Reversible cyclopropane ring-cleavage reactions within etheno-bridged [4.3.1]propelladiene frameworks leading to aza- and oxa-[5.6.5.6] fenestratetraenes

Article abstract of DOI:10.1002/chem.201202903

Opening and closing a chemical window: Oxidation of the etheno-bridged [4.3.1]propelladienol 1 with pyridinium chlorochromate (PCC) affords oxa[5.6.5.6]fenestratetraene 2. The reduction of 2 with diisobutylaluminum hydride (DIBAl-H) leads to the regeneration of its precursor (1; see scheme). These transformations most likely involve a [3,5]-sigmatropic rearrangement process. Copyright

Full text of DOI:10.1002/chem.201202903

COMMUNICATION
DOI: 10.1002/chem.201202903
Reversible Cyclopropane Ring-Cleavage Reactions within Etheno-Bridged
[5.3.1]Propelladiene Frameworks Leading to Aza- and Oxa-
[5.6.5.6]Fenestratetraenes
Nora Heinrich, Anthony C. Willis, Ian A. Cade, Junming Ho, Michelle L. Coote, and
Martin G. Banwell*[a]
In 2008 Takeya and co-workers reported[1] on the isolation
and structural elucidation of the [5.3.1]propellane-containing
and biologically active neoclerodane salvileucalin B (1,
Figure 1). In connection with efforts to probe the origins of
reaction mixture with Rochelleꢀs salt. It was anticipated that
this would provide the corresponding aldehyde. However,
the only products of this reaction were the aza-
AHCTUNGTREG[NNUN 5.6.5.6]fenestratetrene 5 (15%) and its oxa analogue 6 (var-
iable yields), both of which proved to be rather unstable
Figure 1. The natural product, salvileucalin B (1), and related substruc-
tures. Tf=trifluoromethanesulfonyl.
Figure 2. Fenestranes 5, 6, and 7.
compounds (Figure 2). When the reaction mixture was
quenched with aqueous HCl (rather than Rochelleꢀs salt)
then compound 6 (49%) was the exclusive product of reac-
tion. The structure of compound 5 was established by single-
crystal X-ray analysis whereas that of congener 6 followed
from the equivalent analysis of the readily obtained 4-
phenyl-1,2,4-triazoline-3,5-dione (PTAD)[4] adduct 7 (53%),
itself a bridged fenestrane.[5]
The bond angles, as determined by the above-mentioned
X-ray analyses, about the central carbon atoms (C12 in each
case) of compounds 5 and 7 are shown in Table 1. These
data reveal that in each case there is some significant[3] devi-
ation from tetrahedral geometry imposed on the central
carbon atom by the surrounding framework, both in terms
of angle compression (down to 1008) and angle widening
the cytotoxic effects of this natural product, we recently de-
veloped a synthesis of the key substructure 2.[2] Part of the
associated reaction sequence involved reduction of nitrile 3
to primary alcohol 4. This proved to be an unexpectedly
complex process because of the intervention of remarkable
and reversible cyclopropane ring-cleavage processes leading
to the formation (and disassembly) of aza- and oxa-
[5.6.5.6]fenestratetraenes. These compounds represent un-
usual new examples of fenestranes that, as a class, continue
to attract considerable attention because of the capacity
they provide to examine the “plasticity” of the tetrahedral
geometry normally imposed on sp3-hybridized carbon atoms
by virtue of the bonding in such systems.[3] Details are pre-
sented herein.
Our initial attempts to effect the conversion of 3 into 4 in-
volved treating the first compound with DIBAl-H in di-
chloromethane (CH2Cl2) at À408C and then quenching the
Table 1. Bond angles, as determined by X-ray analysis, about the central
carbon atoms (C12) of fenestranes 5 and 7.
Compound 5
Compound 7
[a] N. Heinrich, Dr. A. C. Willis, Dr. I. A. Cade, Dr. J. Ho,
Prof. M. L. Coote, Prof. M. G. Banwell
Research School of Chemistry
Atom array
Bond angle [8]
Atom array
Bond angle [8]
C2-C12-C7
112.7(2)
102.60(19)
104.61(19)
115.1(2)
100.5(2)
121.2(2)
C2-C12-C7
113.5(2)
101.5(2)
107.2(2)
109.4(2)
100.1(2)
126.0(2)
C2-C12-C11
C2-C12-C13
C7-C12-C11
C7-C12-C13
C11-C12-C13
C2-C12-C11
C2-C12-C13
C7-C12-C11
C7-C12-C13
C11-C12-C13
The Australian National University
Canberra, ACT 0200 (Australia)
E-mail: mgb@rsc.anu.edu.au
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202903.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Scheme 2. The interconversion of compounds 4 and 6.
Scheme 1. Possible pathways to fenestratetraenes 5 and 6 from nitrile 3.
in 88% yield. Attempts to effect the DIBAl-H-mediated re-
duction of the readily obtained PTAD adduct, 16 (87%
yield), of nitrile 3 failed, presumably because of competing
reactions involving the carbonyl-containing residues within
the former compound.
(up to 1268) and thus further emphasizing the “plasticity” in
the spatial arrangements of groups attached to an sp3-hybri-
dized carbon atom.
Two possible pathways to fenestratetraenes 5 and 6 are
outlined in Scheme 1. Thus, either the initially formed imine
8 or the corresponding aldehyde 9 could engage in a [3,3]-
sigmatropic rearrangement to deliver the cyclopropane-ring-
cleavage products 10 and 11, respectively. These could, in
turn, participate in either a nonconcerted 1,3-nitrogen or a
nonconcerted 1,3-oxygen shift (allylic rearrangement) to
give compounds 12 and 6, respectively. Compound 12 was
not isolated because of the intervention of a Tf-group trans-
fer reaction leading to the observed product (5) and (pre-
sumably) the vinylogous lactam 13, a compound we have
not been able to isolate from the reaction mixture. Another
possible pathway to compounds 5 and 6 involves the direct
[3,5]-sigmatropic rearrangement of imine 8 and aldehyde 9
to give tetraenes 12 and 6, respectively. Theoretical studies
by Houk and co-workers[6] on the [3,5]-sigmatropic reaction
of the parent all-carbon system suggest that diradical inter-
mediates are likely to be involved. Studies by Kohmoto
et al.[7] also indicate that this type of reaction can occur
within the norcaradiene framework. Our own calculations
(see below) suggest that a [3,5]-sigmatropic process is proba-
bly operating during the course of the transformations de-
scribed herein.
Interestingly, the conversion of 9 into 6 proposed above
seems to occur in a reversible manner. Thus, when a pure
sample of compound 6 was resubjected to reaction with
DIBAl-H in CH2Cl2 at À408C (Scheme 2) and the reaction
mixture quenched with aqueous HCl, then alcohol 4 was ob-
tained in 66% yield. Similarly, oxidation of compound 4
with pyridinium chlorochromate (PCC)[8] in CH2Cl2 at 188C
for 2 hours regenerated, presumably via the initially formed
aldehyde 9, the fenestratetraene 6 in 76% yield (based on
recovered starting material). In contrast, when the readily
obtained PTAD adduct, 14 (76% yield), of alcohol 4 was
oxidized with PCC then the stable aldehyde 15 was obtained
In a further attempt to probe the nature of the conversion
of 9 into 6, the dihydro form of the former compound was
sought on the basis that it could undergo a [3,3]- but not a
[3,5]-sigmatropic rearrangement reaction. To these ends,
compound 4 was subjected to reaction with dihydrogen in
the presence of palladium on carbon (Scheme 3) and the de-
Scheme 3. The synthesis and oxidative rearrangement of [5.3.1]propelle-
nol 17.
sired diene 17 thereby obtained, albeit in low yield (14%)
because of the competing formation of over-reduced and
isomeric materials.[2] Upon treatment of compound 17 with
PCC or the Ley–Griffith reagent—tetrapropylammonium
perruthenate
(TPAP)/N-methylmorpholine
N-oxide
(NMO)[9]—a product containing
a
newly incorporated
oxygen atom (in the form of a ketone) was obtained. On the
basis of the spectroscopic data, we believe alcohol 17 is ini-
tially converted into aldehyde 18 (the sought-after dihydro
form of aldehyde 9) and that this itself undergoes an allylic
oxidation to give enone 19.[10] Finally, compound 19 engages
in a [3,3]-sigmatropic rearrangement to give the observed
product 20 (61%), which itself is a fenestrane.
&
2
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Reversible Cyclopropane Ring-Cleavage Reactions
COMMUNICATION
Figure 3. Gibbs-free-energy profiles for the [3,5]- and [3,3]-sigmatropic rearrangements of [5.3.1]propelladiene 9 in which the OTf group is replaced with
a hydrogen atom (the values in parentheses refer to Me2HAl-complexed analogues whereas * denotes the lowest energy conformer of 9).
Because the above-mentioned experiments failed to dif-
ferentiate between the two possible reaction pathways
shown in Scheme 1, namely between a direct [3,5]-sigma-
tropic rearrangement and a [3,3]-sigmatropic rearrangement
followed by a 1,3-migration process, computational assess-
ments of the energetics of the interconversion of aldehyde 9
and fenestratetraene 6 were carried. These were conducted
the atomic charge on the carbonyl oxygen is 0.024 e more
negative. These characteristics might account for the differ-
ent stabilities of the Me2HAl-complexed transition states.
The transition states detected were all closed-shell singlets
and various attempts to locate either singlet biradical transi-
tion states or intermediates were unsuccessful. This result is
unexpected because the [3,5]-sigmatropic rearrangement is
thermally disallowed and when it does occur, it is thought to
proceed through a stepwise pathway involving biradical in-
termediates.[6] Inspection of the frontier molecular orbitals
(Figure 4) appears to provide an explanation in that the mo-
at a high level of theory [G3ACTHNUGTRNEUNG(MP2)RAD(+) and under
“conditions” reflecting the experimental ones][11] albeit on
compounds in which the OTf group was replaced by a hy-
drogen atom. These computations (Figure 3) revealed that
the ground state energies of the two compounds were com-
parable (9.3 kJmolÀ1 versus 4.9 kJmolÀ1) whereas that of
compound 11 was considerably higher (58.9 kJmolÀ1). A
qualitatively similar situation was observed for the Me2HAl-
complexed forms of compounds 6 and 11, namely the com-
plex of 11 was also of higher energy than that of aldehyde 9.
It should be stressed that while the calculations presented
here indicate that the equilibrium lies in favor of compound
9 (with 6 being 4.9 kJmolÀ1 higher in energy), this is entirely
due to the model chosen where the OTf group was replaced
with a hydrogen atom to reduce the computational cost.
Qualitative assessment at a lower level of theory [B3LYP/6-
31+G(d)] revealed that inclusion of the OTf group leads to
compound 6 being the predominant (thermodynamically
more stable) product, an outcome that is consistent with the
experimental findings.
Figure 4. Frontier molecular orbitals of intermediate 9 (see Scheme 1) ob-
tained at the B3LYP/6-31G(d) level of theory.
lecular geometry facilitates overlap between the HOMO
(the p orbital of the diene) and the p* orbital of the carbon-
yl group (CO p*). This overlap results in an activation barri-
er of approximately 100 kJmolÀ1. This is relatively low in
comparison to the values that are typical for thermally al-
lowed hydrocarbon-based pericyclic reactions, which have
activation barriers of approximately 140 kJmolÀ1.[6]
Under the “reaction conditions” used in the computation-
al work (À408C), the calculated barriers associated with
direct conversion of both 9 into 6 and 9 into 11 translate
The transition states for the conversion of 9 into 6 and the
conversion of 9 into 11 were calculated to be of similar ener-
gies (101.6 kJmolÀ1 versus 103.4 kJmolÀ1) but very signifi-
cant differences were encountered for the corresponding
Me2HAl-complexed compounds (58.2 kJmolÀ1 versus
84.4 kJmolÀ1). The calculated natural-bond-order (NBO)
atomic charges and solvation free energies indicate a signifi-
cantly more polarized transition state in which, for example,
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

M. G. Banwell et al.
Kramps, J. Am. Chem. Soc. 2006, 128, 11620, and references cited
therein; c) R. Keese, Chem. Rev. 2006, 106, 4787; d) H. Hopf, Clas-
sics in Hydrocarbon Chemistry, Wiley-VCH, Weinheim, 2000,
pp. 88–94.
into half-lives of greater than 260 years. By comparison, the
corresponding barriers associated with the Me2HAl-com-
plexed compounds confer half-lives that are consistent with
the time-scale of the actual experiments. On this basis, acid-
catalyzed [3,5]-sigmatropic rearrangements are believed to
be operative in the conversions of 3 into 6, 4 into 6,[12] and 6
into 4.
The above-mentioned transformations represent new and
unconventional means of generating fenestranes and suggest
that oxa- and aza-based variants of the [3,5]-sigmatropic re-
arrangement could be used for assembling complex hetero-
cyclic frameworks. In light of these results, it is conceivable
that the “retro-Claisen” rearrangement product observed
[4] R. C. Cookson, S. S. Gupte, I. D. R. Stevens, C. T. Watts, Org. Synth.
1971, 51, 121.
[5] Crystallographic data for compound 5: C15H11F6NO5S2, M=463.38,
¯
T=200 K, triclinic, space group P1, Z=2, a=6.8778(2), b=
10.8844(4), c=12.8790(4) ꢂ, a=99.1558(19)8, b=102.708(2)8, g=
104.325(2)8; V=887.71(5) ꢂ3, Dx =1.733 gcmÀ3, 4063 unique data
(2qmax =55.28), R=0.051 [for 3003 reflections with I> 2.0s(I)]; Rw=
0.132 (all data), S=0.95. Crystallographic data for compound 7:
¯
C22H16F3N3O6S, M=507.45, T=200 K, triclinic, space group P1, Z=
2, a=6.6517(2), b=11.4024(6), c=14.7561(8) ꢂ, a=79.894(2)8, b=
83.096(3)8, g=78.323(3)8; V=1074.84(9) ꢂ3, Dx =1.568 gcmÀ3, 4921
unique data (2qmax =558), R=0.060 [for 3027 reflections with I>
2.0s(I)]; Rw=0.153 (all data), S=0.97. CCDC-891377 (5) and
891378 (7) contain the supplementary crystallographic data for this
paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_
request/cif
during the course of the recently reported total synthesis of
[13]
salvileucalin B (1)
may result from the operation of a
[3,5]-sigmatropic rather than a [3,3]-sigmatropic process.
[6] A. G. Leach, S. Catak, K. N. Houk, Chem. Eur. J. 2002, 8, 1290.
[7] S. Kohmoto, N. Nakayama, J.-i. Takami, K. Kishikawa, M. Yamamo-
to, K. Yamada, Tetrahedron Lett. 1996, 37, 7761, and references
cited therein.
[8] E. J. Corey, J. W. Suggs, Tetrahedron Lett. 1975, 16, 2647.
[9] S. V. Ley, J. Norman, W. P. Griffith, S. P. Marsden, Synthesis 1994,
639.
Acknowledgements
We thank the Australian Research Council, the Institute of Advanced
Studies, and the Australian National Computational Infrastructure for
generous financial support. The important observations of Josh Buckler
are gratefully acknowledged.
[10] The possibility that the allylic oxidation step takes place first cannot
be discounted.
[11] Gas phase Gibbs free energies were calculated with G3
RAD(+)//B3LYP/6-31+G[d]. Solvation free energies were calculat-
ed using the SMD model at the M06–2X/6-31+G(d,p) level of
ACHTUNGTRENNUNG(MP2)-
Keywords: cleavage reactions
norcaradienes · salvileucalin B · sigmatropic rearrangement
·
fenestratetraenes
·
AHCTUNGTRENNUNG
theory in CH2Cl2 on solution-phase optimized geometry; the solva-
tion contribution to the reaction energies (at 233 K) was approxi-
mated from those obtained at 298 K.
[1] Y. Aoyagi, A. Yamazaki, C. Nakatsugawa, H. Fukaya, K. Takeya, S.
Kawauchi, H. Izumi, Org. Lett. 2008, 10, 4429.
[2] N. Heinrich, M. G. Banwell, A. C. Willis, I. A. Cade, R. J. Capon,
X.-C. Huang, Aust. J. Chem., in press.
[3] For useful points of entry into the literature on fenestranes, see:
a) M. Charpenay, A. Boudhar, G. Blond, J. Suffert, Angew. Chem.
2012, 124, 4455; Angew. Chem. Int. Ed. 2012, 51, 4379, and referen-
ces cited therein; b) S. E. Denmark, J. I. Montgomery, L. A.
[12] In the case of this rearrangement, we assume the acidic nature of
the reaction medium resulting from the presence of PCC facilitates
the conversion of aldehyde 9 into oxa-[5.6.5.6]fenestratetraene 6.
[13] S. Levin, R. R. Nani, S. E. Reisman, J. Am. Chem. Soc. 2011, 133,
774.
Received: August 14, 2012
Published online: && &&, 0000
&
4
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Reversible Cyclopropane Ring-Cleavage Reactions
COMMUNICATION
Opening and closing a chemical
window: Oxidation of the etheno-
bridged [5.3.1]propelladienol 1 with
pyridinium chlorochromate (PCC)
affords oxa-[5.6.5.6]fenestratetraene 2.
The reduction of 6 with diisobutylalu-
minum hydride (DIBAl-H) leads to
the regeneration of its precursor (1;
see scheme). These transformations
most likely involve a [3,5]-sigmatropic
rearrangement process.
Sigmatropic Rearrangements
N. Heinrich, A. C. Willis, I. A. Cade,
J. Ho, M. L. Coote,
M. G. Banwell* ............... &&&& — &&&&
Reversible Cyclopropane Ring-
Cleavage Reactions within Etheno-
Bridged [5.3.1]Propelladiene Frame-
works Leading to Aza- and Oxa-
[5.6.5.6]Fenestratetraenes
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!