Syntheses of racemic and scalemic cis-chrysanthemic acid from β,γ-unsaturated cyclohexanol

Article abstract of DOI:10.1002/hlca.201200393

2,2,5,5-Tetramethylcyclohexane-1,3-dione is a valuable starting-material precursor of cis-chrysanthemic acid. The (1S)-stereoisomer is a precursor of pyrethrin I, the most active natural insecticide from Chrysanthemum cinerariifolium, whereas the (1R)-stereoisomer is efficiently transformed to deltamethrin, the most active commercially available pyrethroid insecticide. Several intermediates have been identified and used with variable success for that purpose.

Full text of DOI:10.1002/hlca.201200393

Helvetica Chimica Acta – Vol. 95 (2012)
2123
Syntheses of Racemic and Scalemic cis-Chrysanthemic Acid from
b,g-Unsaturated Cyclohexanol
by Alain Kriefa), Ste´phane Jeanmarta)b), Humaira Y. Gondala)c), and Adrian Kremer*a)d)
a
`
´
) Laboratoire de Chimie Organique de Synthese, Facultes Universitaires N.-D. de la Paix,
61 rue de Bruxelles, B-5000 Namur
b) Syngenta Crop Protection Mꢀnchwilen AG, Schaffhauserstrasse CH-4332 Stein
c) Chemistry Department, University of Sargodha, Pakistan
d
´
´
´
) Laboratoire de Chimie Organique et des Materiaux Supramoleculaires, Facultes Universitaires N.-D.
de la Paix, 61 rue de Bruxelles, B-5000 Namur
(phone: þ 32-81-725415; e-mail: adrian.kremer@fundp.ac.be)
Dedicated to Professor Dieter Seebach on the occasion of his 75th birthday
2,2,5,5-Tetramethylcyclohexane-1,3-dione is a valuable starting-material precursor of cis-chrysan-
themic acid. The (1S)-stereoisomer is a precursor of pyrethrin I, the most active natural insecticide from
Chrysanthemum cinerariifolium, whereas the (1R)-stereoisomer is efficiently transformed to deltameth-
rin, the most active commercially available pyrethroid insecticide. Several intermediates have been
identified and used with variable success for that purpose.
Introduction. – 2,2,5,5-Tetramethylcyclohex-3-en-1-one (1) possesses a C10 skel-
eton, as chrysanthemic acid, a CO group, and two C-atoms each bearing geminal Me
groups in suitable position to generate, after proper reactions, cis-chrysanthemic acid
(2cis) possessing the above mentioned CO group as part of its carboxy group and
geminal Me groups located at its vinylic C-atom and on its cyclopropane ring
(Scheme 1). cis-Chrysanthemic acid (2cis) is in turn an efficient precursor of
deltamethrin, the most powerful commercially available pyrethroid insecticides [1].
Scheme 1. Synthesis of 2cis via Addition of X2 to C¼C Bond of 1
We have reported, in the recent past, preliminary work involving the transformation
of 1 for that purpose [2], but the experimental part was missing in most of the related
papers [2a – 2c], which also provided fragmentary information about the topic. We
disclose in this article i) an integrated view on the whole topic and ii) scope and
limitation of each of the transformations, and provide, any time it has not yet been
done, experimental details.
ꢀ 2012 Verlag Helvetica Chimica Acta AG, Zꢁrich

2124
Helvetica Chimica Acta – Vol. 95 (2012)
Thus, the transformation of tetramethyl-cyclohexenone 1 to cis-chrysanthemic acid
(2cis) requires three distinct steps i) halogenation of its C¼C bond leading to 3,4-
dihalogeno-2,2,5,5-tetramethylcyclohexanones 3X (X ¼ Br: 1 equiv. Br2, CH2Cl2, ꢀ 788;
quant. yield, or 1 equiv. Br2, 0.1 equiv. AcNH2, CCl4, 08, 1 h, 98% yield; X ¼ Cl: 1 equiv.
Cl2, 0.1 equiv. AcNH2, CCl4, ꢀ 408, titration; 92% yield) [3] (Scheme 1). ii) Reaction
of the resulting 3X with a base, producing an enolate which, after subsequent
elimination of the halogen atom located three C-atom away, generates the cyclo-
propane ring, part of 4-halogeno-3,3,6,6-tetramethylbicyclo[3.1.0]hexan-2-ones 4X
(X ¼ Br or Cl, Scheme 1) [3]. iii) Reaction of 4X with an appropriate source of HOꢀ
ion, acting as a nucleophile on its CO group and subsequent fragmentation which
releases the halide ion and produces the desired vinyl cyclopropane carboxylate
precursor of cis-chrysanthemic acid (2cis) after acid hydrolysis (Scheme 1) [3].
The synthesis of 4-halogeno-3,3,6,6-tetramethylbicyclo[3.1.0]hexan-2-ones 4X
(X ¼ Br or Cl) has been efficiently achieved at low temperature using stoichiometric
amounts of iPr2NLi (LDA) in THF (Table 1, Entries a and d; Conditions 1) or tBuOK in
THF (Table 1, Entries b and e; Conditions 1) acting as a base [3]. The same
transformation carried out on 3Br required excess of MeOLi in MeOH and led to
bicyclic derivative 4Br at higher temperature and in lower yield (Table 1, Entry c;
Conditions 1). Therefore, the sub-stoichiometric amount of reagent used does not
promote under such drastic conditions the Grob fragmentation [4] which would have
led instead to methyl chrysanthemate [5].
Table 1. Synthesis of cis-Chrysanthemic Acid (2cis) Involving 3,4-Dihalogeno-2,2,5,5-tetramethylcyclo-
hexanones 3X. WPH ¼ Aq. KOH in DMSO; APH ¼ tBuOK/H2O.
Entry
X
Base
Conditions 1 for 3 ! 4 Yield of Conditions 2 for 4 ! 1
4 [%]
Yield of
2cis [%]
a
b
c
d
e
Br 1 equiv. LDA
THF, ꢀ 788, 1 h
86
WPH, DMSO, 708, 0.8 h 87
APH, THF, 208, 0.5 h 94
APH, DMSO, 208, 0.3 h 53
WPH, DMSO, 708, 0.8 h 75
t
Br 1.2 equiv. BuOK THF, ꢀ 78 to 208, 1 h 94
Br 6 equiv. MeOLi
MeOH, 658, 48 h
THF, ꢀ 788, 1 h
49
88
Cl 1 equiv. LDA
t
Cl 1.2 equiv. BuOK THF, ꢀ 78 to 208, 1 h 89
APH, THF, 208, 0.5 h
91
The synthesis of cis-chrysanthemic acid (2cis) from 4X (X ¼ Br or Cl) has been
achieved at 708 using WPH (i.e., KOH in aqueous DMSO (DMSO/H2O 4 :1); Table 1,
Entries a and d; Conditions 2) or better at 208 using SwanꢀGassman reagent [6],
generated in situ from tBuOK and H2O (tBuOK/H2O 3 :1 (APH); Table 1, Entries b,c,
and e; Conditions 2). We have found that, whereas 4Cl is less reactive than 4Br towards
WPH, they react both very fast, when APH is used, and provide chrysanthemic acid 2cis
in very high yield especially if the reactions are carried out in THF (Table 1, cf.
Entries b,c,e; Conditions 2) [7].
It was tempting to perform both the cyclization and fragmentation steps in the same
pot using KOH acting sequentially as a base first then as a nucleophile. This has been
successfully achieved from 3 whatever the conditions used (WHP-DMSO [3], APH-
THF, or APH-DMSO [8]; Scheme 2 and Table 2). APH in THF, however, is preferred
since it generates 2cis under more reliable and milder conditions than those involving

Helvetica Chimica Acta – Vol. 95 (2012)
2125
Scheme 2. ꢁOne-Potꢂ CyclizationꢀFragmentation of 3X
Table 2. ꢁOne-Potꢂ CyclizationꢀFragmentation of 3X
Entry X1 ¼ X2 Substrate Conditions
Yield of 2cis [%]
a
b
c
Br
Cl
Br
Cl
Br
3Br
3Cl
3Br
3Cl
3Br
i) 6 equiv. KOH, DMSO/H2O (4 : 1), 708, 2 h. ii) H3Oþ.
i) 6 equiv. KOH, DMSO/H2O (4 : 1), 708, 4 h. ii) H3Oþ.
87
64
94
80
t
i) 6 equiv. BuOK, 3 equiv. H2O, THF, 208, 0.4 h. ii) H3Oþ.
t
d
e
i) 6 equiv. BuOK, 3 equiv. H2O, THF, 208, 1 h. ii) H3Oþ.
t
i) 6 equiv. BuOK, 3 equiv. H2O, DMSO, 208, 0.5 h. ii) H3Oþ. 65
WHP or AHP in DMSO [8]. As expected 3Br reacts faster than 3Cl towards both AHP
and WHP (Scheme 2, cf. Entry a to b and c to d).
It is interesting to notice that performing the reaction of 5Br, the didemethyl analog
of 3Br, with WPH produces didemethyl-isopyrocine 6 instead of the didemethyl-vinyl-
cyclopropanecarboxylic acid 7cis. The latter is in fact produced when the reaction is
carried out with APH in THF (Scheme 3).
Scheme 3. ꢁOne-Potꢂ CyclizationꢀFragmentation of 5Br
The transformation of 5Br to 7cis can be rationalized by assuming that the HOꢀ ion
acts first as a base towards 5Br as outlined in Scheme 2 (X1, X2 ¼ Br; R ¼ H). The
transformation of 5Br to 6 is in turn rationalized (Scheme 4) by assuming that the HOꢀ
ion reacts instead as a nucleophile on the CO group of 5Br leading to 8 that subsequently
fragments to 9Br and collapses to 6 (R ¼ H) via an intramolecular substitution of the
allyl bromide by the carboxylate intermediate [8].
Results and Discussion. – We became next interested to carry out a related
transformation on the 3,4-epoxycyclohexanone 11, according to the retrosynthetic
Scheme 5.
The challenging transformation is without doubt that of 1 to 12, since the
transformation of the latter to 2cis has been already achieved (Scheme 6) [9]. It involves
mesylation of the alcohol 12 (1.2 equiv. MsCl, 1.5 equiv. Et3N, CH2Cl2, ꢀ 108, 0.75 h,

2126
Helvetica Chimica Acta – Vol. 95 (2012)
Scheme 4. Rationalization for the Transformation of 5Br to 6
Scheme 5. Retrosynthetic Scheme Involving 11
Scheme 6. Synthesis of 2cis Involving 12?
t
i) 1 equiv. Br2, CCl4, 08, 2 h. ii) 1.2 equiv. BuOK, ꢀ 788 to r.t., 1 h. iii) 1,2-dimethoxyethane (DME),
ꢀ 788, 1 h, then 208, 1 h. iv) 1 equiv. NaBH4, 1 equiv. CeCl3 · 7 H2O, MeOH, ꢀ 788, 1.5 h. v) 1.1 equiv.
t
MsCl, 1.5 equiv. NEt3, CH2Cl2, ꢀ 108, 0.75 h. vi) a) 6 equiv. BuOK/3 equiv. H2O, THF, 208, 0.3 h;
t
b) 6 equiv. BuOK/3 equiv. H2O, DMSO, 208, 0.4 h.

Helvetica Chimica Acta – Vol. 95 (2012)
2127
86% yield) [9a] leading to 4OMs, followed by reaction of the HOꢀ ion, leading via
HallerꢀBauer reaction/Grob fragmentation to cis-chrysanthemic acid (2cis) after acid
hydrolysis. We have shown that this transformation cannot be achieved using WPH in
DMSO [9a], in contrast to what has been achieved with the related bicyclic halides 4X
(Scheme 1 and Table 1, Entries a and d; Conditions 2) [3], since polymeric substances
are formed, beside minute amounts (7%) of cis-chrysanthemic acid (2cis) [10]. The
latter can be, however, efficiently produced using APH in THF or better in DMSO
(Scheme 6) [10].
The more efficient transformation of 1 to 12 would have involved epoxidation of 1
and base promoted tandem metalationꢀepoxide ring opening of the resulting 7-
oxabicyclo[4.1.0]heptane 11. A related transformation has been successfully achieved
on 9-oxabicyclo[6.1.0]nonane 16, the higher homologue of 11, which is expected,
however, to involve a much less strained transition state (Scheme 7; 17K to 18) [11].
Scheme 7. Synthesis of 18 from 9-Oxabicyclo[6.1.0]nonane 16
3,4-Epoxycyclohexanone 11 has been prepared from 1 and mCPBA (2 equiv.,
4 equiv. NaHCO3, CH2Cl2, 208, 40 h, 82% yield) [2c] but we were unable to achieve the
t
epoxide ring opening and the subsequent carbocyclization using a) BuOK as
successfully used for its higher homolog 16 (Scheme 8, compare to Scheme 7); b)
KOH in DMSO (5 equiv., 208); c) EtONa in EtOH at reflux, or d) LDA in THF (ꢀ 78
to 208) [12].
Scheme 8. Attempt to Synthesize 12 from 3,4-Epoxycyclohexanone 11
Since this could be ascribed to the extreme strain generated for the required
alignment of the orbitals in the transition state of the planned substitution reaction, the
only options left for achieving the desired transformation was to destroy, even partially,
the C(4)ꢀO bond of the epoxide ring, expecting the intermediate formation of a
carbocation. The transformation of 11 to 12 could be achieved on treatment of i) the
preformed lithium enolate 19Li (from LDA or LiHMDS on 11) using Lewis acids such
as BF3 · Et2O (1 to 5 equiv.), Et2AlCl (5 equiv.), scandium(III) triflate (0.2 equiv.) or
ii) the related enol 19H generated on reaction of 11 with BF3 · Et2O in CHCl3 or with
super acids such as (BF3 · Et2OꢀHBF4) or (BF3 · Et2OꢀHBr) in the same solvent. Both
approaches proved unsuccessful, the former leading to the recovery of 11 after
acidification, and the latter to its destruction [5].

2128
Helvetica Chimica Acta – Vol. 95 (2012)
We, however, found that the required epoxide ring opening could be successfully
achieved stepwise by reacting the lithium enolate 19Li, prepared from 11 and LDA
(1 equiv.) with Me3SiI (TMSI, 2 equiv.). We obtained evidence that the latter first traps
the enolate intermediate then effects the epoxide ring opening [13] to finally produce a
mixture of the two regioisomeric bis(silyloxy)-cyclohexenones 20’ and 20’’ (Scheme 9).
Scheme 9. ꢁOne-Potꢂ Synthesis of Iodo-bis(silyloxy)-cyclohexenes 20’ and 20’’. LiHMDS ¼ Lithium
hexamethyldisilazide, LDA ¼ lithium diisopropylamide; LiTMP ¼ lithium tetramethylpiperidide.
Lithium tetramethylpiperidide (LiTMP) [14], the strongest among the different
bases tested (pKa ca. 37) [15], provides the best yield of 20. It nevertheless does not
produce the highest percentage of the regioisomeric 20’, the potential precursor of the
bicyclo[3.1.0] derivative 4OMs and of 2cis (Scheme 9). A slightly improved ratio (20’/20’’
58 :42) has been achieved by reacting TMSI with the preformed silylenolate 19TMS
,
freed from the amine by-product resulting from the metalation of 11 (Scheme 10). We
have not been able however, to promote the regioselective formation of 20’ using the
strategy outlined in Scheme 5.
Scheme 10. Two-Steps Synthesis of 20’ and 20’’
The synthesis of the silylenolate 19TMS is better achieved on reaction of the lithium
enolate 19Li with Me3SiCl (TMSCl), since it is not only cheaper than TMSI, but also it
does not affect the epoxide ring opening observed above, even if it used in excess, a
property that it shares with TMSBr (Scheme 10).
We have also successfully performed the epoxide ring opening on 11 using HBr
[16], HCl [17], or Li2NiBr4 [18], in THF; Ph3PHBr [19], POCl3 þ DMAP, titanium
tetrahalides (X ¼ Cl, Br) [20], or beryllium dichloride [21] in CH2Cl2, and have
generated the corresponding regioisomeric b-chloro- and b-bromo-hydroxy-cyclo-
hexanones 21’ and 21’’ in reasonably good yields but as a stereoisomeric mixture
(Scheme 11 and Table 3).

Helvetica Chimica Acta – Vol. 95 (2012)
2129
Scheme 11. Synthesis of Halogeno-hydroxy-cyclohexanones 21’ and 21’’
Table 3. Synthesis of b-Halogeno-hydroxy-cyclohexanones 21’ and 21’’
Entry Reagent
Solvent Equiv. T [8] t [h]
X
21X Yield of 21X [%] 21’/21’’ Ratio
a
b
c
d
e
f
TiCl4
TiBr4
TiBr4
TiBr4
CH2Cl2
CH2Cl2 0.5
CH2Cl2
CH2Cl2
THF
1
20
20
72
2
5
1
2
Cl 21Cl 92
Br 21Br 96
Br 21Br 55
Br 21Br 96
Br 21Br 75
Cl 21Cl 75
Br 21Br 41
Cl 21Cl 83
58 : 42
43 : 57
55 : 45
45 : 55
65 : 35
66 : 34
75 : 25
80 : 20
1
1
2
2
5
5
ꢀ 78
20
HBr [16]
HCl [17]
Li2NiBr4 [18] THF
BeCl2 [21] CH2Cl2
20
20
20 168
20 100
THF
3
g
h
Similarly, reaction of epoxy-cyclohexanone 11 with TMSI in THF and acid
hydrolysis of the intermediate b-silyloxy-iodo-cyclohexanones 22 using aqueous HCl
acid, also led to hydroxy-iodo-cyclohexanones 21I, but as an almost 1:1 mixture of
regioisomers (Scheme 12).
Scheme 12. Synthesis of Hydroxy-iodo-cyclohexanones 21’I and 21I’’
We have not been able to achieve the regioselective ring opening of 11, and, except
rare cases which are listed in Table 3, the reaction produces an almost 1:1 regioisomeric
mixture 21. The most valuable reagent for our purpose was beryllium dichloride but the
de of the resulting products (60%) was far from excellent (Scheme 11; Table 3, Entry h)
[21].
We, however, synthesized regioselectively 4-bromo-3-hydroxy-2,2,5,5-tetramethyl-
cyclohexanone (21’Br) by a different strategy which involves a multistep process in
which reduction of the CO group of 1 leads to the homoallylic alcohol 23 whose OH

2130
Helvetica Chimica Acta – Vol. 95 (2012)
group is able to direct the face selective epoxidation of the C¼C bond and to achieve
the regioselective ring opening of the resulting allylic epoxide 24 (Scheme 13). This was
effectively achieved by: i) reduction of 1 with NaBH4 (1.2 mol equiv., MeOH, 08, 1.3 h,
82% yield); ii) stereoselective Mo-catalyzed epoxidation of the resulting homoallylic
t
alcohol 23 (1.5 equiv. BuOOH, 0.015 equiv. Mo(CO)6, C6H6, 808, 2 h) leading to the
allylic epoxide 24 with the OH and epoxy moieties at the same face (88% yield, de
96%; Scheme 13) [22]; iii) regio- and stereoselective epoxide ring opening using the
couple Ti(OiPr)4 and Br2 [23], which almost exclusively leads to the diol 25B’ r bearing
the two OH moieties in a 1,3-cis-relationship, and a Br-atom with trans-relationships
with the two OH groups, as unambiguously assessed by X-ray diffraction [24]
(1.1 equiv. Br2, 1.1 equiv. Ti(OiPr)4, CH2Cl2, 0 to 208, 5 h, 93% yield, de 86%,
Schemes 13 and 14; and Table 4, Entry a) [23], and iv) regioselective oxidation of the
OH group of 25’Br bearing in d position the Br-atom, when carried out by the Jones
reagent [25] (0.66 equiv. H2CrO4, acetone, 08, titration, Scheme 13). This reagent
generated 21’Br in very high yield 92%, beside trace amount of 4-bromo-2,2,5,5-
tetramethylcyclohexane-1,3-dione 14 (4%).
Scheme 13. Synthesis of Bromo-hydroxy-cyclohexanone 21B’ r
Scheme 14. Synthesis of Diols 25’X and 25X’’
Table 4. Regiocontrolled Epoxide Ring Opening of 24 Using Ti(OiPr)4 and Halogen
Entry Reagent
Equiv. T [8]
t [h]
X
Yield of 25’ þ 25’’ [%] 25’X/25X’’
Ratio
’’
a
b
c
Ti(OiPr)4 þ Br2 1.1
0 – 20
20
5
72
30
24
Br 93
Br 87
I
25’Br/25
25’Br/25
25’I/25’I’
25’Cl/25
93 : 7
32 : 68
73 : 27
Br
’’
TiBr4
0.5
1.1
Br
Ti(OiPr)4 þ I2
20
20
66
0
’’
d
Ti(OiPr)4 þ Cl2 1.1
Cl
–
Cl

Helvetica Chimica Acta – Vol. 95 (2012)
2131
This series of transformations is quite exceptional, since, as it will be reported
below, control of the regio- and stereochemistry cannot be easily achieved using
reagents different from those we have selected. Thus, for example, epoxidation of 23
with mCPBA proved to be completely unselective [26] and led to 2,2,5,5-tetramethyl-7-
oxabicyclo[4.1.0]heptan-3-ol (24) in excellent yield, but as a 1:1 mixture of the two
diastereoisomers (1.5 equiv. mCPBA, 2 equiv. NaHCO3, CH2Cl2, 208, 4 h, 81% yield,
t
de 0%) and reagents involving H2O2 or BuOOH proved to be inefficient. This is, for
t
t
example, the case of i) BuOOH and (tBuO)3Al (1.5 equiv. BuOOH, 1.5 equiv.
(tBuO)3Al, C6H6, 808, 5 h, 0% yield) [27], ii) H2O2 in the presence of dicyclohex-
ylcarbodiimide (8.8 equiv. H2O2, 2 equiv. DCC, 2 equiv. KHCO3, MeOH, 208, 23 h,
t
25%, de 94%) [28], or iii) BuOOH and catalytic amounts of VO(acac)2 (1.5 equiv.
tBuOOH, 0.015 equiv. VO(acac)2, C6H6, 808, 5 h, 26% yield, de 98%). Although the
latter reagents allowed the highly diastereoselective epoxidation of 23 to the 3,4-
epoxycyclohexanol 24 bearing the epoxide ring on the same face as the OH group, as
described when the reaction was instead catalyzed by Mo(CO)6 (Scheme 13), the
conversion rate was low, and using VO(acac)2 competing formation of the epoxy-
cyclohexanone 11 was observed. Such behavior has been already reported with alcohols
whose OH group similarly is in equatorial position [29].
Similarly, we have been unable to achieve regiocontrolled epoxide ring opening
using TiBr4 (0.5 equiv. TiBr4, 208, 72 h; Scheme 14; Table 4, Entry b) or the related
couples Ti(OiPr)4 and I2 (Table 4, Entry c), or Cl2 (Table 4, Entry d) [23], as we did
using instead Ti(OiPr)4 and Br2, which is known to produce BrTi(OiPr)3) [23]
(Schemes 13 and 14; and Table 4, Entry a compared to Entries b – d).
We assume that the regioselective synthesis of 25’Br from BrTi(OiPr)3 (Scheme 14,
and Table 4, Entry a) results, according to FꢀrstꢀPlattner rule [30], from the trans-
diaxial nucleophilic ring opening of the epoxide ring of the less stable conformer 26b
(Scheme 15) stabilized through chelation by TiIV of the O-atom of the alcohol and the
epoxide ring as well.
Scheme 15. trans-Diaxial Nucleophilic Ring Opening of 26
The same rule [30], applied to the non-chelated and more stable conformer 26a
(Scheme 15), can rationalize the reversed selectivity observed when TiBr4 is instead
used (Scheme 14, and Table 4, Entry b). Work is in progress to understand these
discrepancies.
Finally, selective oxidation of the 1,3-diol 25’Br to the 3-hydroxyketone 21’Br which
has been so efficiently achieved using the Jones reagent [25] (Scheme 13) proved to be
extremely challenging, since the oxidation should take place selectively on one of the
two alcohols whose OH groups are both equatorial, and competing ꢂover-oxidationꢃ
and ꢂretro-aldol reactionꢃ on 21’Br should be avoided. The first results involving
pyridinium chlorochromate (PCC) [31] were disappointing, since the starting material

2132
Helvetica Chimica Acta – Vol. 95 (2012)
was recovered unchanged even after standing in CH2Cl2 for more than 3 d at 208.
tBuOOH in the presence of VO(acac)2 [29], which was expected to produce 3,4-epoxy-
2,2,5,5-tetramethylcyclohexanone (11) on reaction of 2,2,5,5-tetramethylcyclohex-3-
enol (23) to 2,2,5,5-tetramethyl-7-oxabicyclo[4.1.0]heptan-3-ol (24), proved to be only
slightly better although it effectively furnished the desired ketone 21’Br with complete
regiocontrol, but the reaction was very slow and the yield extremely modest (C6H6, 808,
84 h, 19%).
The transformation of 21’Br to 12, which was the last goal to achieve, was not a
simple task, since the base required to produce 12 could competitively generate the
epoxide 11 instead (Scheme 16).
Scheme 16. Selective Transformation of 21’Br to 12. BM ¼ Base (M is the counter ion).
These transformations involve first metallation of 21’Br which could either take
place on its OH group, leading to 27’Br, or on one of the CH2 H-atoms leading to 28’Br
under kinetic control then react to produce 11 by O-alkylation and the bicyclic
cyclopropane derivative 12 by C-alkylation directly or after a pre-equilibrium
(thermodynamic control; Scheme 17). Those can also be formed through the
dimetallated intermediate 29’Br especially if more than 1 equiv. of base is used
(Scheme 17). The nature of the leaving group X and of the counter ion M, the number
of equivalent of base used, as well as the experimental conditions should, therefore,
have a crucial impact on the chemioselectivity of the process. Furthermore the
conditions should be mild enough to avoid a ꢂretro-aldol reactionꢃ on the resulting 30
which is known to lead to polymeric materials [10].
We first performed the reaction of lithium amides on 21’Br, since it is not only the
potential precursor of the required bicyclic derivative 30 but also the only halohydrin of
the series which we have synthesized free from its regioisomer (Scheme 18). The
highest percentage of bicyclo[3.1.0] derivative 12 was observed by adding, at ꢀ 258 21B’ r
in THF to an excess (ꢁ2 equiv.) of base in the same solvent (Table 5, Entries d and h;
Conditions R). Best results have been obtained by using LiTMP (Table 5, Entry d;
Conditions R), the strongest among the various lithium amides used (LiTMP (pKa ¼
37) [14][15] > LDA (pKa ¼ 36) [32] > lithium hexamethyldisilazide (LiHMDS,
pKa ¼ 30) [12]).
The finding that some epoxide 11 was formed if the reverse order of addition was
used (Table 5, Conditions N, Entries c and d) especially if a single equivalent of LiTMP
was used (Table 5, Conditions N, Entries b and d) supports the assumption that
metallation occurs first on the OH group of 21’Br, and that epoxide formation is slow
enough, to allow, if another equivalent of strong base is present, a second metallation to
occur leading to the corresponding enolate 29’Br. Apparently, substitution of the Br-
atom by O-alkylation leading to epoxide formation is slow enough from 27’Br and 29B’ r

Helvetica Chimica Acta – Vol. 95 (2012)
2133
Scheme 17. O-Alkylation vs. C-Alkylation from 21’Br. BH ¼ Conjugate acid.
Scheme 18. Synthesis of 11 and 12 from 21B’ r
Table 5. Synthesis of 11 and 12 from 21B’ r
a
Entry Reagent Equiv. Conditions
) T [8] t [h] 12/11 Ratio Yield of 12 [%] Yield of 11 [%]
a
b
c
d
e
f
g
h
i
LiTMP
LiTMP
LiTMP
LiTMP
LDA
LDA
LDA
LDA
1
1
2
2
1
2
2
3
3
3
3
N
N
N
R
N
N
R
R
R
R
R
20 0.5
10 : 90
17: 83
77: 23
100 : 0
4 : 96
13 : 87
70 : 30
100 : 0
30 : 70
0 : 100
67: 33
–
–
77
83
–
86
83
21
–
92
83
21
–
ꢀ 25
1
1
1
ꢀ 25
ꢀ 25
20 0.5
20 0.5
–
ꢀ 25
1
1
3
57
80
32
–
ꢀ 25
LiHMDS
KHMDS
KHMDS
20
63
83
29
j
k
20 0.3
ꢀ 78
1
66
a) N ¼ normal addition (of the base to 21’), R ¼ reverse addition (of 21’ to the base).

2134
Helvetica Chimica Acta – Vol. 95 (2012)
to favor C-alkylation leading to the selective cyclopropane formation. Epoxide
formation, however, successfully compete if the reactivity of the alkoxide is enhanced
(K instead of Li; Table 5, Entries i and j), if the lithium amide is not strong enough to
achieve the second metallation rapidly (Table 5, Entries g,i, and Entries d and h) or
missing (Table 5, Entries a,b, and e). We also found that, as expected, the temperature
at which the reaction is carried out, influences the nature of the compounds formed
(Scheme 18 and Table 5).
The transformation of the resulting bicyclic hydroxy ketone 12 to cis-chrysantemic
acid (2cis) has been successfully achieved according to published procedures as outlined
in Scheme 19 [10].
Scheme 19. Synthesis of 2cis from 21B’ r
The next goal to achieve was to carry out the desired reaction on the mixture 21’/21’’
expecting to find conditions to selectively transform 21’ to bicyclic 12, precursor of cis-
chrysantemic acid 2cis, and its regioisomer 21’’ to the epoxide 11, respectively. This
would take advantage of the easy separation of 12 from 11 by chromatography on SiO2
due to the different functional groups present in each of the compounds and, therefore,
allowing efficient recycling of 11 through the intermediate formation of the
regioisomer mixture of 21’/21’’ (Scheme 20).
Scheme 20. Synthesis of 11 and 12 from a Mixture of 21’X and 21X’’
We were rather surprised to find that the conditions successfully used to transform
’’
21’Br to 12 do not apply to the 1:1 mixture 21B’ r/21 , which provide an intractable
Br

Helvetica Chimica Acta – Vol. 95 (2012)
2135
mixture of unidentified compounds, and this proved also to be the case if similar
reaction was carried out on the regioisomer mixture 21’I/21I’’.
’’
We could, however, transform the mixture 21’Cl/21 to the mixture of the
Cl
cyclopropane derivative 12 and epoxide 11 (Scheme 21; and Table 6, Entry d). This
was fortunate, since chlorohydrins were the only halohydrins, among those we have
produced, in which the amount of the regioisomer 21’ precursor of the bicyclo[3.1.0]
’’
carbocycle 12, was the highest (21’Cl/21 80 :20) [2b].
Cl
Scheme 21. Selective Cyclization of a Mixture of 21’Cl and 21C’ l
’’
Table 6. Selective Cyclization of a Mixture 21’Cl/21
Cl
’’
Entry
Reagent; equiv.;
mode of addition
T [8]
t [h]
12/11/21’Cl/21 Ratio
Yield of
isolated
12 [%]
Yield of
isolated
11 [%]
Cl
a
b
c
LiTMP; 1; N
LiTMP; 1; N
LiTMP; 2; R
i) LiTMP; 2; R;
ꢀ 25
20
1
0.5
1
21 : 0 : 58 : 21
67: 8 : 8 : 16
80 : 0 : 0 : 20
80 : 20 : 0 : 0
20
66
68
74
ꢀ 25
d
ꢀ 25
1.5
19
t
ii) BuOK; 2; N
’’
We found that the mixture of 21’Cl/21 (80 :20) reacted with 2 equiv. of LiTMP to
Cl
’’
furnish the cyclopropane derivative 12 and recovered 21 after acid hydrolysis,
Cl
indicating that alkylation exclusively took place of the bromohydrin 21’Cl, whereas O-
’’
alkylation of its stereoisomer 21 , which would have lead to the epoxide 11, did not
Cl
take place (Scheme 21, and Table 6, Entry c). The latter reaction was effectively
achieved by replacing the Liþ cation by Kþ after subsequent addition of an excess of
tBuOK (Scheme 21, and Table 6, Entry d). This strategy is very convenient, since, after
acid hydrolysis, it led not only to the desired compounds 11 and 12 in the same pot but
also allowed their easy separation by chromatography on SiO2 owing the difference of
the functional groups present on each of the two products.
The formation of the bicyclo[3.1.0] compound 12 when reacting the same mixture
with a single equivalent of LiTMP (20% at ꢀ 258; 66% at 208; Scheme 21; Table 6,
Entries a and b) suggests that each of the two b-chloro-alcoholate intermediates, 27C’ l
’’
and 27 , does not cyclize to the epoxide 11 and generate the enolate 28’, if not directly
Cl
formed, that efficiently cyclize already at ꢀ 258. Therefore, the behavior of LiTMP is
different towards 21’Cl than towards 21B’ r (compare Scheme 21, Table 6, Entries a and b,
to Scheme 18, Table 5, Entries a and b).
We outline in Scheme 22 the whole process that allows the synthesis of cis-
chrysanthemic acid (2cis) from the epoxide 11 via the intermediate formation of
’’
’’
chlorohydrins 21’Cl and 21 , and recycling of the unwanted 21 regioisomer [2b].
Cl
Cl

2136
Helvetica Chimica Acta – Vol. 95 (2012)
’’
Scheme 22. Synthesis of 2cis Involving a Mixture of 21’Cl and 21
Cl
t
i) 5 equiv. BeCl2, CH2Cl2, 208, 100 h. ii) a) 2 equiv. LiTMP; b) 2 equiv. BuOK, THF, ꢀ 258, 1.5 h. iii)
Separation by chromatography on SiO2. iv) 1.2 equiv. MsCl, 1.5 equiv. Et3N, CH2Cl2, ꢀ 108, 0.75 h. v)
t
6 equiv. BuOK/3 equiv. H2O, DMSO, 208, 0.4 h.
Finally, we could transform the 58 :42 regioisomer mixture 20’/20’’ (Scheme 10) to
the bicyclic b-hydroxy ketone 12 (90% from 20’; Scheme 23, a) and the epoxide 11
(83% from 20’’, Scheme 23,a), respectively, by using Bu4NF (2.2 equiv., THF, ꢀ 258,
then acidic treatment).
Scheme 23. Selective Cyclization of a Mixture of 20’ and 20’’
a) 1. 2.2 equiv. Bu4NF, THF, ꢀ 258, 0.8 h. 2. H3Oþ. b) 1. 1 equiv. Bu4NF, THF, ꢀ 258, 0.8 h. 2. H3Oþ.
To gain more insight into the intimate mechanism of such process, we reacted the
mixture 20’/20’’ with a single equiv. of Bu4NF (Scheme 23,b) and observed the

Helvetica Chimica Acta – Vol. 95 (2012)
2137
formation, after acid hydrolysis, of the bicyclic 12 in very good yield (90% from 20’) and
21’I’ (67% from 20’’). This fully supports the hypothesis that the first equiv. of Bu4NF
selectively reacts on the silylenolate moieties of 20’ and 20’’ leaving their silylether
moieties untouched that play the role of protecting groups favoring the exclusive
carbocyclization 31 (Scheme 23,b). It is interesting to notice that, in the presence of the
second equiv. of Bu4NF, carbocyclization of 31 occured faster than deprotection of its
silylether moiety (Scheme 23,a) [2b].
The second route to cis-chrysanthemic acid (2cis) that we have devised offers the
advantage over the former to produce cis-chrysanthemic acid (2cis) from both
regiosiomers of 21. It involves, as outlined in Scheme 24: i) activation of the
regioisomer mixture 21’/21’’ to afford the mixture of cyclohexanones 32 bearing two
leaving groups at g,d-positions, ii) metallation of 32 favoring the annelation reaction
leading to the bicyclo[3.1.0] cyclohexanones 4, and iii) nucleophilic attack of a HOꢀ
anion on the CO group of the ketones, followed by fragmentation of the intermediate
to produce cis-chrysanthemic acid (2cis) after acidic workup, as we already reported
(Schemes 1, 19, and 22).
Scheme 24. Synthesis of 2cis from Both Regioisomers of 21
The transformation of the mixtures of cyclohexanones 32 to chrysanthemic acid 2
could be achieved most conveniently by using:
1) A single reagent acting sequentially as a base to achieve, in a single pot, the
selective carbocyclization of each of the regioisomers of 32 leading to 4X and 4OMs
,
respectively, then as a nuleophile to affect the selective fragmentation of each of them
to generate 2. The choice is strictly limited to anhydrous KOH (APH) that is the only
reagent able to transform the ꢂobligedꢃ intermediate 4OMs to 2. APH, however, has been
rarely used as a base [33], and most of the reports emphasize its exceptional
nucleophilicity [4a].
2) Two distinct reagents that require the selection among various bases that are the
most susceptible to achieve the selective carbocyclization of 32 to the mixture 4X/4OMs
.
Two options still remain to achieve the synthesis of 2 from the crude mixture 4X/4OMs a)
without separation in a single pot, the transformations again requires to use anhydrous
KOH (APH) at the last stage (see above) or b) after separation of 4X from 4OMs that

2138
Helvetica Chimica Acta – Vol. 95 (2012)
can only be achieved in distinct pots. It should be easily achieved by chromatography
owing the different polarity of 4X and 4OMs. In such case, still APH is required for the
transformation of 4OMs to 2, but either WPH or APH can achieve the transformation of
4X to 2.
The latter route, although slightly longer, offers the advantage to synthesize
scalemic 2cis from 4X and 4OMs when those possess ꢂopposite chiralityꢃ as it will be
disclosed in the last part of the article.
Reaction of the mixture of regioisomeric cyclohexanones 21 with MsCl in the
presence of DMAP (2 equiv. MsCl, 2 equiv. Et3N, 0.1 equiv. DMAP, CH2Cl2, ꢀ 10 to
’’
208, 6 h) led to a mixture of cyclohexanones 32 in 80 – 86% yield. Reaction of 32’Br/32
Br
(55 :45) with KHMDS provided a mixture of 4OMs/4Br (55 :45) in high overall yield
(83%) that has been easily separated, by chromatography on SiO2 (4Br: Rf (pentane/
Et2O 95 :5) 0.52 and 4OMs: Rf (pentane/Et2O 60 :40) 0.19). Each of them was
transformed to cis-chrysanthemic acid (2cis) using WPH or APH according to
procedures described above. Each of the regioisomers 32’ and 32’’ also separately
reacted with an excess of commercial KOH in wet DMSO (DMSO/H2O 4 :1; WPH).
The results are compiled in Scheme 25 and Table 7 [8] and commented below.
Scheme 25. Reactivity of 2,2-Dimethylcyclohexanones 32X towards Aq. KOH in DMSO (WPH)
Table 7. Reactivity of 2,2-Dimethylcyclohexanones 32X towards Aq. KOH in DMSO (WPH)
Entry
32X
X1
X2
T [8]
t [h]
4
Yield of
Yield of
Yield of
10/2cis
4 [%]
2cis [%]
10 [%]
Ratio
a
b
c
d
e
f
g
h
i
32’Cl
32’Br
MsO
MsO
MsO
Cl
Cl
Br
Br
I
I
Cl
Br
I
MsO
MsO
MsO
MsO
MsO
MsO
70
70
70
20
70
20
70
20
70
0.6 h
0.6 h
0.5 h
1 h
6 h
8 h
1 h
3 h
1 h
4OMs
4OMs
4OMs
4Cl
71
65
64
76
6
32’I
’’
32
32
32
32
Cl
’’
4Cl
57
7
26
4
Cl
’’
61
53
67
66
90 : 10
67: 33
95 : 5
Br
’’
Br
32’I’
32’I’
7
90 : 10
The reaction involving 4-halogeno-3-(mesyloxy)-2,2,5,5-tetramethylcyclohexa-
nones 32’X was best achieved at 708 for a short time (0.6 h) and furnished 3,3,6,6-
tetramethyl-4-oxobicyclo[3.1.0]hex-2-yl methanesulfonate (4OMs) in reasonably good
yields in all the cases (Scheme 25, and Table 7, Entries a – c). Performing the reaction of
32’X with WPH for a much longer time led to the disappearance of 4OMs and formation,
as expected, of a polymeric material [10].

Helvetica Chimica Acta – Vol. 95 (2012)
2139
WPH could also achieve the carbocyclization from the regioisomeric 3-chloro-4-
’’
(mesyloxy)-2,2,5,5-tetramethylcyclohexanone (32 ) to afford, at low temperature, 4-
Cl
chloro-3,3,6,6-tetramethylbicyclo[3.1.0]hexan-2-one (4Cl) chemoselectively and in high
yield (208, Scheme 25; and Table 7, Entry d), and cis-chrysanthemic acid (2cis), if the
reaction is carried out at higher temperature (708 instead of 208; Scheme 25, Table 7,
Entry e).
The reaction took however, another course, when performed on the related
’’
bromide 32 and iodide 32’I’, since it delivered, after acid hydrolysis, the unsaturated
Br
lactone 10 beside some cis-chrysanthemic acid (2cis; Table 7, Entries f – i). Apparently,
KOH in aqueous DMSO (WPH) no longer acts as a base towards 32’B’ r and 32I’’ but as a
nucleophile as it has been already observed for the dibromide 5Br (Schemes 3 and 4). It
is interesting to notice that a higher percentage of cis-chrysanthemic acid (2cis) is
’’
observed if the transformation is performed at higher temperature on 32 , but that this
Br
is no longer the case for the related iodinated derivative 32’I’.
To achieve the desired transformations of 32’’ to cis-chrysanthemic acid (2cis), we
performed the reaction with APH that proved so successful to produce the vinyl-
cyclopropane derivative 7cis instead of the vinyl-lactone 6 from the related 5Br
(Scheme 3). We were delighted to find that, as we had expected, APH in THF
allowed, under very mild conditions (208), efficient access to chrysanthemic acid 2,
from the whole series of cyclohexanones 32’’, even those that proved reluctant to do so
using WPH instead (Scheme 26 and Table 8, Entries a – c, compare to Scheme 25,
Table 7), and, in only one case, a few percent of vinyl-lactone was also produced
(Scheme 26 and Table 8, Entry c). The reactions were so fast that isolation of the
’’
bicyclic intermediates 4X was not feasible except in the case of the chloride 32 which
Cl
delivered small amounts of the bicyclic derivative 4Cl, besides cis-chrysanthemic acid
2acis (Table 8, Entry a).
Scheme 26. Reactivity of 2,2-Dimethylcyclohexanones 32 towards tBuOK/H2O (APH)
Performing the reaction in DMSO instead of THF did not offer advantages, since it
was slower and gave lower yield of chrysanthemic acid, especially from the iodide 32I.
Surprisingly, however, in the case of 32’B’ r and 32I’’, chrysanthemic acid was obtained as a
cis/trans diastereoisomer mixture.
We could not find the origin of the trans-chrysanthemic acid (2trans) formed in this
process. We have, however, secured, by independent reactions, that ꢂAHPꢃ in DMSO is

2140
Helvetica Chimica Acta – Vol. 95 (2012)
Table 8. Reactivity of 2,2-Dimethylcyclohexanones 32 towards tBuOK-H2O (APH)
Entry 32X X1
X2
Time [h] Solvent Yield of
Yield of Yield of Yield of 1 [%] 1/10
10 [%] (cis/trans) Ratio
1 þ 4 þ 10 [%] 4 [%]
’’
a
b
c
d
e
f
32 Cl
MsO 1.5
MsO 0.6
MsO 1
MsO 2
MsO 2
MsO 2
THF
THF
THF
90
89
93
18
0
0
0
0
0
0
0
0
0
0
0
0
0
4
0
5
5
0
0
0
0
0
0
72 (100 : 0)
89 (100 : 0)
89 (100 : 0)
75 (100 : 0)
76 (83 : 17)
36 (30 : 70)
38 (100 : 0)
48 (100 : 0)
10 (100 : 0)
67 (100 : 0)
80 (100 : 0)
30 (100 : 0)
100 : 0
100 : 0
96 : 4
100 : 0
94 : 6
88 : 12
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
100 : 0
Cl
’’
32 Br
Br
32’I’
I
’’
32 Cl
DMSO 75
DMSO 81
DMSO 41
Cl
’’
32 Br
Br
32’I’
I
g
h
i
j
k
l
32’Cl MsO Cl
32’Br MsO Br
32’I MsO I
32’Cl MsO Cl
32’Br MsO Br
32’I MsO I
1.5
1.6
2
THF
THF
THF
38
48
10
2.5
0.75
2
DMSO 67
DMSO 80
DMSO 30
unable to i) epimerize potassium cis- to potassium trans-chrysanthemate or to ii)
transform the lactone 10 to 2trans [8].
Another hypothesis, which involves a completely different mechanism, is tenta-
tively presented in Scheme 27. It implies metallation of the intermediate 33OMs, resulting
from the attack of ꢂAPHꢃ on the CO group of 32’X’ , followed by cyclization of 34OMs. We
have not been able, however, to experimentally support this proposal [8].
Scheme 27. Hypothesis for the Formation of 2trans
We also carried out the reaction of 32’ with APH to get the whole reactivity pattern
of each of the two partner in order to select the best conditions to perform in fine the
tandem carbocyclization/HallerꢀBauer reaction/Grob fragmentation reaction on the
mixture 32’’/32’.
We were delighted to find that chrysanthemic acid 2cis was chemoselectively and
stereoselectively formed from all 32’ derivatives, whatever the solvent used (Scheme 26,
and Table 8, Entries g – l). We were, however, surprised to observe that, except in the
case of 32’Br (Table 8, Entry k), the yields in 2, were lower than those obtained from
their regioisomers 32’’ under similar conditions (Table 8, compare Entries g – l to a – f).
Surprisingly better yields were generated when the reaction of 32’ with APH was

Helvetica Chimica Acta – Vol. 95 (2012)
2141
carried out in DMSO instead of THF (Table 8, compare Entries j – l to Entries g – i)
and, therefore, the behavior of 32’ towards APH markedly differs from that of their
regioisomers 32’’ (Table 8, compare Entries j – l to Entries d – f and to Entries a – c).
Accordingly, the reaction of the epoxy-cyclohexanone 11 sequentially with TiBr4,
H2O and MsCl led to the mixture of the two isomeric bromo-methanesulfonates
(57:43) in high yield (80%), which was transformed in a single pot to cis-chrysanthemic
acid (2cis) on reaction with APH in THF (70% yield, Scheme 28).
’’
Scheme 28. Synthesis of 2cis Involving a Mixture of 21’Br and 21
Br
i) 0.5 equiv. TiBr4, CH2Cl2, 208, 2 h, then aq. NaHCO3. ii) 2 equiv. MsCl, 2 equiv. Et3N, 0.1 equiv. 4-
(dimethylamino)pyridine (DMAP), ꢀ 10 to 208, 6 h. iii) 32/tBuOK/H2O, 1:7.6 :2.3, THF, 208, 2 h.
We were interested to rationalize the difference between WPH and APH to act as
nucleophile or base toward 32. The formal mechanism outlined in Schemes 2 and 4, to
be more accurate, should take into account stereochemical features. Those are
presented in Schemes 29 and 30. Thus, attack of HOꢀ anion can take place at the H-
atom at C(6) or at the CO C-atom of 32.
We assume that:
1) Metallation of 32 is best achieved at the more acidic HaxꢀC(6) and should
provide, under kinetic control, the corresponding enolate Ba by constant overlap of the
orbitals involved (Scheme 29,a) [34]. Metallation of Heq, however, the less acidic of the
two H-atoms at C(6), achieves the antiperiplanar arrangement of atoms and bonds
suitable for the cyclization leading to cyclopropane ring present in cis-chrysanthemic
acid (2cis; Scheme 29, Entry) and, therefore, successful cyclization is expected to involve
an equilibrium [8].
2) Addition of the HOꢀ ion onto the CO C-atom [35] from the top-face is expected
to generate the intermediate 35a, whose alkoxide ion is adequately positioned to favor
through an antiperiplanar conformation [36], the fragmentation reaction leading to the
open-chain intermediate 33X, precursor of the vinyl lactone 10 (Scheme 30,a). This
approach could be hampered with a bulky Me group (1,3-diaxial interaction as in 3Br
(Scheme 2) compared to 5Br (Scheme 4)). Addition from the bottom-face, however,
provides instead 35e, whose fragmentation is expected to be much less favored
(Scheme 30,b) [8].

2142
Helvetica Chimica Acta – Vol. 95 (2012)
Scheme 29. Rationalization for the Formation of 2cis
Scheme 30. Rationalization for the Formation of 10
APH apparently reacts faster in the process outlined in Scheme 29,a, than WPH. It
favors all the time, the cis-vinyl-cyclopropanecarboxylic acid 2cis, even when WPH
instead produces the vinyl lactone 10, probably because the ꢂmore reactiveꢃ APH is
more able than WPH to react via the process described in Scheme 29,b, on Heq of Ae,
leading to 4X, the precursor of cis-chrysanthemic acid (2cis; Scheme 29,b).
To have a better insight on the more suitable species required for successful
synthesis of cis-chrysanthemic acid (2cis), we carried out the reaction of 3-bromo-4-
’’
(mesyloxy)cyclohexanone 32 with KOH generated by i) dehydration, on heating, of
Br
powdered commercial KOH, or ii) reacting stoichiometric amounts of H2O on KH

Helvetica Chimica Acta – Vol. 95 (2012)
2143
t
(1 equiv. KH, 1 equiv. H2O, THF, 208) [37]. These reagents lack BuOK as well as
tBuOH.
We found that both reactions carried out in THF at 208, were slower (18 h and 66 h,
resp.) than that involving APH (0.5 h). Using dried KOH, the b-bromo bicyclo[3.1.0]
cyclohexanone 4Br (70%) and the vinyl-lactone 10 (30%) were formed, whereas using
KOH from KH, 4Br (19%) and cis-chrysanthemic acid (2cis; 81%) were obtained after
acid hydrolysis.
ꢂDried KOHꢃ in THF behaves as KOH in aqueous DMSO, whereas KOH from KH
in THF behaves as AHP in chemoselectivity but not in reactivity. AHP, therefore,
possesses an exceptional reactivity, which differentiates it from the two ꢂanhydrous
KOHꢃ and wet KOH reagents tested in this study [8].
We took the opportunity of this work to slightly modify the routes disclosed in order
to achieve the enantioselective synthesis of (1R)-cis-chrysanthemic acid ((1R)-2cis), or
its methyl ester, (1R)-2cis-Me, respectively, which, after adequate replacement of the
isopropylidene moiety by the dibromovinyl, allows the synthesis of deltamethrin (36),
the most powerful commercial insecticide for outdoor uses (Scheme 31, Path a). We
also synthesized (1R)-trans-chrysanthemic acid ((1R)-2trans), the constituent of the
naturally occurring pyrethrin I or the related commercially available insecticide (S)-
bioallethrin (37), suitable for indoor use. The latter transformation involves
epimerization at C(1) of (1S)-cis-chrysanthemic acid (1S)-2cis, or better of one of its
t
ester, especially the butyl ester (Scheme 31, Path b). The individual steps used for
those transformations are outlined in Scheme 31.
Thus, the most straightforward route to those compounds would have been the
enantioselective additions on the C¼C bond of 2,2,5,5-tetramethylcyclohex-3-enone 1
which would have led to the corresponding scalemic epoxide 11*, dihalides 3*, or
pseudo-halides 32*. We have been, however, unable to achieve it. We could, however,
perform the desired transformations in a three-step sequence which involves i)
enantioselective reduction of the CO group of 1 leading to the scalemic (1S)-2,2,5,5-
tetramethylcyclohex-3-enol (S)-23, ii) diastereoselective additions across the C¼C
bond of the resulting homoallyl alcohol, iii) regioselective oxidation of the resulting
saturated alcohol to the functionalized ketones 11*, 3*, and 32*.
The enantioselective reduction of 2,2,5,5-tetramethylcyclohex-3-en-1-one (1) to
scalemic (1S)-2,2,5,5-tetramethylcyclohex-3-enol (S)-23, which is the key step in this
transformation, was conveniently achieved in high yield (85%) and very high
stereocontrol (ee > 97%) by using (ꢀ)-b-chlorodiisopinocampheylborane (Ipc2BCl)
[38] (i) 1.05 equiv., neat, 258, 48 h, ii) 2.2 equiv. diethanolamine, Et2O, 258, 85% yield,
ee > 97%, Scheme 32).
Vicinal dibromination of the homoallyl alcohol (S)-23 proved to be highly
stereoselective if carried out at very low temperature (1 equiv. Br2, CH2Cl2, ꢀ 958,
0.33 h) and led to (R,R)-38 in high yield and with very high stereocontrol (95% yield;
de 97%; Scheme 32). Its structure has been secured by X-ray crystallography of the
corresponding acetate (R,R)-39 (Scheme 33) [2d].
The selective formation of (R,R)-38 over (S,S)-38 could be rationalized by
assuming the attack of the Brꢀ ion at C(4) of 40a or at C(3) of its diastereoisomer 40b
to reach a chair (Scheme 34,a and b) rather than a twisted boat transition state
(Scheme 34,c and d) [39].

2144
Helvetica Chimica Acta – Vol. 95 (2012)
Scheme 31. Synthesis of Deltamethrine (36) and S-Bioallethrin (37) from cis-Chrysanthemic Acid (2cis
)
i) a) O3, MeOH, ꢀ 808; b) Me2S, ꢀ 408 to 208; c) aq. AcOH, 808, 0.25 h. ii) CBr4, PPh3, CH2Cl2, 258,
0.15 h. iii) tBuOK, THF, 208, 2 h. iv) KOH, MeOH, 658, 24 h, then H3Oþ.
We have used these results to produce (1R)-cis-dibromovinyl-cyclopropanecarbox-
ylic acid 41H in more than 98% ee conducting known reactions that we already
disclosed (Scheme 32).
The dibromination of (S)-23 involves i) attack of the bromonium ion on its C¼C
bond syn or anti to the OH group, and leading to the intermediate formation of
bromoniums 40a or 40b, ii) ring opening of 40 which can take place either at the g- or d-
positions relative to the C-atom bearing the OH group as already discussed
(Scheme 34). To distinguish amongst those possibilities, we have carried out a model
study involving the bromohydroxylation of (S)-42, the acetate of (S)-23 (Scheme 35).
Best results were obtained by using NBS in aqueous acetone [40]. The reaction was
slower than that involving the same reagent in aqueous tBuOH [41] but provided better

Helvetica Chimica Acta – Vol. 95 (2012)
2145
Scheme 32. Enantioselective Synthesis of cis-Chrysanthemic Acid (1R)-2cis Involving (1S)-2,2,5,5-
Tetramethylcyclohex-3-en-1-ol ((S)-23). Diethanolamine ¼ 2,2’-Iminodiethanol, Ipc2BCl ¼ chloro-
bis(2,6,6-trimethylbicyclo[3.1.1]heptan-3-yl)borane; PDC ¼ pyridinium dichromate.
Scheme 33. Synthesis of Acetate (R,R)-39
Scheme 34. Hypothesis for the Selective Formation of (R,R)-38 over (S,S)-38

2146
Helvetica Chimica Acta – Vol. 95 (2012)
Scheme 35. Bromohydroxylation of (S)-42. NBS ¼ N-Bramosuccinimide.
Table 9. Bromohydroxylation of (S)-42
Entry x equiv. of NBS Conditions
T [h]
Yield of
Yield of Yield of
43’ þ 43’’ [%] 42 [%]
39 [%]
a
b
c
1.1
2
4
tBuOH/H2O 1 : 2, 208
0.75 57
0
34
24
0
6
27
19
DMSO þ 2 equiv. H2O, 208 168
26
34
86
DME/H2O 4 : 1, 208
168
168
d
1.5
Acetone/H2O 1 : 4
ca. 2
yield of 43 than when carried out in aqueous DMSO [42] or DME [43] (Scheme 35,
Table 9).
The 1H-NMR studies coupled with the results disclosed in Schemes 37 and 38, led us
to conclude that i) the resulting crude compound is in fact a mixture (R,R)-43’’/(R,R)-
43’/(S,S)-43’’/(S,S)-43’ in a 55 :32 :10 :3 ratio, and the formation of bromonium 44a,
which results from the attack of the bromonium ion on the face opposite to that where
the Ac group lies, is slightly favored (58%) over the other (42%; Scheme 36).
On the basis of these results, we devised new routes to scalemic cis-chrysanthemic
acid (2cis) from the mixture 43’/43’’ (43*) (Schemes 37 and 38).
Thus reaction of mesylchloride with 43’/43’’ led to the b-bromomesylates 45’/45’’
(45*) whose deacetylation, followed by oxidation with PDC, provided a mixture of
’
’’
*
regioisomeric ketones, 32Br/32Br. Reaction of 32Br mixture with APH in THF, followed
by acid hydrolysis, provided in one pot (1R)-cis-chrysanthemic acid ((1R)-2cis) in 64%
yield and 76% ee (Scheme 37).
*
Alternatively, reaction of the same mixture of 32Br with KHMDS afforded a
*
*
mixture of the bicyclic derivatives, 4OMs/(S)-4Br in 78% yield. Those have been easily
separated by column chromatography on SiO2 ((S)-4OMs: 36%, ee 82%; (S)-4Br: 64%,
ee 74%; Scheme 38). These results support those disclosed in Scheme 36, indicating that
asymmetric induction in the reaction of (S)-42 with NBS in aqueous acetone that leads
to the bromohydrin 43’ is higher than the one which instead produces in the same pot its
regioisomer 43’’ (Scheme 36).
We have then achieved the transformation of (S)-4OMs and (S)-4Br to cis-
chrysanthemic acid by using APH in DMSO (Scheme 6) and THF (Scheme 1, and
Table 1, Entry b), respectively (Scheme 38).
We next synthesized enantioselectively (1S)-cis-chrysanthemic acid ((1S)-2cis) from
scalemic homoallyl alcohol (S)-23 using the series of reactions outlined in Schemes 13
and 19, which involve i) stereoselective epoxidation of (S)-23 leading to 24*
(Scheme 13), ii) regioselective ring opening of the epoxide 24* leading exclusively to

Helvetica Chimica Acta – Vol. 95 (2012)
2147
Scheme 36. Rationalization for the Bromohydroxylation of (S)-42
*
Scheme 37. Synthesis of scalemic cis-Chrysanthemic Acid (2cis) from the Mixture of the Bromohydrines
43* Involving ꢁOne-Potꢂ CyclizationꢀFragmentation
i) 2 equiv. MsCl, 2 equiv. Et3N, 0.2 equiv. DMAP, CH2Cl2, ꢀ 10 to 208, 55 h. ii) 5 equiv. K2CO3, THF/
MeOH 1:1, 0 to 208, 5.5 h. iii) 1.4 equiv. PDC, mol. sieves (MS; 4 ꢄ), CH2Cl2, 0 to 208, 4 h. iv) 7.6 equiv.
tBuOK, 2.3 equiv. H2O, THF, 208, 2.5 h, then H3Oþ.
the bromohydrin 25B’*r (Scheme 13), iii) regioselective oxidation to 21B’*r (Scheme 13), iv)
carbocyclization producing the bicyclic b-hydroxy-cylopentanone 12* (Scheme 19),
and v) cascade HallerꢀBauer reaction-Grob fragmentation of the corresponding
*
methanesulfonate 4OMs leading finally to (1S)-cis-chrysanthemic acid (1S)-2cis in very

2148
Helvetica Chimica Acta – Vol. 95 (2012)
*
Scheme 38. Synthesis of Scalemic cis-Chrysanthemic Acid (2cis) from the Mixture of the Bromohydrines
43* Involving Sequential CyclizationꢀFragmentation. KHMDS ¼ Potassium hexamethyldisilazide.
good yield (90%), and extremely high de (100%) and ee (93%) values (Scheme 19)
[2a].
A variation of Scheme 13 implies the oxidation of the scalemic homoallyl alcohol
(S,R)-24 to the corresponding b-oxo epoxide (S,R)-11 (90% overall yield, 94% ee;
Scheme 39) which, on reaction with TiBr4 as disclosed in the racemic version
’
’’
*
(Scheme 28), produced a mixture of scalemic bromohydrins 21Br/21Br (21Br), resulting
*
from epoxide ring opening (Scheme 39). Mesylation of mixture of 21Br, followed by
’’
treatment of the resulting methanesulfonate 32’Br/32 (32*) sequentially with APH in
Br
THF and HCl led to (1S)-cis-chrysanthemic acid ((1S)-2cis) in good overall yield but
very poor enantioslectivity (9%, Scheme 39) [2c] due to 10% excess of the
methanesulfonate (R)-4OMs over the bromide (S)-4Br that are pseudo-enantiomers.
A slight change in the above strategy that involves i) isolation of the mixture (S)-
4Br/(R)-4OMs using KHMDS instead of APH, ii) easy separation by column chroma-
tography on SiO2, and iii) fragmentation of each bicyclic derivative using APH ((S)-4Br
in THF; Scheme 40; (R)-4OMs in DMSO; Scheme 40) [2c] furnished, in high yield and
very high ee, (1R)-cis-chrysanthemic acid ((1R)-2cis), precursor of deltamethrin
(Scheme 31) and (1S)-cis-chrysanthemic acid ((1S)-2cis) precursor of (S)-bioallethrin
(Scheme 31), respectively.
*
A related transformation carried out on the scalemic mixture 21Cl obtained by
epoxide ring opening of scalemic (1S,6R)-2,2,5,5-tetramethyl-7-oxabicyclo[4.1.0]hep-
tan-3-one (S,R)-11 by BeCl2 (Scheme 11, and Table 3, Entry g) led to the mixture (S,S)-
21C’ l/(R,R)-21C’’ l (80 :20) that then provided, as outlined in Scheme 41, a mixture of (1S)-
cis-chrysanthemic acid ((1S)-2cis; ee: 94%) and (1S,6R)-2,2,5,5-tetramethyl-7-oxabicy-
clo[4.1.0]heptan-3-one ((S,R)-11; ee: 94%) [2b].
Acid/base treatment allowed the separation of (1S)-2cis, the precursor of S-
bioallethrin (Scheme 31, Entry b) from (S,R)-11, which has been then recycled
(Scheme 41).
Conclusions. – In this article, we have described a series of reactions that allow the
enantioselective transformation of the prochiral 2,2,5,5-tetramethylcyclohex-3-enone 1

Helvetica Chimica Acta – Vol. 95 (2012)
2149
*
Scheme 39. Synthesis of Scalemic cis-Chrysanthemic Acid (2cis) from the Mixture of the Bromohydrines
21* Involving ꢁOne-Potꢂ CyclizationꢀFragmentation
i) 1.5 equiv. tBuOOH, 0.015 equiv. Mo(CO)6, C6H6, 808, 2 h. ii) 1.5 equiv. PDC, MS (4 ꢄ) CH2Cl2, 0 to
208, 18 h. iii) 0.5 equiv. TiBr4, CH2Cl2, 208, 2 h. iv) 2 equiv. MsCl, 2 equiv. Et3N, 0.1 equiv. DMAP,
CH2Cl2, ꢀ 10 to 208, 18 h. v) 7.6 equiv. tBuOK, 2.3 equiv. H2O, THF, 208, 2 h. vi) H3Oþ.
*
Scheme 40. Synthesis of Both Enantiomers of Scalemic cis-Chrysanthemic Acid (2cis) from the Mixture of
the Bromohydrines 21* Involving Sequential CyclizationꢀFragmentation
i) 1.2 equiv. KMHDS, THF, 08, 1 h. ii) 7.6 equiv. tBuOK, 2.3 equiv. H2O, THF, 208, 0.6 h, then H3Oþ. iii)
7.6 equiv. tBuOK, 2.3 equiv. H2O, DMSO, 208, 0.6 h, then H3Oþ.

2150
Helvetica Chimica Acta – Vol. 95 (2012)
*
Scheme 41. Synthesis of Scalemic cis-Chrysanthemic Acid (2cis) from the Mixture of the Chlorohydrines
21*
t
i) 5 equiv. BeCl2, CH2Cl2, 208, 100 h. ii) a) 2 equiv. LiTMP; b) 2 equiv. BuOK, THF, ꢀ 258, 1.5 h. iii)
Separation by chromatography on SiO2. iv) 1.1 equiv. MsCl, 1.5 equiv. Et3N, CH2Cl2, ꢀ 108, 2 h. v)
6 equiv. tBuOK, 3 equiv. H2O, DMSO, 208, 1 h, then H3Oþ.
to (1S)-2,2,5,5-tetramethylcyclohex-3-enol ((S)-23) using (ꢀ)-b-chlorodiisopinocam-
phenylborane [38] then to i) (1R)-cis-chrysanthemic acid ((1R)-2cis), the precursor of
deltamethrin 38 (Scheme 31, a), via diastereoselective dibromination of (S)-23
(Scheme 32) [2d] or ii) (1S)-cis-chrysanthemic acid ((1S)-2cis), the precursor of (S)-
bioalletrin 40 (Scheme 31,b), by a) diastereoselective epoxidation of (S)-23, followed
by epoxide ring opening on the epoxyalcohol (S,R)-24 to produce, after regioselective
oxidation, the ketone (S,S)-21’Br (scalemic version of Schemes 13 and 19) [2a] or b)
epoxide ring opening of the related epoxy ketones (S,R)-11 by BeCl2 to give mainly
(S,S)-21’Cl (Scalemic version of Scheme 22) [2b].
It is interesting to notice that the strategy described above is very flexible, since
starting from the same 2,2,5,5-tetramethylcyclohex-3-enone 1, (1S)-cis-chrysanthemic
acid ((1S)-2cis) can be obtained using the reactions disclosed in i in the above
paragraph), and its regioisomer (1R)-cis-chrysanthemic acid ((1R)-2cis) can be
prepared by using the reactions disclosed in ii in the above paragraph by simply using
the dextrogyre (þ)-B-chlorodiisopinocampheylborane [38] instead of its enantiomer.
Both are commercially available for about the same cost.

Helvetica Chimica Acta – Vol. 95 (2012)
2151
Experimental Part
General. For the procedures of syntheses presented in Schemes 32 and 33, see reference [2d].
Reactions requiring an inert atmosphere were carried out in two-neck flask fitted with a magnetic stirrer,
a rubber cap and an Ar-inflated balloon. Unless otherwise mentioned, solvents and commercial reagents
were used without further purification. Et2O and THF were distilled from Na/benzophenone/cetyl, C6H6
was refluxed over Na, and CH2Cl2 was refluxed over P2O5. Column chromatography (CC): 30 g of silica
gel (SiO2; Merck silica 7734) per g crude sample. M.p.: Bꢀchi B-545 in open capillary tubes; uncorrected.
Specific rotations [a]D: Perkin-Elmer 241 polarimeter; at 208; with a Na lamp (589 nm) and a sample tube
1
of 10-cm length. H- and 13C-NMR: Jeol JNM EX-400 at 400.0 and 100.4 MHz, respectively; chemical
1
shifts refer to d(TMS) ¼ 0 for H and to d(CDCl3) ¼ 77.0 ppm for 13C. GC/MS: Hewlett-Packard 5890
series II gas chromatograph coupled to a Hewlett-Packard 5989B mass spectrometer, at 70 eV.
2,2,5,5-Tetramethylcyclohex-3-en-1-one (1). In a 250-ml round-bottom neck flask, TsNHNH2 (9.3 g,
50 mmol) was added to a stirred soln. of 2,2-dimethyldimedone (13; 8.4 g, 50 mmol) in abs. EtOH
(100 ml) at r.t. without any special care. The TsNHNH2 did not solubilize totally, and more precipitate
appeared during the reaction. After 7 d, the mixture was placed under reduced pressure to furnish a white
solid as crude product. In a 500-ml round-bottom two-neck flask fitted with a condenser under Ar, the
crude product was added to a soln. of ethylene glycol/sodium glycolate (from 250 ml of ethylene glycol
and 5.75 g of Na, 250 mmol) and was heated at 1808 until evolution of N2 had ceased (ca. 0.5 h). The
cooled soln. was poured onto 350 ml of ice and extracted with pentane (5 ꢂ 120 ml). The extracts were
combined, dried (MgSO4), filtered, and concentrated by rotary evaporation to give a brown liquid (the
product obtained was quite volatile). The crude product was purified by CC (pentane/Et2O 98 :2) to
furnish 1.44 g (50%) of slightly yellow, volatile liquid. Spectral properties were identical with those
described in [44].
3,4-Dibromo-2,2,5,5-tetramethylcyclohexanone (3Br) in CH2Cl2 at ꢀ 788. In a 25-ml round-bottom
two-neck flask under Ar, at ꢀ 788, a 1m soln. of Br2 in CH2Cl2 (1 ml, 1 mmol) was added to a stirred soln.
of 1 (152 mg, 1 mmol) in CH2Cl2 (5 ml). At the end of the addition, the mixture was evaporated under
reduce pressure, while it was allowed to warm to r.t. to furnish 311 mg (100%) of 3Br. White solid.
Spectral properties were identical with those described in [2d].
Compound 3Br in CCl4 at 08. In a 50-ml round-bottom two-neck flask under Ar, acetamide (30 mg,
0.5 mmol) was added to a stirred soln. of 1 (760 mg, 5 mmol) in CCl4 (20 ml). The soln. was cooled to 08
before adding dropwise a soln. of Br2 (800 mg, 5 mmol) diluted with CCl4 (5 ml). After 1 h at 08, aq. sat.
NaHCO3 (10 ml) was added, and the mixture extracted with Et2O (3 ꢂ 30 ml). The combined org.
extracts were washed with aq. sat. Na2S2O5 (10 ml), dried (MgSO4), filtered, and evaporated under
reduced pressure to furnish 1.53 g (98%) of 3Br. White solid. Spectral properties were identical with those
described in [2d].
3,4-Dichloro-2,2,5,5-tetramethylcyclohexanone (3Cl). In a 50-ml round-bottom two-neck flask under
Ar, acetamide (47 mg, 0.8 mmol) was added to a stirred soln. of 1 (1.22 g, 8 mmol) in CCl4 (8 ml). The
soln. was cooled to ꢀ 408 before adding dropwise a soln. of Cl2 (1.3m in CCl4, 6.15 ml, 8.0 mmol). At the
end of the addition, the reaction was quenched with aq. sat. NaHCO3 (25 ml), and the mixture was
allowed to warm to r.t. and extracted with CH2Cl2 (3 ꢂ 20 ml). The combined org. extracts were washed
with aq. sat. Na2S2O5 (10 ml), dried (MgSO4), filtered, and evaporated under reduced pressure to yield
1.70 g (95%) of 3Cl. White solid. M.p. 998. IR (KBr): 2987, 2937, 2879, 1708, 1462, 1391, 1370, 1306, 1263,
1
1248, 1196, 1123, 1080, 973, 949, 900, 874, 817, 784, 740. H-NMR (400 MHz, CDCl3): 4.22 (d, J ¼ 11.2,
1 H); 4.01 (d, J ¼ 11.2, 1 H); 2.74 (dd, J ¼ 14.4, 0.8, 1 H); 2.27 (d, J ¼ 14.4, 1 H); 1.27 (s, 3 H); 1.25 (s,
3 H); 1.24 (s, 3 H); 0.95 (s, 3 H). 13C-NMR (100 MHz, CDCl3): 209.0; 71.2; 70.8; 52.2; 50.0; 39.1; 29.9;
22.5; 21.7; 20.0.
4-exo-Bromo-3,3,6,6-tetramethylbicyclo[3.1.0]hexan-2-one (4Br) with LDA. In a 25-ml round-bottom
two-neck flask under Ar, at ꢀ 788, a 0.57m soln. of LDA in THF (1.75 ml, 1 mmol) was added dropwise to
a soln. of 3Br (312 mg, 1 mmol) in THF (3 ml). After 1 h at ꢀ 788, the mixture was hydrolyzed with 20 ml
of H2O and extracted with Et2O (3 ꢂ 20 ml). The combined org. extracts were washed with brine (2 ꢂ
10 ml), dried (MgSO4), filtered, and evaporated under reduced pressure to furnish 199 mg (86%) of
4Br. Colorless liquid. IR (film): 3043, 2969, 2931, 2871, 1731, 1462, 1380, 1360, 1280, 1236, 1195, 1109, 1028,

2152
Helvetica Chimica Acta – Vol. 95 (2012)
1
1008, 974, 870, 837, 785, 746. H-NMR (400 MHz, CDCl3): 4.01 (d, J ¼ 1.2, 1 H); 2.15 (dd, J ¼ 5.6, 1.6,
1 H); 1.98 (d, J ¼ 6.0, 1 H); 1.23 (s, 3 H); 1.16 (s, 3 H); 1.01 (s, 3 H); 1.00 (s, 3 H). 13C-NMR (100 MHz,
CDCl3): 212.3; 56.1; 54.4; 38.9; 38.0; 27.5; 26.8; 26.4; 19.0; 16.9. Due to the instability of the product,
further analyses were not performed.
t
t
Compound 4Br with BuOK. In a 250-ml round-bottom two-neck flask under Ar, at ꢀ 788, BuOK
(2.69 g, 24 mmol) was added in one portion to a stirred soln. of 3Br (6.24 g, 20 mmol) in dry THF
(200 ml), and the mixture was stirred while warming up to r.t. After 1 h, the reaction was quenched with
aq. sat. NH4Cl (20 ml), the org. layer was decanted and the aq. layer extracted with Et2O (3 ꢂ 50 ml). The
combined org. extracts were dried (MgSO4), filtered, and evaporated under reduced pressure. The crude
product was purified by CC (pentane/Et2O 95 :05) to furnish 4.34 g (94%) of 4Br. Colorless liquid.
Spectral properties were identical with those already reported.
Compound 4Br with MeOLi. In a 25-ml round-bottom two-neck flask equipped with a reflux
condenser under Ar, BuLi (1.6m in hexane, 3.75 ml, 6 mmol) was added to freshly dried MeOH (5 ml).
The mixture was then evaporated under reduce pressure in order to remove the solvents. Freshly dried
MeOH (1 ml) and 3Br (332 mg, 1 mmol) were then added successively, and the soln. was heated at 658
under Ar. After 48 h, the reaction was quenched with aq. sat. NH4Cl (15 ml), and the mixture was
extracted with Et2O (3 ꢂ 20 ml). The combined org. extracts were washed with brine (20 ml), dried
(MgSO4), filtered, and evaporated under reduced pressure to furnish 113 mg (49%) of 4Br. Colorless
liquid. Spectral properties were identical with those already reported.
4-exo-Chloro-3,3,6,6-tetramethylbicyclo[3.1.0]hexan-2-one (4Cl) with LDA. In a 25-ml round-bottom
two-neck flask under Ar, at ꢀ 788, a 0.57m soln. of LDA in THF (1.75 ml, 1 mmol) was added dropwise to
a soln. of 3Cl (223 mg, 1 mmol) in THF (3 ml). After 1 h at ꢀ 788, the mixture was hydrolyzed with 20 ml
of H2O and extracted with Et2O (3 ꢂ 20 ml). The combined org. extracts were washed with brine (2 ꢂ
10 ml), dried (MgSO4), filtered, and evaporated under reduced pressure to furnish 164 mg (88%) of
4Cl. Colorless liquid. IR (film): 3042, 2970, 2931, 2871, 1732, 1463, 1381, 1360, 1279, 1263, 1198, 1110, 1029,
1
974, 882, 842, 788, 712. H-NMR (400 MHz, CDCl3): 3.98 (s, 1 H); 1.98 (d, J ¼ 0.8, 2 H); 1.18 (s, 3 H);
1.16 (s, 3 H); 1.04 (s, 3 H); 1.02 (s, 3 H). 13C-NMR (100 MHz, CDCl3): 213.3; 62.8; 56.4; 38.8; 37.7; 26.9;
25.6; 24.4; 19.2; 17.0. Due to the instability of the product, further analyses were not performed.
t
t
Compound 4Cl with BuOK. In a 50-ml round-bottom two-neck flask under Ar, at ꢀ 788, BuOK
(404 mg, 3.6 mmol) was added in one portion to a stirred soln. of 3Cl (669 mg, 3 mmol) in dry THF
(30 ml), and the mixture was stirred while warming up to r.t. After 1 h, the reaction was quenched with
aq. sat. NH4Cl (10 ml), the org. layer was decanted, and the aq. layer was extracted with Et2O (3 ꢂ
20 ml). The combined org. extracts were dried (MgSO4), filtered, and evaporated under reduced
pressure. The crude product was purified by CC (pentane/Et2O 90 :10) to furnish 500 mg (89%) of 4Cl.
Colorless liquid. Spectral properties were identical with those already reported.
cis-Chrysanthemic Acid (2cis) from 4Br with WPH (General Procedure 1 (GP 1)). In a 25-ml round-
bottom two neck flask equipped with a reflux condenser under Ar, 4Br (116 mg, 0.5 mmol) was added to a
stirred soln. of KOH (168 mg, 3 mmol) in DMSO/H2O 4 :1 (2 ml). The mixture was then heated to 708
for 0.8 h. A orange-brown color appeared, and the reaction was monitored by TLC (pentane/Et2O
80 :20). The mixture was acidified to pH 2 with aq. HCl (10%; discoloration) and extracted with Et2O
(4 ꢂ 15 ml). The combined org. extracts were washed with H2O (2 ꢂ 5 ml), dried (MgSO4), filtered, and
evaporated under reduce pressure. The crude product was purified by CC (pentane/Et2O 80 :20) to
furnish 73 mg (87%) of 2cis. White solid. Spectral properties were identical with those described in [2d].
Compound 2cis from 4Br with APH in THF (General Procedure 2 (GP 2)). In a 25-ml round-bottom
two-neck flask under Ar, H2O (21 mg, 1.15 mmol) was added to a stirred soln. of freshly sublimed tBuOK
(426 mg, 3.8 mmol) in dry THF (4 ml), and the mixture was stirred at r.t. After 10 min, a soln. of 4Br
(116 mg, 0.5 mmol) in dry THF (2 ml) was added dropwise. A yellow color appeared, and the reaction
was monitored by TLC (pentane/Et2O 80 :20). After 0.5 h at r.t., ice (8 ml) was added, the mixture was
acidified to pH 2 with aq. HCl (10%; discoloration) and extracted with Et2O (4 ꢂ 15 ml). The combined
org. extracts were washed with H2O (2 ꢂ 5 ml), dried (MgSO4), filtered, and evaporated under reduce
pressure. The crude product was purified by CC (pentane/Et2O 80 :20) to furnish 79 mg (94%) of 2cis.
White solid. Spectral properties were identical with those described in [2d].