Synthesis of Polysubstituted γ-Butenolides via a Radical Pathway: Cyclization of α-Bromo Aluminium Acetals and Comparison with the Cyclization of α-Bromoesters at High Temperature

Article abstract of DOI:10.1002/chem.201501294

Polysubstituted butenolides were obtained in good to high yields from α-bromoesters derived from propargyl alcohols by a one-pot reaction involving the radical cyclization of α-bromo aluminium acetals, followed by the oxidation of the resulting cyclic aluminium acetals in an Oppenauer-type process and migration of the exocyclic C-C bond into the α,β-position. Comparison with the direct cyclization of α-bromoesters at high temperature and under high dilution conditions is described. Deuterium-labelling experiments allowed us to uncover "invisible" 1,5-hydrogen atom transfers (1,5-HATs) that occur during these cyclization processes, together with the consequences of the latter in the epimerization of stereogenic centres. Compared to the classical approach, the cyclization of aluminium acetals proved to be highly chemoselective and its efficiency was illustrated by the short total syntheses of optically enriched γ-butenolides isolated from Plagiomnium undulatum and from Kyrtuhrix maculans.

Full text of DOI:10.1002/chem.201501294

DOI: 10.1002/chem.201501294
Full Paper
&
Radical Reactions |Hot Paper|
Synthesis of Polysubstituted g-Butenolides via a Radical Pathway:
Cyclization of a-Bromo Aluminium Acetals and Comparison with
the Cyclization of a-Bromoesters at High Temperature
Romain BØnØteau,[a] Carole F. Despiau,[a] Jean-Christophe Rouaud,[a] Anne Boussonni›re,[a, b]
Virginie Silvestre,[a] Jacques Lebreton,[a] and Fabrice DØn›s*[a]
Abstract: Polysubstituted butenolides were obtained in
good to high yields from a-bromoesters derived from prop-
argyl alcohols by a one-pot reaction involving the radical
cyclization of a-bromo aluminium acetals, followed by the
oxidation of the resulting cyclic aluminium acetals in an Op-
penauer-type process and migration of the exocyclic C=C
bond into the a,b-position. Comparison with the direct cycli-
zation of a-bromoesters at high temperature and under
high dilution conditions is described. Deuterium-labelling ex-
periments allowed us to uncover “invisible” 1,5-hydrogen
atom transfers (1,5-HATs) that occur during these cyclization
processes, together with the consequences of the latter in
the epimerization of stereogenic centres. Compared to the
classical approach, the cyclization of aluminium acetals
proved to be highly chemoselective and its efficiency was il-
lustrated by the short total syntheses of optically enriched g-
butenolides isolated from Plagiomnium undulatum and from
Kyrtuhrix maculans.
Introduction
g-Butenolides represent an important class of compounds as
part of many natural compounds, which display a wide range
of biological activities (Figure 1). Moreover, they can be regard-
ed as functionalized chiral building blocks for the elaboration
of more complex structures, for these cyclic compounds can
be easily converted into diastereomerically pure material.[1]
Among the synthetic methods available to build or decorate
the butenolide skeleton,[2] the ring-closing metathesis ap-
proach has proved to be particularly efficient.[3] This strategy,
however, is limited to the preparation of butenolides for which
the C=C bond is di- or trisubstituted, as the synthesis of the
more sterically demanding tetrasubstituted butenolides
proved inefficient.[4] For this purpose, the ruthenium-catalysed
cyclocarbonylation[5] and the carbocupration-carboxylation[6] of
allenyl alcohols, as well as the transition-metal catalysed cycli-
zation of 2,3-dienyl carboxylic acid derivatives[2e,f,7] have
proved to be much more valuable.
Figure 1. Selected examples of natural compounds containing a g-buteno-
lide skeleton.
Another approach that could be envisaged to form the bu-
tenolide moiety is the radical cyclization of a-bromoesters de-
rived from simple propargylic alcohols. Although the strategy
seems appealing, this type of radical cyclization involving elec-
trophilic radical derived from a-bromoesters is known to be
hampered by the slow equilibrium between the reactive,
minor, s-cis conformer and major s-trans conformer. As a result
the cyclization of this type of bromoesters in the presence of
a hydrogen-atom donor usually leads to dehalogenated, non-
cyclized products.[8] This explains why high dilution (ca.
10À2 m), slow addition techniques, and high temperature (typi-
cally refluxing benzene or toluene) are required in order for
[a] R. BØnØteau, Dr. C. F. Despiau, J.-C. Rouaud, A. Boussonni›re, Dr. V. Silvestre,
Prof. Dr. J. Lebreton, Dr. F. DØn›s
CEISAM UMR 6230 - UFR des Sciences et des Techniques
UniversitØ de Nantes, 2 rue de la Houssiniere BP 92208-44322
Nantes Cedex 3 (France)
E-mail: fabrice.denes@univ-nantes.fr
[b] A. Boussonni›re
Current Address: UniversitØ du Maine and CNRS UMR 6283
Institut des MolØcules et MatØriaux du Mans
FacultØ des Sciences et Techniques, avenue Olivier Messiaen
72085 Le Mans Cedex 9 (France)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201501294.
Chem. Eur. J. 2015, 21, 11378 – 11386
11378
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
the electrophilic radical intermediate to have a sufficient life-
time to undergo cyclization prior to intermolecular hydrogen-
atom abstraction from the chain-carrier reagent (typically a tin
hydride derivative).[9] Pattenden and co-workers reported ex-
amples of such radical cyclizations, which, combined with
a base-mediated migration of the exocyclic C=C bond into the
a,b-position, gave access to the g-butenolide subunit
(Scheme 1).[9a]
Scheme 2. Preparation of butenolides through a radical cyclization under
Pattenden’s condition.[9a]
We believed that the failure of the reaction with substrates
such as the one used in Scheme 2 [Eq. (2)] was not due to the
radical-cyclization process itself, as the size and the nature of
the side-chain were not expected to have a dramatic influence
on the rate of the cyclization step. The absence of formation
of the desired butenolides in some cases was more likely the
result of the outcome of the highly reactive alkenyl radical in-
termediate. We decided then to conduct a systematic study,
using a tin deuteride derivative as the chain-carrier reagent,
with the hope that it would allow us to determine the fate of
the alkenyl radical intermediate in these reactions. The results
are shown in Scheme 3.
Scheme 1. Pattenden’s approach to the butenolide unit.
Both the high-dilution (ca. 10À2 m) and the slow-addition
technique required to achieve this radical cyclization make this
approach unattractive, especially on large scale. Moreover,
under these reaction conditions the highly reactive alkenyl rad-
ical intermediates resulting from the 5-exo-dig cyclization pro-
cess are very likely to lead to side-reactions prior to hydrogen-
atom abstraction from the chain-carrier reagent, the concentra-
tion of which in the medium is voluntarily maintained suffi-
ciently low in order not to interfere with the cyclization pro-
cess. For all these reasons, we were intrigued by the report by
Pattenden and co-workers and decided to reinvestigate this re-
action in detail. This study allowed us to shed light on “invisi-
ble” hydrogen-atom transfers, which might result in complica-
tions after the cyclization process. We propose herein an alter-
native route to circumvent these chemoselectivity issues by
the means of a radical cyclization at low temperature of alumi-
nium acetal intermediates. The mildness of this novel approach
is illustrated by the total synthesis of two enantioenriched, nat-
urally occurring butenolides.
Results and Discussion
We started our investigations by repeating some of the exam-
ples described in the original paper by Pattenden and co-work-
ers. Under the reaction conditions reported in their seminal
contribution, we were able to achieve the radical cyclization of
some a-bromoesters, with limited reduction of the electrophil-
ic radical intermediate (ca. 10–15%), in good agreement with
the observations by these authors. Following in situ treatment
with a base, several tetrasubstituted butenolides were indeed
isolated in good yields and high purity by flash-chromatogra-
phy over silica gel/KF (9:1)[10] [Scheme 2, Eq. (1)]. However, as
we suspected, the success of this reaction proved to strongly
depend upon the nature of the bromoester used as precursor
and, with some substrates not even traces of butenolide were
observed [Scheme 2, Eq. (2)].
Scheme 3. Deuterium-labelling experiment.
The results obtained in the presence of tributyltin deuteride
clearly indicate that 1,5-hydrogen atom transfer (1,5-HAT)[11,12]
could take place under the reaction conditions described by
Pattenden and co-workers. In some cases, these translocation
processes did not interfere with the formation of the desired
products, as illustrated by the preparation of compounds 2a–
d, which were isolated in moderate to good yields, with or
without incorporation of the deuterium atom on the side-
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11379
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
chain. As expected, the translocation process is accelerated by
an increase in the stabilization of the carbon-centred radical
(Figure 2) resulting from the 1,5-hydrogen shift (see deuterium
incorporation in 2a, 2b, vs. 2c and 2d), and this effect coun-
ter-balances effectively the statistical effect (number of hydro-
gen atoms that can be abstracted from primary, secondary and
tertiary positions).[13]
It is worth noting that while only trace amounts of deuteri-
um atom were incorporated at the 5-position (isopropyl side-
chain) in 2a, the butenolide was isolated with only about 30%
of deuterium atom incorporation at the benzylic position. Simi-
larly, compound 2b was obtained with only 52% of deuterium
incorporation, exclusively at the 5-position (n-pentyl side-
chain). So far two possibilities can be envisaged to explain the
incorporation of hydrogen atom at the benzylic position:
either the alkenyl radical abstracts a hydrogen atom from a un-
known source of hydrogen atom (from the alkyl chains of the
tin hydride or from the products resulting from the thermal
decomposition of AIBN), or more likely the base used in this
study (DBU) is capable of abstracting a proton (or deuterium)
from the benzylic position. If the protons at the benzylic posi-
tion were exchangeable in the presence of a base, then it
would explain why less than 100% of deuterium atom incor-
poration was measured for all the substrates that did not un-
dergo quantitative 1,5-HATs. This was later confirmed by the
use of a milder base to promote the C=C bond migration (vide
infra). Butenolides 2e and 2 f could not be observed under the
aforementioned reaction conditions and the cyclization led in
these cases to complex mixtures. It is likely that the alkenyl
radical resulting from the radical cyclization of 1e underwent
either 1,5-HAT or 6-exo-trig cyclization, or both. In the case of
compound 1 f, the resulting alkenyl radical either evolved by
1,5-HAT (on both possible sites), 6-exo-trig cyclization, or more
likely by ring-opening of the cyclopropylalkenyl radical inter-
mediate. Unfortunately, the complex reaction mixture did not
allow us to isolate any of these compounds. In some cases,
such as the cyclization of compound 1g, the carbon-centred
radical underwent 4-exo-trig cyclization to give bicyclic com-
pounds such as 4 (Scheme 5).
Figure 2. Effect of the stabilization of the resulting carbon-centred radical on
the translocation process.
Interestingly, both sp-hybridized and sp2-hybridized alkenyl
radicals underwent 1,5-HAT with similar efficiency (Scheme 4),
as indicated by the relative percentage of deuterium incorpo-
ration in 2c and 2d. Contrary to what is commonly observed
in translocation processes leading to 1-hexenyl radical, for
which the carbon-centred radical is located in a suitable posi-
tion to undergo 5-exo-trig cyclization and then deliver 5-mem-
bered rings,[11c] the radicals formed in the translocation process
in this study are 1-pentenyl radicals and they do not cyclize
very rapidly neither in a 5-endo-trig mode,[14,15] nor in a 4-exo-
trig mode. Accordingly, these newly formed carbon-centred
radicals abstract instead the deuterium atom from the tributyl-
tin deuteride to give, after treatment with a base, the desired
butenolides. In this case, the translocation process remains “in-
visible” for the reactions carried out with nBu3SnH. In some
other cases, however, 4-exo-trig cyclization was observed (vide
infra).
Scheme 5. Formation of bicyclic compounds by a translocation process.
The formation of bicyclic compound 4 is the result of a radi-
cal cascade involving the cyclization of electrophilic radical 5
to give alkenyl radical 6 (Scheme 6). The latter undergoes 1,5-
HAT to generate carbon-centred radical 7, which either ab-
stracts a hydrogen atom from the tributyltin hydride or cyclizes
onto the activated carbon=carbon double bond in a 4-exo-trig
mode to give benzylic radical 8. Hydrogen atom abstraction
from the chain carrier reagent leads to bicyclic lactone 4. We
assume that the presence of substituents at vicinal positions in
1g allows for an increase in the rate of the 4-exo-trig cycliza-
Scheme 4. Translocation process with sp- and sp2-hybridized alkenyl radicals.
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11380
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 8. TEMPO/BAIB oxidation of methylene-g-lactols and DBU-mediated
migration of the C=C bond to give butenolides.
Not satisfied by these results, we investigated the possibility
to directly access the desired butenolides from a-bromoesters,
without isolation of the sensitive methylene-g-lactols. The se-
quence is based on the reactivity of the cyclic aluminium ace-
tals formed as products of the radical cyclization. These alumi-
nium species were found to react with nucleophiles, but they
could also transfer a hydride onto the carbonyl group of
a ketone or an aldehyde.[19] The Oppenauer-type oxidation of
the cyclic aluminium acetals proved to be very efficient for the
preparation of g-lactones from a-bromoesters. Only a two- to
threefold excess of a cheap, commercially available aldehyde
was required to perform the oxidation in situ and obtain g-lac-
tones in high yields, as previously illustrated by the synthesis
of (À)-trans-cognac lactone (Scheme 9).[20]
Scheme 6. Proposed mechanism for the formation of bicyclic lactone 4.
tion process. A similar case of rate enhancement thanks to a vi-
cinal effect has been reported for 1,5-HAT.[13a] This would ex-
plain why this type of rearrangement was not observed with
related precursors 1b–d.
Alkylidene butenolide 3 presumably results from butenolide
2g (not observed under these reaction conditions) by the elim-
ination of the methoxy group. This observation confirms that
the base used in this sequence (DBU) is basic enough to de-
protonate, at least partially, these g-butenolides (vide infra).
These results prompted us to investigate a new route to bu-
tenolides based upon our previous work on the radical cycliza-
tion of aluminium acetals. The latter are generated from the
corresponding a-bromoesters by reduction at low temperature
with diisobutylaluminium hydride (DIBAL-H) and they proved
to be suitable substrates for a radical cyclization at low tem-
perature. This reaction was found to be particularly efficient to
prepare g-lactols[16] and methylene-g-lactols[17] (Scheme 7).
Scheme 9. Sequence reduction/cyclisation/oxidation leading to g-lactones.
Here, this one-pot sequence was combined with the base-
mediated migration of the C=C bond and applied to the cycli-
zation of a-bromoesters 1a–i. Gratifyingly, butenolides 2a–
i presenting a tetrasubstituted C=C bond were obtained in
good to high yields in most cases. The results are reported in
Table 1.
Scheme 7. Preparation of methylene-g-lactols by the radical cyclization of a-
bromoaluminium acetals at low temperature.
We first searched for reaction conditions to oxidize the
methylene-g-lactols into the corresponding methylene-g-lac-
tones. Among the different oxidizing reagent tested, only
TEMPO/BAIB ([bis(acetoxy)iodo]benzene)[18] gave the desired
compounds without migration of the C=C bond (91–99%
yields, NMR monitoring). In our hands, these compounds could
not be purified by flash chromatography as the migration oc-
curred on silica gel. The crude reaction mixtures were then
treated with a base to give the desired butenolides 2i–k in
low to moderate yields over the two steps (29–52%)
(Scheme 8).
Bromoesters 1a–1d were efficiently converted into the cor-
responding aluminium acetals by treatment with DIBAL-H in
toluene at À708C. The conversion was monitored by thin-layer
chromatography (TLC) and, after completion nBu3SnH, Et3B
and air were added slowly. The reaction mixture was stirred at
low temperature until complete disappearance of the propar-
gylic alcohol (resulting from the decomposition of the alumini-
um acetals on TLC). The formation of methylene-g-lactols
could be observed by TLC. Addition of a threefold excess of
a cheap, commercially available aldehyde into the reaction
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11381
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
allowed us to prove the mildness of these reaction conditions,
as no 1,5-HAT took place in this case. More interestingly, bute-
nolides 2e and 2 f could be obtained in good yields (82 and
63% yields, respectively), while the cyclization at high temper-
ature gave only complex mixtures, with no traces of these two
compounds (vide supra; Table 1, entries 5 and 6). As previously
observed for the synthesis of methylene-g-lactols,[17] butenolide
2 f was obtained without ring-opening of the cyclopropyl
unit.[21] Surprisingly, the reduction of bromoester 1g required
an excess of DIBAL-H (3 equiv) and the one-pot sequence gave
butenolide 2g only in a modest 33% yield (Table 1, entry 7),
together with the propargylic alcohol resulting from the ther-
mal decomposition of the aluminium acetal intermediate (ca.
40% yield). Under these reaction conditions, and contrary to
what was observed at high temperature under Pattenden’s
conditions (vide supra), no rearrangements were observed. As
the presence of an alkoxy substituent in the bromoester pre-
cursors did not seem to interfere neither in the reduction step,
nor in the cyclization process in previous studies,[16,17,20] we
wondered whether the problems encountered here were
simply due to steric hindrance. We decided to test this hypoth-
esis with compound 1h, which possess a structure similar to
1a but with bulkier substituents. The one-pot sequence, how-
ever, proceeded without any problems and butenolide 2h was
obtained in high yield (Table 1, entry 8). It seems then that it is
the combination of an alkoxy group and a secondary carbon
atom directly linked to the propargylic position that makes the
reduction more difficult and decreases the stability of the alu-
minium acetal intermediate. Finally, butenolide 2i was ob-
tained in good yield from bromoester 1i (Table 1, entry 9).
Encouraged by these results we decided to study the prepa-
ration of optically enriched compounds in order to make sure
that our reaction conditions would allow for the preparation of
optically active compounds. Indeed, it is well-known that g-
substituted butenolides can undergo epimerization under
basic conditions,[22] in particular in the presence of relatively
strong bases such as DBU.[23] We decided to investigate the
possibility to achieve the C=C bond migration without loss of
optical activity and we first turned our attention to the synthe-
sis of enantioenriched butenolides from natural sources. Bute-
nolides 2l was isolated from Plagiomnium undulatum in
2001[24] and its absolute configuration was determined by
Brückner, Kçnig and co-workers who proposed the first total
synthesis of the natural compound.[25] The original synthesis re-
quired seven steps from heptanal and we believed our meth-
odology might offer an alternative and shorter route to this
natural compound. The precursor for the radical cyclization, a-
bromoester 1l, was obtained in high yield from commercially
available (+)-oct-1-yn-3-ol (99.0% ee) and (Æ)-2-bromopropion-
ic acid (Scheme 10). Gratifyingly, butenolide 2l was obtained in
good yields and with high levels of enantiomeric excesses
from a-bromoester 1l under our standard reaction conditions,
using a variety of bases to achieve the migration of the exocy-
clic C=C bond. Among the different bases tested to perform
the migration of the C=C bond into the a,b-position, iPr2NEt
gave the best results, both in terms of yield and enantiomeric
excess. The use of aqueous HCl to promote the migration of
Table 1. Preparation of g-butenolides with a tetrasubstituted C=C bond
by radical cyclization of aluminium acetals, followed by oxidation and C=
C bond migration.
Entry
Substrate
Butenolides[a]
Yield [%][b]
1
75
2
3
4
5
88
77
73
82
6
63
7
8
9
33[c,d]
88
74
[a] All reactions were carried out in toluene on 2 mmol scale using 1.2–
1.5 equiv DIBAL-H at À708C (unless otherwise stated), 1.2–1.5 equiv
nBu3SnH and 0.3–0.5 equiv Et3B (1m in hexanes); oxidation was carried
out with 3.0 equiv of iPrCHO; C=C migration was achieved with 2.0 equiv
DBU. [b] After purification by chromatography on silica gel (silice/KF 9:1).
[c] 3 equiv of DIBAL-H was used. [d] 40% of propargylic alcohol recov-
ered.
mixture (e.g., PhCHO or iPrCHO) at low temperature, followed
by warming to room temperature, resulted in the non-reversi-
ble hydride transfer from the cyclic aluminium acetal to the al-
dehyde. Once the oxidation of the aluminium acetal into the
corresponding methylene-g-lactone was complete (TLC or GC
monitoring), a base (typically DBU) was added to achieve the
migration of the exocyclic C=C bond to the a,b-position.
This one-pot, four-step sequence led to tetrasubstituted bu-
tenolides 2a–2d in good to high yields (73–88%, Table 1, en-
tries 1–4). This approach compares favourably with the direct
cyclization at high temperature described in Scheme 3. More-
over, the use of nBu3SnD in the one-pot, four-step sequence,
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11382
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
argylic alcohol 11 in 92% yield and with good level of enantio-
meric excess (96.0% ee).[30] The coupling between propargylic
alcohol 11 and racemic bromoacid 12 to give a-bromoester 13
was achieved in high yield (95%) using classical conditions
(DCC, DMAP cat.). Bromoester 13 then engaged in the key
step, which consisted first in the reduction with DIBAL-H to
give the corresponding aluminium acetal. This thermally labile
intermediate underwent radical cyclization at low temperature
(below À708C) in the presence of nBu3SnH, using Et3B/air as
an initiator. The resulting cyclic aluminium acetal was then
oxidized in the presence of iPrCHO into the corresponding
arylidene-g-lactone. The mixture of stereoisomers was then
converted into maculalactone A by addition of iPr2NEt
(Scheme 11). Under these reaction conditions maculalactone A
could be obtained with a high level of enantiomeric excess
(94.8% ee), close to the one determined for the natural com-
pound. In this case, the migration of the C=C bond into the
a,b-position proved to be extremely slow and it required ex-
tended reaction time to reach good yields. The use of DBU
also led to maculalactone A, in much shorter reaction time
than with iPr2NEt, but with complete racemization, in sharp
contrast with what was previously observed for the prepara-
tion of the tetrasubstituted butenolide from P. undulatum (vide
supra). The use of aqueous HCl failed to promote the C=C
bond migration with this substrate. Although we decided not
to explore this route, an even more straightforward approach
to propargylic alcohol 11 would require enantioselective addi-
tion of alkynylmetals onto phenylacetaldehyde. This would
shorten the enantioselective synthesis of maculalactone A
down to only three steps.
Scheme 10. Expeditive synthesis of tetrasubstituted butenolide from P. undu-
latum.
the C=C bond also led to 2l in excellent enantiomeric excess,
albeit in slightly lower yield. This represents the shortest syn-
thesis to date of this natural compound.
The high chemoselectivity of our approach was further dem-
onstrated by the synthesis of another representative of the bu-
tenolide family. Maculalactone A (Figure 3) is the most abun-
dant secondary metabolite produced by the marine cyanobac-
terium Kyrtuhrix maculans. This compound was isolated in
1996 by Brown and co-workers and it presents an unusual tri-
benzyl butenolide skeleton,[26] which is common to the other
maculalactones B and M (Figure 3).[27]
This route was next compared with the direct radical cycliza-
tion of bromoester 13, under reaction conditions used by Pat-
tenden and co-workers. A solution of bromoester (Æ)-13 in
benzene (c=10À2 m) was heated at reflux and the solutions of
Figure 3. Maculalactone A and some other members of the family.
Interestingly, the authors noticed that the natural compound
was biosynthesized in a partially racemized form (90–95% ee
only).[4,28,29] Maculalactone A and its derivatives present promis-
ing activity as antifouling agents,[4,28] a property that is used by
the long-lived, extremely slow-growing cyanobacterium
K. maculans to protect itself indirectly from herbivores. Our
synthesis started with commercially available acyl chloride 9.
The latter was converted into the corresponding Weinreb
amide in 89% yield (Scheme 11). Addition of the Grignard re-
agent of phenylacetylene onto the Weinreb amide gave enone
10 in 72% yield. Several other routes have been explored to
prepare directly enone 10 from acyl chloride 9, including the
use of 1-TMS-phenylacetylene in the presence of AlCl3 or the
use of copper acetylide, but the best result in terms of yield
and purity was obtained in the two-step sequence via the
Weinreb amide. The configuration of the chiral centre of prop-
argylic alcohol 11 was set using asymmetric reduction of
ynone 10 using Noyori’s ruthenium-based catalyst (see the
Supporting Information for details). This reduction led to prop-
Scheme 11. Total synthesis of maculalactone A using a radical cyclization at
low temperature.
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11383
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
nBu3SnH and AIBN in benzene were then added slowly over
16 h. Interestingly, under these reaction conditions no traces of
maculalactone A could be detected either by TLC or by
1H NMR spectroscopic analysis of the crude reaction mixture.
The main products that could be characterized from the com-
plex reaction mixture were lactone 14a and ester 15, isolated
in 28 and 20% yields, respectively (Scheme 12). The presence
of 14a indicates that the alkenyl radical intermediate was not
trapped rapidly enough by the tin hydride under these reac-
tion conditions. As a result side reactions such as the intramo-
lecular addition onto the aromatic rings present on the side-
chains took place. Although we were not able to assess with
certitude the structures of the other by-products, it seems
plausible that cyclization did not take place exclusively at only
one of the two aromatic rings located at the side-chains. Ac-
cordingly, although not isolated, lactone 14b might also be
present in the reaction mixture, as well as other products re-
sulting from intermolecular addition onto the solvent (ben-
zene).[31]
Scheme 13. 1,5-HATs resulting in epimerization during radical cyclization at
high temperature.
reaction conditions, the cyclization of aluminium acetals at low
temperature led to syn-2m and anti-2m in good yields from
syn-1l and anti-1m, respectively, with only minor loss (if any)
of the stereochemical information [Scheme 13, Eqs. (2) and
(3)].
Finally, we turned our attention to precursor 1n, easily pre-
pared in an optically pure manner from l-isoleucine
(Scheme 14, see the Supporting Information for details). The
cyclization of 1n was carried out at low temperature, via its
aluminium acetal, under our optimized reaction conditions to
give 2n in good yield and as a single diastereoisomer
[Scheme 14, Eq. (1)]. When nBu3SnH was replaced with
nBu3SnD, the cyclization at low temperature, followed by oxi-
dation on the presence of isobutyraldehyde and iPr2NEt-pro-
moted C=C bond migration, butenolide [D]2n was obtained
with deuterium incorporation exclusively at the benzylic posi-
tion [Scheme 14, Eq. (2)]. The direct radical cyclization of 1n
under the reaction conditions described by Pattenden and co-
workers led to [D]2n’ as a nearly 1:1 mixture of diastereoiso-
mers. As clearly indicated by deuterium-labelling experiments,
epimerization occurred almost exclusively at a remote position
on the side-chain and not at the butenolide moiety
[Scheme 14, Eq. (3)]. The high level of deuterium incorporation
on the side-chain proves that nearly quantitative 1,5-HAT oc-
curred after 5-exo-dig cyclization. This radical cascade led to
a complete epimerization of a stereogenic centre located at
a position that was not suspected at first glance to be sensi-
tive. The relative configuration in 1n (syn or anti) has no influ-
ence on the 1,5-HAT process, as shown by the racemization
observed during the cyclization of a mixture of syn- and anti-
1n at high temperature, which led to a racemic mixture of bu-
tenolides syn- and anti-2n [see chiral HPLC analysis in the Sup-
porting Information; Scheme 14, Eq. (2)].
Scheme 12. Attempt at preparing maculalactone A under “classical” reaction
conditions (Pattenden’s conditions).
The chemoselectivity of our methodology compared to the
classical approach at high temperature was illustrated further
with the attempted cyclization of compounds syn-1m and
anti-1m (Scheme 13). The reactions carried out at high temper-
ature from syn-1m and anti-1m were expected to result in the
epimerization of the target butenolides due to undesired 1,5-
HAT.[32] The translocation process was demonstrated by using
tributyltin deuteride. In the case of syn-1m and anti-1m, how-
ever, the expected butenolides [D]2m were only observed as
minor components of a complex crude reaction mixture. Nev-
ertheless the disappearance of the characteristic signal for the
proton CH(OMe) indicates that deuterium has been incorporat-
ed at this position [Scheme 13, Eqs. (1) and (2)]. On the contra-
ry, the radical cyclization of aluminium acetals at low tempera-
ture proved to be extremely chemoselective and allowed for
a mild cyclization reaction to take place. Under the optimized
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11384
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Hemez (CEISAM, HRMS measurement) for their help in the
analysis of the compounds prepared in this study. The authors
thank “LABERCA, unitØ de recherche d’Oniris (Ecole Nationale
VØtØrinaire, Agroalimentaire et de l’Alimentation de Nantes At-
lantique)”, the “Plateforme de SpectromØtrie de Masse de l’IRS-
UN (Nantes)”, and “Plateforme Biopolym›res, IntØractions, Bio-
logie Structurale (BIBS) de l’INRA (Nantes)” for the access to
HRMS instruments. The authors thank the “Minist›re de la Re-
cherche et de l’Enseignement SupØrieur” (MESR grant for A.B.)
and the “Agence Nationale de la Recherche” (grant ANR-10-
JCJC-0712 “Radic[Al]” for R.B., and ANR Blanc 13 BS07-0010-01
“UZFul-Chem” for C.D. and J.-C.R.) for funding.
Keywords: butenolides · deuterium-labeling experiments ·
hydrogen transfer · maculalactone A · radical reactions
[1] For selected examples of post-functionalization of g-butenolides, see:
a) A. Kondoh, A. Arlt, B. Gabor, A. Fürstner, Chem. Eur. J. 2013, 19, 7731–
7738; b) R. M. OrtuÇo, R. Merce, J. Font, Tetrahedron 1987, 43, 4497–
4506; c) A. A. Jakubowski, F. S. Guziec Jr., M. Sugiura, C. Chan Tam, M.
Tishler, J. Org. Chem. 1982, 47, 1221–1228; d) A. A. Jakubowski, F. S. Gu-
ziec Jr., M. Tishler, Tetrahedron Lett. 1977, 18, 2399–2402; e) G. Casira-
ghi, L. Colombo, G. Rassu, P. Spanu, J. Org. Chem. 1991, 56, 6523–6527.
[2] For selected reviews, see: a) D. W. Knight, Contemp. Org. Synth. 1994, 1,
287–315; b) X. Jusseau, L. Chabaud, C. Guillou, Tetrahedron 2014, 70,
2595–2615; c) Q. Zhang, X. Liu, X. Feng, Current Org. Synth. 2013, 10,
764–785; d) T. M. Ugurchieva, V. V. Veselovsky, Russian Chem. Rev. 2009,
78, 337–373; e) S. Ma, Acc. Chem. Res. 2009, 42, 1679–1688; f) S. Ma,
Acc. Chem. Res. 2003, 36, 701–712.
[3] For selected recent examples of the use of metathesis in the synthesis
of butenolides, see: a) C. Bürki, J. Bonjoch, B. Bradshaw, G. Villa, P.
Renaud, Chem. Eur. J. 2015, 21, 395–401; b) J.-c. Han, F. Li, C.-c. Li, J.
Am. Chem. Soc. 2014, 136, 13610–13613; c) R. A. Fernandes, P. H. Patil,
A. K. Chowdhury, Asian J. Org. Chem. 2014, 3, 58–62; d) K.-i. Takao, R.
Nanamiya, Y. Fukushima, A. Namba, K. Yoshida, K.-i. Tadana, Org. Lett.
2013, 15, 5582–5585; e) M. Billamboz, J.-C. Legeay, F. Hapiot, E. Monfli-
er, C. Len, Synthesis 2012, 44, 137–143; f) Y. Wang, W.-M. Dai, Tetrahe-
dron 2010, 66, 187–196; g) M. Bassetti, A. D’Annibale, Current Org.
Chem. 2013, 17, 2654–2677.
Scheme 14. Epimerization of remote position by 1,5-HATs during radical cyc-
lization at high temperature.
Conclusions
We have reported herein a general access to the sterically de-
manding tetrasubstituted g-butenolides from easily accessible
a-bromoesters as precursors. The target butenolides were ob-
tained in good to high yields using a one-pot sequence involv-
ing the formation of a thermally labile aluminium acetal inter-
mediate, followed by a radical cyclization and oxidation of the
resulting cyclic aluminium acetals by the use of a cheap, com-
mercially available aldehyde. The migration of the C=C bond
was then achieved under basic or acidic conditions. This ap-
proach allows for the preparation of optically active buteno-
lides as illustrated by the efficient synthesis of two natural
compounds. The mechanistic study also shed light on “invisi-
ble” 1,5-hydrogen-atom transfers that occur during the radical
cyclization under previously reported reaction conditions for
the direct cyclization of a-bromoesters (high temperature and
high dilution). As illustrated with several examples of the pres-
ent study, these 1,5-HATs are no longer “invisible” when epime-
rization of stereocentres occurs, or when rearrangements lead-
ing to by-products are possible.
[4] S. L. Bader, M. U. Luescher, K. Gademann, Org. Biomol. Chem. 2015, 13,
199–206.
[5] a) E. Yoneda, T. Kaneko, S.-W. Zhang, K. Onitsuka, S. Takahashi, Org. Lett.
2000, 2, 441–443; b) M. Tsubuki, K. Takahashi, T. Honda, J. Org. Chem.
2009, 74, 1422–1425; c) R. W. Bates, S. Sridhar, J. Org. Chem. 2008, 73,
8104–8105; d) R. W. Bates, S. Sridhar, J. Org. Chem. 2011, 76, 5026–
5035; e) Y. Wang, L. Zhu, Y. Zhang, R. Hong, Angew. Chem. Int. Ed. 2011,
50, 2787–2790; Angew. Chem. 2011, 123, 2839–2842.
[6] S. Li, B. Miao, W. Yuan, S. Ma, Org. Lett. 2013, 15, 977–979.
[7] J. Li, Y. Liu, C. Li, X. Jia, Chem. Eur. J. 2011, 17, 7409–7413.
[8] A. L. J. Beckwith, S. A. Glover, Aust. J. Chem. 1987, 40, 157–173.
[9] a) J. M. Clough, G. Pattenden, P. G. Wight, Tetrahedron Lett. 1989, 30,
7469–7472; b) J. L. Belletire, N. O. Mahmoodi, Tetrahedron Lett. 1989,
30, 4363–4366; c) S. Hanessian, R. Di Fabio, J.-F. Marcoux, M. Prud’
homme, J. Org. Chem. 1990, 55, 3436–3438.
[10] D. C. Harrowven, I. L. Guy, Chem. Commun. 2004, 1968–1969.
[11] For selected reviews on hydrogen-atom transfers, see: a) L. Feray, N.
Kuznetsov, P. Renaud, Radicals in Organic Synthesis Vol. 2 (Eds.: P.
Renaud, M. P. Sibi), Wiley-VCH, Weinheim 2001, pp. 246–278; b) J. Rob-
ertson, J. Pillai, R. K. Lush, Chem. Soc. Rev. 2001, 30, 94–103; c) F. DØn›s,
F. Beaufils, P. Renaud, Synlett 2008, 2389–2399; d) M. Nechab, S.
Mondal, M. P. Bertrand, Chem. Eur. J. 2014, 20, 16034–16059.
[12] For selected examples of related 1,5-HAT transfers involving diradical
species, see: a) P. J. Wagner, Acc. Chem. Res. 1989, 22, 83–91; b) C. F. Bill-
era, R. D. Little, J. Am. Chem. Soc. 1994, 116, 5487–5488; c) G. L. Carroll,
A. K. Allan, M. K. Schwaebe, R. D. Little, Org. Lett. 2000, 2, 2531–2534.
For examples of 5-exo-dig radical cyclization followed by 1,5–HAT lead-
Experimental Section
See the Supporting Information for details.
Acknowledgements
F.D. thanks Prof. A. Alexakis and Dr. C. Meyer for very helpful
discussions. The authors thank I. Louvet (CEISAM, determina-
tion of the enantiomeric excesses by chiral HPLC) and J.
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11385
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
ing to silicon-centered radical, see: d) D. L. J. Clive, M. Cantin, J. Chem.
Soc., Chem. Commun. 1995, 319–320; e) M. Sannigrahi, D. L. Mayhew,
D. L. J. Clive, J. Org. Chem. 1999, 64, 2776–2788; f) D. L. J. Clive, W.
Yang, A. C. MacDonald, Z. Wang, M. Cantin, J. Org. Chem. 2001, 66,
1966–1983; g) D. L. J. Clive, E.-S. Ardelean, J. Org. Chem. 2001, 66,
4841–4844; h) A. Martinez-Grau, D. P. Curran, J. Org. Chem. 1995, 60,
8332–8333; i) A. Martinez-Grau, D. P. Curran, Tetrahedron 1997, 53,
5679–5698. For related examples of 5-exo-dig radical cyclization fol-
lowed by 1,n-HAT (n=5, 6) and subsequent cyclization, see: j) S. Bogen;
M. Journet, M. Malacria, Synlett 1994, 958–960; M. Journet, M. Malacria,
Synlett 1994, 958–960; k) S. Bogen, M. Malacria, J. Am. Chem. Soc. 1996,
118, 3992–3993; l) S. Bogen, L. Fensterbank, M. Malacria, J. Org. Chem.
1999, 64, 819–825;m) M. P. Bertrand, D. Crich, R. Nouguier, R. Samy, D.
Stien, J. Org. Chem. 1996, 61, 3588–3589; n) D. Stien, D. Crich, M. P. Ber-
trand, Tetrahedron 1998, 54, 10779–10788.
121–125; c) E. Mainetti, L. Fensterbank, M. Malacria, Synlett 2002, 923–
926; d) K. K. Milnes, S. E. Gottschling, K. M. Baines, Org. Biomol. Chem.
2004, 2, 3530–3534; e) S. E. Gottschling, T. N. Grant, K. K. Milnes, M. C.
Jennings, K. M. Baines, J. Org. Chem. 2005, 70, 2686–2695.
[22] a) P. Duret, B. Figad›re, R. Hocquemiller, A. CavØ, Tetrahedron Lett. 1997,
38, 8849–8852; b) X. Franck, B. Figad›re, Tetrahedron Lett. 2002, 43,
1449–1451. c) For a useful method for the determination of enantio-
meric purity of butenolides by NMR spectroscopy, see: S. Latypov, X.
Franck, J.-C. Jullian, R. Hocquemiller, B. Figad›re, Chem. Eur. J. 2002, 8,
5662–5666.
[23] a) K. H. Kim, H. S. Lee, S. H. Kim, K. Y. Lee, J.-E. Lee, J. N. Kim, Bull. Korean
Chem. Soc. 2009, 30, 1012–1020; b) D. R. Williams, A. L. Nold, R. J. Mul-
lins, J. Org. Chem. 2004, 69, 5374–5382.
[24] Y. Saritas, M. Mekem Sonwa, H. Iznaguen, W. A. Kçnig, H. Muhle, R.
Mues, Phytochemistry 2001, 57, 443–457.
[13] For related examples of 1,5-HAT at primary positions, see: a) F. Beaufils,
F. DØn›s, P. Renaud, Org. Lett. 2004, 6, 2563–2566; b) S. Bogen, M.
Gulea, L. Fensterbank, M. Malacria, J. Org. Chem. 1999, 64, 4920–4925.
[14] For some examples of radical cascades involving 1,5-HAT followed by
5-endo-trig cyclization, see: a) C. Chatgilialoglu, T. Gimisis, G. P. Spada,
Chem. Eur. J. 1999, 5, 2866–2876; b) A. Kittaka, T. Asakura, T. Kuze, H.
Tanaka, N. Yamada, K. T. Nakamura, T. Miyasaka, J. Org. Chem. 1999, 64,
7081–7093. c) See also Ref. [12d–g] and [13b].
[25] T. Kapferer, R. Brückner, A. Herzig, W. A. Kçnig, Chem. Eur. J. 2005, 11,
2154–2162.
[26] W.-Y. Tsui, G. A. Williams, G. D. Brown, Phytochemistry 1996, 43, 1083–
1085.
[27] S.-C. Lee, G. D. Brown, J. Nat. Prod. 1998, 61, 29–33.
[28] For a review, see: G. D. Brown, H.-F. Wong, N. Hutchinson, S.-C. Lee,
B. K. K. Chan, G. A. Williams, Phytochem. Rev. 2004, 3, 381–400.
[29] For previous syntheses, see: a) G. D. Brown, H.-F. Wong, Tetrahedron
2004, 60, 5439–5451; b) A. Kar, S. Gogoi, N. P. Argade, Tetrahedron
2005, 61, 5297–5302; c) R. J. Duffy, K. A. Morris, R. Vallakati, W. Zhang,
D. Romo, J. Org. Chem. 2009, 74, 4772–4781.
[15] For a discussion on the 5-endo-trig cyclization processes, see: C. Chatgi-
lialoglu, C. Ferreri, M. Guerra, V. Timokhin, G. Froudakis, T. Gimisis, J. Am.
Chem. Soc. 2002, 124, 10765–10772.
[16] A. Boussonni›re, F. DØn›s, J. Lebreton, Angew. Chem. Int. Ed. 2009, 48,
9549–9552; Angew. Chem. 2009, 121, 9713–9716.
[30] K. Matsumura, S. Hashiguchi, T. Ikariya, R. Noyori, J. Am. Chem. Soc.
1997, 119, 8738–8739.
[17] A. Boussonni›re, R. BØnØteau, N. Zimmermann, J. Lebreton, F. DØn›s,
Chem. Eur. J. 2011, 17, 5613–5627.
[18] A. De Mico, R. Margarita, L. Parlanti, A. Vescovi, G. Piancatelli, J. Org.
Chem. 1997, 62, 6974–6977.
[19] A. Boussonni›re, R. BØnØteau, J. Lebreton, F. DØn›s, Eur. J. Org. Chem.
2013, 7853–7866.
[31] For a review on the importance of solvent in free radical chemistry, see:
G. Litwinienko, A. L. J. Beckwith, K. U. Ingold, Chem. Soc. Rev. 2011, 40,
2157–2163.
[32] For a recent report suggesting epimerization “that could potentially
occur” by 1,5-HAT, see: K. J. Hale, M. Maczka, A. Kaur, S. Manaviazar, M.
Ostovar, M. Grabskı, Org. Lett. 2014, 16, 1168–1171.
˘
[20] R. BØnØteau, J. Lebreton, F. DØn›s, Chem. Asian J. 2012, 7, 1516–1520.
[21] For previous studies highlighting the difference in the rate of fragmen-
tations between cyclopropylvinyl- and cyclopropylmethyl radicals, see:
a) J. K. Crandall, G. L. Tindell, A. Manmade, Tetrahedron Lett. 1982, 23,
3769–3772; b) T. G. Back, K. R. Muralidharan, J. Org. Chem. 1989, 54,
Received: April 1, 2015
Published online on June 26, 2015
Chem. Eur. J. 2015, 21, 11378 – 11386
www.chemeurj.org
11386
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim