Radical cascade processes leading to fused- and spiro-bicyclic ring systems

Article abstract of DOI:10.1016/S0040-4020(00)00848-6

The course of the reaction of 2-propargyl-2-(2-carboethoxyvinyl)cyclohexanone with tributyltin hydride and AIBN either neat or in benzene solution is discussed. The structure of the spirocyclic major product is shown to differ from that previously suggest

Full text of DOI:10.1016/S0040-4020(00)00848-6

TETRAHEDRON
Pergamon
Tetrahedron 56 (2000) 8959±8965
Radical Cascade Processes Leading to Fused-
and Spiro-Bicyclic Ring Systems
*
Jeremy Robertson, Hon Wai Lam, Sokol Abazi, Stephen Roseblade and Rachel K. Lush
The Dyson Perrins Laboratory, South Parks Road, Oxford OX1 3QY, UK
Received 4 July 2000; revised 30 August 2000; accepted 14 September 2000
AbstractÐThe course of the reaction of 2-propargyl-2-(2-carboethoxyvinyl)cyclohexanone with tributyltin hydride and AIBN either neat or
in benzene solution is discussed. The structure of the spirocyclic major product is shown to differ from that previously suggested and
plausible reaction pathways are presented to account for the outcome. q 2000 Elsevier Science Ltd. All rights reserved.
Introduction
We set out to investigate this reaction further in order to ®nd
support for some of these interesting mechanistic possi-
bilities and in this report we show that, in fact, a different
sequence of events takes place and that none of the
compounds 2, 3, or 4 are produced.
Recently we had occasion to repeat the reaction of enyne 1
with tributyltin hydride under homolytic conditions, a
process that has been reported1 to generate spirocycle 3
after treatment of the proposed intermediate 2 with PPTS
(Scheme 1).
Results and Discussion
This process intrigued us in view of the likely steric impedi-
ment towards attack at the b-position of the enoate alkene,
which bears an adjacent quaternary centre, particularly
because an unhindered terminal alkyne is present in the
same molecule. Although radical additions to alkynes are
generally slower2 than those to comparable alkenes, we
envisaged that the steric bias in this example was such
that a more likely reaction pathway would initiate with
attack of stannyl radical at the exposed sp-centre.3
However, in order to form the reported product (3), a stereo-
electronically unfavourable 5-endo-trig4 cyclisation is
required (Scheme 2); alternatively 4-exo-trig cyclisation
and ring-expansion5 would generate the same product
although this would have to involve a bicyclo[2.1.0]pentane
intermediate that is expected to be inaccessible under these
conditions. This scheme has the advantage that a vinyl
stannane (4) would be generated which should be much
more readily protiodestannylated6 than the proposed inter-
mediate 2.
The required precursor 1 was prepared by a slight modi®-
cation of the literature procedure, the Wadsworth±Emmons
ole®nation being run directly on the crude product obtained
by propargylation of 2-formylcyclohexanone to minimise
deformylation processes. Subjecting this material to the
conditions reported for the radical reaction [tributyltin
hydride (ca. 1.2 equiv.), AIBN (0.48 equiv.), neat, 808C,
15 min] followed by treatment with PPTS in dichloro-
methane (RT, 48 h) produced hydrostannylated compound
5 as a mixture of E- and Z- isomers (21%), known allyl-
cyclohexanone derivative 67 (24%), and a third component,
assigned as structure 78 (35%) (Scheme 3). This result
supported the view that stannyl radical attack at the alkyne
occurs in preference to attack at the (relatively hindered)
alkene.
Next, in an effort to promote a radical cascade sequence, the
tin hydride was added over 6 h (syringe pump) to a 10 mM
Scheme 1.
Keywords: cyclopropanes; radicals and radical reactions; rearrangements; spiro compounds.
* Corresponding author. Tel.: 144-1865-275660; fax: 144-1865-275674; e-mail: jrobert@ermine.ox.ac.uk
0040±4020/00/$ - see front matter q 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)00848-6

8960
J. Robertson et al. / Tetrahedron 56 (2000) 8959±8965
Scheme 2.
Scheme 3. Reagents: (i) Bu3SnH, AIBN, neat; (ii) PPTS, CH2Cl2.
Scheme 4.
solution of substrate 1 in benzene at re¯ux. Once more, the
intermediate stannylated compounds were not isolated at
this point but directly subjected to PPTS treatment. Under
these conditions the major isolated product (42% in the ®rst
ketoester. In an effort to simplify the spectra and allow
correlation with known diketone 99 the alkene was cleaved
ozonolytically giving, we presumed, diketoester 8 but ester
hydrolysis and decarboxylation of this material failed.
Whilst hydrolysis could be effected there was no sign of
loss of carbon dioxide, even under forcing conditions, and
it appeared likely that the structure of ketoester 8 and, by
implication, that of its precursor (3) were incorrect. On the
basis of extensive NMR investigations the regioisomeric
structures 10 and 11 were proposed for 3 and 8 respectively
(Scheme 4).
1
run) exhibited H and 13C NMR data that appeared con-
sistent with the proposed structure (3) for the spirocyclic
Crucially, in the HMBC spectrum of alkene 10, the
resonance corresponding to CHCO2Et showed two- and
three-bond couplings to the highlighted carbons (z) shown
Figure 1. Carbons marked (z) show a strong cross-peak to the ester a-proton
in the HMBC spectrum.
Scheme 5. Reagents: (i) Bu3SnH, AIBN, 7.6 mM in PhH; (ii) PPTS, CH2Cl2.

J. Robertson et al. / Tetrahedron 56 (2000) 8959±8965
8961
Scheme 6.
cyclisation is far from obvious; 5-endo-trig cyclisation
would give the spirocycle directly (after reduction) but
this is likely to be slow.10 In principle, a reversible but
kinetically preferred 3-exo-trig process could generate
vinyl cyclopropane intermediate 16 whose evolution to the
product is conceivable by a vinyl cyclopropane±cyclo-
pentene rearrangement11 although further work is needed
before ®rm conclusions can be drawn.
Although spirocycle 17 may well not be an intermediate in
this reaction, the possibility that a stannyl radical-catalysed
rearrangement of this compound operates to give 12 led us
to prepare a series of structurally related vinyl cyclopropane
substrates (Fig. 2) whose reactivity might shed light on the
mechanism.12 In fact, these studies13 revealed complex
behaviour in which a fascinating interplay of electronic
and steric effects operates to dictate the course of the
rearrangements and a substantial discussion of the results
will be presented in a separate report.
Figure 2.
in Fig. 1, there being no cross-peak to the alkene CH2 which
discounts structure 3 as a possibility.
Further investigation revealed that a second component 14
(5.5%, radical reaction run at a ®nal substrate concentration
of 7.6 mM) was formed in addition to the spirocycle 10
(54%), and this became the major product when the concen-
tration of the radical reaction was increased to ca. 40 mM.
Full characterisation of the stannanes leading to these two
products was hampered by their instability but, on the basis
of their spectroscopic data and their products of protio-
destannylation, the structures 12 and 13 (Scheme 5) are
proposed (see Experimental for details).
However, the preparation and reaction of cyclopropane 20
(Scheme 7) are worth noting in this context as this substrate
proved to possess the correct combination of structural
features for a vinyl cyclopropane±cyclopentene rearrange-
ment catalysed by stannyl radical. Thus, enolate alkylation
of cyclohexanone with 4-chloro-1-iodobut-2-ene and treat-
ment of the crude product with potassium t-butoxide under
equilibrating conditions afforded separable vinyl cyclo-
propanes14 18 and 19 (2:1 ratio). These were ole®nated
separately under Wadsworth±Emmons conditions, the
trans-isomer (18) being converted in high yield into a 3:1
ratio of alkene diastereomers 20 and 21. Interestingly,
under the same conditions, the more crowded cis-cyclo-
propane isomer (19) reacted only very slowly with the
ole®nation reagent and the so-formed enoate underwent
The mechanism of this transformation remains obscure but,
taking all the results together, the early part of the sequence
is most readily explained by a 4-exo-trig cyclisation as
originally suggested (Scheme 2) but the reaction pathway
diverges at this point (Scheme 6). Fragmentation would lead
to a-carbonyl radical 15 which, in a concentration-depen-
dent process, could be either reduced directly to give diene
13 or could cyclise to give spirocycle 12. The mode of
Scheme 7. Reagents: (i) LDA, 4-chloro-1-iodobut-2-ene, THF; (ii) KOt-Bu, THF; (iii) NaH, (EtO)2P(O)CH2CO2Et, DME.

8962
J. Robertson et al. / Tetrahedron 56 (2000) 8959±8965
Scheme 8. Reagents: Bu3SnH (0.25 equiv.), AIBN, PhH.
divinylcyclopropane rearrangement15 under the reaction
conditions to give bicyclic ester 22.
15 min then, after cooling, dichloromethane (20 mL) and
PPTS (2.15 g, 8.4 mmol) were added and the mixture was
stirred at RT for 48 h. The solvent was removed in vacuo
and the residue was triturated repeatedly with a 2:3 mixture
of petrol and ethyl acetate (10£10 mL). The combined
extracts were concentrated in vacuo to give a viscous yellow
oil. Puri®cation by chromatography (12:1 petrol:ether) gave
2-(2-carboethoxyvinyl)-2-[3-(tributylstannyl)prop-2-enyl]-
cyclohexanone (5) in which the (E)-isomer predominated
(230 mg, 21%), 2-allyl-2-(2-carboethoxyvinyl)cyclohexa-
none7 (6) (420 mg, 24%) and 2-(4-carboethoxy-2-methyl-
enebutyl)cyclohexanone (7) (170 mg, 35%) as colourless
oils. Data for 5 [(E)-isomer]: Found: C, 59.41; H, 9.01.
C26H46O3Sn requires: C, 59.44; H, 8.83%; Rf 0.66 (1:1
petrol:ether); nmax/cm21 2929s, 2871m, 1715s, 1644m,
1597m, 1464s, 1312s, 1271s, 1180s, 1039s, 989m, 864m;
dH (500 MHz, CDCl3) 0.87±0.92 (15H, m), 1.27±1.34 (9H,
m), 1.45±1.50 (6H, m), 1.75±1.82 (4H, m), 1.89±2.03 (2H,
m), 2.41±2.45 (2H, m), 2.51±2.53 (2H, m), 4.20 (2H, q,
Jˆ7.0 Hz), 5.75 (1H, d, Jˆ16.5 Hz) overlays 5.74 (1H, dt,
Jˆ18.5, 7.0 Hz), 5.96 (1H, d, Jˆ18.5 Hz), 7.06 (1H, d,
Jˆ16.5 Hz); dC (50.3 MHz, CDCl3) 9.3, 13.5, 14.1, 21.2,
26.8, 27.1, 29.0, 35.4, 39.7, 45.1, 54.6, 60.5, 122.1, 133.7,
143.1, 150.9, 166.4, 211.2; m/z (CI) 527 (M(120Sn)H1,
29%), 526 (M(119Sn)H1, 15), 525 (M(118Sn)H1, 28), 524
(M(117Sn)H1, 12), 523 (M(116Sn)H1, 16), 469 (100), 291
(45), 235 (43), 177 (44), 121 (14), 83 (13). Data for 6:7 Rf
0.57 (1:1 petrol:ether); nmax/cm21 2938m, 2866m, 1713s,
1643m, 1448m, 1368m, 1314s, 1272m, 1188s, 1036m,
986m, 919w; dH (500 MHz, CDCl3) 1.29 (3H, t, Jˆ
7.0 Hz), 1.72±1.81 (4H, m), 1.89±1.94 (1H, m), 1.98±
2.04 (1H, m), 2.36±2.49 (4H, m), 4.20 (2H, q, Jˆ7.0 Hz),
5.04 (1H, dd, Jˆ10.0, 1.5 Hz), 5.06 (1H, dd, Jˆ17.0,
1.5 Hz), 5.64 (1H, ddt, Jˆ17.0, 10.0, 7.5 Hz), 5.76 (1H, d
Jˆ16.5 Hz), 7.03 (1H, d, Jˆ16.5 Hz); dC (50.3 MHz,
CDCl3) 14.0, 21.2, 26.8, 35.6, 39.6, 41.1, 54.4, 60.5,
118.8, 122.4, 133.1, 150.6, 166.3, 211.0; m/z (CI) 254
(MNH14 , 100%), 237 (MH1, 64), 191 (22), 61 (15). Data
for 7: Rf 0.51 (1:1 petrol:ether); Accurate mass: Found
239.1647, C14H23O3 (MH1) requires 239.1647; nmax/cm21
2999m, 2942s, 1736s, 1711s, 1462s, 1440m, 1369s, 1230s,
1184s, 1039w, 707w; dH (400 MHz, CDCl3) 1.23 (3H, t,
Jˆ7.0 Hz), 1.25±1.33 (1H, m), 1.61±1.69 (2H, m), 1.87
(1H, dd, Jˆ14.5, 9.0 Hz), 1.79±1.87 (1H, m), 2.02±2.13
(2H, m), 2.22±2.35 (3H, m), 2.38 (1H, t, Jˆ4.0 Hz),
2.41±2.50 (3H, m), 2.59 (1H, dd, Jˆ14.5, 5.0 Hz), 4.11
(2H, q, Jˆ7.0 Hz), 4.71 (1H, br s), 4.76 (1H, br s); dC
(100.6 MHz, CDCl3) 14.2, 24.9, 28.0, 30.6, 32.5, 33.4,
35.8, 42.0, 48.4, 60.3, 111.0, 145.5, 173.2, 212.6; m/z (CI)
256 (MNH14 , 5%), 240 (14), 239 (MH1, 100), 238 (14), 193
(100).
Treatment of a 5 mM benzene solution of diastereomer 20
with 0.25 molar equivalents of tributyltin hydride in the
presence of AIBN resulted in a smooth transformation
(90%) to give the hexahydroindene derivative 23 (Scheme
8). We suggest a catalytic cascade sequence consisting of (i)
addition of stannyl radical to the unconjugated alkene; (ii)
cyclopropyl carbinyl fragmentation; (iii) 5-exo-trig cycli-
sation of the so-formed allylic radical; (iv) re-ejection of
stannyl radical to continue the chain.
Conclusion
Further studies are required in order to be con®dent of the
mechanism for the transformation of cyclohexanone 1 into
spirocycle 12, in particular whether or not a direct 5-endo-
trig pathway from intermediate 15 can be ruled out as a
possibility. Whichever mechanism proves to be supported
in future work, our studies have already led to the identi®-
cation of novel rearrangement processes and we hope that,
ultimately, the results will prove to be of synthetic value as
well as mechanistic interest.
Experimental
General
1H and 13C NMR spectra were recorded on Varian Gemini
200, Bruker AC 200, Bruker DPX 400, or Bruker AM 500
spectrometers with 1H assignments being made on the basis
of a combination of chemical shift, coupling constant,
COSY, and NOE data as appropriate; 13C assignments
were supported by DEPT and HMQC experiments.
Coupling constants (J) are quoted to the nearest 0.5 Hz.
Infrared spectra were recorded on either Perkin±Elmer
1750 or Perkin±Elmer Paragon 1000 FT spectrometers as
thin ®lms. Mass spectra were recorded on VG Micromass
ZAB 1F, Masslab 20-250, VG Platform, or VG TRIO-1
spectrometers. Accurate mass data were obtained by the
EPSRC Mass Spectrometry Service Centre at Swansea.
All solvents and commercially available reagents were
dried and puri®ed according to standard procedures. `Petrol'
refers to the fraction of light petroleum ether boiling in the
range 30±408C; `ether' refers to diethyl ether. All experi-
ments involving air-sensitive reagents were performed
under an argon atmosphere.
Treatment of enyne 1 with tributyltin hydride without
solvent. A mixture of enyne 11 (490 mg, 2.1 mmol), AIBN
(170 mg, 1 mmol) and tributyltin hydride (ca. 70% hydride
content by NMR, 970 ml, 2.5 mmol) was stirred at 808C for
Treatment of enyne 1 with tributyltin hydride in ben-
zene. To a solution of the enyne 1 (110 mg, 0.47 mmol) and
AIBN (10 mg) in degassed benzene (52 mL) at re¯ux was

J. Robertson et al. / Tetrahedron 56 (2000) 8959±8965
8963
added a solution of tributyltin hydride (ca. 70% hydride
content by NMR, 219 ml, 0.57 mmol) in degassed benzene
(10 mL) over 3 h (syringe pump). The mixture was left at
re¯ux for 1 h and then the solvent was removed in vacuo to
afford a yellow oil. In a separate experiment (for character-
isation purposes) puri®cation of a small portion of this oil by
chromatography (15:1 petrol:ether) yielded 1-carboethoxy-
3-(tributylstannyl)methylspiro[4.5]dec-2-en-6-one (12) as a
labile, colourless oil. Rf 0.48 (2:1 petrol:ether); nmax/cm21
2926s, 2870s, 2880s, 1732s, 1712s, 1645w, 1464s, 1178s,
1038m, 669m; dH (500 MHz, CDCl3) 0.85±0.98 (15H, m),
1.25 (3H, t, Jˆ7.0 Hz) overlays 1.23±1.35 (6H, m), 1.35±
1.56 (6H, m), 1.62±1.80 (4H, m), 1.86±1.90 (1H, m), 1.92±
2.00 (1H, m), 2.36±2.56 (6H, m), 4.12 (2H, q, Jˆ7.0 Hz),
4.33 (1H, d, Jˆ1.5 Hz), 5.03 (1H, d, Jˆ1.5 Hz); dC
(125 MHz, CDCl3) 9.5, 12.5, 13.7, 14.3, 22.3, 26.7, 27.5,
29.1, 34.4, 39.2, 47.3, 54.3, 59.3, 60.2, 116.8, 143.4, 173.7,
210.6. Also obtained was ethyl 4-[(2-oxocyclohexyl)-
methyl]-5-tributylstannylpenta-2,4-dienoate (13) as a labile,
colourless oil. nmax/cm21 2929s, 2855s, 1714s, 1619s,
1555w, 1464m, 1280s, 1176s, 1043m, 866m; dH (400
MHz, CDCl3) 0.82±0.94 (9H, m), 0.99 (6H, app. t,
Jˆ8.0 Hz), 1.20±1.33 (9H, m), 1.40±1.51 (6H, m), 1.52±
1.62 (3H, m), 1.75±1.86 (1H, m), 1.96±2.14 (3H, m), 2.23±
2.35 (1H, m), 2.36±2.47 (2H, m), 3.03 (1H, dd, Jˆ14.5,
4.0 Hz), 4.18 (2H, q, Jˆ7.0 Hz), 5.80 (1H, d, Jˆ16.0 Hz),
6.40 (1H, s), 7.22 (1H, d, Jˆ16.0 Hz); dC (100 MHz,
CDCl3) 10.6, 13.6, 14.2, 25.0, 27.2, 27.9, 28.9, 33.4, 35.4,
42.1, 49.0, 60.2, 118.3, 146.9, 147.6, 150.5, 167.0, 211.9.
The crude product from the radical reaction was dissolved in
dichloromethane (15 mL), PPTS (472 mg, 1.88 mmol) was
added and the mixture was stirred at RT for 48 h. The
solvent was removed in vacuo and the residue triturated
with a 2:3 mixture of petrol and ethyl acetate (10£5 mL).
The combined extracts were concentrated in vacuo to afford
a viscous yellow oil that was puri®ed by chromatography
(12:1 petrol:ether) to yield 1-carboethoxy-3-methylene-
spiro[4.5]decan-6-one (10) (60 mg, 54%). Rf 0.38 (2:1
petrol:ether); Accurate mass: Found 237.1491, C14H21O3
(MH1) requires 237.1491; nmax/cm21 2935m, 2866m,
1729s, 1710s, 1448w, 1371m, 1264w, 1181m, 1041m,
879w; dH (500 MHz, CDCl3) 1.24 (3H, t, Jˆ7.0 Hz),
1.62±1.74 (3H, m), 1.75±1.83 (2H, m), 1.95±2.01 (1H,
m), 2.36±2.50 (3H, m), 2.62±2.66 (3H, m), 3.64 (1H, t,
Jˆ8.5 Hz), 4.06±4.21 (2H, m), 4.88±4.89 (2H, m); dC
(125 MHz, CDCl3) 14.2, 21.9, 26.1, 32.4, 33.2, 39.2, 43.3,
47.9, 58.4, 60.3, 107.5, 147.0, 173.5, 211.3; m/z (CI) 254
(MNH14 , 10%), 238 (14), 237 (MH1, 100), 191 (39), 190
(28), 163 (28), 162 (18), 146 (21). Also obtained was ethyl
4-[(2-oxocyclohexyl)methyl]penta-2,4-dienoate (14) (6 mg,
5.5%). Accurate mass: Found 237.1486, C14H21O3 (MH1)
requires 237.1491; nmax/cm21 2937s, 2863m, 1711br s,
1631s, 1603m, 1448m, 1367m, 1308s, 1274s, 1177s,
1038s, 986m, 866m; dH (400 MHz, CDCl3) 1.30 (3H,
t, Jˆ7.0 Hz), 1.56±1.71 (3H, m), 1.82±1.89 (1H, m),
2.02 (1H, dd, Jˆ14.5, 9.0 Hz) overlaying 2.02±2.14
(2H, m), 2.26±2.37 (1H, m), 2.40±2.49 (2H, m), 2.91
(1H, ddd, Jˆ14.5, 4.0, 0.5 Hz), 4.21 (2H, q, Jˆ7.0 Hz),
5.30 (1H, s), 5.43 (1H, s), 5.85 (1H, d, Jˆ16.0 Hz),
7.28 (1H, d, Jˆ16.0 Hz); dC (100 MHz, CDCl3) 14.3,
25.1, 27.9, 31.4, 33.8, 42.1, 48.6, 60.4, 118.4, 125.0, 142.4,
146.2, 167.0, 211.9; m/z (CI) 237 (MH1, 11%), 191 (100),
163 (9).
1-Carboethoxyspiro[4.5]decan-3,6-dione (11). Ozone gas
was bubbled through a solution of spirocyle 10 (150 mg,
0.64 mmol) in a solution of ethanol:dichloromethane (1:1,
12 mL) at 2788C, until the solution became pale blue (ca.
5 min). Oxygen gas was then bubbled through the solution
to remove residual ozone then triphenylphosphine (167 mg,
0.63 mmol) was added and the mixture stirred for 2 h at RT.
The solvent was removed in vacuo to give an oil which was
puri®ed by chromatography (2:1 petrol:ether) to yield the
title compound (11) as a colourless oil (99 mg, 66%). Rf 0.32
(1:3 petrol:ether); Accurate mass: Found 239.1283,
C13H19O4 (MH1) requires 239.1283; nmax/cm21 2939m,
2867m, 1752s, 1727s, 1706s, 1373m, 1337m, 1262m,
1194m, 1161m, 1126m, 1033m; dH (500 MHz, CDCl3)
1.28 (3H, t, Jˆ7.0 Hz), 1.65±1.73 (1 H, m), 1.78±1.97
(5 H, m), 2.43±2.59 (6 H, m), 3.66 (1 H, dd, Jˆ7.5,
5.0 Hz), 4.13±4.24 (2H, m); dC (125 MHz, CDCl3) 14.2,
21.8, 26.4, 34.8, 38.4, 40.1, 45.7, 46.2, 55.8, 61.0, 172.9,
210.6, 213.2; m/z (CI) 256 (MNH14 , 60%), 239 (MH1, 100),
193 (17), 138 (27).
(Z)-2-(4-Chlorobut-2-enyl)cyclohexanone. To a stirred
solution of diisopropylamine (3.24 mL, 23 mmol) in THF
(88 mL) at 08C was added n-butyllithium (9.24 mL of a
2.5 M solution in hexanes, 23.1 mmol) followed, after
30 min, by the addition of cyclohexanone (0.95 mL,
22 mmol). After stirring for a further 30 min at 08C Z-4-
chloro-1-iodobut-2-ene (4.93 g, 23 mmol) was added and
the mixture was allowed to warm slowly to RT. The reaction
was quenched by the addition of saturated aqueous ammo-
nium chloride solution (150 mL), the separated aqueous
layer was extracted with ether (3£100 mL) then the
combined organic portions were dried (magnesium sulfate),
®ltered, and concentrated in vacuo to give an orange/brown
oil (5.3 g) that was used directly in the next reaction. nmax
/
cm21 2936m, 2862m, 1710s, 1448m, 1129m, 972m; dH
(400 MHz, CDCl3) 1.31±1.44 (1H, m), 1.53±1.72 (2H,
m), 1.82±1.93 (1H, m), 1.97±2.18 (3H, m), 2.27±2.44
(3H, m), 2.53 (1H, dt, Jˆ14.5, 6.0 Hz), 4.02 (2H, d,
Jˆ7.0 Hz), 5.59±5.68 (1H, m), 5.70±7.80 (1H, m); dC
(125 MHz, CDCl3) 25.0, 27.9, 32.0, 33.5, 42.1, 45.2, 50.2,
127.7, 133.5, 212.2; m/z (CI), 151 (100%).
1-Vinylspiro[2.5]octan-4-one14 (18) and (19). To a stirred
solution of the crude chloride from the previous reaction
(,23 mmol) in THF (50 mL) at RT was added potassium
t-butoxide (2.6 g, 23 mmol) and stirring continued for 6 h.
Saturated aqueous ammonium chloride solution (100 mL)
was added, the separated aqueous layer was extracted with
ether (3£100 mL) and the combined extracts were dried
(magnesium sulfate), ®ltered, and concentrated in vacuo.
The obtained yellow/brown oil was puri®ed by chroma-
tography (40:1 petrol:ether) to give the two diastereomers
18 (1.35 g, 41% from cyclohexanone) and 19 (645 mg, 20%
from cyclohexanone) as colourless oils. Data for 18: Rf 0.28
(9:1 petrol:ether); nmax/cm21 2933m, 1697s, 1634w, 1448m,
1362m, 1118m, 996m, 901m; dH (400 MHz, CDCl3) 0.75±
0.85 (1H, m), 1.35 (1H, dd, Jˆ14.0, 2.5 Hz), 1.64±1.88 (5H,
m), 1.98±2.13 (3H, m), 2.46 (1H, br dd, Jˆ13.5, 3.0 Hz),
4.95 (1H, dt, Jˆ10.0, 1.5 Hz), 5.15 (1H, dt, Jˆ17.0, 1.5 Hz),
5.31 (1H, dddt, Jˆ17.0, 10.0, 8.5, 1.5 Hz); dC (100 MHz,
CDCl3) 17.9, 23.9, 25.4, 35.0, 35.8, 37.4, 41.7, 115.6, 135.6,
208.4; m/z (CI) 168 (MNH14 , 10%), 151 (MH1, 100). Data

8964
J. Robertson et al. / Tetrahedron 56 (2000) 8959±8965
for 19: Rf 0.21 (9:1 petrol:ether); nmax/cm21 3082w, 2940s,
2863m, 1692s, 1635m, 1449m, 1360m, 1136m, 987m,
904m; dH (400 MHz, CDCl3) 0.68 (1H, dd, Jˆ6.5,
4.0 Hz), 1.65 (1H, dd, Jˆ9.0, 4.0 Hz) overlaying 1.60±
1.82 (4H, m), 1.84±1.95 (2H, m), 2.01 (1H, app. q, Jˆ
8.0 Hz), 2.33±2.46 (2H, m), 5.14 (1H, ddd, Jˆ10.5, 1.5,
0.5 Hz), 5.19 (1H, ddd, Jˆ17.0, 1.5, 0.5 Hz), 5.59 (1H,
ddd, Jˆ17.0, 10.5, 8.0 Hz); m/z (CI) 168 (MNH14 , 10%),
151 (MH1, 100).
1-Carboethoxy-2-vinyl-2,3,4,5,6,7-hexahydro-1H-indene
(23). A degassed solution of vinylcyclopropane 20 (40 mg,
0.18 mmol), tributyltin hydride (12.5 ml, 0.045 mmol), and
AIBN (6.0 mg, 0.036 mmol) in benzene (36 mL) was heated
at re¯ux for 14 h. The cooled reaction mixture was concen-
trated in vacuo and the residue puri®ed by chromatography
(20:1 petrol:ether) to give the title compound (23) as a
colourless oil (36 mg, 90%) and as an inseparable mixture
of diastereomers. Rf 0.42 (9:1 petrol:ether); Accurate mass:
Found 221.1537, C14H21O2 (MH1) requires 221.1541; nmax
/
4-Carboethoxymethylene-1-vinylspiro[2.5]octane (20)
and (21). To a suspension of sodium hydride (88 mg of a
60% dispersion in mineral oil, 2.2 mmol) in anhydrous
DME (3 mL) at RT was added triethylphosphonoacetate
(0.46 mL, 2.2 mmol). The mixture was stirred for 1 h then
a solution of ketone 18 (220 mg, 1.46 mmol) in DME
(1 mL) was added at such a rate that the temperature of
the mixture remained below 308C. After a further 14 h,
the reaction mixture was diluted with saturated aqueous
ammonium chloride solution (15 mL), extracted with ether
(3£10 mL), dried (magnesium sulfate), ®ltered, and con-
centrated in vacuo. Puri®cation by chromatography (40:1
petrol:ether) afforded the diastereomeric alkenes 20
(168 mg, 52%) and 21 (59 mg, 19%) as colourless oils.
Data for 20: Rf 0.41 (9:1 petrol:ether); Accurate mass:
Found 221.1542, C14H21O2 (MH1) requires 221.1541;
cm21 2929s, 1734s, 1641w, 1445m, 1368m, 1336m, 1162s,
1035m, 913m; dH (400 MHz, CDCl3) 1.21±1.28 (3H,
2£overlapping t), 1.55±1.68 (4H, m), 1.86±2.05 (4H, m),
2.11±2.18 (0.5H, m), 2.37 (1H, app. br d, Jˆ8.0 Hz), 2.52±
2.58 (0.5H, m), 3.12±3.22 (1.5H, m), 3.41 (0.5H, app. d,
Jˆ8.5 Hz), 4.06±4.20 (2H, m), 4.94±5.10 (2H, m), 5.88
(1H, app. ddd, Jˆ17.0, 10.0, 7.5 Hz); dC (100 MHz,
CDCl3) 14.4 (two peaks), 22.5, 22.7 (two peaks), 22.8,
24.2, 24.6, 25.7 (two peaks), 41.2, 41.6, 44.9, 45.4, 58.3,
59.6, 59.9, 60.3, 113.8, 115.3, 131.4, 132.1, 137.2, 138.3,
139.1, 141.5, 173.3, 174.6; m/z (CI) 238 (MNH14 , 17%), 221
(MH1, 100), 147 (29).
Acknowledgements
n
max/cm21 2981m, 2932s, 2857m, 1716s, 1642s, 1447m,
We thank the Swiss National Science Foundation for a
fellowship (S. A.), the EPSRC and GlaxoWellcome for a
CASE Award (R. K. L.), the EPSRC Mass Spectrometry
Service Centre for accurate mass measurements, and Pro-
fessor Jack Baldwin for helpful discussions.
1372m, 1305m, 1178s, 1038s, 901m; dH (400 MHz,
CDCl3) 0.60 (1H, app. t, Jˆ5.5 Hz), 1.22 (1H, dd, Jˆ8.5,
5.5 Hz), 1.28 (3H, t, Jˆ7.0 Hz), 1.43±1.71 (6H, m), 1.76±
1.83 (1H, m), 2.36±2.43 (1H, m), 3.42 (1H, app. dt, Jˆ13.5,
4.0 Hz), 4.14 (2H, q, Jˆ7.0 Hz), 5.08 (1H, ddd, Jˆ10.0, 2.0,
0.5 Hz), 5.16 (1H, ddd, Jˆ17.0, 2.0, 0.5 Hz), 5.52 (1H, s),
5.62 (1H, ddd, Jˆ17.0, 10.0, 8.5 Hz); dC (100 MHz, CDCl3)
14.3, 18.5, 24.7, 27.1, 29.2, 29.5, 32.3, 33.3, 59.6, 110.9,
115.7, 136.5, 165.7, 167.3; m/z (CI) 238 (MNH14 , 19%), 221
(MH1, 100), 147 (33). Data for 21: Rf 0.36 (9:1 petrol:
ether); nmax/cm21 2981m, 2934s, 2857m, 1720s, 1644s,
1445m, 1228m, 1179s, 1116m, 1037m, 901m; dH (400
MHz, CDCl3) 0.78 (1H, app. t, Jˆ5.5 Hz), 1.09 (1H, dd,
Jˆ8.5, 5.5 Hz), 1.29 (3H, t, Jˆ7.0 Hz), 1.42±1.85 (7H, m),
2.25±2.29 (2H, m), 4.08±4.22 (2H, m), 5.06 (1H, dd,
Jˆ10.5, 2.0 Hz), 5.14 (1H, ddd, Jˆ17.0, 2.0, 0.5 Hz), 5.57
(1H, s), 5.74 (1H, ddd, Jˆ17.0, 10.5, 8.5 Hz); dC (100 MHz,
CDCl3) 14.2, 21.1, 24.8, 29.2, 29.5, 30.9, 32.7, 37.9, 60.0,
112.9, 115.0, 137.1, 162.2, 166.5; m/z (CI) 238 (MNH14 ,
4%), 221 (MH1, 100), 175 (7), 147 (15).
References
1. Shanmugam, P.; Srinivasan, R.; Rajagopalan, K. Tetrahedron
1997, 53, 11685±11692.
2. Giese, B.; Lachhein, S. Angew. Chem., Int. Ed. Engl. 1982, 21,
768±775.
3. (a) Stork, G.; Mook, Jr., R. Tetrahedron Lett. 1986, 27, 4529±
4532; (b) Stork, G.; Mook, Jr., R. J. Am. Chem. Soc. 1987, 109,
2829±2831.
4. Baldwin, J. E.; Cutting, J.; Dupont, W.; Kruse, L.; Silberman,
L.; Thomas, R. C. J. Chem. Soc., Chem. Commun. 1976, 736±738.
5. Dowd, P.; Zhang, W. Chem. Rev. 1993, 93, 2091±2115.
6. Pereyre, M.; Quintard, J. -P.; Rahm, A. Tin in Organic Syn-
thesis; Butterworths: London, 1987, pp 129±134.
7. Janardhanam, S.; Shanmugam, P.; Rajagopalan, K. J. Org.
Chem. 1993, 58, 7782±7788.
2-Carboethoxybicyclo[5.4.0]undeca-1(7),4(5)-diene (22).
Following a procedure analogous to that used for the
preparation of 20 and 21 but starting with diastereomer 19
(150 mg, 1 mmol), the title compound (22) was obtained as
a colourless oil (41 mg, 19%), the mass balance consisting
of recovered starting ketone. Accurate mass: Found
221.1542, C14H21O2 (MH1) requires 221.1541; nmax/cm21
2928s, 1735s, 1439m, 1370m, 1310m, 1195m, 1155s,
1096m, 1027m; dH (400 MHz, CDCl3) 1.27 (3H, t,
Jˆ7.0 Hz), 1.49±1.68 (4H, m), 1.87±2.08 (4H, m), 2.32±
2.49 (3H, m), 3.02 (1H, br. d, Jˆ19.0 Hz), 3.60 (1H, dd,
Jˆ9.5, 3.0 Hz), 4.17 (2H, q, Jˆ7.0 Hz), 5.54±5.64 (2H, m);
dC (100 MHz, CDCl3) 14.3, 22.8 (two peaks), 27.7, 28.4,
32.3, 33.5, 47.2, 60.3, 126.9, 128.1, 129.9, 134.7, 173.9; m/z
(CI) 221 (MH1, 100%), 147 (16).
8. This material seems to originate from hydrostannylation of
diene 13 (Scheme 5) and double protiodestannylation and was
probably generated in this reaction because of an underestimation
of the purity of the tin hydride; it was not observed in later experi-
ments where the radical reaction was run in solution.
9. (a) Janaki, S. N.; Subba Rao, G. S. R. J. Chem. Soc., Perkin
Trans 1 1997, 195±200. (b) Minor, K. P.; Overman, L. E.
Tetrahedron 1997, 53, 8927±8940.
10. A sequence closely related to the one described in Scheme 6 is
described in: Journet, M.; Rouillard, A.; Cai, D.; Larsen, R. D.
J. Org. Chem. 1997, 62, 8630±8631. It is noteworthy that most
5-endo-trig radical cyclisations originate from attack by electro-
philic radicals (e.g. thiyl, a-carbonyl, etc.).
11. Hudlicky, T.; Kutchan, T. M.; Naqvi, S. M. Org. React. 1985,

J. Robertson et al. / Tetrahedron 56 (2000) 8959±8965
33, 247ff. We suggest that catalysis could proceed through addition
8965
of stannyl radical to the carbonyl oxygen, fragmentation, and cycli-
sation of the so-formed allylic radical onto the stannyl enol ether
with
e