Retinoic acid oxidation at high oxygen pressures: Evidence for spin-forbidden direct addition of triplet molecular oxygen

Full text of DOI:10.1021/ja970774o

9560
J. Am. Chem. Soc. 1997, 119, 9560-9561
Scheme 1
Retinoic Acid Oxidation at High Oxygen Pressures:
Evidence for Spin-Forbidden Direct Addition of
Triplet Molecular Oxygen1
K. Brady Clark,*,†,‡ J. A. Howard,‡ and Alan R. Oyler§
Sussex Research Laboratories Inc., 100 Sussex DriVe
Ottawa, Canada, K1A OR6
Steacie Institute for Molecular Sciences, National
Research Council of Canada, Ottawa, Canada K1A 0R6
R. W. Johnson Pharmaceutical Research Institute
Raritan, New Jersey 08869-0602
we report our preliminary results concerning the mechanism
for the formation of the endoperoxide 4.
Preliminary oxygen uptake studies showed that reaction of 1
with O2 is self-initiated and autocatalytic under ambient
conditions in solvents such as benzene. Despite this, 1 initially
reacts with O2 rather slowly under similar conditions in 90%
ethanol, the solvent system of this study. Reactions were,
therefore, carried out at higher oxygen concentrations by
increasing the oxygen pressure.5 It is our contention that the
results of previous retinoic acid oxidation studies were obscured
by prolonged oxidation resulting in mixtures dominated by
numerous secondary products.4 In the present study, reaction
mixtures were analyzed6 at low (<5%) conversions of starting
material. Surprisingly simplified product distributions (eq 1)
ReceiVed March 10, 1997
Introduction
Retinoic acid (1), a biologically active polyene vital for
mammalian development with activity against a number of
cancers and dermatological diseases in humans,2 is known to
react rapidly with molecular oxygen. Retinoic acid has been
shown to undergo initiated autoxidation in solution by a
mechanism involving free radical chain reactions.3 A thorough
kinetic analysis was carried out for this process; however, the
mechanism of retinoid autoxidation put forward was never
clarified since a thorough product analysis was not carried out.3
With few exceptions the resulting complex reaction mixtures
from other studies have not been well characterized, and this
has led to the assumption that abstraction of an activated
secondary H atom from the substituted cyclohexenyl ring by a
peroxyl radical is responsible for the observed product distribu-
tion (Scheme 1).
A comprehensive study has been reported for the oxidation
of all-trans-retinoic acid (1) in 90% ethanol in which the major
“dark” products were characterized.4 Among the products
common to retinoid oxidation, e.g., the furan 2 and the epoxide
3, were a number of products that had not previously been
identified. These included minor quantities of the cyclic
peroxides 4 and 5 and high yields of the olefinic fragmentation
compounds 6-9. It was suggested that the cyclic peroxide 4
and 5 and the fragmentation products 6-9 were produced by a
radical-cation chain mechanism rather than by autoxidation.4
were observed and consisted of three major and several minor
products. Two of the major products were identified7 as 5,6-
epoxy-5,6-dihydroretinoic acid (3) and 5,8-epidioxy-5,8-dihy-
droretinoic acid (4). The third major product was identified8
as 5-hydroxy-8-oxo-6,7-dihydroretinoic acid (10), which to the
best of our knowledge has not previously been reported as an
oxidation product of 1. In addition to other unidentified
compounds, trans- and cis-5,8-epoxy-5,8-dihydroretinoic acid
(2a and 2b, respectively), 2-methyl-6-oxo-2,4-heptadienal (6),
â-ionone (7), cyclocitral (8), and dihydroactinidiolide were
identified in minor amounts. All compounds, with the exception
of 10, have previously been identified and fully characterized
by Oyler et al. as products of all-trans-retinoic acid oxidation
but in quite different relative concentrations. Thus, Oyler et
al. reported that 2 and 6-9 comprised the bulk of the product
mixture while 3 and 4 were minor components,4 whereas 2 and
6-9 are all minor products and 3 and 4 are major components
(along with 10) at the low substrate conversions used in the
present work.
To minimize the importance of autoxidation, reaction of 1
with O2 was carried out under conditions identical to those
described above but in the presence of 1-10× molar excess of
2,6-di-tert-butyl-4-methylphenol (BMP), an efficient peroxyl
radical scavenger. Under these conditions, yields of 4 and 10
(5) Oxidations employed 3 mM solutions of 1 (0.1% by weight) in 90%
ethanol. For inhibited runs, the reaction mixtures generally contained 1-10×
molar equivalents of antioxidant. Handling of photolabile retinoic acid
oxidation solutions was carried out under Gold light to prevent photoi-
somerization and photoinitiation.
(6) Analytical HPLC was performed on a Hewlett-Packard Series II 1090
Liquid Chromotograph equipped with diode array UV-vis detector.
Separations were achieved by using a 4.6 mm × 250 mm Spherisorb ODS2
5 m column operated with a tertiary gradient solvent system.
(7) On the basis of elution times, co-injection and comparison of UV
specta of authentic, independently synthesized compounds.
We have recently been investigating the initial stages of the
reaction of all-trans-retinoic acid with molecular oxygen with
a view of clarifying the mechanism(s) of oxygenation. Herein,
† Sussex Research Laboratories Inc.
‡ National Research Council of Canada
§ R. W. Johnson Pharmaceutical Research Institute.
(1) Issued as NRCC No. 40833.
(2) The Retinoids: Biology, Chemistry, and Medicine, 2nd ed.; Sporn,
M. B., Roberts, A. B., Goodman, D. S., Eds.; Raven Press: New York,
1994.
(3) Finklelshtein, E. I.; Rubchinskaya, Y. M.; Kozlov, E. I. Zhurn. Org.
Chem. 1981, 17, 936.
(8) On the basis of its spectral and molecular weight data after isolation
of the pure compound from oxidized 1. UV-vis λmax ) 327 nm
1
(acetonitrile); H NMR (CDCl3, 500 MHz) vinyl protons: 5.80 (s br, 1H,
(4) Oyler, A. R.; Motto, M. G.; Naldi, R. E.; Facchine, K. L.; Hamburg,
P. F.; Burinsky, D. J.; Dunphy, R.; Cotter, M. L. Tetrahedron 1989, 45,
7679-7694.
H-14), 6.12 (s br, 1H, H-7), ca. 6.15 (d, 1H, J10,11 ) 11.3 Hz, H-10, doublet
obscured by H-7 (6.12)), 6.28 (d, 1H, J11,12 ) 15.3 Hz, H-12), 6.97 (dd,
1H, J11,12 ) 15.3, J10,11 ) 11.3 Hz, H-11); MS (ES) 332.1 (M+•).
S0002-7863(97)00774-9 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 40, 1997 9561
Scheme 2
3
were unaffected while formation of 3 was almost completely
inhibited. There was also neglible formation of 2a,b, 6-9, and
some other minor unidentified products even at >25% conver-
sion of 1. This suggests that 4 and 10 are not formed by
autoxidation. However, it should be noted that similar con-
centrations of more efficient antioxidants, such as R-tocopherol9
resulted in lower yields of both 4 and 10.
oxygen encounter complex [1, O2].16 It is well-known that
oxygen-saturated organic solvents produce charge-transfer
spectra, the intensity of which in many cases has been shown
to be proportional to the oxygen concentration.17 For most
organic compounds (M), the spectrum is collisional in nature
with no stable complex between O2 and the substrate being
formed. The importance of charge transfer may increase
These results suggest that, at early stages of oxidation, the
product mixture is relatively simple and that two concurrent
processes are at play, one involving an inhibitable radical chain
process (autoxidation) which leads to formation of 3, and another
where 1 apparently undergoes a direct reaction with triplet
molecular oxygen to give 4 and 10. Formation of 4 and 10 is,
however, quickly overshadowed, as the oxidation proceeds by
production of 2, 3, and 6-9 (among others) by autoxidation.
Although BMP does not inhibit the formation of 4 and 10, the
ability of tocopherols to partially inhibit their formation suggests
that oxygen-centered radical intermediates are involved in their
formation. The experimental evidence collected here does not
support the participation of radical-cation chains, at least in the
formation of 4 and 10, although work supporting such inter-
mediates has been published for other unsaturated systems.12
Kinetic data13 would seem to preclude reaction of 1•+ and O2
in any solvent system, let alone one comprised of ethanol
containing 10% water (where H2O behaves as a weak nucleo-
philic radical-cation quencher with a bimolecular rate constant
of 3.5 × 105 M-1 s-1).13 Our preliminary solvent dependence
studies suggest that yields of 4 (as well as 10) are actually higher
in aqueous systems.
3
significantly for [1, O2] complexes with the enhanced donor
strength of retinoic acid, an electron-rich polyene with a low
ionization potential.18 Oxygenation may be initiated when
electron donor-acceptor complexes collapse reversibly to form
3
triplet biradicals (carbon-peroxyl biradical) D and/or zwitte-
rionic intermediates as depicted in Scheme 2. For simplicity,
we have not included the possibility of zwitterionic intermediates
in Scheme 2. Fast ring closure of biradical 3D after obligatory
intersystem crossing would result in endoperoxide 4 formation
(path a), a process that must occur very rapidly to avoid
3
hydrogenation via BMP. At high oxygen concentrations, D
may be intercepted by O2, resulting in an unstable tetroxide
which would be expected to undergo cleavage to 10 via a
mechanism similar to that described by Russell for bimolecular
reaction of primary and/or secondary peroxyls.19 Formation of
10 would be expected to increase as the concentration of O2
increases if this mechanism is correct. Experiments designed
to follow the formation of 10 as a function of O2 concentration
are underway. Preliminary results reveal that formation of 10
may be attenuated at lower O2 concentrations.
3
Inhibitable radical chain autoxidation is initiated when D
The possibility that 10 is derived from rearrangement of
endoperoxide 4 was considered; however, control experiments,
where 3 mM solutions of 4 were subjected to oxidation
conditions (retinoic acid absent) similar to those for retinoic
acid, did not result in any detectable formation of 10.
We can only speculate as to the exact nature of the process
that gives rise to direct addition of molecular oxygen and
products 4 and 10. Although considered rare, there have been
a number of “direct” olefinic oxidations reported that do not
appear to involve radical chains or singlet oxygen.12,14-16
At this point we prefer to think of the direct addition of O2
in terms of a mechanism involving a retinoic acid-ground state
(Scheme 2) abstracts hydrogen from solvent or substrate,
resulting in DH, a radical which in turn would undergo
intramolecular homolytic substitution (SHi), yielding epoxide 3
and a chain carrying hydroxyl, HO• (path c, Scheme 2).
Although according to this mechanism epoxide 3 is formed in
the initiating steps, the bulk of 3 that is formed in the uninhibited
autoxidation of 1 likely arises from a chain propagating step in
which a peroxyl RO2 adds to C5 of 1 followed by an SHi
reaction to produce the epoxide.
Further work, aimed at examining the role of radical
intermediates in the propagation process, is currently in progress.
•
Acknowledgment. The authors thank Dr. K. U. Ingold for helpful
discussions, and Mauro Tomietto for assisting in the semipreparative
HPLC isolation of compounds.
(9) R-Tocopherol (R-T) scavenges peroxyls ∼200 times faster than
BMP.10 In aqueous media R-T is about 3 times more reactive than BMP.11
(10) Burton, G. W.; Ingold, K. U. J. Am. Chem. Soc. 1981, 103, 6472.
(11) Pryor, W. A.; Strickland, T.; Church, D. F. J. Am. Chem. Soc. 1988,
110, 2224-2229.
(12) Correa, P. E.; Hardy, G.; Riley, D. P. J. Org. Chem. 1988, 53, 1695-
1702.
JA970774O
(13) Bobrowski, K.; Das. P. K. J. Phys. Chem. 1985, 89, 5079-5085.
(14) Seip, M.; Brauer, H.-D. J. Am. Chem. Soc. 1992, 114, 4486-4490.
(15) Bartlett, P. D.; Banavali, R. J. Org. Chem. 1991, 56, 6043-6050.
(16) Turro, N. J. Tetrahedron 1985, 41, 2089-2098.
(17) Evans, D. F. J. Chem. Soc. 1953, 345.
(18) Kochi, J. K. AdV. Phys. Org. Chem. 1994, 29, 185-272.
(19) Russell, G. A. J. Am. Chem. Soc. 1957, 79, 3871.