A general oxidative cyclization of 1,5-dienes using catalytic osmium tetroxide

Article abstract of DOI:10.1002/anie.200390253

Dihydroxylation of 1,5-dienes under acidic conditions is sufficient to induce an efficient oxidative cyclization that produces stereochemically defined tetrahydrofurans in one step [Eq. (1)]. This method is very general and efficient for a range of dienes

Full text of DOI:10.1002/anie.200390253

Communications
Roush, A. E. Walts, L. K. Hoong, J. Am. Chem. Soc. 1985, 107,
8186; h) W. R. Roush, W. L. Banfi, J. Am. Chem. Soc. 1988, 110,
3979; i) M. T. Reetz, T. Zierke, Chem. Ind. 1988, 663; j) R. P.
Short, S. Masamune, J. Am. Chem. Soc. 1989, 111, 1892; k) E. J.
Corey, C.-M. Yu, S. S. Kim, J. Am. Chem. Soc. 1989, 111, 5495;
l) J. W. Faller, D. L. Linebarrier, J. Am. Chem. Soc. 1989, 111,
1937; m) A. Hafner, R. O. Duthaler, R. Marti, G. Rihs, Rothe-P.
Streit, F. Schwarzenbach, J. Am. Chem. Soc. 1992, 114, 2321.
[2] For a recent review of enantioselective Lewis acid catalyzed
allylmetal additions, see: a) A. Yanagisawa in Comprehensive
Asymmetric Catalysis, Vol. II (Eds.: E. N. Jacobsen, A. Pfaltz, H.
Yamamoto), Springer, Heidelberg, 1999, chap. 27. Enantioselec-
tive Lewis base catalysis: b) S. E. Denmark, D. M. Coe, N. E.
Pratt, B. D. Griedel, J. Org. Chem. 1994, 59, 6161; c) S. E.
Denmark, J. Fu, J. Am. Chem. Soc. 2000, 122, 12021; d) S. E.
Denmark, J. Fu, J. Am. Chem. Soc. 2001, 123, 9488; e) K. Iseki, Y.
Kuroki, M. Takahashi, Y. Kobayashi, Tetrahedron Lett. 1996, 37,
5149; f) K. Iseki, Y. Kuroki, M. Takahashi, S. Kishimoto, Y.
Kobayashi, Tetrahedron 1997, 53, 3513; g) K. Iseki, S. Mizuno, Y.
Kuroki, Y. Kobayashi, Tetrahedron Lett. 1998, 39, 2767; h) K.
Iseki, S. Mizuno, Y. Kuroki, Y. Kobayashi, Tetrahedron 1999, 55,
977; i) M. Nakajima, M. Saito, M. Shiro, S. Hashimoto, J. Am.
Chem. Soc. 1998, 120, 6419; j) A. V. Malkov, M. Orsini, D.
Pernazza, K. W. Muir, V. Langer, P. Meghani, P. Kocovsky, Org.
Lett. 2002, 4, 1047.
reaction time (typically 20 h) would ideally be shorter, and
aromatic and conjugated aldehydes tend to require even
longer reaction times (Table 3), the four conditions for an
ideal reagent outlined above have otherwise been met fully. In
this context it is especially noteworthy, and bears repeating,
that reagent 3 is a readily prepared stable solid that may be
briefly handled in air with no apparent decomposition, and
may be stored in a freezer under N2 or Ar for long periods of
time (> 1 month). Investigations into the mechanistic basis
for the sluggish reactivity of some aromatic and conjugated
aldehydes and a method to overcome this limitation have
been initiated.
Experimental Section
Preparation of reagent (R,R)-3: To a cooled (08C) solution of
allyltrichlorosilane (2.05 mL, 14.1 mmol) and 1,8-diazabicy-
clo[5.4.0]undec-7-ene (4.24 mL, 28.4 mmol) in dichloromethane
(50 mL) was added (R,R)-N,N’-bis-(4-bromobenzyl)cyclohexane-
1,2-diamine (5.37 g, 11.9 mmol) in dichloromethane (20 mL) over
50 min. After 2 h, the mixture was warmed to room temperature, and
was stirred for 13 h. The reaction mixture was concentrated. After
diethyl ether (60 mL) was added, the mixture was stirred for 1 h and
filtered through a pad of celite, and the residue was washed with
diethyl ether (2 ” 10 mL). The filtrate was concentrated. Benzene
(10 mL) was added, and the solution was concentrated. This
procedure was repeated to give the product as an oil (5.37 g, 88%).
Upon standing (under Ar) in a freezer, the oil solidified to a white
solid that may be stored in a freezer (under Ar) and used as needed.
1H NMR (300 MHz, C6D6): d = 7.43 (d, J = 8.4 Hz, 2H; Ar-H), 7.42
(d, J = 8.4 Hz, 2H; Ar-H), 7.18 (d, J = 8.5 Hz, 2H; Ar-H), 7.17 (d, J =
[3] J. W. A. Kinnaird, P. Y. Ng, K. Kubota, X. Wang, J. L. Leighton, J.
Am. Chem. Soc. 2002, 124, 7920.
[4] The cyclohexane diamine is treated with the appropriate sub-
stituted benzaldehyde to give the bis-Schiff base, which is reduced
to the bis-(p-X-benzyl)diamine. The diamine is then treated with
allyltrichlorosilane to give the active reagent. See the Supporting
Information.
[5] a) M. T. Reetz, K. Kesseler, A. Jung, Tetrahedron Lett. 1984, 25,
729. For a comprehensive discussion regarding the tendency of b-
alkoxyalkanals to favor the anti diastereomer in allylation and
aldol reactions, see: b) D. A. Evans, M. J. Dart, J. L. Duffy, M. G.
Yang, J. Am. Chem. Soc. 1996, 118, 4322.
¼
¼
8.5 Hz, 2H; Ar-H), 5.72 (m, 1H; CH CH2), 5.00–4.92 (m, 2H; CH
CH2), 3.98 (d, J = 16.2 Hz, 1H; one of NCH2Ar), 3.95 (d, J = 15.1 Hz,
1H; one of NCH2Ar), 3.65 (d, J = 15.1 Hz, 1H; one of NCH2Ar), 3.64
(d, J = 16.2 Hz, 1H; one of NCH2Ar), 2.63–2.75 (m, 2H; two of
¼
CHN), 1.42–1.79 (m, 6H; four of Cy and SiCH2CH CH2), 0.83–
1.05 ppm (m, 4H; four of Cy); 13C NMR (75 MHz, C6D6): d = 141.7,
140.7, 131.7, 131.4, 130.3, 129.5, 128.7, 121.2, 120.9, 116.6, 66.8, 65.8,
48.3, 47.5, 31.1, 30.7, 25.1, 25.0 ppm; 29Si NMR (60 MHz, C6D6): d =
À4.4 ppm.
General procedure for the reaction of (R,R)-3 with aldehydes: To
a cooled (À108C) solution of (R,R)-3 in CH2Cl2 (0.2m) was added the
aldehyde (1.0 equiv). The reaction mixture was transferred to a
freezer (À108C) and maintained at that temperature for 20 h. To this
cooled solution was added 1n HCl and EtOAc, and the mixture was
vigorously stirred at room temperature for 15 min. The layers were
separated and the aqueous layer was extracted with EtOAc three
times. The combined organic layers were diluted with hexane, dried
(MgSO4), filtered, and concentrated. The homoallylic alcohol prod-
ucts may be purified further by chromatography on silica gel. All
yields listed in Tables 1–3 are for chromatographed, analytically pure
material.
One-Step Route to Tetrahydrofurans
A General Oxidative Cyclization of 1,5-Dienes
Using Catalytic Osmium Tetroxide**
Timothy J. Donohoe* and Sam Butterworth
The oxidative cyclization of 1,5-dienes to produce tetrahy-
drofurans has been known for some time, and is a unique
method for making these heterocycles. The reaction is
Received: October 24, 2002 [Z50425]
[*] Dr. T. J. Donohoe, S. Butterworth
Dyson Perrins Laboratory
[1] a) S. E. Denmark, N. G. Almstead in Modern Carbonyl Chemistry
(Ed.: J. Otera), Wiley-VCH, Weinheim, 2000, chap. 10; b) S. R.
Chemler, W. R. Roush in Modern Carbonyl Chemistry (Ed.: J.
Otera), Wiley-VCH, Weinheim, 2000, chap. 11; c) T. Herold,
R. W. Hoffmann, Angew. Chem. 1978, 90, 822; Angew. Chem. Int.
Ed. Engl. 1978, 17, 768; d) H. C. Brown, P. K. Jadhav, J. Am.
Chem. Soc. 1983, 105, 2092; e) P. K. Jadhav, K. S. Bhat, P. T.
Perumal, H. C. Brown, J. Org. Chem. 1986, 51, 432; f) U. S.
Racherla, H. C. Brown, J. Org. Chem. 1991, 56, 401; g) W. R.
South Parks Road, Oxford, OX1 3QY (UK)
Fax: (+44)1865-275-674
E-mail: timothy.donohoe@chem.ox.ac.uk
[**] We thank the EPSRC and Merck, Sharp, and Dohme for funding this
project. AstraZeneca, GlaxoSmithKline, Pfizer, and Novartis are
thanked for unrestricted donations.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
948
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1433-7851/03/4208-0948 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2003, 42, No. 8

Angewandte
Chemie
particularly powerful because it involves the stereospecific
suprafacial addition of two oxygen atoms across each of the
two alkenes, coupled with the steroselective formation of a
cis-substituted tetrahydrofuran ring (Scheme 1). Typically,
and 31% yield, respectively, but was not completely stereo-
selective for cis-tetrahydrofurans.
Our own work in this area had concentrated on the use of
osmium tetroxide, which we believe is the oxidant with the
most potential for the oxidative cyclizations of dienes. Initial
results using the combination of OsO4/tetramethylethylenedi-
amine (TMEDA) with a diene had yielded an osmate ester
R3
R5
À
4, which could be persuaded to cyclize (by adding the RO
R2
R1
R2
R5
R6
[O]
R6
O
¼
Os O unit across a distal alkene) under acidic conditions
R4
R3
R1
R4
HO
OH
(Scheme 2).[5]
1
2
Scheme 1. Stereoselective formation of cis-tetrahydrofurans.
[O]=KMnO4, OsO4, RuO4.
this type of oxidation requires stoichiometric amounts of
powerful oxidants such as KMnO4 or CrO3 derivatives,[2]
[1]
which usually leads to low yields because of over-oxidation.
Recently, conditions have been reported for accomplishing
this reaction using catalytic osmium tetroxide[3] or ruthenium
tetroxide[4] in conjunction with NaIO4 reoxidant. However,
while both of these processes are admirable, neither is
particularly general or high yielding. For example, oxidation
of geranyl acetate using Picialli's conditions (cat. OsO4,
NaIO4, DMF[3]) gave 55% yield of a tetrahydrofuran, and a
slightly lower yield with the isomeric neryl substrate (see
entries 3 and 4, Table 1, see below); in our experience, these
two substrates are often the ones that give the highest yields.
In contrast, catalytic ruthenium tetroxide gave the same
products as above from both geranyl and neryl acetates in 61
Scheme 2. Acid-promoted cyclization of osmate esters.
Although the yields of oxidative cyclization were good
using this protocol (typically, 60–80%), the use of stoichio-
metric transition metals was unattractive and we sought out
catalytic variants. Our early work had clearly shown that it
was the acidic conditions that promoted cyclization of the
osmate ester 4 to form the five-membered ring. Therefore, we
examined the oxidative cyclization of a range of 1,5-dienes
using catalytic OsO4 (5%), Me3NO (4 equiv), and either
camphorsulfonic
acid
(CSA;
6 equiv) or trifluoroacetic acid
(TFA; excess) to lower the pH.
Remarkably, the dihydroxylation/
oxidative cyclization sequence
worked very well and provided
several stereochemically defined
tetrahydrofurans in good yield (Ta-
ble 1), and as single stereoisomers.
The conditions reported here are
general for a range of 1,5-diene
substrates and are a broad solution
to the problems of yield and applic-
ability, as outlined above. In fact,
the lowest yield in Table 1 (60%) is
a consequence of the volatility of
the diene starting material (see
entry 1). The stereoselective oxida-
tion of both cis and trans isomers
(compare entries 3 with 4 and 5 with
6, Table 1) is also noteworthy.[6]
Table 1: Osmium-catalyzed cyclizations of 1,5-dienes.
Entry
1
R1
H
R2
H
R3
H
R4
H
R5
H
R6
H
Product
Yield [%]
60[a]
2
3
4
5
6
H
H
Me
H
Me
Me
Me
H
H
H
72[a]
88[b]
95[b]
78[b]
81[b]
Me
Me
H
Me
Me
Bu
H
H
CH2OBn
H
CH2OBn
H
H
H
Bu
H
In terms of mechanism, we sug-
Bu
H
H
Bu
gest a similar scheme to that descri-
bed above in Scheme 2. The initial
reaction is dihydroxylation of the
diene to yield an osmate ester
(Scheme 3); this may then be cy-
clized to a tetrahydrofuan before
the osmium is reoxidized back to
7
H
-(CH2)4-
H
H
H
H
71[a]
[a] CSA conditions in CH2Cl2. [b] TFA conditions in acetone/water (9:1). Bn=benzyl.
Angew. Chem. Int. Ed. 2003, 42, No. 8
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 1433-7851/03/4208-0949 $ 20.00+.50/0
949

Communications
Scheme 3. Possible mechanism of oxidative cyclization.
OsVIII [7]
. The exact role of the acid in promoting cyclization is
unknown, but we suggest that it protonates an oxo ligand on
the osmium center, making it a more electron-deficient
partner in the ensuing cycloaddition.
Scheme 5. Synthesis of anhydro-d-glucitol and d-chitaric acid.
TEMPO= 2,2,6,6-tetramethyl-1-piperidinyloxyl.
Our next objective was to search for methods of making
these products enantiomerically pure.[8] We turned our
attention to the role of stereodirecting groups on the diene
backbone, between the two alkene groups: this is the first
time that this possibility has been explored. The oxidation of
substrate 12 is instructive for two reasons (Scheme 4). First, it
To conclude, we report here for the first time a general
and high-yielding oxidative cyclization of 1,5-dienes to cis-
tetrahydrofurans. Our method is particularly attractive be-
cause it uses catalytic quantities of transition metal, in
conjunction with an amine oxide as reoxidant and an acid to
promote cyclization. Moreover, the potential of internal
substituents on the diene backbone to act as stereodirecting
groups has been explored and a short synthesis of highly
functionalized enantiopure tetrahydrofuran natural products
has been developed.
Received: November 12, 2002 [Z50535]
[1] E. Klein, W. Rojahn, Tetrahedron 1965, 21, 2353; J. E. Baldwin,
M. J. Crossley, E.-M. M. Lehtonen, J. Chem. Soc. Chem.
Commun. 1979, 918; D. M. Walba, M. D. Wand, M. C. Wilkes,
J. Am. Chem. Soc. 1979, 101, 4396.
Scheme 4. Oxidation of substrate 12. TBDPS= tert-butyldiphenylsilyl.
[2] Chromium-promoted cyclization of diols derived from 1,5-diene
give the same THF products, see: D. M. Walba, G. S. Stoudt,
Tetrahedron Lett. 1982, 23, 727.
indicates that bulky silyl protecting groups are compatible
with the oxidative regimen. Second, the lack of stereo-
selectivity with respect to the initial chiral center probably
means that the reaction initiates through dihydroxylation of
the unsubstituted alkene to give a mixture of osmate esters[9]
(presumably, this reaction is nonstereoselective because the
chiral center is too far removed). However, the presence of
two cis-tetrahydrofuran products, 13 and 14, means that the
penchant for forming cis stereochemistry during cyclization
clearly overrides any bias imposed by the substituent at C3.
Hence, two cis isomers are formed in equal amounts, one from
each isomer of the osmate ester.[10]
One consequence of this mechanism is that the readily
available C2-symmetric compound (þ)-15[11] should be oxi-
dized to give a single stereoisomer, whatever the outcome
from the initial dihydroxylation (Scheme 5). This was indeed
the case, and compound (þ)-16 was isolated pure in 84%
yield from the oxidation reaction.
[3] V. Piccialli, Tetrahedron Lett. 2000, 41, 3731.
[4] P. H. J. Carlsen, T. Katsuli, V. S. Martin, K. B. Sharpless, J. Org.
Chem. 1981, 46, 3936; V. Piccialli, N. Cavallo, Tetrahedron Lett.
2001, 42, 4695.
[5] T. J. Donohoe, J. J. G. Winter, M. Helliwell, G. Stemp, Tetrahe-
dron Lett. 2001, 42, 971.
[6] Compound 5 is known, see: M. E. Jung, I. D. Trifunovich, A. W.
Sledeski, Heterocycles 1993, 35, 273. Compound 6 was derivat-
ized to a bis-Moshers ester that was no longer symmetrical;
compounds 8, 10, and 11 were characterized by X-ray crystal-
lography. The structure of 7 is assigned by analogy to that of 8,
while compound 9 was converted into an authentic sample of 10
by a double Mitsunobu inversion/hydrolysis sequence.
[7] Preliminary experiments have shown that both first and second
cycle osmium species are capable of undergoing the oxidative
cyclization reaction and therefore we find it difficult to discrim-
inate between the two.
[8] The pH of these oxidation reactions has been measured at < 1;
in this range we do not (nor would we expect to) observe
asymmetric induction from a chiral cinchona alkaloid ligand.
Previous attempts to make the cyclization products enantiopure
have used acyl-derived auxiliaries, see: D. M. Walba, C. A.
Przybyla, C. B. Walker, Jr., J. Am. Chem. Soc. 1990, 112, 5624;
P. J. Kocienski, R. C. D. Brown, A. Pommier, M. Procter, B.
Schmidt, J. Chem. Soc. Perkin Trans. 1 1998, 9; A. R. L. Cecil,
R. C. D. Brown, Org. Lett. 2002, 4, 3715; R. C. D. Brown, C. J.
Bataille, R. M. Hughes, A. Kenney, T. J. Luker, J. Org. Chem.
The utility of this substrate was illustrated when com-
pound (þ)-16 was deprotected (H2, Pd/C) to give (þ)-
anhydro-d-glucitol,[12] or then monoprotected with BnBr/
Ag2O to yield (þ)-17. When compound 17 was oxidized to the
carboxylic acid, and deprotected by hydrogenolysis, (þ)-d-
chitaric acid[13] was produced in a short sequence (Scheme 5).
950
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1433-7851/03/4208-0950 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2003, 42, No. 8

Angewandte
Chemie
2002, 67, 8079; or chiral phase-transfer catalysis, see: R. C. D.
Brown, J. F. Keily, Angew. Chem. 2001, 113, 4628; Angew. Chem.
Int. Ed. 2001, 40, 4496.
ketone. Both 13 and 14 gave the same ketone, which showed
reciprocal NOE enhancements between H(C2) and H(C5).
[11] S. D. Burke, G. M. Sametz, Org. Lett. 1999, 1, 71.
[12] The spectroscopic data from our material matched exactly that
in the literature, see: R. Persky, A. Albeck, J. Org. Chem. 2000,
65, 5632.
[13] E. Hardegger, F. Lohse, Helv. Chim. Acta 1957, 40, 2383; H. C.
Cox, E. Hardegger, F. Kꢀgl, P. Liechti, F. Lohse, C. A. Salemink,
Helv. Chim. Acta 1958, 41, 229; E. Hardegger, H. Furter, J. Kiss,
Helv. Chim. Acta 1958, 41, 2401.
[9] Based on Sharpless' rate data, we estimate that the C1–C2
alkene unit is approximately ten times less reactive than the
other towards dihydroxylation: H. C. Kolb, M. S. VanNieu-
wenhze, K. B. Sharpless, Chem. Rev. 1994, 94, 2483. Using OsO4/
TMEDA we have been able to isolate the two (inseparable)
osmate ester diastereosiomers shown in Scheme 4 and show that
they cyclize to a mixture of 13 and 14.
[10] The structures of 13 and 14 were proven by protection of the
primary hydroxy groups, desilylation, and oxidation to the C3
Angew. Chem. Int. Ed. 2003, 42, No. 8
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 1433-7851/03/4208-0951 $ 20.00+.50/0
951