Highly diastereoselective reactions using masked allylic zinc reagents

Article abstract of DOI:10.1039/a805953e

Substituted allylic organozinc reagents have been prepared using a novel fragmentation reaction; the resulting allylic zinc species are configurationally stable and give excellent regio- and diastereo-selectivities.

Full text of DOI:10.1039/a805953e

Highly diastereoselective reactions using masked allylic zinc reagents
Philip Jones and Paul Knochel*
Fachbereich Chemie der Universität, Hans-Meerwein Strasse, D-35032 Marburg, Germany.
E-mail: knochel@ps1515.chemie.uni-marburg.de
Received (in Liverpool, UK) 29th July 1998, Accepted 22nd September 1998
OH
OH
Substituted allylic organozinc reagents have been prepared
using a novel fragmentation reaction; the resulting allylic
zinc species are configurationally stable and give excellent
regio- and diastereo-selectivities.
Me
Me
6, 84%, 96:4, anti : syn
7, 76%, 97:3, anti : syn
We have recently documented an entirely new approach for the
generation of allylic zinc reagents based on a retro-allylation, an
allylation sequence starting from the sterically hindered tertiary
alcohol 1 (Scheme 1).1,2 Generation of the zinc alkoxide of this
homoallylic alcohol results in fragmentation and formation of
an allyl zinc reagent 2 in situ; this is itself able to react with a
range of other electrophiles. This method avoids the problems
associated with Wurtz coupling.
OH
OH
Me
Me
9, 92%, >98:2, anti : syn
8, 86%, >98:2, anti : syn
Fig. 1
(i) BuLi
OH
OH
(ii) ZnBr2
PhCHO
BrZn
But
But
anti-diastereomer was observed by 1H and 13C NMR spectros-
copy. In all cases, none of the g-substituted product was
detected.
Ph
1
2
Scheme 1
Similar results were obtained when we placed other sub-
stituents in the a-position. 3-(tert-Butyl)-2,2-dimethyl-4-ethyl-
hex-5-en-3-ol 10 was prepared using identical chemistry as used
in the preparation of homoallylic alcohol 3.4 This too was found
to exhibit high levels of anti-selectivity. Reaction with BuLi,
benzaldehyde and zinc chloride (Scheme 3) gave 2-ethyl-
1-phenylbut-3-en-1-ol 11 in 91% as a 91:9 mixture of anti:syn
isomers. Likewise, reaction with cyclohexanecarbaldehyde and
2-ethylbutyraldehyde gave rise to the homoallylic alcohols 12
and 13 in 83 and 81% yields, respectively, both solely as the
anti-diastereomers.
Further synthetic investigation revealed that a benzyl group
could easily be incorporated into the a-position of the
homoallylic alcohol (Scheme 4). The precursor was readily
prepared form 2, 2-dimethylhex-5-en-3-one 14 by deprotona-
tion with LDA at 278 °C, followed by a-alkylation with BnBr
in THF–HMPA in 65% yield. Subsequent reaction with ButLi
gave the required precursor 15 in 85% yield. As in all previous
cases, deprotonation and reaction with benzaldehyde in the
presence of zinc chloride gave the desired benzylic alcohol 16
in 89%, again with excellent diastereoselectivity (95:5).
Aliphatic aldehydes also reacted well giving the homoallylic
alcohols 17 and 18 in 88 and 80% yields, respectively.
Surprised by these results we were interested to investigate
the outcome of placing substituents in the g-position of the
Substituted allyl zinc reagents had previously been prepared
by Tamura from allyl benzoates but with mixed results;3 we
hoped that our mild method would give improved selectiv-
ities.
Accordingly, the tertiary alcohol 3 bearing an a-methylallyl
group was prepared (Scheme 2).4 However, we were disap-
pointed to find that upon deprotonation of this material at room
temperature with BuLi a rapid isomerisation to the g-substituted
isomer 4 occurred.5 Undeterred, the deprotonation was at-
tempted at 278 °C, and we were pleased to find no
isomerisation. Addition of benzaldehyde followed by a solution
of zinc chloride gave within 1 h at 278 °C the benzylic alcohol
5 in 83% isolated yield.† More interestingly, the product was
isolated as a 94:6 mixture of anti:syn diastereomers.6 This
reaction is in stark contrast to the addition of crotylzinc bromide
to benzaldehyde, which gives approximately 1:1 mixtures of
diastereomers.7
Inspired by this reaction, a range of aldehydes were screened
(Fig. 1). Reaction of the homoallylic alcohol with cyclohex-
anecarbaldehyde gave the alcohol 6 in 84% yield, again with the
anti diastereomer in excess (96:4). Similarly, reaction with
2-butylacrolein gave solely the product of 1,2-addition, giving
the homoallylic alcohol 7 in 76% isolated yield as a 97:3
mixture of diastereomers. The reaction with 2-ethylbutyr-
aldehyde and 1-naphthaldehyde gave the expected products 8
and 9 in 86 and 92% yields, respectively; in both cases only the
OH
OH
(i) BuLi, –78 °C
(ii) PhCHO
But
But
Ph
OH
OH
(iii) ZnCl2, –78 °C
Me
Me
But
But
But
But
Me
BuLi
10
11, 91%, 91:9, anti : syn
room temp.
4
Me
Me
3
3
OH
OH
OH
OH
(i) BuLi, –78 °C
(ii) PhCHO
But
But
Ph
Me
Me
(iii) ZnCl2, –78 °C
Me
13, 81%, >98:2, anti : syn
12, 83%, >98:2, anti : syn
Scheme 3
5, 83%, 94:6, anti : syn
Scheme 2
Chem. Commun., 1998, 2407–2408
2407

R
R
OH
But
But
But
H
But
H
(i) LDA, –78 o
(ii) BnBr
(iii) ButLi
C
O
O
O
Allylic
Transposition
But
Me
But
Me
ZnX
ZnX
But
O
O
Ph
14
15
OH
26
27
(i) BuLi, –78 o
(ii) PhCHO
(iii) ZnCl2
C
Ph
H
H
Ph
Allylic
Transposition
ZnX
R
OH
R
O
16, 89%, 95:5, anti : syn
Me
Me
OH
OH
H
H
29
28
Scheme 6
Ph
17, 88%, >98:2, anti : syn
Ph
18, 80%, >98:2, anti : syn
These results suggest a plausible mechanism involving a
double allylic transposition. Generation of the zinc alkoxide of
the alcohol 26 results in a cyclic six-membered intermediate
where the zinc is complexed to the reacting carbonyl compound
(Scheme 6). Allylic transposition gives rise to a crotylzinc
reagent 27 complexed to the parent bis(tert-butyl) ketone and
the reaction partner, the zinc reagent bearing solely a trans-
configuration. At 278 °C this species is stable and undergoes
no isomerisation.8 Owing to the complexation of the reacting
partner with the zinc, a new six-centred intermediate 28 is
possible, whereby all the substituents lie in equatorial positions;
allylic transposition then gives rise to the product 29, predom-
inately as the anti-diastereomer. This mechanism is supported
by the unreactive nature of the g-substituted isomer, whereby
steric congestion prevents the first allylic transposition from
occurring.
In summary, we have developed a novel method for the
preparation of substituted allylic zinc reagents. This method is
extremely selective, giving both excellent regiochemical se-
lectivity and excellent diastereoselectivity. The method is
extremely mild and avoids Wurtz coupling products.9
The authors thank the DFG (SFB 260 and Leibniz program)
for generous financial support, and the Royal Society for an
award (to P. J.) under the European Science Exchange
Programme.
Scheme 4
allylic system. A new synthetic route needed to be adopted and
oxirane 19 was prepared and opened with a variety of lithium
acetylides to give the homopropargyl compounds 20a,b in
reasonable yields (84 and 65%) (Scheme 5). Hydrogenation
with palladium on barium sulfate gave the Z-isomer 21
quantitatively, whilst treatment with LiAlH4 gave the required
E-isomer 22. With these two compounds to hand the migration
was investigated. The resulting zinc alkoxides were found to be
less reactive. Whereas the a-substituted compounds migrated at
278 °C, these compounds required warming to room tem-
perature before migration could be observed. With 21 migration
occurred in 52% yield after 48 h (86% based on recovered
starting material) to give the desired homoallylic alcohol 23,
while the E-isomer 22 gave the corresponding alcohol 24 in
23% yield after 12 h (87% based on recovered starting material).
In both cases the products were isolated as 2:1 mixture of Z:E
isomers. The g-disubstituted compound 25 was also prepared;
however, this compound proved to be stable and no migration
was detected.
R
OH
R
O
BuLi, HMPA
But
But
19
But
But
20a R = Pr
b R = Bu
Notes and references
† Typical procedure: anti-2-methyl-1-phenylbut-3-en-1-ol 5: A solution of
BuLi (2.52 mmol) in pentane (1.6 M, 1.58 ml) was added dropwise over 5
OH
min to a stirred solution of 3-tert-butyl-2,2,4-trimethylhex-5-en-3-ol (ref. 4)
3 (500 mg, 2.52 mmol) in THF (4 ml) at 278 °C under argon. The resulting
solution was then stirred for 15 min and benzaldehyde (256 ml, 2.52 mmol)
was added and stirred for a further 15 min; finally a solution of zinc chloride
(343 mg, 2.52 mmol) in THF (2 ml) was added over 3 min. The reaction was
stirred at 278 °C for 1 h then allowed to warm to room temperature. The
reaction was worked up as usual to give a crude residue, which was then
purified by column chromatography on silica using 10% Et2O–light
petroleum as eluent to give the desired alcohol (ref. 10) 5 (341 mg, 83%) as
a pale yellow oil.
H2
But
But
Pd/BaSO4
20a
Pr
21
22
OH
OH
But
But
Bu
LiAlH4
20b
(i) BuLi
(ii) PhCHO
(iii) ZnCl2
1 Y. Yamamoto and N. Asao, Chem. Rev., 1993, 93, 2207.
2 P. Jones, N. Millot and P. Knochel, Chem. Commun., 1998, 2405.
3 M. Shimizu, M. Kimura, S. Tanaka and Y. Tamaru, Tetrahedron Lett.
1998, 39, 609 and references cited therein.
4 R. A. Benkeser, M. P. Siklosi and E. C. Mozdzen, J. Am. Chem. Soc.,
1978, 100, 2134.
OH
Pr
But
But
Ph
23, 52% (86%)
Pr
21
22
(i) BuLi
(ii) PhCHO
(iii) ZnCl2
OH
OH
5 F. Gérard and P. Miginiac, Bull. Chim. Soc. Fr., 1974, 2527; F. Barbot
and P. Miginiac, Bull. Chim. Soc. Fr., 1977, 113.
Bu
But
But
Bu
Ph
6 All diastereomeric excesses were determined by 1H NMR analysis.
7 S. R. Wilson and M. E. Guazzaroni, J. Org. Chem., 1989, 54, 3087.
8 bis (3-Methylallyl)zinc is known to be a rapidly isomerising system at
room temperature, hence explaining the 1:1 anti:syn selectivity in
additions to aldehydes. R. Benn, E. G. Hoffmann, H. Lehmkuhl and H.
Nehl, J. Organomet. Chem., 1978, 146, 103.
24, 23% (87%)
OH
But
But
Me
9 A patent has been filed with Chemetall GmbH (Frankfurt).
10 S. Kobayashi and K. Nishio, J. Org. Chem., 1994, 59, 6620.
Me
25
Scheme 5
Communication 8/05953E
2408
Chem Commun., 1998