Construction of a C(30-38) dioxabicyclo[3.2.1]octane subtarget for (+)-sorangicin A, exploiting a regio- and stereocontrolled acid-catalyzed epoxide ring opening

Article abstract of DOI:10.1021/ol049644s

In this paper, we report assembly of the novel dioxabicyclo[3.2.1]octane subtarget (-)-2, comprising the signature structural element of the potent antibiotic (+)-sorangicin A (1). The synthesis was achieved in 15 steps (1.5% overall yield) via a series o

Full text of DOI:10.1021/ol049644s

ORGANIC
LETTERS
2004
Vol. 6, No. 9
1477-1480
Construction of a C(30−38)
Dioxabicyclo[3.2.1]octane Subtarget for
(+)-Sorangicin A, Exploiting a Regio-
and Stereocontrolled Acid-Catalyzed
Epoxide Ring Opening
Amos B. Smith, III* and Richard J. Fox
Department of Chemistry, Laboratory for Research on the Structure of Matter,
and Monell Chemical Senses Center, UniVersity of PennsylVania,
Philadelphia, PennsylVania 19104
smithab@sas.upenn.edu
Received February 27, 2004
ABSTRACT
In this paper, we report assembly of the novel dioxabicyclo[3.2.1]octane subtarget (−)-2, comprising the signature structural element of the
potent antibiotic (+)-sorangicin A (1). The synthesis was achieved in 15 steps (1.5% overall yield) via a series of acid-catalyzed epoxide ring
openings. The first, facilitated by the complex of alkyne (+)-3 with Co2(CO)8, proceeded in a highly regio- and stereoselective fashion.
In a world of ever-increasing antibiotic resistance, the dis-
covery and development of new, potent antibiotics is of
utmost importance. In 1985, Ho¨fle and co-workers described
the isolation of a family of novel macrolide antibiotics termed
the sorangicins from Sorangium cellulosum, the same bacte-
rial strain that furnished the epothilone antitumor agents.1
Sorangicin A (1), the most potent congener, proved to be
highly effective against a broad panel of both Gram-positive
and Gram-negative bacteria, displaying average MIC values
of 10 ng/mL and 10 µg/mL, respectively. At the molecular
level, sorangicin A inhibits RNA polymerase in Escherichia
coli and Staphylococcus aureus. Importantly, eukaryotic cells
are unaffected. Thus, rats infected with virulent E. coli under-
went marked improvement when dosed with sorangicin A.2
Given the novel architecture, including the signature
dioxabicyclo[3.2.1]octane ring moiety, the cis,cis,trans-
trienoate and the 31-membered macrolide ring, in conjunction
with the reported, significant instability of the natural product
to a variety of reagents (e.g., fluoride ion, DDQ, and the
dissolving metal sodium amalgam),3 the sorangicin class of
natural products, and in particular sorangicin A (1), represent
timely synthetic targets. In this paper, we disclose an effec-
tive, stereocontrolled assembly of a C(30-38) dioxabicyclo-
[3.2.1]octane subtarget 2.
Our synthetic analysis, illustrated in Scheme 1, called
initially for construction of 2 via an acid-catalyzed cascade
of epoxide openings, the first facilitated and stereocontrolled
by an alkyne-Co2(CO)6 complex of bis-epoxide 3. Subse-
quent installation of the vinyl iodide and oxidation would
then complete construction of the bicyclic fragment.
(1) (a) Jansen, R.; Wray, V.; Irschik, H.; Reichenbach, H.; Ho¨fle, G.
Tetrahedron Lett. 1985, 26, 6031. (b) Jansen, R.; Irschik, H.; Reichenbach,
H.; Schomburg, D.; Wray, V.; Ho¨fle, G. Liebigs Ann. Chem. 1989, 111.
(2) Irschik, H.; Jansen, R.; Gerth, K.; Ho¨fle, G.; Reichenbach, H. J.
Antibiot. 1987, 40, 7.
(3) Schummer, D.; Irschik, H.; Ho¨fle, G. Liebigs Ann. Chem. 1993, 293.
10.1021/ol049644s CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/30/2004

1,4-diol (PMBCl, NaH), followed by ozonolysis to afford
aldehyde 55 (Scheme 3, 50%, two steps). Asymmetric Brown
Scheme 1
Scheme 3
In 1994, Mukai, Hanaoka, and co-workers4 reported the
regio- and stereoselective formation of 2-ethynyl-3-hydroxy-
tetrahydrofuran derivatives via an endo-ring closure of 3,4-
epoxy-6-substituted-5-yn-1-ols (Scheme 2). They discovered
crotylation6 then furnished homoallylic alcohol (+)-67 in
excellent yield and with high stereoselectivity (84%; d.r.,
e.r. >20:1).8 Protection of the homoallylic hydroxyl and
Sharpless dihydroxylation9 led to diol (-)-7 in good yield,
after careful removal of the minor diastereomer by flash
chromatography.
Scheme 2
Enyne (-)-8 was next prepared by one-flask Fraser-Reid
epoxide formation,10 followed by reaction with the lithium
anion derived from 1,4-bis(trimethylsilyl)-1,3-butadiyne,11
and in turn chemo- and stereoselective reduction of the
(5) Keck, G. E.; Wager, C. A.; Wager, T. T.; Savin, K. A.; Covel, J. A.;
McLaws, M. D.; Krishnamurthy, D.; Cee, V. J. Angew. Chem., Int. Ed.
2001, 40, 231.
that the epoxide opening can proceed either by retention or
by inversion of stereochemistry at the propargylic stereo-
center, depending on the reaction conditions.
With this precedent in mind, our point of departure for
the construction of 3 entailed bis-protection of cis-2-butene-
(6) Brown, H. C.; Bhat, K. S. J. Am. Chem. Soc. 1986, 108, 5919.
(7) Zheng, W.; DeMattei, J. A.; Wu, J.-P.; Duan, J. J.-W.; Cook, L. R.;
Oinuma, H.; Kishi, Y. J. Am. Chem. Soc. 1996, 118, 7946.
(8) The absolute stereochemistry and enantiomeric ratio of (+)-6 was
determined via the Kakisawa analysis of the derived Mosher esters: (a)
Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512. (b) Ohtani, I.;
Kusumi, T.; Kashman, Y.; Kakisawa, H. J. Am. Chem. Soc. 1991, 113,
4092. Also see the Supporting Information.
(4) (a) Mukai, C.; Sugimoto, Y.-i.; Ikeda, Y.; Hanaoka, M. J. Chem.
Soc., Chem. Commun. 1994, 1161. (b) Mukai, C.; Sugimoto, Y.-i.; Ikeda,
Y.; Hanaoka, M. Tetrahedron 1998, 54, 823.
(9) (a) Crispino, G. A.; Jeong, K.-S.; Kolb, H. C.; Wang, Z.-M.; Xu, D.;
Sharpless, K. B. J. Org. Chem. 1993, 58, 3785. (b) Kolb, H. C.;
VanNiewenhze, M. S.; Sharpless, K. B. Chem. ReV. 1994, 94, 2483.
1478
Org. Lett., Vol. 6, No. 9, 2004

internal triple bond.12 With (-)-8 in hand, removal of the
PMB groups, followed by a second Fraser-Reid10 protocol
led to epoxide (+)-9. Reagent-controlled diastereoselective
epoxidation exploiting the elegant Shi conditions13 furnished
a mixture of diastereomers (12:1), which upon HPLC
purification afforded the requisite cyclization precursor bis-
epoxide (+)-3 (41%, 64% BORSM).
We initiated cyclization studies with the expectation of
retention of configuration based on the precedent of Mukai,
Hanaoka, and co-workers.4 To this end, treatment of (+)-3
with 1.1 equiv of Co2(CO)8 at ambient temperature, followed
at -78 °C by a catalytic amount of BF3‚OEt2, yielded
alkyne-cobalt complex (+)-11 (Scheme 4). Importantly, the
OEt2 reaction conditions did not result in any change, thus
suggesting that the cyclizations proceed under kinetic control.
Based on the X-ray crystal structure of (+)-11, we
reasoned that the steric bulk of the cobalt complex might
disfavor the desired 6-exo-tet pathway. To reduce the steric
incumbrance, as well as to attenuate the ease of epimeriza-
tion, we removed the cobalt moiety from the alkyne prior to
attempting the second cyclization (Scheme 4). This three-
step operation, performed in a single flask, furnished epoxide
(-)-15 in 88% yield.
Scheme 4
With epoxide (-)-15 in hand, ring opening of the second
epoxide was reinvestigated (Table 1). Initially, we explored
Table 1. Examining the 6-exo- and 7-endo-tet Cyclizations of
(-)-15
t
time
(h)
(-)-16/ convb
entry
conditionsa
(°C)
(-)-17b
(%)
1
2
3
4
5
6
7
8
CSA (0.1 equiv)
BF3‚OEt2 (0.1 equiv) -78 25
BF3‚OEt2 (0.1 equiv)
BF3‚OEt2 (0.1 equiv)
BF3‚OEt2 (1.0 equiv) -78
BF3‚OEt2 (2.0 equiv) -78
BF3‚OEt2 (2.0 equiv)
5 M LiClO4‚OEt2,
CSA (0.1 equiv)
25 28
1:1
1:6
1:1
1:1
1:3
1:3
1:1
1:1.1
97
10
100
100
80
93
100
10
reaction proceeded with complete chemoselectivity at the
activated propargylic epoxide. Single-crystal X-ray analysis
verified that the first epoxide opening had indeed proceeded
with retention of stereochemistry.14
0
1
40
0.5
1
0.75
0.3
1.5
To determine the feasibility of performing the second
epoxide opening without isolation of (+)-11, we explored a
variety of added acids and bases. Unexpectedly, (+)-11 failed
to produce any of the desired bicycle resulting from 6-exo-
tet cyclization. Instead, with either CSA or PPTS, only
selective formation of bicycle (-)-12, resulting from 7-endo-
tet cyclization, was observed! Surprisingly, treatment of (+)-
11 with BF3‚OEt2 led to mixtures of the 7-membered bicycle
(-)-12, along with the 6- and 7-membered bicycles (-)-13
and (-)-14, respectively, both epimerized at the propargylic
center! Bicycles (-)-12, (-)-13, and (-)-14 were distin-
0
25
a Entries 1-7 were run in CH2Cl2. b Ratios and conversions determined
by 500 MHz H NMR.
1
the cyclization under “standard” acidic conditions (see entries
1-7, Table 1). Unlike cobalt complex (+)-11, exposure of
(-)-15 to either CSA or BF3‚OEt2 furnished the desired
6-exo-tet product (-)-16, albeit at best as a mixture (ca. 1:1)
with the 7-endo-tet bicycle (-)-17. Importantly, no epimer-
ized products were observed. Attempts to increase this ratio
with other Lewis acids (ca. TiCl4 or CeCl3) or under basic
conditions (ca. KH, DMSO) proved unsuccessful. In an
attempt to increase the amount of carbocation character in
the ring-opening transition state, anticipated to favor attack
at the epoxide secondary carbon, we explored the extremely
polar solvent system of Grieco and co-workers15 (LiClO4‚
OEt2) (see entry 8, Table 1). Although these conditions
dramatically increased the reaction rate (entry 1 vs 8, Table
1), the ratio of (-)-16 to (-)-17 did not improve.
1
guishable through H, COSY, and D2O exchange experi-
ments. Reexposure of the individual bicycles to the BF3‚
(10) (a) Hicks, D. R.; Fraser-Reid, B. Synthesis 1974, 203. (b) Cink, R.
D.; Forsyth, C. J. J. Org. Chem. 1995, 60, 8122.
(11) (a) Holmes, A. B.; Jennings-White, C. L. D.; Schulthess, A. H.
Chem. Commun. 1979, 840. (b) Holmes, A. B.; Jones, G. E. Tetrahedron
Lett. 1980, 21, 3111.
(12) Doolittle, R. E. Synthesis 1984, 730.
(13) (a) Wang, Z.-X.; Tu, Y.; Frohn, M.; Zhang, J.-R.; Shi, Y. J. Am.
Chem. Soc. 1997, 119, 11224. (b) Wang, Z.-X.; Cao, G.-A.; Shi, Y. J. Org.
Chem. 1999, 64, 7646.
(14) For a mechanistic explanation of this double-inversion process see
ref 5b.
Org. Lett., Vol. 6, No. 9, 2004
1479

We next reasoned that, since reducing the steric bulk of
the alkyne led at least to some of the desired product, per-
forming the second epoxide opening on the terminal acety-
lene might improve the yield of the desired 6-exo-tet bicycle.
To this end, the TMS group was removed to furnish (-)-18
(Table 2).
result in a change. Single-crystal X-ray analysis verified both
the structure and stereochemistry of (-)-19.
With (-)-19 in hand, final elaboration to coupling partner
2 was accomplished via radical-promoted hydrostannyla-
tion,16 iodination, and Dess-Martin oxidation (Scheme 5).17
This three step sequence furnished (-)-2 in 77% yield.
Scheme 5
Table 2. Examining the 6-exo- and 7-endo-tet Cyclizations of
(-)-18
In summary, an effective 15-step synthesis of dioxabicyclo-
[3.2.1]octane (-)-2, comprising the C(30-38) signature
fragment of sorangicin A (1), has been achieved. The cor-
nerstone of the synthetic strategy involved a kinetically con-
trolled, regio- and stereoselective epoxide opening facilitated
by the Co2(CO)6 complex of alkyne (+)-3. Importantly, in
the second epoxide cyclization, both reducing the alkyne
steric bulk and increasing the reaction rate led to a higher
yield of the desired 6-exo-tet cyclization. Studies to improve
the synthesis of (-)-2 and to complete the total synthesis of
(+)-sorangicin A (1) will be reported in due course.
t
time
(h)
(-)-19/ convb
entry
conditionsa
(°C)
(-)-20b
(%)
1
2
BF3‚OEt2 (0.1 equiv)
BF3‚OEt2 (10.0 equiv)
(inverse addition)
BF3‚OEt2 (10.0 equiv)
(inverse addition)
0
40
1
1.4:1
2.1:1
100
100c
5 min
4 min
1.5
3
4
80
1:1.2
100
BF3‚OEt2 (0.1 equiv)
25
1.8:1
100
a Entries 1 and 2 were run in CH2Cl2. Entries 3 and 4 were run in
benzene. b Ratios and conversions determined by 500 MHz 1H NMR.
c Isolated yield for (-)-19 was 65%.
Acknowledgment. Support was provided by the National
Institutes of Health through Grant No. GM-29028 and an
Eli Lilly and Dissertation Graduate Fellowships to R.J.F.
Pleasingly, treatment of (-)-18 under conditions identical
to those employed in entry 3 (Table 1) increased the ratio
of the desired bicycle (-)-19 to 1.4:1 (see entry 1, Table 2).
The results in Tables 1 and 2 further suggest that as the
reaction rate increases, the amount of the desired 6-exo-tet
bicycle increases (see entries 6 and 7, Table 1). With this in
mind, our best conditions to date for the formation of (-)-
19 call for the dropwise addition of (-)-18 to a solution of
CH2Cl2 at reflux containing 10.0 equiv of BF3‚OEt2. Under
these conditions, the ratio increased to 2.1:1 (entry 2, Table
2); the isolated yield of (-)-19 was 65%. Again, reexposure
of either (-)-19 or (-)-20 to the reaction conditions did not
Supporting Information Available: Spectroscopic and
analytical data for 2-9, 11, and 15-20 and selected
experimental procedures. This material is available free of
charge via the Internet at http://pubs.acs.org.
OL049644S
(15) Grieco, P. A.; Nunes, J. J.; Gaul, M. D. J. Am. Chem. Soc. 1990,
112, 4595.
(16) (a) Tolstikov, G. A.; Miftahov, M. S.; Danilova, N. A.; Vel’der, Y.
L. Synthesis 1986, 496. (b) Jung, M. E.; Light, L. A. Tetrahedron Lett.
1982, 23, 3851.
(17) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4156.
1480
Org. Lett., Vol. 6, No. 9, 2004