Selected regiocontrolled transformations applied to the synthesis of (1S)-cis-chrysanthemic acid from (1S)-3,4-epoxy-2,2,5,5-tetramethylcyclohexanol

Article abstract of DOI:10.1039/b808695h

(1S)-cis-Chrysanthemic acid has been prepared in a few steps with complete control of the relative and absolute stereochemistry using regiocontrolled epoxide ring opening, diol mono-oxidation and cyclopropanation. The Royal Society of Chemistry.

Full text of DOI:10.1039/b808695h

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Selected regiocontrolled transformations applied to the synthesis of
(1S)-cis-chrysanthemic acid from (1S)-3,4-epoxy-2,2,5,5-
tetramethylcyclohexanolw
Alain Krief,*a Humaira Y. Gondalab and Adrian Kremera
Received (in Cambridge, UK) 22nd May 2008, Accepted 10th July 2008
First published as an Advance Article on the web 26th August 2008
DOI: 10.1039/b808695h
(1S)-cis-Chrysanthemic acid has been prepared in a few steps
with complete control of the relative and absolute stereoche-
mistry using regiocontrolled epoxide ring opening, diol mono-
oxidation and cyclopropanation.
In the course of a study involving the enantioselective syn-
thesis of chrysanthemic acids1,2 from 2,2,5,5-tetramethylcyclo-
hex-3-enone 2, we needed to open the epoxide ring present in 5
to generate the bromohydrin 7. This reaction has been carried
out with several reagents but it was unselective and produced a
Scheme 2
mixture of the two regioisomeric bromohydrins 6 and 7
(Scheme 1).2a
We therefore developed a novel strategy, shown in Scheme 2.
related (1R)-trans-chrysanthemic acid 1Rtrans, a precursor of
the natural pyrethrin I or S-bioallethrin, the most powerful
indoor insecticide1 (Scheme 3a).
This takes advantage of the efficient enantioselective reduction
of the b,g-unsaturated enone 2 using (À)-Ipc2BCl2,3 and the
Furthermore, this strategy can be easily adapted (by using
molybdenum-catalysed epoxidation of the resulting homoallyl
alcohol 3 by t-butylhydroperoxide that we have already per-
(+)-Ipc2BCl3) to the stereoselective synthesis of (1R)-cis-chry-
santhemic acid 1Rcis, a precursor of scalemic cis-deltametrinic
formed (Scheme 1).2a,4 We expected that the hydroxyl group in
acid 1RDcis (Scheme 3b) and to deltamethrin, the most active
4, which is cis to the epoxide ring, will control the regio-
chemistry of the epoxide ring opening, leading to 8.
outdoor insecticide.1
A successful synthesis faces the following challenges: (i)
Oxidation of the 1,3-diol 8 to the b-hydroxyketone 7,
followed by base-promoted carbocyclisation, was expected to
Regioselective epoxide ring opening (4 to 8) (Scheme 4a); (ii)
Oxidation of an alcohol to a ketone (8 to 7) with the require-
produce 11, a known5 precursor of (1S)-cis-chrysanthemic
ment that it should exclusively involve the hydroxyl group
acid 1Scis. The latter delivers, on epimerisation at C-1, the
farthest from the halogen atom (Scheme 4b); (iii) Selective
1,3-carbocyclisation of 7 to 11 possessing the cyclopropane
ring, rather than 1,3-O-cycloalkylation generating the epoxide
present in 5 (Scheme 4c).
We expected that the halogen would play an important role
in each of the processes shown in Scheme 4, and therefore we
tested a series of reagents able to perform the epoxide ring
opening regioselectively (Scheme 5). The yields of halohydrins
were usually good to excellent, but the regioselectivity was
extremely poor. For example, reaction of 4 with titanium
tetrabromide6 mainly provided the bromohydrin 9Br posses-
sing the unwanted regiochemistry (Table 1, entry a).
Scheme 1 Reagents and conditions: (i) 1.05 eq. (À)-Ipc2BCl, neat,
25 1C, 48 h; (ii) 2.2 eq. diethanolamine, Et2O, 25 1C; (iii) 1.5 eq.
t-BuOOH, 0.015 eq. Mo(CO)6, C6H6, 80 1C, 2 h; (iv) PDC, CH2Cl2,
20 1C, 0.33 h; (v) 0.5 eq. TiBr4, CH2Cl2, 20 1C, 2 h.
a
`
´
Laboratoire de Chimie Organique de Synthese, Facultes
Universitaires N.-D. de la Paix, 61 rue de Bruxelles, 5000 Namur,
Belgium. E-mail: alain.krief@fundp.ac.be; Fax: +32 (81)724536;
Tel: +32 (81)724539
b On leave from the Chemistry department of the University of
Sargodha, Pakistan
w Dedicated with great appreciation to Prof. E. J. Corey on the
occasion of his 80th birthday.
Scheme 3
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 4753–4755 | 4753

View Article Online
Scheme 4
Scheme 7
The first results involving pyridinium chlorochromate
(PCC)11 were disappointing, since the starting material was
recovered unchanged even after standing in dichloromethane
for more than 3 days at 20 1C. tert-Butyl hydroperoxide in the
presence of vanadium di(acetylacetonate),12 which so readily
produced 3,4-oxido-2,2,5,5-tetramethylcyclohexanone when
we tried to epoxidise (1S)-2,2,5,5-tetramethylcyclohexanol to
3,4-oxido-2,2,5,5-tetramethylcyclohexanol,2b proved only
slightly better: although it effectively produced the desired
ketone 7 with complete regiocontrol, the reaction was very
slow and the yield very modest (C6H6, 80 1C, 84 h, 19%).
We also tested an alternative method involving selective
Scheme 5
The reagent pair titanium isopropoxide and bromine (giving
BrTi(OiPr)3)7 was, among those tested, the only one that
delivered 8Br in extremely good yield with high regiocontrol
(Table 1, entry b); the structure of 8Br was unambiguously
assessed by XRD.8 In addition, 8Br was easily separated from
9Br by silica gel chromatography.9
We assume that the regioselective synthesis of 8Br from
BrTi(OiPr)3 (Table 1, entry b) results, according to the
Furst–Plattner rule,10 from the trans-diaxial nucleophilic ring
monoacetylation of
8 and oxidation of the resulting
3-(acetyloxy)cyclohexanol. Since acetylation was extremely slow
and poorly regioselective, we did not follow this ‘‘protection–
deprotection’’ strategy.
We ultimately found that the Jones reagent13 led to formation
of 7 in 91% yield (0.66 eq. H2CrO4, acetone, 0 1C, 0.75 h),
together with 4% 4-bromo-2,2,5,5-tetramethylcyclohexane-1,3-
dione. By following the reaction by GC, we found that over-
oxidation of 7 only occurs once 8Br has completely disappeared.
The final goal was to find conditions for the 1,3-elimination of
¨
opening of the epoxide ring of the less stable conformer 4B
stabilised through chelation of the oxygen atoms of the
alcohol and of the epoxide by Ti(IV) (Scheme 6).
The same rule,10 applied to the non-chelated and more stable
conformer 4A, can rationalise the reversed selectivity observed
when TiBr4 is instead used (Scheme 6; Table 1, entry a). Work
is in progress to understand these discrepencies.
The selective oxidation of the 1,3-diol 8Br to the 3-hydroxy-
ketone 7 (Scheme 4b) was the next goal to achieve. This step
was extremely challenging since the oxidation needed to take
place selectively at one of the two alcohols, which are both
equatorial. Furthermore, competing over-oxidation and
potential ‘‘retro-aldol’’ reaction of 7 needed to be avoided.
7
providing the cyclopropane moiety present in exo-
4-hydroxy-3,3,6,6-tetramethylbicyclo[3.1.0]hexan-2-one 11. This
transformation is not a simple task, since mono-metallation is
expected to occur under kinetically controlled conditions at the
hydroxyl hydrogen (Scheme 7a) rather than a to the carbonyl
group (Scheme 7c) of 7, suggesting that the unwanted epoxide 5
will be produced at the expense of 11.
Table 1 Epoxide ring-opening of 4 (see Scheme 5)
Table 2 Reaction of 7 with base (see Scheme 7)
t/
h
8 + 9
(%)
Entry Reagent
Eq. T/1C
X
8/9
11
5 (%) (%)
Entry Reagent
Eq. N/Ra T/1C t/h 5/11
a
b
c
TiBr4
0.5
20
72 Br 87
Br 87
32 : 68
93 : 7
73 : 27
—
Ti(OiPr)4 + Br2 1.1
Ti(OiPr)4 + I2 1.1
Ti(OiPr)4 + Cl2 1.1
0 - 20
20
20
5
30
a
b
c
d
e
f
LDA
LDA
LDA
LDA
LiTMP
LiTMP
LiTMP
LiHMDS17
KHMDS17
KOH
1
2
2
3
1
2
2
3
3
2
N
N
R
R
N
N
R
R
R
N
20
20
20
20
0.5 96 : 4
92
—
—
60
85
—
77
83
32
—
—
I
66
0
0.5 87 : 13 83
0.5 24 : 76 20
d
24 Cl
0.3 5 : 95
0.5 90 : 10 86
0
20
À25
À25
20
20
20
1
1
3
23 : 77 21
0 : 100
70 : 30 63
g
h
i
—
0.3 100 : 0 83
12 90 : 10 90
j
a
N = normal addition; R = reverse addition.
Scheme 6
ꢀc
This journal is The Royal Society of Chemistry 2008
4754 | Chem. Commun., 2008, 4753–4755

View Article Online
Scheme 8 Reagents and conditions: (i) 1.05 eq. tBuO2H, 0.015 eq. Mo(CO)6, C6H6, 80 1C, 2 h; (ii) Ti(OiPr)4 + Br2, CH2Cl2, 0 - 20 1C, 5 h;
(iii) 0.66 eq. H2CrO4, acetone, 0 1C, 0.75 h; (iv) 2 eq. LiTMP, reverse addition, À25 1C, 1 h; (v) 1.1 eq. MsCl, CH2Cl2; (vi) 6 eq. tBuOK, 3 eq. H2O;
(vii) aq. HCl.
This proved to be the case when metallation of 7 was carried
out by adding a single equivalent of LDA14,15 or LiTMP15,16
to 7 dissolved in THF (Scheme 7; Table 2, entries a and e),
since 5 was produced almost exclusively (5/11: 96 : 4). Similar
results were obtained when potassium amides were used
instead under similar conditions (Table 2, entry i).
In the course of this work, we found that the results
described for the bromohydrin 7 cannot be systematically
transposed to the related chlorohydrin because the lithium
alcoholate has a lower propensity to cyclise to 5. Those results
will be reported in due course.
Taking into account those preliminary results, we initially
considered trapping the first-formed alcoholate using tri-
methylsilyl chloride, expecting to prevent epoxide formation
and allowing the synthesis of the cyclopropane ring present in
11. However, we did not favour this option because it would
require a lengthy protection–deprotection strategy.
Notes and references
1 (a) A. Krief, in Stereocontrolled Organic Synthesis. A ‘Chemistry
for the 21st Century’ Monograph, ed. B. M. Trost, International
Union of Pure and Applied Chemistry, Blackwell Scientific,
Oxford, 1994, pp. 337–397 and references cited therein; (b) F.
Naumann, in Chemistry of Plant Protection, ed. W. S. Bowers, W.
Ebing, D. Martin and R. Wegler, Springer-Verlag, Berlin, 1990;
(c) J. Tessier, Chem. Ind. (London), 1984, 199–204.
2 (a) A. Krief, S. Jeanmart and A. Kremer, Heterocycles, 2008
76, accepted; (b) A. Krief, S. Jeanmart and A. Kremer, to be
submitted.
3 (a) H. C. Brown, J. Chandrasekharan and P. V. Ramachandran,
J. Am. Chem. Soc., 1988, 110, 1539; (b) 3 [a]2D0 = À64.1 (c = 1.02 in
CHCl3).
4 K. B. Sharpless and R. C. Michaelson, J. Am. Chem. Soc., 1973,
95, 6136–6137.
We then envisaged a strategy involving metallation of the
b-alkoxyketone 12 to generate the dialkoxide 13, expecting
that C-alkylation (producing the cyclopropane ring) would
favorably compete with the epoxide formation. This would
only be feasible if enolate formation (12 to 13) was faster than
epoxide formation (12 to 5). In order to preclude epoxide
formation, we decided to perform the reaction with an excess
of base. The choice of reverse addition of the reactants became
obvious (addition of 7 to the base; ‘‘R’’ mode), and proved to
be highly beneficial.
5 A. Krief and A. Kremer, Synlett, 2007, 607–610.
6 For related reactions involving TiCl4, see: (a) M. Chini, P. Crotti,
L. A. Flippin and F. Macchia, J. Org. Chem., 1990, 55, 4265–4272;
(b) M. Shimizu, A. Yoshida and T. Fujisawa, Synlett, 1992,
204–206.
We found that the whole process could be successfully
achieved by simply performing the addition (‘‘R’’ mode) of
two equivalents of particularly strongly basic LiTMP
(Scheme 7; Table 2, entry g). We were not surprised to find
that LDA (Table 2, entry c) and LiHMDS (Table 2, entry h),
which are not as strong bases as LiTMP,16a deliver, under
similar conditions, a lower amount of 11. Although the
formation of a much higher amount of 11 can be achieved
when 3 equivalents of LDA are used (Table 2, entry d), this is
not the case for LiHMDS, which is known to be an even
poorer base than LDA (Table 2, entry h).
7 E. Alvarez, M. T. Nunez and V. S. Martin, J. Org. Chem., 1990, 55,
3429–3431.
8 A. Kremer, A. Krief, J. Wouters and B. Norberg, Acta Crystallogr.,
Sect. E: Struct. Rep. Online, 2000, to be submitted.
9 Separation of the 8Br + 9Br mixture was carried out by chromato-
graphy on silica gel, eluting with pentane–ether 50 : 50. 8: Rf 0.30;
9: Rf 0.56.
10 (a) E. L. Eliel, N. L. Allinger, S. J. Angyal and G. A. Morisson,
Conformational Analysis, Interscience, New York, 1965, p. 102;
(b) A. Furst and P. A. Plattner, Abstracts of Papers of the 12th
¨
International Congress of Pure and Applied Chemistry, 1951, p. 409.
11 E. J. Corey and J. W. Suggs, Tetrahedron Lett., 1975, 16,
2647–2650.
12 T. Itoh, K. Jitsukawa, K. Kaneda and S. Teranishi, J. Am. Chem.
Soc., 1979, 101, 159–169.
13 (a) K. Bowden, I. M. Heilbron, E. R. H. Jones and B. C. L.
Weedon, J. Chem. Soc., 1946, 39–45; (b) T. Ray, in Encyclopedia of
Reagents for Organic Synthesis, ed. L. Paquette, John Wiley &
Sons, 1995, vol. 2, p. 1261.
The beneficial effect of using the ‘‘R’’ rather than the ‘‘N’’
mode of addition of the reagents (so that 7 is always kept in an
excess of base) proved to be, as expected, extremely important
(Table 2, compare entries g to f and c to b), especially when the
less reactive LDA is used.
The cooperative effect of the lithium cation has to be
pointed out, since to a certain extent it is playing the role of
a hydroxyl ‘‘protecting group’’ (as could have the trimethyl-
silyl group), avoiding or lowering competing epoxide ring
formation. This is not the case when potassium bases are used
(Table 2, entries i and j), since the higher ionic character of the
first-formed metal alcoholate increases the rate of epoxide
formation, precluding the formation of 13 (Scheme 7; Table 2,
entry i; compare with entry h).
14 (a) M. Hamell and R. J. Levine, J. Org. Chem., 1950, 15, 162–168;
(b) W. I. I. Bakker, P. L. Wong and V. Snieckus, in Encyclopedia of
Reagents for Organic Synthesis, ed. L. Paquette, John Wiley &
Sons, 1995, vol. 5, p. 3096.
15 (a) H. Ahlbrecht and G. Schneider, Tetrahedron, 1986, 42,
4729–4741; (b) F. J. Vinick, M. C. Desai, S. Jung and P. Thadeio,
Tetrahedron Lett., 1989, 30, 787–788.
16 (a) C. L. Kissel and B. Rickborn, J. Org. Chem., 1972, 37, 2060;
(b) M. Campbell and V. Snieckus, in Encyclopedia of Reagents for
Organic Synthesis, ed. L. Paquette, John Wiley & Sons, 1995,
vol. 5, p. 3166.
17 We trapped the mixture with MeOD and found incorporation of one
equivalent of deuterium a to the carbonyl group of 5.
18 (a) [a]2D0 = À51.8 (c = 1.11 in CHCl3; (b) [a]2D0 = À78.1 (c = 1.00
in CHCl3).
We proved the identity of compound 11 by comparison with an
authentic sample,18a and transformed it according to a known
procedure to (1S)-cis-chrysanthemic acid 1Scis (Scheme 8).2a,5,18b
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 4753–4755 | 4755