First Enantiospecific Total Synthesis of the Antitubercular Marine Natural Product Pseudopteroxazole. Revision of Assigned Stereochemistry

Article abstract of DOI:10.1021/ja0378916

A concise, enantiospecific synthesis of pseudopteroxazole (3), which had originally been assigned structure 1, has been accomplished starting from S-(-)-limonene. The known cyclohexanone 5 was converted in five steps to the α,β-enone 8 by a modified Robin

Full text of DOI:10.1021/ja0378916

Published on Web 10/10/2003
First Enantiospecific Total Synthesis of the Antitubercular
Marine Natural Product Pseudopteroxazole. Revision of
Assigned Stereochemistry
James P. Davidson and E. J. Corey*
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
Received August 12, 2003; E-mail: corey@chemistry.harvard.edu
Abstract: A concise, enantiospecific synthesis of pseudopteroxazole (3), which had originally been assigned
structure 1, has been accomplished starting from S-(-)-limonene. The known cyclohexanone 5 was
converted in five steps to the R,â-enone 8 by a modified Robinson annulation. Transformation of 8 to the
orthogonally protected amino phenol 11 was accomplished by a new modification of the Wolff-Semmler
rearrangement. The synthesis was completed by cationic cyclization to form 14 diastereoselectively and
subsequent introduction of the terminal oxazole subunit.
The report1 by Rodriguez and co-workers of a novel tetra-
cyclic oxazole, pseudopteroxazole, from the West Indian
gorgonial coral Pseudopteragorgia elisabethae that exhibited
strong inhibition of Mycobacterium tuberculosis in vitro further
stimulated the interest of synthetic chemists in this class of com-
pounds. As one of the world’s leading pathogens, Mycobacte-
rium tuberculosis infects an estimated 2 billion people and kills
some 3 million a year. Currently, the standard treatment for M.
tuberculosis infection is a 6-month regimen of antibiotics, con-
sisting in some cases of more than 10 pills per day. Because of
this demanding treatment schedule, patient compliance is low,
and several different multi-drug-resistant strains have evolved.
Thus, there is a clear need for new and efficacious therapeutic
agents. We reported earlier three enantiospecific and stereocon-
trolled synthetic routes to the pseudopterosins,2,3 and a variety
of synthetic approaches have been developed in other labora-
tories.4 Most recently, our synthetic and spectroscopic studies
have allowed the revision of the structures originally proposed
for several members of the pseudopterosin family.5 This work
also called into question the structure originally assigned to
pseudopteroxazole (1).1 An unambiguous total synthesis of 1
and its C(1)-diastereomer (2) later showed that pseudopterox-
azole was neither of these and strengthened the case for structure
3.6 We describe herein an enantiospecific total synthesis of 3
which confirms this conclusion and unequivocally establishes
this structure for pseudopteroxazole through comparison with
the natural product. In addition, we have also synthesized the
C(1)-diastereomer of pseudopteroxazole (4) by modifying final
stages of the synthesis of 3. Thus, all of the diastereomers 1-4
are available for studies of antimicrobial activity.
(1) Rodr´ıguez, A. D.; Ram´ırez, C.; Rodr´ıguez, I. I.; Gonza´lez, E. Org. Lett.
1999, 1, 527.
The pathway of the synthesis leading to 3, which is outlined
in Scheme 1, began with the known cyclohexanone 5, available
in three steps from inexpensive S-(-)-limonene.3 Protection of
the primary hydroxyl as the tert-butyl-diphenylsilyl ether
(TBDPS) gave 6 in 96% yield. Kinetic deprotonation of the
cyclohexanone 6 selectively at the R-methylene and trapping
with chlorotrimethylsilane provided the required silyl enol ether,
which, upon treatment with 1-benzyloxy-3-methyl-but-3-en-2-
(2) (a) Corey, E. J.; Carpino, P. J. Am. Chem. Soc. 1989, 111, 5472. (b) Corey,
E. J.; Carpino, P. Tetrahedron Lett. 1990, 31, 3857.
(3) Corey, E. J.; Lazerwith, S. E. J. Am. Chem. Soc. 1998, 120, 12777.
(4) (a) Broka, C. A.; Chan, S.; Peterson, B. J. Org. Chem. 1988, 53, 1584. (b)
McCombie, S. W.; Cox, B.; Lin, S.-I.; Ganguly, A. K.; McPhail, A. T.
Tetrahedron Lett. 1991, 32, 2083. (c) McCombie, S. W.; Ortiz, C.; Cox,
B.; Ganguly, A. K. Synlett 1993, 541. (d) Buszek, K. R. Tetrahedron Lett.
1995, 36, 9125. (e) Buszek, K. R.; Bixby, D. L. Tetrahedron Lett. 1995,
36, 9129. (f) Gill, S.; Kocienski, P.; Kohler, A.; Pontiroli, A.; Qun, L. J.
Chem. Soc., Chem. Commun. 1996, 1743. (g) Majdalani, A.; Schmalz, H.-
G. Tetrahedron Lett. 1997, 38, 4545. (h) Majdalani, A.; Schmalz, H.-G.
Synlett 1997, 1303. (i) Kato, N.; Zhang, C.-S.; Matsui, T.; Iwabachi, H.;
Mori, A.; Ballio, A.; Sassa, T. J. Chem. Soc., Perkin Trans. 1 1998, 2475.
(5) Lazerwith, S. E.; Johnson, T. W.; Corey, E. J. Org. Lett. 2000, 2, 2389.
(6) Johnson, T. W.; Corey, E. J. J. Am. Chem. Soc. 2001, 123, 4475.
9
13486
J. AM. CHEM. SOC. 2003, 125, 13486-13489
10.1021/ja0378916 CCC: $25.00 © 2003 American Chemical Society

Total Synthesis of Pseudopteroxazole
A R T I C L E S
Scheme 1. Synthetic Pathway to Pseudopteroxazole 1a
a (a) TBDPSCl, imidazole, DMF, 96%. (b) LDA, TMS-Cl, -78 °C, 100%. (c) 1-Benzyloxy-3-methyl-but-3-en-2-one, SnCl4, CH2Cl2, -78 °C, 61%. (d)
KOH, EtOH, -10 °C, 83%. (e) SOCl2, pyridine, 83%. (f) NH2OH-HCl, pyridine, 69%. (g) Pivaloyl chloride, pyridine, 96%. (h) Acetyl chloride, toluene,
80 °C, 64%. (i) H2, Pd(OH)2-C, EtOH, 99%. (j) Carbonyldiimidazole, Et3N; then aqueous NaHCO3, 94%. (k) HF-pyridine, 95%. (l) TPAP, NMO, 82%.
(m) Wittig olefination, 81%. (n) MeSO3H, acetic acid, 18 °C, 80%. (o) BOC2O, DMAP, 96%. (p) MeMgBr, -78 to 23 °C; then TFA, HC(OEt)3, 90%.
one and SnCl4, gave the Mukaiyama-Michael adduct 7 in 61%
yield (82% based on recovered 6) as a mixture of diastereomers
(approximate ratio 1:1). Cyclization of the 1,5-diketone 7 in
dilute solution (0.01 M) using potassium hydroxide in ethanol
(83%), and elimination of the tertiary hydroxyl group with
thionyl chloride in pyridine, proceeded smoothly to give a
mixture of diastereomeric R,â-enones 8 in 83% yield. Trans-
formation of this diastereomeric mixture into the corresponding
oximes 9 (65-70%)7 followed by acylation with pivaloyl
chloride provided the oxime pivalate 10 in 96% yield.
Next, we turned our attention to the aromatization of 10. It
was clear from the outset that conventional Wolff-Semmler
conditions, which consist of heating with excess acetyl chloride
and concentrated HCl in refluxing acetic anhydride,8 would be
far too harsh for this acid-sensitive substrate. We experimented
extensively with different O-protected oximes, acylating agents,
additives, and solvents (for discussion, see below). Ultimately,
we found that 1 equiv of acetyl chloride in toluene at 80 °C in
a tightly sealed reaction vessel effectively promoted the aro-
matization of the O-pivaloyl protected oxime 10 to give the
orthogonally protected ortho aminophenol 11 reliably in 60-
65% yield. It should be noted that both oxime diastereomers
undergo aromatization equally well, regardless of whether they
are subjected to the reaction conditions individually or as a
mixture.
cyclic carbamate 12 in 94% yield after a mildly basic aqueous
workup (NaHCO
3) to remove the N-acetyl group. Desilylation
of 12 with hydrofluoric acid-pyridine complex (95%), followed
by mild oxidation to the aldehyde with tetrapropylammonium
perruthenate and 4-methylmorpholine-N-oxide (TPAP-NMO,
82%),9 and Wittig-Vedejs E-selective olefination,3,10 gave diene
13 (81%), the key intermediate for the formation of the third
carbocyclic ring of 3 by cationic cyclization. Previous experience
with this type of ring closure provided the insights for control
of stereochemistry at C(1) by the relative electron donation of
the substituents at C(9) and C(10) of the aromatic ring.3 We
speculated that the slightly superior electron-donating properties
of the nitrogen atom at position 10 would direct the stereo-
chemistry at C(1) to the desired S configuration. Surprisingly,
it was found after experimentation that either C(1)-diastereomer
of 14 could be accessed selectively by changing the reaction
conditions for the cyclization. With acetic acid as solvent,
treatment of 13 with 3 equiv of methanesulfonic acid at 19 °C
for 72 h gave 14 in 80% yield as approximately a 4:1 ratio
favoring the required S configuration at C(1). Alternatively, with
CH2Cl2 as solvent and methanesulfonic acid at -30 °C for 6 h,
the diastereomer of 14 at C(1) was obtained in 95% isolated
yield with a 4:1 predominance over 14. Tetracycle 14 and its
C(1)-diastereomer were separated chromatographically on a
Chiralcel OD column and converted individually by parallel
processes to pseudopteroxazole (3) and the diastereomer 4,
respectively. This was accomplished by the sequence: (1)
acylation of the free NH of 14 with di-tert-butyl dicarbonate
(BOC2O) to give 15 in 96% yield; (2) reaction with excess
methylmagnesium bromide (MeMgBr) to cleave the cyclic
carbamate; and (3) addition of the resulting dianion to a mixture
of trifluoroacetic acid and triethylorthoformate to give 3 in 85-
90% yield. Comparison of the spectral data for synthetic 3 with
that originally reported for the natural material revealed identical
Hydrogenolysis of the benzyl ether (99%), followed by
cyclization of the phenol with carbonyldiimidazole, gave the
(7) Under conventional oxime forming conditions (excess hydroxylamine
hydrochloride in ethanol), no oxime was observed with 8. Heating the
reaction in ethanol with hydroxylamine resulted in a complex mixture
containing only traces of the desired oxime. With excess hydroxylamine
hydrochloride, at 50 °C, with pyridine as solvent the reaction proceeded to
approximately 50% conversion after 4 h, leaving one diastereomer of 8
untouched (55% isolated yield of 9, 84-89% based on recovered 8). The
remaining enone diastereomer was then isolated, and resubjected with excess
hydroxylamine hydrochloride in refluxing pyridine to give predominantly
a different oxime diastereomer in 50% yield (65-70% overall). If the initial
mixture of enones 8 was subjected to excess hydroxylamine hydrochloride
in refluxing pyridine, it was possible to force the reaction to completion.
However, isolated yields of the resulting mixture of oximes were only about
55%. Curiously, the reaction was also not influenced by the addition of
other amine bases or by the removal of water either azeotropically or with
drying agents.
1
rotation, IR, H NMR, 13C NMR, and high-resolution mass
spectra. The pseudopteroxazole diastereomer 4 was synthesized
(9) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D. J. Chem. Soc.,
Chem. Commun. 1987, 1625.
(10) Vedejs, E.; Fang, H. W. J. Org. Chem. 1984, 49, 210.
(8) (a) Semmler, W. Chem. Ber. 1892, 25, 3352. (b) Wolff, L. Liebigs Ann.
1902, 322, 351.
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13487

A R T I C L E S
Davidson and Corey
Scheme 2
Scheme 3 a
a (a) H2, Pd(OH)2-C, 99%. (b) MsCl, Et3N, THF, 95%. (c) HF-pyridine,
THF, 50 °C, 96%. (d) TPAP, NMO, 81%. (e) Wittig olefination, 90%. (f)
MeSO3H, CH2Cl2, 95%. (g) BOC2O, Et3N; then K2CO3, MeOH, 92%. (h)
MeMgBr, -78 °C f 23 °C; then TFA, HC(OEt)3, 42%.
using the analogous sequence from the C(1)-diastereomer of
14 and was demonstrated to be spectroscopically different from
3.
bonding of the carbamate NH to the solvent acetic acid, as
depicted by 20 in Scheme 2. If this explanation is correct, it
has significant implications with regard to synthetic planning
for the use of benz-1,3-oxazol-2-one subunits in synthesis.
The use of the cyclic carbamate derivative 13 was critical to
the successful cyclization to form 14. One advantage of 13 is
that the coplanarity of the oxygen and nitrogen substituents
allows maximum electron supply from these groups into the
aromatic ring as cationic cyclization is taking place. This aspect
of the cyclization was probed further by the synthesis of the
acetamide-mesylate 22, and the corresponding triflate, from 11
via the intermediate 21, as shown in Scheme 3. The methane-
sulfonyl substituent in 22 was chosen to provide strong
attenuation of the electron-donating capabilities of oxygen into
the aromatic ring during direct cationic attack at the para
position. Methanesulfonic acid-catalyzed cyclization of 22 either
in CH2Cl2 or in acetic acid produced 23 with at least 20:1
diastereoselectivity relative to 14. A similar result was obtained
even in the cyclization of the substrate having CF3SO2 instead
of CH3SO2 of 22. These data for 22 and the corresponding
triflate make it clear that the acetamido substituent is a very
poor electron donor in the cyclization of 22 as a consequence
of its noncoplanarity with the aromatic ring due to ortho
substitution.
The transformation of the R,â-enone 8 into the aromatic
amino phenol derivative 11 (Scheme 1) presented a major
challenge to the execution of the synthetic plan. As mentioned
earlier, standard Wolff-Semmler conditions8 were inappropri-
ate. A number of exploratory experiments with oximes of model
2-cyclohexenones and with 8 provided the following useful
information: (1) the best results were obtained with pivalate
esters of the oximes, (2) the best reagent for the oxime pivalate
aromatization was acetyl chloride in toluene at 80 °C, and (3)
hydrogen chloride was an essential catalyst for the process. In
the Wolff-Semmler transformation of the oxime pivalate 10
to the protected acetamide-phenol 11 (Scheme 1), a byproduct,
identified as 24, was isolated in yields of between 5% and 15%.
Control experiments using HCl or triphosgene in toluene
revealed that 24 could be derived from the aromatization product
11, likely through an acid-catalyzed cleavage of the benzyl ether,
and subsequent cyclization and dehydration of the N-acetyl.
Efforts to remove excess hydrogen chloride from the reaction,
Determination of the stereochemistry at C(1) was relatively
straightforward. Our previous work on the pseudopterosin A-F
aglycon, pseudopterosin G-J aglycon, helioporin E, as well as
1
1 and 2, has demonstrated that the C(1)-H signal in the H
NMR spectrum is of significant diagnostic value.3,5,6 In each
case where the C(1) proton resides on the R (bottom) face of
the molecule the peak width is smaller, and the signal appears
nearly rectangular with no discernible splitting pattern. On the
other hand, for each case where the C(1) proton resides on the
â (top) face of the molecule the proton signal clearly appears
as a quartet. Additionally, the proton signal for an R-oriented
H at C(1) appears ∼0.1 ppm upfield from the âH of the C(1)
diastereomer.11
The stereocontrolled cyclization 13 f 14, a key step in the
synthesis of pseudopteroxazole, is of considerable mechanistic
interest. Protonation of the 1,3-diene subunit of 13 by meth-
anesulfonic acid initiates the cyclization by forming the
stabilized allylic carbocation 16 (Scheme 2). Ring closure of
16 to form the spirocyclopentyl cation 17 was expected to be
favored by two factors: (1) an intrinsic stereoelectronic effect
leading to a preference for five-membered ring closure,12 and
(2) the greater electron-donating ability of the carbamate NH
substituent on the benzenoid ring as compared with the
carbamate O substituent. The S stereochemistry at C(1) of 17
is favored over the diastereomeric R arrangement because it
involves less steric repulsion about the isobutenyl group. A 1,2-
shift of C(1) in 17 to the next position on the benzenoid ring
and subsequent deprotonation produces 14 stereoselectively
when acetic acid is used as solvent. In methylene chloride as
solvent with methanesulfonic acid as catalyst, the balance is
shifted in favor of the alternative direct six-membered cycliza-
tion pathway 16 f 18 as shown in Scheme 2. In this ring
closure, the R configuration at C(1) is clearly favored for steric
reasons and so the final product is the C(1)-(R) diastereomer of
14, compound 19 in Scheme 2. We believe that the reason for
preferential formation of 14 with HOAc as solvent versus 19
with CH2Cl2 as solvent may simply be greater stabilization of
the transition state for cyclization of 16 to 17 by hydrogen
(11) For representative proton spectra, see Supporting Information and ref 5.
(12) See: Corey, E. J.; Sauers, C. K. J. Am. Chem. Soc. 1957, 79, 248.
9
13488 J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003

Total Synthesis of Pseudopteroxazole
A R T I C L E S
designed to inhibit the formation of 24, resulted in significantly
depressed yields. The yield of 11 was maximized by limiting
the temperature and duration of the reaction, and the conditions
described herein for the conversion of 10 to the acetamido
phenol ether 11 are superior to those previously recorded.
The completion of the first synthesis of pseudopteroxazole
as shown in Scheme 1 firmly establishes the structure of this
interesting bioactive marine product as 3. It also demonstrates
a biomimetic cyclization which provides a basis for understand-
ing the origin of the stereochemical diversity of the pseudopterosin
family and how the natural diastereomeric structures are formed
in nature.
Acknowledgment. We are grateful to the National Institutes
of Health for a postdoctoral fellowship to J.P.D.
Supporting Information Available: Experimental procedures,
and spectral data for new compounds (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
JA0378916
9
J. AM. CHEM. SOC. VOL. 125, NO. 44, 2003 13489