Highly regioselective copper-catalyzed cis- and trans-1-propenyl Grignard cleavage of hindered epoxides. Application in propionate synthesis

Article abstract of DOI:10.1021/jo060833t

Hindered protected and unprotected epoxy alcohols were regioselectively cleaved using copper-catalyzed cis- and trans-1-propenylmagnesium bromide. The reaction exhibited good yield and excellent regioselectivity in systems where organocuprates and organoalanes failed. The cis Grignard reagent displayed no double-bond isomerization, whereas the trans isomer showed partial trans-to-cis equilibration, which was minimized by controlling the reagent formation conditions. The reaction was shown to be highly useful for the elaboration of the C10-C15 Streptovaricin D ansa chain fragment.

Full text of DOI:10.1021/jo060833t

frequently, vinyl groups. Nonetheless, the regioselective ring
opening of epoxides by Grignard reagents has been restricted
mostly to unhindered epoxides and those activated by an
adjacent vinyl or aryl group.4,5 Disubstituted epoxides present
inherent regioselectivity problems, which are also found with
other organometallic reagents, such as organocuprates6 and
organoalanes.7 Moreover, given the presence of Lewis acids
(magnesium halide salts) associated with the Schlenk equilib-
rium,8 the formation of side products (halohydrin, rearrange-
ment, elimination, etc.) may also be observed when an epoxide
is treated with a Grignard reagent.9
Highly Regioselective Copper-Catalyzed cis- and
trans-1-Propenyl Grignard Cleavage of Hindered
Epoxides. Application in Propionate Synthesis
David Rodr´ıguez, Marlenne Mulero, and Jose´ A. Prieto*
Department of Chemistry, UniVersity of Puerto Rico, R´ıo
Piedras Campus, P.O. Box 23346, San Juan,
Puerto Rico 00931-3346
japrieto@uprrp.edu
The cleavage of disubstituted epoxides represents a valuable
transformation for the stereoselective generation of two contigu-
ous stereogenic centers. In this regard, we recently reported a
study on the cleavage of hindered 3,4-epoxy alcohols by an
alkynyl aluminum reagent, as part of our quest for developing
a general epoxide-based methodology for the preparation of
polypropionates.7e In our three-step approach, an epoxide is
submitted to a sequence of diethylpropynylalane-mediated
oxirane cleavage, alkyne reduction, and stereoselective epoxi-
dation to yield a 3,4-epoxy alcohol (Scheme 1). This methodol-
ogy may be repeated to produce a new 3,4-epoxy alcohol, which
allows for chain elongation in a reiterative fashion.
This successful approach resulted in a series of diastereomeric
stereotetrads; nonetheless, some limitations were encountered
in the regioselectivity of the cleavage of some epoxy alcohol
precursors. For example, the trans-epoxy alcohol 1 was ef-
fectively cleaved under the alane reaction conditions to produce
stereotetrad 2 with high regioselectivity (89:11) and a reasonable
68% yield after separation; however, the closely related epoxy
alcohol 5 afforded a 56:44 mixture of the expected 1,3-diol 6
and the 1,4-diol 7 in a very low yield (ca. 28%, Scheme 2).
Moreover, when the epoxy alcohol 1 was protected as the methyl
ether 3 and subjected to the alkynyl alane conditions, no reaction
was observed. Epoxy alcohol 3 represents a potentially useful
precursor for the introduction of the C(27) methoxy group
present in the rifamycin S ansa chain.
ReceiVed April 20, 2006
Hindered protected and unprotected epoxy alcohols were
regioselectively cleaved using copper-catalyzed cis- and
trans-1-propenylmagnesium bromide. The reaction exhibited
good yield and excellent regioselectivity in systems where
organocuprates and organoalanes failed. The cis Grignard
reagent displayed no double-bond isomerization, whereas the
trans isomer showed partial trans-to-cis equilibration, which
was minimized by controlling the reagent formation condi-
tions. The reaction was shown to be highly useful for the
elaboration of the C10-C15 Streptovaricin D ansa chain
fragment.
Grignard reagents are among the most classical organome-
tallic reagents which continue to see extensive use in C-C bond
formation reactions.1 In addition to their typical use in carbonyl
and related nucleophilic addition reactions, they have found
widespread applications in transition-metal-catalyzed cross-
coupling reactions.2,3 Grignard reagents have also been used as
nucleophiles to perform the cleavage of oxiranes.4 In this regard,
copper(I) salts have been shown to catalyze the addition of
Grignard reagents to epoxides to introduce alkyl, allyl, and, less
These findings prompted us to explore other organometallic
pathways to overcome these limitations and to expand the
generality of our epoxide-based approach to polypropionates.
(5) (a) Martin, L. D.; Stille, J. K. J. Org. Chem. 1982, 47, 3630-3633.
(b) Tius, M. A.; Fauq, A. H. J. Org. Chem. 1983, 48, 4131-4132. (c) Erdik,
E. Tetrahedron 1984, 40, 641-657. (d) Millar, J. G.; Underhill, E, W. J.
Org. Chem. 1986, 51, 4726-4728. (e) Pineschi, M. New J. Chem. 2004,
28, 657-665.
(6) (a) Lipshutz, B. H.; Barton, J. C. J. Org. Chem. 1988, 53, 4495. (b)
Marshall, J. A. Chem. ReV. 1989, 89, 1503-1511. (c) Lipshutz, B. H.;
Barton, J. C. J. Am. Chem. Soc. 1992, 114, 1084-1086. (d) Guanti, G.;
Merlo, V.; Narisano, E. Tetrahedron 1994, 50, 12245-12258. (e) Chauret,
D. C.; Chong, J. M.; Ye, Q. Tetrahedron: Asymmetry 1999, 10, 3601-
3614.
(7) (a) Chini, M.; Crotti, P.; Flippin, L. A.; Gardelli, C.; Giovani, E.;
Macchia, F.; Pineschi, M. J. Org. Chem. 1993, 58, 1221-1227. (b) Sasaki,
M.; Tanino, K.; Miyashita, M. J. Org. Chem. 2001, 66, 5388-5394. (c)
Pena, P. C. A.; Roberts, S. M. Curr. Org. Chem. 2003, 7, 555-571. (d)
Restorp, P.; Somfai, P. Chem. Commun. 2004, 2086-2087. (e) Tirado, R.;
Torres, G.; Torres, W.; Prieto, J. A. Tetrahedron Lett. 2005, 797-801. (f)
Lewinski, J.; Horeglad, P.; Tratkiewicz, E.; Justyniak, I.; Ochal, Z. J.
Organomet. Chem. 2005, 690, 3697-3699.
(1) (a) Wakefield, B. J. Organomagnesium Methods in Organic Synthesis;
Academic Press: London, 1995. (b) Handbook of Grignard Reagents;
Silverman, G. S., Rakita, P. E., Eds.; Marcel Dekker: New York, 1996.
(c) Grignard Reagents: New DeVelopments; Richey, H. G., Jr., Ed.;
Wiley: New York, 1999.
(2) Recent reviews: (a) Knochel, P.; Dohle, W.; Gommermann, N.;
Kneisel, F. F.; Kopp, F.; Korn, T.; Sapountzis, I.; Vu, V. A. Angew. Chem.,
Int. Ed. 2003, 42, 4302-4320. (b) Shinokubo, H.; Oshima, K. Eur. J. Org.
Chem. 2004, 2081-2091.
(3) Recent applications: (a) Rottlander, M.; Boymond, L.; Berillon, L.;
Lepretre, A.; Varchi, G.; Avolio, S.; Laaziri, H.; Queguiner, G.; Ricci, A.;
Cahiez, G.; Knochel, P. Chemistry 2000, 6, 767-770. (b) Scheiper, B.;
Bonnekessel, M.; Krause, H.; Fuerstner, A. J. Org. Chem. 2004, 69, 3943-
3949. (c) Itami, K.; Higashi, S.; Mineno, M.; Yoshida, J. Org. Lett. 2005,
7, 1219-1222. (d) Limmert, M. E.; Roy, A. H.; Hartwig, J. F. J. Org.
Chem. 2005, 70, 9364-9370.
(4) (a) Taber, D. F.; Green, J. H.; Geremia, J. M. J. Org. Chem. 1997,
62, 9342-9344. (b) Hodgson, D. M.; Fleming, M. J.; Stanway, S. J. J. Am.
Chem. Soc. 2004, 126, 12250-12251. (c) Tanaka, T.; Hiramatsu, K.;
Kobayashi, Y.; Ohno, H. Tetrahedron 2005, 61, 6726-6742.
(8) Schlenk, W.; Schlenk, W., Jr. Ber. Dtsch. Chem. Ges. 1929, 62, 920-
924.
(9) (a) Sarangi, C.; Das, N. B.; Nanda, B.; Nayak, A.; Sharma, R. P. J.
Chem. Res. 1997, 180. (b) Bonini, C.; Righi, G. Synlett 1992, 204-206.
10.1021/jo060833t CCC: $33.50 © 2006 American Chemical Society
Published on Web 06/27/2006
5826
J. Org. Chem. 2006, 71, 5826-5829

SCHEME 1. General Reiterative Epoxide-Based Approach
to Polypropionates
without detectable double-bond isomerization. The retention of
the propenyl geometry was established by its 13C NMR
spectrum, which exposed the diagnostic signal for the allylic
methyl carbon at 13.0 ppm.11 Similarly, the out-of-plane infrared
absorption band at 680 cm-1 corroborates this result. The cis-
propenyl-substituted alcohol 8 corresponds to the C(24)-C(27)
stereotetrad fragment of rifamycin S.
This success taken together with the evident lack of reactivity
of epoxide 3 toward the organoalane and organocuprate reagents
encouraged us to investigate in more detail the copper-mediated
cleavage of other epoxides of synthetic interest for our polypro-
pionate studies12 by cis-1-propenylmagnesium bromide (Table
1). In the case of the primary epoxide 9 (used as a control
substrate), the reaction occurred as expected in good yield (ca.
>70%) and without detectable double-bond isomerization (entry
1). Next, we focused our attention on the cis- and trans-
disubstituted epoxides 13 and 15, which correspond to the
starting substrates used in our epoxide-based polypropionate
synthesis. Again, the epoxides were cleaved in comparable
yields and excellent regioselectivity (entries 4 and 5). The less
hindered benzyl- and SEM-protected epoxides 17 and 19 showed
somewhat diminished regioselectivity (entries 6 and 7). These
results demonstrate that useful regioselectivities are attainable
even in systems with less significant steric bias. Fortunately,
the monoprotected epoxy diol 21 exhibited excellent regiose-
lectivity, which produced only the alkene diol 22 in 85% yield
(entry 9). Diol 22 represents a potentially useful precursor for
the synthesis of the streptovaricin D polypropionate chain, which
possesses an unusual carbomethoxy group at the C(10) position.
To evaluate the extent of the catalytic nature of the reaction,
a series of catalyst loading versus concentration of the Grignard
reagent was studied. The cleavage was first performed with the
unhindered epoxide 9 as a model. The copper catalyst was varied
in the range from 5 to 130 mol %. The reaction proceeded
smoothly to afford the homoallylic alcohol in comparable yields
for all attempts, even at a catalyst loading of 5 mol %. To discard
the possibility that the Grignard reagent alone might promote
the epoxide cleavage, as has been observed in some allylic
cases,4a the reaction was performed without the catalyst. After
several attempts, the Grignard addition was not observed and
only the bromohydrin product was formed, which confirms
copper catalysis. To further examine this aspect, the catalyst
loading was lowered to 10 mol % for compounds 11, 19, and
3 (Table 1, entries 3, 8, and 11). Under these conditions, similar
yields (and regioselectivity in the case of 19) were obtained.
Since the free hydroxy group is required for the cleavage of
3,4-epoxy alcohols by propynyl alane reagents (Scheme 2), we
decided to apply the copper-catalyzed Grignard conditions (10
mol % of CuI) to the epoxy alcohol 1. In this case, the expected
1,3-diol product 23 was obtained as the only regioisomer in
good yield (entry 12). This result shows an improved regiose-
lectivity for epoxy alcohol 1 compared to the one obtained using
the propynylalane conditions (89:11).7e The examples in entries
9 and 12 manifest that the copper-catalyzed propenyl Grignard
SCHEME 2. Alkynyl Alane Approach to the Cleavage of
Disubstituted Epoxide
SCHEME 3. Cleavage of Epoxide 3 by cis-1-Propenyl
Organometallic Reagents
Herein we describe a general procedure for the reaction of
disubstituted epoxides with readily available propenyl Grignard
reagents under Cu catalysis to form directly stereodefined
alkenyl propionates.
After the unsuccessful attempts to cleave epoxide 3 by means
of a propynylalane, we decided to explore the cleavage with a
propenyl organometallic reagent (Scheme 3). A potential
advantage of this concept would be direct formation of a
homoallylic alkenol (8) in one step, without the need for the
subsequent alkyne to alkene reduction required in the alkynyl
alane method.7e We first explored higher-order propenyl orga-
nocuprates, which have been used in related but less hindered
systems, although with moderate regioselectivity.6a,c Thus, when
epoxide 3 was subjected to a higher-order cis-1-propenyl
organocuprate reagent, only the starting epoxide was recovered,
even after an extensive variation of the reaction conditions, the
preparation of the organometallic reagent,6 and the order of
addition.10
We then proceeded to test the cleavage of epoxide 3 with an
alkenyl Grignard reagent. To our delight, when epoxide 3 was
exposed to cis-1-propenylmagnesium bromide (6 equiv) in the
presence of CuI (1.3 equiv), the desired homoallylic alcohol 8
was obtained in 85% yield and excellent regioselectivity.
Moreover, the reaction produced exclusively the cis alkenol
(10) Nicolaou, K. C.; Buggan, M. E.; Ladduwahetty, T. Tetrahedron Lett.
1984, 25, 2069.
(11) All compounds were fully characterized by 1D and 2D NMR (COSY
and HMQC) methods. For known compound, the spectral data were
compared with the literature values reported in refs 7e and 12.
(12) (a) Tirado, R.; Prieto, J. A. J. Org. Chem. 1993, 58, 5666-5673.
(b) Torres, G.; Torres, W.; Prieto, J. A. Tetrahedron 2004, 60, 10245-
10251. (c) Arbelo, D. O.; Prieto, J. A. Tetrahedron Lett. 2002, 43, 4111-
4114.
J. Org. Chem, Vol. 71, No. 15, 2006 5827

TABLE 1. Yields and Regioselectivities of the Cu-Catalyzed
Cleavage of Epoxides with cis-1-Propenylmagnesium Bromide
TABLE 2. cis/trans Product Isomerization Studies for Epoxides 9
and 11
propenyl
time
trans/cis
entry geometrya epoxide
T (°C)
25
25
25
0 to 25
-25 to 25
-25 to 25
25
(min)b product
ratioc
1
2
3
4
5
6
7
8
trans
trans
trans
trans
trans
trans
cis
9
11
9
9
9
11
9
11
60
60
10a,b
12a,b
10a,b
10a,b
10a,b
12a,b
10a
50:50
50:50
63:37
77:23
86:14
71:29
<5:95f
6:94
<1d
40e
40e
40e
360
180
trans/cisg
25
12a,b
a The cis/trans geometry of the 1-bromopropene staring material was
determined by NMR spectroscopy. b The Grignard reagent was prepared
in THF at room temperature and was allowed to stir for the specified time
and temperature after its initial reflux before being transferred to the CuI/
ether slurry at -78 °C. c The trans/cis ratio was determined by NMR
spectroscopy. d With initial reflux and immediate transfer of the reaction
mixture to the CuI/ether slurry. e No initial reflux; after the formation of
the Grignard reagent had been initiated, the reaction mixture was cooled.
f No other isomer was detected. g 40:60 trans/cis.
a 1:1 cis/trans mixture of ring-opened products was obtained
(entry 2). These unexpected results urged us to explore a
systematic temperature and time variation for the formation of
the propenyl Grignard reagent. When the organometallic reagent
was quickly transferred to the copper iodide slurry (immediately
after the initial reflux) and treated with the epoxide 9, a mixture
of cis and trans alkenols was again formed in a moderate
63:37 trans/cis selectivity. The Grignard reagent was then
prepared without initial reflux (cooling to 0 or -25 °C after
initial reagent formation). The cleavage occurred with improved
stereoselectivity (entries 4-6); for example, a 86:14 trans/cis
ratio was obtained for epoxide 9. An attempt to prepare the
Grignard reagent at an initial temperature below room temper-
ature did not produce the organometallic reagent (data not shown
in Table 2). These results reveal that the double-bond isomer-
ization occurs during the formation of the trans-propenyl
Grignard reagent. It depends on the initial temperature at which
the reagent is prepared (reflux versus temperature control), as
well as on the time it is allowed to stir before transfer to the
copper slurry. Moreover, the isomerization process also takes
place, but slowly, during the epoxide cleavage step, as suggested
by the differing trans/cis product ratio for epoxides 9 and 11.
Since the secondary epoxide 11 reacts slower than 9, a lower
trans/cis ratio is expected because there is more time for the
isomerization.
a Prepared by the procedures in ref 12. b Condition A: 6 equiv of
Grignard reagent, 1.3 equiv of CuI, -78 °C. Condition B: 6 equiv of
Grignard reagent, 0.1 equiv of CuI, -78 °C. c Characterized by NMR
spectroscopy. d Isolated product yield after distillation or chromatography.
e Ratio of external epoxide carbon attack versus internal attack, when
observed. f A 2:1 mixture of cleavage product/starting material eluted from
the chromatography column. g Yield of crude product; the diol partially
decomposed on a silica gel column. The alcohol was transformed to the
bistriethylsilyl ether for purification and characterization.
cleavage of hindered epoxides does tolerate a free hydroxy
group, diminishing the need for a protection-deprotection
sequence.
To test the possibility of double-bond isomerization by
starting from the cis-propenyl Grignard reagent, a solution was
prepared as usual and was allowed to stir for up to 6 h at room
temperature before treatment with epoxide 9. Again, only the
cis-alkenol product was observed (entry 7). Even more signifi-
cant, when a commercially available 40:60 trans/cis mixture
of 1-bromopropenes was used to prepare the Grignard reagent
and treated with epoxide 10 for 3 h, a 6:94 trans/cis mixture
was obtained (entry 8).
A similar contra-thermodynamic double-bond isomerization
of trans-1-propenylmagnesium bromide to its cis isomer, but
not the reverse, was reported by Beak13 in a detailed mechanistic
study on the addition of trans- and cis-propenyl organometallic
reagents to carbonyl and thiocarbonyl compounds. In this
The Grignard reagent derived from trans-1-bromopropene
was studied next under the same conditions as employed for
the cis-propenyl Grignard reagent, to produce the complemen-
tary trans homoallylic alcohols. Although no major complica-
tions were expected, when the trans-1-propenylmagnesium
bromide was added to epoxide 9, a 1:1 cis/trans mixture was
produced. Consequently, to elucidate this unexpected isomer-
ization process, we examined the model epoxides 9 and 11
(Table 2). The propenyl cis/trans geometry of the products was
again established by 13C NMR spectroscopy, which exposed
the diagnostic allylic methyl signals at 12.8-13.5 ppm, com-
pared to 17.9-18.1 ppm for the trans isomers.
When epoxide 11 was treated with the trans Grignard reagent
under the standard copper-mediated conditions (1.3 equiv), also
5828 J. Org. Chem., Vol. 71, No. 15, 2006

SCHEME 4. Reaction of Other Alkenyl Grignard
Derivatives with Disubstituted Epoxides
an alcohol oxidation state. This fortunate result demonstrates
the scope and efficacy of the copper-catalyzed Grignard
methodology as a convenient means and as a feasible tool of
regioselectively cleaving hindered epoxides for the elaboration
of stereodefined polypropionate chains.
The work presented herein, especially the good yields and
high regioselectivities, offers a convenient and promising
synthetic methodology of applying the copper-catalyzed Grig-
nard reaction to the cleavage of hindered disubstituted epoxides.
We emphasize the fact that this useful regio- and stereocon-
trolled protocol for the synthesis of polypropionate is shorter
than the alkynyl alane method and solves the regioselectivity
problems encountered with the latter.
SCHEME 5. Application of the Copper-Catalyzed Grignard
Approach for the Cleavage of Epoxide 26
Experimental Section
Representative Procedure for the Copper-Catalyzed Cleavage
of Epoxides Using Alkenylmagnesium Bromides. Preparations
of (2R*,3S*,4S*,5R*,6Z)-2-Methoxy-3,5-dimethyl-1-[(triisopro-
pylsilyl)oxy]-6-octen-4-ol (8). Mg powder (0.22 g atom, 5.4 g, 14
equiv) was added to a round-bottom flask assembled with a dry
ice condenser. Freshly distilled THF (100 mL, 2.3 M with respect
to Mg) was added to the flask and spiked with I2. After several
minutes, the reaction mixture turned from yellow to turbid gray.
Then, the dry ice condenser was cooled to -78 °C and cis-1-
bromopropene (11.38 g, 94 mmol, 6 equiv) was added. After several
minutes, the reaction warmed and refluxed vigorously. The reaction
was stirred for 1 h. The solution was transferred via a double-ended
needle to a three-neck round-bottom flask containing CuI (0.30 g,
1.6 mmol, 0.10 equiv) in ether (40 mL) at -78 °C. The CuI was
previously heated overnight under reduced pressure to remove
moisture. After 30 min, the epoxide (5.00 g, 15.8 mmol) was added
at -78 °C. The reaction mixture was allowed to reach room
temperature and was stirred overnight. At this point, the reaction
mixture was quenched with a saturated solution of NH4Cl (40 mL)
and diluted in ether (50 mL). The layers were separated, and the
aqueous phase was extracted with ether (2 × 10 mL). The combined
organic phase was dried over MgSO4, filtered through Celite, and
concentrated under reduced pressure to yield the crude alkenol. The
crude product was purified by silica gel flash chromatography (12:1
hexane/EtOAc) providing 4.25 g (75%) of pure product. 1H NMR
(CDCl3): δ 5.40 (dq, J ) 11.0, 6.7 Hz, 1H), 5.23 (ddq, J ) 11.0,
8.1, 1.6 Hz, 1H), 3.85 (dd, J ) 12.2, 7.9 Hz, 1H), 3.63 (dd, J )
12.2, 5.3 Hz, 1H), 3.60 (ddd, J ) 7.9, 5.3, 2.6 Hz, 1H), 3.50 (s,
3H), 3.27 (dd, J ) 7.7, 4.4 Hz, 1H), 2.60 (m, 1H), 1.92 (m, 1H),
1.65 (dd, J ) 6.7, 1.6 Hz, 3H), 1.05 (d, J ) 6.4 Hz, 3H), 1.05
(21H), 0.99 (d, J ) 7.2 Hz, 3H). 13C NMR (CDCl3): δ 133.7 (CH),
122.9 (CH), 83.3 (CH), 80.0 (CH), 64.1(CH2), 58.4 (CH3), 36.1
(CH), 35.7 (CH), 17.9 (CH3), 16.6 (CH3), 13.0 (CH3), 12.0 (CH3),
11.8 (CH). IR (neat, cm-1): ν 3500 (OH), 2940-2865 (CH), 1109,
1066 (C-O), 881 (Si-O), 680 (HCdCH). HRMS-FAB (m/z):
[M+H]+ calcd for C20H43O3Si, 359.2981; found, 359.2978. Anal.
Calcd for C20H42O3Si: C, 66.98; H, 11.80. Found: C, 67.33; H,
11.60.
investigation, only about 15% trans-to-cis propenyl isomeriza-
tion was obtained (in our examples, the trans-to-cis isomeriza-
tion is as much as 50%!) for the trans-configured Grignard
reagent, whereas complete retention was found for the cis
reagent. Although no detailed explanation for this unusual
behavior was offered, the authors ascribed this trend to the
difference in reactivity between the propenyl Grignard diaster-
eomers (the cis isomer reacts 1.7 times faster than the trans).
The greater extent of double-bond isomerization observed by
us for the epoxide cleavage reaction compared to Beak’s
nucleophilic addition to thioketones relates again to the lower
reactivity of the epoxides since for the latter several hours are
required for completion at low temperature (-78 °C to room
temperature).
To further assess the scope of the copper-catalyzed epoxide
cleavage reaction, we studied two other potentially useful vinyl
Grignard derivatives. These included vinylmagnesium bromide
and R-(trimethylsilyl)vinylmagnesium bromide (Scheme 4).5a
In both cases, the disubstituted epoxides were cleaved in good
yields (73-75%) and excellent regioselectivities. The vinyl TMS
group provides a potentially useful access to other vinyl
derivatives, such as R-substituted vinyl bromides.5a These
examples reveal that this copper-catalyzed alkenyl Grignard
cleavage of hindered epoxides is not limited to 1-propenyl
Grignard reagents, showing the potential for expansion to other
alkenyl Grignard derivatives.
As a final point, we decided to test this copper-catalyzed cis-
propenyl Grignard approach on the hindered epoxide 26, related
to the currently ongoing synthesis of the Streptovaricin D ansa
chain.14 The challenge here is that several previous attempts to
cleave the epoxide ring in 26 were unsuccessful with the alkynyl
alane protocol (Scheme 5). Indeed, the epoxide 26 was ef-
ficiently cleaved to form the tripropionate fragment 27 in good
yield (74%), without any detectable regioisomer or double-bond
isomerization.11 Product 27, which has six contiguous chirality
centers, corresponds to the C15-C10 fragment of the strepto-
varicin D ansa chain and contains the C(10) methyl group in
Acknowledgment. This work was supported by NIH RISE
(1R25-GM-61151-01A1) and SCORE (2S06GM-08102-29)
Programs. We thank E. Valent´ın and W. Torres for their work
on epoxides 19 and 26, and D. Castillo for her help with various
starting materials. Partial support for the NMR facility was
provided by the UPR-MCC. We also thank Professor Waldemar
Adam for fruitful discussions.
Supporting Information Available: Experimental details,
spectroscopic data, and copies of 1H and 13C NMR for all epoxide
cleavage products. This material is available free of charge via the
Internet at http://pubs.acs.org.
(13) Beak, P.; Yamamoto, J.; Upton, C. J. J. Org. Chem. 1975, 40, 3052-
3062.
(14) Epoxide 26 was prepared from epoxide 1 by the corresponding
alkynyl alane methodology (refs 12a,b).
JO060833T
J. Org. Chem, Vol. 71, No. 15, 2006 5829