Intramolecular anodic olefin coupling reactions and the synthesis of cyclic amines

Article abstract of DOI:10.1021/ja910586v

Anodic olefin coupling reactions using a tosylamine trapping group have been studied. The cyclizations are favored by the use of a less-polar radical cation and more basic reaction conditions. The most significant factor for obtaining good yields of cyclic product is the use of the more basic reaction conditions. However, a number of factors including the nature of both the solvent and the electrolyte used can influence the yield of the cyclizations. The cyclizations allow for the rapid synthesis of both substituted proline and pipecolic acid type derivatives.

Full text of DOI:10.1021/ja910586v

Published on Web 02/04/2010
Intramolecular Anodic Olefin Coupling Reactions and the
Synthesis of Cyclic Amines
Hai-Chao Xu and Kevin D. Moeller*
Department of Chemistry, Washington UniVersity, St. Louis, Missouri 63130
Received December 16, 2009; E-mail: moeller@wustl.edu
Abstract: Anodic olefin coupling reactions using a tosylamine trapping group have been studied. The
cyclizations are favored by the use of a less-polar radical cation and more basic reaction conditions. The
most significant factor for obtaining good yields of cyclic product is the use of the more basic reaction
conditions. However, a number of factors including the nature of both the solvent and the electrolyte used
can influence the yield of the cyclizations. The cyclizations allow for the rapid synthesis of both substituted
proline and pipecolic acid type derivatives.
Scheme 1
Introduction
The anodic oxidation of an enol ether, vinyl sulfide, or ketene
acetal generates a radical cation that can be used to trigger a
number of interesting cyclization reactions.1 Typically, the
intermediates are trapped with either an electron-rich olefin, an
aromatic ring, or an alcohol nucleophile in order to generate a
variety of carbocyclic, tetrahydrofuran, and tetrahydropyran
products. In a retrosynthetic analysis, the cyclizations can be
recognized by noting that they involve umpolung reactions
between the normally nucleophilic carbon alpha to a carbonyl
and a second nucleophile.
In principle, anodic cyclization reactions of this type should
also be useful for the synthesis of cyclic amino acid derivatives
and a variety of peptidomimetics (Scheme 1).2-6 In this scenario,
the target amino acid would be dissected between the carbon
alpha to the carboxylic acid and a nitrogen-based nucleophile
to afford an acyclic electrolysis substrate 2.7 The cyclization
would then offer the opportunity to construct the amino acid
derivative with control over the stereochemistry of the R carbon,
even in cases where a tetrasubstituted carbon was needed at
this position.8 The substrates for the electrolyses would be
available in an asymmetric fashion by taking advantage of the
previous route to alcohol-based substrates like 3.8a With this in
(5) For synthetic routes to peptidomimetics containing cyclic amino acid
derivatives: (a) Scott, W. L.; Alsina, J.; Kennedy, J. H.; O’Donnell,
M. J. Org. Lett. 2004, 6, 1629. (b) Palomo, C.; Aizpurua, J. M.; Benito,
A.; Miranda, J. I.; Fratila, R. M.; Matute, C.; Domercq, M.; Gago, F.;
Martin- Santamaria, S.; Linden, A. J. Am. Chem. Soc. 2003, 125,
16243. (c) Colombo, L.; Di Giacomo, M.; Vinci, V.; Colombo, M.;
Manzoni, L.; Scolastico, C. Tetrahedron 2003, 59, 4501. (d) Dolbeare,
K.; Pontoriero, G. F.; Gupta, S. K.; Mishra, R. K.; Johnson, R. L.
J. Med. Chem. 2003, 46, 727. (e) Khalil, E. M.; Pradhan, A.; Ojala,
W. H.; Gleason, W. B.; Mishra, R. K.; Johnson, R. L. J. Med. Chem.
1999, 42, 2977. (f) Aube, J. AdV. Amino Acid Mimetics Peptidomi-
metics 1997, 1, 193. (g) Tong, Y.; Olczak, J.; Zabrocki, J.; Gershen-
gorn, M. C.; Marshall, G. R.; Moeller, K. D. Tetrahedron 2000, 56,
9791. (h) Duan, S.; Moeller, K. D. Tetrahedron 2001, 57, 6407. (i)
Liu, B.; Brandt, J. D.; Moeller, K. D. Tetrahedron 2003, 59, 8515.
(6) For examples of the use of electrochemistry to functionalize cyclic
amino acids, see (a) Beal, L. M.; Liu, B.; Chu, W.; Moeller, K. D.
Tetrahedron 2000, 56, 10113 and references cited therein. (b) Fobian,
Y. M.; d’Avignon, D. A.; Moeller, K. D. Bioorg. Med. Chem. Lett.
1996, 6, 315. (c) Fobian, Y. M.; Moeller, K. D. Peptidomimetic
Protocols; Methods in Molecular Medicine, Vol. 23; Humana Press:
Totowa, NJ, 1999; p 259. (d) Tong, Y.; Fobian, Y. M.; Wu, M.; Boyd,
N. D.; Moeller, K. D. J. Org. Chem. 2000, 65, 2484. (e) Cornille, F.;
Fobian, Y. M.; Slomczynska, U.; Beusen, D. D.; Marshall, G. R.;
Moeller, K. D. Tetrahedron Lett. 1994, 35, 6989. (f) Cornille, F.;
Slomczynska, U.; Smythe, M. L.; Beusen, D. D.; Moeller, K. D.;
Marshall, G. R. J. Am. Chem. Soc. 1995, 117, 909. (g) Slomczynska,
U.; Chalmers, D. K.; Cornille, F.; Smythe, M. L.; Beusen, D. D.;
Moeller, K. D.; Marshall, G. R. J. Org. Chem. 1996, 61, 1198. (h)
Simpson, J. C.; Ho, C.; Shands, E. F. B.; Gershengorn, M. C.; Marshall,
G. R.; Moeller, K. D. Bioorg. Med. Chem. 2002, 10, 291. (i) Li, W.;
Hanau, C. E.; d’Avignon, A.; Moeller, K. D. J. Org. Chem. 1995, 60,
8155.
(1) For a recent account, see (a) Moeller, K. D. Synlett 2009, 1208. For
reviews of early work, see (b) Moeller, K. D. Tetrahedron 2000, 56,
9527. (c) Moeller, K. D. Top. Curr. Chem. 1997, 185, 49.
(2) For a review of cyclic amino acid derivatives, see Park, K.-H.; Kurth,
M. J. Tetrahedron 2002, 58, 8629.
(3) For recent references, see (a) Mitsunaga, S.; Ohbayashi, T.; Sugiyama,
S.; Saitou, T.; Tadokoro, M.; Satoh, T. Tetrahedron: Asymmetry 2009,
20, 1697. (b) Wang, Y.-G.; Mii, H.; Kano, T.; Maruoka, K. Bioorg.
Med. Chem. Lett. 2009, 19, 3795. (c) Kaname, M.; Yamada, M.;
Yoshifuji, S.; Sashida, H. Chem. Pharm. Bull. 2009, 57, 49. (d)
Dickstein, J. S.; Fennie, M. W.; Norman, A. L.; Paulose, B. J.;
Kozlowski, M. C. J. Am. Chem. Soc. 2008, 130, 15794. (e) Prazeres,
V. F. V.; Castedo, L.; Gonzalez-Bello, C. Eur. J. Org. Chem. 2008,
23, 3991. (f) Simila, S. T. M.; Martin, S. F. Tetrahedron Lett. 2008,
49, 4501. (g) Undheim, K. Amino Acids 2008, 34, 357, and references
therein.
(4) For leading references concerning the use of lactam-based peptido-
mimetics containing cyclic amino acid derivatives, see (a) Cluzeau,
J.; Lubell, W. D. Biopolymers 2005, 80, 98. (b) I-lalab, L.; Gosselin,
F.; Lubell, W. D. Biopolymers 2000, 55, 101. (c) Hanessian, S.;
McNaughton-Smith, G.; Lombart, H.-G.; Lubell, W. D. Tetrahedron
1997, 53, 12789. For additional lead references, see (d) Polyak, F.;
Lubell, W. D. J. Org. Chem. 1998, 63, 5937. (e) Curran, T. P.;
Marcaurell, L. A.; O’Sullivan, K. M. Org. Lett. 1999, 1, 1225. (f)
Gosselin, F.; Lubell, W. D. J. Org. Chem. 2000, 65, 2163. (g) Polyak,
F.; Lubell, W. D. J. Org. Chem. 2001, 66, 1171. (h) Feng, Z.; Lubell,
W. D. J. Org. Chem. 2001, 66, 1181.
(7) For a preliminary account of this work, see Xu, H.-C.; Moeller, K. D.
J. Am. Chem. Soc. 2008, 130, 13542.
9
10.1021/ja910586v  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 2839–2844 2839

A R T I C L E S
Xu and Moeller
Scheme 2
Scheme 4
Scheme 3
a less polarized radical cation might improve the cyclizations
highlighted in Scheme 2.
mind, the key question was whether or not a nitrogen-based
nucleophile would be compatible with the oxidative coupling
reaction.
Initial Studies
In order to test this idea and give the study the best chance
for success, a p-toluenesulfonamide trapping group was selected
for the cyclizations. The sulfonamide trapping group was chosen
because of its propensity for serving as a nucleophile.11 Four
substrates were synthesized as outlined in Scheme 4.12 One
aspect of this work deserves comment. Initially, the enol ether
substrate (9a) could not be made via the very effective
Mitsunobu reaction-t-Boc deprotection strategy used to make
substrates 9b,c because the enol ether group was not stable to
the deprotection reaction. For this reason, the alternative low-
yielding sequence illustrated in Scheme 4 was used. Recently,
this problem has been solved by developing an alternative
method for the deprotection. This method involves treatment
of the t-Boc-protected sulfonamide with methyllithium in ether
at -20 °C for 10 min. With these conditions, the t-Boc group
can be removed from the sulfonamide in 90% yield even in the
presence of an enol ether. These conditions are now used for
the deprotection reaction in all circumstances.
The initial substrate oxidized was enol ether 9a (Table 1,
entry 1).7 The reaction was carried out by use of a reticulated
vitreous carbon anode, a Pt cathode, 0.1 M tetraethylammonium
tosylate in 30% MeOH/tetrahydrofuran (THF) electrolyte solu-
tion, 2,6-lutidine as a proton scavenger, and a constant current
of 6 mA. The reaction was conducted until 2.2 F/mol current
had been passed (10% more than the theoretical amount needed
for the two-electron oxidation). As in the previous studies, the
cyclization was not very successful, affording the desired
product in only 20% yield.
Initial attempts to answer this question were not encouraging.
A number of substituent patterns were tried for the reactions,
with the most successful being illustrated in Scheme 2. In this
case, a methoxyenol ether was tethered to an acylated amine in
order to generate a five-membered ring product upon cyclization.
A low yield of the desired product was obtained (Scheme 2).
Varying the reaction conditions did not improve the cycliza-
tion. Neither did changing the protecting group on the nitrogen.
Even the known alternative approach of triggering related
cyclizations by oxidizing an amine or hydroxylamine derivative
met with failure.9 Evidently, the enol ether is not compatible
with such an approach.
Fortunately, recent efforts have demonstrated that the nature
of the substituents on a radical cation intermediate can have a
profound influence on its ability to react with various trapping
groups.10 Morepolarizedradicalcationstendtofavorcarbon-carbon
bond forming reactions, while less polarized radical cations tend
to favor reactions with heteroatomic trapping groups. This
observation was used nicely to optimize a coupling reaction
between an electron-rich olefin and an electron-rich furan ring
during our synthesis of the arteannuin ring system (Scheme 3).8c
In this case, the use of a more polarized ketene acetal-derived
radical cation dramatically improved the yield of the cyclization
relative to the use of an enol ether-derived radical cation.
Having used a more polarized radical cation to improve the
cyclization highlighted in Scheme 3, we wondered if the use of
An immediate improvement in the cyclization was observed
when the vinyl sulfide substrate 9b was used. In this case, the
reaction afforded a 54% isolated yield of the cyclic product 11b
under reaction conditions identical to those employed for the
oxidation of 9a. The improvement in the yield was ascribed to
the decrease in polarization of the radical cation derived from
the vinyl sulfide. A feel for the lower polarization of the vinyl
sulfide-derived radical cation can be gained by examining the
polarity of the substrates by 13C NMR (Figure 1). For example,
the presence of the sulfide does not polarize the double bond in
the starting material, leaving both olefinic carbons with the same
(8) For selected examples of reactions leading to tetrasubstituted carbons,
see ref 7 in addition to (a) Xu, H.-C.; Brandt, J. D.; Moeller, K. D.
Tetrahedron Lett. 2008, 49, 3868. (b) Tang, F.; Moeller, K. D. J. Am.
Chem. Soc. 2007, 129, 12414. (c) Wu, H.; Moeller, K. D. Org. Lett.
2007, 9, 4599. (d) Mihelcic, J.; Moeller, K. D. J. Am. Chem. Soc.
2004, 126, 9106–9111. (e) Liu, B.; Duan, S.; Sutterer, A. C.; Moeller,
K. D. J. Am. Chem. Soc. 2002, 124, 10101. (f) Reddy, S. H. K.; Chiba,
K.; Sun, Y.; Moeller, K. D. Tetrahedron 2001, 57, 5183. (g) Frey,
D. A.; Reddy, S. H. K.; Wu, N.; Moeller, K. D. J. Org. Chem. 1999,
64, 2805–2813. (h) Tinao-Wooldridge, L. V.; Moeller, K. D.; Hudson,
C. M. J. Org. Chem. 1994, 59, 2381.
(9) (a) Tokuda, M.; Miyamoto, T.; Fujita, H.; Suginome, H. Tetrahedron
1991, 47, 747. (b) Tokuda, M.; Fujita, H.; Miyamoto, T.; Suginome,
H. Tetrahedron 1993, 49, 2413. (c) Karady, S.; Corley, E. G.;
Abramson, N. L.; Amato, J. S.; Weinstock, L. M. Tetrahedron 1991,
47, 757. (d) Abou-Elenien, G. M.; El-Anadouli, B. E.; Baraka, R. M.
J. Chem. Soc.; Perkin Trans. 2 1991, 1377.
(11) (a) Liu, G.; Stahl, S. S. J. Am. Chem. Soc. 2007, 129, 8328. (b) Fix,
S. R.; Brice, J. L.; Stahl, S. S. Angew. Chem., Int. Ed. 2002, 41, 164.
(12) For complete experimental details concerning syntheses of the
substrates, see the Supporting Information.
(10) See ref 8b as well as Huang, Y.-T.; Moeller, K. D. Tetrahedron 2006,
62, 6536.
9
2840 J. AM. CHEM. SOC. VOL. 132, NO. 8, 2010

Intramolecular Anodic Olefin Coupling Reactions
A R T I C L E S
Table 1
Scheme 5
product meant that the toluenesulfonamide group was not nearly
as good a nucleophile for trapping the radical cation as was an
alcohol.
Electrolyses under More Basic Conditions
The intramolecular reaction was not competing effectively
with the intermolecular trapping group. One method for solving
this problem would be to increase the nucleophilicity of the
nitrogen trapping group. In order to accomplish this goal, the
reactions were repeated under more basic reaction conditions.
The idea was to take advantage of the acidity of the sulfonamide
relative to that of the methanol solvent. To this end, the 2,6-
lutidine used in the initial reactions was replaced with lithium
methoxide. For an electrolysis reaction in an undivided cell,
acid is produced at the anode and equal amount of base produced
at the cathode by the reduction of methanol. The electrolysis
remains at the same pH throughout the entire course of the
reaction. Hence, the use of lithium methoxide in place of 2,6-
lutidine raises the pH of the reaction medium used for the
electrolysis from start to finish. The lithium methoxide was
either produced in situ by adding n-BuLi to the methanol-based
solvent mixture or introduced as a THF solution (available from
Aldrich).
chemical shift. For comparison, oxygen serves as a clear electron
donor to the π-system. The improvement in the cyclization
observed with 9b suggested that the use of the vinyl sulfide
might allow for the development of a synthetically useful
oxidative approach to cyclic amino acids.
The more basic reaction conditions had a dramatic influence
on the reactions (Table 2). For example, the oxidation of
substrate 9a (entry 1) led to the formation of cyclic product
11a in a 70% isolated yield when 30% MeOH/THF was used
as the solvent for the reaction. When the solvent was changed
to methanol, the isolated yield of 11a climbed to 82%. Once
again, the reaction utilizing the thioenol ether (9b) led to a higher
yield of product (entry 2) than did the reaction using the
methoxyenol ether. In this case, the yield of the reaction was
85% when 30% MeOH/THF was used as solvent and 90% when
methanol was used.
The role of LiOMe in these reactions is thought to be
deprotonation of the toluenesulfonamide in order to make a
better coupling partner for the reaction. This can occur either
by the anion of the sulfonamide serving as a better nucleophile
for the electrochemically generated radical cation, or by the
oxidation taking place at the nitrogen in order to form a nitrogen-
based radical that then adds to the olefin (Scheme 6). These
two pathways cannot be readily distinguished from the oxidation
potentials of the two coupling partners because they have the
potential to equilibrate through a fast intramolecular electron
transfer.14 Both pathways lead to the same cyclic radical product
18. Radical 18 is then oxidized to form a stabilized cation 19
that is trapped with methanol solvent to form the final cyclic
product. It is thought that the reactions benefit from the use of
pure methanol as solvent because of a need to trap cation 19
Figure 1
A suggestion for how the reactions could be further optimized
arose from the study of ketene dithioacetal substrates 9c and
9d. Ketene dithioacetals have proven to be very effective
coupling partners for alcohol nucleophiles in anodic olefin
coupling reactions.13 However, the anodic oxidation of 9c under
the same conditions described above led to only a 14% yield
of cyclic product 11c along with 4% of an overoxidized product.
The overoxidized product appeared to arise from the elimination
of proton from the initial cyclization product (15) competing
with the desired methanol trapping reaction. The elimination
was followed by oxidation of the resulting ketene acetal (16)
and subsequent methanol trapping of the radical cation generated
(Scheme 5).
In order to avoid the elimination reaction, the oxidation of
substrate 9d was examined. In this case, a methyl group replaced
the proton on the ꢀ-carbon of the electron-rich olefin. The
reaction again gave rise to a poor yield (19%) of the desired
cyclic product. The reaction led to a six-membered ring product
(13) formed by competitive trapping of the radical cation with
methanol solvent. The formation of the six-membered ring
9
J. AM. CHEM. SOC. VOL. 132, NO. 8, 2010 2841

A R T I C L E S
Xu and Moeller
Table 2
Table 3
entry
reaction conditions
yield (%)
1
2
3
4
5
6
7
0.5 equiv of n-BuLi, MeOH, 0.1 M Et4NOTs,
6 mA, 9.4 F/mol
0.5 equiv of n-BuLi, 30% MeOH/THF, 0.1 M Et4NOTs,
6 mA, 2.4 F/mol
0.5 equiv of n-BuLi, 60% MeOH-THF, 0.1 M Et4NOTs,
6 mA, 2.4 F/mol
0.5 equiv of n-BuLi, MeOH, 0.03 M Et4NOTs,
6 mA, 3.3 F/mol
0.5 equiv of n-BuLi, MeOH, 0.003 M Et4NOTs,
6 mA, 3.8 F/mol
0.5 equiv of n-BuLi, MeOH, 0.03 M LiCIO4 6 mA,
4 F/mol
2,6-lutidine, MeOH, 0.03 M LiCIO4, 6 mA,
4 F/mol
39
33
34
66
47
10
18
Even with the more basic reaction conditions and the use of
methanol solvent to accelerate the trapping of the cyclic cation
19, the cyclization originating from 9c remained problematic
(Table 3). While the overall yield of cyclic product could be
raised to 66% by reducing the concentration of the electrolyte
to further increase the concentration of MeOH close to the
surface of the anode, the elimination reaction leading to
overoxidation of the product could not be avoided.
The use of the more basic reaction conditions did solve the
problems initially encountered with substrate 9d (Table 2, entry
3). In this case, the reaction utilizing the 30% MeOH/THF
conditions formed the desired cyclization product in 72%
isolated yield without any of the methanol trapping product.
The use of MeOH as the solvent did lead to some of the
competing six-membered ring MeOH trapping product but did
not lower the yield of the desired five-membered ring product
obtained. Evidently, the loss of product through trapping of the
radical cation with methanol solvent was counteracted by the
improved yield of the cyclic product by methanol trapping of
19.
The need for having sufficient methanol present to trap the
cyclic cation 19 was acutely observed when substrate 9e was
oxidized (Table 2, entry 4).12,15 The oxidation of 9e in 30%
MeOH/THF led to only a 28% yield of the desired cyclic
product. The isolated yield of the reaction was raised to 80%
by switching to MeOH solvent. The product was formed as a
single diastereomer whose stereochemistry was assigned with
the use of a nuclear Overhauser effect spectroscopy (NOESY)
experiment. Formation of the major diastereomer can be
rationalized by the preference for the transition state in which
steric interactions between the methyl and dithiane groups are
a Reaction conditions: RVC anode, Pt wire cathode, 0.5 equiv of
LiOMe, 30% MeOH/THF, 0.1 M Et4NOTs, 6 mA, 2.0-2.4 F/mol.
b MeOH as solvent. c 0.7 equiv of LiOMe (no additional electrolyte was
used), MeOH.
Scheme 6
(13) In addition to refs 8a and 8e, please see (a) Brandt, J. D.; Moeller,
K. D. Heterocycles 2006, 67, 621. (b) Brandt, J. D.; Moeller, K. D.
Org. Lett. 2005, 7, 3553. (c) Sun, Y.; Liu, B.; Kao, J.; d’Avignon, A.;
Moeller, K. D. Org. Lett. 2001, 3, 1729.
before subsequent rearrangement can take place. In this way,
the reactions require a balance between having the cyclization
occur faster than methanol trapping of the original radical cation
while still having methanol present in sufficient quantity to
efficiently form the final product.
(14) For examples of intramolecular electron transfers in anodic oxidations,
see (a) Duan, D.; Moeller, K. D. J. Am. Chem. Soc. 2002, 124, 9368.
(b) Moeller, K. D.; Wang, P. W.; Tarazi, S.; Marzabadi, M. R.; Wong,
P. L. J. Org. Chem. 1991, 56, 1058.
(15) For a previous example of the final trapping step playing a large role
in the yield of an anodic olefin coupling reaction, see ref 13c.
9
2842 J. AM. CHEM. SOC. VOL. 132, NO. 8, 2010

Intramolecular Anodic Olefin Coupling Reactions
A R T I C L E S
Scheme 7
Scheme 8
Figure 2
Table 4
Scheme 9a
yield (%)
21 22
entry
reaction conditions
1
0.5 equiv of n-BuLi, MeOH, 0.1 M Et4NOTs, 6 mA, 20
2.3 F/mol
9
8
5
2
3
0.5 equiv of n-BuLi, 30% MeOH/THF, 0.1 M
Et4NOTs, 6 mA, 2.3 F/mol
c ) 0.006 M, 0.5 equiv of n-BuLi, MeOH, 0.1 M
Et4NOTs, 6 mA, 14 F/mol
27
17
a Reaction conditions: (a) RVC anode, 0.5 equiv of LiOMe, 0.1 M
Et4NOTs, MeOH, 6 mA, 2.4 F/mol. (b) RVC anode, 0.5 equiv of LiOMe,
0.1 M LiClO4, MeOH, 6 mA, 2.4 F/mol.
minimized (Figure 2). The stereochemistry of the reaction is
consistent with earlier cyclizations using oxygen nucleophiles.16
Reactions originating from thioenol ether (Table 2, entry 5)
and methoxyenol ether (Table 2, entry 6) sustrates were also
compatible with the generation of tetrasubstituted carbons. These
reactions again benefited from the use of methanol as the solvent
and the less polar thioenol ether substrate.
The observation that the cyclizations proceed well with less-
polarized radical cations suggested that they may be compatible
with the use of a variety of olefins. To test this idea, substrates
9h and 9i were studied.12 In both cases, anodic oxidation of
the substrate led smoothly to the cyclized product. With the
styrene-based substrate (Table 2, entry 7), an 88% isolated yield
of product was obtained. Oxidation of the allylsilane-based
substrate (Table 2, entry 8) led to a 90% isolated yield of
product.
Finally, the cyclizations were compatible with the use of a
diene substrate (9j). In this reaction, 0.7 equiv of LiOMe was
used. No additional electrolyte was employed. The reaction led
to a 70% isolated yield of the desired five-membered ring
product. Conditions with LiOMe and no additional electrolyte
were also used for the oxidation of 9b. In this case, a 91%
isolated yield of the cyclic product was obtained. It appears the
LiOMe used in the reaction can fulfill the role of both the 2,6-
lutidine base scavenger and the tetraethylammonium tosylate
electrolyte used in the earlier examples.
generated. A suggested mechanism for the formation of 22 is
shown in Scheme 7. Alternatively, the elimination reaction might
arise from either an intramolecular deprotonation of the radical
cation by the toluenesulfonamide anion or an intramolecular
hydrogen atom abstraction by a nitrogen-based radical.
The elimination reaction was also a problem when attempts
were made to use the oxidation of other nonpolar, electron-
rich olefins to trigger the formation of six-membered ring
pipecolic acid derivatives. For example, the use of an allylsilane
coupling partner for the toluenesulfonamide led to a 25% yield
of the desired six-membered ring product along with 20% of a
product derived from the elimination reaction (Scheme 8).
Insight into the nature of the elimination reaction was gained
from examining the electrolyses of substrates 26a and 26b
(Scheme 9).12 In these substrates, a methyl group was added to
the electron-rich olefin in an attempt to make a tetrasubstituted
carbon. In both cases, a low yield of the desired product 27
was obtained along with a product (28) derived from the
elimination. In these cases, the elimination occurred from both
allylic positions, leading to both a seven- and a five-membered
ring product. The elimination of a proton from the methyl group
suggests that the elimination reaction involves an intermolecular
deprotonation (Scheme 7). An intramolecular deprotonation with
the toluenesulfonamide anion serving as the base would lead
to a five-membered ring product.
Cyclizations Affording Six-Membered Rings
With the success of the cyclizations leading to five-membered
rings, attention was turned toward the more difficult to generate
six-membered ring products.17 Initially, a thioenol ether substrate
(20) was used along with the optimized conditions developed
above (Table 4). The reaction led to only a 20% isolated yield
of the six-membered ring product 21. In addition, a five-
membered ring product (22) resulting from an elimination
reaction involving the initial radical cation intermediate was
The cyclization reaction originating from the oxidation of 26b
formed the six-membered ring product as a single diastereomer
with the bulky dithio methylorthoester trans to the neighboring
methyl group. The yield of this product could be raised to 51%
(along with 10% of the product derived from elimination) by
changing the electrolyte to lithium perchlorate. The change to
lithium perchlorate electrolyte from the tetraethylammonium
tosylate allows for a higher concentration of methanol close to
the surface of the anode. The dependence of the isolated yield
of product on this change supports the earlier observation that
the final methanol trapping step is important for optimizing the
yield of the reaction.
(16) See ref 13 in addition to Duan, S.; Moeller, K. D. Org. Lett. 2001, 3,
2685.
(17) For evidence showing slower anodic olefin coupling reactions leading
to six-membered ring products, see refs 8f and 8h.
9
J. AM. CHEM. SOC. VOL. 132, NO. 8, 2010 2843

A R T I C L E S
Xu and Moeller
Scheme 10
Figure 3
Scheme 12
Scheme 11
are effective olefin coupling partners for the reactions. If a
trisubstituted allylsilane were employed in the cyclization, then
an A1,3-interaction in the transition state for the reaction would
force the allylic R group into a pseudoequatorial position and
the allylic proton into a pseudoaxial position perpendicular to
the π-system of the radical cation (Figure 3). The poor overlap
between the allylic proton and the radical cation would be
expected to lead to a slow elimination reaction and more time
for the desired cyclization.
In practice, this idea worked very nicely (Scheme 12). When
a substrate having an trisubstituted allylsilane coupling partner
and an allylic methyl group (34) was oxidized, a 71% isolated
yield of the cyclized product was obtained as a single diaste-
reomer. Clearly, the use of the larger allylsilane group pushed
the reaction toward the cyclized product.
The oxidation of substrates 29a and 29b was examined in
order to elucidate the success of six-membered ring formation
in the absence of the elimination reaction (Scheme 10).12
The oxidation of 29a proceeded well and led to an 81%
isolated yield of the six-membered ring product. The oxidation
of 29b provided a control experiment to determine whether the
success of the cyclization resulting from 29a was really due to
the gem-methyl groups stopping the elimination reaction or if
it was the result of a faster cyclization due to the gem-dialkyl
effect.18 While the presence of the gem-methyl groups in 29b
improved the cyclization relative to the reaction originating from
20 (Table 4), it was clear that the 81% yield of product 30a
was mainly due to the gem-methyls in 29a preventing the
elimination reaction.
With the knowledge that the six-membered ring cyclizations
can be successful, attention was turned toward developing a
strategy for the synthesis of 3-substituted pipecolic acid deriva-
tives. The plan was to take advantage of sterics to slow down
the elimination reaction and increase the time available for the
cyclization. It was hoped that a single substituent on the allylic
carbon of the substrate would accomplish this task (Scheme
11).12 Since substrates like 32a and 32b can be synthesized in
an asymmetric fashion, a successful cyclization would allow
access to the chiral amino acid derivatives.
The cyclizations met with limited success. When a methyl
group was placed on the allylic carbon of the substrate, the yield
of cyclized product improved to 44%. The yield could be
improved to 62% by placing a larger t-butyldiphenylsiloxy group
on the allylic carbon of the substrate.
While the reaction with the t-butyldiphenylsiloxy group was
successful, the need for such a large group limited the types of
pipecolic acid derivatives that could be made with the cycliza-
tions. What was needed was a method for increasing the
effective steric size of any group used at the allylic position of
the substrate. One strategy for accomplishing this goal would
be to take advantage of the earlier observation that allylsilanes
Conclusions
We have found that anodic olefin coupling reactions can be
used to generate new carbon-nitrogen bonds and synthesize
substituted pyrrolidine and piperidine rings. The cyclization
reactions benefit from the use of LiOMe as a base and methanol
solvent. For the synthesis of five-membered ring products, the
cyclizations proceed well with methoxyenol ether-, thioenol
ether-, and dithioketene acetal-derived substrates. Reactions
using a less polar olefin as a coupling partner afford higher yields
of cyclized product.
Coupling reactions leading to six-membered rings are more
difficult because of the competing elimination of a proton from
the carbon alpha to the radical cation intermediate. This problem
can be minimized by manipulating the sterics of the reactions,
an observation that suggests the use of a trisubstituted allylsilane
as an optimized olefin coupling partner for the reactions.
Overall, the cyclizations hold promise as a new method for
constructing both proline and pipecolic acid-based amino acid
derivatives. Work along these lines is continuing.
Acknowledgment. We thank the National Science Foundation
(CHE-0809142) for their generous support of our work. We also
gratefully acknowledge the Washington University High Resolution
NMR facility, partially supported by NIH Grants RR02004,
RR05018, and RR07155, and the Washington University Mass
Spectrometry Resource Center, partially supported by NIHRR00954,
for their assistance.
(18) For examples, see (a) Sperry, J. B.; Wright, D. L. J. Am. Chem. Soc.
2005, 127, 8034. (b) Jung, M. E.; Gervay, J. J. Am. Chem. Soc. 1991,
113, 224. (c) Lightstone, F. C.; Bruice, T. C. J. Am. Chem. Soc. 1994,
116, 10789. (d) Parrill, A. L.; Dolata, D. P. J. Mol. Struct.
(THEOCHEM) 1996, 370, 187. (e) Kirby, A. J. AdV. Phys. Org. Chem.
1980, 17, 183. For early references, see (f) Eliel, E. L.; Allinger, J. L.;
Angyal, S. J.; Morrison, G. A. Conformational Analysis; Wiley-
Interscience: New York, 1967; p 191. (g) Allinger, J. L.; Zalkow, V.
J. Org. Chem. 1960, 25, 701.
Supporting Information Available: Full experimental and
characterization data for all substrates and products, including
copies of proton and carbon NMR. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA910586V
9
2844 J. AM. CHEM. SOC. VOL. 132, NO. 8, 2010