Strained to the limit: When a cyclobutyl moiety becomes a thermodynamic sink in a protolytic ring-opening of photogenerated oxetanes

Article abstract of DOI:10.1021/ol101297b

(Equation Presented). Strained polycyclic oxetanes generated photochemically from the Diels-Alder adducts of cyclic dienes and enones undergo deep skeletal rearrangements under protolytic ring-opening conditions offering expeditious access to chlorohydrins and other products of unique skeletal topology.

Full text of DOI:10.1021/ol101297b

ORGANIC
LETTERS
2010
Vol. 12, No. 15
3398-3401
Strained to the Limit: When a Cyclobutyl
Moiety Becomes a Thermodynamic Sink
in a Protolytic Ring-Opening of
Photogenerated Oxetanes
Roman A. Valiulin, Teresa M. Arisco, and Andrei G. Kutateladze*
Department of Chemistry and Biochemistry, UniVersity of DenVer,
DenVer, Colorado 80208
akutatel@du.edu
Received June 4, 2010
ABSTRACT
Strained polycyclic oxetanes generated photochemically from the Diels-Alder adducts of cyclic dienes and enones undergo deep skeletal
rearrangements under protolytic ring-opening conditions offering expeditious access to chlorohydrins and other products of unique skeletal topology.
As popularized by Sauers,1 acyl norbornenes undergo an
intramolecular Paterno`-Bu¨chi (P.-B.) reaction yielding
reactive oxetanes which can be further modified by ring-
opening either via acid-catalyzed reactions or reduction.
This harvesting of the strain installed by a photochemical
step to access targets otherwise not easily accessible via
the ground state chemistry has been used in a number of
examples, including Rawal’s reductive fragmentation
providing access to diverse di- and triquinanes,2 photo-
induced oxametathesis,3,4 etc.
Z3 did not produce the expected oxetane Z5, but rather gave
a product Z4 rationalized in terms of the π-π* excited state
of the alkene moiety (Scheme 1).1e
Scheme 1a
Sauers also examined the photochemistry of “polycyclic
acylnorbornenes,” poised to further increase the strain. For
example, the spiro-connected ketone Z1 gave the P.-B.
product Z2. However, the “fused” cyclopentenone adduct
(1) (a) Sauers, R. R.; Kelly, K. W. J. Org. Chem. 1970, 35, 498–501.
(b) Sauers, R. R.; Whittle, J. A. J. Org. Chem. 1969, 34, 3579–3582. (c)
Sauers, R. R.; Schinski, W.; Mason, M. M. Tetrahedron Lett. 1969, 79–82.
(d) Sauers, R. R.; Kelly, K. W.; Sickles, B. R. J. Org. Chem. 1972, 37,
537–543. (e) Sauers, R. R. J. Org. Chem. 1974, 39, 1850–1853.
a From ref 1e.
10.1021/ol101297b  2010 American Chemical Society
Published on Web 07/07/2010

This failure to form Z5 has probably discouraged subse-
quent P.-B. studies of fused polycycles, as there is only a
partially relevant 1977 Paddon-Row’s account5a found in the
literatureslater confirmed by Coxon5bswhere a P.-B. chan-
nel competes with an all-carbon [2π+2π] photocyclization
in a hemicyclone-benzoquinone adduct.
We hypothesized that a modest increase in the ring size
of either the dienophile or diene could relieve the strain and
restore the Paterno`-Bu¨chi photoreactivity in fused polycyclic
systems of type Z3. Our Density Functional Theory (DFT)
calculations at the B3LYP/6-311+G(d,p) level confirmed that
oxetane Z5 relaxes by as much as 11 kcal/mol when its
cyclopentyl moiety is expanded by a single methylene group,
suggesting that the Diels-Alder (D.-A.) adducts of six-
membered and larger cycloalkenones can be P.-B. photore-
active.
monooxetane, unlike the similarly sized cyclohexano-nor-
bornene 1, which is photoactive. This further confirms that
flexibility of the ketone-containing ring is critical. Unlike 1,
the six-membered ring in 12 is further constrained by the
second rigidly fused norbornyl moiety.
According to our DFT calculations these polycyclic
oxetanes are still far more strained than their spiro counter-
parts (such as Z2 in Scheme 1). As such, they readily
undergo acid-catalyzed ring-opening and subsequent cationic
rearrangements, conforming to certain computationally pre-
dictable topologies consistent with the structure of the initial
polycyclic enone (Scheme 3).
Scheme 3
We now report that the D.-A. adducts of 2-cyclohexenone
and 2-cycloheptenone6 are indeed capable of forming strained
polycyclic oxetanes upon irradiation (Scheme 2).7 Strikingly,
even the bis-adduct of cyclohexadiene and benzoquinone 98
was found to be photoactive, producing C2-symmetric
dioxetane 11 via the monooxetane 10 when irradiated.
Scheme 2
When treated with HCl, oxetanes usually produce 1,3-
chlorohydrins. However, in the case of strained oxetanes 5-8
and 11, which can potentially open to form both tertiary or
secondary carbocations, the produced chlorohydrins resulted
from more elaborate cationic rearrangements.9
Relative stability of the tertiary cation depends on the
degree of its pyramidalization, which in these systems is
defined primarily by the size of the enone ring. Our rationale
is that for larger rings (m > 1) the acid-catalyzed oxetane
ring-opening preferentially produces the expected tertiary
cation of type A1, Scheme 4A. While relatively more stable
In the bis-series, the photoreactivity limit is reached with
the cyclopentadiene adduct 12, which does not form even a
(2) (a) Rawal, V. H.; Dufour, C. J. Am. Chem. Soc. 1994, 116, 2613–
2614. (b) Dvorak, C. A.; Dufour, C.; Iwasa, S.; Rawal, V. H. J. Org. Chem.
1998, 63, 5302–5303.
(7) In a typical procedure, 3-10 mM solution of ene-one precursors
1-4, 9 in CH3CN (or benzene) was irradiated in a quartz tube in the Rayonet
reactor equipped with RPR-3000 UV lamps (broadband 250-350 nm UV
source with peak emission at 300 nm) for 24-72 h. Conversion was
monitored by NMR, the oxetanes were used without further purification.
Oxetanes 5-8, 10-11 have a very characteristic low field multiplet at
4.0-4.5 ppm, W1/2 ≈ 7 Hz. We also were able to accurately calculate their
NMR spectrassee SI. Attempts to further purify reaction mixtures by
column chromatography were mainly unsuccessful, as the strained polycyclic
oxetanes 5-8 are not stable on silica gel, producing varying amounts of
rearranged products.
(3) Valiulin, R. A.; Kutateladze, A. G. Org. Lett. 2009, 11, 3886–3889.
(4) Pe´rez-Ruiza, R.; Miranda, M. A.; Alleb, R.; Meerholzb, K.;
Griesbeck, A. G. Photochem. Photobiol. Sci. 2006, 5, 51–55.
(5) (a) Warrener, R. N.; McCay, I. W; Paddon-Row, M. N. Aust.
J. Chem. 1977, 30, 89–94. (b) Coxon, J. M.; O’Connell, M. J.; Steel, P. J.
Aust. J. Chem. 1986, 39, 1537–1557.
(6) Synthesis of D.-A. adducts 1-4, 9, 12 is described in the literature:
(a) Fringuelli, F.; Guo, M.; Minuti, L.; Pizzo, F.; Taticchi, A.; Wenkert, E.
J. Org. Chem. 1989, 54, 710–712. (b) Northrup, A. B.; MacMillan, D. W. C.
J. Am. Chem. Soc. 2002, 124, 2458–2460. (c) Rathore, R.; Kochi, J. J.
Org. Chem. 1995, 60, 4399–4411.
(8) see Rathore, R.; Kochi, J. J. Org. Chem. 1995, 60, 4399–4411.
Org. Lett., Vol. 12, No. 15, 2010
3399

Scheme 4
Scheme 5
It appears that bis-oxetane 11, as most extremely strained,
offers a variety of deep skeletal rearrangements under
protolytic ring-opening conditions. We note, however, that
perhaps for this very reason the selectivity and reproducibility
of the reaction shown in Scheme 5 was poor. In several runs
alkene 17 was forming at 40-50% yield, whereas the yields
of products 18-21 never exceeded 10%.
Chlorohydrins 19 and 20 are expectedly related to product
15, which further confirms the generality of the rearrange-
ment shown in Scheme 4B for the oxetanes derived from
six membered dienes and dienophiles.
The alkene moiety of 19 probably resulted from an acid-
catalyzed cycloreversion in the initial oxetane, whereas
chloride 20 has a rearranged olefinic half. A plausible
mechanism for this rearrangement yielding the bicyclo[3.2.1]
moiety in 20 and, by analogy in 17 and 18, is shown in
Scheme 6. We hypothesize that cyclopropane C3 is the key
intermediate in the formation of 20. While formation of
nortricyclenes is common in solvolytic transformations of
bicyclo[2.2.1]heptanes, similar tricyclenes have been reported
in the bicyclo[2.2.2]octane series as well.10
than the alternative secondary cation, A1 is still destabilized
by forced pyramidalization. It relaxes by (remarkably)
rearranging into A2 possessing a cyclobutyl moiety. A2 is
then trapped by the chloride anion to yield the observed
products 13-14 and 16.
We note that the stereochemical outcome of this two step
rearrangement is inversion at each step, which implies that
the whole sequence might be a concerted reaction or nearly
so.
For m ) 1, that is, the smallest ring size at which Paterno`-
Bu¨chi cyclization is possible in this series, the tertiary cation
is even more destabilized. It is plausible that the two cations
are formed reversibly and competitively. The outcome in this
case depends on “n”, that is, the size of the top bridge. With
n ) 2, Scheme 4B, upon formation of a secondary cation
B1, the two carbon bridge undergoes 1,2-shift to give B2,
which is trapped by the chloride anion. Again, all steps in
the B-sequence occur with inversion of configuration, which
hints at a semiconcerted mechanism.
In the case when both m ) 1 and n ) 1 as in 5, the single
methylene bridge fails to migrate and pathway A prevails,
yielding cyclobutane 13.
As is clearly demonstrated, protolytic heterolysis of a C-O
bond in oxetanes 5-8 provides insufficient relief of strain
installed at the photochemical step, and therefore subsequent
carbocationic rearrangements accompany the oxetane ring-
opening.
Scheme 6
The polycyclic oxetanes are very reactive; they do not
generally survive column chromatography. This is especially
true for bis-oxetane 11. We found that unlike oxetanes 5-8,
bis-oxetane 11 undergoes a number of competing ring-
openings and rearrangements under acid-catalyzed condi-
tions, Scheme 5. The major product is an isomer of the
starting diene 9 for which we tentatively assign structure 17.
The rearranged bicyclo[3.2.1] moiety of 17 (and of its minor
isomer 18) is inferred from the NMR data and by analogy
with the X-ray structure of chlorohydrin 20 which has this
moiety. Structures of chlorohydrin 19 and hemiacetal 21 were
also elucidated by X-ray crystallography.
(9) A typical protolytic ring opening procedure: To a 3-10 mM solution
of oxetanes 5-8, or 11 in CH2Cl2, 1-3 mol equiv of HCl (4.0 M solution
in dioxane) was added. The resulting mixture was stirred at ambient
temperature for 24 h, the solvent was removed in vacuum, and the products
were purified by chromatography on a silica gel column using hexane-
ethyl acetate (or hexane-ethanol) as eluent .
According to a force field estimate, the formation of C3
from its precursor oxetane is about 20 kcal/mol downhill.
(10) Alder, R. W.; Carta, F.; Reed, C. A.; Stoyanova, I.; Willis, C. L.
Org. Biomol. Chem. 2010, 8, 1551–1559.
3400
Org. Lett., Vol. 12, No. 15, 2010

Acid-catalyzed cyclopropyl ring-opening in C3 followed by
Grob fragmentation gives the cis-fused alkene C4, which
equilibrates via an enol intermediate into 20. Product 20 has
trans-stereochemistry and is 1.5 kcal/mol more stable than
C4.11
Scheme 8
We suggest that the rearranged major alkene 17 (and the
minor isomer 18) are formed via a similar mechanism. Since
a small amount of the starting diene 9 is always formed by
acid-catalyzed cycloreversion, one notes that alternatively
the [2.2.2] f [3.2.1] transformation can be achieved by
simply protonating the alkenyl moiety in diene 9 as shown
in Scheme 7.
Scheme 7
However, this mechanism was ruled out because diene 9,
when treated with HCl under the same reaction conditions,
does not yield rearranged diene 17.
Arguably, the most spectacular transformation in this series
produces hemiacetal 21. Our rationale for its formation is
presented in Scheme 8, which starts with the retro P.-B.
(consistent with the formation of both 19 and 21).
The critical D3fD4 Grob-type fragmentation is set up
by two consecutive 1,2-hydride shifts, where the first shift,
D1fD2, amounts to a retro-pinacol rearrangement. Acid-
catalyzed oxetane ring-opening in D4, accompanied by the
capture of the generated carbocation with the formyl group
and subsequent conjugate electrophilic addition to the double
bond in the bicyclo[2.2.2]octadiene moety of D7, furnishes
the spiro-product 21, (its X-ray structure is shown).
In summary, strained polycyclic oxetanes can be synthe-
sized photochemically from the D.-A. adducts of cyclic
dienes and enones generally as long as the dienophile has a
ring size greater than five. Such oxetanes harvest and store
excited state energy, which can be utilized to access unique
functionalized polycyclic scaffolds via acid-catalyzed oxetane
ring-opening and subsequent deep cationic rearrangements.
Given that during the past decade enantioselectiVe catalytic
Diels-Alder reactions of cycloalkenones have been devel-
oped, starting with MacMillan’s pionering studies,6b the
photoprotolytic one-pot transformations described here may
offer expeditious access to a variety of enantiopure func-
tionalized polycycles.
Acknowledgment. Support of this research by the NIH
(GM093930) and by the donors of the Petroleum Research
Fund, administered by the American Chemical Society
(49785-ND4), is gratefully acknowledged.
Supporting Information Available: Experimental details,
NMR and X-ray characterization data. This material is
available free of charge via the Internet at http://pubs.acs.org.
(11) Notice that the same outcome can be achieved with an alternating
sequence of (1,3-hydride)-(1,2-alkyl)-(1,3-hydride) shifts.
OL101297B
Org. Lett., Vol. 12, No. 15, 2010
3401