Diastereoselective SmI2-mediated 3-exo-trig cyclisation of δ-oxo-alkylidene-malonates to cyclopropanols

Article abstract of DOI:10.1055/s-2005-872683

In the presence of samarium diiodide and a proton donor (tert-butanol or phenol), δ-oxo-γ,γ-disubstituted-alkylidenemalonates readily cyclise according to the 3-exo-trig mode to give, depending on the exact nature of the substrate, trans and cis cycloprop

Full text of DOI:10.1055/s-2005-872683

2362
LETTER
Diastereoselective SmI2-Mediated 3-exo-trig Cyclisation of d-Oxo-Alkylidene-
malonates to Cyclopropanols
Cyclisation of
d
-O
i
xo-A
l
a
kylidenema
d
lonates toC
h
yclopropanols Zriba, Sophie Bezzenine-Lafollée,* François Guibé,* Marie-George Guillerez
Institut de Chimie Moléculaire et des Matériaux, Laboratoire de Catalyse Moléculaire, UMR CNRS-8075, Bât 420, Université Paris-Sud,
91405 Orsay, France
Fax +33(1)69155259; E-mail: fraguibe@icmo.u-psud.fr; E-mail: sbezzenine@icmo.u-psud.fr
Received 13 May 2005
R3
CO2R'
R3
Abstract: In the presence of samarium diiodide and a proton donor
(tert-butanol or phenol), d-oxo-g,g-disubstituted-alkylidenema-
lonates readily cyclise according to the 3-exo-trig mode to give, de-
pending on the exact nature of the substrate, trans and cis
cyclopropanols or lactones derived from the cis isomers. Total ste-
reoselectivity towards the formation of trans cyclopropanols is ob-
served with di-tert-butyl alkylidenemalonates when PhOH is used
as the protonating agent.
X
R1
SmI2 (2 equiv)
CO2R'
(1)
R1
t-BuOH (3–4 equiv)
R2
R2
THF, 0 °C to r.t.
1
2
X = Br, I
R1 = H, alkyl, aryl; R2, R3 = H, alkyl
HO
R1
Key words: samarium, radical reactions, cyclopropanol, cyclisa-
tions (3-exo-trig), alkylidenemalonates
CO2R'
R2
R2
O
SmI2 (2 equiv)
4
(2)
+
CO2R'
R1
t-BuOH (3–4 equiv)
We previously reported that d-iodo or d-bromo-a,b-un-
saturated esters of general formula 1 readily cyclise
(Scheme 1, Equation 1) according to the 3-exo-trig mode
to cyclopropane compounds 2 in the presence of two
equivalents of samarium diiodide and a proton source
(typically tert-butanol).1 A very similar reaction was
simultaneously reported by Lange and Merica.2 More
recently, we found that, under the same conditions, d-oxo-
g,g-disubstituted-a,b-unsaturated esters 3 also cyclise
leading to cyclopropanols trans-4 and/or lactones 5
derived from the cis cyclopropanols (Scheme 1, Equa-
tion 2).3
Recently and in a related field, T. Inokuchi4 has reported
that in the presence of samarium dichloride (in situ
formed from samarium iodide and lithium chloride) g-
acetyl-g-carbalkoxy-cyclohexenones lead in a transient
manner to cyclopropanols in the course of a tandem
cyclisation/retroaldolisation reaction. Gaunsäuer and co-
workers5 and Fernandez–Mateos and co-workers6 have
shown that in the presence of Cp2TiCl d,e-epoxy-a,b-un-
saturated carbonyl compounds cyclise to cyclopropane-
carbinols.
O
R2 R2
3
THF, 0 °C to r.t.
O
R1
R1 = H, alkyl, aryl; R2 = alkyl
R2
R2
5
Scheme 1
Lehnert’s procedure.7 Alkylidenemalonates were usually
obtained in moderate to good yields, except 7e for which
yield was low (ca 30%). For some unclear reason, this
procedure failed in the case of the condensation of 3-oxo-
2,2-dimethylbutyraldehyde (8, R1 = Me) with dimethyl
malonate. Compound 6a was finally obtained, albeit in a
poor 23% yield, from the corresponding di-tert-butyl
alkylidenemalonate 7a, by transesterification (anhyd
MeOH, Dowex 50 H+, 15 h, 75 °C in a tightly stoppered
vial).
O
O
O
H
CO2t-Bu
CO2Me
R1
R1
R1
O
In the present letter, we report on the extension of our
Me Me CO2t-Bu
cyclisation studies to d-oxo-alkylidenemalonates.
Me Me CO2Me
Me Me
6
7
8
Dimethyl and di-tert-butyl alkylidenemalonates 6 and 7
(Scheme 2) were investigated. They were conveniently
prepared by Knoevenagel condensation of b-ketoal-
dehydes 8 and, respectively, dimethyl or di-tert-butyl
malonates. The condensations were effected in the
presence of TiCl4 and pyridine in THF according to
a R1 = Me
b R1 = i-Pr
c R1 = c-Pr
d R1 = Ph
a R1 = Me d R1 = Ph
e R1 = 2-furyl b R1 = i-Pr e R1 = Bn
c R1 = c-Pr
Scheme 2
For each substrate 6 or 7, the cyclisation reactions were
run several (usually three) times with good reproducibili-
ty. Products were purified by column chromatography.
SYNLETT 2005, No. 15, pp 2362–2366
Advanced online publication: 07.09.2005
DOI: 10.1055/s-2005-872683; Art ID: G12605ST
© Georg Thieme Verlag Stuttgart · New York
1
9
.0
9
.2
0
0
5

LETTER
Cyclisation of d-Oxo-Alkylidenemalonates to Cyclopropanols
2363
The cyclisations of dimethyl alkylidenemalonates 6 in the respectively, the exo and endo forms. From them and the
presence of tert-butanol as the proton donor were found to Karplus equation,11 values of 0.1 Hz and 5.7 Hz could be
give the trans cyclopropanol adducts 9 and/or lactones 10 calculated for the corresponding coupling constants.
undoubtedly arising from the cis adduct8 (Table 1).
These observations and calculations are consistent with
those made by Barluenga and coworkers on other structur-
ally related cyclopropane g-lactones.12
A good or total diastereoselectivity in favour of the trans
adduct is observed with alkyl ketones 6a–c. With aryl ke-
tones, the stereoselectivity is poor, being still in favour of The cyclisations of di-tert-butyl alkylidenemalonates 7 in
the trans adduct for R1 = phenyl (6d) but in favour of the the presence of tert-butanol resulted in substantially better
cis adduct with R1 = furyl (6e). In most cases, overall overall yields. They lead to trans and/or cis cyclopropanol
yields are only fair but the fact that alkylidenemalonates adducts (Table 2).13 Unlike the corresponding methyl es-
do cyclise is nevertheless remarkable. Indeed in a (unfor- ters, the cis adducts do not lactonise or lactonise only to a
tunately) single example, a z-oxo-alkylidenemalonate has very small extent under our reaction or work-up condi-
been reported9 to lead, upon treatment with samarium io- tions. The stereochemistry of the cyclopropanol adducts
dide/HMPA in the presence of tert-butanol, to the product could nevertheless be in each case unambiguously deter-
of simple reduction of the double bond to the exclusion of mined by NOE experiments. Representative results are
5-exo-trig cyclisation products.
given in Figure 1.
In our cases, products of reduction of the double bond
could never be detected despite the fact that radical 3-exo-
trig cyclisations, besides being reversible, are notoriously
slower than 5-exo-trig ones.10 In the case of compound 6c
bearing a cyclopropyl group next to the carbonyl, for
which yields are particularly low, two other products were
isolated by chromatography, accounting each for ca. 30%
of the mass of starting material. Their structure could not
be determined, but inspection of the NMR spectra re-
vealed that the terminal cyclopropyl ring has been main-
tained in one of them but opened in the other one.
Concerning lactones 10, it is worth noting that they are
obtained in a single diastereoisomeric form which was
identified as exo, on the basis of a 3J (H,H) £ 1 Hz between
the two protons a and b to carbonyl groups. AM1 semiem-
pirical calculations, performed at RHF level using Hyper-
chem 5.1 program, confirms this attribution. For instance,
for 10d values of 100.1° and 33.0° are obtained for the di-
hedral angle formed by the Ca–H and the Cb–H bonds in,
HO
H
H
HO
Me
C(CO2t-Bu)2
Me
C(CO2t-Bu)2
Me
H
Me
H
Me
Me
12a
11a
H
HO
HO
H
C(CO2t-Bu)2
Me
Me
Hb
Ha
C(CO2t-Bu)2
Me2C
H
H
H
H
Me
Me
H
H
12 c
11b
Figure 1 Representative NOE correlations (selected examples).
NOE correlations that could not be evidenced with total certainty due
to peak overlaps are not represented.
Table 1 SmI2-Mediated Cyclisation of Dimethyl Alkylidenemalonates
O
O
HO
SmI2 (2 equiv)
CO2Me
CO2Me
CO2Me
R1
CO2Me
O
H
R1
t-BuOH (3 equiv)
+
THF, 0 °C to r.t.
R1
CO2Me
9
6
10
Starting compound
Product
R1
Trans cyclopropanol (yield, %)a
Lactone (yield, %)a
10a (–)
Trans:cis (9:10)
100:0
6a
Me
i-Pr
c-Pr
Ph
9a (49)
9b (45)
9c (32)b
9d (65)
9e (11)
6b
10b (8)
85:15
6c
10c (–)
100:0
6d
10d (33)
66:34
6e
2-Furyl
10e (54)
17:83
a Isolated yields.
b Two other unidentified products accounting to 60% of the mass of starting material were also isolated (see the text).
Synlett 2005, No. 15, 2362–2366 © Thieme Stuttgart · New York

2364
R. Zriba et al.
LETTER
Table 2 SmI2-Mediated Cyclisation of Di-tert-butyl Alkylidenemalonates
CO2t-Bu
CO2t-Bu
O
HO
R1
HO
SmI2 (2 equiv)
CO2t-Bu
R1
CO2t-Bu
CO2t-Bu
R1
AH
+
CO2t-Bu
THF, 0 °C to r.t.
7
11
12
AH = t-BuOH (4 equiv) or PhOH (2 equiv)
Starting compound
Product
R1
Alcohol
t-BuOH
t-BuOH
t-BuOH
t-BuOH
t-BuOH
PhOH
Trans cyclopropanol (yield, %)a
Cis cyclopropanol (yield, %)a
Trans:cis (11:12)
7a
Me
i-Pr
c-Pr
Ph
11a (<5)
12a (61)
12b (–)
12c (80)b
12d (–)
12e (45)
12a (–)
12b (–)
12c (–)
12d (–)
12e (–)
<(10:90)
100:0
0:100
100:0
46:54
100:0
100:0
100:0
100:0
100:0
7b
11b (73)
7c
11c (–)
7d
11d (87)
7e
Bn
11e (38)
7a
Me
i-Pr
c-Pr
Ph
11a (85)
7b
PhOH
11b (62)
7c
PhOH
11c (60)
7d
PhOH
11d (70)
7e
Bn
PhOH
11e (98)
a Isolated yields.
b The corresponding lactone was also isolated in 10% yield.
Inspection of Table 2 reveals in most cases, very clear-cut chemical course of the reactions. Indeed, we were very
but also very puzzling stereoselectivities. Thus, for 7a and please to find that in the presence of phenol (2 equiv),
7c (R = Me or c-Pr), total or almost total selectivity to- trans selectivity was now the rule for all four alkylketones
wards the formation of cis cyclopropanols was observed. 7a–c,e under study (R = methyl, cyclopropyl, isopropyl,
In one (over three) run, however, a very small amount (ca benzyl) as well as for phenylketone 7d.
5%) of trans isomer was also obtained with 7a (R = Me).
In our previous work on the cyclisation of d-oxo-a,b-un-
By contrast and quite surprisingly, 7b (R = i-Pr) was
saturated esters 3,3 we had found that alkylketonic com-
found to give exclusively and reproductibly the trans al-
pounds (3, R1 = alkyl) gave an approximately balanced
cohol. An intermediate situation is encountered with com-
mixture of trans cyclopropanols and lactones in the
presence of tert-butanol. Given the above results, we
pound 7e (R = Bn) whose cyclisation gives comparable
amounts of the two diastereoisomeric cyclopropanols. Fi-
hoped that better, or hopefully total, selectivity towards
nally, total trans stereoselectivity takes place in the cycli-
trans cyclopropanol could be attained by switching to
phenol. Unfortunately, this does not seem to be the case.
We tested this possibility on two compounds. As shown in
sation of 7d (R = Ph) the only aromatic compound
investigated in this study.
We finally run the cyclisation of di-tert-butyl alkylidene- Table 3, in both cases comparable results were obtained
malonates in the presence of phenol, a proton donor much whatever the nature (PhOH or t-BuOH) of the protonating
more acidic than tert-butanol. Reissig and coworkers14 agent. These disappointing results cast doubts about the
have shown that some SmI2 promoted cyclisation exact role of PhOH. However, if the reaction starts by the
reactions of carbonyl compounds could be dramatically reduction of the conjugated double bond,3 the radical
improved by using phenol instead of tert-butanol as the anion produced could perhaps be too short-lived and
protonating agent. More precisely, PhOH was used by cyclise before protonation by PhOH.
these authors to avoid troublesome competitive retro-
Michael fragmentations induced by tert-butoxide and
In conclusion, di-tert-butyl d-oxo-g,g-disubstituted-alkyl-
idenemalonates readily cyclise in the presence of SmI2
other tertiary alcoholates formed in the course of the reac-
and a proton source to cyclopropanols. Complete dia-
stereoselectivity towards trans adducts is obtained when
phenol, a relatively acidic species, is used as the protonat-
tion.14a Our own cyclisations are likely to involve,3 when
run in the presence of t-BuOH, anionic radical inter-
mediates. We thought that PhOH could rapidly protonate
ing agent. These results nicely complement those obtained
these intermediates and change drastically the stereo-
with d-oxo-a,b-unsaturated esters3 for which good (albeit
Synlett 2005, No. 15, 2362–2366 © Thieme Stuttgart · New York

LETTER
Cyclisation of d-Oxo-Alkylidenemalonates to Cyclopropanols
2365
Table 3 SmI2-Mediated Cyclisation of d-Oxo-a,b-Unsaturated Esters 3 in the Presence of t-BuOH or PhOH
O
O
HO
SmI2 (2 equiv)
R2
CO2R'
CO2R'
R1
O
R1
AH
R2
R2
R2 R2
+
R2
THF, 0 °C to r.t.
R1
3
4
5
AH = t-BuOH (4 equiv) or PhOH (2 equiv)
Starting compound
Protonating agent
Alcohol
Product
R1
R2
Overall yield (%)a
Trans:cis (4:5)
74:26
3a
Me
Me
c-Pr
c-Pr
(CH2)5
(CH2)5
Me,Me
Me,Me
t-BuOH
PhOH
94
86
95
90
3a
64:36
3b
t-BuOH
PhOH
0:100
3b
0:100
a Isolated yields.
General Procedure for Cyclisation
not total) stereoselectivity towards the formation of trans
adducts could be attained with aryl ketonic compounds
but not with alkyl ketonic compounds.
To a solution of 1 mmol of d-oxo-alkylidenemalonate and 4 mmol
of tert-BuOH (or 2 mmol of PhOH) in 5–6 mL of dry degassed THF
at 0 °C are added dropwise over ca. 2 min 22 mL of a freshly
prepared3 0.1 M solution of SmI2 in THF. The reaction mixture is
stirred at r.t. with monitoring by TLC or IR spectroscopy on ali-
quots. Reactions are usually complete within 2–3 h. After quench-
ing with dilute aq HCl, the reaction products are extracted in Et2O.
Purification is achieved by flash column chromatography on silica
(heptane–EtOAc mixtures as the eluents).
As already stated, the stereochemical results obtained in
this study are, on the whole, quite puzzling. Clearly, any
mechanistic discussion would at the present time be
mainly speculative. It is hoped that further experimental
work will allow a better understanding of these reactions.
Before closing this paper, we would like to add some
comments regarding the equilibrium between the 4-carb-
alkoxy-but-3-enyl radical 13 and the cyclised radical 14
(Scheme 3). Equilibria of this type are very likely to be
involved in the SmI2 promoted cyclisation of d-halo-a,b-
unsatured esters 1 to cyclopropane compounds 2.1
Acknowledgment
We thank Dr. J.-P. Baltaze for performing all the NMR differential
NOE experiments.
References
kc
(1) (a) David, H.; Bonin, M.; Doisneau, G.; Guillerez, M. G.;
Guibé, F. Tetrahedron Lett. 1999, 40, 8557. (b) Villar, H.;
Guibé, F.; Aroulanda, C.; Lesot, P. Tetrahedron: Asymmetry
2002, 13, 1465.
CO2R
CO2R
k–c
14
13
Scheme 3
(2) Lange, G. A.; Merica, A. Tetrahedron Lett. 1999, 40, 7897;
we became aware of this work only very recently.
(3) Bezzenine-Lafollée, S.; Guibé, F.; Villar, H.; Zriba, R.
Tetrahedron 2004, 60, 6931.
(4) Inokuchi, T. J. Org. Chem. 2005, 70, 1497.
(5) (a) Gansäuer, A.; Lauterbach, T.; Geich-Gimbel, D. Chem.
Eur. J. 2004, 10, 4983. (b) Friedrich, J.; Dolg, M.;
Gansäuer, A.; Geich-Gimbel, D.; Lauterbach, T. J. Am.
Chem. Soc. 2005, 127, 7071.
Our assumption1a that this equilibrium lies on the side of
the homoallylic species 13 was recently rejected5b by
Friedrich and Gansäuer who claimed that the reverse situ-
ation holds true. Aside from their own calculations, their
statement was based on kinetic studies by Beckwith and
Bowry15 that give slightly higher values for kc over k-c
(kc = 1.6 × 107 s–1, k–c = 1.2 × 107 s–1 at 80 °C). In a more
recent study, however, Newcomb and coworkers16 have
convincingly shown that k–c has been underestimated by
Beckwith and that the actual value is at least ten times
higher and probably more. Of course, it could be argued
that kc has also been underestimated. However, this would
mean that radical 13 would now cyclise much more rapid-
ly than does the 4-phenyl-3-butenyl radical for which
kc80 °C equals 2.8 × 107 s–1.15 This appears unlikely since
methyl radical adds only slightly faster (approximately
1.5 times) to acrylic esters than to styrene.17
(6) Fernandez-Mateos, A.; Mateos Buron, L.; Martin de la
Nava, E. M.; Rabanedo Clemente, R.; Rubio Gonzalez, R.;
Sanz Gonzalez, F. Synlett 2004, 2553.
(7) (a) Lehnert, W. Tetrahedron 1973, 29, 635. (b) See also:
Noguchi, M.; Yamada, H.; Takamura, S.; Okada, K.;
Kakeki, A.; Yamamoto, H. Tetrahedron 2000, 56, 1299.
(8) Selected spectroscopic data.
Compound 9a: 1H NMR (200 MHz, CDCl3): d = 3.74 (3 H,
s), 3.72 (3 H, s), 2.94 (1 H, d, J = 12.0 Hz), 1.29 (3 H, s), 1.21
(3 H, s), 1.15 (1 H, d, J = 12.0 Hz), 0.93 (s, 3 H). IR (CHCl3):
1734 cm–1.
Synlett 2005, No. 15, 2362–2366 © Thieme Stuttgart · New York

2366
R. Zriba et al.
LETTER
Compound 9d: 1H NMR (400 MHz, CDCl3): d = 7.51–7.39
(5 H, m), 3.83 (3 H, s), 3.71 (3 H, s), 2.95 (1 H, d, J = 11.5
Hz), 1.64 (1 H, d, J = 11.5 Hz), 1.43 (3 H, s), 0.92 (3 H, s).
IR (CHCl3): 1734 cm–1.
(13) Selected spectroscopic data.
Compound 14a: 1H NMR (400 MHz, CDCl3): d = 2.72 (1 H,
d, J = 11.5 Hz), 1.91 (1 H, br s, OH), 1.48 and 1.47 [(9 + 9)
H, 2 s], 1.31 (3 H, s), 1.25 (3 H, s), 1.08 (1 H, d, J = 11.5 Hz),
0.97 (3 H, s). IR (CHCl3): 1721 cm–1.
Compound 10d: 1H NMR (250 MHz, CDCl3): d = 7.59–7.35
(5 H, m), 3.85 (3 H, s), 3.60 (1 H, s), 2.08 (1 H, s), 1.24 (3 H,
s), 0.96 (3 H, s). IR (CHCl3): 1781, 1739 cm–1.
Compound 15a: 1H NMR (400 MHz, CDCl3): d = 3.13 (1 H,
d, J = 11.0 Hz), 1.45 and 1.44 [(9 + 9) H, 2 s], 1.40 (3 H, s),
1.06 (6 H, br s), 0.71 (1 H, d, J = 11.0 Hz). IR (CHCl3): 1722
cm–1.
(9) Hon, Y.-S.; Lu, L.; Chu, K.-P. Synth. Commun. 1991, 21,
1981.
Compound 14b: 1H NMR (400 MHz, CDCl3): d = 2.78 (1 H,
d, J = 12.0 Hz), 2.73 (1 H, br s, OH), 1.57–1.49 (1 H, sept,
J = 8.0 Hz), 1.39 (18 H, br s), 1.18 (3 H, s), 1.11 (1 H, d,
J = 12.0 Hz), 0.94 (3 H, s), 0.93 (3 H, d, J = 8.0 Hz), 0.89 (3
H, d, J = 8.0 Hz). IR (CHCl3): 1721 cm–1.
(10) (a) This at least is the case for olefinic carbinyl radicals with
kc values10b,c of 0.8 × 104 s–1 and 2.3 × 105 s–1 for cyclisations
of, respectively, the parent 3-butenyl and 5-hexenyl radicals
at 25 °C. Two radical processes may lead to cyclopropanols
and cyclopentanols, namely the 3-exo and 5-exo cyclisations
of ketyl radicals onto olefinic bond or, conversely, the 3-exo
and 5-exo-cyclisations of carbinyl radicals onto carbonyl
groups. Unfortunately, to the best of our knowledge only kc
value of the 5-exo-trig cyclisation of 5-oxo-pentyl radical is
known.10d (b) Newcomb, M.; Glenn, A. G.; Williams, G. W.
J. Org. Chem. 1989, 54, 2675. (c) Chatgilialoglu, C.;
Ingold, K. U.; Scaiano, J. C. J. Am. Chem. Soc. 1981, 103,
7739. (d) Beckwith, A. L. J.; Hay, B. P. J. Am. Chem. Soc.
1989, 111, 2674.
Compound 14c: 1H NMR (250 MHz, CDCl3): d = 3.15 (1 H,
br s, OH), 3.06 (1 H, d, J = 11.5 Hz), 1.78–1.63 (1 H, m),
1.52 and 1.50 [(9 + 9) H, 2 s], 1.26 (3 H, s), 1.17 (1 H, d,
J = 11.5 Hz), 1.13 (3 H, s), 0.66–0.49 (2 H, m), 0.48–0.41 (1
H, m), 0.40–0.32 (1 H, m). IR (CHCl3): 1723 cm–1.
Compound 15c: 1H NMR (400 MHz, CDCl3): d = 3.10 (1 H,
d, J = 11.0 Hz), 2.32 (1 H, br s, OH), 1.44 and 1.43 [(9 + 9)
H, 2 s], 1.31–1.15 (1 H, m), 1.12 (3 H, s), 1.06 (3 H, s), 0.66
(1 H, d, J = 11.0 Hz), 0.55–0.46 (2 H, m), 0.45–0.36 (1 H,
m), 0.35–0.26 (1 H, m). IR (CHCl3): 1722 cm–1.
(14) (a) Gross, S.; Reissig, H.-U. Synlett 2002, 2027. (b) Berndt,
M.; Gross, S.; Hölemann, A.; Reissig, H.-U. Synlett 2004,
422 and references cited therein.
(11) Karplus, M. J. Chem. Phys. 1959, 30, 11.
(12) Barluenga, J.; Aznar, F.; Guttiérez, I.; Martin, J. A. Org.
Lett. 2002, 4, 2719.
(15) Beckwith, A. L. J.; Bowry, V. W. J. Am. Chem. Soc. 1994,
116, 2710.
(16) Horner, J. H.; Tanaka, N.; Newcomb, M. W. J. Am. Chem.
Soc. 1998, 120, 10379.
(17) Zytowsky, T.; Fischer, H. W. J. Am. Chem. Soc. 1996, 118,
437.
Synlett 2005, No. 15, 2362–2366 © Thieme Stuttgart · New York