Development of highly regioselective amidyl radical cyclization based on lone pair-lone pair repulsion

Article abstract of DOI:10.1021/jo800845b

(Chemical Equation Presented) The substituent effect on the reactivity and regioselectivity of N-(4-pentenyl)amidyl radical cyclization was investigated. Exclusive 6-endo cyclization was observed for N-(4-pentenyl)amidyl radicals with internal vinylic heteroatom substitution (Cl, Br, I, OMe, SEt). The substituent on the carbonyl group also showed a significant influence on the reactivity of amidyl radicals, which increases in the order of Ph < Me < OEt. As a result, the photostimulated reactions of N-(4-halopent-4-enyl)amides and carbamates (X = Cl, Br, I) with DIB/I₂ or Pb(OAc)₄/I ₂ led to the efficient and exclusive formation of the corresponding piperidines while those of N-(5-halopent-4-enyl)amides afforded the pyrrolidine products only. The halogen-substitution effect also allowed the 6-exo and 7-endo amidyl radical cyclization to proceed in a highly regioselective manner. The above experimental results, in combination with theoretical analyses, revealed that the lone pair-lone pair repulsion between the nitrogen radical and the vinylic heteroatom played an important role in controlling the regioselectivity of cyclization.

Full text of DOI:10.1021/jo800845b

Development of Highly Regioselective Amidyl Radical Cyclization
Based on Lone Pair-Lone Pair Repulsion
Xinting Yuan, Kun Liu, and Chaozhong Li*
Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences,
354 Fenglin Road, Shanghai 200032, People’s Republic of China
clig@mail.sioc.ac.cn
ReceiVed April 18, 2008
The substituent effect on the reactivity and regioselectivity of N-(4-pentenyl)amidyl radical cyclization
was investigated. Exclusive 6-endo cyclization was observed for N-(4-pentenyl)amidyl radicals with internal
vinylic heteroatom substitution (Cl, Br, I, OMe, SEt). The substituent on the carbonyl group also showed
a significant influence on the reactivity of amidyl radicals, which increases in the order of Ph < Me <
OEt. As a result, the photostimulated reactions of N-(4-halopent-4-enyl)amides and carbamates (X ) Cl,
Br, I) with DIB/I2 or Pb(OAc)4/I2 led to the efficient and exclusive formation of the corresponding
piperidines while those of N-(5-halopent-4-enyl)amides afforded the pyrrolidine products only. The
halogen-substitution effect also allowed the 6-exo and 7-endo amidyl radical cyclization to proceed in a
highly regioselective manner. The above experimental results, in combination with theoretical analyses,
revealed that the lone pair-lone pair repulsion between the nitrogen radical and the vinylic heteroatom
played an important role in controlling the regioselectivity of cyclization.
Introduction
a great potential in the synthesis of natural products such as
alkaloids.5 The oxidative entry to these transient species allows
the direct use of the N-H amides, which in turn makes this
radical strategy of more synthetic value. For example, Nicolaou
and co-workers reported the successful o-iodoxybenzoic acid
(IBX)-mediated 5-exo cyclization of unsaturated N-aryl
amides.3a Studer et al. extended this methodology to the
oxidative cyclization of acylated alkoxyamines.3c
The cyclization of amidyl radicals has two modes, to form
lactams and to form cyclic amines. While most of the literature
works were centered at the generation of lactams, the formation
of cyclic amines received far less attention and the chemistry
is not well understood. Kinetic studies7 by Newcomb et al.
revealed that the rate constant for 5-exo cyclization of N-(4-
pentenyl)butanamidyl radical is 1 order of magnitude lower than
The application of nitrogen-centered radicals in organic
synthesis has gained increasing popularity in the past few years.1
Among various nitrogen-centered radicals, amidyl radicals have
received considerable attention. The highly reactive and elec-
trophilic nature allows amidyl radicals to add to electron-rich
CdC double bonds leading to the direct construction of CsN
bonds.2 In particular, the intramolecular addition provides a
unique entry to N-heterocycles of various sizes,3,4 thus offering
(1) For reviews, see: (a) Neale, R. S. Synthesis 1971, 1. (b) Esker, J.;
Newcomb, M. In AdVances in Heterocyclic Chemistry; Katritzky, A. R., Ed.;
Academic Press: New York, 1993; Vol. 58, p 1. (c) Zard, S. Z. Synlett 1996,
1148. (d) Fallis, A. G.; Brinza, I. M. Tetrahedron 1997, 53, 17543. (e) Stella, L.
In Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-VCH:
Weinheim, Germany, 2001; Vol. 2, p 407.
(2) (a) Guin, J.; Much-Lichtenfeld, C.; Grimme, S.; Studer, A. J. Am. Chem.
Soc. 2007, 129, 4498. (b) Kemper, J.; Studer, A. Angew. Chem., Int. Ed. 2005,
44, 4914. (c) Walton, J. C.; Studer, A. Acc. Chem. Res. 2005, 38, 794.
(3) (a) Nicolaou, K. C.; Baran, P. S.; Zhong, Y.-L.; Barluenga, S.; Hunt,
K. W.; Kranich, R.; Vega, J. A. J. Am. Chem. Soc. 2002, 124, 2233. (b) Gagosz,
F.; Moutrille, C.; Zard, S. Z. Org. Lett. 2002, 4, 2707. (c) Janza, B.; Studer, A.
J. Org. Chem. 2005, 70, 6991.
(5) (a) Sharp, L. S.; Zard, S. Z. Org. Lett. 2006, 8, 831. (b) Cassayre, J.;
Gagosz, F.; Zard, S. Z. Angew. Chem., Int. Ed. 2002, 41, 1783.
(6) (a) Mackiewicz, P.; Furstoss, R.; Wageel, B.; Cote, R.; Lessard, J. J.
Org. Chem. 1978, 43, 3746. (b) Newcomb, M.; Esker, J. L. J. Org. Chem. 1993,
58, 4933.
(4) (a) Tang, Y.; Li, C. Org. Lett. 2004, 6, 3229. (b) Chen, Q.; Shen, M.;
Tang, Y.; Li, C. Org. Lett. 2005, 7, 1625. (c) Lu, H.; Li, C. Tetrahedron Lett.
2005, 46, 5983. (d) Hu, T.; Shen, M.; Chen, Q.; Li, C. Org. Lett. 2006, 8, 2647.
(7) (a) Horner, J. H.; Musa, O. M.; Bouver, A.; Newcomb, M. J. Am. Chem.
Soc. 1998, 120, 7738. (b) Martinez, E., II; Newcomb, M. J. Org. Chem 2006,
71, 557.
6166 J. Org. Chem. 2008, 73, 6166–6171
10.1021/jo800845b CCC: $40.75 2008 American Chemical Society
Published on Web 07/24/2008

Cyclization Based on Lone Pair-Lone Pair Repulsion
that of N-butyl-4-pentenamidyl radical.7a With regard to the
implementation of cyclization reactions, an early study by
Lessard and co-workers indicated that the 5-exo cyclization of
unsaturated acetamidyl radicals is much less efficient while the
6-exo cyclization did not proceed at all.6a As a comparison,
Newcomb et al. showed that the N-(4-pentenyl)benzamidyl
radical underwent efficient 5-exo cyclization.6b However, the
effect of different carbonylic substituents on the reactivity of
amidyl radicals is not clear. It should also be noted that only
5-exo cyclization was implemented so far while cyclizations in
other modes remain unknown or unsuccessful. Due to the easy
removal of the acyl group of amides to provide free amines, it
is certainly of interest to develop efficient and highly regiose-
lective amidyl radical-based strategies toward the synthesis of
cyclic amines.
TABLE 1. Regioselectivity of Cyclization of N-(4-Pentenyl)amidyl
Radicals
product/yield (%)a
entry
1
R
X
2
3 or 4
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1f
Me
OEt
Me
OEt
Me
OEt
OEt
OEt
H
H
Me
Me
Cl
Cl
OMe
SEt
2a (7)
3a (71)
3b (45)
3c (28)
3d (20)
b
b
b
b
2b (20)
2c (47)
2d (54)
2e (60)
2f (70)
2g (79)
2h (90)
During our investigation on the chemistry of amidyl radicals,4
we found that the regioselectivity of 5-hexenamidyl radical
cyclization could be well controlled by vinylic halogen substitu-
tion, leading to the regiospecific 6-exo and 7-endo cyclization.4d
This remarkable halogen effect was then successfully extended
to the control of regioselectivity in sulfonamidyl8a and primary
aminyl8b radical cyclizations. We were motivated to find if (1)
this halogen effect could be extended to the N-alkenylamidyl
radical cyclization system and if (2) other heteroatoms could
have a similar effect in controlling the regioselectivity of
cyclization. Herein we report that efficient and highly regiose-
lective 5-exo, 6-endo, 6-exo, and even 7-endo cyclization of
(N-alkenyl)amidyl radicals could be achieved not only by vinylic
halogen substitution, but also by other vinylic heteroatom (O
and S) substitution. Theoretical calculations in combination with
the experimental results provide us a clear understanding on
the behaviors of amidyl radicals.
1g
1h
a Isolated yield based on 1. b 3 or 4 was not detected by 1H NMR
(300 MHz).
1
by the crude H NMR. The difference between Me and Cl in
controlling the regioselectivity of cyclization is remarkable.
Next, we replaced the chlorine atom in 1 by an ethylthio or a
methoxy group (1g and 1h). Exclusive 6-endo cyclization was
again observed and the yields of 2g (79%) and 2h (90%) were
even higher (entries 7 and 8, Table 1).
As a comparison, amide 5 with a terminal chlorine substitu-
tion was prepared and subjected to the same treatment as above.
A clean reaction was observed and only the 5-exo cyclization
was isolated in 84% yield (eq 1). By comparison with the result
of 1b, the terminal chlorine substitution is highly effective in
directing the cyclization to proceed in a 5-exo mode.
Results and Discussion
We first examined the vinylic substitution effect on the
regioselectivity of N-(4-pentenyl)amidyl radical cyclization. Two
typical types of amidyl radicals were employed as the models:
acetamidyl and carbamidyl radicals. The N-phenylthio-substi-
tuted amides 1 were chosen as the radical precursors and their
reactions with Bu3SnH/AIBN were carried out in benzene at
80 °C according to Newcomb’s procedure.9 The results are
summarized in Table 1. The N-(4-pentenyl)acetamidyl radical
derived from precursor 1a gave mainly pyrrolidine 3a (71%)
along with a small amount of piperidine 2a (7%), indicating
the predominance of 5-exo cyclization (entry 1, Table 1). The
carbamidyl radical generated from 1b also showed the prefer-
ence of 5-exo cyclization (entry 2, Table 1). However, the ratio
was only about 2.2 to 1, indicating the strong influence of the
substituent R on the regioselectivity, which will be further
discussed later. With an internal methyl substitution at the CdC
bond, both the unsaturated radicals afforded the mixture of 5-exo
and 6-endo cyclization products with the slight preference of
6-endo cyclization (entries 3 and 4, Table 1). On the other hand,
switching the methyl to a chlorine atom resulted in the exclusive
formation of 6-endo cyclization products 2e and 2f (entries 5
and 6, Table 1). In these two cases, the 5-exo cyclization
products 3e and 3f were expected to undergo HCl elimination
to give, upon workup with water, the corresponding ketones 4e
and 4f, respectively. However, neither 3 nor 4 could be detected
The above results demonstrated that the regioselectivities of
amidyl cyclization could be well controlled by the vinylic
heteroatom substitution. It is of interest to learn if other halogen
atoms such as Br or I would have the same effect. However,
the use of Bu3SnH prevented the investigation because vinyl
bromides or iodides are also easily reduced by Bu3SnH. It is
also highly desirable to carry out these transformations with
the parent N-H amides rather than the phenylthio-substituted
amides 1 as the substrates; this should be of more synthetic
value. In the meantime, the effect of carbonylic substituents (R
in 1) on the reactivity of amidyl radicals has yet to be examined.
Thus, we next chose the N-H amides 7 bearing an internal
vinylic bromine substituent as the substrates to explore this
4d,8,10
possibility. The amides 7 were treated with Pd(OAc)4/I2
upon continuous UV irradiation at room temperature and the
results are listed in Table 2. The trifluoroacetamide 7a gave no
expected cyclization product while a large amount (72%) of
the starting material was recovered (entry 1, Table 2). The
reaction of benzamide 7b was complicated under the experi-
(9) Esker, J. K.; Newcomb, M. Tetrahedron Lett. 1993, 34, 6877.
(10) Barton, D. H. R.; Beckwith, A. L. J.; Goosen, A. J. Chem. Soc. 1965,
181.
(8) (a) Lu, H.; Chen, Q.; Li, C. J. Org. Chem. 2007, 72, 2564. (b) Liu, F.;
Liu, K.; Yuan, X.; Li, C. J. Org. Chem. 2007, 72, 10231.
J. Org. Chem. Vol. 73, No. 16, 2008 6167

Yuan et al.
TABLE 4. 6-Exo versus 7-Endo Amidyl Radical Cyclization
TABLE 2. Reactions of N-(4-Bromo-4-pentenyl)amides with
Pd(OAc)4/I2
yield (%)a
entry
7
R
7
8
1
2
3
4
5
7a
7b
7c
7d
7e
CF3
Ph
Me
OEt
NMe2
72
15
0
0
0
0
20
60
83
0
a Isolated yield based on 7.
a Isolated yield based on 13. b Pb(OAc)4/I2 were used. c E/Z ) 3/1.
d E/Z ) 4/1. e Two stereoisomers in ∼3:2 ratio.
TABLE 3. 5-Exo versus 6-Endo Amidyl Radical Cyclization
all decomposed presumably because of the possible demethy-
lation of 7e under the oxidative conditions (entry 5, Table 2).
The results in Table 2 indicated that unsaturated acetamides
and carbamates such as 7c and 7d could be the ideal candidates
for amidyl radical cyclization reactions. Our further efforts in
the optimization of reaction conditions showed that, with the
use of 2 equiv of (diacetoxyiodo)benzene (DIB) and 1.5 equiv
of iodine,4d,9,11 the UV irradiation of 7d in CH2Cl2 at rt resulted
in a clean reaction leading to the isolation of the product 8d in
92% yield. On the other hand, under this milder condition (DIB/
I2), the reaction of 7c gave the cyclized product 8c in only 22%
yield. These data implied that carbamidyl radicals are more
reactive than the corresponding acetamidyl radicals.12
We then prepared a number of carbamates and alkanamides
and subjected them to the reaction with DIB/I2 or Pb(OAc)4/I2.
The results are summarized in Table 3. The Br or I substitution
showed the same effect as Cl in controlling the regioselectivity
of cyclization. With internal vinylic halogen subsitution, the
piperidine products were obtained while no corresponding 5-exo
cyclization products similar to 3 or 4 could be detected (entries
1-6, Table 3). The ethyl carbamates gave higher product yields
than the corresponding alkanamides (entries 2 and 4-6, Table
3). On the other hand, with terminal halogen substitution, only
the 5-exo cyclization product pyrrolidines were isolated (entries
7 and 8, Table 3). The cyclic carbamates 9 and 11 showed
similar behaviors (entries 9 and 10, Table 3).
We then extended the above cyclization reactions to the 6-exo
versus 7-endo cyclization system (Table 4). The same halogen
effect was observed. The gem-dimethyl-substitution in 13a-d
was designed to avoid the possible intramolecular 1,5-H
abstraction of carbamidyl radicals. As a comparison, substrate
15 without dimethyl substitution failed to give the expected
cyclization product. Our attempt to further extend the above
strategy to the cyclization in a 7-exo or 8-endo mode with the
use of amides 15-18 as substrates was unsuccessful.
a Isolated yield based on the starting amide. b DIB/I2 were used.
c Pb(OAc)4/I2 were used. d E/Z
) )
85:15. e E/Z 1:1. f Two
stereoisomers in ∼2:1 ratio. g Two stereoisomers in ∼1:1 ratio. h Two
stereoisomers in ∼3:2 ratio.
mental condition and the expected 6-endo cyclization product
8b was isolated in only 20% yield (entry 2, Table 2). As a
comparison, the acetamide 7c afforded the expected piperidine
8c in 60% yield under the same experimental conditions (entry
3, Table 2). The best result was obtained in the case of ethyl
carbamate 7d in which 8d was achieved in 83% isolated yield
(entry 4, Table 2). With the use of urea 7e as the substrate, no
cyclized product could be detected and the starting material was
The above results clearly demonstrated that efficient and higly
regioselective cyclization could be achieved for unsaturated
6168 J. Org. Chem. Vol. 73, No. 16, 2008

Cyclization Based on Lone Pair-Lone Pair Repulsion
TABLE 5. Calculated (UB3LYP/6-31G*) Activation Energies for
the Cyclization of Radicals A
in the order of CF3 < OEt < Me ≈ n-Bu ≈ i-Pr < Ph < NMe2,
ranging from 5.6 to 12.7 kcal/mol (entries 1-7, Table 5). A
similar trend is also seen for the corresponding 5-exo cyclization.
This remarkable substituent effect can be well interpreted based
on electronic and steric effects. In terms of electronic effect,
the computed Mulliken charges on the nitrogen atom of amidyl
radicals A (X ) Br) are summarized in Table 6. The lowest
Mulliken charge is observed in the case of R ) CF3, while the
highest one is observed for R ) NMe2. The changes in the
Mulliken charges closely parallel the changes in the activation
energies for cyclization, indicating the dominance of the
electronic effect on the reactivity of amidyl radicals. Thus, the
more electrophilic the amidyl radicals, the more reactive toward
the electron-rich CdC double bond. The only difference is that
the Mulliken charges are about the same for carbamidyl and
alkanamidyl radicals while the activation energies for 6-endo
cyclization show an approximately 2 kcal/mol difference. This
can be analyzed via steric effect according to the computed
transition state structures for cyclization. The transition states
for 6-endo cyclization, as exemplified by structure 19 (R ) OEt,
X ) Br), are in a chair conformation with both the carbonyl
and the X groups at the axial positions. Thus, the bulkiness of
R becomes an important factor. With the ethyl group swinging
away to avoid the steric congestion and the ether oxygen smaller
in size than Me, the ethoxy group causes, in fact, less steric
hindrance than a methyl group. As a result, the activation energy
for 6-endo cyclization of carbamidyl radicals is lower than that
of acetamidyl radicals. The transition states for 5-exo cyclization,
as exemplified by structure 20 (R ) OEt, X ) Br), are in an
envelop conformation. With the carbonyl group at the equatorial
position in 20, the steric effect of R is now less obvious. Note
that the calculated energies of 5-exo cyclization are about the
same for both carbamidyl and acetamidyl radicals (entries 2-5,
Table 5).
entry
R
X
∆Gq(5-exo)a
∆Gq(6-endo)a
1
2
3
4
5
6
7
8
CF3
OEt
Me
n-Bu
i-Pr
Ph
NMe2
Me
Me
Br
8.3
10.6
11.1
11.0
11.0
12.0
13.6
8.0
5.6
7.0
9.2
8.7
9.4
10.3
12.7
10.2
8.4
9.7
8.6
Br
Br
Br
Br
Br
Br
H
Me
Cl
H
Me
Cl
I
OMe
SEt
9
9.0
12.0
7.3
10
11
12
13
14
15
16
Me
OEt
OEt
OEt
OEt
OEt
OEt
7.8
7.1
7.7
7.6
7.1
11.2
11.1
8.5
11.4
6.1
a In kcal/mol.
TABLE 6. Computed (UB3LYP/6-31G*) Mulliken Charges (MC)
on the Nitrogen in Radicals A (X ) Br)
R )
CF3
OEt
Me
n-Bu
i-Pr
Ph
NMe2
MC -0.313 -0.329 -0.322 -0.333 -0.337 -0.359 -0.381
acetamidyl and carbamidyl radicals. To gain more insight into
the reactivity of N-(4-pentenyl)amidyl radicals, we turn to
density functional calculations for help, which have been
demonstrated to be a fairly accurate tool in the study of radical
reactions.13,14 The transition state structures and energies were
fully optimized at the UB3LYP/6-31G* level. Once convergence
is reached, the harmonic vibration frequencies were calculated
at this point to confirm the geometry obtained to be a true first-
order saddle point. The zero-point vibrational energy and thermal
corrections were also obtained at the UB3LYP/6-31G* level.
The computed activation free energies (∆Gq) for 5-exo and
6-endo cyclization are summarized in Table 5. The calculated
Mulliken charges on the nitrogen atom of amidyl radicals A
(X ) Br) are listed in Table 6 (see also the Supporting
Information for details).
The above calculated data are in good agreement with the
experimental results shown in Table 2 except for the case of
CF3. While theoretical calculations point out that trifluoroac-
etamidyl radicals are the most reactive, no expected product
could be isolated in the reaction of trifluoroacetamide 7a. A
possible explanation for this inconsistency is that, because of
the powerful electron-withdrawing nature of CF3, the amide
nitrogen is much less nucleophilic and the oxidative generation
of the corresponding N-I intermediate as the radical precursor
did not occur at all.
We next turn our attention to the control of regioselectivity.
For N-(4-pentenyl)acetamidyl radical A (R ) Me, X ) H), the
activation energy difference between 5-exo and 6-endo cycliza-
tion is computed to be 2.2 kcal/mol in favor of 5-exo cyclization
(entry 8, Table 5). With an internal methyl substitution (X )
Me), the energy difference is lowered to 0.6 kcal/mol (entry 9,
Table 5). However, switching the methyl to a chlorine atom
resulted in a difference of 2.3 kcal/mol in the activation energy,
First we focus on the influence of carbonylic substituents R
on the reactivity of amidyl radicals. The calculated activation
energies for 6-endo cyclization of radicals A (X ) Br) increases
(11) Concepcion, J. I.; Francisco, C. G.; Hernandez, R.; Salazar, J. A.; Suarez,
E. Tetrahedron Lett. 1984, 25, 1953.
(12) It should be noted that the ionic iodocyclization via a 5-exo manner
prevailed when the reactions of 7 with Pb(OAc)4/I2 or DIB/I2 were performed
in the dark. For example, the treatment of 7b with DIB/I2 in the dark at rt for
2 h afforded N-(5-iodo-4-oxopentyl)benzamide in 48% yield. See also ref 8a.
(13) For review articles on use of calculations in radical reactions, see: (a)
Eksterowicz, J. E.; Houk, K. N. Chem. ReV. 1993, 93, 2439. (b) Schiesser, C. H.;
Skidmore, M. A. In Radicals in Organic Synthesis; Renaud, P., Sibi, M. P.,
Eds.; Wiley-VCH: Weinheim, Germany, 2001; Vol. 1, p 337.
(14) (a) Chen, Q.; Li, C. Organometallics 2007, 26, 223. (b) Itoh, Y.; Houk,
K. N.; Mikami, K. J. Org. Chem. 2006, 71, 8918. (c) O’Neil, L. L.; Wiest, O.
J. Org. Chem. 2006, 71, 8926. (d) Hartung, J.; Daniel, K.; Rummey, C.;
Bringmann, G. Org. Biomol. Chem. 2006, 4, 4089. (e) Lin, H.; Chen, Q.; Cao,
L.; Yang, L.; Wu, Y.-D.; Li, C. J. Org. Chem. 2006, 71, 3328. (f) Liu, L.; Chen,
Q.; Wu, Y.-D.; Li, C. J. Org. Chem. 2005, 70, 1539. (g) Speybroeck, V. V.; De
Kimpe, N.; Waroquier, M. J. Org. Chem. 2005, 70, 3674. (h) References 4b
and 4d.
J. Org. Chem. Vol. 73, No. 16, 2008 6169

Yuan et al.
indicating the predominance of 6-endo cyclization (entry 10,
Table 5). The effect of Cl substitution is even more obvious in
the cases of carbamidyl radicals (entries 11-13, Table 5). An
activation energy difference of >3.0 kcal/mol is also computed
for carbamidyl radicals with other halogen atom substitution
(X ) Br, I). Compared to Cl-substitution, the vinylic OMe or
SEt substitution lowers the activation energies for 6-endo
cyclization (entries 15 and 16, Table 5). This could be attributed
to the enhanced electron density of the CdC bond with OMe
or SEt substitution. Note that the changes in activation energy
coincide well with the outcome of cyclization observed experi-
mentally (entries 6-8, Table 1). The above calculation data are
in good agreement with the experimental results. The only
exception is that, with OMe substitution, the calculated activa-
tion energy difference is only 1.4 kcal/mol while exclusive
6-endo cyclization was observed. A plausible explanation is that
thermodynamic factors might also contribute to the control of
regioselectivity in this case.15 For radical A (R ) OEt, X )
H), the 5-exo and 6-endo cyclization are exothermic by 13.2
and 16.2 kcal/mol, respectively. As a comparison, the corre-
sponding cyclizations of radical A (R ) OEt, X ) OMe) are
less exothermic (7.2 and 12.3 kcal/mol, respectively) and the
gap between 5-exo and 6-endo cyclization becomes larger, which
increase the extent of thermodynamic control on the regiose-
lectivity.
The calculations also help us to elucidate the unique role of
heteroatom substitution. While the activation energies for 6-endo
cyclization of carbamidyl radicals remain in the range of 7.0-7.7
kcal/mol, significant changes in the activation energy for 5-exo
cyclization are computed from methyl to halogen substitution
(entries 2 and 12-14, Table 5). This is unlikely to result from
the steric factors (see also structure 20). The difference between
the methyl moiety and a heteroatom is that the latter bears lone
pair electrons. Once a nitrogen-centered radical adds to a CdC
bond having an internal heteroatom substituent, it will face the
lone pair-lone pair electron repulsion between the nitrogen
radical and the heteroatom. As can be seen from the transition
state structures 19 and 20, the N-Br distance in 20 (5-exo) is
computed to be 3.069 Å, well within the range of close
interaction. On the other hand, the N-Br distance in 19 (6-
endo) is much longer (3.802 Å). Therefore, the lone pair electron
repulsion is much stronger in 5-exo cyclization than in 6-endo
cyclization. As a result, the activation energies for 5-exo
cyclization are significantly increased by the presence of the
heteroatom, leading to the overwhelming predominance of
6-endo cyclization. A similar discussion based on lone pair
electron repulsion was found to successfully explain not only
the regioselectivity of sulfonamidyl and aminyl radical cycliza-
tion, but also the chemoselectivity in the electrophilic halocy-
clization of unsaturated amides with vinylic halogen substitu-
tion,16 implying the ubiquity of lone pair electron repulsion in
the chemistry of vinyl halides toward heteroatom-centered
reactive species.
that carbamidyl radicals are the superior choice for efficient
cyclization. Moreover, the regioselectivity of cyclization can
be well controlled by vinylic heteroatom substitution. As a result,
highly regioselective 5-exo, 6-endo, 6-exo, and even 7-endo
cyclization could be implemented. This finding should be an
important application in the synthesis of cyclic amines.
Experimental Section
Typical Procedure for Bu3SnH-Initiated Amidyl Radical
Cyclization Reactions. Ethyl N-phenylthio-4-methylpent-4-enyl-
carbamate (1d, 279 mg, 1.0 mmol) was dissolved in 80 mL of
anhydrous benzene, and the solution was brought to reflux. The
mixture of Bu3SnH (0.4 mL, 1.5 mmol) and AIBN (49 mg, 0.3
mmol) in benzene (20 mL) was added over a period of 4 h with
the aid of a syringe pump under nitrogen atmosphere. The mixture
was refluxed for an additional 1 h and then cooled to room
temperature. After removal of the solvent under reduced pressure,
the residue was chromatographed on silica gel with hexane-ethyl
acetate (20:1, v:v) as the eluent to give piperidine 2d (92 mg, 54%
yield) and pyrrolidine 3d (34 mg, 20% yield). 2d: Colorless liquid.
1H NMR (300 MHz, CDCl3) δ 0.87-0.89 (3H, m), 1.03-1.10 (1H,
m), 1.23-1.28 (3H, m), 1.42-1.61 (3H, m), 1.77-1.80 (1H, m),
2.39 (1H, br), 2.67-2.76 (1H, m), 4.00-4.13 (4H, m). 13C NMR
(75.4 MHz, CDCl3) δ 14.6, 18.7, 24.7, 30.8, 32.8, 44.1, 51.1, 60.9,
155.4. EIMS: m/z (rel intensity) 171 (M+, 15), 156 (2), 142 (100),
126 (9), 116 (18), 98 (58), 69 (13), 56 (28). Anal. Calcd for
C9H17NO2: C, 63.13; H, 10.01; N, 8.18. Found: C, 63.01; H, 9.74;
N, 8.19. 3d: Colorless liquid. 1H NMR (300 MHz, CDCl3) δ
1.22-1.30 (3H, m), 1.40/1.34 (6H, s), 1.72-1.80 (4H, m),
3.40-3.51 (2H, m), 4.04-4.19 (2H, m). 13C NMR (75.4 MHz,
CDCl3) δ 14.4/14.3, 21.8/21.5, 25.7/26.7, 41.4/42.5, 47.3/48.2, 59.5/
60.0, 60.2/59.7, 153.7/155.0. EIMS: m/z (rel intensity) 171 (M+,
2), 156 (61), 128 (17), 112 (18), 84 (100), 70 (8), 55 (21), 47 (27).
Anal. Calcd for C9H17NO2: C, 63.13; H, 10.01; N, 8.18. Found: C,
62.82; H, 9.86; N, 8.24.
Typical Procedure for the Reactions of Amides with Pb-
(OAc)4/I2. To the solution of Pb(OAc)4 (155 mg, 0.35 mmol) in
dry dichloromethane (3 mL) was added iodine (64 mg, 0.25 mmol)
at rt under nitrogen atmosphere. The mixture was stirred at rt for
5 min. N-(4-Bromopent-4-enyl)acetamide (7c, 21 mg, 0.1 mmol)
was then added. The reaction mixture was irradiated at rt for 1 h
with the aid of a 125 W high-pressure mercury lamp. The light
was then turned off and aqueous Na2S2O3 (10 mL) was added. The
two layers were separated and the aqueous phase was extracted
with CH2Cl2. The combined organic layer was washed with aqueous
Na2CO3 and brine and then dried over anhydrous Na2SO4. After
removal of the solvent under reduced pressure, the crude product
was purified by column chromatography on silica gel with hexane/
ethyl acetate (20:1, v:v) as the eluent to give the product 8c as a
yellowish oil. Yield: 20 mg (60%). 1H NMR (300 MHz, CDCl3) δ
1.55-1.86 (2H, m), 2.18/2.14 (3H, s), 2.42-2.50 (1H, m),
2.64-2.73 (1H, m), 3.46-3.73 (2H, m), 3.86/4.04 (1H, 2d, J )
14.4 Hz), 4.06/4.25 (1H, 2d, J ) 14.4 Hz). 13C NMR (75.4 MHz,
CDCl3) δ 21.4, 25.3/26.1, 38.0/36.3, 40.9/45.8, 50.2/50.0, 65.4/
60.2, 169.2/169.0. EIMS: m/z (rel intensity) 332 (M+ + 1, 9), 252
(1), 206 (53), 204 (50), 164 (70), 162 (69), 83 (100), 43 (74). HRMS
calcd for C7H11OBrIN (M) 330.9069, found 330.9067.
Typical Procedure for the Reactions of Unsaturated Amides
with DIB/I2. To the solution of DIB (64 mg, 0.2 mmol) in dry
CH2Cl2 (3 mL) was added iodine (38 mg, 0.15 mmol) at rt under
nitrogen atmosphere. The mixture was stirred at rt for 5 min. Ethyl
4-iodopent-4-enylcarbamide (7g, 32 mg, 0.1 mmol) was added and
the resulting mixture was irradiated at rt for 2 h with the aid of a
125 W high-pressure mercury lamp. Aqueous Na2S2O3 (5 mL) was
then added. The two layers were separated and the aqueous phase
was extracted with CH2Cl2. The combined organic layer was washed
with aqueous Na2CO3 and brine, and then dried over anhydrous
Conclusion
The chemistry detailed above has provided a clear under-
standing on the behaviors of unsaturated amidyl radicals. The
substituents attached to the carbonyl (R in A) significantly alter
the reactivity of amidyl radicals toward intramolecular addition
to the CdC bond. More electrophilic amidyl radicals are
generally more prone to cyclization. Our experimental investiga-
tion in combination with theoretical analyses has pointed out
6170 J. Org. Chem. Vol. 73, No. 16, 2008

Cyclization Based on Lone Pair-Lone Pair Repulsion
Na2SO4. After removal of the solvent, the crude product was purified
by column chromatography on silica gel with hexane/ethyl acetate
(20:1, v:v) as the eluent to give the product 8g as a yellowish oil.
Yield: 27 mg (67%). 1H NMR (300 MHz, CDCl3) δ 1.23 (3H, t, J
) 7.0 Hz), 1.57-1.62 (2H, m), 2.41 (2H, t, J ) 5.6 Hz), 3.45 (2H,
t, J ) 5.1 Hz), 3.82 (2H, s), 4.12 (2H, q, J ) 6.9 Hz). 13C NMR
(75.4 MHz, CDCl3) δ 7.7/6.7, 14.7, 26.6, 43.0/43.3, 52.1, 61.8,
64.7/64.5, 154.7. ESI-MS: m/z (rel intensity) 410 (M+ + 1), 432
(M+ + Na). HRMS calcd for C8H14NO2I2 (M + H) 409.9108, found
409.9117.
by the Shanghai Municipal Committee of Science and Technol-
ogy (Grant No. 07XD14038).
Supporting Information Available: Characterizations of
compounds 1-18 and calculation results on A-C. This material
is available free of charge via the Internet at http://pubs.acs.org.
JO800845B
(15) For a similar discussion, see: Alabugin, I. V.; Manoharan, M. J. Am.
Chem. Soc. 2005, 127, 9534.
Acknowledgment. This project was supported by the Na-
tional NSF of China (Grant Nos. 20672136 and 20702060) and
(16) Hu, T.; Liu, K.; Shen, M.; Yuan, X.; Tang, Y.; Li, C. J. Org. Chem.
2007, 72, 8555.
J. Org. Chem. Vol. 73, No. 16, 2008 6171