Allene epoxidation: Synthesis of functionalized lactones by the DMDO oxidation of allenic acids

Article abstract of DOI:10.1016/S0040-4020(02)00698-1

A series of allenic carboxylic acids has been converted to functionalized lactones by oxidation-cyclization promoted by dimethyldioxirane. These transformations are rationalized by the involvement of unisolated allene oxide and spirodioxide intermediates. The structures of the starting allenic acids and the reaction conditions determine which of these two intermediate species serves as the source of the isolated products. The use of prepared solutions of the oxidant generally proceed via spirodioxides; whereas in situ reactions normally give products derived from the allene oxides. When products are formed directly from allene oxides, keto lactones are formed. The corresponding spirodioxide intermediates give lactones with appropriately situated α-hydroxyketone moieties. An understanding of the regiochemistry of the epoxidations and the subsequent cyclizations of the reactive intermediates is developed, as are methods for the manipulation of the experimental conditions to favor the desired products.

Full text of DOI:10.1016/S0040-4020(02)00698-1

Tetrahedron 58 (2002) 7027–7036
Allene epoxidation: synthesis of functionalized lactones by the
DMDO oxidation of allenic acids
*
Jack K. Crandall and Elisa Rambo
Department of Chemistry, Indiana University, Bloomington, IN 47405-7102, USA
Received 5 April 2002; accepted 20 May 2002
Abstract—A series of allenic carboxylic acids has been converted to functionalized lactones by oxidation–cyclization promoted by
dimethyldioxirane. These transformations are rationalized by the involvement of unisolated allene oxide and spirodioxide intermediates. The
structures of the starting allenic acids and the reaction conditions determine which of these two intermediate species serves as the source of
the isolated products. The use of prepared solutions of the oxidant generally proceed via spirodioxides; whereas in situ reactions normally
give products derived from the allene oxides. When products are formed directly from allene oxides, keto lactones are formed. The
corresponding spirodioxide intermediates give lactones with appropriately situated a-hydroxyketone moieties. An understanding of
the regiochemistry of the epoxidations and the subsequent cyclizations of the reactive intermediates is developed, as are methods for
the manipulation of the experimental conditions to favor the desired products. q 2002 Elsevier Science Ltd. All rights reserved.
1. Introduction
conditions,3,5 which typically expose them to the carboxylic
acids derived from the peracid oxidizing agent. As a result,
it has been difficult to control the various competitive and
consecutive pathways that occur during the epoxidation of
allenes to give a desired product in an efficient manner.
The details of the epoxidation chemistry of allenes that are
summarized in Scheme 1 have been elaborated in studies
dating back to the late 1960s.1,2 The reaction of allenes with
peracids and other oxygen-transfer reagents has been
demonstrated to proceed by way of allene oxide (methylene-
oxirane) intermediates of type 1.3–5 While these reactive
species have been shown to isomerize to the more stable, but
reactive cyclopropanones of general formulation 2 in certain
instances,5,6 they generally react with nucleophiles present
in the reaction mixture to give functionalized ketones of
structure 3 as isolated products. Under some oxidation
conditions, a second epoxidation reaction can become
competitive with other transformations of 1. This sequential
oxidation leads to the formation of spirodioxides (1,4-
dioxaspiropentanes) of parent structure 4 as a further
reactive intermediate.7 The intermediate moiety 4 also
interacts readily with nucleophiles to give stable acyclic
hydroxy ketones of type 5, in addition to other reaction
processes. The overall conversion of allenes to the highly
substituted ketones 5 results in an interesting differential
functionalization of the three carbons of the original allene
carbon triad, a process of some synthetic interest for the
rapid incorporation of multiple functionality into a
molecule. While examples of compounds of types 1, 2,
and 4 have been isolated and characterized in special cases,
these highly reactive species are ordinarily converted to
stable products such as 3 and 5 under the reaction
The discovery of a method for the preparation of dilute
solutions of the highly reactive and selective oxidizing
agent dimethydioxirane8 (DMDO) has provided a whole
new dimension to the epoxidation chemistry of allenes. This
remarkable reagent generally permits clean, sequential
epoxidation of allenes to isolable spirodioxides,9 since it
is typically used under neutral, non-nucleophilic conditions.
These intermediates can then be treated with a range of
Keywords: epoxidation of allenes; lactone synthesis; allenic acids;
dimethyldioxirane.
*
Corresponding author. Fax: þ1-812-855-8300;
e-mail: jkcran@indiana.edu
Scheme 1.
0040–4020/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(02)00698-1

7028
J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
external nucleophiles under non-acidic conditions to
produce hydroxy ketones of type 5 possessing a variety of
Z substituents. The stereochemistry of the spirodioxides
derived from substituted allenes can be understood in terms
of the normal propensity for the initial epoxidation of
DMDO to occur at the more substituted double bond,10 with
stereochemistry being controlled by attack of the epoxidiz-
ing agent from the p-face of the double bond situated away
from a substituent attached to the non-reacting double
bond.9 The substituted allene oxide intermediate thus
generated undergoes a rapid, but frequently stereochemi-
cally less discriminating, second epoxidation leading to a
substituted spirodioxide. Ring-opening of substituted spiro-
dioxides in the absence of acid appears to take place by
nucleophilic attack at the less substituted spirodioxide
terminus with inversion of configuration at this carbon site.9
This is, of course, the normal reactivity pattern for
nucleophilic additions to epoxides.
presence of a weak acid or a more aggressive nucleophile,
respectively, in these two situations might well be expected
to influence the oxidation chemistry of these allenic
compounds in a significant manner.
2. Results and discussion
The DMDO oxidations of a number of 3,4-dienoic acids
were studied in detail. Thus, acid 6a with gem-dimethyl
substitution at the remote allenic carbon gave hydroxy
lactone 7a (Scheme 2) in 87% yield. The pathway to 7a
undoubtedly involves sequential formation of allene oxide
8a and spirodioxide 9a, prior to intramolecular nucleophilic
attack by the pendant carboxylic acid moiety to furnish
lactone 7a. With the specific substitution pattern of allene
6a, the initial epoxidation is expected to occur preferentially
at the more substituted double bond from the p-face
opposite to the acid side-chain to give largely the
stereoisomer of 8a shown.9 The spirodioxide 9a formed
by further epoxidation of 8a would then have a geometry
appropriate for ring-opening with backside participation of
the acid function. This is the suggested mechanistic mode of
cyclization. However, given the weakly acidic conditions of
this reaction, acid-catalyzed ring-opening of 9a to a
carbocation that is subsequently captured by the acid
group is an alternative mechanism that cannot be rigorously
excluded, although there is no firm evidence for such a
pathway.
The oxidation of a series of allenic alcohols with DMDO has
provided information concerning the behavior of substrates
with appended nucleophilic groups.11 With few exceptions,
these reactions proceed by way of spirodioxide inter-
mediates of type 4, which are not stable under the reaction
conditions, but which undergo rapid nucleophilic cycliza-
tion. This is not the case for allenic alcohols with other types
of oxidizing reagents, which tend to favor reactions of the
allene oxide intermediate.12 Furthermore, the regio-
chemistry of this intramolecular nucleophilic process is
largely controlled by the relative position of the hydroxy
group, so as to selectively generate more stable five and six-
membered heterocyclic rings. In the present work, we
compare the behavior of the analogous carboxylic acids,
which can be oxidized either as the free carboxylic acid or in
the form of the corresponding carboxylate anion.13 The
When acid 6a was oxidized in situ by incorporating it into a
mixture of Oxone,w acetone, aqueous sodium bicarbonate
and methylene chloride11 (conditions for the preparation of
DMDO), a new lactone 10a was formed in 67% yield with
little, if any, of its hydroxy analog 7a being observed. In this
Scheme 2.

J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
7029
instance, cyclization is clearly occurring at the initial allene
oxide stage prior to a second epoxidation. Two features of
the in situ reaction probably account for this distinct change
in the product-forming intermediate from allene oxide 8a to
spirodioxide 9a. First, 8a is present as its sodium salt under
the in situ reaction conditions, providing for more efficient
cyclization onto the epoxide moiety of 8a by classical
displacement at the remote epoxide center by the more
nucleophilic carboxylate anion. Secondly, the concentration
of DMDO, the presumed oxidant, is undoubtedly much
lower under the in situ circumstances, thereby retarding the
overall rate of the second epoxidation, once again providing
a better opportunity for competing cyclization of allene
oxide 8a. Consistent with this explanation, the oxidation of
an acetone solution of acid 6a that had been previously
treated with sodium bicarbonate to generate the correspond-
ing carboxylate anion with a solution of DMDO gave a
mixture of the two types of lactones 7a (52% yield) and 10a
(10% yield). Thus, products derived from cyclization of
either the mono or the di-epoxidized intermediates from 6a
can be achieved simply by modifying the oxidation
conditions. In situ reactions give the simple lactone 10a,
whereas the use of DMDO solutions leads to the hydroxy
lactone 7a.
Scheme 3.
initial epoxidation, which now occurs preferentially at the
more substituted, internal double bond of allene 12 to
generate allene oxide 14. While the unsubstituted double
bond of allene oxide 14 is expected to undergo epoxidation
less readily than that of the more highly substituted analogs
discussed above, it is nonetheless surprising that ring-
closure of 14 to the strained b-lactone 13 is able to compete
so favorably with further epoxidation. The exclusive
nucleophilic attack of the both the free acid and the
corresponding carboxylate anion at the proximate epoxide
site of 14 suggests that lactone formation occurs directly
from the intact allene oxide even in the case of the free acid.
An alternate route involving prior conversion to hydroxy
allyl cation 15 by protonation and ring-opening of 14 would
be expected to give at least some of the more stable isomeric
d-lactone, unless the requisite syn isomer of 15 is totally
inaccessible for some reason. The steric bulk of the acid-
bearing side-chain of 15 in this specific case makes it
difficult to rule out exclusive reaction through the more
stable anti isomer of 15, which can only cyclize to b-lactone
13. Nonetheless, the simplest explanation consistent with
the facts is a single-step nucleophilic displacement for the
free acid, as well as the corresponding carboxylate anion.
The closely related allenic acid 6b, which possesses an
additional methyl group at the 2-position, behaved quite
analogously to 6a. DMDO oxidation of the free acid gave
lactone 7b in 92% yield as a 3:2 mixture of cis and trans
isomers. The assignment of stereochemistry to the minor
isomer is based on a 13 Hz coupling constant between the
two vicinal protons on the ring in the NMR spectrum of the
trans compound. (The cis isomer has a measured coupling
constant of 6 Hz.) Molecular mechanics calculations14 were
used to predict the stable conformations and coupling
constants of the two isomers. This verifies that only the
trans compound is consistent with the large observed
coupling constant. Thus, the methyl group at the 2-position
has little controlling influence on the stereochemistry of the
product as might have been expected. Interestingly, an
attempt at silylation of this isomeric mixture with TBDMS
triflate and 2,6-lutidine, in order to facilitate separation,
resulted in conversion to a mixture of the silylated trans
alcohol 7b (50% yield) and the dehydrated lactone 11 (34%
yield). Elimination and epimerization accompany the
silylation of 7b under these conditions.
The addition of p-toluenesulfonic acid to a DMDO
oxidation of 12 was also examined, in view of the
observation that strong acid has been found to encourage
the cyclization of the analogous alcohols at the allene oxide
stage.11 In this case, a similar process would involve
diverting 14 to the corresponding hydroxy allyl cation 15. In
fact, this reaction gave two unexpected products that had
incorporated acetone, as well as a 15% yield of b-lactone
13. The new solids were assigned as the cyclic ketals 16
(43% yield) and 17 (32% yield) on the basis of their unique
spectroscopic properties. While the pathway to these
products is not entirely clear, there are a number of logical
routes that ketal formation and migration could follow. The
more highly oxidized ketal 17 appears to be derived from
the further oxidation of 16, since higher oxidation products
of 12 that might lead directly to 17 are not observed in any
of the other reactions of 12.
Once again, in situ oxidation gave the less highly
functionalized d-lactone 10b efficiently (76% yield).
Oxidation of the carboxylate anion of 6b with a solution
of DMDO in the presence of NaHCO3 resulted in the
formation of a mixture of lactone 10b with the cis and trans
hydroxy lactones 7b as found in the previous experiment.
The oxidation chemistry of the terminal allenic acid 12
showed a significantly different pattern (Scheme 3). In this
case, product formation from the allene oxide intermediate
was dominant, since 12 gave the b-lactone 13 as the major
product, regardless of the oxidation conditions: DMDO
(52% yield), DMDO plus NaHCO3 (84% yield) and in situ
reaction (72% yield). The formation of a strained four-
membered ring lactone in this instance is striking, but this
selectivity is surely governed by the regiochemistry of the
The 3,4-dienoic acid 6c, with gem-dimethyl substitution at
both the allene terminus and the 2-position, was found to
give crystalline hydroxy lactone 7c in 96% yield under the

7030
J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
usual DMDO oxidation conditions (Scheme 2). Unexpect-
edly, flash chromatography on silica gel isomerized 7c to a
secondary product assigned structure 18. This new lactone
18 could also be prepared in 80% yield simply by stirring 7c
with a slurry of silica gel in THF. Similar isomerizations
were noted in our work on the oxidative cyclization of
alcohols.11 This ketol rearrangement undoubtedly involves
tautomerization to an enediol intermediate that provides for
equilibration of 7c with isomeric keto alcohol 18. The
displacement of this equilibrium strongly in favor of the
3-keto isomer 18 is remarkable, but finds precedence in
related situations.11,15
sites for intramolecular nucleophilic attack. In this instance,
the more aggressive carboxylate anion appears to favor
b-lactone formation relative to the cyclization of the free
acid, which gives only d-lactone. Geminal dimethylation at
the 2-position is probably an important factor in promoting
this regiochemistry by the well-known Thorpe–Ingold
effect,16 which may be more effective in orienting the
molecule for the more constrained cyclization to the
b-lactone relative to that for the d-lactone. It was also
considered that b-lactone 19 might have been derived from
a minor stereoisomer of spirodioxide with an inappropriate
geometry for intramolecular carboxylate displacement with
inversion at the remote site to give hydroxy lactone 7c.
However, little, if any, of the wrong spirodioxide dia-
stereomer is expected from 6c because of the influence of
the sterically large acid-bearing side-chain on epoxidation.
The b-lactone 19 was shown to be converted to d-lactone 7c
upon stirring with K2CO3 in acetone. On the other hand,
weak acid did not promote this conversion, since exposure
to acetic acid in acetone did not result in an observable
change in the 7c/19 ratio of a mixture of the two lactones. It
is quite conceivable that some of the d-lactone 7c that is
observed in the oxidation of 6c with DMDO in the
presence of NaHCO3 is derived from base-promoted
equilibration of b-lactone 19 and d-lactone 7c under the
reaction conditions.
Once again in situ oxidation of 6c cleanly produced the less
highly oxidized lactone 10c (76% yield). The addition of
p-toluenesulfonic acid to a DMDO oxidation modified the
oxidation to give a mixture of the two types of d-lactones 7c
and 10c in a ratio of 1:4. In this case, the presence of a strong
acid has diverted some of the allene oxide 8c to lactone 10c
as anticipated, although the strong acid did not prevent
substantial competing epoxidation of 8c to spirodioxide 9c,
the precursor of lactone 7c. Unlike the related reaction of
12, only d-lactones were formed; the corresponding
b-lactones were not observed as significant products.
Interestingly, DMDO oxidation of 6c in the presence of
NaHCO3 resulted in a mixture of d-lactones 7c (44% yield)
and 10c (11% yield), along with an 11% yield of a
crystalline hydroxy b-lactone 19. Thus, oxidation of 6c in
the form of its carboxylate anion once again results in some
attack at the proximal carbon of the original allene to give
the strained lactone 19. However, this time cyclization to a
b-lactone takes place at the stage of spirodioxide inter-
mediate 9c, where the carboxylate anion has a choice of
Two examples of 3,4-dienoic acids bearing a single
substituent at the remote allenic center were also examined
(Scheme 4). As expected, DMDO oxidation of acid 20a with
gem-dimethyl substitution at the 2-position gave hydroxy
d-lactone 21a as a mixture of cis (36% yield) and trans
(53% yield) isomers. The lack of stereocontrol, while
disappointing, is consistent with previous results that
Scheme 4.

J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
7031
showed poor stereoselection9 in the formation of
spirodioxides from 1,3-disubstituted allenes. Although
post-cyclization epimerization at either side of the ketone
function would also explain the randomization of stereo-
chemistry here, enolization at the alcohol center would
probably be expected to promote transposition of the keto
alcohol unit in a manner similar to the 7c to 18
rearrangement. There is no evidence for significant
involvement of such an isomerization during this oxidation
reaction, so it is likely that the initial product is a similar
mixture.
3,4-Hexadienoic acid (20b) similarly generated a 2:3
diasteriomeric mixture of cis and trans hydroxy lactones
21b in 76% yield upon reaction with DMDO. An in situ
oxidation gave only the simple d-lactone 22b albeit in a
rather mediocre isolated 42% yield and then only if the pH
was carefully controlled to avoid decomposition of the
starting material. While this result might be construed as an
indication that only the remote double bond of 20b had been
epoxidized, this is certainly not consistent with the usual
lack of regioselectivity in the epoxidation of 1,3-disubsti-
tuted allenes. The low isolated yield in this case leaves open
the possibility that some of the unobserved b-lactone 23b
was actually formed in this reaction, but that this species
selectively decomposed to uncharacterized products under
the reaction conditions. On the other hand, it should be
noted that b-lactones have so far only been observed from
acids bearing gem-dimethyl groups at the 2-position.
Whether this influence is operative in promoting cyclization
or in retarding further transformations of the strained
b-lactones is not clear.
In situ oxidation of 20a gave a mixture of the simple
d-lactone 22a (53% yield) and isomeric b-lactone 23 (53%
yield). If it is assumed that ring-closure under the conditions
of this reaction occurs by a single-step nucleophilic attack of
carboxylate anion on an allene oxide, then the two allene
oxide regioisomers 24a and 25a that are precursors for 22a
and 23, respectively, must have been generated in roughly a
2:1 ratio. The dominant stereochemistry of the reactive
allene oxides is expected to be that shown in structures 24a
and 25a, owing to the controlling steric influence of the
substitutent on the non-reacting double bond. In both cases
backside nucleophilic cyclization is possible. The same
degree of substitution at the two double bonds of the allene
unit accounts for the relatively low level of regioselectivity
in allene oxide formation, although there appears to be a
slight discrimination in favor of epoxidation at the double
bond with the sterically smaller substituent.
Several allenic acids with longer carbon-chains separating
the two functional units were also explored in order to
further define the scope of the oxidative cyclization reaction
(Scheme 5). The gem-dimethyl substituted 4,5-dienoic acid
28a gave hydroxy g-lactone 29a via the intermediate
spirodioxide 30a in good yield upon DMDO oxidation as
the free acid (82% yield), with DMDO in the presence of
cesium carbonate (100% yield), or even by the in situ
procedure (81% yield). The regiochemical course of the
lactonization process is completely reversed in this
situation. Epoxidation with this substitution pattern should
preferentially generate allene epoxide 31a. However, for the
first time under in situ conditions, cyclization of the allene
oxide intermediate 31a is not competitive with further
epoxidation, owing no doubt to the remoteness of the two
potential reaction centers that greatly slows down this
process kinetically. However, once a second epoxidation
has occurred to give spirodioxide 30a, a geometrically more
favorable nucleophilic displacement can take place at the
internal carbon center of the reactive moiety, which,
furthermore, has no other important reaction pathway
available to it. Not surprisingly, cyclization occurs via a
five-center array to give the observed hydroxy g-lactone
DMDO oxidation of the pre-formed carboxylate salt
(NaHCO3) of 20a generated a mixture of four lactones:
22a (9% yield), 23 (7% yield), 21a (37% yield), and a new
b-lactone 26 (10% yield). This result again indicates
competitive intramolecular carboxylate trapping of the
isomeric allene oxides with further oxidation to the
diastereomeric spirodioxides 27a under these conditions.
In addition to the cis and trans hydroxy d-lactones 21a that
were generated in the simple DMDO oxidation, the hydroxy
b-lactone 26 was also obtained from allenic acid 20a as a
2:1 mixture of diasteromers. This parallels the results with
7c. Thus, b-lactone formation occurs from both allene oxide
and spirodioxide intermediates in the oxidation of 20a with
DMDO and NaHCO3.
Scheme 5.

7032
J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
Scheme 6.
29a. There was no evidence for the other regiochemical
mode of cyclization to produce a seven-membered lactone
in this instance.
Finally, the observation of participation of appended
carbonyl groups of aldehydes and ketones in the ring-
opening of allene oxide and spirodioxide intermediates
during the epoxidation of allenic derivatives,17 prompted an
examination of the DMDO oxidation of 3,4-dienoate ester
37, since an analogous process involving the ester function
would generate cyclic intermediates that could, in principle,
lead to lactone products.18 Reaction of 37 with DMDO in
the presence of molecular sieves produced the acyclic
dihydroxy ketone 38 in 80% yield (Scheme 7). An in situ
reaction gave a very similar result (69% yield of 38). Ketone
38 is the hydrolysis product of spirodioxide 39, which is
formed even when only traces of water are present, as in the
first experiment. On the basis of earlier studies,9 a simple
spirodioxide with this alkyl substitution pattern should have
been stable enough to isolate, so it is quite possible that
hydrolysis of spirodioxide 39 is somehow facilitated by
nucleophilic participation of the neighboring ester function.
In any event, the exclusive formation of acyclic products in
this favorable case suggests that allenic esters are not
promising as alternate starting materials to the correspond-
ing acids for the synthesis of lactones.
The homologous 5,6-dienoic acid 28b with the same
substitution pattern behaved in a similar manner to generate
the analogous hydroxy d-lactone 29b, either with DMDO
alone (71% yield) or in the presence of Cs2CO3 (68% yield).
However, extension of the chain by an additional carbon in
6,7-dienoic acid 28c did not result in cyclization upon
DMDO oxidation, but rather gave a mixture from which
enone 32 and oxetanone 33 were isolated after treatment
with diazomethane. These rearranged materials are the
typical acid-catalyzed products expected from isomeriza-
tion of a simple trisubstituted spirodioxide of this type.9 The
corresponding acids were present in the crude reaction
mixture. Obviously, cyclization to a seven-member lactone
cannot compete with the acid-catalyzed decomposition of
spirodioxide 30b under these conditions.
A different outcome was demonstrated with the unsubsti-
tuted 4,5-hexadienoic acid (34a), which reacted with
DMDO to give the simple g-lactone 35a as the only
important isolated product, albeit in a very low 18% isolated
yield (Scheme 6). Other major products were not observed
here, but might have been lost because of their high polarity.
The homologous 5,6-heptadienoic acid (34b) gave the
corresponding d-lactone 35b in a better 68% yield with
DMDO in the presence of Cs2CO3. Interestingly, the
unsymmetrically disubstituted 6,7-dienoic acid 34c, bearing
an additional methyl group at the internal carbon of the
allene, was smoothly converted by DMDO in the presence
of K2CO3 to the seven-membered 1-lactone 35c in 61%
yield. The substitution pattern of allenic acids 34 is the
determining factor in this series. The lack of substitution at
the allene terminus results in the epoxidation at the internal
double bond to give allene oxides 36 as intermediates on the
way to the simple five to seven-ring lactones 35. A
geometrically favorable intramolecular nucleophilic attack
on the allene oxide unit of intermediates 36 is the
predominant reaction pathway here, especially with the
more reactive carboxylate anions. Thus, once again there is
a clear preference for cyclization over further epoxidation to
a spirodioxide with these terminal allenic acids as was
observed with allenic acid 12.
In summary, the DMDO oxidation of a variety of allenic
acids has been shown to provide highly functionalized
lactones with a range of structural parameters. These
transformations proceed via unisolated allene oxides and
spirodioxides as intermediates. The exact nature of the
product(s) depends on a number of parameters which
control the product-determining intermediates and the mode
of their further transformations. The structure of the
reactant, particularly the substitution pattern, is important
in controlling the site of the initial epoxidation and the
propensity for a second epoxidation relative to competitive
cyclizations by intramolecular attack by the acid function.
More substituted allenes typically undergo sequential
epoxidation prior to lactone formation, whereas terminal
allenes tend to react at the allene oxide stage. The
lactonization regiochemistry is usually determined by the
relative location of the allene and acid groups in the starting
material and shows a decided preference for the formation
of normal-sized rings when reacting via the spirodioxides.
The reaction conditions can also be critical to the course of
the reaction. In general, most oxidations of the free acids
proceed via spirodioxides using pre-prepared solutions of
Scheme 7.

J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
7033
DMDO. On the other hand, the in situ method typically
involves cyclization at the allene oxide stage. Oxidation of
carboxylate anions with DMDO usually gave mixtures of
lactones derived from both types of reactive intermediates,
but in several cases these conditions resulted in strained
b-lactones. At this stage, reasonable predictions can be
made on the course of oxidative cyclizations of allenic
acids.
3.2.2. 3-Hydroxy-2,5-dimethyl-4-oxo-5-hexanolide (7b).
Reaction of 75 mg of 6b with 40 ml of DMDO gave after
recrystallization (hexane) 85 mg (92%) of a white solid 3:2
mixture of cis and trans 7b: IR 3154, 1722, 1279,
1093 cm21; HRMS (CI) m/z (rel intensity) 173 (87), 155
(100), 137 (36), 116 (43), 114 (14), 86 (35), 71 (19); exact
1
mass 173.0813; calcd for C8H13O4 173.0813; cis 7b: H
NMR d 4.84 (dd, 1, J¼6.5, 3.5 Hz), 3.37 (qd, 1, J¼7.8,
6.5 Hz), 3.32 (d, 1, J¼3.5 Hz), 1.60 (s, 3), 1.56 (s, 3), 1.20
(d, 3, J¼7.8 Hz); 13C NMR d 209.2, 170.4, 85.6, 70.4, 41.4,
1
27.4, 25.7, 11.2; trans 7b: H NMR d 4.29 (dd, 1, J¼13.2,
3. Experimental
3.1. General procedures
3.0 Hz), 3.39 (d, 1, J¼3.0 Hz), 2.67 (dq, 1, J¼13.4, 6.7 Hz),
1.59 (s, 3), 1.56 (s, 3), 1.47 (d, 3, J¼6.7 Hz); 13C NMR d
209.2, 170.2, 85.3, 72.1, 41.0, 26.8, 26.0, 12.7.
All nuclear magnetic resonance (NMR) spectra were
1
To a solution of 77 mg (0.45 mmol) of this mixture and
79 ml (0.68 mmol) of 2,6-lutidine in 5 ml of CH2Cl2 at
2788C was added 180 ml of TBDMSOTf. After stirring for
4 h, the solution was diluted with ether, washed with sat.
CuSO4 and brine, dried (MgSO4) and concentrated. Flash
chromatography (20% ether/hexane) afforded 65 mg (50%)
of the TBDMS ether of trans 7b: IR 1742, 1257, 1168,
obtained on CDCl3 solutions unless otherwise noted. H
and 13C NMR spectra were recorded on a Varian VXR-400
spectrometer at 400 and 100 MHz or a Bruker AM-500
spectrometer at 500 and 125 MHz, respectively. Infrared
(IR) spectra were obtained with a Perkin-Elmer 298 IR
spectrometer or a Mattson 4020 Galaxy Series FT-IR.
Samples were prepared as thin films on NaCl plates. Mass
spectra (HRMS) were obtained on a Kratos MS 80 RFAQQ
spectrometer using either chemical (CI) or electron-impact
(EI) ionization. Melting points were determined on a
Thomas-Hoover Unimelt apparatus and are uncorrected.
1
1121 cm21; H NMR d 4.27 (d, 1, J¼12 Hz), 2.82 (dq, 1,
J¼12, 6 Hz), 1.56 (s, 3), 1.51 (s, 3), 1.38 (d, 3, J¼6 Hz),
0.93 (s, 9), 0.17 (s, 3), 0.06 (s, 3); 13C NMR d 207.2, 171.0,
85.5, 73.7, 41.5, 26.9, 26.3, 25.6, 18.5, 13.0; HRMS (CI) m/z
(rel intensity) 287 (2), 229 (8), 201 (12), 173 (18), 157 (42),
143 (100), 129 (68), 115 (39), 75 (52); exact mass 287.166;
calcd for C14H27O4 287.1678 and 23 mg (34%) of 3,5-
dimethyl-4-oxo-5-hex-2-enolide (11): IR 1735, 1690, 1623,
Diethyl ether and tetrahydrofuran (THF) were distilled from
sodium benzophenone ketyl under a nitrogen atmosphere.
Dichloromethane (CH2Cl2) was distilled over calcium
hydride under nitrogen. 2,6-Lutidine was distilled from
potassium hydroxide and stored over molecular sieves.
t-Butyldimethylsilyl triflate (TBDMSOTf) was distilled
before use. Reagent grade acetone was used in the
preparation of dimethyldioxirane (DMDO). High Pressure
Liquid Chromatography (HPLC) was performed on a Rainin
HPLC using a Dynamax-60A silica gel column with HPLC
grade solvents. Silica gel 60 (230–400 mesh, E. Merck) was
used for column chromatography. Preparative thin-layer
chromatography was performed on Kieselgel 60 F-254
silica gel on 10£20 cm2 plates of 0.25 mm thickness.
1262, 1110 cm21
;
1H NMR d 6.58 (s, 1), 2.19 (d, 3,
J¼1 Hz), 1.58 (s, 6).
3.2.3. 3-Hydroxy-2,2,5-trimethyl-4-oxo-5-hexanolide
(7c). Reaction of 100 mg of 6c with 30 ml of DMDO gave
116 mg (96%) of 7c as a white solid after recrystallization
(hexane), mp 95–968C; IR 3520, 2980, 1760, 1745,
1
1120 cm21; H NMR d 4.54 (d, 1, J¼4 Hz), 3.33 (d, 1,
J¼4 Hz), 1.62 (s, 3), 1.57 (s, 3), 1.49 (s, 3), 1.08 (s, 3); 13C
NMR d 209.0, 173.1, 85.4, 74.4, 45.2, 27.6, 25.4, 23.6, 19.0;
Anal. calcd for C9H14O4: C, 58.05; H, 7.58. Found: C,
58.22; H, 7.63.
3.2. General procedure for DMDO oxidations of allenic
acids
3.2.4. 2,2-Dimethyl-4-oxo-3-pentanolide (13). Reaction of
100 mg of 12 with 50 ml of DMDO gave after flash
chromatography (20% ether/hexane) 62 mg (52%) of 13 as a
To 1 equiv. of allenic acid was added an excess
(6–10 equiv.) of a cold solution of DMDO11 in acetone.
After 0.5–2 h, the reaction mixture was concentrated and
the residue was diluted with methylene chloride, dried
(MgSO4), and concentrated. The product was purified by
recrystallization, flash chromatography, HPLC, etc.
1
white solid: IR 1833, 1728, 1184, 1098 cm21; H NMR d
4.54 (s, 1), 2.30 (s, 3), 1.54 (s, 3), 1.23 (s, 3); 13C NMR d
204.4, 173.1, 82.6, 57.7, 27.7, 22.4, 17.6; HRMS (CI) m/z
(rel intensity) 143 (11), 117 (11), 114 (23), 99 (30), 85 (35),
83 (100), 72 (52), 70 (28); exact mass 143.070; calcd for
C7H11O3 143.0708.
3.2.1. 3-Hydroxy-5-methyl-4-oxo-5-hexanolide (7a).
Reaction of 50 mg of 6a with 40 ml of DMDO gave
55 mg (87%) of 7a after recrystallization (hexane) as a
3.2.5. 3-Hydroxy-2,2-dimethyl-4-oxo-5-octanolide (21a).
Reaction of 100 mg of 20a with 50 ml of DMDO gave a
mixture that was separated by HPLC (30% ethyl acetate/
hexane) to give 36 mg (30%) of trans 21a as an oil: IR 3507,
2968, 1751, 1734, 1116 cm21; 1H NMR d 4.81 (ddd, 1, J¼6,
3, 1 Hz), 4.42 (d, 1, J¼1 Hz), 3.38 (br s, 1), 2.09 (m, 1), 1.62
(m, 3), 1.46 (s, 3), 1.08 (s, 3), 0.97 (t, 3, J¼7 Hz); 13C NMR
d 206.1, 173.1, 82.5, 76.1, 44.5, 34.5, 23.3, 19.4, 18.7, 13.4;
Anal. calcd for C10H16O4: C, 59.98; H, 8.05. Found: C,
60.11; H, 7.94 and 63 mg (53%) of cis 21a as a white solid,
white solid: IR 3530–3300, 1753, 1737, 1270, 1130 cm21
;
1H NMR d 4.62 (dd, 1, J¼14, 7 Hz), 3.30 (dd, 1, J¼17,
7 Hz), 2.72 (dd, 1, J¼17, 14 Hz), 2.15 (s, 1), 1.58 (s, 3), 1.55
(s, 3); 13C NMR d 209.2, 166.8, 85.7, 67.6, 36.4, 26.9, 25.9;
Anal. calcd for C7H10O4: C, 53.16; H, 6.37. Found: C,
52.73; H, 6.29. Characteristic NMR signals corresponding
to a small amount of 2-hydroxy-2-methylpent-4-en-3-one19
were observed in the crude product.

7034
J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
mp 79–808C: IR 3511, 2980, 1753, 1736, 1105 cm21; H
NMR d 4.74 (t, 1, J¼6 Hz), 4.33 (s, 1), 3.70 (s, 1), 1.88 (m,
2), 1.52 (m, 2), 1.46 (s, 3), 1.07 (s, 3), 0.95 (t, 3, J¼7 Hz);
13C NMR d 207.7, 173.2, 81.9, 75.8, 44.4, 34.7, 22.8, 17.9,
17.6, 13.5; MS (CI) m/z (rel intensity) 201 (11), 183 (3), 144
(5), 83 (6), 72 (100); exact mass 201.113; calcd for
C10H16O4 201.1127.
3.2.11. Methyl 3,5-dihydroxy-2,2,5-trimethyl-4-oxo-hex-
anoate (38). Reaction of 50 mg of 37 and 55 ml of DMDO
˚
in the presence of 1 g of powdered 3 A molecular sieves
1
gave after flash chromatography (10% EtOAc/hexane)
52 mg (80%) of 38 as an oil: IR 3464, 1721, 1153,
1047 cm21; 1H NMR d 4.43 (br s, 1), 3.37 (s, 3), 3.58 (br s,
1), 1.40 (s, 3), 1.38 (s, 3), 1.30 (s, 3), 1.18 (s, 3); 13C NMR d
212.4, 178.0, 78.3, 77.8, 52.3, 27.4, 26.3, 22.5, 22.2; HRMS
(CI) m/z (rel intensity) 219 (6), 187 (12), 169 (18), 132 (41),
114 (89), 99 (37), 83 (66), 71 (31), 59 (73), 48 (100); exact
mass 219.122; calcd for C10H19O5 219.1232
3.2.6. 3-Hydroxy-4-oxo-5-hexanolide (21b). Reaction of
32 mg of 20b with 30 ml of DMDO gave 30 mg (76%) of
21b as a 2:3 mixture of stereoisomers: IR 3510, 1753,
1726 cm21; HRMS (CI) m/z (rel intensity) 145 (100), 144
(10), 129 (8), 127 (54), 109 (40), 101 (10), 88 (62); exact
mass 145.048; calcd for C6H9O4 145.0500; One isomer
3.3. DMDO oxidations in the presence of carbonate
1
shows: H NMR d 5.03 (q, 1, J¼6 Hz), 4.60 (m, 1), 3.08
The oxidations were performed in a manner similar to the
other DMDO reactions except that the acids were stirred
with NaHCO3 (0.5 g), K2CO3 (1.5 equiv.), or Cs2CO3
(1 equiv.) in 1–5 ml of acetone prior to the addition of
DMDO.
(m,1), 2.99 (m, 1), 1.43 (d, 3, J¼6 Hz); 13C NMR d 205.2,
1
170.9, 79.3, 68.2, 37.5, 16.7. The other isomer shows: H
NMR d 5.03 (q, 1, J¼6 Hz), 4.52 (m, 1), 3.22 (m, 1), 2.93
(m, 1), 1.45 (d, 3, J¼6 Hz); 13C NMR d 204.8, 170.7,
79.0, 68.5, 37.0, 17.8. The 13C NMR data was assigned
to the two isomers on the basis of a correlation
between the proton and carbon data from a HETCOR
NMR study.
3.3.1. Oxidation of 6a in the presence of NaHCO3. From
100 mg of 6a was obtained after flash chromatography (20%
ethyl acetate/hexane) 6 mg (10%) of 10a and 33 mg (52%)
of 7a.
3.2.7. 6-Hydroxy-6-methyl-5-oxo-4-heptanolide (29a).
Reaction of 75 mg of 28a with 30 ml of DMDO gave
after molecular distillation 100 mg (82%) of 29a as a clear
oil: IR 3612, 3508, 1784, 1730, 1157 cm21; 1H NMR d 5.49
(m, 1), 3.06 (br s, 1), 2.53 (m, 3), 2.26 (m, 1), 1.43 (s, 3),
1.42 (s, 3); 13C NMR d 209.6, 176.6, 77.9, 76.8, 26.8, 26.7,
25.1; HRMS (CI) m/z (rel intensity) 173 (78), 155 (46), 113
(11), 95 (41), 85 (27), 69 (11); exact mass 173.082; calcd for
C8H13O4 173.0814.
3.3.2. Oxidation of 6b in the presence of NaHCO3. From
100 mg of 6b was obtained after flash chromatography
(10% ethyl acetate/hexane) 31 mg (28%) of 10b and 66 mg
(54%) of a 3:2 mixture of cis and trans 7b.
3.3.3. Oxidation of 6c in the presence of NaHCO3. From
300 mg of 6c was obtained after flash chromatography
159 mg (44%) of 7c, 37 mg (11%) of 10c, and 41 mg (11%)
of 5-hydroxy-2,2,5-trimethyl-4-oxo-3-hexanolide (19) as a
white solid, mp 92–938C: IR 3620, 2984, 1840, 1740,
3.2.8. 7-Hydroxy-7-methyl-6-oxo-5-octanolide (29b).
Reaction of 57 mg of 28b with 40 ml of DMDO gave
after molecular distillation 48 mg (71%) of 29b as a clear
oil: IR 3477, 1742, 1728, 1110 cm21; 1H NMR d 5.53 (t, 1,
J¼6 Hz), 2.94 (br s, 1), 2.56 (m, 2), 2.16 (m, 2), 1.87 (m, 2),
1.43 (s, 3),1.41 (s, 3); 13C NMR d 210.0, 170.2, 79.5, 77.2,
29.5, 27.1, 26.9, 24.1, 17.2; Anal. calcd for C9H14O4: C,
58.05; H, 7.58. Found: C, 57.99; H, 7.42.
1
1100 cm21; H NMR d 5.18 (s, 1), 2.24 (s, 1), 1.55 (s, 3),
1.42 (s, 3), 1.41 (s, 3), 1.23 (s, 3); 13C NMR d 207.5, 173.2,
81.2, 77.1, 58.6, 27.6, 25.9, 21.7, 18.1; Anal. calcd for
C9H14O4: C, 58.05; H, 7.58. Found: C, 57.99; H, 7.63.
3.3.4. Oxidation of 12 in the presence of NaHCO3.
Reaction of 100 mg of 12 gave after flash chromatography
(20% ether/hexane) 100 mg (84%) of 13.
3.2.9. Oxidation of 8-methyl-6,7-nonadienoic acid (28c)
with DMDO. Reaction of 132 mg of 28c and 35 ml of
DMDO gave a crude product that was dissolved in 10 ml of
methanol and treated with an ethereal solution of diazo-
methane until the yellow color persisted. Flash chromato-
graphy (40% ether/hexane) gave 42 mg (25%) of
8-carbomethoxy-4-hydroxy-2-methylocta-1-en-3-one (32)
3.3.5. Oxidation of 20a in the presence of NaHCO3.
Reaction of 100 mg of 20a gave after flash chromatography
(50% ether/pentane) 8 mg (7%) of 22a, 10 mg (9%) of 23,
44 mg (37%) of 21a as a 2:3 mixture of cis and trans
isomers and 12 mg (10%) of a 2:1 mixture of diastereomers
of 5-hydroxy-2,2-dimethyl-4-oxo-3-pentanolide (26) as a
clear oil: IR 3522, 1835, 1724 cm21; 1H NMR d 4.65 (m, 2),
4.11 (m, 1, J¼7, 2 Hz), 4.04 (m, 1), 2.33 (d, 1, J¼3 Hz),
1.75 (d, 1, J¼3 Hz), 1.74–1.32 (m, 4), 1.32–1.31 (four
overlapping singlets, 12), 0.98 (t, 3, J¼7 Hz), 0.96 (t, 3,
J¼7 Hz); 13C NMR d 212.0, 211.1, 178.0, 85.3, 71.7, 70.7,
44.7, 44.2, 35.6, 33.7, 22.0, 21.8, 19.4, 19.2, 18.9, 18.6,
13.7, 13.6; HRMS (CI) m/z (rel intensity) 201 (16), 183 (9),
155 (7), 128 (46), 113 (6), 100 (100), 73 (30), 70 (33), 55
(64), 41 (27); exact mass 201.112; calcd for C10H17O4
201.1126.
as an oil: IR 3410, 3105, 1734, 1671, 1631, 1215 cm21
;
1H NMR (d 5.91 (m, 2), 4.80 (br s, 1), 3.66 (s, 3), 3.47 (s, 1),
2.30 (t, 2, J¼7 Hz), 1.93 (dd, 3, J¼2, 1 Hz), 1.81 (m, 2),
1.65 (m, 2), 1.50 (m, 2) and 24 mg (14%) of 4-(4-
carbomethoxybutyl)-2,2-dimethyl-3-oxacyclobutanone (33)
as an oil: IR 1816, 1734, 1215 cm21; 1H NMR d 5.28 (t, 1,
J¼7 Hz), 3.66 (s, 3), 2.32 (t, 2, J¼7 Hz), 1.81 (m, 2), 1.66
(m, 2), 1.50 (m, 2), 1.47 (s, 3), 1.45 (s, 3).
3.2.10. 5-Oxo-4-hexanolide (35a).20 Reaction of 50 mg of
34a with 30 ml of DMDO gave 10 mg (18%) of 35a as an
1
oil: IR 1785, 1718 cm21; H NMR d 4.81 (m, 1), 2.54 (m,
3), 2.30 (m, 2), 2.25 (m, 2).
3.3.6. Oxidation of 28a in the presence of Cs2CO3.
Reaction of 26 mg of 28a gave 32 mg (100%) of 29a.

J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
7035
3.3.7. Oxidation of 28b in the presence of Cs2CO3.
Reaction of 57 mg of 28b gave after molecular distillation
25 mg (83%) of 29b.
after flash chromatography (25% ether/pentane) 210 mg
(53%) of 2,2-dimethyl-4-oxo-5-octanolide (22a) as a white
solid, mp 74–758C: IR 1747, 1734, 1122 cm21; 1H NMR d
4.70 (dd, 1, J¼8, 4 Hz), 2.69 (d, 1, J¼16 Hz), 2.51 (d, 1,
J¼16 Hz), 1.85 (m, 2), 1.50 (m, 2), 1.36 (s, 3), 1.28 (s, 3),
0.94 (t, 3, J¼7 Hz); 13C NMR d 206.3, 174.5, 84.0, 48.3,
38.5, 34.3, 26.4, 24.7, 17.9, 13.6; HRMS (CI) m/z (rel
intensity) 185 (97), 167 (18), 156 (29), 142 (50), 128 (25),
83 (24), 71 (100); Anal. calcd for C10H16O3: C, 65.19; H,
8.75. Found: C, 65.22; H, 8.72 and 100 mg (25%) of 2,2-
dimethyl-4-oxo-3-octanolide (23) as a clear oil: IR 1836,
3.3.8. 6-Oxo-5-heptanolide (35b). Reaction of 94 mg of
K2CO3 and 57 mg of 34b gave after preparative TLC (ether)
45 mg (68%) of 35b as an oil: IR 1727, 1716, 1286, 1172,
1091, 1020 cm21; 1H NMR d 4.75 (t, 1, J¼6 Hz), 2.60 (m,
2), 2.31 (s, 3), 2.15 (m, 1), 1.89 (m, 3); 13C NMR d 206.4,
173.1, 80.5, 38.2, 30.6, 27.1, 26.9; HRMS (CI) m/z (rel
intensity) 143 (20), 99 (100), 84 (26), 71 (89), 55 (81); exact
mass 143.069; calcd for C7H11O3 143.0708.
1
1716, 1096, 913, 747 cm21; H NMR d 4.55 (s, 1), 2.57
(sym m, 2), 1.59 (m, 2), 1.54 (s, 3), 1.35 (m, 2), 1.20 (s, 3),
0.92 (t, 3, J¼7 Hz).
3.3.9. 6-Methyl-7-oxo-6-octanolide (35c). Reaction of
166 mg of K2CO3 and 148 mg of 34c gave 100 mg (61%)
of 35c as an oil: IR 2942, 1728, 1715, 1172, 1091, 1020,
913 cm21; 1H NMR d 2.72 (m, 1), 2.37 (m, 1), 2.26 (m, 2),
2.25 (s, 3), 1.85 (m, 1), 1.74 (m, 1), 1.61 (m, 2), 1.50 (s, 3);
13C NMR d 207.4, 174.0, 87.8, 36.7, 36.6, 26.9, 25.0, 24.8,
22.5; Anal. calcd for C9H14O3: C, 63.51; H, 8.29. Found: C,
63.20; H, 7.98.
3.4.6. 4-Oxo-5-hexanolide (21b).20 This reaction was
modified by the addition of 1 g of NaHCO3 after each 7-g
portion of Oxone using 200 mg of 20b to give after flash
chromatography (20% ethyl acetate/hexane) 96 mg
(42%) of 22a as a white solid, mp 74–758C: IR 1730,
1
1274, 1123, 1072 cm21; H NMR d 4.78 (q, 1, J¼6 Hz);
2.93 (m, 2), 2.74 (m, 2), 1.51 (d, 3, J¼6 Hz); 13C NMR d
205.4, 170.0, 79.5, 33.2, 28.2, 15.9; HRMS (CI) m/z (rel
intensity) 129 (100), 167 (7), 111 (90), 100 (28), 86 (42), 84
(64), 56 (86); exact mass 129.056; calcd for C10H17O3
129.0552.
3.4. In situ oxidations
To a mixture of the allenic acid, 8 g (96 mmol) of sodium
bicarbonate, 20 ml of acetone, 10 ml of methylene chloride
and 60 ml of water was slowly added 100 g (160 mmol) of
Oxone with vigorous stirring. After 2 h, the mixture was
filtered through Celite and the filtrate was extracted with
ether. The extracts were dried (MgSO4), concentrated and
purified.
3.4.7. 6-Hydroxy-6-methyl-5-oxo-4-heptanolide (29a).
Reaction of 200 mg of 28a gave after flash chromatography
(30% ether/hexane) 200 mg (81%) of 29a.
3.4.8. Methyl 3,5-dihydroxy-2,2,5-trimethyl-4-oxo-
hexanoate (38). Reaction of 200 mg of 37 gave after
chromatography (30% ethyl acetate/hexane) 160 mg (69%)
of 38.
3.4.1. 5-Methyl-4-oxo-5-hexanolide (10a). Reaction of
160 mg of 6a gave after flash chromatography (25% ethyl
acetate/hexane) 110 mg (67%) of 10a as an oil: IR 1733,
1727, 1130 cm21; 1H NMR d 2.90 (m, 2), 2.75 (m, 2), 1.52
(s, 3), 1.51 (s, 3); 13C NMR d 207.3, 169.1, 86.4, 33.3, 28.4,
25.9; HRMS (CI) m/z (rel intensity) 143 (100), 142 (30), 125
(54), 114 (53), 70 (13); exact mass 143.071; calcd for
C7H10O3 143.0708.
3.5. Oxidations in the presence of p-toluenesulfonic acid
Reactions were performed as usual except that a solution of
1 equiv. of allenic acid in 5–10 ml of methylene chloride
was added to a solution of 2 equiv. of p-toluenesulfonic acid
in 6–10 equiv. of DMDO in acetone.
3.4.2. 2,5-Dimethyl-4-oxo-5-hexanolide (10b). Reaction of
200 mg of 6b gave after recrystallization (hexane) 164 mg
((74%) of 10b as a white solid, mp 72–758C: IR 1734, 1726,
3.5.1. Oxidation of 6c. Reaction of 50 mg of 6c gave 53 mg
(88%) of a 1:4 mixture of 10c and 7c.
1
1210 cm21; H NMR d 2.83 (m, 1), 2.62 (dd, 1, J¼17,
5 Hz), 2.55 (dd, 1, J¼17, 13 Hz), 1.46 (s, 3), 1.42 (s, 3), 1.26
(d, 3, J¼7 Hz); 13C NMR d 207.8, 172.4, 87.0, 41.0, 33.3,
26.3, 25.5, 16.0; HRMS (CI) m/z (rel intensity) 157 (20),
139 (40), 128 (37), 70 (59), 69 (46), 59 (100); exact mass
157.087; calcd for C8H13O3 157.0865.
3.5.2. Oxidation of 12. Reaction of 100 mg of 12 after
pouring into sat. NaHCO3, followed by drying and
concentration of the organic layer and flash chromatography
(30% ethyl acetate/hexane) of the residue gave 17 mg (15%)
of 13, 68 mg (43%) of 16 as a white solid, mp 50–518C: IR
1
3.4.3. 2,2,5-Trimethyl-4-oxo-5-hexanolide (10c). Reaction
of 480 mg of 6c gave after recrystallization (hexane)
400 mg (76%) of 10c as a white solid, mp 94–958C; IR
1773, 1282, 1223, 1086 cm21; H NMR d 4.17 (s, 1), 1.76
(s, 3), 1.45 (s, 3), 1.41 (s, 3), 1.32 (s, 3), 1.30 (s, 3); 13C
NMR d 178.8, 112.8, 110.6, 87.9, 46.1, 27.7, 27.1, 23.9,
23.7, 19.0; HRMS (CI) m/z (rel intensity) 201 (27), 185 (42),
143 (38), 115 (41), 98 (60), 83 (100); exact mass 201.112;
calcd for C10H17O4 201.1126 and 55 mg (32%) of 17 as a
white solid, mp 105–1078C: IR 3620, 3475, 1778, 1218,
1
2980, 1737, 1726, 1120 cm21; H NMR d 2.67 (s, 2), 1.54
(s, 6), 1.32, (s, 6); 13C NMR d 207.9, 174.4, 87.9, 46.9, 39.6,
26.5, 26.4; Anal. calcd for C9H14O3: C, 63.51; H, 8.29.
Found: C, 63.77; H, 8.36.
1
1078 cm21; H NMR d 4.48 (d, 1, J¼10 Hz), 4.14 (d, 1,
3.4.4. 2,2-Dimethyl-4-oxo-3-pentanolide (13). Reaction of
200 mg of 12 gave 170 mg (72%) of 13.
J¼10 Hz), 4.09 (d, 1, J¼4 Hz), 2.64 (d, 1, J¼4 Hz), 1.52 (s,
3), 1.44 (s, 3), 1.36 (s, 3), 1.19 (s, 3); 13C NMR d 179.2,
113.2, 111.7, 79.9, 70.3, 44.7, 26.3, 25.9, 23.4, 18.2; HRMS
(CI) m/z (rel intensity) 217 (100), 201 (74), 199 (18), 159
3.4.5. Oxidation of 20a. Reaction of 360 mg of 20a gave

7036
J. K. Crandall, E. Rambo / Tetrahedron 58 (2002) 7027–7036
Hydroxyl Groups and their Sulphur Analogues; Patai, S.,
Ed.; Wiley: New York, 1980; pp 859–879. (b) Chan, T. H.;
Ong, B. S. Tetrahedron 1980, 36, 2269–2289.
2. Crandall, J. K.; Machleder, W. H. Tetrahedron Lett. 1966,
6037–6041.
(55), 117 (29), 114 (25), 101 (23), 72 (90), 59 (63); exact
mass 217.107; calcd for C10H17O5 217.1075.
3.5.3. 4-Hydroxy-2,2,5-trimethyl-3-oxo-5-hexanolide
(19). Silica gel (ca. 2 g) was added to a solution of 50 mg
of 7c in 10 ml of THF until a thick slurry was formed. After
24 h, the mixture was filtered through Celite and concen-
trated to give a yellow liquid. Preparative TLC (30%
ether/pentane) gave 40 mg (80%) of 19: IR 3750, 2985,
1753, 1738, 1125 cm21; 1H NMR d 4.50 (s, 1), 2.15 (br s,1),
1.41 (s, 3), 1.39 (s, 3), 1.33 (s, 3), 1.32 (s, 3); 13C NMR d
211.2, 177.7, 88.3, 72.1, 44.4, 26.5, 24.9, 22.2, 19.3; Anal.
calcd for C9H14O4: C, 58.05; H, 7.58. Found: C, 58.04; H,
7.65.
3. Crandall, J. K.; Machleder, W. H.; Sojka, S. A. J. Org. Chem.
1973, 38, 1149–1154.
4. Camp, R. L.; Greene, F. D. J. Am. Chem. Soc. 1968, 90, 7349.
5. Crandall, J. K.; Conover, W. W.; Komin, J. B.; Machleder,
W. H. J. Org. Chem. 1974, 39, 1723–1729.
6. Crandall, J. K.; Machleder, W. H. J. Am. Chem. Soc. 1968, 90,
7347–7349.
7. Crandall, J. K.; Machleder, W. H.; Thomas, M. J. J. Am. Chem.
Soc. 1968, 90, 7346–7347.
8. Murray, R. W.; Jeyaraman, R. J. Org. Chem. 1985, 50,
2847–2853.
3.6. Isomerization of 19 to 7c
9. Crandall, J. K.; Batal, D. J.; Sebesta, D. P.; Lin, F. J. Org.
Chem. 1991, 56, 1153–1166.
3.6.1. Isomerization by K2CO3. A mixture of 21 mg of 19
and 20 mg of K2CO3 in 20 ml of acetone was stirred for
40 min. Concentration, dilution with CH2Cl2, drying
(MgSO4), and concentration gave a sample whose 1H
NMR spectrum indicated a 2:1 mixture of lactones 7c and
19.
10. (a) Baumstark, A. L.; McCloskey, C. J. Tetrahedron Lett.
1987, 28, 3311–3314. (b) Baumstark, A. L.; Vasquez, P. C.
J. Org. Chem. 1988, 53, 3437–3439.
11. Crandall, J. K.; Batal, D. J.; Lin, F.; Reix, T.; Nadol, G. S.; Ng,
R. A. Tetrahedron 1992, 48, 1427–1448.
12. Bertrand, M.; Dulcere, J. P.; Gil, G. Tetrahedron Lett. 1980,
21, 1945–1948. Bertrand, M.; Dulcere, J. P.; Gil, G.
Tetrahedron Lett. 1977, 18, 3807–3810. Bertrand, M.;
Dulcere, J. P.; Gil, G.; Grimaldi, J.; Sylvestre-Panthet, P.
Tetrahedron Lett. 1976, 17, 3305–3308.
3.6.2. Stability of 19 to acetic acid. To 30 mg of a mixture
of lactones 7c and 19 in an NMR tube was added 5 ml of
water, 10 ml of acetic acid and 1 ml of acetone-d6. After 2 h,
1
the ratio of 7c and 19 was unchanged by H NMR.
13. For a preliminary publication, see: Crandall, J. K.; Rambo, E.
J. Org. Chem. 1990, 55, 5929–5930.
Acknowledgments
14. PCMODEL v7.5; Serena Software, Box 3076, Bloomington,
IN 47402-3076.
This work was supported by a grant from the PHS (GM-
39988). E. R. received a GAANN fellowship on a grant
from the Department of Education. We thank the expert
help from Dr Feng Lin in conducting heteronuclear
correlated 2D NMR spectroscopy and Dr Kevin Gilbert
with the molecular modeling.
15. Crandall, J. K.; Reix, T. Tetrahedron Lett. 1994, 35,
2513–2516.
16. Kirby, A. J. Adv. Phys. Org. Chem. 1980, 17, 183–278.
17. Crandall, J. K.; Rambo, E. Tetrahedron Lett. 1994, 35,
1489–1492.
18. Kolsaker, P.; Boerresen, S. Acta Chem. Scand., Ser. 1979,
B33, 133–137.
19. Hoff, S.; Brandsma, L.; Arens, J. F. Rec. Trav. Chim. 1968, 87,
159–161.
References
20. Jensen, J. E.; Torssell, K. Acta Chem. Scand., Ser. 1978, B32,
457–459.
1. (a) Stang, P. J. In The Chemistry of Functional Groups,
Supplement E. The Chemistry of Ethers, Crown Ethers,