A curious benzenoid deacylation reaction: Neighbouring group participation in acylhydroxy[2.2]paracyclophanes

Article abstract of DOI:10.1016/S0040-4039(01)01607-0

Attempted carbonyl reduction of acyldihydroxy[2.2]paracyclophanes with sodium borohydride led instead to competing aromatic deacylation. The effect was traced to a neighbouring group participation that can be correlated with carbonyl IR stretching frequencies. Deuterium labelling studies implicate the formal intermediacy of a solvated [2.2]paracyclophane anion via an S_E1 mechanism.

Full text of DOI:10.1016/S0040-4039(01)01607-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 7527–7529
A curious benzenoid deacylation reaction: neighbouring group
participation in acylhydroxy[2.2]paracyclophanes
D. Christopher Braddock,a,* Iain D. MacGilpb and Benjamin G. Perrya
aDepartment of Chemistry, Imperial College of Science Technology and Medicine, South Kensington, London SW 7 2AY, UK
bProcess Chemistry, GlaxoSmithKline R&D, Temple Hill, Dartford, Kent DA1 5AH, UK
Received 20 July 2001; accepted 24 August 2001
Abstract—Attempted carbonyl reduction of acyldihydroxy[2.2]paracyclophanes with sodium borohydride led instead to competing
aromatic deacylation. The effect was traced to a neighbouring group participation that can be correlated with carbonyl IR
stretching frequencies. Deuterium labelling studies implicate the formal intermediacy of a solvated [2.2]paracyclophane anion via
an SE1 mechanism. © 2001 Elsevier Science Ltd. All rights reserved.
During the course of our studies we had reason to
prepare diacyldihydroxy[2.2]paracyclophane 1† and
O
OH
OH
OH
attempt its direct reduction to bismethylene[2.2]-
paracyclophane 2. For this reduction we were attracted
by the method of Bell who has shown that the use of
sodium borohydride in refluxing aqueous sodium
hydroxide solution cleanly reduces o-hydroxyaryl-
ketone 3 to methylene compound 4 via the presumed
intermediacy of an exomethylene quinone.1 However,
we were surprised to discover that treatment of
analogously substituted diacyldiol 1 under these condi-
tions resulted in the predominant formation of diol 52
bearing no acyl groupings.
OH
O
1
2
OH
O
OH
3
4
This result was intriguing since deacylation of aromat-
ics is rare,3 especially under basic conditions which
require a strong base such as potassium t-butoxide,4
and typically only substrates containing non-enolisable
protons are suitable.5 We therefore decided to investi-
gate this system further. At the outset, we suspected a
neighbouring group participation (NGP) by either (or
both) of the ortho and pseudo-gem hydroxyl
functionalities.
OH
OH
5
NaBH4 in refluxing MeOH only.‡ Under these condi-
tions diacyldihydroxy[2.2]paracyclophane 1 was con-
sumed to again provide diol 5 (isolated yield 20%) but
the major component was monodeacylolefin 6 (65%
For further experiments we moved away from the
aqueous system described by Bell and instead employed
‡ Representative procedure: To a stirred solution of 1 in MeOH
(0.035 M) was added NaBH4 (25 equiv.). The solution was heated
to reflux for 14 h, allowed to cool to rt and treated with an excess
of aqueous HCl (2.5 M). The mixture was extracted with EtOAc
(2×20 mL) and the combined organic layers were washed with water
(1×20 mL), saturated brine solution (1×20 mL), dried (MgSO4),
evaporated and chromatographed (CH2Cl2 through to 2% MeOH
in CH2Cl2), to give 6 (65%) as a colourless oil and 5 (20%) as a
yellow solid.
Keywords: [2.2]paracyclophane; deacylation; neighbouring group par-
ticipation.
* Corresponding author. E-mail: c.braddock@ic.ac.uk
† All the [2.2]paracyclophanes described are racemic modifications
and have been satisfactorily characterised.
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)01607-0

7528
D. C. Braddock et al. / Tetrahedron Letters 42 (2001) 7527–7529
isolated yield). This compound was clearly evident by
the appearance of a doublet triplet at 6.66 ppm (J=
16.3, 1.5 Hz) and a double triplet at 5.65 ppm (J=16.3,
6.9 Hz) in the 1H NMR spectrum for the a- and
b-styryl protons, respectively. Presumably, this com-
pound is formed by deacylation of one acyl grouping
followed by reduction of the other ketone to the alco-
hol and elimination during the acidic work-up.
OH
O
12
13
OR
OMe
OR'
8; R = R' = H
OH
9; R = R' = Me
10; R = H, R' = Me
11; R = Me, R' = H
OMe
O
OMe
OH
OH
OMe
The results from these experiments suggest that both an
ortho and a pseudo-gem hydroxyl group on the para-
cyclophane ring are required to be present to direct this
novel deacylation reaction. This can be further vali-
dated by inspection of the carbonyl stretching frequen-
cies in the infrared spectra of the various
paracyclophane compounds 8–11. For bis-methoxy
paracyclophane 9, the IR spectrum displays an un-
remarkable stretch at 1682 wavenumbers, characteristic
of an aryl ketone. Interestingly, the effect of a free
hydroxyl group pseudo-gem to the carbonyl in para-
cyclophane 10 shifts the stretching frequency to 1668
cm−1. Intramolecular hydrogen bonding in orthogonally
protected methoxyparacyclophane 11 leads to a stretch
at 1606 cm−1, and in the dihydroxyparacyclophane 8
the carbonyl stretch is observed at 1600 wavenumbers.
The weakening of a carbonyl double bond by an elec-
tron rich substituent in a pseudo-gem position in
[2.2]paracyclophane chemistry (via donation of a lone
pair into the carbonyl p* orbital) has been previously
noted,6 but to date there are no examples of a hydroxyl
group participating in this manner.
O
6
7
In order to explore the notion that the hydroxyl groups
were intimately involved in this deacylation reaction, a
protected diol substrate was prepared. Exposure of
diacyldimethoxy[2.2]paracyclophane 7 to the standard
conditions resulted in no reaction and the starting
material was recovered intact. This is a somewhat
startling observation but simple steric factors may be
responsible for the lack of reactivity. In the absence of
any NGP effects, normal reactivity would result in
alcohol formation. In this case it is apparent that
approach of the borohydride would be from the least
hindered side of the paracyclophane, leading to strong
steric compression in the transition state as the sub-
stituents on one paracyclophane ring move towards
those on the other.
To separate the effect of the positioning of the hydroxyl
groups relative to the ketone functionality,
monoacyl[2.2]paracyclophanes 8–11 were prepared.
Non-protected acyl compound 8 smoothly deacylated
under the reaction conditions to give diol 5 as the
major component (73%) along with some alkene 6
(18%). The diprotected compound 9 was found to be
unreactive, again presumably due to steric factors as
per dimethoxyparacyclophane 7. For the monopro-
tected systems, compound 10 underwent reduction to
alkene 12 (49%), unreacted starting material was recov-
ered (25%) but no deacylparacyclophane 13 could be
detected. For orthogonally protected paracyclophane
11 a complex reaction mixture was produced but no
deacyl compound 13 was detected.
The mechanism of the deacylation was further probed
by a series of deuterium labelling experiments. Mono-
acylparacyclophane 8 was exposed to the standard condi-
tions, but using either sodium borodeuteride or MeOD
in lieu of their non-deuterated reagents. When 8 was
allowed to react with sodium borodeuteride in MeOH,
diol 5 was isolated (50%) with no-deuterium incorpora-
tion, along with monodeuterated alkene 14 (30%; >90%
D-incorporation).§ Conversely, in the reaction of 8 with
NaBH4 in MeOD, which broadly gave the same
product distribution, no deuterium incorporation was
observed in the produced alkene 6, and deuterium
incorporation at the ortho position in paracyclophane
15 was observed.
D
OH
D
OH
OH
OH
H OMe
14
15
C4H9
C4H9
OH
O
OH
O
§ The resonance at 6.66 ppm as seen in non-deuterated alkene 6 was
diminished by 90% relative to the other resonances in the NMR
spectra, and the clean double triplet at 5.65 ppm had collapsed to a
broadened multiplet.
OH
O
A
B
Figure 1.

D. C. Braddock et al. / Tetrahedron Letters 42 (2001) 7527–7529
7529
Therefore, mechanistically it seems probable that there
Acknowledgements
are two competing reaction pathways operating in the
above reactions. In the first, normal borohydride reduc-
tion proceeds by attack on the ketone resulting in alkenes
after elimination. In the second, we postulate that the
ortho-hydroxyl group activates the carbonyl group via
hydrogen bonding, which undergoes transannular-type
attack by the pseudo-gem hydroxyl group (Fig. 1; A). The
tetrahedral intermediate can collapse by formal loss of
a paracyclophane anion where the incipient build-up of
negative charge is probably solvated by MeOH (or
MeOD) leading to H+ (or D+) transfer (Fig. 1; 2B). The
ester so produced from the transannular acyl transfer is
evidently cleaved rapidly.¶ Overall, this second pathway
may best be described as an SE1-type mechanism.7
We thank GlaxoSmithKline plc for a CASE award (to
B.G.P.) and the EPSRC.
References
1. Bell, K. H. Aust. J. Chem. 1969, 22, 601–605.
2. Reich, H. J.; Cram, D. J. J. Am. Chem. Soc. 1969, 91,
3527–3533.
3. Cram, D. J. Fundamentals of Carbanion Chemistry; Aca-
demic Press: New York, 1965; p. 138.
4. March, J.; Plankl, W. J. Chem. Soc., Perkin Trans. 1 1977,
460–464.
5. Gassman, P. G.; Lumb, J. T.; Zalar, F. V. J. Am. Chem.
Finally the clean conversion of acylparacyclophane 1
into the methylene compound 2 was achieved by the
application of Clemmensen reduction conditions. Bis-
methylene compound 2 was isolated in 70% yield after
application of standard conditions (Zn/Hg, HCl, EtOH/
PhMe, reflux, 20 h)8 and no deacyl[2.2]para-
cyclophanes were observed.
Soc. 1967, 89, 946–952.
6. Reich, H. J.; Cram, D. J. J. Am. Chem. Soc. 1969, 91,
3534–3543.
7. March, J. Advanced Organic Chemistry, 4th ed.; Wiley,
1992; p. 569.
8. Cram, D. J.; Dalton, C. K.; Knox, G. R. J. Am. Chem.
Soc. 1963, 85, 1088–1093.
¶ In a control experiment the dipentanoylester of diol 5 was exposed
to the standard conditions at rt. Within 5 min all the ester had been
cleaved to diol 5.