Divergence en route to nonclassical annonaceous acetogenins. Synthesis of pyranicin and pyragonicin

Article abstract of DOI:10.1021/jo052233k

Syntheses of the nonclassical annonaceous acetogenins, pyranicin, and pyragonicin from common late-stage intermediates are presented. The construction of key elements relies on asymmetric HWE reactions, including the desymmetrization of a meso-dialdehyde

Full text of DOI:10.1021/jo052233k

Divergence en Route to Nonclassical Annonaceous Acetogenins.
Synthesis of Pyranicin and Pyragonicin†
Daniel Strand,‡ Per-Ola Norrby,§ and Tobias Rein*,‡,|
KTH Chemical Science and Engineering, Organic Chemistry, SE-100 44 Stockholm, Sweden,
Department of Chemistry, Technical UniVersity of Denmark, Building 201, KemitorVet,
DK-2800 Kgs. Lyngby, Denmark, and DiscoVery Chemistry, Research and DeVelopment,
AstraZeneca, SE-151 85 So¨derta¨lje, Sweden
Tobias.Rein@astrazeneca.com
ReceiVed October 27, 2005
Syntheses of the nonclassical annonaceous acetogenins, pyranicin, and pyragonicin from common late-
stage intermediates are presented. The construction of key elements relies on asymmetric HWE reactions,
including the desymmetrization of a meso-dialdehyde and a parallel kinetic resolution of a racemic
aldehyde. A stereoconvergent Pd-catalyzed substitution serves to install the C4 stereocenter in protected
form with different orthogonal protective groups. A divergent strategy to form 1,4- and 1,6-diols, employing
stereoselective Zn-mediated alkynylations, is used for completion of the core structures. Notably, the
stereoselective coupling reaction toward pyragonicin proceeds with highly functionalized fragments. The
methodology is further expanded by a divergent synthesis of all stereoisomers of the 2,3,6-trisubstituted
tetrahydropyran subunit.
Introduction
ability to inhibit multiple drug resistant (MDR) tumor cell lines.3
The cytotoxic effects exhibited by ACGs are at least in part
due to the inhibition of NADH-ubiquinone oxidoreductase
(mitochondiral complex I) of the respiratory chain.4 This
inhibition results in a depletion of ATP levels which causes
arrest in the cell cycle at the G1 phase, and subsequently
apoptosis is induced.5 The exact details of this mechanism still
remains elusive however.2a
Since the discovery of uvaricin in 1982,1 the annonaceous
acetogenins (ACGs) class of natural products has been a center
of attention for chemists and biologists alike.2 This interest stems
from the structural diversity and potent biological effects
exhibited by these structures, including antimicrobial, pesticidal,
and cytotoxic properties. An especially attractive feature is their
Structurally, the ACGs are a series of C-35 or C-37 fatty
acid derivatives, typically bearing one or more tetrahydrofuran
† This paper is dedicated, with respect and gratitude, to Professor Stephen F.
Martin on the occasion of his 60th birthday.
‡ KTH Chemistry.
§ Technical University of Denmark.
(3) Oberlies, N. H.; Croy, V. L.; Harrison, M. L.; McLaughlin, J. L.
Cancer Lett. 1995, 115, 73-79.
| AstraZeneca.
(1) Tempesta, M. S.; Kriek, G. R.; Bates, R. B. J. Org. Chem. 1982, 47,
3151-3153.
(4) The ACGs are among the most potent inhibitors of NADH-
ubiquinone oxidoreductase known; see ref 2a.
(2) (a) Bermejo, A.; Figadere, B.; Zafra-Polo, M.-C.; Barrachina, I.;
Estornell, E.; Cortes, D. Nat. Prod. Rep. 2005, 22, 269-303. (b) Alali, F.
Q.; Liu, X.-X.; McLaughlin, J. L. J. Nat. Prod. 1999, 62, 504-540. (c)
Zafra-Polo, M. C.; Figadere, B.; Gallardo, T.; Tormo, J. R.; Cortes, D.
Phytochemistry 1998, 48, 1087-1117. (d) Casiraghi, G.; Zanardi, F.;
Battistini, L.; Rassu, G.; Appendino, G. Chemtracts 1998, 11, 803-827.
(5) (a) Yuan, S.-S. F.; Chang, H.-L.; Chen, H.-W.; Yeh, Y.-T.; Kao,
Y.-H.; Lin, K.-H.; Wu, Y.-C.; Su, J.-H. Life Sci. 2003, 72, 2853-2861. (b)
Morre, D. J.; de Cabo, R.; Farley, C.; Oberlies, N. H.; McLaughlin, J. L.
Life Sci. 1995, 56, 343-348. (c) Hollingworth, R. M.; Ahammadsahib, K.
I.; Gadelhak, G.; McLaughlin, J. L. Biochem. Soc. Trans. 1994, 22, 230-
233.
10.1021/jo052233k CCC: $33.50 © 2006 American Chemical Society
Published on Web 02/02/2006
J. Org. Chem. 2006, 71, 1879-1891
1879

Strand et al.
FIGURE 1. Pyranicin, pyragonicin, and some structurally related natural products.
(THF) or tetrahydropyran (THP) rings, often flanked by adjacent
hydroxyl groups (Figure 1). The lion’s share of the synthetic
efforts toward these structures has been directed toward the
mono-THF and adjacent bis-THF subclasses.6 Some of these
studies have employed a modular approach to access several
natural ACGs7 or synthetic analogues8 from common synthetic
intermediates or through a common synthetic sequence, includ-
ing an elegant example of a mix/split fluorous tag strategy by
Curran and co-workers in the synthesis of murisolin and isomers
thereof.9
Pyranicin (1) and the related pyragonicin (2) are members
of the nonclassical subgroup of acetogenins,10 and both were
isolated from the stem bark of Goniothalamus giganteus by
McLaughlin and co-workers in 1998.11 When tested against a
panel of human carcinoma solid tumor cell lines, pyranicin was
shown to exhibit selective inhibitory action against PACA-2
(human pancreatic cancer) cell lines with a potency of about
10 times that of clinically used adriamycin. Pyranicin was also
found to be generally about 10 times more potent compared to
pyragonicin, which is in good agreement with the observed
structure activity relationship (SAR) that optimum potency is
achieved in acetogenins bearing ring systems starting at C15.12
A previous synthesis of pyranicin has been reported by
Takahashi and co-workers.13 We have recently reported con-
vergent strategies for the synthesis of pyranicin14 and pyragoni-
cin15 (the latter published back-to-back with a coincident
independent synthesis by Takahashi et al.16). Herein we report
the full details of these syntheses.
Results and Discussion
Synthesis Plan. We envisioned the construction of both 1
and 2 from a common butenolide fragment 12 joined to either
of two THP fragments 7 or 8 which, in turn, would be accessed
through a common route (Scheme 1). Thus, the key disconnec-
tion in our retrosynthetic analysis is the C-C bond joining C15
and C14 in 1 and C13 and C12 in 2, respectively. For the
pyranicin case we envisioned, in the synthetic direction, the
asymmetric addition of an acetylenic unit to each of the two
fragments 8 and 13; a subsequent metal-catalyzed cross-coupling
would then provide the desired 1,6-diol motif. In the pyragonicin
case, the related 1,4-diol would arise from a direct stereoselec-
tive coupling of an identical butenolide fragment 12 to an
intermediate 8, differing from the pyranicin intermediate 8 only
in the length of the terminal aliphatic chain, thus enabling late-
stage divergence to the two targets.
Previously, we have shown that an asymmetric HWE de-
symmetrization of meso-dialdehydes17 such as 10 can be utilized
for the construction of analogues of 818 via a substrate-controlled
intramolecular oxa-Michael cyclization.19 However, to access
the desired all-cis stereochemistry in the THP ring of pyranicin/
pyragonicin, the meso symmetry of 10 has to be modified via
inversion of one stereocenter. This extra step is well compen-
sated for by the divergent nature of such an approach, allowing
access to several diastereomers of the 2,3,6-trisubstituted THP
subunit from a common precursor. The C4 stereocenter of the
butenolide fragment would originate from racemic acrolein
dimer (14), which is subjected to a parallel kinetic resolution20
(6) For recent examples see: (a) Tinsley, J. M.; Roush, W. R. J. Am.
Chem. Soc. 2005, 127, 10818-10819. (b) Nattrass, G. L.; Diez, E.;
McLachlan, M. M.; Dixon, D. J.; Ley, S. V. Angew. Chem., Int. Ed. 2005,
44, 580-584. (c) Han, H.; Sinha, M. K.; D’Souza, L. J.; Keinan, E.; Sinha,
S. C. Chem. Eur. J. 2004, 10, 2149-2158. (d) Crimmins, M. T.; She, J. J.
Am. Chem. Soc. 2004, 126, 12790-12791.
(7) For examples see: (a) Keum, G.; Hwang, C. H.; Kang, S. B.; Kim,
Y.; Lee, E. J. Am. Chem. Soc. 2005, 127, 10396-10399. (b) Das, S.; Li,
L.-S.; Abraham, S.; Chen, Z.; Sinha, S. C. J. Org. Chem. 2005, 70, 5922-
5931. (c) Marshall, J. A.; Piettre, A.; Paige, M. A.; Valeriote, F. J. Org.
Chem. 2003, 68, 1771-1779. (d) Marshall, J. A.; Piettre, A.; Paige, M. A.;
Valeriote, F. J. Org. Chem. 2003, 68, 1780-1785. (e) Emde, U.; Koert, U.
Eur. J. Org. Chem. 2000, 1889-1904. (f) Sinha, S. C.; Sinha, A.; Yazbak,
A.; Keinan, E. J. Org. Chem. 1996, 61, 7640-7641.
(14) Strand, D.; Rein, T. Org. Lett. 2005, 7, 199-202.
(15) Takahashi, S.; Ogawa, N.; Koshino, H.; Nakata, T. Org. Lett. 2005,
7, 2783-2786.
(16) Strand, D.; Rein, T. Org. Lett. 2005, 7, 2779-2781.
(17) For recent review concerning asymmetric HWE reactions see: Rein,
T.; Pedersen, T. M. Synthesis 2002, 579-594.
(18) Vares, L.; Rein, T. J. Org. Chem. 2002, 67, 7226-7237.
(19) (a) Betancort, J. M.; Martin, V. S.; Padron, J. M.; Palazon, J. M.;
Ramirez, M. A.; Soler, M. A. J. Org. Chem. 1997, 62, 4570-4583. (b)
Betancort, J. M.; Martin, V. S.; Padron, J. M.; Palazon, J. M.; Ramirez, M.
A.; Soler, M. A. J. Org. Chem. 1997, 62, 4570-4583. See also: (c)
Schneider, C.; Schuffenhauer, A. Eur. J. Org. Chem. 2000, 73-82.
(8) Jiang, S.; Li, Y.; Chen, X.-G.; Hu, T.-S.; Wu, Y.-L.; Yao, Z.-J. Angew.
Chem., Int. Ed. 2004, 43, 329-334.
(9) Zhang, Q.; Lu, H.; Richard, C.; Curran, D. P. J. Am. Chem. Soc.
2004, 126, 36-37.
(10) The nonclassical subgroup is defined as acetogenins bearing a THP
and/or a ring-hydroxylated THF moiety.
(11) Alali, F. Q.; Rogers, L.; Zhang, Y.; McLaughlin, J. L. Tetrahedron
Lett. 1998, 54, 5833-5844.
(12) For further details on SAR, see refs 2a,b and references therein.
(13) Takahashi, S.; Kubota, A.; Nakata, T. Org. Lett. 2003, 5, 1353-
1356.
1880 J. Org. Chem., Vol. 71, No. 5, 2006

Synthesis of Pyranicin and Pyragonicin
SCHEME 1. Retrosynthesis of Pyranicin and Pyragonicin
followed by a stereoconvergent reaction sequence.21 In the key
Pd-catalyzed transformation, we hoped to expand our previous
methodology by introducing a protective group suitable for last-
step global deprotection. The butenolide moiety would then be
constructed using well-established methodology through a
condensation with a lactaldehyde derivative.22
SCHEME 2. Synthesis of (Z)-Hetero-Michael Precursor 1926
Synthesis of the Hetero-Michael Precursors 19 and 22.
The first key transformation was an asymmetric HWE desym-
metrization of meso-dialdehyde 15, readily available in multi-
gram scale in five steps using Ba¨ckvalls syn-1,4-diacetoxylation
of cyclohexadiene.18,23 Crucial for the application of this
desymmetrization reaction to the synthesis of 1 and 2 was access
to a practical analogue of the antipode24 of the previously used
(-)-(1R,2S,5R)-8-phenylmenthol. Gratifyingly, the commer-
cially available (1S,2R)-nor-8-phenylmenthol25 showed a virtu-
ally identical behavior to that of 8-phenylmenthol in the
desymmetrization of 15, giving 16 in good yield as a single
detected isomer (77%, dr >98:2, (Z):(E) >98:2) (Scheme 2).
Reduction of aldehyde 16 proceeded as expected using NaBH4
in i-PrOH/THF at 0 °C, with the desired migration of the
secondary TBDPS group to the less hindered primary position,
giving a readily separable 90:10 mixture of secondary/primary
alcohols from which the desired secondary alcohol was isolated
in good yield (70%). A more convenient procedure, in which
crude aldehyde 16 was filtered through a short silica column
and subjected to the reduction/migration conditions, enabled the
isolation of pure secondary alcohol 18 in 70% yield over two
steps. The remaining mixture of secondary/primary alcohol (17/
18) could be reequilibrated in refluxing EtOH with a catalytic
amount of DMAP in 70% yield (isolated secondary alcohol 18)
to further increase the overall yield of the sequence.
(20) Pedersen, T. M.; Jensen, J. F.; Humble, R. E.; Rein, T.; Tanner,
D.; Bodmann, K.; Reiser, O. Org. Lett. 2000, 2, 535-538.
(21) Pedersen, T. M.; Hansen, E. L.; Kane, J.; Rein, T.; Helquist, P.;
Norrby, P.-O.; Tanner, D. J. Am. Chem. Soc. 2001, 123, 9738-9743.
(22) Marshall, J. A.; Jiang, H. J. Nat. Prod. 1999, 62, 1123-1127.
(23) Ba¨ckvall, J. E.; Bystro¨m, S. E.; Nordberg, R. E. J. Org. Chem. 1984,
49, 4619-4631.
(24) (+)-(1S,2R,5S)-8-phenylmenthol is now commercially available. For
preparation, see: Buschmann, H.; Scharf, H. D. Synthesis 1988, 827-830.
(25) Comins, D. L.; Salvador, J. M. J. Org. Chem. 1993, 58, 4656-
4661. See also: Vaulont, I.; Gais, H.-J.; Reuter, N.; Schmitz, E.; Ossenkamp,
R. K. L. Eur. J. Org. Chem. 1998, 805-826.
(26) For clarity, the chiral auxiliaries used are abbreviated: R′ ) (1S,2R)-
nor-8-phenylmenthyl; R′′ ) (1R,2S,5R)-8-phenylmenthyl.
To access the stereochemistry needed to reach 8, an inversion
of the C7 stereocenter of 18 was required. This was ac-
complished using standard Mitsunobu contitions (TPP, DIAD,
chloroacetic acid).27 The use of toluene as a reaction medium
(27) Saiah, M.; Bessodes, M.; Antonakis, K. Tetrahedron Lett. 1992,
33, 4317-4320.
J. Org. Chem, Vol. 71, No. 5, 2006 1881

Strand et al.
SCHEME 3. Synthesis of (E)-Hetero-Michael Precursor
SCHEME 5. Michael Cyclization of the (E)-Substrate 2226
2226
SCHEME 6. Michael Cyclization of the (Z)-Substrate 1926
SCHEME 4. Predicted Kinetic Products of the
Hetero-Michael Cyclizations
proved beneficial over THF, resulting in a cleaner and faster
reaction, giving the inverted chloroacetate in excellent yield
(95%). The selective hydrolysis of the chloroacetate over the
nor-8-phenylmenthol ester was then readily accomplished with
0.4 M LiOH in THF/H2O at ambient temperature in excellent
yield (92%). The more sterically demanding nor-8-phenylmen-
thyl ester is virtually inert to these conditions, and the secondary
alcohol was isolated as a 10:1 separable mixture of 19 and 19a.
To evaluate which substrate was the most suitable for the
hetero-Michael cyclization, we also exploited the (E)-selective
asymmetric HWE desymmetrization of meso-dialdehyde 15,
giving, after reduction, the secondary alcohol 21,18 which was
readily inverted to give 22 over two steps (Scheme 3). This
route however suffers from a lower yield and a somewhat lower
enantiotopic selectivity (dr ) 93:7) in the desymmetrization step,
resulting in the isolation of (4R,7R)-22 and (4S,7S)-22 as an
inseparable 93:7 mixture in 94% over two steps from 21.
Stereoselective Hetero-Michael Cyclizations. In a detailed
study by Mart´ın and co-workers,19a,b it was shown that the
stereochemical outcome of oxa-Michael cyclizations with
substrates similar to 19 and 22 is influenced by the allylic
stereocenter and the alkene geometry, resulting in a preference
for the (E)-alkene substrates to cyclize to 2,3-cis-THP deriva-
tives and for the (Z)-alkenes to give the corresponding 2,3-trans-
THP products under kinetically controlled conditions.28 On the
basis of these findings, we expected the (E)-olefin 23 to give
the desired 2,3-cis-2,6-cis-THP 24 and the (Z)-olefin 25 to give
the 2,3-trans-2,6-trans-THP 26 as the initial products (Scheme
4), although we anticipated 24 to be thermodynamically favored.
Preliminary results showed that the desired 2,3-cis-2,6-cis-
THP 27 is indeed formed as the major product in the cyclization
of (E)-alkene 22, but unfortunately as an inseparable mixture
with its diastereomer (2S,3S,6S)-27 originating from the minor
product of the initial (E)-selective HWE desymmetrization
(Scheme 5).29
tereoselectivity (Scheme 6), thus overall being clearly superior
to (E)-alkene 22 for the purpose of accessing 29. The use of
catalytic rather than stoichiometric amounts of base completely
suppressed the formation of byproducts. Running the reaction
in THF resulted in no protective group migration, and any
primary alcohol 19a present in the starting material was
recovered unaffected by the reaction conditions. Change of
solvent to toluene, however, gave the same level of stereoin-
duction as well as allowed for protective group migration,
resulting in a quantitative yield of 29 from the mixture of
secondary and primary alcohols 19 and 19a obtained from the
hydrolysis of the chloroacetate.
With this result at hand we became interested in the possibility
of stopping the reaction at the presumed transient 2,3-trans-
2,6-trans isomer. By running the reaction at low temperature
we could occasionally at very low conversion isolate this product
as a single diastereomer, but not in a reproducible manner.
However, by using a nonnucleophilic protic cosolvent (1:1
t-BuOH/toluene), we were successful in promoting the cycliza-
tion to give 28 in good yield (76%) in a perfectly stereocon-
trolled manner. An interesting aspect of this result is that we
now, in conjuction with previous results, have controlled access
to all eight stereoisomers of these 2,3,6-trisubstituted THP
derivatives (Figure 2). Isomers 33/32 and 34/31 can be directly
formed via the cyclization of (Z)-alkene 18 or the corresponding
(E)-alkene as previously described,18 whereas isomers 35/24 and
30/26 can be accessed from 18 via a Mitsunobu inversion and
control of the reaction conditions in the subsequent cyclization.
The relevance of this is further emphasized by the presence of
these substructures in, besides pyranicin/pyragoncin, several
other natural products such as decarestrictine L (6),30 jimenezin
(5),31 and mucocin (3)32 (Figures 1 and 2).
We also investigated the possibility of accessing THP 29
directly from chloroacetate 36 (Scheme 7) by in situ hydrolysis
under the cyclization conditions; this was achieved by the use
of hydroxide ions generated from an excess of t-BuOK (3.0
equiv) and water (2.0 equiv).33,34 A similar level of stereose-
Pleasingly, the secondary alcohol 19 derived from the (Z)-
selective HWE desymmetrization and inversion sequence formed,
when exposed to our standard cyclization conditions, the desired
2,3-cis-2,6-cis-THP 29 in quantitative yield with a 96:4 dias-
(30) Grabley, S.; Hammann, P.; Huetter, K.; Kirsch, R.; Kluge, H.;
Thiericke, R.; Mayer, M.; Zeeck, A. J. Antibiot. 1992, 45, 1176-1181.
(31) Chavez, D.; Acevedo, L. A.; Mata, R. J. Nat. Prod. 1998, 61, 419-
421.
(28) This trend was efficiently exploited in the hetero-Michael cycliza-
tions of 18 and 21 as described in ref 18.
(29) Upon cleavage of the chiral auxiliary in the subsequent step, the
relationship between the products becomes enantiomeric, thus preventing
removal of the minor isomer until a further reagent controlled asymmetric
transformation is carried out.
(32) Shi, G.; Alfonso, D.; Fatope, M. O.; Zeng, L.; Gu, Z.-m.; Zhao,
G.-x.; He, K.; MacDougal, J. M.; McLaughlin, J. L. J. Am. Chem. Soc.
1995, 117, 10409-10410.
(33) Chang, F. C.; Wood, N. F. Tetrahedron Lett. 1964, 2969-2973.
1882 J. Org. Chem., Vol. 71, No. 5, 2006

Synthesis of Pyranicin and Pyragonicin
protocol as implemented in MacroModel. Solvation was in-
cluded through the GB/SA continuum model40 in MacroModel,
employing the only organic solvent implemented in our version
of MacroModel, chloroform. For the kinetic preference, we
investigated the conformational space of the anionic reactant,
with the distance between the alkoxide and the â-position
harmonically constrained to 1.5 Å. This approach, which does
not allow actual bond formation, will not give reaction barriers,
but only the steric cost of approaching the nucleophile to the
Michael acceptor. However, for diastereomeric transition states,
the barrier should correlate with the calculated energies.
The computational results are depicted in Scheme 8. Looking
first at the (Z)-reactant 37, we see that the conformational space
is restricted by allylic strain. For the â-γ bond, only one rotamer
is energetically accessible, enforcing nucleophilic attack from
one side only, leading to the 2,3-trans-2,6-trans product 39. In
the presence of suitably acidic protons (t-BuOH), the ring
closure product is immediately neutralized to yield 40. However,
in the absence of a proton source, the Michael attack is
reversible, but free rotation around the R-â bond before the
retro-Michael reaction now leads to the (E)-reactant 43. This
TS lacks the allylic strain and therefore has a lower barrier than
the reverse reaction to the (Z)-reactant 37.
FIGURE 2. THP substructures of natural products accessible from
meso-dialdehyde 12.
SCHEME 7. Direct Cyclization of the (Z)-Chloroacetate
3626
For the (E)-reactant, two rotameric forms 43 and 44 are
accessible, allowing nucleophilic attack from either face of the
Michael acceptor. The more favorable path, as judged from the
near-TS constrained conformations, is the one leading to the
2,3-cis-2,6-cis product 47.
lectivity was observed (dr ) 93:7) in preference of the 2,3-cis-
2,6-cis product, but given that the yield over two separate
reactions was higher, 92% compared to 65% for the one-step
procedure, the two-step route was found superior.
Finally, investigating the possible configurations and con-
formations of the neutral product, we note that the 3- and 6-THP
substituents must always be cis and that one of them therefore
must assume an axial position in a chair conformation. The
alternative boat or twist-boat conformations consistently have
high energies, as expected. Interestingly enough, the preferred
conformation always has the 3-TBDPS-O-substituent in an
axial position, even in the 2,3-trans-2,6-trans-THP, which thus
assumes a 2,3-diaxial conformation 41. Obviously, the repulsion
between the 2- and 3-substituents in the diequatorial conforma-
tion is large enough to compensate for the 1,3-diaxial interac-
tions that each substituent experiences with hydrogen atoms on
the THP ring in the diaxial conformation. Given the axial
preference of the 3-substituent, we get the expected result that
the 2,3-cis-2,6-cis-THP with two equatorial substituents is ca.
6 kJ/mol more stable than the best conformation of the 2,3-
trans-2,6-trans-THP with only one equatorial substituent. The
experimentally observed ratio of ca. 20:1 corresponds to a free
energy difference of ca. 8 kJ/mol. A deviation of 2 kJ/mol is a
very good result considering the size of the system and the
complexity of the conformational space, but not unexpected
from MMFF.37
Completion of the THP Fragments. With the intermediate
29 at hand, we turned to the completion of the THP fragments.
Reduction of ester 29 with DIBAL to the corresponding alcohol
in CH2Cl2 at -78 °C was sluggish and often led to a mixture
of the corresponding aldehyde and alcohol; furthermore, the
aldehyde could not be readily separated from the released nor-
8-phenylmenthol. This was circumvented by reduction to the
aldehyde followed by in situ addition of the appropriate
phosphorus ylide in THF at -78 °C and allowing the reaction
mixture to warm to room temperature (Scheme 9). Workup
Mechanistic Insights into the Hetero-Michael Cyclizations
of 19 and 22. The results of the ring closure reactions prompted
us to perform a computational investigation of the intramolecular
Michael addition. The experimental results indicate that the 2,3-
trans-2,6-trans-THP 28 is the kinetic product from the ring
closure of (Z)-substrate 19 and that the 2,3-cis-2,6-cis-THP 29
is the thermodynamic product. The thermodynamic preference
was addressed by molecular mechanics calculations, employing
the MMFF force field35 in MacroModel 7.2.36 MMFF is one of
the most accurate force fields available, and yields conforma-
tional energies comparable to high-level quantum mechanical
calculations,37 while still being fast enough to allow full
conformational searching. The conformational space was ex-
plored using a combined Monte Carlo38/Lowmode39 search
(34) Procedure: To a stirred solution of chloroacetate 36 (31 mg, 0.032
mmol) in 1.5 mL of toluene was added a freshly prepared solution of water
(1.17 µL, 0.065 mmol) and t-BuOK (9.7 µL, 1.0 M in THF) in 0.7 mL of
toluene at 0 °C. The temperature was brought to room temperature over 1
h. The reaction was then quenched by addition of phosphate buffer pH 7
and subjected to the usual workup giving the cyclisized product 29 (18.3
mg, 65%).
(35) Halgren, T. A. J. Comput. Chem. 1996, 17, 490-519.
(36) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440-467. For more recent versions, see: www.schrod-
inger.com.
(37) (a) Gundertofte, K.; Liljefors, T.; Norrby, P.-O.; Pettersson, I. J.
Comput. Chem. 1996, 17, 429-449. (b) Liljefors, T.; Gundertofte, K.;
Norrby, P.-O.; Pettersson, I. In Computational Medicinal Chemistry for Drug
DiscoVery; Bultinck, P., Tollenaere, J. P., De Winter, H., Langenaeker, W.,
Eds.; Dekker: New York, 2004; pp 1-28.
(38) Chang, G.; Guida, W. C.; Still, W. C. J. Am. Chem. Soc. 1989,
111, 4379-4386.
(39) Kolossva´ry, I.; Guida, W. C. J. Am. Chem. Soc. 1996, 118, 5011-
5019.
(40) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J. Am.
Chem. Soc. 1990, 112, 6127-6129.
J. Org. Chem, Vol. 71, No. 5, 2006 1883

Strand et al.
SCHEME 8. Mechanism for the Intramolecular Michael Addition of 1926
SCHEME 9. Synthesis of the THP Fragments Continued26
FIGURE 3. Expected outcome of a Cram-chelatecontrolled addition.
a closely related model substrate,46 only to give a disappointing
57:43 selectivity. It thus appears that substrates such as 50a,
having a 2,3-cis-2,6-cis configuration, do not give a strong bias
in chelate-controlled additions or do not readily form chelates.
We therefore turned our attention to the reagent-controlled,
zinc-mediated acetylide addition developed by Carreira and co-
workers47 that recently has been successfully used in the addition
of TMS-acetylene to 2-formyltetrahydrofuran derivatives.48 We
were pleased to find that the addition proceeded in a highly
stereocontrolled manner under standard conditions using Zn-
(OTf)2/(-)-NME to give the desired propargylic alcohol in good
yield (83%, dr > 95:5 49) (Scheme 10).
A one-pot TBS protection of the propargylic alcohol using
TBSCl, imidazole, and DMAP in CH2Cl2 at room temperaure
followed by removal of the TMS group by quenching the
reaction with methanol and K2CO3 then gave the protected
propargylic alcohol 53 in good yield (93%). Hydrozirconation
with the Schwartz reagent in THF at room temperature followed
by addition of elemental iodine thereafter gave exclusively the
problems due to formation of insolubles associated with the
Wittig reaction and to some degree the DIBAL reductions was
conveniently avoided by direct evaporation of the reaction
mixture onto silica, and purification by flash chromatography
gave the olefinated products 48a and 48b directly from the ester
in good yields (48a, 75%; 48b, 78%) and (Z):(E) selectivities
(∼10:1 in both cases).
Selective deprotection at the primary position through adsorp-
tion onto activated alumina from hexanes41 proceeded as
expected to give the desired primary alcohols 49a and 49b in
good yields (49a, 83%; 49b, 81%). The subsequent oxidations
using Dess-Martin periodinane in CH2Cl2 then gave the
corresponding aldehydes in good yields (50a, 81%; 50b, 83%).
For series b, toward pyragonicin, this provided the completed
THP fragment 50b.
(45) Michelet, V.; Adiey, K.; Tanier, S.; Dujardin, G.; Genet, J.-P. Eur.
J. Org. Chem. 2003, 2947-2958.
(46) To simplify the determination of the diastereoselecitivty of the
addition, the saturated aldehyde 85 was used as a model.
To complete the pyranicin THP fragment, three additional
steps were needed, starting with a syn-stereoselective addition
of a two-carbon spacer to 50a. The direct addition of organo-
metallic reagents to analogous 2-formyl-THP derivatives has
been studied previously in syntheses of mucocin,42 muconin,43
and jimenezin,44 but with only modest levels of the desired syn-
selectivity, requiring an additional two-step oxidation/reduction
sequence to achieve good stereocontrol. Since a 1,2-cyclic Cram-
chelate controlled addition would afford the desired syn-product
(Figure 3), addition of TMS-acetylene magnesium bromide in
the presence of a 10-fold excess of MgBr245 was attempted on
(47) (a) Boyall, D.; Frantz, D. E.; Carreira, E. M. Org. Lett. 2002, 4,
2605-2606. (b) Anand, N. K.; Carreira, E. M. J. Am. Chem. Soc. 2001,
123, 9687-9688.
(48) (a) Kojima, N.; Maezaki, N.; Tominaga, H.; Yanai, M.; Urabe, D.;
Tanaka, T. Chem. Eur. J. 2004, 10, 672-680. (b) Kojima, N.; Maezaki,
N.; Tominaga, H.; Asai, M.; Yanai, M.; Tanaka, T. Chem. Eur. J. 2003, 9,
4980-4990.
(49) Since intermediates 52-54 and 76 were carried through the synthesis
as mixtures of (E):(Z) isomers, an exact quantification of the stereochemical
outcome of the addition could not be made. However, we did not observe
a minor diastereomer derived from this transformation in any of the
remaining steps of the synthesis. The relative configuration of 52 was
assigned by analogy with literature.47
(41) Feixas, J.; Capdevila, A.; Guerrero, A. Tetrahedron 1994, 50, 8539-
8550.
(42) Takahashi, S.; Nakata, T. J. Org. Chem. 2002, 67, 5739-5752.
(43) Yang, W.-Q.; Kitahara, T. Tetrahedron Lett. 1999, 40, 7827-7830.
(44) Takahashi, S.; Maeda, K.; Hirota, S.; Nakata, T. Org. Lett. 1999,
1, 2025-2028.
1884 J. Org. Chem., Vol. 71, No. 5, 2006

Synthesis of Pyranicin and Pyragonicin
SCHEME 10. Completion of the Pyranicin THP Fragment26
SCHEME 11. Synthesis of the Butenolide Precursor
Intermediatesa,26
(E)-vinyl iodide in 86% yield, thus completing the pyranicin
THP fragment 54.
Synthesis of the Butenolide Precursor 65. The synthesis
of the butenolide fragment began with an asymmetric HWE
parallel kinetic resolution (PKR) of racemic acrolein dimer 14,
in which the substrate enantiomers are known to form products
with opposite alkene geometry (Scheme 11).50 We have previ-
ously performed a PKR with this substrate using a mixture of
the (E)-selective phosphonate 56 and the (Z)-selective phos-
phonate 55 in excellent yield and with good diastereoselectiv-
ity.20
a Yields are given for the mixtures. See Supporting Information for details
regarding the pure isomers.
We were however pleased to find that a simpler protocol,
using slow addition of phosphonate 55 to a slight excess (1.3
equiv) of rac-14 in THF at -78 °C, provided better results in
terms of diastereoselectivity. The olefinated products were
obtained as a 60:40 mixture51 of (2Z,4S)-61 (dr ) 96:4) and
(2E,4R)-57 (dr ) 98:2), respectively, in 77% combined yield.
It should be pointed out that although 57 and 61 are readily
separable at this stage, we chose not to separate the isomers,
but carried them through the subsequent three steps as mixtures,
with similar or better results in terms of yield, since later
intermediates 60 and 64 give the exact same product in the Pd-
(0)-catalyzed step.
SCHEME 12. Stereoconvergent Stepa,26
The subsequent acid-catalyzed addition of water to vinyl
ethers 57 and 61 using 1 M HCl was sluggish, but changing
the proton source to p-TSA provided reproducible conditions
in 1:1 water/THF, giving the hemiacetals 58 and 62 in good
combined yields (85%) as inseparable 34:66 mixtures of
diastereomers. It could be noted that precise temperature control
is of the essence; the reaction was slow at room temperature,
and at 40 °C mainly an unidentified byproduct was isolated,
while a reaction at 32 ( 2 °C gave the desired products. Next,
an HWE olefination using a benzyl ester phosphonate served
to trap the ring-open forms of the hemiacetals, while incorporat-
ing a functionality that could be readily cleaved in the presence
of the 8-phenylmenthol ester. Due to the predisposition of 62/
63 to lactonize under basic conditions, extensive optimization
of the reaction conditions of this step was needed. Using
KHMDS as the base, lactonization could not be suppressed;
with the sodium base the reaction proceeded slowly at -20 °C
with some lactonization, but to our delight the use of LiHMDS
almost completely suppressed the ring closure. When running
at 0 °C in THF, the reaction went to completion overnight. The
secondary alcohols were then activated for the Pd-catalyzed
allylic substitution by conversion into the corresponding phos-
a Reagents and conditions: (a) 60/64, Pd2dba3 (0.2-0.4 equiv Pd), ligand
(0.4-0.6 equiv), cosolvent/R-OH (1:1), rt.
phinate esters 60 and 64 using Ph2P(O)Cl, imidazole and DMAP
in a mixture of 1:1 ClCH2CH2Cl/THF at 60 °C in excellent
yield (96%).
Stereoconvergent Step. In a palladium-catalyzed allylic
substitution (Z)-alkenes are known to often react with soft
nucleophiles with inversion of both the alkene geometry and
the allylic stereocenter through a π-σ-π rearrangement,
whereas (E)-alkenes are known to react with overall retention
at both stereogenic units. By exposing our mixture of (2E,4R)-
60 and (2Z,4S)-64 to a Pd(0) catalyst with a suitable primary
alcohol as nucleophile, we expected, on the basis of our previous
results, both substrate isomers to converge into a single
O-protected product. The reaction proved to be highly dependent
on the reaction conditions, the main byproduct being diene 68
formed via â-elimination of the intermediate π-allyl complex.
To suppress this side reaction, a large excess of the nucleophile
was needed; typically a 1:1 mixture of nucleophile52 and a
cosolvent was used (Scheme 12, Table 1).
(50) For mechanistic insight into the asymmetric HWE reaction see:
Norrby, P.-O.; Brandt, P.; Rein, T. J. Org. Chem. 1999, 64, 5845-5852.
(51) This should be compared to a 54:46 (Z):(E) mixture using a 50:50
mixure of phosphonates 55 and 56; see footnote 20.
(52) It should be noted that the excess TMS(CH2)2OH could be recovered
in 80% by bulb-to-bulb distillation of the reaction crude.
J. Org. Chem, Vol. 71, No. 5, 2006 1885

Strand et al.
yield, %
TABLE 1. Representative Results of the Stereoconvergent Step
nucleophile
ligand
cosolvent
(2E,4R):(2E,4S)a
product
product:68
(TMS)(CH2)2OH
(TMS)(CH2)2OH
(TMS)(CH2)2OH
(TMS)(CH2)2OH
PMBOH
neocuproine
neocuproine
DPPE
DPPP
neocuproine
neocuproine
CH2Cl2
THF
THF
THF
CH2Cl2
THF
97:3b
97:3c
98:2b
94:6b
97:3b,e
-
65
65
65
65
66
94:6b
87:13c
52:48c
61:39c
d
72
65
d
d
89
0
TBSOH
(67)
0:100
a The dr of the product in all cases matches that of the starting material. b Determined by 1H NMR of the crude product. c Determined by 1H NMR after
filtration through a short plug of silica. d Not recorded. e Separable by flash chromatography.
SCHEME 13. Completion of the Shared Butenolide
When evaluating ligands, neocuproine was found superior
to phosphine-based ligands, both in terms of reaction times and
Fragment26
yield. Using neocuproine, the reaction times could be shortened
to 4 h using 5 mol % Pd2dba3, compared to 3 days with DPPE
using 20 mol % Pd2dba3. On a gram scale, as little as 1.25% of
Pd catalyst was used; however a prolonged reaction time (16
h) was required. The cosolvent also had a large impact on the
ratio between the desired product and the elimination product
68, and CH2Cl2 was found superior to THF without affecting
the diastereoselectivity of the reaction. To evaluate the scope
of this step as a means to introduce protected hydroxyl
functionalities in an allylic position, the reaction was carried
out with a set of alcohols: (TMS)(CH2)2OH, PMBOH and
TBSOH. The reaction proceeded similarly well with (TMS)-
(CH2)2OH and PMBOH, the difference in yield mainly reflecting
the tedious purification of the (TMS)(CH2)2O-R reaction. The
TBSOH failed completely to give product, probably for steric
reasons. It could be noted that although (TMS)(CH2)2O-R is
an established protective group for acids, it has rarely been
utilized for secondary alcohols;53 one might speculate this is
partly because of its difficult introduction. This methodology
provides a convenient way to introduce PMBO- and (TMS)-
(CH2)2O- in allylic positions. For our purpose the highly robust
(TMS)(CH2)2- group proved ideal to carry through the remaining
steps of the synthesis.
Completion of the Shared Butenolide Fragment. With the
C4 stereocenter in place we focused on the completion of the
butenolide fragment (Scheme 13). Hydrogenation/hydrogenoly-
selectivity (dr ) 98:2).54 For the conversion of aldehyde 70
sis of 65 was performed with Pd(C) in hexanes. Although the
into propargylic alcohol 71 this reaction proved reliable and
reaction was slow at room temperature, it provided the saturated
should provide a significant improvement over the more
acid in good yield (90%). The use of THF as solvent in this
common three-step procedure involving a nonselective addition
reaction surprisingly resulted in partial reductive removal of the
of TMS-acetylene followed by an oxidation/reduction se-
C4 oxygen, and Rh (5% on alumina) also in THF caused
quence. For a thorough discussion of the performance of this
extensive ring hydrogenation. A subsequent borane reduction
followed by Dess-Martin oxidation provided aldehyde 70 in
good overall yield (82%).
reaction in various complex settings, see ref 55 and references
therein.
The butenolide was contructed using classical methodology,
beginning with an aldol reaction of a preformed enolate derived
To install the C10 stereocenter, we once again utilized the
Carreira methodology to add TMS-acetylene stereoselectively.
This transformation proved more challenging than the corre-
sponding addition to 2-formyl-THP 50a; it appears that an R-oxy
substituent is an activating factor for this transformation. Under
standard conditions (Zn(OTf)2, (+)-NME, and Et3N in toluene
at room temperature) only traces of product could be isolated.
We were however pleased to find that raising the temperature
to 60 °C in a sealed vessel with an excess of TMS-acetylene
increased the conversion significantly, and the propargylic
alcohol 71 was isolated in 75% yield with excellent diastereo-
from 71 and aldehyde 75, readily available from (S)-lactic acid
methyl ester in two steps.56 The aldol product was isolated as
a mixture of isomers and carried through the subsequent two
steps as such. The THP group was then cleaved with subsequent
ring closure to the corresponding lactone under acidic conditions
using catalytic amounts of CSA in refluxing methanol. To access
the completed butenolide, we then needed to perform a bis-
(54) The diastereomeric ratio of 71 was determined by 1H and 19F analysis
of the corresponding (+)- and (-)-MTPA derivatives. The relative
configuration was assigned by analogy with literature.47
(55) Kirkham, J. E. D.; Courtney, T. D. L.; Lee, V.; Baldwin, J. E.
Tetrahedron 2005, 61, 7219-7232.
(53) (a) Burke, S. D.; Pacofsky, G. J.; Piscopio, A. D. Tetrahedron Lett.
1986, 27, 3345-3348. (b) Paquette, L. A.; Backhaus, D.; Braun, R. J. Am.
Chem. Soc. 1996, 118, 11990-11991.
(56) Ito, Y.; Kobayashi, Y.; Kawabata, T.; Takase, M.; Terashima, S.
Tetrahedron 1989, 45, 5767-5790.
1886 J. Org. Chem., Vol. 71, No. 5, 2006

Synthesis of Pyranicin and Pyragonicin
SCHEME 14. End Game toward Pyranicin
SCHEME 15. Attempted End Game toward Pyragonicin
leaving the (TMS)(CH2)2 ether untouched. We were however,
delighted to find that HF(aq) at a slightly elevated temperature
(40 °C) enabled us to obtain unprotected pyranicin (1) in good
yield (85%). Spectroscopic properties of our synthetic material
were in all respects identical to those reported by Takahashi.
Takahashi et al. confirmed McLaughlin’s original assignment
of the relative and absolute configuration of pyranicin by
preparing the known corresponding (+)- and (-)-MTPA
derivatives of 1. There is, however, a strong discrepancy in the
optical rotation reported for the natural product and that of the
activation under conditions where only the C13-ester of 73 was
eliminated. We were pleased to find that the formed lactone
was converted into butenolide 74 under conditions described
by Marshall and co-workers57 using TFAA and E3N. Due to
the purification problems caused by the oxidation byproduct
formed in the reaction between Et3N and TFAA,58 a slightly
modified protocol was used where the TFAA was exchanged
for Cl3CCOCl. The C10 alcohol could then be released during
a mild workup with NaHCO3 without affecting the now
epimerization-prone C34 stereocenter59 to give the completed
butenolide fragment 74 ready for coupling to either of 50b or
54.
synthetic material: natural material, [R]23 -9.7 (c ) 0.008,
D
CHCl3);62 synthetic material (Takahashi), [R]23 +19.5 (c )
D
0.55, CHCl3); our synthetic material, [R]23 +21.1 (c ) 0.24,
D
End Game toward Pyranicin. A Sonogashira cross-coupling
of 54 and 74 giving ene-yne 76 provided the completed
pyranicin framework in good yield (89%) using (PPh3)2PdCl2
and CuI in neat Et3N, without interference from the unprotected
butenolide C10 hydroxy center (Scheme 14). Previously, these
cross-coupling conditions have been used in the final stages of
acetogenin syntheses and shown not to cause epimerization of
the butenolide.57 The subsequent removal of the chain unsat-
urations could not be accomplished using hydrogenation with
Wilkinsons catalyst; instead an inseparable mixture of alkenes
was obtained. In sharp contrast, reduction with diimide generated
in situ from TsNHNH2 and NaOAc in DME/H2O at reflux60
provided triprotected pyranicin in good yield (89%).
A final stage global deprotection would then furnish pyranicin
(1). Using BF3‚Et2O in CH2Cl2 and gradually increasing the
temperature from 0 °C to room temperature resulted in rapid
complete removal of the TBS and (TMS)(CH2)2 ethers but left
the TBDPS group untouched, and a further increase of tem-
perature resulted in decomposition of the substrate. We also
evaluated a protocol of in-situ generation of HF from BF3‚Et2O
and salic aldehyde61 with similar results. On the other hand,
the use of HF (40% aq) in MeCN at room temperature resulted
in rapid selective removal of the TBS and TBDPS groups
CHCl3). A sample of pyranicin was not available to us for
purposes of direct comparison to establish the origin of this
discrepancy. To ensure that the butenolide stereocenter was not
epimerized, we employed Figade`res method.63
End Game toward Pyragonicin. The Carreira methodology
has been used for the construction of 1,4-diols with unprotected
propargylic hydroxyl groups present.64 A direct stereoselective
coupling of 50b and 74 would give the pyragonicin core while
at the same time installing the C13 stereocenter. To circumvent
the possibility of epimerization of the butenolide moiety, we
chose to use the noneliminated lactone 77 for the initial studies
(Scheme 15). At 60 °C the coupling of 50b and 77 proceeded
as expected using an excess of reagents.65 Full conversion of
the aldehyde 50b was achieved; however, the pyragonicin core
78 was isolated as an inseparable mixture with remaining lactone
77. Since the product was obtained as a mixture of stereoisomers
at C2, C33, and the C20 olefin, an exact quantification of the
selectivity could not be accomplished. Hydrogenation of the
mixture of 78 and 77 resulted in a still inseparable mixture of
saturated products. The crude mixture from the hydrogenation
was then exposed to the same elimination conditions as were
used in the final step in the synthesis of butenolide 74 (vide
(57) Marshall, J. A.; Jiang, H. J. Org. Chem. 1999, 64, 971-975.
(58) Schreiber, S. L. Tetrahedron Lett. 1980, 21, 1027-1030.
(59) Mild conditions such as TBDMSOTf, pyridine, and 4-DMAP have
been shown to cause partial epimerisation;63 see also: (b) Yu, Q.; Wu, Y.;
Wu, Y.-L.; Xia, L.-J.; Tang, M.-H. Chirality 2000, 12, 127-129. (c) Duret,
P.; Figadere, B.; Hocquemiller, R.; Cave, A. Tetrahedron Lett. 1997, 38,
8849-8852.
(62) One might speculate that the very low concentration used in this
analysis would render the outcome sensitive to experimental error and
impurities.
(63) Latypov, S.; Franck, X.; Jullian, J.-C.; Hocquemiller, R.; Figadere,
B. Chem. Eur. J. 2002, 8, 5662-5666.
(64) Amador, M.; Ariza, X.; Garcia, J.; Ortiz, J. Tetrahedron Lett. 2002,
43, 2691-2694.
(60) Hart, D. J.; Hong, W. P.; Hsu, L. Y. J. Org. Chem. 1987, 52, 4665-
(65) Due to the small scale an excess of reagents was used under
otherwise standard conditions: Zn(OTf)2, 3.0 equiv; (-)-NME, 5.0 equiv;
Et3N, 10.0 equiv; butenolide 77, 1.3 equiv; and aldehyde 50b, 1.0 equiv.
4673.
(61) Mabic, S.; Lepoittevin, J. P. Synlett 1994, 851-852.
J. Org. Chem, Vol. 71, No. 5, 2006 1887

Strand et al.
SCHEME 16. End Game toward Pyragonicin
for 85% of the total amount of butenolide fragment 74 used in
the coupling. The spectroscopic properties of 2 (IR, 1H and 13
C
NMR) were identical in all respects to those reported by
Takahashi16 and McLaughlin.11 However, a strong discrepancy
was found in the optical rotation for the synthetic products
compared to that of the natural material, similar to what was
found for pyranicin; natural material, [R]25 -25.6 (c 0.008,
D
CHCl3);62 synthetic material (Takahashi), [R]24D +13.8 (c 0.11,
CHCl3); our synthetic material, [R]25D +10.5 (c 0.125, CHCl3).
To ensure that the C34 stereocenter was not epimerized in the
coupling step, we again employed Figade`re’s method63 to
confirm that within the limits of detection no epimerization had
occurred.70 Takahashi and co-workers further investigated the
origin of the discrepancy in optical rotation, by preparing the
corresponding (+)- and (-)-MTPA esters of pyragonicin.16
Discrepancies for these derivatives were found in the chemical
shift of C10, prompting the authors to synthesize 10-epi-
pyragonicin. This product had however significant deviances
from the corresponding data reported by McLaughlin. Although
derivatives of 2 clearly fit the data reported by McLaughlin
better than 10-epi-2, there remains some room for ambiguity
regarding the relative stereochemistry of the natural product.
To clarify this matter, access to the natural product would be
necessary; unfortunately, a sample was not available to us for
purposes of direct comparison.
Conclusions. In summary, we have accomplished convergent
stereoselective syntheses of the proposed structures of pyranicin
(1) and pyragonicin (2) through a common synthetic route. The
longest linear sequence to pyranicin comprises 19 isolated
intermediates starting from cyclohexadiene, with an overall yield
of 6.3%. The longest linear sequence to pyragonicin comprises
16 isolated intermediates from acrolein dimer (14) with an
overall yield of 5.5% (16 isolated intermediates, 6.4% overall
from cyclohexadiene). Highlights of the syntheses are (i) an
asymmetric HWE desymmetrization and subsequent diastereo-
selective transformations to provide access to all eight stereoi-
somers of the 2,3,6-substituted THP-ring system; (ii) a stereo-
convergent sequence to directly install a protected hydroxyl
group using an uncommon, yet highly efficient, protective group
((TMS)(CH2)2); (iii) the employment of Carreiras asymmetric
acetylide additions to construct 1,4- and 1,6-diols, including a
stereoselective coupling reaction involving sensitive function-
alities and unprotected adjacent hydroxyl groups. Further studies
toward the synthesis of related acetogenins as well as investiga-
tions of their biological properties will be reported in due course.
supra). Unfortunately we obtained, besides the desired elimina-
tion, an unwanted cyclization of the C10/C13 positions in a
71:29 ratio.66
Encouraged by the successful joining of 50b and 77, we chose
to evaluate the more straightforward direct coupling of the
completed butenolide 74 with THP-fragment 50b. It has been
shown that Et3N does not cause epimerization of butenolides
at ambient temperature in aprotic solvents;67 therefore to further
minimize the risk of epimerization, we chose to run the coupling
at ambient temperature. Pleasingly, the reaction proceeded
cleanly, albeit to 40% conversion68 using an excess of reagents,69
1.5 equiv of aldehyde 50b and 1.0 equiv of acetylene 74. It
should be noted that unreacted 74 was recovered from the
reaction mixture, unfortunately as an inseparable mixture with
the desired product product 80, which in this case prevented
recycling. The stereoselectivity could not be directly measured
due to the (E):(Z) mixture of the alkene and the presence of 74,
but at no point in the subsequent steps could we detect any
minor diastereomer attributed to the C13 stereocenter. A
coupling reaction of this kind involving unprotected propargylic
alcohols is known to be favored by high concentration and
elevated temperature;64 thus, on a larger scale, with potentially
recoverable substrates, we expect this reaction type to be an
even more efficient alternative for coupling of sensitive and
highly functionalized substrates with or without unprotected
hydroxyl functionalities.
Experimental Section
The final steps proceeded as expected, and diimide reduction
of a mixture of 80 and 74 gave a still inseparable mixture of
reduced products (Scheme 16). Global deprotection under
identical conditions as were used in the pyranicin synthesis
afforded pyragonicin (2) in 34% yield over three steps, along
with readily separated butenolide 81 (51%), thus accounting
(1S,2R)-2-(1-Methyl-1-phenylethyl)cyclohexyl (2Z,4R,7S)-4,7-
Bis{[tert-butyl(diphenyl)silyl]oxy}-8-oxooct-2-enoate (16). To a
stirred solution of phosphonate 20 71 (1.48 g, 2.94 mmol) in THF
(130 mL) was added NaHMDS (4.45 mL, 2.67 mmol, 0.6 M in
(70) For further details, see Supporting Information.
(71) Prepared using a protocol similar to that used in: Hatakeyama, S.;
Satoh, K.; Sakurai, K.; Takano, S. Tetrahedron Lett. 1987, 28, 2713-2716.
Selected data: Yield 93%; IR (film, cm-1) 2964, 2918, 1730, 1300, 1268,
1180, 1070, 960; 1H NMR (500 MHz, selected data) δ 7.31-7.28 (m, 4H),
7.16-7.11 (m, 1H), 4.81 (ddd [app td], J ) 10.5, 4.5, 1H), 4.45-4.27 (m,
4H), 2.27 (ddd, J ) 26.0, 20.5, 16 Hz, 2H), 2.12 (app td, J ) 11.5, 3.5 Hz,
1H), 1.94 (br d, J ) 13 Hz, 1H), 1.87-1.82 (m, 1H), 1.77-1.67 (m, 2H),
1.30 (s, 3H), 1.18 (s, 3H); 13C NMR (125 MHz) δ 164.1, 151.9, 128, 1,
125.3, 125.1, 123.7, 121.1, 76.3, 62.4 (qd, J ) 19.6, 5.5 Hz), 62.3 (qd, J )
19.6, 5.5 Hz), 50.5, 39.5, 39.5, 33.4, (d, J ) 14.4 Hz), 32.9, 30.0, 26.6,
25.8, 24.6, 22.1. See also: Vares, L. Ph.D. Thesis, Tartu University, Estonia,
2000.
(66) Determined by 1H NMR of the reaction crude. The isomers are
separable by flash but were not identified. One might expect that this
cyclization could be circumvented by reversing the order of reactions,
starting with elimination of 78 and then performing a diimide reduction of
the product.
(67) Marshall, J. A.; Jiang, H. J. Org. Chem. 1999, 64, 971-975.
(68) The conversion was estimated as the ratio of unreacted butenolide
74 to the coupled product 80 from the 1H NMR spectrum after filtration
through a short plug of silica to remove any remaining aldehyde.
(69) Due to the small scale, an excess of reagents was used: Zn(OTf)2,
2.0 equiv; (-)-NME, 3.0 equiv; and Et3N, 5.0 equiv.
1888 J. Org. Chem., Vol. 71, No. 5, 2006

Synthesis of Pyranicin and Pyragonicin
toluene) dropwise at -78 °C. The resulting solution was stirred
for 30 min and then transferred to a precooled solution of dialdehyde
15 (2.00 g, 3.21 mmol) in THF (70 mL). After 4 h the reaction
was quenched by addition of AcOH (1 M in MeOH) followed by
phosphate buffer (pH 7) and partitioned between EtOAc and
phosphate buffer (pH 7). The combined organic phases were dried
(MgSO4), filtered, and concentrated in vacuo to afford a crude oil.
Purification by flash chromatography (6.25% EtOAc/heptane)
furnished olefin 16, a single detected stereoisomer, as a clear oil
7.12 (m, 2H), 4.71 (dt, J ) 10.5, 4.3 Hz, 1H), 3.96 (td, J ) 9.2,
4.6 Hz, 1H minor diastereomer), 3.72 (dd, J ) 10.2, 4.8 Hz, 1H),
3.61-3.54 (m, 2H), 3.44-3.36 (m, 1H major diastereomer), 3.20
(dd, J ) 9.1, 3.5 Hz, 1H), 2.28 (dd, J ) 15.7, 9.0 Hz, 1H), 2.02
(dt, J ) 11.8, 3.4, 1H), 1.86-0.86 (m, 11H), 1.28 (s, 3H), 1.16 (s,
3H), 1.08 (s, 9H), 1.07 (s, 9H); 13C NMR (125 MHz, CDCl3) δ
170.6, 151.9, 135.97, 135.95, 135.6, 135.5, 134.4, 133.89, 133.85,
133.6, 129.6, 129.55, 129.52, 129.4, 127.7, 127.59, 127.53, 127.4,
125.3, 124.8, 78.0, 76.6, 74.4, 67.8, 67.2, 50.8, 39.6, 38.3, 33.0,
31.8, 30.2, 28.0, 27.1, 26.9, 26.8, 26.0, 24.6, 24.3, 22.6, 22.3, 19.6,
19.2; HRMS (FAB, M + H+) calcd for C55H71O5Si2 867.4840,
found 867.4852.
23
(2.00 g, 77% based on 20 72): [R]D -21.8 (c ) 0.91, CH2Cl2);
IR (film) 3072 (w), 2931 (s), 2858 (s), 1753 (s), 1710 (s), 1427
1
(s), 1191 (s), 1110 (s), 700 (s); H NMR (500 MHz, CDCl3) δ
9.53 (d, J ) 1.4 Hz, 1H), 7.63-7.70 (m, 4H), 7.62-7.55 (m, 4H),
7.49-7.42 (m, 2H), 7.43-7.35 (m, 6H), 7.35-7.19 (m, 8H), 7.27-
7.18 (m, 1H), 5.95 (dd, J ) 11.7, 8.0 Hz, 1H), 5.43-5.35 (m, 1H),
4.83 (dd, J ) 11.7, 0.7 Hz, 1H), 4.64 (dt, J ) 10.4, 4.2 Hz Hz,
1H), 4.07 (t, J ) 4.7 Hz, 1H), 2.02-1.94 (m, 1H), 1.88-0.81 (m,
12H), 1.23 (s, 3H), 1.18 (s, 3H), 1.15 (s, 9H), 1.06 (s, 9H); 13C
NMR (125 MHz, CDCl3) δ 203.7, 164.8, 151.9, 151.8, 136.25,
136.22, 136.1, 134.4, 134.2, 133.5, 133.4, 130.3, 130.0, 129.9,
128.3, 128.2, 128.1, 127.9, 127.8, 125.7, 125.4, 118.9, 78.4, 74.6,
69.8, 51.3, 40.2, 33.7, 32.5, 28.4, 27.6, 27.49, 27.42, 26.3, 26.0,
25.1, 19.8, 19.7; HRMS (ES+, M + Na+) calcd for C55H68O5Si2-
Na 887.4503, found 887.4523.
(2R,3R,6R)-3-{[tert-Butyl(diphenyl)silyl]oxy}-6-{[tert-butyl-
(diphenyl)silyl]oxymethyl}-2-[(Z)-dodec-2-enyl]tetrahydro-2H-
pyran (48a). To a stirred solution of ester 29 (100 mg, 0.116 mmol)
in CH2Cl2 (2 mL) at - 78 °C was added DIBAL-H (90.8 µL, 0.136
mmol, 1.5 M in toluene) dropwise over 5 min. The resulting mixture
was stirred for 35 min, after which a preformed (30 min) solution
of decyltriphenylphosphonium bromide (164 mg, 0.340 mmol) and
NaHMDS (0.45 mL, 0.227 mmol, 0.6 M in toluene) in THF (4
mL) at 0 °C was added via a cannula. The temperature was raised
to 0 °C over 12 h, and the reaction was then quenched by
evaporation onto silica. Purification by flash chromatography
(0.78-3.13% EtOAc/heptane) furnished olefin 48a as a clear oil
(69 mg, 75%): (E):(Z) ∼1:10; [R]D23 +19.0 (c ) 1.0, CH2Cl2); IR
(film) 3070 (w), 2927 (s), 2856 (s), 1471 (m), 1427 (m), 1112 (s),
701 (s); 1H NMR (500 MHz, CDCl3) δ 7.82-7.63 (m, 8H), 7.48-
7.29 (m, 12H), 5.39-5.21 (m, 2H), 3.82 (dd, J ) 10.3, 5.3 Hz,
1H), 3.71 (s, 1H), 3.66 (dd, J ) 10.2, 5.4 Hz, 1H), 3.51-3.42 (m,
1H), 3.17 (dd, J ) 8.3, 5.2 Hz, 1H), 2.48-2.24 (m, 1H), 2.21-
1.98 (m, 1H), 1.95-1.68 (m, 4H), 1.57-1.35 (m, 3H), 1.35-0.99
{(2S,3R,6R)-3-{[tert-Butyl(diphenyl)silyl]oxy}-6-{[tert-butyl-
(diphenyl)silyl]oxymethyl}tetrahydro-2H-pyran-2-yl}acetic Acid
(1S,2R)-2-(1-Methyl-1-phenylethyl)cyclohexyl Ester (28). To a
solution of secondary alcohol 19 (28 mg, 0.032 mmol) in a mixture
of toluene (2 mL) and t-BuOH (1.5 mL) at 0 °C was added t-BuOK
(16 µL, 0.1 mmol, 1.0 M in THF). The reaction was stirred for 15
min and then brought to room temperature. After an additional 2
h, the reaction was quenched by pouring onto water and partitioned
between EtOAc and water. The combined organic phases were dried
(MgSO4), filtered, and concentrated in vacuo. Purification by flash
chromatography (3.13-6.25% EtOAc/heptane) gave the cyclized
product 28 as a clear oil (21 mg, 75%): Diastereomeric ratio (2S:
(m, 13H), 1.10 (s, 9H), 1.09 (s, 9H), 0.90 (t, J ) 6.8 Hz, 3H); 13
C
NMR (125 MHz, CDCl3) 136.0, 135.7, 136.0, 134.5, 134.0, 133.9,
133.88, 133.81, 133.6, 132.7, 132.6, 131.1, 129.4, 128.6, 128.5,
128.48, 128.44, 128.3, 127.53, 127.47, 127.40, 127.3, 126.2, 80.5,
78.1, 68.0, 67.3, 31.9, 30.7, 30.5, 29.6, 29.57, 29.56, 29.35, 29.33,
27.4, 27.1, 26.8, 22.6, 22.5, 19.6, 19.2, 14.1; HRMS (FAB, M +
Na+) calcd for C50H70NaO3Si2 797.4761, found 797.4762.
23
2R) g 98:2; [R]D -8.3 (c ) 0.67, CH2Cl2); IR (film) 3070 (m),
2931 (s), 2858 (m), 1727 (s), 1471 (m), 1427 (m), 1110 (s), 1031
1
(m); H NMR (400 MHz, CDCl3) δ 7.69-7.61 (m, 8H), 7.47-
(1R)-1-{(2R,5R,6R)-5-{[tert-Butyl(diphenyl)silyl]oxy}-6-[(Z)-
dodec-2-en-1-yl]tetrahydro-2H-pyran-2-yl}-3-(trimethylsilyl)-
prop-2-yn-1-ol (52). To a solution of zinc triflate (97 mg, 0.256
mmol) and (1S, 2R)-N-(-)-methylephedrine (57 mg, 319 mmol)
in toluene (1 mL) was added Et3N (88.8 µL, 0.637 mmol). The
resulting slurry was stirred 1 h 45 min, and the trimethylsilylacety-
lene (150 µL, 1.06 mmol) was added. After 15 min a solution of
aldehyde 50a (120 mg, 0.212 mmol) in toluene (1 mL) was added
via a cannula (rinsed with 0.5 mL toluene).73 After stirring for 20
h at room temperature the reaction mixture was evaporated onto
silica. Purification by flash chromatography (6.25% EtOAc/heptane)
7.32 (m, 12H), 7.23-7.12 (m, 4H), 7.08-7.02 (m, 1H), 4.70 (dt,
J ) 10.5, 4.3 Hz, 1H), 3.92 (td, J ) 9.6, 4.7 Hz, 1H), 3.72-3.63
(m, 2H), 3.59-3.50 (m, 1H), 3.40 (m, 1H), 2.05 (dd, J ) 14.4, 4.3
Hz, 1H), 1.96 (dt, J ) 11.5, 3.3 Hz, 1H), 1.86-0.81 (m, 13H),
1.24 (s, 3H), 1.16 (s, 3H), 1.05 (s, 9H), 1.04 (s, 9H); 13C NMR
(100 MHz, CDCl3) δ 171.0, 151.9, 136.24, 136.23, 136.06, 136.00,
134.6, 134.2, 134.0, 133.9, 130.08, 130.02, 130.00, 128.2, 128.0,
127.9, 125.8, 125.3, 75.0, 74.9, 71.5, 70.4, 65.2, 51.1, 40.2, 37.4,
33.6, 27.9, 27.5, 27.3, 27.2, 26.3, 25.6, 25.0, 23.7, 19.68, 19.64;
HRMS (FAB, M + Na+) calcd for C55H70NaO5Si2 889.4659, found
889.4668.
23
afforded propargylic alcohol 52 as a clear oil (110 mg, 83%); [R]D
+19.8 (c ) 1.0, CH2Cl2); IR (film) 3478 (br, w), 2956 (s), 2927
(s), 2856 (s), 2177 (w), 1427 (w), 1249 (m), 1110 (s), 842 (s), 701
(s); 1H NMR (500 MHz, CDCl3) δ 7.78-7.65 (m, 4H), 7.50-7.37
(m, 6H), 5.43-5.20 (m, 2H), 4.32 (d, J ) 7.7 Hz, 1H), 3.76-3.70
(m, 1H), 3.40 (ddd, J ) 10.1, 7.9, 2.0 Hz, 1H), 3.23 (dd, J ) 9.1,
4.3 Hz, 1H), 2.98 (s, 1H), 2.53-2.44 (m, 1H), 2.02-1.68 (m, 5H),
1.67-1.52 (m, 2H), 1.41-0.99 (m, 14H), 1.13 (s, 9H), 0.92 (t, J
) 6.9 Hz, 3H), 0.20 (s, 9H); 13C NMR (125 MHz, CDCl3) δ 136.4,
134.2, 132.1, 130.0, 128.00, 127.96, 126.0, 103.4, 91.2, 81.1, 80.7,
68.2, 66.8, 32.3, 30.9, 30.8, 30.06, 30.00, 29.79, 29.76, 27.8, 27.5,
23.1, 22.1, 20.0, 14.5, 0.2; HRMS (FAB, M + Na+) calcd for
C39H61NaO3Si2 633.4159, found 633.4153.
{(2R,3R,6R)-3-{[tert-Butyl(diphenyl)silyl]oxy}-6-{[tert-butyl-
(diphenyl)silyl]oxymethyl}tetrahydro-2H-pyran-2-yl}acetic Acid
(1S,2R)-2-(1-Methyl-1-phenylethyl)cyclohexyl Ester (29). To a
solution of secondary alcohol 19 (or a mixture of secondary/primary
alcohol 19:19a (92:8)) (440 mg, 0.597 mmol) in toluene (10 mL)
at 0 °C was added t-BuOK (99 µL, 0.1 mmol, 1.0 M in THF)
dropwise over 5 min. The reaction was stirred for 50 min, then
quenched by addition of phosphate buffer (pH 7), and partitioned
between EtOAc and phosphate buffer (pH 7). The combined organic
phases were dried (MgSO4), filtered, and concentrated in vacuo to
afford the cyclized product 29 as a clear oil, pure by NMR (440
23
mg, quant.): Diastereomeric ratio (2R:2S) ) 96:4; [R]D -0.8 (c
) 1.0, CH2Cl2); IR (film) 3070 (w), 2931 (s), 2851 (s), 1725 (s),
1427 (m), 1110 (s), 701 (s); 1H NMR (500 MHz, CDCl3) δ 7.75-
7.65 (m, 8H), 7.50-7.34 (m, 13 H), 7.26-7.22 (m, 2H), 7.18-
(1R,2S,5R)-5-Methyl-2-(1-methyl-1-phenylethyl)cyclohexyl (2E)-
3-[(2R)-3,4-Dihydro-2H-pyran-2-yl]acrylate (57) and (1R,2S,5R)-
5-Methyl-2-(1-methyl-1-phenylethyl)cyclohexyl (2Z)-3-[(2S)-3,4-
(72) Alternatively, the crude 16 obtained after filtration through a short
plug of silica could be directly subjected to reduction/protective group
migration. This protocol afforded 18 in 70% overall yield, based on 20.
(73) Due to the higher reactivity of aldehyde 50a compared to that of
aldehyde 70, the reaction could be run at room temperature using standard
inert techniques.
J. Org. Chem, Vol. 71, No. 5, 2006 1889

Strand et al.
Dihydro-2H-pyran-2-yl]acrylate (61). To a stirred solution of
phosphonate 55 (1.97 g, 3.80 mmol) and 18-crown-6 (2.61 g, 9.88
mmol) in THF (150 mL) was added KHMDS (7.24 mL, 3.62 mmol,
0.5 M in toluene) dropwise at -78 °C. The resulting solution was
stirred for 30 min and then added via a cannula to a precooled
solution of acrolein dimer rac-14 (533 mg, 4.94 mmol) in THF
(70 mL) over 5 h at -78 °C. After an additional 2 h the reaction
was quenched by addition of AcOH (1 M, MeOH) followed by
phosphate buffer (pH 7) and partitioned between EtOAc and
phosphate buffer (pH 7). The combined organic phases were dried
(MgSO4), filtered, and concentrated in vacuo to afford a crude oil.
Purification by flash chromatography (3.13-6.25% EtOAc in
heptane) afforded a 60:40 mixture of (2Z,4S)-61 and (2E,4R)-57
as a clear oil (1.54 g, 77% based on 55): dr 61, (2Z,4S):(2Z,4R) )
96:4; 57, (2E,4R):(2E,4S) ) 98:2.74
CH2Cl2); IR (film) 3436 (br, s), 2952 (s), 2929 (s), 2961 (s), 2169
(w), 1727 (s), 1249 (s), 1174 (m), 840 (s); H NMR (500 MHz,
1
CDCl3) δ 7.33-7.27 (m, 4H), 7.18-7.11 (m, 1H), 4.84 (dt, J )
10.7, 4.4 Hz, 1H), 4.38 (t, J ) 6.5 Hz, 1H), 3.52-3.40 (m, 2H),
3.17-3.08 (m, 1H), 2.06-1.98 (m, 1H), 1.97-0.81 (m, 27H), 1.34
(s, 3H), 1.23 (s, 3H), 0.20 (s, 9H), 0.03 (s, 9H); 13C NMR (125
MHz, CDCl3) δ 173.5, 152.0, 128.3, 125.8, 125.4, 107.2, 89.7, 78.3,
74.3, 66.3, 63.2, 50.7, 42.2, 40.1, 38.1, 34.9, 34.2, 31.7, 30.7, 29.8,
29.1, 28.1, 27.0, 25.7, 25.6, 25.5, 22.2, 19.0, 0.3, -0.8; HRMS
(FAB, M + H+) calcd for C36H63O4Si2 615.4265, found 615.4269.
(5S)-3-{(2R,8R)-8-Hydroxy-2-[2-(trimethylsilyl)ethoxy]dec-9-
yn-1-yl}-5-methylfuran-2(5H)-one (74). To a solution of ester 71
(100 mg, 0.17 mmol) in THF (5 mL) at -78 °C was added LDA
(249 µL, 0.499 mmol, 2 M). The resulting mixture was stirred for
35 min and a precooled solution (-78 °C) of 76 (131 mg, 0.83
mmol) in THF (3 mL) was added dropwise via a cannula. After
stirring for 1 h the reaction was quenched by addition of MeOH (2
mL). The reaction was then brought to room temperature, and K2-
CO3 (200 mg, 1.44 mmol) was added. The resulting suspension
was stirred another 12 h after which the reaction mixture was poured
into HCl (1 M, aq). The mixture was partitioned between EtOAc
and HCl (1 M, aq). The combined organic phases were then dried
(MgSO4) and concentrated in vacuo. The crude product 72 was
dissolved in MeOH and 10-camphorsulfonic acid (cat.) was added.
The resulting mixture was heated to reflux. After 60 min the reaction
was cooled to room temperature and partitioned between EtOAc
and NaHCO3(sat.,aq). The combined organic phases were dried
(MgSO4) and filtered through a short plug of silica (50% EtOAc/
heptane) to remove (-)-8-phenylmenthol. The resulting crude oil
(59.9 mg) was isolated as a mixture of diastereomers. Of this crude
lactone 44 mg was dissolved in CH2Cl2 (3 mL) and Et3N (159.2
µL, 0.572 mmol) was added, followed by trichloroacetyl chloride
(38.3 µL, 0.343 mmol) at 0 °C. The resulting mixture was stirred
at room temperature for 24 h after which THF (6 mL) followed by
NaHCO3 (5 mL, sat., aq) was added. The resulting mixture was
stirred for 3 h and then partitioned between EtOAc and water. The
combined organic phases were dried (MgSO4), filtered, and
concentrated in vacuo to afford a crude oil. Purification by flash
chromatography (12.5-50% EtOAc/heptane) afforded butenolide
74 as a clear oil (33 mg, 79% overall from 71): [R]D23 +17.7 (c )
0.82, CH2Cl2); IR (film) 3430 (br, m), 3297 (w), 2935 (s), 2859
(m), 1752 (s), 1319 (m), 1247 (m), 1076 (s), 1076 (s), 837 (s); 1H
NMR (500 MHz, CDCl3) δ 7.16 (d, J ) 1.3 Hz, 1H), 5.08-5.00
(m, 1H), 4.39 (dq, J ) 6.6, 2.1 Hz, 1H), 3.58-3.46 (m, 3H), 2.50-
2.45 (m, 3H), 1.85 (d, J ) 5.6 Hz, 1H), 1.80-1.66 (m, 2H), 1.58-
1.25 (m, 11H), 0.99-0.82 (m, 2H), 0.02 (s, 9H); 13C NMR (125
MHz, CDCl3) δ 174.0, 151.4, 130.7, 84.9, 77.5, 76.6, 72.8, 66.3,
62.2, 37.5, 34.0, 29.8, 29.2, 25.1, 24.8, 19.1, 18.6, -1.3; HRMS
(FAB, M + H+) calcd for C20H35O4Si 367.2305, found 367.2308.
Pyranicin (1). To a stirred solution of triprotected pyranicin 84
(12.0 mg, 0.0114 mmol) in MeCN (1.2 mL) was added HF (50
µL, 40%, aq) at room temperature. The reaction was heated to 45
°C during 22 h and then quenched by addition of NaHCO3 (sat.,
aq) and partitioned between EtOAc and NaHCO3. The combined
organic phases were dried (MgSO4), filtered, and concentrated in
vacuo. Purification by flash chromatography (3-5% MeOH/EtOAc)
afforded unprotected pyranicin (1) as a clear oil (5.7 mg, 85%):
[R]D23 +21.1 (c ) 0.24, CHCl3); IR (film) 3392 (br, m), 2921 (s),
2850 (s), 1757 (m), 1743 (m), 1644 (m), 1467 (m), 1321 (m), 1205
(w), 1079 (s), 1027 (m); 1H NMR (500 MHz, CDCl3) δ 7.19 (dd,
2.5, 1.2 Hz, 1H), 5.07 (dq, J ) 6.8, 1.3 Hz, 1H), 3.89-3.82 (m,
1H), 3.65-3.58 (m, 2H), 3.49-3.43 (m, 1H), 3.35 (dd, J ) 8.1,
5.6 Hz, 1H), 3.20 (ddd, J ) 10.8, 7.0, 2.2 Hz, 1H), 2.69 (s, 1H),
2.54 (tdd, J ) 15.1, 3.1, 1.4 Hz, 1H), 2.41 (tdd, J ) 15.2, 8.3, 1.2
Hz, 1H), 2.3 (br s, 1H), 2.01 (ddd, J ) 13.0, 5.9, 3.0 Hz, 1H), 1.84
(d, J ) 8.2 Hz, 1H), 1.77-1.10 (m, 45H), 1.44 (d, J ) 6.8 Hz,
10-Benzyl 1-[(1R,2S,5R)-5-Methyl-2-(1-methyl-1-phenylethyl)-
cyclohexyl] (2E,4R,8E)-4-[2-(trimethylsilyl)ethoxy]deca-2,8-di-
enedioate (65). Neocuproine (427 mg, 0.608 mmol) and Pd2dba3
(15.7 mg, 0.015 mmol) were dissolved in CH2Cl2 (1 mL) at room
temperature and stirred for 30 min. The resulting clear orange-red
solution was transferred to a stirred solution of 60 or 64 (or a
mixture of the two), dissolved in 2-(trimethylsilyl)ethanol (2 mL)
and CH2Cl2 (1 mL) at room temperature. The resulting clear yellow
solution was then stirred at room temperature for 3 h during which
time the color changed to light brown. The reaction mixture was
poured onto phosphate buffer (pH 7) and partitioned between CH2-
Cl2 and phosphate buffer (pH 7). The combined organic phases
were dried (MgSO4) followed by removal of CH2Cl2 under reduced
pressure. Recovery of 2-(trimethylsilyl)ethanol was accomplished
using bulb-to-bulb distillation (0.01 mmHg, 60 °C) to give 1.50 g,
80%. Repeated purification by flash chromatography (10% Et2O/
heptane) furnished the (2E,8E)-diene 65 as a clear oil (263.4 mg,
23
72%): Diastereomeric ratio (4R):(4S) ) 97:3; [R]D +16.4 (c )
1.0, CH2Cl2); IR (film) 3058 (w), 2952 (s), 2925 (s), 1714 (s), 1652
(m), 1440 (m), 1230 (m), 1230 (s), 1199 (s); 1H NMR (400 MHz,
CDCl3) δ 7.40-7.20 (m, 9H), 7.13-7.07 (m, 1H), 7.01 (td, J )
15.6, 6.9 Hz, 1H), 6.53 (dd, J ) 15.8, 6.5 Hz, 1H major
diastereomer), 6.35 (dd, J ) 15.8, 6.3 Hz, 1H minor diastereomer),
5.89 (td, J ) 15.6, 1.5, 1H), 5.44 (dd, J ) 15.8, 1.1 Hz, 1H), 5.18
(s, 2H), 4.87 (dt, J ) 10.7, 4.4 Hz, 1H), 3.70 (q, J ) 5.5 Hz, 1H),
3.49 (ddd, J ) 10.1, 9.4, 6.1 Hz, 1H), 3.29 (ddd, J ) 10.2, 9.4, 6.3
Hz, 1H), 2.26-2.18 (m, 2H), 2.03 (ddd, J ) 12.3, 10.6, 3.2 Hz,
1H), 1.96-1.89 (m, 1H), 1.76-0.74 (m, 12H), 1.31 (s, 3H), 1.23
(s, 3H), 0.87 (d, J ) 6.5 Hz, 3H), 0.00 (s, 9H); 13C NMR (100.6
MHz, CDCl3) δ 167.7, 166.7, 152.7, 150.6, 149.3, 137.4, 129.8,
129.55, 129.50, 129.2, 126.7, 126.3, 123.6, 122.7, 79.4, 75.9, 67.9,
67.4, 51.8, 43.0, 41.1, 35.9, 35.8, 33.4, 32.6, 28.3, 28.0, 27.4, 25.1,
23.1, 19.7, 0.0; HRMS (FAB, M + H+) calcd for C38H54O5Si
619.3819, found 619.3822.
(1R,2S,5R)-5-Methyl-2-(1-methyl-1-phenylethyl)cyclohexyl (4R,-
10R)-10-Hydroxy-12-(trimethylsilyl)-4-[2-(trimethylsilyl)ethoxy]-
dodec-11-ynoate (71). To a solution of zinc triflate (79.7 mg, 0.21
mmol) and (1R,2S)-N-(+)-methylephedrine (46.9 mg, 0.26 mmol)
in toluene (1 mL) was added Et3N (73 µL, 0.52 mmol) through a
septum.75 The resulting slurry was stirred 1 h 45 min and then
trimethylsilylacetylene (246 µL, 1.744 mmol) was added. After 15
min a solution of aldehyde 70 (90 mg, 0.17 mmol) in toluene (1
mL) was added via a cannula (rinsed with 0.5 mL of toluene). The
reaction vessel was then sealed with a screw cap and heated to 60
°C. After 20 h the reaction mixture was evaporated onto silica.
Purification by flash chromatography (12.5-25% EtOAc/heptane)
afforded propargylic alcohol 71 as a clear oil (80 mg, 75%):
23
Diastereomeric ratio (10R:10S) ) 98:2;76 [R]D -2.8 (c ) 1.0,
(74) For analytical data of 57 and 61, see: Pedersen, T. M.; Jensen, J.
F.; Humble, R. E.; Rein, T.; Tanner, D.; Bodmann, K.; Reiser, O. Org.
Lett. 2000, 2, 535-538.
(75) Sealing the reaction vessel using a septum was not sufficient to
prevent TMS-acetylene from escaping from the reaction.
(76) As determined by 1H and 19F analysis of the corresponding (+)-
and (-)-MTPA derivatives.
1890 J. Org. Chem., Vol. 71, No. 5, 2006

Synthesis of Pyranicin and Pyragonicin
23
3H), 0.89 (t, J ) 6.9 Hz, 3H); 13C NMR (125 MHz, CDCl3) δ
174.5, 151.8, 131.1, 81.2, 80.0, 77.9, 74.0, 71.7, 69.9, 66.1, 37.35,
37.30, 37.2, 33.4, 32.2, 31.9, 31.6, 30.5, 29.67, 29.66, 29.65, 29.63,
29.60, 29.57, 29.50, 29.3, 25.63, 25.60, 25.51, 25.3, 22.6, 21.5,
19.1, 14.1; HRMS (FAB, M + H+) calcd for C35H65O7 597.4730,
found 597.4732.
oil (4.6 mg, 34%): [R]D + 10.3 (c ) 0.13, CHCl3); IR (film)
3409 (br, m), 2915 (s), 2848 (s), 1741 (s), 1461 (m), 1317 (w),
1081 (m); H NMR (500 MHz, CDCl3) δ 7.19 (q, J ) 1.5 Hz,
1
1H), 5.07 (qq, J ) 6.7, 1.2 Hz, 1H), 3.88-3.81 (m, 1H), 3.69-
3.58 (m, 2H), 3.51 (dt, J ) 7.5, 2.5 Hz, 1H), 3.36 (dd, J ) 7.6, 5.8
Hz, 1H), 3.24 (ddd, J ) 10.5, 7.4, 2.3 Hz, 1H), 2.53 (dt, J ) 15.1,
3.2, 1.6, 1H), 2.40 (ddt J ) 15.1, badly resolved, 1H), 2.02 (ddd,
J ) 13.4, 5.7, 3.0 Hz, 1H), 1.83-1.18 (m, 47H), 1.44 (d, J ) 6.8
Hz, 3H), 0.88 (t, J ) 6.9 Hz, 3H); 13C NMR (125 MHz, CDCl3) δ
174.6(weak), 151.8, 131.2, 81.0, 80.1, 78.0, 74.3, 71.6, 69.9, 66.0,
37.3, 37.2, 33.4, 33.3, 31.9, 31.6, 30.5, 29.69, 29.66, 29.64, 29.61,
29.5, 29.4, 29.3, 28.4, 25.63, 25.61, 25.5, 22.7, 21.6, 19.1, 14.1;
HRMS (FAB, M + H+) calcd for C35H65O7 597.4730, found
597.4739.
Pyragonicin (2). To a solution of zinc triflate (16.6 mg, 0.046
mmol) and (1S,2R)-N-(-)-methylephedrine (12.3 mg, 0.068 mmol)
in toluene (100 µL) was added Et3N (16 µL, 0.11 mmol). The
resulting slurry was stirred for 1 h 45 min, and then butenolide 74
(8.4 mg, 0.022 mmol) in toluene (100 µL) was added. After 15
min, a solution of aldehyde 50b (19.3 mg, 0.034 mmol) in toluene
(100 µL) was added. The reaction was stirred for 20 h, and the
reaction mixture was filtered through a plug of silica (25-37.5%
EtOAc/heptane) to remove remaining aldehyde. The crude oil
(∼10.0 mg) was dissolved in DME (2.4 mL) and tosylhydrazine
(300 mg, 1.6 mmol) was added. The reaction was heated to reflux,
and sodium acetate (160 mg, 2.0 mmol) in water (3.0 mL) was
added over 4 h using a syringe pump. The reaction was then poured
onto water and partitioned between EtOAc and water. The combined
organic phases were dried (MgSO4), filtered, concentrated in vacuo,
and filtered through a plug of silica (25-37.5% EtOAc/heptane).
The remaining crude oil was dissolved in MeCN (2.5 mL) and HF
(50 µL, 40%, aq) was added at room temperature. The reaction
was heated to 40 °C during 22 h and then quenched by addition of
NaHCO3(sat.,aq) and partitioned between EtOAc and NaHCO3. The
combined organic phases were dried (MgSO4), filtered, and
concentrated in vacuo. Purification by flash chromatography (0-
5% MeOH/EtOAc) afforded unprotected pyragonicin (2) as a clear
Acknowledgment. We thank AstraZeneca R&D So¨derta¨lje
and Aulin-Erdtman foundation for financial support. Professor
Paul Helquist and the Helquist group are gratefully acknowl-
edged for support during a research visit of D.S. at the University
of Notre Dame.
Supporting Information Available: Characterization data and
experimental procedures for compounds 18, 82, 19, 86, 22, 27, 48b,
49a,b, 50a,b, 53, 54, 58, 62, 59, 63, 60, 64, 66, 69, 83, 70, 76, 84,
and 81. H and 13C NMR spectra for compounds 16, 28, 29, 48a,
1
52, 65, 71, 74, 1, and 2. This material is available free of charge
via the Internet at http://pubs.acs.org.
JO052233K
J. Org. Chem, Vol. 71, No. 5, 2006 1891