Oxonia-Cope rearrangement and side-chain exchange in the Prins cyclization

Article abstract of DOI:10.1021/ol0102850

(Formula Presented) TMSOAc OAc OAc 2 3 Evidence is presented here for the mechanism of the Prins cyclization of benzylic homoallylic alcohols, which shows that the outcome of the reaction is dependent upon the substituents on the aromatic ring. The presence of an electron-rich aromatic ring favors an oxonia-Cope rearrangement yielding a symmetrical tetrahydropyran as the major product formed via a side-chain exchange process. In contrast, with electron-deficient aromatic rings the expected 2,4,6-trisubstituted tetrahydropyran is formed.

Full text of DOI:10.1021/ol0102850

ORGANIC
LETTERS
2002
Vol. 4, No. 4
577-580
Oxonia-Cope Rearrangement and
Side-Chain Exchange in the Prins
Cyclization
Stuart R. Crosby,† John R. Harding,‡ Clare D. King,‡ Gregory D. Parker,† and
,†
Christine L. Willis*
School of Chemistry, UniVersity of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K.,
and AstraZeneca UK Ltd., Mereside, Alderley Park, Macclesfield, SK10 4TG, U.K.
chris.willis@bristol.ac.uk
Received December 5, 2001
ABSTRACT
Evidence is presented here for the mechanism of the Prins cyclization of benzylic homoallylic alcohols, which shows that the outcome of the
reaction is dependent upon the substituents on the aromatic ring. The presence of an electron-rich aromatic ring favors an oxonia-Cope
rearrangement yielding a symmetrical tetrahydropyran as the major product formed via a side-chain exchange process. In contrast, with
electron-deficient aromatic rings the expected 2,4,6-trisubstituted tetrahydropyran is formed.
In recent years, there has been widespread interest in the
use of Prins cyclizations for the stereocontrolled synthesis
of 2,4,6-trisubstituted tetrahydropyrans.1 These reactions are
believed to proceed via formation of an oxocarbenium ion
(generated in situ either from reaction of a homoallylic
alcohol with an aldehyde or from a homoallylic acetal) that
undergoes an intramolecular cyclization, giving the tetra-
hydropyran with all three substituents located in an equatorial
position. It has been suggested that cationic oxonia-Cope
rearrangements may participate in Prins cyclisations and
related reactions as illustrated in Figure 1.2
Figure 1. Oxonia-Cope rearrangement.
† University of Bristol.
‡ AstraZeneca UK Ltd.
(1) For example: (a) Cloninger, M. J.; Overman, L. E. J. Am. Chem.
Soc. 1999, 121, 1092. (b) Yang, J.; Viswanathan, G. S.; Li, C.-J. Tetrahedron
Lett. 1999, 40, 1627. (c) Rychnovsky, S. D.; Hu, Y.; Ellsworth, B.
Tetrahedron Lett. 1998, 39, 7271. (d) Marko´, I. E.; Chelle´, F. Tetrahedron
Lett. 1997, 38, 2895.
Most literature examples of the synthesis of tetrahydro-
pyrans by such cyclizations have utilized homoallylic alco-
hols with aliphatic side chains as the starting material. At
(2) Examples of such rearrangements include: (a) Jaber, J. J.; Mitsui,
K.; Rychnovsky, S. D. J. Org. Chem. 2001, 66, 4679. (b) Semeyn, C.;
Blaauw, R. H.; Hiemstra, H.; Speckamp, W. N. J. Org. Chem. 1997, 62, 2,
3426. (c) Lolkema, L. D. M.; Semeyn, C.; Ashek, L.; Hiemstra, H.;
Speckamp, W. N. Tetrahedron 1994, 50, 7129. (d) Gasparski, C. M.;
Herrington, P. M.; Overman, L. E.; Wolfe, J. P. Tetrahedron Lett. 2000,
41, 9431. (e) Loh, T.-P.; Hu, Q.-Y.; Ma, L.-T. J. Am. Chem. Soc. 2001,
123, 2450. (f) Brown, M. J.; Harrison, T.; Herrington, P. M.; Hopkins, M.
H.; Hutchinson, K. D.; Mishra, P.; Overman, L. E. J. Am. Chem. Soc. 1991,
113, 5365.
10.1021/ol0102850 CCC: $22.00 © 2002 American Chemical Society
Published on Web 01/30/2002

the start of our investigations, we proposed that if the
homoallylic alcohol has a side chain with an electron-rich
aromatic ring such as 1a (Table 1), then it would be
reasonable to expect an oxonia-Cope rearrangement to be
favored through stabilization of the positive charge.3
TMSOAc to act as a fluoride trap gave yields of between
50 and 70%.5
Thus, cyclization of homoallylic alcohol 1a with propanal
was investigated using BF3‚OEt2 in the presence of AcOH
and TMSOAc in cyclohexane at room temperature to give
the expected tetrahydropyran 2a as a single diastereomer but
in only 20% yield (entry i, Table 1). Interestingly, three
further products were isolated including the allylic acetate
4a and the parent aldehyde 5a, and of particular interest was
another tetrahydropyran 3 formed in 21% yield. The plane
of symmetry was clearly revealed in the 13C NMR spectrum
of 3, which contained only seven signals. To the best of our
knowledge, this is the first example of a Prins-type cycliza-
tion involving a homoallylic alcohol and aldehyde (with
different side chains) to give a symmetrical 2,4,6-trisubsti-
tuted tetrahydropyran and must involve an allyl transfer
process. Side-chain exchanges have been reported in other
cyclizations. For example, a recent paper by Roush and
Dilley6 describes the preparation of 2,6-disubstituted di-
hydropyrans from allylsilanes, and Li and co-workers7
detected symmetrical 2,4,6-thiacyclohexanes by GC/MS on
treatment of homoallylic thiols with an aldehyde in the
presence of indium trichloride.
Table 1. Reaction of Alcohols 1a-g with BF3‚OEt2 (2 Equiv),
AcOH (5 Equiv), and TMSOAc (4 Equiv) in C6H12 at Room
Temperature
To further investigate the effect that the nature of the
aromatic ring has on the outcome of the Prins cyclizations,
a series of substituted aromatic homoallylic alcohols 1b-g
was prepared in good yield by reaction of allyl iodide with
the substituted benzaldehyde in the presence of indium
powder in water.8 Each homoallylic alcohol was treated with
propanal in the presence of BF3‚OEt2, AcOH, and TMSOAc
at room temperature, and following workup, the products
were purified by flash chromatography (Table 1). In accord
with the results from cyclization of the 3,4-methylenedioxy
derivative 1a, the electron-rich anisaldehyde derived homo-
allylic alcohol 1b gave rise to a greater proportion of the
symmetrical product 3 than the trisubstituted heterocycle 2b
(entry ii, Table 1). In addition, it is interesting to note that
the homoallylic acetate 4b and the parent aldehyde 5b were
also isolated. Cyclization of homoallylic alcohols 1c and 1d
(prepared from benzaldehyde and 3-fluorobenzaldehyde,
respectively) gave the tetrahydropyrans 2c and 2d, respec-
tively; in each case, the symmetrical compound 3 was only
a minor product. In contrast, when homoallylic alcohols 1e,
1f, and 1g (derived from 2-chlorobenzaldehyde, 4-nitroben-
zaldehyde, and 3,4-diacetoxybenzaldehyde, respectively)
were treated with propanal under the standard cyclization
conditions, the only products isolated in each case were the
expected tetrahydropyrans 2e, 2f, and 2g along with recov-
ered starting material.
entry
R )
2 (%) 3 (%) 4 (%) 5 (%) 1 (%)
i
ii
iii
iv
v
1a ; 3,4-OCH2O
1b; 4-OMe
1c; H
1d ; 3-F
1e; 2-Cl
20
15
54
46
57
36
67
21
21
24
9
26
21
17
18
23
9
23
36
43
a
vi
vii
1f; 4-NO2
1g; 3,4,-diOAc
a
7% of the 4-F analogue was isolated due to extended reaction time
(16 h).
To test this hypothesis, homoallylic alcohol 1a was
prepared in 92% yield from commercially available piperonal
and allylmagnesium bromide. Several methods have been
reported for the introduction of oxygenated substituents at
C-4 of tetrahydropyrans in Prins cyclizations with varying
success.4 In recent studies into the stereocontrolled synthesis
of 4-hydroxy-2,5-disubstituted tetrahydropyrans, we have
shown that hydrolysis of the esters formed from reaction of
homoallylic acetals with either trifluoroacetic acid or with
BF3‚OEt2 in the presence of AcOH as the nucleophile and
Since homoallylic acetates are only detected as products
in the cyclization studies with electron-rich aromatic rings
(entries i and ii, Table 1), a stabilized carbocation intermedi-
(3) We are grateful to Professor S. D. Rychnovsky for sharing his
fascinating results with us prior to publication in which he showed that the
intermediate from an oxonia-Cope rearrangement may be trapped by
reduction with Bu3SnH: Rychnovsky, S. D.; Marumoto, S.; Jaber, J. J.
Org. Lett. 2001, 3, 3815.
(4) (a) Nishizawa, M.; Shigaraki, T.; Takao, H.; Imagawa, H.; Sugihara,
T. Tetrahedron Lett. 1999, 40, 1153. (b) Zhang, W.-C.; Viswanathan, G.
S.; Li, C.-J. Chem. Commun. 1999, 291. (c) Zhang, W.-C.; Li, C.-J.
Tetrahedron 2000, 56, 2403. (d) Hu, Y.; Skalitzky, D. J.; Rychnovsky, S.
D. Tetrahedron Lett. 1996, 37, 8679. (e) Viswanathan, G. S.; Yang, J.; Li,
C.-J. Org. Lett. 1999, 1, 993. (f) Rychnovsky, S. D.; Yang, G.; Hu, Y.;
Khire, U. R. J. Org. Chem. 1997, 62, 3022.
(5) Al-Mutairi, E. H.; Crosby, S. R.; Darzi, J.; Harding, J. R.; Hughes,
R. A.; King, C. D.; Simpson, T. J.; Smith, R. W.; Willis, C. L. Chem.
Commun. 2001, 835.
(6) Roush, W. R.; Dilley, G. J. Synlett. 2001, 955.
(7) Yang, X. F.; Mague, J. T.; Li, C.-J. J. Org. Chem. 2001, 66, 739.
(8) Chan, T. H.; Lee, C. J.; Lee, M. C.; Wei, Z. Y. Can. J. Chem. 1994,
72, 1181.
578
Org. Lett., Vol. 4, No. 4, 2002

Figure 2. Proposed mechanism for formation of products 2b, 3, 4b, and 5b on treatment of 1b with EtCHO, TMSOAc, and AcOH +
BF3‚OEt2.
ate I is implicated (Figure 2). To probe for such an
intermediate, first homoallylic alcohols with either an
electron-rich (R ) 4-OMe) or electron-poor (R ) 3-F)
aromatic ring, 1b and 1d, respectively, were each simply
treated with BF3‚OEt2, AcOH, and TMSOAc in cyclohexane
at room temperature (i.e., no propanal was added). In the
case of the 4-methoxy substrate 1b, the acetate 4b was
formed in quantitative yield, whereas the reaction with the
3-fluoro derivative 1d was slower giving a 2:1 mixture of
starting alcohol 1d and acetate 4d. These results are in accord
with the proposal that the acetates 4a and 4b are formed via
the stabilized carbocation intermediate I.
The mechanism of the reaction was further investigated
using the enantioenriched homoallylic alcohol (S)-1b, which
was prepared in good yield (77%) and 89% ee (as determined
via analysis of the Moshers ester derivative) by a modifica-
tion of the method of Brown and co-workers.10 Reaction of
(S)-1b with propanal, BF3‚OEt2, AcOH, and TMSOAc gave
the four expected products, which were readily separated by
flash chromatography (Scheme 1). Homoallylic acetate 4b
was racemic in accord with the proposed mechanism
involving a stabilized carbocation intermediate I (Figure 2).
In addition, we found that the 2,4,6-trisubstituted tetrahy-
dropyran 2b had very low ee (<5%), as determined by chiral
HPLC. This was rather unexpected as there are literature
examples of Prins cyclizations in which the enantioenrich-
ment in the parent alcohol is retained in the oxygen
heterocycles produced.2c,6 In this case, however, the initially
produced oxocarbenium ion II may be formed either by
attack of I by propanal or directly from alcohol 1b by
reaction with propanal and dehydration, thus accounting for
the low ee of 2,4,6-trisubstituted tetrahydropyran 2b (Figure
2).
If, as expected, the oxonia-Cope rearrangement of II to
III is favored in the case of an electron-rich aromatic side
chain due to resonance stabilization of the positive charge,
then the symmetrical tetrahydropyran 3 may be formed by
the mechanism shown in Figure 2. Nucleophilic attack (by
either water or acetate) on III would lead to the parent
aldehyde 5 and the homoallylic alcohol 6 with an aliphatic
side chain, hence effecting the necessary allyl transfer
reaction for the synthesis of the symmetrical product. The
symmetrical tetrahydropyran and aldehyde are produced in
an approximately a 1:1 ratio, in accord with the proposed
mechanism.
Scheme 1. Cyclization of Enantioenriched Alcohol (S)-1b.
Next, we turned our attention to the cyclization of an
enantioenriched homoallylic alcohol with an electron-
deficient aromatic ring, and we selected the 2-chloro deriva-
(9) Allyl-transfer reactions of homoallylic alcohols have been noted
previously; see, for example: (a) Nokami, J.; Yoshizane K.; Matsuura, H.;
Sumida, S. J. Am. Chem. Soc. 1998, 120, 6609. (b) Loh, T.-P.; Tan, K.-T.;
Hu, Q.-H. Angew. Chem., Int. Ed. 2001, 40, 2921. (c) Nokami, J.; Ohga,
M.; Nakamoto, H.; Matsubara, M.; Nakamoto, H.; Matsubara, T.; Hussain,
I.; Kataoka, K. J. Am. Chem. Soc. 2001, 123, 9168.
(10) Brown, H. C.; Bhat, K. S.; Randad, R. S. J. Org. Chem. 1987, 52,
319.
Org. Lett., Vol. 4, No. 4, 2002
579

cases of homoallylic alcohols possessing an aromatic side
chain. The outcome of the reaction is dependent upon the
substituents on the aromatic ring. With an electron-deficient
aromatic ring, the Prins cyclization proceeds to give the
expected 2,4,6-trisubstituted tetrahydropyran 2, which virtu-
ally retains the enantiopurity during the reaction. None of
the symmetrical tetrahydropyran 3 is detected. In contrast,
with an electron-rich aromatic ring, the symmetrical tetra-
hydropyran 3 is formed as the major product via a side-
chain exchange process and the 2,4,6-trisubstituted tetra-
hydropyran 2 produced has very low ee (<5%). Various
alternative mechanisms to those shown in Figure 2 could be
proposed to account for these observations. For example, a
pathway involving either an oxonia-Cope rearrangement
between 1b and 5b would lead to racemization of the starting
material or an oxonia-Cope rearrangement of 6 with propanal
would racemize 6 and the reversible formation of III and
cyclization would lead to racemized product 2b. Further
studies are currently underway to investigate the mechanisms
of these reactions. In light of these results, care should be
taken when using the Prins cyclization in the synthesis of
natural products.
Scheme 2 Cyclization of Enantioenriched Alcohol (S)-1e
tive (S)-1e, which was readily prepared in 77% yield and
94% ee (Scheme 2). In this case, formation of cation I would
be less favorable than in the case of 1b such that tetrahy-
dropyran 2e produced in the cyclization process would be
expected to be enantioenriched. Indeed, treatment of (S)-1e
with propanal, BF3‚OEt2, AcOH, and TMSOAc gave the
2,4,6-trisubstituted tetrahydropyran (-)-2e as the sole prod-
uct. None of the symmetrical tetrahydropyran 3 was detected
indicating that an oxonia-Cope rearrangement is less favored
with an electron-deficient aromatic ring. Unlike tetrahydro-
pyran 2b, in this case the product (-)-2e had 79% ee, i.e.,
only a slightly reduced enantiopurity compared to the starting
homoallylic alcohol (S)-1e.
Acknowledgment. We thank BBSRC, EPSRC, and
AstraZeneca for funding the work.
Supporting Information Available: The preparation and
characterization of the compounds described in the paper
are included. This material is available free of charge via
the Internet at http://pubs.acs.org.
In conclusion, we have presented results that indicate that
the mechanisms of Prins cyclizations are not simple in the
OL0102850
580
Org. Lett., Vol. 4, No. 4, 2002