Synthesis of the functionalized tricyclic core of lactonamycin by oxidative dearomatization.

Article abstract of DOI:10.1021/ol006531+

[reaction: see text] We report in this Letter a synthesis of the densely oxygenated CDEF ring system (27) corresponding to that found in the recently discovered antibiotic lactonamycin. The key steps in the synthesis consist of an intramolecular Wessely oxidative lactonization of acid 18, followed by a hydroxyl-directed epoxidation of enol ether 21.

Full text of DOI:10.1021/ol006531+

ORGANIC
LETTERS
2000
Vol. 2, No. 22
3493-3496
Synthesis of the Functionalized Tricyclic
Core of Lactonamycin by Oxidative
Dearomatization
,†,‡
Christopher Cox† and Samuel J. Danishefsky*
Department of Chemistry, Columbia UniVersity, HaVemeyer Hall,
New York, New York 10027, and Laboratory for Bioorganic Chemistry,
Sloan-Kettering Institute for Cancer Research, 1275 York AVenue,
New York, New York 10024
s-danishefsky@ski.mskcc.org
Received August 31, 2000
ABSTRACT
We report in this Letter a synthesis of the densely oxygenated CDEF ring system (27) corresponding to that found in the recently discovered
antibiotic lactonamycin. The key steps in the synthesis consist of an intramolecular Wessely oxidative lactonization of acid 18, followed by
a hydroxyl-directed epoxidation of enol ether 21.
The isolation of lactonamycin (Figure 1) from Streptomycies
rishirienes was recently described by Mastumoto and col-
leagues.1 Of primary interest at the biological level is its
powerful antimicrobial activity against both methacilin- and
vancomycin-resistant organisms.
report a solution to this problem in the context of the
synthesis of a tetracyclic construct.
Given the potential importance of lactonamycin as a
prospective matrix for a new class of antibiotics, we were
particularly desirous of a concise strategy which would gain
rapid access to the most functionality-laden section of the
Viewed from the perspective of chemical synthesis, the
structure of lactonamycin inherently poses issues of some
scope. The stereospecific attachment of a 2-deoxy sugar to
a deactivated tertiary acceptor constitutes a challenge at the
frontier of difficult glycosylations. Moreover, formation of
the hexacyclic aglycone ensemble, even in a gross sense, is
far from a simple matter. Particularly challenging is the
densely oxygenated DEF substructure. In this Letter, we
† Columbia University.
‡ Sloan-Kettering Institute for Cancer Research.
(1) (a) Isolation and biological evaluation: Matsumoto, Y.; Tsuchida,
T.; Maruyama, M.; Kinoshita, N.; Homma, Y.; Iinuma, H.; Sawa, T.;
Hamada, M.; Takeuchi, T.; Heida, N.; Yoshioka, T. J. Antibiot. 1999, 52,
269. (b) Structure determination: Matsumoto, Y.; Tsuchida, T.; Nakamura,
H.; Sawa, R.; Takahashi, Y.; Naganawa, H.; Inuma, H.; Sawa, T.; Takeuchi,
T.; Shiro, M. J. Antibiot. 1999, 52, 276.
Figure 1.
10.1021/ol006531+ CCC: $19.00 © 2000 American Chemical Society
Published on Web 10/04/2000

molecule. An approach to this domain, which suggested itself
in idealized form, is summarized in the progression of 2 f
3 f 4 (Scheme 1). A dihydroisobenzofuran system, properly
ment of the DEF substructure, hoping that it could be
introduced after assembly of the ring system was complete.
In this vein, treatment of 6 with triethylsilane-trifluor-
acetic acid led to reduction of the hemiacetal and deprotection
of the ester moiety. Following debenzylation, the Wessely
precursor 9b was in hand. Indeed, treatment of this com-
pound with Pb(OAc)4 afforded 10, albeit in low yield. The
poor outcome of this reaction can be attributed to the
vulnerability of the product, which was not stable to
purification on silica, Florisil or even basic alumina.5
Having gained some small measure of confidence as to
the feasibility of our route, we pursued a more advanced
model comprising the CDEF sector. We hoped that the benzo
equivalent of ring C would help confer greater stability to
the otherwise labile pre-DEF intermediates. Starting with the
known6 phosphonate 11, phthalide 12 was constructed by a
tandem Michael addition-cyclization sequence (Scheme 3).7
Scheme 1
substituted with geminal methoxy and acetic acid substituents
(cf. 2), was expected to undergo intramolecular Wessely
oxidation2,3 to produce 3 with the desired cis-fused EF
assembly. This step might, in turn, set the stage for oxidation
of the enol ether linkage by attack of an oxidizing agent
(Ox+) anti to the angular methoxy group, to give rise to the
cis-fused DE system with ring D presented as the 1,4
diketone. Such a scenario would be in keeping with a
growing theme in our laboratory, which seeks to exploit
oxidative dearomatizations toward various synthetic ends.
Scheme 3a
In our earliest attempts at reduction to practice, we
identified a system of type 9a as a subgoal (Scheme 2).
Scheme 2a
a (a) LiOtBu, THF, then 11, then 2(5H)-furanone (80%); (b)
Me2SO4, K2CO3, acetone, reflux (91%); (c) B-iodo-9-BBN, CH2Cl2,
reflux (79%); (d) BnBr, K2CO3, acetone, reflux (97%); (e) LDA,
tBuOAc, then 15; (f) TES, TFA, CH2Cl2 (70% for two steps); (g)
1,4-cyclohexadiene, 10% Pd/C, EtOH (97%); (h) Pb(OAc)4, CH2Cl2
(74%); (i) DMDO, CH2Cl2 (80% of 20b); TES ) triethylsilane;
TFA ) trifluoroacetic acid.
t
a (a) LDA, BuOAc, then 5 (85%); (b) Amberlite IR 120(plus),
MeOH, 5 min (88%); (c) 6, TES, TFA, CH2Cl2 (95%); (d) H2, 5%
Pd/C, MeOH (99%); (e) 9b, Pb(OAc)4, CH2Cl2 (see ref 5); TES )
triethylsilane; TFA ) trifluoroacetic acid.
After differentiation of the phenolic ethers, the general theme
Unfortunately, we were unable to reach this system via esters
corresponding to 6 or 7, due to the very high lability of the
benzylic, anomeric-like hydroxy and methoxy functions.4
Accordingly, we forfeited, for the moment, the incorporation
of the angular methoxy functionality prior to the establish-
of Scheme 2 was followed to reach Wessely precursor 18.
(3) Although oxidative dearomatizations of advanced intermediates with
Pb(OAc)4 are scarce in the chemical literature, the analogous use of
hypervalent iodine reagents is well studied. For the use of I(III) reagents in
synthesis, see: (a) Varvoglis, A. Tetrahedron 1997, 53, 1179. (b) Kitamura,
T.; Fujiwara, Y. Org. Prep. Proc. Int. 1997, 29, 409. (c) Wirth, T.; Hirt, U.
H. Synthesis 1999, 1271.
(2) The Wessely oxidation was introduced originally as a way to access
R-acetoxy ketones by the oxidative dearomatization of phenols in acetic
acid with Pb(OAc)4, see: (a) Wessely, F.; Sinwell, F. Montasch. Chem.
1950, 81, 1055. This venerable reaction was advanced by the pioneering
work of Yates to include intermolecular trapping by various carboxylic acids
in inert solvents, see: (b) Yates, P.; Auksi, H. J. Chem. Soc., Chem.
Commun. 1976, 1016. (c) Yates, P. Auksi, H. Can. J. Chem. 1979, 57,
2853.
(4) The esters 6 and 7 were extremely sensitive to both acid and base
due to facile formation of the corresponding R,â-unsaturated ester, which
also comes into conjugation with the aromatic ring. Acetal 7 was prepared
with a number of other ester protecting groups, including allyl, benzyl,
2-(trimethylsilyl)ethyl, and tert-butyldiphenylsilyl. However, these too could
not be removed without concurrent destruction of the acetal. Direct addition
of the dianion of acetic acid also failed due to competing deprotonation of
the highly acidic benzylic protons in 5.
3494
Org. Lett., Vol. 2, No. 22, 2000

Upon exposure to Pb(OAc)4 in CH2Cl2, 18 underwent
smooth, stereospecific oxidative cyclization to provide the
stable lactone 19 in 74% yield. Not surprisingly, treatment
of 19 with DMDO cleanly led to a hydroxyketone (80%
yield). The diastereofacial selectivity of this reaction was in
excess of 95% as judged by proton NMR analysis.
Various arguments and counter arguments could be posited
in anticipation of the likely course of facial preference in
this epoxidation.8 In fact, X-ray analysis revealed the major
isomer to be the undesired 20b, wherein the DMDO oxidant
had approached the enol ether from its â-face.9 While we
could improve the ratio of 20a to 20b by altering epoxidation
reagents, we could not render it the major product in
satisfactory yield.10
We were not unmindful that, in principle, diasteromers
20a and 20b could be interconverted through a retroaldol-
realdol sequence (see arrow).11 In practice, however, attempts
to realize this possibility were not successful. At this juncture,
we sought to exploit the directing ability of the free hydroxyl
moiety12 which would be unveiled if the γ-lactone could be
opened. Accordingly, compound 19 was treated with lithium
methoxide, thereby affording the allylic alcohol 21 (Scheme
4). The success of this reaction under such mild conditions
mixture favoring the desired R-diastereomer.13 By performing
the epoxidation in benzene, the ratio was increased to 7:1.
However, the greatest success was realized through the use
of trifluoroperacetic acid (TFPAA), which provided a 20:1
mixture of 20a:20b. In this fashion, 20a became available
as a single isomer in 51% yield from 21, following
recrystallization (mp ) 173 °C).
Ultimately, it would be our goal to incorporate and
maintain the angular methoxy group prior to construction
of the CDEF system. However, for the present, we undertook
its installation post facto. GiVen the strain of the γ-lactone
discussed aboVe, it was anticipated that installation of R,â-
unsaturation would enable smooth Michael-type addition of
MeOH. In pursuing this conjecture, we attempted to generate
the discrete enolate of TMS-protected lactone 22 (Scheme
5). Such efforts, instead, resulted in opening of the E ring
Scheme 5a
Scheme 4a
a (a) LiHMDS, MeOH, CH2Cl2, -35 °C (51% + 40% recovered
19); (b) TFPAA, Na2CO3, CH2CL2, 0 °C; (c) TsOH, benzene, reflux
(51% for two steps). LiHMDS ) lithium bis(trimethylsilyl)amide;
TFPAA ) trifluoroperacetic acid.
a (a) HMDS, imidazole, cat. TMSCl, CH2Cl2 (84%); (b) 20b,
TMSCl, THF, then LiHMDS, then NIS (86%; see ref 15); (c)
DMDO, CH2Cl2 (71% + 17% recovered 23); (d) Tf2O, Hu¨nig’s
base, CH2Cl2 (60%); (e) CSA, MeOH (88%); (f) LiI, THF, HOAc,
reflux (55%). HMDS ) 1,1,1,3,3,3-hexamethyldisilazane; Tf2O )
trifluoromethanesulfonic anhydride; CSA ) camphorsulfonic acid.
probably reflects significant strain associated with the
γ-lactone in the context of this tetracyclic construct. In the
event, treatment of 21 with mCPBA in CH2Cl2, followed
by acid-catalyzed ring closure, provided ketols 20 as a 3:1
via â-elimination. However, upon addition of LiHMDS to
lactone 20 in the presence of TMSCl,14 the corresponding
silyl ketene acetal could be generated, as evidenced by its
reaction with a number of trapping agents.
Among the electrophiles that reacted successfully with the
presumed silyl ketene acetal was N-iodosuccinimide, which
(5) We could obtain analytically pure 10 in 10-20% yield after quickly
passing the crude material through Florisil, or in up to 50-60% of less
pure material by an aqueous wash of the crude reaction mixture. Disap-
pointingly, attempted epoxidation led predominately to decomposition,
providing only a trace of material with NMR and LRMS signals consistent
with those of the desired R-hydroxyketone.
(6) Napolitano, E.; Spinelli, G.; Fiaschi, R.; Marsili, A. Synthesis 1985,
38.
(7) Watanabe, M.; Morimoto, H.; Nogami, K.; Ijichi, S.; Furukawa, S.
Chem. Pharm. Bull. 1993, 41, 968.
(10) For instance, epoxidation with trifluoroperacetic acid provided an
approximately 1:1 ratio of 20a:20b; however, the yield was low, presumably
due to competing processes such as protonation of the enol ether and/or
Baeyer-Villiger oxidation.
(8) While approach from the bottom face is significantly hindered by
the lactone ring, epoxidation from the top face would provide the highly
strained trans-fused 6:5 ring system. Additionally, π-overlap between the
double bond and the proximal â-disposed ketone might have favored attack
from the R-face: cf. Woodward, R. B.; Bader, F. E.; Bickel, H.; Frey, A.
J.; Kierstead, R. W. Tetrahedron 1958, 2, 1.
(9) Crystallographic data for the structures reported in this paper have
been deposited with the Cambridge Crystallographic Data Center as
supplementary publication numbers CCDC-149057 (20b) and CCDC-
149058 (27).
(11) As expected, density functional calculations (pBP/DN*) confirm
that the desired 20a is favored thermodynamically with respect to 20b (3.1
kcal/mol). We thank Professor James Leighton for access to his SGI O2
running Spartan 5.1.2 to perform this analysis.
(12) For an excellent review of substrate-directed reactions in organic
synthesis, see: (a) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. ReV.
1993, 93, 1307. Some years ago we made use of the sequence of ring
opening of a lactone, hydroxyl-directed epoxidation, and re-lactonization
in the total synthesis of vernolepin, see: (b) Danishefsky, S.; Kitahara, T.;
Schuda, P. F.; Etheredge, S. J. J. Am. Chem. Soc. 1976, 98, 3028.
Org. Lett., Vol. 2, No. 22, 2000
3495

provided R-iodolactones 23 as a 2:1 mixture favoring 23a.15
The halogen had thus been preferentially introduced syn to
the bridgehead proton by approach from the convex face of
the lactone ring. This mixture of diastereomers was then
treated with DMDO in an effort to initiate the syn elimination
of a putative alkyl iodoso intermediate.16 In the event, a single
product identified as R-ketolactone 24 was isolated in 71%
yield (along with 17% of recovered starting material). The
formation of 24 can be viewed from the perspective of an
iodoso Pummerer-like transformation.17 Though not our
intended target, the fortuitous generation of the R-keto
lactone was used to advantage, commencing with its conver-
sion to enol triflate 25. As predicted (vide supra), in the
presence of a catalytic amount of camphorsulfonic acid, 25
suffered smooth conjugate addition of methanol, giving rise
to 26 as a single diastereomer (88% yield).18 Finally,
reductive removal of the R-triflate with lithium iodide under
Fleet’s conditions19 provided the target system 27 in 55%
yield. Assignment of the relative stereochemistry implied in
structure 27 was subsequently corroborated by X-ray crystal-
lography.9
In summary, we have reported herein the synthesis of the
tricyclic DEF ring system of lactonamycin in the setting of
a CDEF construct. The melding of this dense arrangement
of functionality is surely a major issue in any projected total
synthesis of lactonamycin. Central to the strategy followed
here was the use of an intramolecular Wessely oxidation and
subsequent hydroxyl-directed epoxidation of enol ether 21.
The application of these lessons to the larger total synthesis
problem is a matter of continuing interest.
(13) Lactones 20a,b were isolated as their m-chlorobenzoate esters when
formed through the action of mCPBA. This is most likely a result of the
highly nucleophilic benzoic acid byproduct adding to the carbon containing
the methoxy ether with simultaneous opening of the epoxide, followed by
intramolecular acyl transfer. An analogous acyl transfer, albeit under acidic
conditions, is seen in Kishi’s synthesis of tetrodotoxin; see: (a) Kishi, Y.;
Aratani, M.; Fukuyama, T.; Nakatsubo, F.; Goto, T.; Inoue, S.; Tanino, H.;
Sugiura, S.; Kakoi, H. J. Am. Chem. Soc. 1972, 94, 9217. For acyl transfer
from the epoxidation of silyl enol ethers, see: (b) Rubottom, G. M.; Gruber,
J. M.; Boeckman, R. K., Jr.; Ramaiah, M.; Medwid, J. B. Tetrahedron Lett.
1978, 4603. (c) Jefford, C. W.; Rimbault, C. G. J. Am. Chem. Soc. 1978,
100, 6437.
Acknowledgment. This work was supported by the
National Institutes of Health (HL25848). C.C. gratefully
acknowledges the NIH for support in the form of a
Postdoctoral Fellowship (F32 CA84758). The authors thank
Mr. David Churchill and Prof. Ged Parkin of Columbia
University for solving the X-ray structures of 20b and 27
and Ms. Sachiyo Nomura for high-resolution mass spectral
analyses.
(14) Corey, E. J.; Gross, A. W. Tetrahedron Lett. 1984, 495.
(15) Although analysis of the crude reaction indicated that the product
initially contained a > 5:1 mixture favoring the syn diastereomer (23a),
equilibration occurred upon chromatography such that the isolated product
was a 2:1 mixture favoring the syn isomer. The assignment of the major
isomer as the syn compound was based on the lack of J-coupling between
the adjacent protons on the lactone ring, whereas the minor component
exhibited J ) 6.5 Hz.
(16) (a) Reich, H. J.; Peake, S. L. J. Am. Chem. Soc. 1978, 100, 4888.
Fuchs has recently re-awakened interest in this seldom used transforma-
tion: (b) Kim, S.; Fuchs, P. L. Tetrahedron Lett. 1994, 7163. (c) Mahadevan,
A.; Fuchs, P. L. J. Am. Chem. Soc. 1995, 117, 3272.
(17) A number of other mechanisms can be envisioned for this
transformation, including one which proceeds by an initial syn-elimination,
followed by epoxidation of the highly strained double bond with subsequent
rearrangement (see, for example, Murray, R. W.; Shiang, D. L.; Singh, M.
J. Org. Chem. 1991, 56, 3677). Alternatively, a more simplistic explanation
would include tautomerization of the proposed iodoso intermediate (to form
the iodo-analogue of an oxime), followed by hydrolysis.
Supporting Information Available: Experimental pro-
cedures and full characterization data for 13-27. This
information is available free of charge via the Internet at
http://pubs.acs.org.
OL006531+
(18) The stereochemistry at the lactone R-carbon was not determined.
(19) Elliott, R. P.; Fleet, G. W. J.; Gyoung, Y. S.; Ramsden, N. G.; Smith,
C. Tetrahedron Lett. 1990, 3785.
3496
Org. Lett., Vol. 2, No. 22, 2000