On the synthesis of vinyl isonitriles

Article abstract of DOI:10.1039/a706830a

A reversal of the stereochemistry of base mediated isomerisation of some isonitrile epoxides to hydroxy-vinyl isonitriles on changing from lithium diisopropylamide to lithium bis(trimethylsilyl)amide is reported and applied to a synthesis of racemic desepoxyaerocyanidin.

Full text of DOI:10.1039/a706830a

View Article Online / Journal Homepage / Table of Contents for this issue
On the synthesis of vinyl isonitriles
Jack E. Baldwin, Deqi Chen and Andrew T. Russell
The Dyson Perrins Laboratory, South Parks Road, Oxford, UK OX1 3QY
NC
O
A reversal of the stereochemistry of base mediated
isomerisation of some isonitrile epoxides to hydroxy-vinyl
isonitriles on changing from lithium diisopropylamide to
lithium bis(trimethylsilyl)amide is reported and applied to a
synthesis of racemic desepoxyaerocyanidin.
KOBut–ButOH
R
Me
R
Me
O
5
6
KOBut
–
hydrolysis
H
As part of our studies on the synthesis of the epoxy isonitrile
antibiotic aerocyanidin 11 we required access to vinyl isonitrile
2.2 We considered that base mediated isomerisation of an
epoxide of general structure 3 could provide a stereocontrolled
access to these systems (Scheme 1). During these studies we
noted a pronounced dependence of the stereochemical outcome
of the isomerisation on the choice of base.
–
C
–
ButOH
+
N
O
N
O
N
O
R
R
R
Me
Me
Me
7
R = (CH2)6Me
Scheme 3
O
Me
(CH2)7
CO2H
it to rearrange with a 12:1 (9:10) selectivity and 96% yield. We
speculated that this isomerisation proceeded through an anti-
elimination via an open transition state and found that the
relative rate of the anti-elimination could be increased by the
addition of a small quantity of 12-crown-4 which resulted in an
increase in selectivity for 9 to 20:1. Consistent with this is the
observation that while LDA causes rearrangement of 5 with a
1:7 (9:10) ratio addition of 12-crown-4 depresses the ratio to
2:3.6
NC OH
Aerocyanidin 1
Me
(CH2)7
CO2H
O
NC
Base
3
Me
?
(CH2)7
CO2H
With this result in hand we undertook the synthesis of 2
(Scheme 5). Arndt–Eistert homologation of commercially
available 11 afforded after bromination and elimination alkyne
acid 12 in 66% overall yield. The dilithium anion of 12 was
quenched with acetaldehyde and after conversion to the methyl
ester introduction of the required amino group was effected
under Mitsunobo conditions to give 13 (47% from 12).7
Hydrogenation over a Lindlar catalyst afforded the correspond-
ing Z-alkene which, after deprotection of the phthalimide with
methylamine, was smoothly formylated with acetic formic
anhydride in 60% yield for the three steps. Taking advantage of
the allylic strain in the Z-vinylformamide 14 we used MCPBA
to affect a directed epoxidation to yield 15 as a single
stereoisomer (79%).8 Dehydration with trifluoromethane-
sulfonic anhydride9 afforded the corresponding isonitrile ep-
oxide (83%) which was saponified with Ba(OH)2 (87%) to
afford 16 for the key isomerisation reaction. Treatment of 16
with lithium bis(trimethylsilyl)amide at 215 °C in the presence
of a small quantity 12-crown-4 afforded 2 which proved to be an
unstable compound decomposing slowly to an a,b-unsaturated
ketone corresponding to 6. Therefore the lithium alkoxide of 2
was protected in situ as a tert-butyldimethylsilyl ether 17 (57%)
NC OH
Scheme 1
2
Following our successful application of this approach to the
synthesis of the spirocyclic isonitrile 4, in which a proton anti to
the epoxide was removed (Scheme 2),3 we treated isonitrile
epoxide 5 with ButOK in ButOH but obtained only the a,b-
unsaturated ketone 6. This presumably arises by rearrangement
of the initially formed oxyanion 7 as proposed in Scheme 3.
This result reflects observations made by Scho¨llkopf et al.
regarding the stability of g-hydroxy isonitriles in saturated
acyclic systems.4 They observed that under basic protic
conditions cyclisation to the corresponding 5,6-dihydro-
4H-1,3-oxazine occurs whereas under basic aprotic conditions
the starting material was recovered. Thus we turned to the use of
strong bases in an aprotic solvent system. In this context it has
been shown that epoxides can be isomerised to allylic alcohols
with lithium amide bases via a syn-elimination process and a
cyclic six-membered transition state has been postulated to
account for this.5 Treatment of 8 with LDA in THF at 220 °C
gave a 2:1 mixture of 9:10 with 9 presumably arising via a syn-
b-elimination (Scheme 4).† In an attempt to optimise the
selectivity of this process we exchanged the LDA for lithium
bis(trimethylsilyl)amide and observed a reversal of stereo-
selectivity with a 9:10 ratio of 1:7. To take advantage of this
enhanced selectivity we returned to diastereoisomer 5 and found
NC
Me
Me
R
R
H
H
O
Base
Me
CN
+
R
R
(see text)
NC
NC OH
Me OH
NC
NC
O
O
i
9
10
O
NC
Scheme 2 Reagents and conditions: i, KOBut, THF, 278 °C, 56%
O
HO
8
R = (CH2)6Me
4
Scheme 4
Chem. Commun., 1997
2389

View Article Online
i, ii
In summary we have noted a reversal of the stereochemistry
of base mediated isomerisation of a series of isonitrile epoxides
to allylic alcohols on switching from LDA to lithium bis-
(trimethylsilyl)amide. Using this observation we have prepared
racemic desepoxyaerocyanidin 2.
(CH2)7
CO2H
(CH2)6
CO2H
11
12
iii–v
R
R = o-C6H4(CO)2N-
NHCHO
vi, vii
Footnotes and References
Me
Me
(CH2)8CO2Me
(CH2)9CO2Me
* E-mail: jack.baldwin@dpl.ox.ac.uk
† Given the established acidity of protons adjacent to isonitriles10 we cannot
preclude the possibility that the isomerisation goes, at least in part, via an
E1cb mechanism. In circumstances where a substantial portion of the
product derived from this mechanism the differences in stereochemical
outcome between the diastereoisomers would require that ring opening of
the epoxide is faster than bond rotation.
14
15
13
viii
NC
NHCHO
ix, x
Me
(CH2)8CO2H
Me
(CH2)8CO2Me
xi, xii
O
O
H
16
1 W. L. Parker, M. L. Rathnum, J. H. Johnson, J. S. Wells, P. A. Principe
and R. B. Sykes, J. Antibiot., 1988, 41, 454.
2 J. E. Baldwin, R. M. Adlington, I. A. O’Neil, A. T. Russell and
M.-L. Smith, Chem. Commun., 1996, 41.
3 J. E. Baldwin, R. M. Adlington, J. Chondrogianni, M. S. Edenborough,
J. W. Keeping and C. B. Ziegler, J. Chem. Soc., Chem. Commun., 1985,
816.
xi
H
Me
Me
(CH2)8CO2H
(CH2)8CO2H
NC OSiMe2But
NC OH
17
2
4 U. Scho¨llkopf and R. Jentsch, Angew. Chem., Int. Ed. Engl., 1973, 12,
323.
5 J. K. Crandall and M. Apparu, in Organic Reactions, Wiley, New York,
1983, vol. 29, p. 345; R. P. Thummel and B. Rickborn, J. Am. Chem.
Soc., 1970, 92, 2064.
6 K. M. Morgan and J. J. Gajewski, J. Org. Chem., 1996, 61, 820.
7 J. A. Marshall and A. W. Garofalo, J. Org. Chem., 1993, 58, 3675.
8 A. H. Hoveyda, D. A. Evans and G. C. Fu, Chem. Rev., 1993, 93, 1307;
W. R. Roush, J. A. Straub and R. J. Brown, J. Org. Chem., 1987, 52,
5127.
Scheme 5 Reagents and conditions: i, (COCl)2, DMF, CH2Cl2; ii, (a)
CH2N2, Et2O, Et3N then Ag2O, Na2S2O5, dioxane, H2O; (b) Br2, Et2O then
Na, liq. NH3, 66% overall; iii, BunLi, THF, 220 °C then MeCHO, THF,
25 °C, 58%; iv, CH2N2, Et2O, 0 °C, 100%; v, phthalimide, Ph3P, DEAD,
THF, 16 h, 81%; vi, H2, Lindlar’s catalyst, quinoline, hexane, EtOAc, 24 h,
90%; vii, MeNH2 (40% aq.) CH2Cl2, 16 h then AcOCHO, CH2Cl2, 67%
(two steps); viii, MCPBA, CH2Cl2, 79%; ix, Tf2O, Pri2NEt, CH2Cl2,
278 °C, 83%; x, Ba(OH)2, MeOH then citric acid, 87%; xi, LHMDS, THF,
215 °C; xii, TBSOTf, 57% (for two steps)
9 J. E. Baldwin and I. A. O’Neil, Synlett, 1990, 603.
10 G. E. Niznik, W. H. Morrison and H. M. Walborsky, J. Org. Chem.,
1974, 39, 600.
to permit full characterisation. The geometry of the double bond
was supported by NOE spectroscopy, a 6% enhancement in the
intensity of the vinylic proton resonance on irradiation of the
vinyl methyl resonance being observed.
Received in Cambridge, UK, 22nd September, 1997; 7/06830A
2390
Chem. Commun., 1997