Prying open a Thiele cage: Discovery of an unprecedented extended pinacol rearrangement

Article abstract of DOI:10.1039/c8cc08862d

The first extended pinacol rearrangement across an sp³-sp³ bond is reported. The reaction appears to be both stereo- and regio-specific, and results in an extremely rare example of cage opening for a 1,3-bishomocubane structure deriv

Full text of DOI:10.1039/c8cc08862d

ChemComm
View Article Online
View Journal
COMMUNICATION
Prying open a Thiele cage: discovery of an
unprecedented extended pinacol rearrangement†
Cite this:DOI: 10.1039/c8cc08862d
a
Nathan Dao, ‡a Jonathan K. Sader,‡a Allen G. Oliverb and Jeremy E. Wulff
*
Received 7th November 2018,
Accepted 10th January 2019
DOI: 10.1039/c8cc08862d
rsc.li/chemcomm
The first extended pinacol rearrangement across an sp3–sp3 bond is
reported. The reaction appears to be both stereo- and regio-
specific, and results in an extremely rare example of cage opening
for a 1,3-bishomocubane structure derived from Thiele’s ester.
Thiele’s acid and ester are Diels–Alder heterodimers resulting from
equilibrating mixtures of carboxylated cyclopentadiene (Fig. 1A).1
While in principle up to 72 different Diels–Alder adducts could
be produced from such a reaction, in practice only a single major
product predominates in both the acid and ester cases.2,3 The
rigid molecular architecture of Thiele’s acid and derivatives
thereof suggests as-yet unrealized opportunities in the synthesis
of conformationally restricted drug candidates for biological
screening applications.4,5
Fig. 1 Origin of Thiele’s acid and ester, and a series of conformationally
tuned scaffolds derived from them. Blue arrows indicate angles of projec-
tion for the carboxylic acid or ester substituents, and numbers in blue
indicate cleft angles calculated as in ref. 6.
Our group recently described an improved synthesis of
Thiele’s ester,6 established the first chiral resolution of the parent
diacid,7 and developed a suite of conformationally tuned molecular
clefts based upon the Thiele framework (Fig. 1B).6 While these shall refer to here as a Thiele cage (Fig. 2A).8 Marchand and others
products have the appropriate structural characteristics for use in have exploited this reaction to make a wide variety of structures
supramolecular applications or perhaps as b-turn mimics, they all incorporating the 1,3-bishomocubane motif,4,5,9–12 but here too,
display essentially the same core structure, and therefore lack the deep-seated structural changes for cages derived from Thiele’s acid
deep-seated scaffold alterations that would be desirable for the or ester are exceedingly rare.13
development of a screening library.
Although Thiele’s acid and ester have been known for over a
century, relatively little reaction chemistry has been developed
around them. Nevertheless, Dunn and Donohue showed many
years ago that exposure of Thiele’s ester to ultraviolet light results in
an efficient intramolecular [2+2] reaction that transforms the ‘open’
Thiele’s ester into a ‘closed’ 1,3-bishomocubane structure that we
a Department of Chemistry, University of Victoria, PO Box 3065 STN CSC,
Victoria, British Columbia, V8W 3V6, Canada. E-mail: wulff@uvic.ca
b Molecular Structure Facility, Department of Chemistry and Biochemistry,
University of Notre Dame, 251 Nieuwland Science Hall, Notre Dame,
IN 46556, USA
† Electronic supplementary information (ESI) available: Fig. S1 and S2, full experimental
Fig. 2 Generation of a 1,3-bishomocubane cage structure from Thiele’s
ester, and representative structural rearrangements for keto- and acetoxy-
substituted 1,3-bishomocubanes.
details, spectroscopic data for all new compounds. CCDC 1871599 (3). For ESI and
crystallographic data in CIF or other electronic format see DOI: 10.1039/c8cc08862d
‡ These authors contributed equally to the preparation of the manuscript.
This journal is ©The Royal Society of Chemistry 2019
Chem. Commun.

View Article Online
Communication
ChemComm
Containing two cyclobutane units, 1,3-bishomocubanes are
strained systems, and in fact several 1,3-bishomocubane derivatives
have been investigated as explosives and high energy density
fuels.12,14 More functionalized bishomocubanes—especially those
incorporating ketone or acetoxy groups within the pentacyclo-
[5.3.0.02,5.03,9.04,8]decane core—are known to undergo rearrange-
ment reactions leading to partially open seco-cage structures.15–17
Although such rearrangements have not previously been shown for
Thiele’s acid derivatives (which necessarily lack functionality within
the 1,3-bishomocubane core), we hypothesized that a selective
ring-opening reaction could be identified, which would result in
deep-seated changes to the Thiele cage core structure.18,19
Here we report the results of our study aimed at the
identification of such a reaction, culminating in the discovery
of an unprecedented extended pinacol rearrangement that opens
Scheme 2 Synthesis of diol substrates, and observation of a new extended
pinacol rearrangement reaction. Red arrows indicate NOE interactions.
the Thiele cage. The reaction proceeds with perfect control over (2a–2d). The conversion in each case was high although purifica-
both regiochemistry and stereochemistry, and results in a sub- tion challenges contributed to modest isolated yields for three of
stantial change to the three-dimensional structure of the overall the four products.
scaffold architecture.
We next attempted thermal or photochemical ring opening
Operating under a hypothesis that increasing steric strain might for cages 2a–2d. Photochemical irradiation at either 254, 300,
favour ring opening of the Thiele cage, we began by preparing or 350 nm in benzene-d6 or methanol-d4 led to no detectable
substrates of increasing steric bulk (1a–1d, Scheme 1). Thiele’s reaction, and to recovery of starting material for all four
methyl ester (1b) and benzyl ester (1c) were prepared from sodium cages.21 Thermal activation was studied at 50 1C and 100 1C,
cyclopentadienylide and the appropriate carbonate, following our in a variety of solvents (toluene, xylene, chloroform, ethyl
previously developed route.6 The methyl ester was saponified (using acetate, ethanol, acetic acid, or dimethyl sulfoxide). Once again,
a hindered solvent to avoid unwanted conjugate addition) to afford however, no detectable ring opening was observed. In each case
Thiele’s acid (1a). Accessing the corresponding adamantyl ester (1d) the substrate was recovered unchanged.
was more challenging, but we were ultimately able to achieve this
substrate (albeit in low yield) through the use of Lebel’s amine system, we next reacted ester 2b with phenyl lithium to afford
diazotization protocol.20 the very hindered diol 3 (Scheme 2).22 The constrained nature
In an effort to further increase the steric demands for the
Compounds 1a–1d were irradiated at 254 nm. Each of the sub- of this molecule is highlighted by the fact that the b-proton at
strates cyclized efficiently to the corresponding 1,3-bishomocubane C10 of the pentacyclo[5.3.0.02,5.03,9.04,8]decane core appears at
1
À0.64 ppm in the H NMR spectrum—evidence of substantial
magnetic shielding by the phenyl groups on C11.
Diol 3 also did not react under photochemical conditions, or
upon heating in aprotic solvents. Gratifyingly, however, when
the compound was incubated in acidic conditions (neat AcOH
at 100 1C or 2.5 equiv. of H2SO4 in various solvents) we observed
rearrangement to a single major product. The less hindered
analogue 6 did not undergo a similar reaction, affording only
acetylated derivative 7 upon heating in acetic acid, or recovered
starting material when exposed to H2SO4.
Isolation and characterization of the rearrangement product
from 3 immediately revealed that the C2–C5 bond of the cage
had been selectively cleaved, and that the two diphenylmethanol
substituents had been exchanged for a diphenyl alkylidene
(13C NMR: d 136 and 144 ppm) and a phenyl ketone (13C NMR:
d 198 ppm; IR n 1661 cmÀ1). These observations (together with
additional spectroscopic and MS data; refer to the ESI,† for full
details) were sufficient to identify the product as arising from an
extended pinacol rearrangement, from which two different
regioisomeric products could be envisioned (4 and 5). For each
regioisomer, two diastereomers were considered (i.e. 4a and 4b;
5a and 5b; structures shown in Scheme 3).
Detailed analysis of the 1H, 13C, COSY, DEPT-135, HSQC and
Scheme 1 Synthesis of Thiele acid and ester substrates of increasing
steric demand, and photocyclization to the corresponding cage structures.
HMBC data – together with comparison of 1H and 13C chemical
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2019

View Article Online
ChemComm
Communication
We briefly investigated this mechanism computationally, but
could find little purchase to support it on either thermodynamic or
kinetic grounds. Compound 4a is not the lowest in energy of the
four potential rearrangement products shown in Scheme 3A, and
there doesn’t appear to be sufficient difference between the
energies of C and D to account for the high degree of regiochemical
control observed in the reaction. Moreover, the geometry-optimized
structure of intermediate D provides no indication for why the
preferred trajectory should be the one leading to 4a over 4b.
With the stepwise mechanism thus unable to explain the
regio- and stereochemical outcome, we considered the possibility of
a concerted reaction. Because the two alcohol groups are held near
one another, we reasoned that protonation of either alcohol could
be stabilized through hydrogen bonding to the other oxygen atom,
akin to the situation with proton sponge base. The existence of this
O–H–O hydrogen bond places an additional constraint on the
system, limiting the number of accessible conformers.
DFT calculations on this equilibrating mixture (F and G)
suggested the existence of a low-energy conformer wherein one
of the two phenyl groups on C11 is held antiperiplanar to the
C2–C5 bond (see Fig. S1, ESI† for optimized structure). At the
Scheme 3 Proposed mechanisms for the extended pinacol rearrange-
ment. Bonds shown in blue are approximately coplanar.
shifts to estimated resonances for each structure (see ESI†) same time, the C12–O bond is approximately synperiplanar,
allowed us to rule out regioisomer 5 and unambiguously something that presumably would facilitate the movement of
confirm the connectivity shown in structure 4 as corresponding electrons as shown in Scheme 3B. Reaction via this conformer
to the isolated product.
would lead exclusively to species 4a.
Establishing the relative stereochemistry for the molecule was
Seeking additional evidence, we obtained an X-ray structure
more difficult, and required us to first use 1D TOCSY NMR to resolve of precursor 3 (Fig. S2, ESI†). This revealed that even in the
the overlapping aromatic signals from the 1H NMR spectrum. This unprotonated form, the molecule is pre-organized to undergo
revealed that the ortho protons on the four aromatic rings (i.e. H14/ rearrangement, with the migrating phenyl group (C13–C18)
18, H20/24, H26/30, and H32/36 using the numbering system in held anti to the C2–C5 bond (y = 1771), and with the C2–C5
Scheme 2) were all sufficiently well resolved in the 1H NMR to permit bond itself exhibiting a significantly longer distance (1.66 Å)
cross-relaxation experiments. A series of 1D NOE spectra showed than the other six cyclobutane bonds in the structure (1.56 Æ
that the ortho protons on the phenyl ketone (H20/24) were in close 0.01 Å)—a feature that was also predicted in our calculated
proximity to the ortho protons of each of the other aromatic rings. structure of the putative reactive conformer (Table S1, ESI†).
This is only possible for the 1R, 2S, 3S, 4R, 7S, 8R, 9S diastereomer, This structure also explained the dramatic upfield shift for the
which allowed us to conclusively establish the product from the C10 b-proton (vide infra), since the latter is positioned directly
rearrangement as 4a.
in the p system of the C13–C18 phenyl group, which is notably
In seeking to understand the reaction from a mechanistic distorted from planarity; the observation of this shift in the
perspective, we first considered the stepwise pathway described solution-state NMR thus supports the relevance of the solid
in Scheme 3A. In this putative mechanism, protonation of state structure in interpreting solution-state reactivity.
either of the two alcohol groups would lead to loss of water
These data would be consistent with a concerted extended
and formation of a pair of regioisomeric carbocation inter- pinacol pathway, but might also serve to rationalize the stereo-
mediates: A and B. Which intermediate forms first would chemical outcome of a stepwise reaction, if the intermediate
probably be inconsequential in this case, since the two species carbocation D is sufficiently short-lived as to facilitate the
could interconvert, either via tetrahydrofuran C or through migration of the phenyl group before structural relaxation
recombination with water to regenerate the starting material. occurs. Further mechanistic study will be required to distinguish
Fragmentation of the C2–C5 cyclobutane bond in either A or B between the two related possibilities.
would be accompanied by generation of an olefin, affording a
Pinacol (or semipinacol) rearrangements are among the most
new pair of regioisomeric carbocation intermediates: D and E. well-known structural reorganization reactions in organic chemistry,
The preferential formation of either D or E would presumably and have been extensively exploited in the synthesis of complex
dictate the regiochemical outcome for the reaction, since these two natural products and other targets.23 p-Extended pinacol rearrange-
species are less likely to interconvert prior to the semipinacol ments (i.e. 8 - 9, Scheme 4) are likewise well-studied,24 and an
rearrangement taking place. The last step—1,2-migration of the enantioselective variant has been described.24d A single example of
benzene ring—should be essentially irreversible and will dictate an analogous reaction across a cyclopropane ring (i.e. 10 - 11) was
the stereochemical outcome for the overall process, by establishing reported,25 but since the sp5 carbon atoms in the cyclopropane
the new stereogenic centre at C2.
contribute significant p character to the bonds, this does not
This journal is ©The Royal Society of Chemistry 2019
Chem. Commun.

View Article Online
Communication
ChemComm
Conflicts of interest
There are no conflicts to declare.
Notes and references
1 J. Thiele, Chem. Ber., 1901, 34, 68.
2 (a) I. Fleming, Frontier Orbitals and Organic Chemical Reactions,
John Wiley & Sons, Chichester, 1976, pp. 136, 167–168; (b) I. Fleming,
Molecular Orbitals and Organic Chemical Reactions, Reference Edition, John
Wiley & Sons, Chichester, 2010, pp. 321–322; (c) G. Deslongchamps and
P. Deslongchamps, Tetrahedron, 2013, 69, 6022; (d) J. Chen and
J. E. Wulff, Org. Biomol. Chem., 2016, 14, 10170; (e) J. Chen, L. Lu and
J. E. Wulff, Synlett, 2017, 28, 2777.
Scheme 4 Known, unknown, and novel extended pinacol rearrangements.
3 A. P. Marchand, I. N. N. Namboothiri, S. B. Lewis, W. H. Watson and
M. Krawiec, Tetrahedron, 1998, 54, 12691.
4 O. K. Onajole, S. Sosibo, P. Govender, T. Govender, P. D. van Helden,
constitute a reaction across a true alkyl bond. Indeed, to the best of
our knowledge, there are no prior examples of extended pinacol
rearrangements occurring across sp3C–sp3C bonds.
´
G. E. M. Maguire, K. Mlinaric-Majerski, I. Wiid and H. G. Kruger,
Chem. Biol. Drug Des., 2011, 78, 1022.
5 A. S. Sklyarova, V. N. Rodionov, C. G. Parsons, G. Quack, P. R. Schreiner
and A. A. Fokin, Med. Chem. Res., 2013, 22, 360–366.
6 J. Chen, B. Kilpatrick, A. G. Oliver and J. E. Wulff, J. Org. Chem.,
2015, 80, 8979.
7 J. Chen, X. Sun, A. G. Oliver and J. E. Wulff, Can. J. Chem., 2017, 95, 234.
8 G. L. Dunn and J. K. Donohue, Tetrahedron Lett., 1968, 9, 3485.
9 A. P. Marchand, I. N. N. Namboothiri, B. Ganguly, W. H. Watson and
S. G. Bodige, Tetrahedron Lett., 1999, 40, 5105.
10 A. P. Marchand, H. K. Hariprakasha and I. N. N. Namboothiri, Synth.
Commun., 2001, 31, 1863.
11 S. G. Bodige, W. H. Watson, A. P. Marchand and V. S. Kumar,
J. Chem. Crystallogr., 1999, 29, 1261.
12 (a) S. Rajkumar, R. S. Choudhari, A. Chowdhury and I. N. N.
Namboothiri, Thermochim. Acta, 2013, 563, 38; (b) S. Lal, L. Mallick,
S. Rajkumar, O. P. Oommen, S. Reshmi, N. Kumbhakarna, A. Chowdhury
and I. N. N. Namboothiri, J. Mater. Chem. A, 2015, 3, 22118.
13 A. P. Marchand, D. Zhao, T.-K. Ngooi and V. Vidyasagar, Tetra-
hedron, 1993, 49, 2613.
14 L. Mallick, S. Lal, S. Reshmi, I. N. N. Namboothiri, A. Chowdhury
and N. Kumbhakarna, New J. Chem., 2017, 41, 920.
15 (a) A. J. H. Klunder, W. C. G. M. de Valk, J. M. J. Verlaak, J. W. M.
Schellekens, J. H. Noordik, V. Parthasarathi and B. Zwanenburg, Tetra-
hedron, 1985, 41, 963; (b) P. M. Ivanov, E. Osawa, A. J. H. Klunder and
B. Zwanenburg, Tetrahedron, 1985, 41, 975.
16 T. Ogino, K. Awano and Y. Fukazawa, J. Chem. Soc., Perkin Trans. 2,
1990, 1735.
17 T. Nigo, T. Hasegawa, Y. Kuwatani and I. Ueda, Bull. Chem. Soc. Jpn.,
1993, 66, 2068.
18 Simple reductions of the cyclobutane bonds in Thiele’s ester and
related 1,3-bishomocubanes are known, using either dissolving
metal reductions (ref. 13) or hydrogenation (ref. 19). However, we
opted not to pursue this strategy here.
19 K. Hirao, T. Iwakuma, M. Taniguchi, O. Yonemitsu, T. Date and
K. Kotera, J. Chem. Soc., Perkin Trans. 1, 1980, 163.
20 C. Audubert and H. Lebel, Org. Lett., 2017, 19, 4407.
21 The broad spectrum n - p transition associated with carbonyl
groups is known to facilitate activation with unfiltered light centred
on 254 nm; see: Y. Liu, X. Lan, S. Gao, Z. Shen, J. Lu and X. Ni, Proc.
SPIE, 2003, 5254, 526.
22 I. N. N. Namboothiri and O. P. Oommen, Indian Pat., Appl. IN
2008MU00098 A 20091002, 2009.
23 Z.-L. Song, C.-A. Fan and Y.-Q. Tu, Chem. Rev., 2011, 111, 7523.
24 (a) R. E. Lutz, R. G. Bass and D. W. Boykin, J. Org. Chem., 1964,
29, 3660; (b) K. Saito, Y. Horie, T. Mukai and T. Toda, Bull. Chem.
Soc. Jpn., 1985, 58, 3118; (c) L.-L. Zhu, X.-X. Li, W. Zhou, X. Li and
Z. Chen, J. Org. Chem., 2011, 76, 8814; (d) H. Wu, Q. Wang and
J. Zhu, Angew. Chem., 2016, 55, 15411.
To gain insight into the scope and limitations of the reaction,
we prepared and tested three additional substrates: 14, 16, and 18
(Scheme 4). As with 11 alcohol 6, the simple 31 alcohol 14 (contain-
ing four methyl groups) failed to produce a product analogous to
4a. Likewise the very electron-poor 18 (containing four CF3 groups
at the para-positions of the aromatic rings) did not react, even at
increased temperatures. However, the more electron-rich substrate
16 (containing four p-methoxy groups) efficiently rearranged to
afford ketone 17. These data would be consistent with carbocation
character in the rate-determining step, but further study will be
required to elucidate the precise mechanistic details.
In summary, the extended pinacol reaction described here
represents a novel ring opening reaction for the Thiele bishomo-
cubane, and constitutes one of the first selective cage openings for
any pentacyclo[5.3.0.02,5.03,9.04,8]decane scaffold derived from the
venerable Thiele acid scaffold. We envision that pairs of tetraary-
lated molecules before and after cage opening (i.e. derivatives of 3
and 4a) will be interesting to evaluate for their biological properties,
and plan to undertake such a study in the future. In addition to
switching the hydrogen-bond donor and acceptor groups on the
compounds’ surface, the rearrangement reaction results in a sub-
stantial change in the shape of the rigid backbone, altering the
projection vectors for the substituents. These changes will pro-
foundly perturb interactions with protein targets.
Of perhaps even greater importance, the synthesis of 4a
represents a significant expansion of known pinacol-type pro-
cesses. While extended pinacol reactions across p systems and
cyclopropanes are described elsewhere, we believe this to be the
first report of such a rearrangement occurring across an alkyl
C–C bond. The apparent regio- and stereo-specificity indicates a
constrained transition state in which an intramolecular hydro-
gen bond may serve to position a distal migrating group to
control the reaction trajectory.
We acknowledge NSERC Canada for operating funds, and
for a fellowship to J. S. We also thank the Canada Research
Chairs Program, the Michael Smith Foundation for Health
Research, and the University of Victoria for ongoing support.
¯
25 R. A. Darby and R. E. Lutz, J. Org. Chem., 1957, 22, 1353.
Chem. Commun.
This journal is ©The Royal Society of Chemistry 2019