Intermolecular additions of cyclobutanone derived radicals. A convergent, highly efficient access to polycyclic cyclobutane containing structures

Article abstract of DOI:10.1016/j.tetlet.2003.09.042

α-Xanthyl cyclobutanones undergo intermolecular radical additions to olefins bearing various functional groups. Adducts containing a phosphonate can be made to cyclise by an internal Wittig-Horner-Emmons reaction to give cyclohexene and cycloheptene rings fused to the cyclobutane.

Full text of DOI:10.1016/j.tetlet.2003.09.042

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 7703–7706
Intermolecular additions of cyclobutanone derived radicals.
A convergent, highly efficient access to polycyclic cyclobutane
containing structures
Gre´gori Binot and Samir Z. Zard*
Laboratoire de Synthe`se Organique associe´ au CNRS, Ecole Polytechnique, F-91128 Palaiseau, France
Received 16 May 2003; accepted 5 September 2003
Abstract—a-Xanthyl cyclobutanones undergo intermolecular radical additions to olefins bearing various functional groups.
Adducts containing a phosphonate can be made to cyclise by an internal Wittig–Horner–Emmons reaction to give cyclohexene
and cycloheptene rings fused to the cyclobutane.
© 2003 Elsevier Ltd. All rights reserved.
Cyclobutane derivatives play an important role in
organic synthesis, both as synthetic targets, since a
sizeable number of natural products and especially
terpenes contain a cyclobutane subunit, and as syn-
thetic intermediates, in view of the variety of useful and
sometimes specific transformations undergone by the
strained 4-membered carbocycle.1 In this second con-
text, cyclobutanones are of special prominence. Their
chemistry is amazingly rich2 and they are readily acces-
sible through the powerful and highly flexible [2+2]
cycloaddition of ketenes and ketene iminium salts.3 The
latter approach is particularly well suited for enantiose-
lective syntheses.4 However, in contrast to larger
cycloalkanones, alkylation of enolates derived from
cyclobutanones is complicated by self-condensation and
by the possibility for the enolate to undergo a thermal
electrocyclic opening. This imposes quite severe con-
straints on the nature of the alkylating agents and the
experimental conditions that can be used. The very few
examples described in the literature concern only highly
reactive alkylating agents such as methyl iodide and
allyl and propargyl halides.5 In one example, a less
reactive alkyl bromide could be introduced in modest
yield by using the magnesium salt of the derived
cyclobutanone cyclohexylimine.5d It is not surprising,
therefore, that the introduction of side chains onto a
cyclobutanone usually necessitates an indirect approach
such as a sequence of aldol reaction, dehydration, and
reduction. We have now found that the xanthate trans-
fer radical addition reaction we uncovered a few years
ago6 nicely complements existing methods and allows
entry into cyclobutane structures hitherto unavailable
or accessible only with difficulty.
Bromide 2 was easily prepared from cyclobutanone 1
by the reaction with bromine in chloroform or by
treatment of the corresponding enol silyl ether with
NBS.7 We found the latter procedure more general.
Displacement of the bromide to give the desired xan-
thate 3 was accomplished in good yield by the action of
potassium O-ethyl xanthate in acetone (Scheme 1). The
analogous substitution of the less reactive but more
readily accessible chloride (by partial reduction of the
dichloroketene adduct) could not be achieved in satis-
factory yield. We were pleased to find that refluxing a
solution of 3 with allyl acetate in 1,2-dichloroethane in
the presence of a small amount of lauroyl peroxide
resulted in a smooth radical addition to give compound
4 in 67% isolated yield as a mixture of diastereoisomers.
It must be noted that the stereochemistry of the xan-
thate precursor is of no consequence since both epimers
lead to the same radical. Reductive removal of the
xanthate group with tributyl stannane gave a 9:1 mix-
ture of epimers 5 (87%). The same reductive dexanthy-
lation could be accomplished in 78% yield under
tin-free conditions using stoichiometric amounts of lau-
royl peroxide in 2-propanol.8 The trans epimer domi-
nates since the radical addition to the olefin occurs
from the least hindered side opposite to the butyl
group. The utility of this sequence leading to branched
cyclobutanones is highlighted by the conversion of 5
Keywords: radical additions; xanthate transfer; cyclobutanones;
intramolecular Horner–Emmons reaction.
* Corresponding author. Tel.: +33 (0)169334872; fax: +33
(0)169333851; e-mail: zard@poly.polytechnique.fr
0040-4039/$ - see front matter © 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.09.042

7704
G. Binot, S. Z. Zard / Tetrahedron Letters 44 (2003) 7703–7706
Scheme 1. Reagents and conditions: (i) Br2/CHCl3 or TBSOTf, Et3N, CH2Cl2; −78°C then NBS, THF, 0°C; (ii) EtOCSSK/acetone,
rt; (iii) lauroyl peroxide (5–20 mol%), 1,2-CH2ClCH2Cl, reflux; (iv) n-Bu3SnH (AlBN), benzene, reflux; (v) lauroyl peroxide (140
mol%), 2-propanol, reflux; (vi) m-CPBA, CH2Cl2, rt; (vii) NaH, THF, rt.
into lactone 6 in 69% yield by the regio- and stereo-
selective Bayer–Villiger reaction.
tion by resonance in the cyclobutenoxyl canonical
structure because of increased strain, in comparison
with the essentially strain-free cyclohexenoxyl
mesomeric form. A further, stark reminder of the spe-
cial nature of cyclobutanones arose when we attempted
to prepare xanthate 17. The reaction of bromide 16
with potassium O-ethyl xanthate at room temperature
was more sluggish and somewhat less clean than with
the analogous case of bromide 2, presumably a conse-
quence of increased hindrance from the bulkier sub-
stituent. We therefore repeated the reaction in refluxing
acetone in the hope of accelerating the reaction and
isolated a compound in 43% yield, which had the
correct molecular weight by mass spectrometry, a pat-
tern of signals in the NMR spectrum roughly compat-
ible with the expected structure 17, but the IR did not
show the carbonyl absorption typical of a cyclobu-
tanone. The prominent band in the carbonyl region
appeared at 1713 cm−1. The 13C spectrum showed two
signals at 216.1 and 69.0 ppm, hinting at the presence
of a thiocarbonyl group. This was confirmed by an IR
band at 1051 cm−1. These data are in accord with
structure 21, resulting from further reaction of the
desired xanthate under the harsher experimental condi-
tions. Thus, reversible addition of the xanthate salt to
the reactive carbonyl group of the cyclobutanone gives
intermediate 18 which collapses by opening of the
strained ring to give transient anion 19, which under-
goes rearrangement to the more stabilised thiolate 20
and then ring-closure to the observed derivative 21, as
outlined in Scheme 2.
In a similar fashion, xanthate 8 was prepared from the
indene-derived bromide 7 and made to react with allyl
acetate. The adduct, 9, obtained in 81% yield, was a
mixture of only two isomers, which gave the same
reduced derivative 10 upon exposure to tributylstan-
nane. Clearly in this case, the radical addition occured
exclusively from the least hindered exo face. Allyl ace-
tate can be replaced by other olefins. For example, in
the presence of vinyl pivalate, a good yield of adduct 11
is obtained containing a geminal xanthate and pivaloxy
representing in fact a masked aldehyde function. We
have used such adducts recently in a flexible synthesis
of pyrroles and of various carbocycles.9 Perhaps more
interestingly, radical addition to methyl 2-(diethylphos-
phono)-4-pentenoate furnished xanthate 12 in 68%
yield and this compound underwent a smooth ring
closure upon treatment with sodium hydride in THF to
give tetracyclic derivative 13 in 66% yield.10 The seven-
membered ring homologue 15 could be constructed by
simply incorporating one more carbon into the olefinic
trap, as illustrated by the synthesis of adduct 14 and its
ring closure by the intramolecular Horner–Emmons
reaction in 68% and 55% yield respectively.
We found the radical next to the carbonyl group in a
cyclobutanone to be noticeably more reactive towards
ordinary olefins than the analogous radical in a cyclo-
hexanone. This may be due to a diminished stabilisa-

G. Binot, S. Z. Zard / Tetrahedron Letters 44 (2003) 7703–7706
7705
Scheme 2. Reagents and conditions: (i) EtOCSSK/acetone; (ii) neoPnOCSSNa/acetone, rt; (iii) lauroyl peroxide (5–20 mol%),
1,2-CH2ClCH2Cl, reflux; (iv) n-Bu3SnH (AlBN), toluene/cyclohexane, reflux.
We could circumvent this unexpected problem by
replacing the potassium O-ethyl xanthate with sodium
O-neopentyl xanthate and keeping the reaction mixture
at room temperature. The substitution reaction
remained slow but was much cleaner giving the corre-
sponding xanthate 22 in 78% yield after a reaction time
of 5 hours. The bulkier neopentyl xanthate is less
inclined to attack the carbonyl group of the product 22,
which itself is also sterically more congested.11 As for
the radical addition itself, it proceeded nicely as shown
by the efficient addition to Boc-protected allylamine
giving adduct 23 in 76% yield. Reductive removal of the
xanthate group allowed us to estimate the epimeric
ratio as being 80:20 in favour of the trans epimer.
The utility of the neopentyl xanthate was further
demonstrated by the transformations implemented on
the two bicyclic systems portrayed in Scheme 3. Thus,
xanthates 25 and 30 derived from the corresponding
bromide precursors12 underwent clean radical additions
to various olefins. Addition of 25 to the diethyl acetal
of acrolein and to allyl trimethylsilane gave compounds
26 and 27 in 59% and 88% yield respectively, whereas
addition to methyl 2-(diethylphosphono)-4-pentenoate
provided the Horner–Emmons precursor 28 in 65%
yield. The latter was cleanly converted into tricyclic
structure 29 by the action of sodium hydride (82%). In
the case of xanthate 30, addition to N-allyl phthal-
imide, allyl cyanide, and methyl 2-(diethylphosphono)-
Scheme 3. Reagents and conditions: (i) lauroyl peroxide (5–20 mol%), 1,2-CH2ClCH2Cl, reflux; (ii) NaH, THF, rt.

7706
G. Binot, S. Z. Zard / Tetrahedron Letters 44 (2003) 7703–7706
4-pentenoate provided the corresponding adducts 31,
32, and 33 in 73%, 72%, and 69% yield, respectively. In
all of these examples, the radical addition took place
from the more accessible face exclusively or with a very
high selectivity. The complete exo stereoselectivity in
the case of 33 was shown by its cyclisation into 34
(85%) with sodium hydride in THF thus eliminating the
asymmetric centre bearing the phosphonate group and
simplifying the spectral analysis.
Vol. 1, pp. 167–213; (b) Bellus, D.; Ernst, B. Angew.
Chem., Int. Ed. Engl. 1988, 27, 797–827.
3. (a) Snider, B. B. Chem. Rev. 1988, 88, 793–811; (b)
Mehta, G.; Rao, H. S. P. Synth. Commun. 1985, 15,
991–1000; (c) Brady, W. T. Synthesis 1971, 8415–8422;
(d) Kagan, H. B. Ann. Chim. 1965, 10, 203–212.
4. (a) Lee-Ruff, E.; Maldenova, G. Chem. Rev. 2003, 103,
1449–1484 and references there cited therein; (b) Orr, R.
K.; Calter, M. A. Tetrahedron 2003, 59, 3545–3566; (c)
Nemoto, H.; Fukumoto, K. Synlett 1996, 863–875; (d)
Genicot, C.; Ghosez, L. Tetrahedron Lett. 1992, 33,
7357–7360.
5. (a) Medarde, M.; Lo´pez, J. L.; Morillo, M. A.; Tome´, F.;
Adeva, M.; San Feliciano, A. Tetrahedron Lett. 1995, 36,
8097–8098; (b) Poss, A. J.; Belter, R. K. J. Org. Chem.
1987, 52, 4810–4812; (c) Conia, J.-M.; Gore´, J. Bull. Soc.
Chim. Fr. 1964, 1968–1975; (d) Dowd, P.; Zhang, W. J.
Org. Chem. 1992, 57, 7163–7171.
6. (a) Zard, S. Z. Angew. Chem., Int. Ed. Engl. 1997, 36,
672–685; (b) Quiclet-Sire, B.; Zard, S. Z. Phosphorus,
Sulfur and Silicon 1999, 153-154, 137–154; (c) Zard, S. Z.
In Radicals in Organic Synthesis; Renaud, P.; Sibi, M. P.,
Eds.; Wiley-VCH: Weinheim, 2001; Vol. 1, pp. 90–108.
7. Reuss, R. H.; Hassner, A. J. Org. Chem. 1974, 39,
1785–1787.
The above examples, involving olefinic traps containing
various functional groups, underscore the generality,
efficiency, and simplicity of the xanthate transfer radi-
cal addition. None of these transformations can be
easily accomplished by traditional ionic and
organometallic processes or even by other radical based
methods. The ability to combine the radical addition
with an efficient intramolecular Horner–Emmons olefi-
nation and presumably a variety of other ring forming
reactions opens access to numerous polycyclic frame-
works of some complexity. Finally, the unique chem-
istry of the cyclobutanone ring itself represents a
further tremendous synthetic asset. The strain in the
4-membered ring greatly facilitates a variety of highly
selective transformations such as the Bayer–Villiger
reaction (as in example 6 above), ring expansions with
diazo derivatives,2,13 the Beckmann and related rear-
rangements, etc. Further studies along these lines are in
progress.
8. Liard, A.; Quiclet-Sire, B.; Zard, S. Z. Tetrahedron Lett.
1996, 37, 5877–5880.
9. (a) Quiclet-Sire, B.; Quintero, L.; Sanchez-Jimenez, G.;
Zard, S. Z. Synlett 2003, 75–78; (b) Binot, G.; Quiclet-
Sire, B.; Saleh, T.; Zard, S. Z. Synlett 2003, 382–386.
10. Cholleton, N.; Gillaizeau-Gauthier, I.; Six, Y.; Zard, S.
Z. Chem. Commun. 2000, 535–536.
11. For another case where the O-neopentyl xanthate proved
superior, see: Quiclet-Sire, B.; Zard, S. Z. J. Am. Chem.
Soc. 1996, 118, 9190–9191.
Acknowledgements
One of us (G.B.) gratefully acknowledges generous
12. Broom, N.; O’Hanlon, P. J.; Simpson, T. J.; Stephen, R.;
Willis, C. L. J. Chem. Soc., Perkin Trans. 1 1995, 3067–
3072.
financial support from the DGA.
13. See for example: (a) Brocksom, T. J.; Coelho, F.; Depre´s,
J.-P.; Greene, A. E.; Freire de Lima, M. E.; Hamelin, O.;
Hartmann, B.; Kanazawa, A. M.; Wang, Y. J. Am.
Chem. Soc. 1996, 118, 15313–15325 and references there
cited; (b) Biswas, S.; Gosh, A.; Venkateswaran, R. V. J.
Org. Chem. 1990, 55, 3498–3502.
References
1. Namyslo, J. C.; Kaufmann, D. E. Chem. Rev. 2003, 103,
1485–1538 and references there cited therein.
2. (a) Lee-Ruff, E. In Advances in Strain in Organic Chem-
istry; Halton, B., Ed.; JAI Press: Greenwich, CT, 1991;