The effect of Lewis acids on the cycloaddition of 3,3,6-Trimethylcyclohex- 5-ene-1,2,4-trione: Hydrogen transfer versus cycloaddition with cyclopentadiene

Article abstract of DOI:10.1002/ejoc.201300706

Exposure of 3,3,6-trimethylcyclohex-5-ene-1,2,4-trione to catalytic amounts of Lewis acids revealed two disparate reactions in the presence of cyclopentadiene. The expected cycloaddition was found to be reversible for the title compound, and transfer hydrogenation was the preferred pathway over long periods of time. Other tested substrates were able to undergo facile cycloaddition with considerable yields and without the parallel reduction. The work presented here defines the effect of Lewis acid catalysts on the reaction between cyclopentadiene and 3,3,6-trimethylcyclohex-5-ene-1,2,4-trione. Cycloaddition and hydrogenation reactions both occur in parallel. For cycloaddition, the dr was similar for all Lewis acids tested. Hydrogenation was catalyzed efficiently by AlCl₃ and found to be the main mode of reaction for this particular system. Copyright

Full text of DOI:10.1002/ejoc.201300706

SHORT COMMUNICATION
DOI: 10.1002/ejoc.201300706
The Effect of Lewis Acids on the Cycloaddition of 3,3,6-Trimethylcyclohex-5-
ene-1,2,4-trione: Hydrogen Transfer versus Cycloaddition with Cyclopentadiene
Nicholas A. Eddy,[a][‡] Jay J. Richardson,[a][‡‡] and Gabriel Fenteany*[a]
Keywords: Lewis acids / Lewis acid catalysis / Hydrogenation / Cycloaddition / Ketones
Exposure of 3,3,6-trimethylcyclohex-5-ene-1,2,4-trione to
pound, and transfer hydrogenation was the preferred path-
catalytic amounts of Lewis acids revealed two disparate reac- way over long periods of time. Other tested substrates were
tions in the presence of cyclopentadiene. The expected
cycloaddition was found to be reversible for the title com-
able to undergo facile cycloaddition with considerable yields
and without the parallel reduction.
Introduction
Cyclohex-5-ene-1,2,4-triones present an understudied
class of molecules, which offer an interesting scaffold for
diversity-oriented synthesis. Substituents on the cyclic ring,
compatible with the use of Bobbitt’s salt,[1] can lead to fun-
damental building blocks for further transformation. Theo-
retically, the alkene, ketone, and diketone moieties could be
targeted through selective reactions to give a wide array of
compounds, which would lead to further elaboration to-
ward the synthesis of natural products or other biologically
relevant targets. In a recent publication, we detailed a one-
step reaction to generate these substrates and further
showed their use in the Diels–Alder cycloaddition with cy-
clopentadiene to give adducts in good yields.[2] Addition of
a methyl or phenyl substituent on the alkene moiety led to
markedly diminished reactivity, which was overcome
through the use of Lewis acid catalysis. The action of Lewis
acids on cyclohex-5-ene-1,2,4-triones has been unexplored.
Herein, we present the twofold reactivity of Lewis acids on
3,3,6-trimethylcyclohex-5-ene-1,2,4-trione (1), and this
leads to both cycloadduct 2 and reduction to dione 3
(Scheme 1).
Scheme 1. Reaction between 1 and cyclopentadiene catalyzed by
Lewis acids.
ference between the diene HOMO and dienophile LUMO
energies through complexation to Lewis basic sites in the
dienophile.[8] Additionally, a higher selectivity for endo and
exo diastereomers can also be achieved, especially at lower
temperatures.[9] Ligating species also afford a secondary
mode of tuning the reactivity profile.[10] Predominantly, this
lies in the realm of chiral modifiers, which form a complex
with the metal catalyst to provide an environment about the
dienophile that allows for enrichment in enantiomers.[11]
Results and Discussion
En route to optimizing the diastereoselectivity of the re-
action to 2, we discovered that 1 was thermally unstable
and decomposed to a complex mixture of products. This
compelled the use of TiCl4, which has been used previously
as a catalyst, to diminish the cycloaddition energy barrier
in preference to the side products. At room temperature, a
good yield (63%) of 2 was found with a diastereomeric ratio
(dr) of 2:1 (endo/exo). Yields diminished at lower tempera-
tures, but the endo/exo ratio remained largely unchanged.
Other Lewis acids were screened to determine the effect of
the catalyst on the cycloaddition, and the results are pre-
sented in Table 1. All of the screened catalysts gave a near-
equimolar distribution of endo and exo adducts, and the
conversions ranged from 5 to 30%. The best conversions
were found with TiCl4, but AlCl3 provided adequate con-
version to use for optimization later. The highest diastereo-
Lewis acid catalysis is a well regarded tool in organic
chemistry. Although these species affect many reactions
(e.g., Friedel–Crafts,[3] Nazarov,[4] Meerwein–Pondorf–Ver-
ley, and Oppenauer reactions[5]), the Diels–Alder cycload-
dition[6] is the most widely recognized as benefitting from
the addition of Lewis acids.[7] In a typical [π4s+π2s] cycload-
dition, the catalyst has been shown to lessen the energy dif-
[a] Department of Chemistry, University of Connecticut,
Storrs, CT 06269, USA
E-mail: gabriel.fenteany@uconn.edu
Homepage: www.biochemweb.org/fenteany
[‡] Current address: Department of Chemistry, Purdue University,
West Lafayette, IN 47909, USA
[‡‡] Current Address: Zeeco Corp.,
22151 East 91st Street, Broken Arrow, OK 74014, USA
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201300706.
Eur. J. Org. Chem. 2013, 5041–5044
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5041

N. A. Eddy, J. J. Richardson, G. Fenteany
SHORT COMMUNICATION
selectivity was found with (NH4)2ZnCl4 (Table 1, entry 7), InCl3 (Table 2, entry 5), CeCl3 (Table 2, entry 6), RuCl3
although the diminished conversions precluded its viability (Table 2, entry 7), and ZnCl2 (Table 2, entry 8). It was fur-
as a catalyst.
ther found that these specific catalysts gave predominately
3. Increased yields were found for the remaining Lewis ac-
ids. The latter cases seemed to give roughly identical
Table 1. Screen of Lewis acids for conversion and diastereoselec-
tivity.[a]
amounts of 2 and 3, except for the cases of Li+ and Mg2+
,
Entry Lewis acid
T [°C]
22
0
–78
22
0
–78
22
Conv.[b] [%] dr (endo/exo)[c]
which gave almost complete conversion to 3 in fair
amounts. No discernible pattern is apparent from these re-
sults. Our goal of increasing the selectivity of the cycload-
dition was realized through the use of an oxaborazolidinone
(Table 2, entry 4; Ts = para-toluenesulfonyl), which gave 2
and 3 in a 2:1 ratio and the cycloadduct with good selectiv-
ity (1:3dr, 4:1er).[12]
The low yields and high proportion of the exo dia-
stereomer found in 2 led us to believe the reaction was re-
versible under all of the catalysts studied. Titanium rapidly
produced 2 in 1.5 h at room temperature. Longer exposure
to TiCl4 (16 h) showed that 3 formed as the major product.
This confirmed that the reaction was reversible over longer
time periods and that 3 was the preferred product. A pro-
nounced concentration of exo found in all of the species
could be attributed to the steric interference between the
methylene and the C-6 methyl group, as presented by the
Alder–Stein principle (Figure 1).[13]
1
2
3
4
5
6
7
8
9
10
TiCl4
TiCl4
TiCl4
AlCl3
AlCl3
AlCl3
(NH4)2ZnCl4
Cu(acac)2
BF3·OEt2
BF3·OEt2
100 (63[d]
)
2:1
3:2
3:2
79 (25[d]
)
72 (21[d]
)
30
30
10
20
5
1:1.6
1:1
1:1
4.4:1
1:1
22
22
–78
[e]
[e]
–
–
25
2:1
[a] Reactions were performed with ene-triketone (0.05 mmol) and
cyclopentadiene (3 equiv.) at the specified temperature for 1.5 h.
acac = acetylacetonate. [b] Analyzed by HPLC with a MeOH/H2O
gradient (30 to 100% over 15 min, total run time = 30 min). [c] De-
1
termined by H NMR spectroscopy and HPLC. [d] Isolated yield
[%] from a 1 mmol-scale reaction. [e] Polymerized.
The reaction with aluminum chloride was performed at
a higher scale (1 mmol), and upon isolation and characteri-
zation, 3 was found as the major product (22%), along with
59% recovery of the starting material. This paradoxical re-
action piqued our interest, as the reversible reaction was
largely unknown. We felt it necessary to assess the effect of
the catalyst on both pathways. The results are presented in
Table 2.
Table 2. Lewis acid screen for the reaction between 1 and cyclopen-
tadiene.[a]
Figure 1. Transition states for endo and exo cycloadducts.
Entry
Lewis acid
% Yield[b]
2/3[c]
Understanding that the cycloaddition was reversible and
led to the formation of the exo adduct, we turned our atten-
tion to the production of 3. A source of hydrogen must be
present in the system for 3 to be produced, which means
that the solvent, catalyst, or cyclopentadiene must act in
such a fashion to liberate hydrogen for the reduction, none
of which were precedented to undergo this sort of reaction.
It was found that under the same conditions as those for
the cycloaddition (20 mol-% catalyst and 0.5 m solution), 3
was not produced over 24 h in the absence of cyclopentadi-
ene. In the presence of cyclopentadiene, the reaction oc-
curred rapidly. Isolated yields of 3 were identical to the
yields obtained in Table 2. This strongly suggests that cyclo-
pentadiene was acting as the source of hydrogen for this
system. This is the first case of cyclopentadiene acting in
such a manner.
1
2
3
4
TiCl4
AlCl3
AlCl3
63
3:1
22
1:99
1:99
2:1
63[d]
54 (1:3 dr, 4:1 er)
5
6
7
8
InCl3
CeCl3·7H2O
RuCl3
ZnCl2
NiCl2·6H2O
PdCl2
ZrCl4
camphorsulfonic acid
LiBr
MgCl2
TMSCl
20 (1:2.5 dr)
8 (1:2 dr)
10 (1:2 dr)
4 (1:3 dr)
34 (1:2 dr)
40 (1:2 dr)
67 (1:3 dr)
quant. (1:1 dr)
44 (1:2 dr)
51 (1:2.7 dr)
81 (1:3 dr)
1:9
1:4
1:9
1:6
1:3
1:4
1:1
1:2
1:9
1:10
1:2
9
10
11
12
13
14
15
As cyclopentadiene would be unable to act as a hydride
source, we felt that a radical pathway may be involved in
the formation of 3. The addition of the radical scavenger
galvinoxyl in a loading equivalent to that of the catalyst led
to 3 in 63% yield (Table 2, entry 3). This increased the yield
almost threefold and suggested that a radical was involved;
[a] Reaction performed at 1 mmol scale with 20 mol-% catalyst and
at 0.5 m in CH2Cl2 for 24 h at room temperature. [b] Combined
yield. [c] Determined by 1H NMR integrations. [d] 20 mol-%
galvinoxyl added.
Across the Lewis acid series, yields and ratios between however; it was most likely involved in a manner that pro-
2 and 3 were variable, and the yields ranged from 8% to hibits the progress of the reaction. Further support through
quantitative. Low amounts of product were found with deuterium labeling of cyclopentadiene[14] found no deuter-
5042
www.eurjoc.org
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2013, 5041–5044

Cyclopentadiene Acting as Hydrogen Source
ium incorporation into 3, although the high excess and low
level of label could be reasons for this.
There was no directly comparable reaction present in the
literature as noted here. The closest relative would be trans-
fer hydrogenation, which, as the name suggests, is the trans-
fer of hydrogen from one molecule to another.[15,16] Usually,
Hantzsch’s ester is used,[17] although tetralin (4), cyclohex-
ene (5), 1,3-cyclohexadiene (6), and 1,4-cyclohexadiene have
been shown to be able to undergo transfer hydrogena-
tion.[16] In terms of selectivity, this type of reduction would
target the alkene moiety, although the Meerwein–Pondorf–
Verley (MPV) reduction selectively reduces carbonyl
groups.[5] Under the conditions for MPV reduction, 3 was
undetected. With this result, we explored the use of known
transfer-hydrogenation agents to ascertain if they would
give the same product under our conditions. This was found
to be the case; 4 provided 3 in 33% yield, whereas 5 gave 3
in 29% yield, and 6 was found to give 3 in 48% yield
(Scheme 2). No cycloadduct was found with 6.
Scheme 3. Reaction of hydrogen and phenyl-substituted ene-tri-
ketones and cyclopentadiene with aluminum chloride.
Conclusions
The action of Lewis acids on cyclohex-5-ene-1,2,4-
triones and cyclopentadiene exhibits twofold reactivity. The
simplest reactant, 7, showed the expected irreversible for-
mation of 9 without the formation of parent dione 11. Sub-
stitution at C-6, that is, both 1 and 8, showed disparate
ratios of the cycloadducts, 2 and 10, and diones, 3 and 12.
This can be attributed to two ideas: (1) steric interference
between the substituent allows reversible cycloaddition;
(2) a thermodynamically favored product. With the ob-
served diastereoselectivities, the former idea seems justified.
The latter idea stems from the degree of unsaturation con-
tained in 1 and its congeners. Because of the quinone-like
structure, it would be a simple matter of aromatization, if
not for the quaternary carbon. We believe this disrupts the
aromaticity, which leads to a “frustrated aromatic” com-
pound. Under the catalytic conditions for cyclization, 1 un-
dergoes reduction to the parent 1,3-dione, whereas 7 and 8
are able to condense with cyclopentadiene. The major reac-
tivity difference between the three substrates is the fluxional
nature of 2 versus 9 and 10. The pathway leading to the
diones, 3, 11, and 12, is a convoluted problem that we have
only begun to explore, and it is the first case of cyclopenta-
diene acting as a hydrogen source.
Scheme 2. Transfer hydrogenation of 2 to 3 with tetralin, cyclohex-
ene, and cyclohexadiene.
Analogous systems were also exposed to the Lewis acid
conditions to explore the generality of the reaction
(Scheme 3). The reduction was solely noted with 1. Alterna-
tive ene-triketones such as 3,3-dimethylcyclohex-5-ene-
1,2,4-trione (7) and 3,3-dimethyl-6-phenylcyclohex-5-ene-
1,2,4-trione (8) showed predominately the cycloadducts
(i.e., 9 and 10, respectively) with a diminished amount of
their corresponding reduced products (i.e., 11 and 12,
respectively). Exposure of 7 to AlCl3 led to the rapid forma-
tion of 9 over 30 min with no discernible formation of re-
duced product 11. Over 24 h, 11 was formed in trace
amounts. During this time period, no change in the dia-
stereoselectivity was noticed for 9. This suggests that adduct
formation was irreversible, which contrasts the reversibility
of 2. Although a small amount of 12 was present, the Diels–
Experimental Section
General Procedure for Cycloaddition: A vial was charged with the
Lewis acid (0.2 mmol), 1 (1 mmol), and CH2Cl2 (3 mL). Freshly
distilled cyclopentadiene was added, and the solution was stirred
at ambient temperature for 24 h. The solution was then separated
by flash column chromatography on silica gel (20% EtOAc/hex-
anes) to afford a mixture of 2 and 3. HPLC traces and NMR inte-
grations were used to determine molar ratios of the sample.
(4S)-4-(Methylindole)-B-phenyl-N-tosyl-1,2,3-oxaborazolidinone:
The procedure of Corey was used with minor modification.[18]
Alder adduct was the major product if 8 was treated with Tryptophan (10.54 g, 51.7 mmol) was dissolved in THF/H2O (1:9,
150 mL), followed by the addition of Et3N (18.0 mL, 129.1 mmol),
and the resulting solution was cooled to 0 °C. p-Toluenesulfonyl
chloride (10.905 g, 57.2 mmol) was added in a single portion. The
mixture was allowed to stir for 1 h at 0 °C, and then warmed to
room temperature and stirred for an additional 3 h. The solution
was extracted with CH2Cl2 (3ϫ 50 mL), dried with Na2SO4, and
concentrated in vacuo. The material was sufficiently pure to carry
forward. The residue was dissolved in toluene, and phenylboronic
acid (6.15 g, 50.8 mmol) was added. The suspension was then
heated at reflux through a Soxhlet extraction apparatus with CaH2
for 6 h, and then concentrated under reduced pressure and redis-
cyclopentadiene. Presumably, the planarity of the phenyl
ring allows for a more stable adduct and irreversible cyclo-
addition. In the absence of the catalyst, 8 will undergo cy-
cloaddition at elevated temperatures (refluxing toluene);
however, both 10 and 12 are present in an approximately
equimolar distribution. The data support the conclusion
that reduction is a thermodynamic preference that can be
disfavored by irreversible adduct formation. The reversibil-
ity is most likely due to the steric hindrance between the
syn-facial methyl group and the methylene unit of 2.
Eur. J. Org. Chem. 2013, 5041–5044
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5043

N. A. Eddy, J. J. Richardson, G. Fenteany
SHORT COMMUNICATION
[6] O. Diels, K. Alder, Justus Liebigs Ann. Chem. 1928, 460, 98–
122.
solved in toluene (120 mL, 0.43 m theoretical). A small sample for
analysis was separated by flash column chromatography on silica
gel (1% MeOH/CH2Cl2) to afford the oxazoboralidinone as a white
powder (m.p. 131–134 °C). Rf = 0.26 (1% MeOH/CH2Cl2). 1H
NMR (500 MHz, [D6]DMSO): δ = 10.8 (s, 1 H, indole NH), 8.13
(d, J = 8.44 Hz, 1 H, indole CH), 7.47 (d, J = 8.25 Hz, 2 H, aryl
CH), 7.23–7.31 (m, 4 H, aryl CH), 7.17 (m, 3 H, aryl CH), 7.06
(m, 2 H, aryl CH), 6.92 (t, J = 7.87 Hz, 1 H, aryl CH), 3.89 (q, J
= 6.99 Hz, 1 H, CH), 3.07 (m, 2 H, CH2), 2.85 (dd, J = 14.49,
7.78 Hz, 1 H, CH), 2.31 (s, 3 H, Ar-CH3), 2.29 (s, 3 H) ppm. 13C
(125 MHz, [D6]DMSO): δ = 172.6 (C=O), 142.2 (C), 138.0 (C),
136.1 (C), 129.1 (CH), 128.9 (CH), 128.2 (CH), 126.9 (CH), 126.2
(C), 123.9 (CH), 120.8 (CH), 118.3 (CH), 117.8 (CH), 111.4 (CH),
108.8 (CH), 56.5 (CH), 28.3 (CH2), 21.0 (CH3) ppm. 11B
(160 MHz, [D6]DMSO): δ = 20.0 ppm. HRMS (ESI+): calcd. for
[7] a) T. Inukai, T. Kojima, J. Org. Chem. 1967, 32, 869–871; b) T.
Inukai, T. Kojima, J. Org. Chem. 1967, 32, 872–875; c) H. B.
Kagan, O. Riant, Chem. Rev. 1992, 92, 1007–1019; d) M. Laut-
ens, W. Klute, W. Tam, Chem. Rev. 1996, 96, 49–92; e) P. Yates,
P. Eaton, J. Am. Chem. Soc. 1960, 82, 4436–4437.
[8] G. I. Fray, R. Robinson, J. Am. Chem. Soc. 1961, 83, 249.
[9] a) J. M. Coxon, R. D. J. Froese, B. Ganguly, A. P. Marchand,
K. Morokuma, Synlett 1999, 1681–1703; b) R. Gleiter, M. C.
Bohm, Pure Appl. Chem. 1983, 55, 237–244.
[10] T. J. Davis, J. Balsells, P. J. Carroll, P. J. Walsh, Org. Lett. 2001,
3, 2161–2164.
[11] a) H.-U. Blaser, Chem. Rev. 1992, 92, 935–952; b) J. M. Brunel,
Chem. Rev. 2005, 105, 857–897; c) E. Corey, T. Shibata, T. Lee,
J. Am. Chem. Soc. 2002, 124, 3808–3809; d) L. Dias, J. Braz.
Chem. Soc. 1997, 8, 289–332.
C24H22BN2O4S [M + H]+ 445.1393; found 445.1431. [α]D
+16.3585 (c = 0.119, CH2Cl2).
=
[12] The adduct formed had [α]D = –8.3667 (c = 0.2, CH2Cl2).
[13] K. Alder, G. Stein, Angew. Chem. 1937, 50, 510–519.
[14] B. A. Lyons, J. Pfeifer, T. H. Peterson, B. K. Carpenter, J. Am.
Chem. Soc. 1993, 115, 2427–2437.
Supporting Information (see footnote on the first page of this arti-
[15] a) T. Nishiguchi, K. Tachi, K. Fukuzumi, J. Am. Chem. Soc.
1972, 94, 8916–8917; b) G. Brieger, T. J. Nestrick, Chem. Rev.
1973, 73, 567–580; c) J. W. Yang, M. T. Hechavarria Fonseca,
B. List, Angew. Chem. 2004, 116, 6829–6832; Angew. Chem.
Int. Ed. 2004, 43, 6660–6662; d) D. Xue, Y.-C. Chen, X. Cui,
Q.-W. Wang, J. Zhu, J.-G. Deng, J. Org. Chem. 2005, 70, 3584–
3591; e) C. Wang, X. Wu, J. Xiao, Chem. Asian J. 2008, 3,
1750–1770; f) K. Fujita, R. Yamaguchi, Synlett 2005, 560–571;
g) T. Naoto, H. Takaya, S. Murahasi, Chem. Rev. 1998, 98,
2599–2660; h) G. Zassinovich, G. Mestroni, S. Giadiali, Chem.
Rev. 1992, 92, 1051–1069.
[16] a) A. M. Felix, E. P. Heimer, T. J. Lambros, C. Tzougraki, J.
Meienhofer, J. Org. Chem. 1978, 43, 4194–4196; b) T. Nishigu-
chi, K. Tachi, K. Fukuzumi, J. Org. Chem. 1974, 40, 237–240.
[17] For reviews on the use of Hanztsch’s ester in Synthesis see: a)
S.-L. You, Chem. Asian J. 2007, 2, 820–827; b) S. G. Ouellet,
A. M. Walji, D. W. C. MacMillan, Acc. Chem. Res. 2007, 40,
1327–1339.
cle): Spectra of the oxazoboralidinone catalyst.
Acknowledgments
The authors would like to thank Profs. William Bailey (University
of Connecticut) and James Bobbitt (University of Connecticut) for
their assistance with the sealed-tube reactions and helpful dis-
cussions. This work was supported by the National Institutes of
Health (NIH) (grant GM077622, to G. F.).
[1] J. M. Bobbitt, J. Org. Chem. 1998, 63, 9367–9374.
[2] N. A. Eddy, C. B. Kelly, M. A. Mercandante, N. E. Leadbeater,
G. Fenteany, Org. Lett. 2012, 14, 498–501.
[3] N. O. Calloway, Chem. Rev. 1935, 17, 327–392.
[4] C. Santelli-Rouvier, M. Santelli, Synthesis 1983, 429–442.
[5] T. Ooi, T. Miura, Y. Itagaki, H. Ichikawa, K. Maruoka, Syn-
thesis 2002, 279–291.
[18] E. J. Corey, K. Ishihara, Tetrahedron Lett. 1992, 33, 6807–6810.
Received: May 14, 2013
Published Online: June 26, 2013
5044
www.eurjoc.org
© 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2013, 5041–5044