Preparation of an unsymmetrically labeled allylic hydroperoxide and study of its allylic peroxyl radical rearrangement [1]

Full text of DOI:10.1021/ja9723038

11534
J. Am. Chem. Soc. 1997, 119, 11534-11535
Communications to the Editor
Scheme 1
Preparation of an Unsymmetrically Labeled Allylic
Hydroperoxide and Study of Its Allylic Peroxyl
Radical Rearrangement
Jennifer R. Lowe and Ned. A. Porter*
Department of Chemistry, Duke UniVersity
Durham, North Carolina 27708
ReceiVed July 11, 1997
Allylic hydroperoxides are formed in the free radical and sing-
let oxygen oxidation of unsaturated compounds. Allylic hy-
droperoxides are also formed enzymatically, the prostaglandin
G family being one example having this substructure. Allylic
hydroperoxides undergo a rearrangement1 thought to be free
radical in nature (see eq 1) in which oxygen is transferred across
the allyl group. The mechanism of the allylic rearrangement
has been long debated with several mechanisms proposed.
silyl ether 4. Removal of the hydroperoxide protecting group
with 4:2:1 THF/AcOH/H2O at 0 °C gives allylic hydroperoxide
5, formed in 47% overall yield in the six steps from 2.5 Reaction
of 5 with triphenylphosphine gives triphenylphosphine oxide
labeled with 44% 18O and the structurally related alcohol that
contains no label. This establishes the position of the oxygen
label as being in the terminal oxygen6 and furthermore is
consistent with the presumed mechanism of the formation of 1
from a benzoyloxy radical in which the label scrambles between
the two equivalent oxygens.
Rearrangement of 5 in dodecane was initiated by di-tert-butyl
hyponitrite at temperatures of 40 °C or below and by azobis-
isobutyronitrile (AIBN) at higher temperatures.7 Rearrange-
ments were carried out under air, aliquots were taken, and the
product hydroperoxide 6 was separated from 5 by HPLC on
silica gel. Isolated 5 and 6 were reacted with triphenylphos-
phine, and the phophine oxide and product alcohol were
analyzed by GC-MS to determine the isotopic composition of
the terminal and proximal oxygen of the two hydroperoxides.
Data from a typical rearrangement are presented in Figure 1.
Extrapolation of the normalized percent label present in the
proximal and terminal positions of 6 to zero time gives
information not only about the regioselectivity of the rearrange-
ment but also provides a measure of the fraction of label lost
during rearrangement. The data presented in the Figure 1
indicate that, at 60 °C in dodecane, 64% of the terminal label
in 5 is transferred to the proximal position of 6 while 6.4%
transfers to the product terminal position and 29.6% of the label
is lost to the atmosphere.
Evidence supports the notion that the rearrangement is free
radical since it is initiated and inhibited with known radical
initiators and inhibitors. While tracing the course of peroxyl
oxygens during rearrangement would provide valuable mecha-
nistic insight, experiments with compounds having a specific
hydroperoxide oxygen labeled have not been possible because
of the synthetic inaccessibility of the requisite hydroperoxides.
We report here the preparation of an allylic hydroperoxide
labeled with 18O (shown as b) only in the peroxide terminal
position along with our first studies of the allylic rearrangement
of this compound.
Reaction of the hydroxamate ester (Scheme 1), labeled in
the carbonyl oxygen with 90% 18O,2 with nitrosyl chloride gives
perester 1 in yields ranging from 28 to 40%.3 Hydrolysis with
lithium hydroxide affords the known hydroperoxide as a
crystalline solid in 85% yield. Protection of the hydroperoxide
as a perketal and reduction of 2 to the corresponding alcohol is
realized with diisobutyllithium aluminum hydride (DIBAL-H),
and the product alcohol is oxidized with pyridinium dichromate.4
The crude aldehyde can be reacted with the stabilized ylide in
benzene at room temperature for 48 h to give the unsaturated
ester 3. A second DIBAL-H reduction and derivatization with
tert-butyldiphenylsilyl chloride (TBDPS-Cl) gives the allylic
Scheme 2 presents time zero extrapolated product label
information for the rearrangement of 5 f 6 as well as for
rearrangements of 6 f 5 carried out at 20, 40, and 60 °C. The
rearrangement is temperature dependent, reactions carried out
at lower temperatures proceeding with less loss of label to the
atmosphere and with more positional selectivity of the label
* Author to whom correspondence should be addressed at the follow-
ing: tel (919) 660-1550, Fax (919) 660-1605, email porter@chem.duke.edu.
(1) (a) Schenck, G. D.; Neumuller, O. A.; Eisfeld, W. Angew. Chem.
1958, 70, 595. (b) Brill, W. F. J. Chem. Soc., Perkin Trans. 2 1984, 621.
(c) Porter, N.; Zuraw, P. J. Chem. Soc., Chem. Commun. 1985, 1472. (d)
Beckwith, A. L.; Davies, A. G.; Davison, I. G. E.; Maccoll, A.; Mruzek,
M. H. J. Chem. Soc., Perkin Trans. 2 1989, 815. (e) Avila, D. V.; Davies,
A. G.; Davison, I. G. E. J. Chem. Soc., Perkin Trans. 2 1988, 1847. (f)
Mills, K. A.; Caldwell, S. E.; Dubay, G. R.; Porter, N. A. J. Am. Chem.
Soc. 1992, 114, 9689. (g) Porter, N. A.; Caldwell, S. E.; Mills, K. A. Lipids
1995, 30, 277 and references cited therein.
(5) Hydroperoxide 5 is stored as a dilute solution stabilized with 5%
butylated hydroxytoluene (BHT) and is purified by HPLC on silica gel
immediately before use.
(2) The hydroxamate ester was synthesized from 18O-labeled benzoyl
chloride and the corresponding alkoxyl amine. 18O-labeled benzoyl chloride
is prepared in one step from benzotrichloride and H218O in a modified
procedure: Ponticorvo, L.; Rittenberg, D. J. Am. Chem. Soc. 1954, 76, 1705.
(3) (a) Koenig, T.; Deinzer, M. J. Am. Chem. Soc. 1966, 88, 4518. (b)
Koenig, T.; Deinzer, M. J. Am. Chem. Soc. 1968, 90, 7014. (c) Koenig, T.;
Deinzer, M.; Hoobler, J. A. J. Am. Chem. Soc. 1971, 93, 938. (d) Koenig,
T.; Hoobler, J. A.; Mabey, W. R. J. Am. Chem. Soc. 1972, 94, 2514. (e)
Koenig, T.; Hoobler, J. A. Tetrahedron Lett. 1972, 1803. (f) Koenig, T.
Tetrahedron Lett. 1973, 3487.
(6) Caldwell, S. E.; Porter, N. A. J. Am. Chem. Soc. 1995, 117, 8676.
(7) In a typical rearrangement, hydroperoxide concentration is 0.01 M
and initiator concentration is 0.001 M. Rearrangements are carried out over
the course of 10-24 h with aliquots taken at regular intervals. During the
first 5-10 h of rearrangement, only starting material and rearrangement
hydroperoxide can be detected either by HPLC or by NMR. At later times,
termination products corresponding to the alcohol and ketone structurally
analogous to 6 can be observed by both HPLC and NMR. The hydroper-
oxide 6 and these termination products generally account for greater than
95% of consumed 5, even at long reaction times (>24 h). An equilibrium
mixture of 5/6 of 55:45 is approached for the rearrangement started from
either hydroperoxide.
(4) For precedents for the transformations presented in Scheme 1, see:
Dussault, P. Synlett 1995, 997.
S0002-7863(97)02303-2 CCC: $14.00 © 1997 American Chemical Society

Communications to the Editor
J. Am. Chem. Soc., Vol. 119, No. 47, 1997 11535
Scheme 3
to give free allyl and oxygen. The data from Scheme 2 provide
an estimate of the competition rate constants. Thus, at 40 °C,
k1/k2 is approximated by (82 - 3.3)/(2 × 3.3 + 14.7) ) 3.7.9
Estimation of k1/k2 at other temperatures followed by Eyring
analysis10 provides ∆∆Hq, and ∆∆Sq, for the competitive
pathways. Thus, ∆Hq - ∆Hq ) 9.0 kcal/mol and ∆Sq
-
Figure 1. Normalized 18O label in product 6 from the rearrangement
of 5 labeled in the terminal peroxide oxygen. The rearrangement is at
60 °C in dodecane.
2
1
2
∆Sq1 ) 26 eu by this analysis. Analysis of the reverse reaction
in a similar way gives similar activation parameters, ∆Hq
-
-3
∆Hq ) 9.0 kcal/mol and ∆Sq - ∆Sq ) 29 eu. These
-1
-3
-1
Scheme 2
data are consistent with a competition between an associated,
ordered reaction (k1) having a negative activation entropy with
a dissociative process (k2) having a positive activation entropy.
The studies reported here do not distinguish between associative
pathways involving a concerted oxygen transfer or a stepwise
endocyclization-fragmentation. Experiments with analogous
peroxyls do not, however, provide evidence for a carbon radical
intermediate in the rearrangement,1b-d pointing to the concerted
[3,2] oxygen transfer in those systems.
The availability of unsymmetrically labeled hydroperoxides
by the synthetic strategy reported here permits other mechanistic
issues, including the nature of the stereoselectivity of the
associative and disociative pathways,11 to be addressed with this
powerful probe. The results of these studies will be reported
in due course.
transfer. This allows for the isolation of 6 from reactions at 20
°C that has label almost entirely at the proximal position. In
fact, the reverse rearrangements (6 f 5) were carried out with
6 isolated from early aliquots of the 5 f 6 rearrangement at 20
°C and generally had a proximal/terminal label ratio of >100:
1. Another feature of the rearrangement is that the 5 f 6
rearrangement (tertiary to secondary) is apparently more selec-
tive than the 6 f 5 process (secondary to tertiary). Thus, while
the proximal to terminal product ratio for the 60 °C 5 f 6
process is 10:1 and 30% of the label is lost to the atmosphere,
the 6 f 5 rearrangement occurs under identical conditions to
give 5 with terminal/proximal label of 2:1 and with fully two-
thirds of the label lost.
A mechanism for the rearrangement involving competitive
pathways k1 Vs k2 and k-1Vs k-3 is presented in Scheme 3. The
k1 pathways (forward and reverse) connect the peroxyl radicals
directly by a process in which the terminal oxygen of 5 is
transferred to the product proximal position, while the k2 and
k3 pathways involve fragmentation of peroxyls to a pair species8
which can collapse to peroxyl with scrambling of label or escape
Acknowledgment. NSF support and the receipt of an NIH MERIT
Award is gratefully acknowledged. J.R.L. acknowledges receipt of a
C. R. Hauser Award. We thank Dr. Patrick H. Dussault of the University
of Nebraska for helpful discussions.
Supporting Information Available: Synthetic procedures, analyti-
cal data, and Arrhenius plots for rearrangements (8 pages). See any
current masthead page for ordering and Internet access instructions.
JA9723038
(8) The reversible addition of oxygen to resonance-stabilized radicals
has been discussed extensively. See ref 1d. As an example, the equilibrium
of the cumylperoxyl and cumyl radicals and dioxygen has been estimated
to be 10-9 M-1 making carbon radical-carbon radical termination reactions
unlikely. For a discussion of this point, see: Howard, J. A.; Chenier, J. H.
B.; Yamada, T. Can J. Chem. 1982, 60, 2566. Nangia, P. S.; Benson, S.
W. Int. J. Chem. Kinet. 1980, XII, 19.
(9) Microscopic reversibility requires that k1/k-1 ) (k2/k-2)(k3/k-3).
See: Hammett, L. H. Physical Organic Chemistry, 2nd ed.; McGraw Hill:
New York, 1970; p 142.
(10) See the Supporting Information for Arrhenius plots.
(11) Porter, N. A.; Mills, K. A.; Caldwell, S. E.; Dubay, G. R. J. Am.
Chem. Soc. 1994, 116, 6697.