, exposure of 61 to aqueous acid resulted in hydrolysis of
(20) Trost, B. M.; Probst, G. D.; Schoop, A. J. Am. Chem. Soc. 1998, 120,
9228.
(21) See the Supporting Information for
spectroscopic data.
a complete comparison of the
9
13606 J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005

Completion of the Synthesis of (
+
)-Amphidinolide A
A R T I C L E S
Scheme 12. Attempted Conversion of 61 to 59
Scheme 13. Second Route to 59
2H), 5.01 (s, 1H), 4.35-4.43 (m, 3H), 4.26 (t, J ) 7.3 Hz, 1H), 4.19
(d, J ) 8.4 Hz, 1H) 4.07 (d, J ) 8.0 Hz, 1H), 3.92 (m, 1H), 3.06 (dd,
J ) 15.4, 5.2 Hz, 1H), 2.28 (s, 3H), 1.92-1.60 (m, 16H); 13C NMR
(125 MHz, CDCl3) δ 166.9, 152.8, 144.0, 143.3, 142.3, 141.2, 135.6,
134.2, 133.9, 127.7, 127.0, 125.1, 120.0, 119.0, 118.7, 118.0, 117.8,
117.4, 115.6, 84.0, 81.2, 79.7, 79.6, 65.8, 46.8, 37.6, 37.4, 37.3, 37.2,
36.3, 34.6, 23.5, 23.4, 23.3, 13.9; HRMS m/z calcd for C32H32O4
(M+ - C10H16O2) 480.2301, found 480.2327.
43 total steps were required to convert (-)-diethyl tartrate to
amphidinolide A in an overall yield of 0.2%. A reliance on
catalytic asymmetric processes allowed the building blocks for
each isomer to be assembled with relative ease. Moreover, the
convergent nature of the synthesis allowed these blocks to be
assembled without an inordinate investment of time given the
complexity of the targets. Finally, the remarkable ability of
[Cp*Ru(MeCN)3]PF6 to catalyze the macrocyclization of tetrol
67 clearly points to the potential power of the Ru-catalyzed
alkene-alkyne coupling and warrants further study to determine
the scope and generality of this transformation.
(2E,4E)-7-((2R,3R)-3-{1-[(2R,3R)-3-Allyl-1,4-dioxaspiro[4.4]non-
2-yl]vinyl} -1,4-dioxaspiro[4.4]non-2-yl)-3-methylocta-2,4,7-trienoic
Acid (33). To a solution of ester 32 (401 mg, 0.62 mmol) in CH2Cl2
(18 mL) at room temperature was added piperidine (2.1 mL, 1.8 g, 21
mmol). After 2.5 h at room temperature, the reaction mixture was
diluted with ether, water, and 1 M HCl. The aqueous phase was
extracted with ether (3×), and the combined organic extracts were
washed with brine (1×), dried over MgSO4, and concentrated. Purifica-
tion by flash column chromatography on silica gel (40% EtOAc in
petroleum ether) gave acid 33 (262 mg, 90%) as a colorless oil: Rf )
Experimental Section
9H-Fluoren-9-ylmethyl (2E,4E)-7-((2R,3R)-3-{1-[(2R,3R)-3-Allyl-
1,4-dioxaspiro[4.4]non-2-yl]vinyl}-1,4-dioxaspiro[4.4]non-2-yl)-3-
methylocta-2,4,7-trienoate (32). Alkyne 6 (36 mg, 0.104 mmol) and
alkene 5 (159 mg, 0.523 mmol) were dissolved in 300 µL of
dichloroethane and warmed to 50 °C in an oil bath, and [Cp*Ru-
(MeCN)3]PF6 (5.1 mg, 0.01 mmol) was then added in one portion. After
approximately 2 h, the reaction was placed atop a silica gel column.
Purification by flash column chromatography on silica gel (10-30%
Et2O in petroleum ether) gave alkene 32 (31 mg, 46%) as a 3.5:1
mixture of branched and linear products. Further elution with 40-80%
CH2Cl2 in petroleum ether provided recovered alkyne 6 (11 mg, 30%)
and alkene 5 (108 mg, 69%). Data for 32: Rf ) 0.41 (20% Et2O in
petroleum ether); IR (film) 2958, 2877, 1714 cm-1; 1H NMR (500 MHz,
CDCl3) δ 7.77 (d, J ) 7.6 Hz, 2H), 7.62 (d, J ) 7.5 Hz, 2H), 7.42 (t,
J ) 7.5 Hz, 2H), 7.33 (t, J ) 7.5 Hz, 2H), 6.13-6.19 (m, 2H), 5.82-
5.90 (m, 2H), 5.48 (s, 1H), 5.47 (s, 1H), 5.24 (s, 1H), 5.09-5.19 (m,
0.69 (10% MeOH in CHCl3); IR (film) 3076, 2959, 1687 cm-1
;
1H NMR (500 MHz, CDCl3) δ 6.15-6.22 (m, 2H), 5.80-5.88 (m,
1H), 5.75 (s br, 1H), 5.47 (s, 1H), 5.46 (s, 1H), 5.22 (s, 1H), 5.08-
5.13 (m, 2H), 4.99 (s, 1H), 4.39 (d, J ) 8.4 Hz, 1H), 4.18 (d, J )
8.5 Hz, 1H), 4.04 (d, J ) 8.1 Hz, 1H), 3.90 (dt, J ) 7.9, 4.0, 1H), 3.05
(dd, J ) 16.6, 4.4 Hz, 1H), 2.89 (dd, J ) 16.5, 5.0 Hz, 1H), 2.39-
2.44 (m, 1H), 2.29 (s, 3H), 2.24-2.29 (m, 1H), 1.62-1.93 (m, 17H);
13C NMR (125 MHz, CDCl3) δ 172.2, 154.5, 143.2, 135.6, 134.8, 133.9,
119.1, 118.7, 117.8, 117.7, 117.4, 115.7, 84.0, 81.2, 79.7, 79.6, 37.6,
37.4, 37.3, 37.2, 36.3, 34.6, 23.5, 23.4, 23.2, 14.1; HRMS m/z calcd
for C28H38O6 470.2668, found 470.2662; optical rotation [R]23D +40.1
(c 1.2, CH2Cl2).
9
J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005 13607

A R T I C L E S
Trost et al.
Table 3. Optimization of the Macrocyclization of 68
concn
concn
(M)
temp
C)
yield
(%)
entry
catalyst
(mol %)
solvent
additive
(
°
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(MeCN)3]PF6
[CpRu(COD)]Cl
[CpRu(COD)]Cl
[CpRu(COD)]Cl
[CpRu(COD)]Cl
[CpRu(COD)]Cl
11
109
10
12
16
100
15
16
18
11
11
9
11
18
15
15
13
25
27
DCE
DCE
acetone
DME
THF
THF
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.001
0.05
0.01
0.01
0.01
0.001
0.001
0.001
0.001
0.001
0.001
50
50
50
50
50
50
50
50
60
60
60
60
60
80
50
50
50
50
50
0
8
0
<5
<5
0
DMF
0
DCE/DMF (20:1)
MeOH
MeOH
MeOH
MeOH
MeOH
DMF/H2O (3:1)
DCE
DMF
THF
DCE
DCE
<5
<5
<5
<5
<5
<5
0
7
0
<5
31
0
NH4PF6 (50 mol %)
AgPF6 (50 mol %)
In(OTf)3 (17 mol %)
LiCl (21 mol %)
[CpRu(COD)]Cl
[Cp*Ru(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
[Cp*Ru(MeCN)3]PF6
n-Bu4NCl (32 mol %)
Scheme 14. Macrocyclization of Tetrol 67
(3a′R,4a′R,7a′R,16′S,17′R,22a′R)-12′,17′-Dimethyl-16′-{(2S,3R)-
3-[(1R)-1-methylbutyl]oxiran-2-yl}-4′,8′,19′-tris(methylene)-3a′,4′,-
4a′,7a′,8′,9′,16′,17′,18′,1 9′,20′,22a′-dodecahydro-14′H-dispiro[cy-
clopentane-1,2′-bis[1,3]dioxolo[4,5-i:4′,5′-l]oxacycloicosine-6′,1′′-
cyclopentan]-14′-one (34). Ester 34 (24 mg, 0.0362 mmol) was
dissolved in 36 mL of dichloroethane and the resulting solution warmed
to 50 °C in an oil bath. [CpRu(MeCN)3]PF6 was then added. After
1 h, the reaction was quenched with the addition of 50 µL of Et3N and
concentrated. Purification by flash column chromatography on silica
gel (10-30% Et2O in petroleum ether) gave 34 (14 mg, 58%) as a
yellow oil: Rf ) 0.32 (20% Et2O in petroleum ether); IR (film) 2958,
1714 cm-1; 1H NMR (500 MHz, CDCl3) δ 6.16 (d, J ) 16.0 Hz, 1H),
6.07-6.13 (m, 1H), 5.79 (s br, 1H), 5.75 (t, J ) 7.4 Hz, 1H), 5.54 (s,
1H), 5.50 (s, 1H), 5.29 (s, 1H), 5.25 (dd, J ) 15.1, 7.3 Hz, 1H), 5.17
(s, 1H), 4.80 (s, 1H), 4.71 (s, 1H), 5.54 (dd, J ) 7.2, 3.7 Hz), 4.28 (d,
J ) 8.2 Hz, 1H), 4.20 (d, J ) 8.3 Hz, 1H), 4.13 (t, J ) 7.7 Hz, 1H),
3.85 (d, J ) 7.8 Hz, 1H), 3.16 (dd, J ) 13.8, 8.8 Hz, 1H), 3.03 (dd,
J ) 14.0, 5.0 Hz, 1H), 2.96 (dd, J ) 7.2, 2.2 Hz, 1H), 2.71 (dd, J )
7.3, 2.1 Hz, 1H), 2.55-2.66 (m, 2H), 2.44 (dd, J ) 13.8, 6.0 Hz, 1H),
2.30 (s, 3H), 2.12-2.18 (m, 1H), 1.80-1.95 (m, 8H), 1.62-1.79 (m,
8H), 1.22-1.56 (m, 6H), 1.04 (d, J ) 7.2 Hz, 3H), 0.99 (d, J )
6.8 Hz, 3H), 0.93 (t, J ) 7.2 Hz, 3H); 13C NMR (125 MHz, CDCl3)
δ 166.0, 151.9, 146.0, 143.1, 142.5, 135.3, 135.2, 133.9, 133.3, 127.9,
118.9, 118.8, 117.2, 115.1, 111.7, 84.9, 82.3, 81.1, 78.2, 77.7, 61.5,
54.4, 39.3, 38.1, 37.6, 37.4, 37.3, 37.1, 36.7, 35.7, 35.5, 33.9, 29.7,
23.5, 23.4, 23.3, 23.3, 19.9, 17.2, 16.0, 14.3; HRMS m/z calcd for
(1R)-3-O-[(2E,4E)-7-((2R,3R)-3-{1-[(2R,3R)-3-Allyl-1,4-dioxaspiro-
[4.4]non-2-yl]vinyl}-1,4-dioxaspiro[4.4]non-2-yl)-3-methylocta-2,4,7-
trienoyl]-1,2-anhydro-4,5-dideoxy-1-[(1R)-1-methylbutyl]-4-prop-2-
yn-1-yl-L-arabinitol (3). Acid 33 (36 mg, 0.0764 mmol) and [RuCl2(p-
cymene)]2 (4.7 mg, 0.00764 mmol) were dissolved in 2 mL of toluene.
Ethyl ethynyl ether (40% in hexanes, 20 µL, 0.084 mmol) was then
added. After 2 h, the reaction was concentrated to a dark solid. Alcohol
7 (42 mg, 0.1997 mmol) was then added in 300 µL of dichloroethane,
followed by a solution of CSA (1.8 mg, 0.00764 mmol) in 200 µL of
dichloroethane. After 2 h, the reaction was quenched with 10 µL of
Et3N and placed atop a silica gel column. Purification by silica gel
chromatography (20-40% Et2O in petroleum ether) gave ester 3
(33 mg, 66%), recovered epoxide 7 (27 mg, 64%), and acid 33
(5.5 mg, 15%). Data for 3: Rf ) 0.19 (10% Et2O in petroleum ether);
IR (film) 3303, 2559, 2116 cm-1; 1H NMR (500 MHz, CDCl3) δ 6.14
(s br, 2H), 5.81-5.89 (m, 1H), 5.75 (s, 1H), 5.47 (s, 1H), 5.46 (s, 1H),
5.22 (s, 1H), 2.08-5.13 (m, 2H), 4.98 (s, 1H), 4.83 (t, J ) 6.0 Hz,
1H), 4.38 (t, J ) 8.5 Hz, 1H), 4.18 (d, J ) 8.5 Hz, 1H), 4.04 (d, J )
8.1 Hz, 1H), 3.88 (m, 1H), 3.03 (d, J ) 16.5 Hz, 1H), 2.92 (dd, J )
6.0, 2.2 Hz, 1H), 2.88 (dd, J ) 15.9, 4.4 Hz, 1H), 2.68 (dd, J ) 7.4,
2.2 Hz, 1H), 2.36-2.44 (m, 2H), 2.35 (dd, J ) 5.1 Hz, 1H), 2.29 (s,
3H), 2.22-2.28 (m, 2H), 2.05-2.11 (m, 1H), 2.05-2.11 (m, 1H), 2.01
(t, J ) 2.7 Hz, 1H), 1.59-1.90 (m, 13H), 1.20-1.46 (m, 6H), 1.13 (d,
J ) 7.0 Hz, 3H), 0.93 (d, J ) 6.8 Hz, 3H), 0.88 (t, J ) 7.2 Hz, 3H);
13C NMR (125 MHz, CDCl3) δ 166.1, 153.2, 143.3, 142.4, 135.6, 134.2,
133.9, 119.0, 118.7, 117.8, 117.4, 115.6, 84.1, 81.8, 81.2, 79.7, 79.6,
74.7, 70.2, 61.5, 56.8, 51.0, 37.6, 37.5, 37.4, 37.3, 37.2, 36.6, 36.2,
35.3, 35.1, 34.6, 29.7, 23.5, 23.4, 23.3, 20.0, 15.8, 15.2, 14.2; HRMS
C41H58O7 662.4183, found 662.4179; optical rotation [R]23 +59.6
D
(c 1.0, CH2Cl2).
(+)-Amphidinolide A (Proposed Structure, 1). Ketal 34 (7.6 mg,
0.0115 mmol) was suspended in 300 µL of dioxane, and 200 µL of
water was added. Amberlyst-15 (15 mg, 0.0705 mmol) was then added,
and the reaction was warmed to 60 °C. After 60 h, additional Amberlyst-
15 (30 mg, 0.141 mmol) was added and the temperature raised to
85 °C. After 6 h, the reaction was filtered and concentrated. The residue
was the purified by HPLC (70% ethyl acetate in hexanes) to provide
1 (3.7 mg, 61%) as a white foam: Rf ) 0.41 (ethyl acetate); IR (film)
3384, 3284, 2922, 1714 cm-1; 1H NMR (500 MHz, CDCl3) δ 6.18 (d,
J ) 15.6 Hz, 1H), 6.04-6.10 (m, 1H), 5.79 (s, 1H), 5.72-5.76 (m,
m/z calcd for C41H58O7 662.4183, found 662.4187; optical rotation [R]23
+30.9 (c 0.75, CH2Cl2).
D
9
13608 J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005

Completion of the Synthesis of (
+
)-Amphidinolide A
A R T I C L E S
1H), 5.59 (s, 1H), 5.58 (s, 1H), 5.36 (s, 1H), 5.25 (d, J ) 13.4 Hz,
1H), 4.86 (s, 1H), 4.77 (s, 1H), 4.53 (dd, J ) 7.2, 3.8 Hz, 1H), 4.28 (s,
1H), 4.21 (d, J ) 7.2 Hz, 1H), 3.82 (s, 1H), 3.22 (bd, J ) 14.9, 1H),
3.11 (t, J ) 15.5, 1H), 2.91 (d, J ) 6.2 Hz, 1H), 2.74 (d br, J )
13.8 Hz, 1H), 2.63-2.70 (m, 3H), 2.52 (d, J ) 12.5 Hz, 1H), 2.34 (s
br, 1H), 2.29 (s, 3H), 2.11-2.20 (m, 2H), 2.01-2.09 (m, 1H), 1.77 (t,
J ) 11.8 Hz, 1H), 1.10-1.60 (m, 6H), 0.97 (d, J ) 6.7 Hz, 3H), 0.86-
0.92 (m, 6H); 13C NMR (125 MHz, CDCl3) δ 165.6, 152.0, 145.2,
136.3, 134.9, 131.1, 130.2, 119.3, 115.4, 114.8, 113.8, 96.1, 94.3, 73.4,
72.5, 70.9, 69.6, 61.4, 60.4, 54.2, 38.5, 36.7, 35.4, 33.3, 31.9, 29.7,
1.14 (d, J ) 6.6 Hz, 3H), 0.98-0.92 (m, 39H), 0.88 (t, J ) 6.7 Hz,
3H), 0.65-0.55 (m, 24H); 13C NMR (126 MHz, CDCl3) δ 166.1, 153.5,
148.8, 146.5, 136.5, 135.5, 135.0, 117.2, 117.0, 116.6, 112.8, 82.3,
78.0, 75.9, 74.9, 74.2, 69.9, 69.6, 62.6, 55.7, 37.1, 36.6, 35.4, 35.2,
35.0, 22.4, 20.0, 15.7, 15.6, 14.3, 14.2, 7.12, 7.09, 7.06, 7.0, 5.4, 5.30,
5.27, 5.1; ESIMS m/z calcd for C55H103O7Si4 (M+ + H) 987.7, found
987.1; optical rotation [R]23 -18 (c 0.50, CH2Cl2).
D
(5S)-4,5-Anhydro-1,2-dideoxy-5-[(1R)-1-methylbutyl]-2-prop-2-
yn-1-yl-3-O -[(2E,4E,8R,9R,11S,12S)-8,9,11,12-tetrahydroxy-3-meth-
yl-7,10-bis(methylene)pentadeca-2,4,14-trienoyl]-D-ribitol (67). To
a solution of silyl ether 66 (45 mg, 0.046 mmol) in THF (1.0 mL) at
room temperature was added 200 µL of a mixture of TBAF (2.0 mL,
1 M in THF, 2.0 mmol) and acetic acid (115 µL, 121 mg, 2.01 mmol).
The reaction mixture was stirred at room temperature for 30 h and
concentrated. Purification by flash column chromatography on silica
gel (from 30% to 50% to 70% EtOAc in petroleum ether) gave tetrol
67 (19 mg, 79%) as a white foam: Rf ) 0.54 (EtOAc); IR (film from
CH2Cl2) 3424 (br), 3297, 2917, 1716, 1607, 1236, 1150 cm-1; 1H NMR
(500 MHz, CDCl3) δ 6.19-6.14 (m, 2H), 5.85 (dddd, J ) 16.8, 14.3,
10.6, 7.1 Hz, 1H), 5.72 (d, J ) 1.0 Hz, 1H), 5.34 (s, 1H), 5.32 (s, 1H),
5.23 (s, 1H), 5.17 (dm, J ) 7.2 Hz, 1H), 5.14 (dd, J ) 1.1, 1.1 Hz,
1H), 5.02 (d, J ) 1.0 Hz, 1H), 4.71 (dd, J ) 4.2, 2.1 Hz, 1H), 4.33-
4.31 (m, 2H), 4.17 (d, J ) 4.9 Hz, 1H), 3.83 (ddd, J ) 8.3, 4.6,
4.6 Hz, 1H), 3.03 (dd, J ) 17.0, 4.9 Hz, 1H), 2.92 (dd, J ) 16.6,
4.4 Hz, 1H), 2.80 (dd, J ) 7.5, 2.1 Hz, 1H), 2.75 (dd, J ) 6.5, 2.2 Hz,
1H), 2.39 (dm, J ) 14.2 Hz, 1H), 2.33 (ddd, J ) 16.2, 5.6, 2.8 Hz,
1H), 2.26 (d, J ) 1.1 Hz, 3H), 2.25 (dd, J ) 14.5, 8.1 Hz, 1H), 2.18
(dd, J ) 4.3, 2.7 Hz, 1H), 2.13 (m, 1H), 1.98 (t, J ) 2.7 Hz, 1H),
1.48-1.21 (m, 5H), 1.13 (d, J ) 6.7 Hz, 3H), 0.88 (t, J ) 7.2 Hz,
3H), 0.87 (d, J ) 6.7 Hz, 3H); 13C NMR (126 MHz, CDCl3) δ 166.0,
152.9, 147.5, 146.3, 135.6, 134.24, 134.18, 118.5, 117.8, 116.7, 114.0,
82.2, 76.0, 74.8, 74.1, 73.0, 72.3, 69.8, 62.5, 55.6, 37.9, 36.6, 36.0,
35.2, 35.0, 22.5, 19.9, 15.6, 14.3, 14.2, 14.0; ESIMS m/z calcd for
29.3, 22.7, 22.3, 21.0, 14.2; optical rotation [R]23 ) +55.8 (c 0.2,
D
CH2Cl2).
(2E,4E,8R,9R,11S,12S)-3-Methyl-7,10-bis(methylene)-8,9,11,12-
tetrakis[(triethylsilyl)oxy]pentadeca-2,4,14-trienoic Acid (64). To
ketal 62 (76 mg, 0.16 mmol) at room temperature were added acetic
acid (1.5 mL) and water (500 µL). The reaction mixture was heated to
40 °C for 20 h and concentrated to give tetrol 63, which was used in
the next step without further purification. To a solution of tetrol 63,
prepared in the previous step, in THF (4 mL) at 0 °C were added i-Pr2-
NEt (260 µL, 193 mg, 1.5 mmol) and (TES)OTf (230 µL, 269 mg,
1.0 mmol). The reaction mixture was stirred at 0 °C for 10 min,
quenched with 1 M HCl, stirred for 10 min, and diluted with ether and
water. The aqueous phase was extracted with ether (3×), and the
combined organic extracts were washed with saturated KH2PO4 (1×)
and brine (1×), dried over MgSO4, and concentrated. Purification by
flash column chromatography on silica gel (5-10% EtOAc in petroleum
ether) gave silyl ether 64 (106 mg, 83%) as a colorless oil: Rf ) 0.56
(30% EtOAc in petroleum ether); IR (film) 2956, 1684, 1610, 1096,
738 cm-1;1H NMR (500 MHz, CDCl3) δ 6.20 (ddd, J ) 15.6, 6.8, 6.8
Hz, 1H), 6.14 (d, J ) 15.6 Hz, 1H), 5.82 (dddd, J ) 17.0, 14.3, 10.3,
7.1 Hz, 1H), 5.40-5.39 (m, 2H), 5.09 (s, 1H), 5.05 (dd, J ) 9.6,
2.0 Hz, 1H), 5.02 (d, J ) 0.9 Hz, 1H), 4.79 (d, J ) 1.3 Hz, 1H), 4.26
(d, J ) 4.0 Hz, 1H), 4.23 (d, J ) 3.8 Hz, 1H), 4.21 (d, J ) 3.7 Hz,
1H), 3.80 (ddd, J ) 7.9, 3.9, 3.9 Hz, 1H), 3.01 (dd, J ) 17.5, 6.6 Hz,
1H), 2.95 (dd, J ) 17.2, 6.6 Hz, 1H), 2.40 (dm, J ) 14.2 Hz, 1H),
2.31 (d, J ) 1.1 Hz, 3H), 2.30 (m, 1H), 2.14 (ddd, J ) 15.4, 7.7,
7.7 Hz, 1H), 0.99-0.91 (m, 36H), 0.65-0.51 (m, 24H), exchangeable
proton not listed; 13C NMR (126 MHz, CDCl3) δ 171.6, 155.0, 148.7,
146.5, 136.4, 136.1, 135.0, 117.2, 116.9, 116.6, 112.9, 78.1, 75.9, 75.8,
75.0, 37.1, 35.4, 14.0, 7.1, 7.05, 7.04, 7.01, 5.4, 5.32, 5.28, 5.1; optical
C31H46O7Na (M+ + Na) 553.3, found 553.3; optical rotation [R]29
D
+4.8 (c 0.10, CH2Cl2).
(+)-Amphidinolide A (Revised Structure, 59). A solution of alkyne
67 (15.6 mg, 0.029 mmol) in DCE (30 mL) was degassed by bubbling
nitrogen through the solution for 15 min. The reaction mixture was
heated to 50 °C, and [Cp*Ru(MeCN)3]PF6 (3.6 mg, 7.1 µmol) was
added. After 6 h at 50 °C, the reaction mixture was filtered through
silica gel and concentrated. Purification by flash column chromatog-
raphy on silica gel (from 40% to 60% to 80% EtOAc in petroleum
ether) gave (+)-amphidinolide A (59) (5.2 mg, 33%) and recovered
alkyne 67 (1.9 mg, 12%) as amorphous solids. Data for (+)-
amphidinolide A (59): Rf ) 0.33 (80% EtOAc in petroleum ether); IR
rotation [R]23 -1.3 (c 0.20, CH2Cl2).
D
(5S)-4,5-Anhydro-1,2-dideoxy-3-O-{(2E,4E,8R,9R,11S,12S)-3-
methyl-7,10-bis(methylene)-8,9,11,12-tetrakis[(triethylsilyl)oxy]pen-
tadeca-2,4,14-trienoyl}-5-[(1S)-1-methylbutyl]-2-prop-2-yn-1-yl-D-
ribitol (66). To a solution of acid 64 (170 mg, 0.21 mmol) in PhMe
(3.0 mL) at room temperature were added [RuCl2(p-cymene)]2 (12 mg,
0.020 mmol) and ethoxyacetylene (60 µL, 50 wt % in hexanes,
0.34 mmol). The reaction mixture was stirred at room temperature for
2 h and concentrated under a stream of argon. A solution of alcohol
65 (108 mg, 0.51 mmol) in DCE (2.5 mL) was added via cannula
followed by CSA (5.1 mg, 22 µmol). The reaction mixture was stirred
at room temperature for 2 h, filtered through silica gel, and concentrated.
Purification by flash column chromatography on silica gel (from 1%
to 2% to 5% to 20% EtOAc in petroleum ether) gave ester 66
(108 mg, 51%) as a colorless oil: Rf ) 0.36 (5% EtOAc in petroleum
(film from CH2Cl2) 3438 (br), 2927, 1715, 1612, 1235, 1150, 850 cm-1
;
1H NMR (500 MHz, CDCl3) δ 6.28 (d, J ) 15.4 Hz, 1H), 6.10 (ddd,
J ) 15.5, 9.6, 5.1 Hz, 1H), 5.80 (s, 1H), 5.69 (dddd, J ) 15.3, 7.3,
7.3, 1.5 Hz, 1H), 5.50 (dd, J ) 15.4, 4.8 Hz, 1H), 5.49 (s, 1H), 5.38
(d, J ) 1.0 Hz, 1H), 5.36 (d, J ) 1.0 Hz, 1H), 5.21 (s, 1H), 4.88 (s,
1H), 4.79 (s, 1H), 4.72 (dd, J ) 6.1, 3.2 Hz, 1H), 4.59 (s br, 1H), 4.43
(s br, 1H), 4.22 (s br, 1H), 4.06 (t br, J ) 3.8 Hz, 1H), 3.21 (dd, J )
14.9, 5.0 Hz, 1H), 3.11 (dd, J ) 14.7, 9.3 Hz, 1H), 2.85 (dd, J ) 6.2,
2.2 Hz, 1H), 2.76 (dd, J ) 7.5, 2.2 Hz, 1H), 2.76 (s br, 2H), 2.34 (dd,
J ) 15.3, 5.0 Hz, 1H), 2.27 (d, J ) 1.1 Hz, 3H), 2.16 (m, 1H), 1.92
(dd, J ) 14.4, 9.2 Hz, 1H), 1.53-1.24 (m, 5H), 1.06 (d, J ) 7.0 Hz,
3H), 0.93 (d, J ) 6.8 Hz, 3H), 0.91 (t, J ) 7.1 Hz, 3H); 13C NMR
(101 MHz, CDCl3) δ 165.8, 152.8, 147.2, 145.0, 144.8, 136.4, 134.8,
130.9, 130.6, 118.6, 116.1, 114.7, 112.9, 75.8, 74.7, 73.5, 72.5, 70.5,
61.8, 54.2, 39.7, 39.0, 36.7, 36.2, 35.4, 33.3, 20.0, 15.8, 14.9, 14.3,
13.9. HRESIMS m/z calcd for C31H46O7Na (M+ + Na) 553.3141, found
ether); IR (film) 3312, 2946, 1714, 1605, 1146, 1094, 741 cm-1
;
1H NMR (500 MHz, CDCl3) δ 6.20-6.10 (m, 2H), 5.82 (dddd, J )
17.2, 14.4, 10.3, 7.3 Hz, 1H), 5.71 (s, 1H), 5.40-5.39 (m, 2H), 5.08
(s, 1H), 5.05 (dm, J ) 9.0 Hz, 1H), 5.02 (d, J ) 1.2 Hz, 1H), 4.79 (d,
J ) 1.5 Hz, 1H), 4.70 (dd, J ) 6.7, 4.3 Hz, 1H), 4.26 (d, J ) 4.0 Hz,
1H), 4.23 (d, J ) 3.7 Hz, 1H), 4.21 (d, J ) 3.5 Hz, 1H), 3.80 (ddd,
J ) 8.2, 3.9, 3.9 Hz, 1H), 3.00 (dd, J ) 16.6, 5.7 Hz, 1H), 2.94 (dd,
J ) 17.2, 6.4 Hz, 1H), 2.82 (dd, J ) 7.5, 2.2 Hz, 1H), 2.75 (dd, J )
6.7, 2.1 Hz, 1H), 2.40 (dm, J ) 14.2 Hz, 1H), 2.34 (ddd, J ) 15.9,
5.5, 2.7 Hz, 1H), 2.29 (d, J ) 1.1 Hz, 3H), 2.19 (dd, J ) 7.9, 2.7 Hz,
1H), 2.16-2.11 (m, 2H), 1.98 (t, J ) 2.7 Hz, 1H), 1.49-1.22 (m, 5H),
553.3157; optical rotation [R]24 +56 (c 0.05, CHCl3).
D
Acknowledgment. We thank the National Institutes of Health
(Grant GM-33049) for their generous support of our programs.
J.D.C. and S.T.W. were supported by NIH postdoctoral fellow-
9
J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005 13609

A R T I C L E S
Trost et al.
ships. Mass spectra were provided by the Mass Spectrometry
Facility, University of San Francisco, supported by the NIH
Division of Research Resources. We thank Professor Robert
Maleczka, Professor Gerald Pattenden, and Dr. Lamont Terrell
for sharing their spectral data for 1 and 36. We also thank
Professor Kobayashi for performing the comparison of 59 to
natural amphidinolide A.
Supporting Information Available: Experimental procedures
and characterization data for 9, 10, 12-16, 29-31, 43, 45, 49,
51, 53, and 62, and characterization data for 38-41, 54, 56,
57, and 59. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA053365Y
9
13610 J. AM. CHEM. SOC. VOL. 127, NO. 39, 2005