A cascade cyclization route to adjacent bistetrahydrofurans from chiral triepoxyfarnesyl bromides

Article abstract of DOI:10.1021/jo801188w

(Chemical Equation Presented) A number of isomeric bistetrahydrofuran analogues were prepared from triepoxy farnesyl bromides by a zinc-initiated reduction-elimination and in situ Lewis acid-promoted cascade cyclization.

Full text of DOI:10.1021/jo801188w

A Cascade Cyclization Route to Adjacent Bistetrahydrofurans from
Chiral Triepoxyfarnesyl Bromides
James A. Marshall* and Ronald K. Hann
Department of Chemistry, UniVersity of Virginia, P.O. Box 400319, McCormick Road,
CharlottesVille, Virginia 22904
jam5x@Virginia.edu
ReceiVed June 3, 2008
A number of isomeric bistetrahydrofuran analogues were prepared from triepoxy farnesyl bromides by
a zinc-initiated reduction-elimination and in situ Lewis acid-promoted cascade cyclization.
SCHEME 1. Suggested Reaction Sequence for a
Zinc-Initiated Iodo Epoxide Cascade Cyclization
Introduction
Tetrahydrofurans provide a structural motif for numerous
bioactive natural products and a great deal of effort has been
invested in devising methods for their synthesis.1 Many aceto-
genins2 and polyether ionophores,3 a special subgroup of these
natural products, possess connected bis- or less commonly
tristetrahydrofuran moieties. The present contribution describes
studies on the synthesis of bistetrahydrofurans related to
ionophore-type natural products. The methodology that forms
the basis of these studies was discovered in connection with
experiments directed at the total synthesis of the cytotoxic
marine cembranolides uprolides D and E.4,5 In those investiga-
tions we were confronted with a need for an efficient route to
tetrahydrofurans with vinyl and hydroxyethyl substituents at the
2- and 5-positions.
products based upon internal epoxide cleavages. In the course
of those investigations we discovered an efficient cascade
sequence in which the zinc-initiated elimination of an iodom-
ethyl epoxide generates the requisite vinyl group with formation
of a Lewis-acidic iodozinc byproduct that catalyzes a subsequent
internal 5-exo cyclization as depicted in Scheme 1.
Previous polyepoxide routes to tetrahydrofurans have typically
employed acidic conditions to activate the epoxide toward
internal cleavage by oxygen nucleophiles.6 In contrast, the zinc
methodology proceeds under nearly neutral conditions and
produces unsymmetrical tetrahydrofurans with chemically dis-
tinct functionality at the 2- and 5-positions.
We examined a number of alternative possibilities for
constructing the tetrahydrofuran ring of the foregoing natural
(3) Faul, M. M.; Huff, B. E. Chem. ReV. 2005, 105, 4348.
(4) Marshall, J. A.; Chobanian, H. R. Org. Lett. 2003, 5, 1931.
(5) Rodriguez, A. D.; Soto, J. J.; Pina, I. C. J. Nat. Prod. 1995, 58, 1209.
(6) Cf. (a) Fukuyama, T.; Vranesic, B.; Negri, D. P.; Kishi, Y. Tetrahedron
Lett. 1978, 2741. (b) Hoye, T. R.; Ye, Z. J. Am. Chem. Soc. 1996, 118, 1801.
(c) Xiong, Z.; Corey, E. J. J. Am. Chem. Soc. 2000, 122, 9328.
(1) For the most recent comprehensive review, see: Wolfe, J. P.; Hay, M. B.
Tetrahedron 2007, 63, 261.
(2) Bermejo, A.; Figadere, B.; Zafra-Polo, M.-C.; Barrachina, I.; Estornell,
E.; Cortes, D. Nat. Prod. Rep. 2005, 269.
10.1021/jo801188w CCC: $40.75
Published on Web 08/07/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 6753–6757 6753

Marshall and Hann
SCHEME 2. An Application of Zinc-Initiated Cascade Methodology to a Synthesis of the C17-C32 Segment of Ionomycin
SCHEME 3
We subsequently examined a single extended version of this
methodology in which an iodomethyl triepoxide serves as the
precursor of an adjacent bistetrahydrofuran segment of the
polyether antibiotic ionomycin (Scheme 2).7 Success with that
application encouraged additional investigations to determine
the potential of this novel extended sequence as a route to other
bistetrahydrofurans. In the present report we describe the results
of those investigations.
the absence of Zn resulted in the formation of multiple
decomposition products. Thus we conclude that the cascade
sequence requires a special combination of reducing metal and
in situ Lewis acid catalyst.
To probe potential stereochemical requirements of the reac-
tion, we prepared the isomeric triepoxy bromide 8 from farnesol
in a sequence parallel to that for 4 (Scheme 4). When subjected
to the zinc-promoted cascade sequence this bromo epoxide
afforded the cis,anti,cis-bistetrahydrofuran 9 in 63% yield. The
improved yield in this cascade reaction may arise from a cis/
trans preference in the initial ring formation, although we did
not attempt to rigorously establish such a preference.
Our next series of experiments was conducted on the R,ω-
difunctionalized farnesyl epoxide 11. This intermediate was
prepared from farnesol along the lines of McDonald and co-
workers10 by Sharpless asymmetric epoxidation followed by
acetylation and Sharpless SeO2-catalyzed allylic oxidation
(Scheme 5). This latter reaction afforded a 60:40 mixture of
the aforementioned primary allylic alcohol 11 and the internal
secondary alcohol isomers 12, along with small amounts of other
byproducts. Separation of this mixture was facilitated by
selective silylation of the primary alcohol 11 prior to chroma-
tography and subsequent regeneration of the alcohol through
treatment of the purified silyl ether 13 with TBAF. Upon
Sharpless asymmetric epoxidation the allylic alcohol 11 gave
the epoxide derivative 14. This epoxy alcohol served as the
starting point for a study to determine the effect of the ω-alcohol
substituent on the efficiency of the cascade sequence. The first
of these, the Boc ester 18, was prepared from alcohol 14 as
outlined in Scheme 5 by treatment with Boc anhydride. Shi
epoxidation of the diepoxy ester 15 led to the triepoxide 16,
which was selectively cleaved to alcohol 17 and converted to
the bromide 18.11
Results and Discussion
As a starting point for this project we selected the triepoxide
4, which was readily prepared from farnesol (1) by Sharpless
OH-directed asymmetric epoxidation with the ethyl L-(+)-
tartrate ligand8 and subsequent bis-Shi epoxidation9 of the
remaining two double bonds in the derived bromide 3 (Scheme
3). Purification of bromo epoxide 4 was greatly facilitated by
treatment of the reaction mixture with NaBH4 to reduce the Shi
keto hexose catalyst to a more readily separable alcohol. Upon
treatment with zinc dust and Bu4NI in refluxing ethanol, bromo
epoxide 4 was converted to the cis,anti,trans-bistetrahydrofuran
5 in 45% yield. Previously we found that bromo epoxides such
as 4 are significantly more stable than the corresponding iodides,
which tend to decompose upon storage. However, the zinc-
initiated elimination reaction proceeds much more slowly with
bromides than with iodides. Thus we employed a procedure for
in situ conversion of the bromides to the more reactive iodides
with Bu4NI. In those earlier studies we found that DME, DMF,
DMSO, MeCN, or Et2O were ineffective as solvents for the
cascade reaction. In the present investigation we confirmed that
the reactions could be carried out in methanol or ethanol with
comparable results, but the use of THF as a solvent or cosolvent
proved unsatisfactory, resulting in low yields of impure products.
Substitution of Mg for Zn resulted in recovery of the bromo
epoxide 4 along with the corresponding iodide with no evidence
for an elimination reaction. Treatment of the bromo epoxide 4
with TBAI and ZnBr2, ZnI2, or MgBr2 in refluxing ethanol in
For the terminal TBS ether analogue 21 (Scheme 6) we
employed a double Shi epoxidation of the previously prepared
dienyl epoxy acetate 13 and subsequent cleavage of the acetate
19 and bromide formation as described above. An analogous
sequence of reactions was used to convert the diepoxy alcohol
14 to the DPS ether analogue 25 (Scheme 7).
(7) Marshall, J. A.; Mikowski, A. M. Org. Lett. 2006, 8, 4375.
(8) (a) Gao, Y.; Hanson, R. M.; Klunder, J. M.; Ko, S. Y.; Masamune, H.;
Sharpless, K. B. J. Am. Chem. Soc. 1987, 109, 5765. (b) Hanson, R. M.;
Sharpless, K. B. J. Org. Chem. 1986, 51, 1922.
(9) (a) Wang, Z.-X.; Tu, Y.; Frohn, M.; Zhang, J.-R.; Shi, Y. J. Am. Chem.
Soc. 1997, 119, 11224. (b) Tu, Y.; Frohn, M.; Wang, Z.-X.; Shi, Y. Org. Syn.
2003, 80, 1. (c) Tu, Y.; Wang, Z.-X.; Shi, Y. J. Am. Chem. Soc. 1996, 118,
9806.
(10) McDonald, F. E.; Bravo, F.; Wang, X.; Wei, X.; Toganoh, M.;
Rodriguez, J. M.; Do, B.; Niewert, W. A.; Hardcastle, K. I. J. Org. Chem. 2002,
67, 2515.
(11) Garegg, P. J.; Samuelsson, B. J. Chem. Soc., Chem. Commun. 1979,
978.
6754 J. Org. Chem. Vol. 73, No. 17, 2008

Cascade Cyclization Route to Bistetrahydrofurans
SCHEME 4
SCHEME 5
SCHEME 6
Each of the triepoxy bromide analogues 18, 21, and 25 was
subjected to a Zn-initiated cascade cyclization to afford the
corresponding, cis,anti,cis-bistetrahydrofurans 26, 27, and 28
(Scheme 8). We had expected the Boc derivative to coordinate
with the halozinc Lewis acid salts formed in the initial step of
the cascade thereby facilitating the second tetrahydrofuran ring
formation (see Scheme 1). However, the yield of the bistet-
rahydrofuran product 26 was significantly lower than that of
the other two analogues and an inseparable byproduct was
formed. Presumably, as noted by McDonald and co-workers,
the Boc grouping initiates a competing “left-to-right” cascade
resulting in the formation of cyclic carbonates and related
inseparable byproducts. Both silyl ethers 21 and 25 afforded
the readily purified furan products 27 and 28 in satisfactory
yield. Thus the cascade reaction does not appear to be
appreciably facilitated by a coordinating terminal substituent.
The forgoing cascade cyclizations proceed in a “right-to-left”
direction in which oxygen nucleophiles attack the secondary
centers of the epoxide intermediates. In the next phase of these
studies we examined a “left-to-right” cascade in which attack
occurs at the tertiary centers of the two epoxide rings. The
appropriate bromo substrates for these studies were prepared
along the lines described above in which the Sharpless and Shi
epoxidation methodology provide the requisite chiral epoxides
(Scheme 9). To prepare the R,ω-difunctionalized farnesol
precursor 35, we adapted a sequence previously employed in
our synthesis of the ionomycin segment (Scheme 2) starting
from epoxide 29. Conversion of alcohol 35 to the triepoxy
bromide 38 then proceeded along the lines previously employed
for triepoxy bromide 18. The TBS analogue 39 was prepared
analogously (Supporting Information).
By using the enantiomeric tartrate ligand for the various
Sharpless epoxidation steps we also prepared the isomeric
triepoxy bromides 42/43 and 46 (Supporting Information). Upon
treatment with zinc dust and TBAI in refluxing ethanol all five
of these substrates were cleanly converted to the expected
bistetrahydrofurans 40/41, 44/45, and 47 (Scheme 10). The
cyclizations were uniformly more efficient than the alternative
“right-to-left” variations reflecting the more favorable disposition
of the tertiary epoxide center to Lewis acid-catalyzed ring
opening. It should be noted that the formation of both cis and
trans 1,5-disubstituted tetrahydrofuran rings occurs with com-
parable efficiency in all of these cascade reactions.
The bistetrahydrofuran products of these cyclizations were
produced as virtually single diastereomers in accord with the
expected high enantioselectivity of the Sharpless and Shi
J. Org. Chem. Vol. 73, No. 17, 2008 6755

Marshall and Hann
SCHEME 7
SCHEME 8
SCHEME 9
SCHEME 10
neutral condition with a high degree of stereocontrol. In the
present study we prepared 3 of the 16 possible diasteroisomers
starting from the readily available starting material trans,trans-
farnesol. The remaining isomers should be available along
similar lines through variations in the double bond geometry
of the starting triene and the appropriate Sharpless tartrate ligand
or Shi catalyst. We have shown that both 2,5-cis and -trans
tetrahydrofuran rings form with nearly comparable ease but have
not yet extended the methodology to syn related bistetrahydro-
furan rings. A priori there is no reason to believe that such
extensions would fail.
Experimental Section
(E,E,R,R)-2,3-Epoxy-3,7,11-trimethyl-6,10-dodecadien-1-ol (6).
To a suspension of 4Å MS (0.57 g, 0.13 g/mmol substrate) in 39.2
mL of CH2Cl2 at -20 °C in a Cryobath was added D-(-)-diethyl
tartrate (0.35 g, 1.72 mmol), followed by Ti(i-PrO)4 (0.51 mL, 1.74
mmol). The mixture was stirred at -20 °C for 10 min, at which
time tert-butyl hydroperoxide (1.1 mL, 5-6 M in decanes, ca. 6.60
mmol) was added dropwise. The mixture was stirred at -20 °C
for 20 min, then a solution of trans,trans-farnesol (1) (1.03 g, 4.43
mmol) in 4.0 mL of CH2Cl2 was added dropwise over 15 min. The
mixture was stirred at -20 °C for 1.5 h and then quenched with a
minimal amount of water, diluted with EtOAc, warmed to rt, and
filtered through Celite. The filtrate was concentrated under reduced
pressure to ca. 25 mL. A solution of 30% NaOH/brine (1.9 mL)
was added and the mixture was stirred vigorously for 20 min at rt.
epoxidation reactions. The relative stereochemistry of the
bistetrahydrofurans 40, 44, and 47 was confirmed by their 2D
NOESY spectra (Supporting Information). All three showed
cross-peaks indicative of a C6 H/C7 CH3 interaction. The
spectrum of the cis,anti,cis isomer 40 contained additional cross-
peaks for the C3 CH3/C6 H and C7 CH3/C10 H substituents
while the trans,anti,cis isomer 44 showed only one additional
cross-peak for the C3 CH3/C6 H.
The foregoing three-step cascade cyclization of farnesyl
triepoxides to bistetrahydrofurans proceeds under mild, near-
6756 J. Org. Chem. Vol. 73, No. 17, 2008

Cascade Cyclization Route to Bistetrahydrofurans
The mixture was extracted with EtOAc. The combined organic
extracts were washed with brine, dried over MgSO4, filtered, and
concentrated under reduced pressure. The residue was purified by
flash chromatography (gradient elution with hexanes to 20% EtOAc/
phase. The mixture was then extracted with hexanes. The combined
organic phase was washed with brine, dried over magnesium sulfate,
filtered, and concentrated under reduced pressure. Purification by
flash chromatography on silica gel (elution gradient of hexanes to
2% EtOAc/hexanes, buffered with 0.5% Et3N) afforded triepoxide
hexanes) to afford epoxide 6 as a clear oil (0.98 g, 93%): [R]20
D
+5.8 (c 1.05, CHCl3); lit.12 [R]20D +6.53 (c 4.21, CHCl3); IR (neat)
8 as a viscous, amber oil (0.46 g, 84%): [R]20 +0.3 (c 1.59,
D
1
1
ν 3425 (b), 1666, 1451, cm-1; H NMR (300 MHz, CDCl3) δ
CHCl3); IR (neat) ν 1459, 1249, 891, 651 cm-1; H NMR (500
5.14-5.02 (m, 2H), 3.87-3.76 (m, 1H), 3.72-3.60 (m, 1H), 2.97
(dd, J ) 4.2, 6.8 Hz, 1H), 2.12-2.01 (m, 4H), 2.00-1.92 (m, 2H),
1.74-1.61 (m, 1H), 1.66 (d, J ) 1.1 Hz, 3H), 1.59 (s, 6H), 1.45
(ddd, J ) 5.1, 8.8, 16.6 Hz, 1H), 1.29 (s, 3H); 13C NMR (75 MHz,
CDCl3) δ 136.0, 131.7, 124.4, 123.4, 63.3, 61.7, 61.5, 39.9, 38.7,
26.8, 25.9, 23.8, 17.9, 17.0, 16.2.
MHz, CDCl3) δ 3.54 (ddd, J ) 4.9, 10.3, 14.5 Hz, 1H), 3.25 (ddd,
J ) 7.8, 10.4, 21.6 Hz, 1H), 3.11 (dt, J ) 4.6, 5.6 Hz, 1H), 2.78
(dd, J ) 4.1, 8.0 Hz, 1H), 2.70 (dd, J ) 4.6, 7.6 Hz, 1H), 1.87-1.81
(m, 1H), 1.80-1.74 (m, 1H), 1.73-1.63 (m, 2H), 1.60 (ddd, J )
5.1, 7.6, 12.1 Hz, 4H), 1.33 (s, 3H), 1.30 (s, 3H), 1.28 (s, 3H),
1.26 (s, 3H); 13C NMR (75 MHz, CDCl3) δ 64.0, 62.7, 62.5, 61.3,
60.7, 58.7, 35.4, 35.1, 29.8, 25.1, 24.8, 24.6, 18.9, 16.9, 16.4. Anal.
Calcd for C15H25BrO3: C, 54.06, H, 7.56. Found: C, 54.31; H, 7.43.
(3S,6R,7S,10R)-2,6,10-Trimethyl-3,6;7,10-diepoxy-11-dodecen-
2-ol (9). Purified zinc dust13 (158 mg, 2.42 mmol) and TBAI (133
mg, 0.35 mmol) were added to a stirring solution of bromide 8 (78
mg, 0.24 mmol) in 4.0 mL of absolute ethanol. The heterogeneous
mixture was heated to reflux in an oil bath and stirred for 5.5 h.
Halide exchange could be visually observed on a TLC plate under
UV light. Upon reaction completion, as judged by TLC analysis,
the reaction mixture was cooled to room temperature, diluted with
EtOAc, and filtered through a short pad of Celite and the filtrate
was concentrated under reduced pressure. The crude, yellow syrupy
product mixture was dissolved in a minimal volume of CH2Cl2 and
purified by flash chromatography on silica gel (gradient elution with
hexanes to 7% EtOAc/hexanes) to afford bistetrahydrofuran 9 as a
(E,E,2S,3R)-1-Bromo-2,3-epoxy-3,7,11-trimethyl-6,10-dodeca-
diene (7). To a magnetically stirred solution of epoxy alcohol 6
(0.52 g, 2.2 mmol) in 22 mL of CH2Cl2 at 0 °C was added
diisopropylethylamine (2.29 mL, 13.2 mmol), PPh3 (1.74 g, 6.6
mmol), and carbon tetrabromide (2.21 g, 6.6 mmol). The mixture
was stirred overnight at 0 °C in a Cryocool bath. The reaction
mixture was concentrated under reduced pressure and eluted through
a short plug of silica gel. The plug was washed with 300 mL of
1:4 ethyl acetate in hexanes and the filtrate was concentrated under
reduced pressure. Purification of the crude yellow syrup by flash
chromatography on silica gel (gradient elution with hexanes to 2%
EtOAc/hexanes) afforded bromide 7 as a clear pale-yellow oil (0.63
g, 95%): [R]20 -12.3 (c 1.70, CHCl3); IR (neat) ν 1667, 1248
D
cm-1; H NMR (300 MHz, CDCl3) δ 5.17-5.01 (m, 2H), 3.53
1
(dd, J ) 5.9, 10.4 Hz, 1H), 3.23 (dd, J ) 7.7, 10.4 Hz, 1H), 3.08
(dd, J ) 5.9, 7.7 Hz, 1H), 2.17-2.01 (m, 4H), 2.01-1.93 (m, 2H),
1.78-1.68 (m, 1H), 1.67 (d, J ) 1.1 Hz, 3H), 1.63-1.55 (m, 6H),
1.44 (ddd, J ) 7.6, 9.0, 13.7 Hz, 1H), 1.30 (s, 3H); 13C NMR (75
MHz, CDCl3) δ 136.1, 131.6, 124.4, 123.3, 63.3, 61.7, 39.9, 38.6,
30.0, 26.8, 25.9, 23.9, 17.9, 16.3, 16.2.
light-yellow syrup (38 mg, 63%): [R]20 +14.9 (c 1.32, CHCl3);
D
1H NMR (500 MHz, CDCl3) δ 5.90 (ddd, J ) 9.4, 17.4, 27.9 Hz,
1H), 5.24-5.12 (m, 1H), 4.98 (dd, J ) 1.2, 10.8 Hz, 1H), 4.08
(dd, J ) 5.9, 9.4 Hz, 1H), 3.85 (dd, J ) 5.3, 7.7 Hz, 1H), 2.19-2.05
(m, 2H), 2.01-1.86 (m, 3H), 1.84-1.76 (m, 1H), 1.58 (ddd, J )
3.6, 10.3, 11.6 Hz, 1H), 1.51 (ddd, J ) 4.9, 8.6, 11.1 Hz, 1H),
1.31 (s, 3H), 1.24 (s, 3H), 1.15 (s, 3H), 1.07 (s, 3H); 13C NMR
(126 MHz, CDCl3) δ 143.7, 111.6, 85.6, 85.4, 84.3, 83.1, 71.8,
37.4, 31.4, 28.6, 27.9, 26.4, 26.2, 25.1, 24.8; HRMS (ESI, M +
Na+) calcd for [C15H26O3 + Na]+ 277.1780, found 277.1772.
(2S,3R,6R,7R,10R)-1-Bromo-3,7,11-trimethyl-2,3;6,7;10,11-
triepoxydodecane (8). To a magnetically stirred solution of bromide
7 (0.40 g, 1.6 mmol) in 40.7 mL of 2:1 dimethoxymethane:
acetonitrile were added successively
a buffer solution of
Na2B4O7 ·10H2O (16.3 mL, 0.05 M in 0.4 mM aqueous Na2EDTA),
Bu4NHSO4 (57 mg, 0.17 mmol), and the D-fructose derived Shi
catalyst (0.92 g, 3.6 mmol). The resulting mixture was stirred at
room temperature for 5 min and then cooled to 0 °C in an ice bath.
A solution of Oxone (4.59 g, 7.5 mmol) in 34.8 mL of 0.4 mM
aqueous Na2EDTA and a solution of K2CO3 (4.51 g, 32.6 mmol)
in 34.8 mL of water were simultaneously added dropwise, using a
syringe pump over 1.5 h at 0 °C. Once the two solutions were
completely added, the reaction mixture was diluted with 50 mL of
hexanes. The biphasic mixture was poured into a separatory funnel
and NaCl was added until a slight slurry formed in the aqueous
Acknowledgment. We are indebted to Eli Lilly and Company
for partial support of this investigation.
Supporting Information Available: Experimental proce-
dures for all new compound not described in the print version
and copies of 1H and 13C NMR spectra. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO801188W
(13) Zinc dust was purified by following the procedure described in the
following: Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals, 3rd ed.; Butterworth-Heinemann Ltd: Boston, MA, 1994; p 360.
(12) Kigoshi, H.; Ojika, M.; Shizuri, Y.; Niwa, H.; Yamada, K. Tetrahedron
Lett. 1982, 23, 5413.
J. Org. Chem. Vol. 73, No. 17, 2008 6757