Nonbonded, attractive cation-π interactions in azide-mediated asymmetric ring expansion reactions

Article abstract of DOI:10.1021/jo800222r

(Chemical Equation Presented) The influence of attractive, nonbonded interactions on the reactions of 1,2- and 1,3-hydroxyalkyl azides with ketones has been investigated through experimental and computational means. A series of 1,3-hydroxyalkyl azides bearing electronically tuned aromatic groups at the 2 position were prepared and reacted along with several derivatives designed to conformationally restrict the rotational orientation of the aromatic substituent. These studies showed that a cation-π interaction between an aryl moiety and an N₂⁺ leaving group plays a role in determining the stereoselectivity of these reactions. A series of ab initio calculations supported this hypothesis. A computational and experimental analysis suggested a primarily steric model for the analogous reactions of substituted 2-azido-1-ethanol analogues.

Full text of DOI:10.1021/jo800222r

Nonbonded, Attractive Cation-π Interactions in Azide-Mediated
Asymmetric Ring Expansion Reactions
Christopher E. Katz,†,‡ Timothy Ribelin,§ Donna Withrow,| Yashar Basseri,|
Anna K. Manukyan,| Amy Bermudez,| Christian G Nuera,| Victor W. Day,†
Douglas R. Powell, Jennifer L. Poutsma,| and Jeffrey Aubé*,§
Department of Chemistry, UniVersity of Kansas, 1251 Wescoe Hall DriVe, Room 2010, Malott Hall,
Lawrence, Kansas 66045-7582, Department of Medicinal Chemistry, UniVersity of Kansas, 1251 Wescoe
Hall DriVe, Room 4070, Malott Hall, Lawrence, Kansas 66045-7582, and Department of Chemistry and
Biochemistry, Old Dominion UniVersity, 4541 Hampton BouleVard, Norfolk, Virginia 23529
jaube@ku.edu; jradkiew@odu.edu
ReceiVed January 28, 2008
The influence of attractive, nonbonded interactions on the reactions of 1,2- and 1,3-hydroxyalkyl azides
with ketones has been investigated through experimental and computational means. A series of 1,3-
hydroxyalkyl azides bearing electronically tuned aromatic groups at the 2 position were prepared and
reacted along with several derivatives designed to conformationally restrict the rotational orientation of
the aromatic substituent. These studies showed that a cation-π interaction between an aryl moiety and
+
an N2 leaving group plays a role in determining the stereoselectivity of these reactions. A series of ab
initio calculations supported this hypothesis. A computational and experimental analysis suggested a
primarily steric model for the analogous reactions of substituted 2-azido-1-ethanol analogues.
Introduction
molecule’s conformation en route to product. The outcomes of
such reactions are usually predicted by analyzing steric interac-
tions present in intermediates or transition structures leading to
the observed products. Such analyses may be carried out
computationally, but qualitative rationalizations based on com-
mon principles of conformational analysis are common.1,2
The covalent connection of two reactive moieties to facilitate
an intramolecular reaction is a useful strategy in chemical
synthesis.1 In this context, a connecting tether is primarily used
to increase the reaction rate by enforcing proximity between
the functional groups. In addition, a tether can affect stereo-
chemistry by the incorporation of substituents that influence the
A tether may become part of the ultimate product or it may
be temporarily deployed and then detached following the
reaction of interest.3 A third method being explored in our
* To whom correspondence should be addressed. Professor Jeffrey Aubé,
Department of Medicinal Chemistry, 1251 Wescoe Hall Drive, Room 4040b,
Malott Hall, University of Kansas, Lawrence, Kansas 66045-2506. Tel
785.864.4496; Fax 785.864.5326.
(1) The literature of intramolecularity in organic synthesis is vast. Some relevant
reviews include: (a) Oppolzer, W. Synthesis 1978, 793–802. (b) Ciganek, E. Org.
React. 1984, 32, 4–374. (c) Speckamp, W. N.; Hiemstra, H. Tetrahedron 1985, 41,
4367–4416. (d) Schinzer, D. Synthesis 1988, 263–273. (e) RajanBabu, T. V. Acc.
Chem. Res. 1991, 24, 139–145. (f) Namboothiri, I. N. N.; Hassner, A. Top. Curr.
Chem. 2001, 216, 1–49. (g) Maryanoff, B. E.; Zhang; Han-Cheng; Cohen, J. H.;
Turchi, I. J.; Maryanoff, C. A. Chem. ReV. 2004, 104, 1431–1628. (h) Zubkov,
F. I.; Nikitina, E. V.; Varlamov, A. V. Russ. Chem. ReV. 2005, 74, 639–669.
(2) Sugimura, T. Eur. J. Org. Chem. 2004, 1185–1192.
† Department of Chemistry, University of Kansas.
‡ Current Address: Tetraphase Pharmaceuticals, 480 Arsenal St., Ste 110,
Watertown, MA 02472-2805.
§ Department of Medicinal Chemistry, University of Kansas.
| Old Dominion University.
Current address: University of Oklahoma Department of Chemistry and
Biochemistry 620 Parrington Oval, Room 208 Norman, OK 73019-305.
3318 J. Org. Chem. 2008, 73, 3318–3327
10.1021/jo800222r CCC: $40.75 2008 American Chemical Society
Published on Web 04/09/2008

Nonbonded, AttractiVe Cation-π Interactions
SCHEME 1
FIGURE 1. Two possible intermediates in the ring-expansion reaction
of hydroxyalkyl azides, one in which steric interactions are minimized
by placing the phenyl group in an equatorial position (left) and one
with the potential for an attractive cation-π interaction between the
leaving group and a phenyl group (right).
diaxial relationship to one another (Figure 1). Preliminary results
that supported this view were communicated in 2003, including
experiments that showed the dependence of stereoselectivity on
the electronic nature of the phenyl group and conformational
constraints.9 In this account, the full details of this work are
disclosed along with a detailed theoretical treatment of the
substituent effects on the 1,3 cation-phenyl interaction. In
addition, the possible role of these effects in the analogous
reactions of 1,2-hydroxyalkyl azides will be discussed.
laboratory involves connection of the tether under the same
conditions used to effect the reaction of interest.4 We have
examined this “in situ tethering” strategy in the context of
hydroxyalkyl azide-mediated ring expansions of cyclic ketones.5,6
The use of chiral hydroxyalkyl azides in this reaction provides
an efficient asymmetric route to lactams from appropriately
substituted cyclohexanones (Scheme 1).7 In this case, the key
stereochemistry-determining feature is the conformation of the
azide-containing tether: the chairlike ring containing an equato-
rial phenyl substituent leads selectively to a spirocyclic inter-
mediate that rearranges to afford an iminium ether. This latter
intermediate is converted into product by treatment with aqueous
base or by reaction with other nucleophiles.8
Results and Discussion
Initial Observations. Our current view of the mechanism
of the in situ tethered hydroxyalkyl azide Schmidt reaction,
indicated in Scheme 2, is supported by considerable experi-
mental evidence5,9 and computational studies.10 We propose that
the stereoselectivity of the sequence is controlled by the
equilibrium between intermediates A and B, each of which
features equatorial addition of azide to the oxonium ion derived
from the ketone substrate (i.e., formation of a spirocycle having
nitrogen trans to the 4-tert-butyl group) but that differ in which
chairlike conformation of the intermediate is involved. Each
intermediate undergoes rearrangement to an iminium ether
through migration of a methylene group antiperiplanar to the
In exploring the scope of this process, it soon became clear
that results using certain phenyl-containing hydroxyalkyl azides
could not be readily explained solely by analysis of steric
interactions. Specifically, we proposed that the stereoselectivity
of some of the reactions could be affected by attractive,
nonbonded interactions of a substituted aromatic group with the
+
leaving N2 group. Two aspects of this step are noteworthy.
First, the system is set up in such a way that the only reac-
tive conformations have a 1,2-diaxial relationship between the
migrating carbon and leaving group (equatorial N2+would be
antiperiplanar to a nonmigratable C-O bond). Second, we propose,
based on theory10 and on the absence of reasonable alternatives,
that the rates of migration from A or B are comparable, leaving
the sole source of stereoselectivity to be the relative stability
of these intermediates.11 A and B can equilibrate by confor-
mational interconversion or by reversion to an oxonium ion
followed by reformation; which one occurs is not relevant to
this analysis, provided that interconversion is fast relative to
migration. In any event, the stereochemistry-determining step
is completed by formation of the iminium ethers, which are
hydrolyzed to afford lactams.
+
cationic N2 leaving group when both moieties were in a 1,3-
(3) Reviews of disposable tethers in synthetic organic synthesis: (a) Ford,
W. T. CHEMTECH 1987, 17, 246–250. (b) Fensterbank, L.; Malacria, M.; McN.
Sieburth, S. Synthesis 1997, 813–854. (c) Gauthier, D. R., Jr.; Zandi, K. S.; Shea,
K. J. Tetrahedron 1998, 54, 2289–2338. (d) Knapp, S. Chem. Soc. ReV. 1999, 28,
61–72. (e) Cox, L. R.; Ley, S. V. In Templated Organic Synthesis; Diederich, F.,
Stang, P. J., Eds.; Wiley-VCH: Weinheim, 2000; Vol. 1 pp 275–395. (f) Breslow,
R. In Templated Organic Synthesis; Diederich, F., Stang, P. J.; Eds., Wiley-VCH:
Weinheim, 2000; Vol. 1 pp 159–188. (g) Zachary, A. D.; Overman Larry, E. Chem.
Commun. 2004, 253, 65. (h) Wallace, D. J. Angew. Chem., Int. Ed. 2005, 44, 1912–
1915.
(4) For other examples of “in situ” tethering in synthesis, see: (a) Sammakia,
T.; Berliner, M. A. J. Org. Chem. 1995, 60, 6652–6653. (b) Haynes, R. K.;
King, G. R.; Vonwiller, S. C. J. Org. Chem. 1994, 59, 4743–4748. (c) Greico,
P. A.; Kaufman, M. D.; Daeuble, J. F.; Saito, N. J. Am. Chem. Soc. 1996, 118,
2095–2096. (d) Suginome, M.; Iwanami, T.; Ito, Y. J. Org. Chem. 1998, 63,
6096–6097. (e) Cloninger, M. J.; Overman, L. E. J. Am. Chem. Soc. 1999, 121,
1092–1093.
Initial observations that suggested something might be
interesting about intermediates bearing a phenyl group placed
(5) (a) Gracias, V.; Milligan, G. L.; Aubé, J. J. Am. Chem. Soc. 1995, 117,
8047–8048. (b) Gracias, V.; Frank, K. E.; Milligan, G. L.; Aubé, J. Tetrahedron
1997, 53, 16241–16252.
(9) Katz, C. E.; Aubé, J. J. Am. Chem. Soc. 2003, 125, 13948–13949.
(10) Hewlett, N. D.; Aubé, J.; Radkiewicz-Poutsma, J. L. J. Org. Chem. 2004,
69, 3439–3446.
(6) See also Haynes, R. K.; King, G. R.; Vonwiller, S. C. J. Org. Chem.
1994, 59, 4743–4748.
(11) Hoye, T. R.; Ryba, T. D. J. Am. Chem. Soc. 2005, 127, 8256–8257.
(12) An outstanding collection of A values for substituted cyclohexanes can
be found in: (a) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of
Organic Compounds; John Wiley & Sons: New York, 1994, pp 696–697; see
also the numerous references cited. For determination of equatorial/axial
preferences in substituted 1,3-dioxanes, see: (b) Eliel, E. L.; Knoeber, M. C.
J. Am. Chem. Soc. 1968, 90, 3444–3458. (c) Eliel, E. L. Acc. Chem. Res. 1970,
3, 1–8. (d) Shea, K. J.; Dougherty, T. K. J. Org. Chem. 1985, 50, 4439–4442.
(7) For asymmetric versions of the hydroxyalkyl azide-mediated Schmidt
reaction, see ref 5a and: (a) Vidari, G.; Tripolini, M.; Novella, P.; Allegraucci,
P.; Garlaschelli, L. Tetrahedron: Asymmetry 1997, 8, 2893–2903. (b) Furness,
K.; Aubé, J. Org. Lett. 1999, 1, 495–497. (c) Sahasrabudhe, K.; Gracias, V.;
Furness, K.; Smith, B. T.; Katz, C. E.; Reddy, D. S.; Aubé, J. J. Am. Chem.
Soc. 2003, 125, 7914–7922.
(8) Gracias, V.; Milligan, G. L.; Aubé, J. J. Org. Chem. 1996, 61, 10–11.
J. Org. Chem. Vol. 73, No. 9, 2008 3319

Katz et al.
SCHEME 2
FIGURE 2. 2-Aryl-1-hydroxyalkyl-3-azides used in this work.
electron-rich phenyl group and the N2+ leaving group. Cation-π
interactions are commonly invoked in bioorganic chemistry as
features of protein structure and have also been cited as
components of host–guest interactions.13 In contrast, such
interactions have only rarely been invoked as stereocontrolling
elements in selective organic reactions.14,15 We decided to probe
this effect by examining related substrates designed to vary both
the electronic and conformational aspects of the reaction.
Substituted Aryl-Containing 1,3-Hydroxyalkyl Azides. A
series of 2-aryl-1-hydroxyalkyl-3-azides were initially prepared
following the hypothesis that lactam stereoisomer b would be
predominant with more electron-rich systems (Figure 2; see
Cubero16 for a theoretical treatment of the effect of substituents
and polarizability on cation-π interactions). Most of the
reactants were prepared and used in racemic form, although
downstream problems in assessing the stereochemical outcome
of several examples necessitated some asymmetric syntheses.
Throughout, compounds will be depicted in a single enantio-
meric series to simplify comparison of various examples.
Synthetic details and particulars of which stereochemical series
were used are provided in the Supporting Information.
With a series of electronically tuned hydroxyalkyl azides in
hand, the compounds were reacted with 4-tert-butylcyclohe-
xanone in the presence of BF3 ·OEt2, followed by hydrolysis
of the intermediate iminium ethers using aqueous KOH.
Selectivities were determined by HPLC or gas chromatographic
analysis of the crude reaction mixtures (Table 1). Subsequently,
the product lactams were purified by chromatography to afford
the yields shown; these reactions were generally high yielding.
Product stereostructures were determined by X-ray crystal-
lography and retention times of crystalline samples were
compared to those taken from crude reaction mixtures.
in a 1,3-diaxial relationship to the leaving group are summarized
in Scheme 2.7c Thus, the selectivity observed upon modifying
the nature of a substituent placed at carbon 2 in a series of 1,3-
hydroxyalkyl azides did not follow the steric demands of the
substituents, with selectivity increasing in the order Ph (A
value12a 2.8) < Me (1.74) < i-Pr (2.21). Furthermore, the low
selectivity seen for the 2-phenyl reactant was unique to this
series, with high selectivities observed for 1,3-hydroxyalkyl
azides bearing a phenyl substituent in either the 1 position (95:5
ratio) or in the 3 position (90:10 ratio; see Scheme 1). On one
hand, lower selectivity compared to 1- or 3-substituted azides
was not surprising in that the axial R group in intermediate B
only experiences a single 1,3-diaxial interaction with a relatively
(13) For literature of cation-π interactions in macromolecular structure and
molecular recognition, see:(a) Dougherty, D. A. Science 1996, 271, 163168(b)
Ma, J. C.; Dougherty, D. A. Chem. ReV. 1997, 97, 1303–1324. (d) Gokel, G. W.;
De Wall, S. L.; Meadows, E. S. Eur. J. Org. Chem. 2000, 2967–2978. (e) Shi,
Z.; Olson, C. A.; Bell, A. J., Jr.; Kallenbach, N. R. Biopolymers 2001, 60, 366–
380. (f) Gokel, G. W.; Barbour, L. J.; Ferdani, R.; Hu, J. Acc. Chem. Res. 2002,
35, 878–886. (g) Zacharias, N.; Dougherty, D. A. Trends Pharmacol. Sci. 2002,
23, 281–287.
+
(14) Cation-π interactions in organic synthesis have been recently reviewed: (a)
Yamada, S. Org. Biomol. Chem 2007, 5, 2903–2912. For lead references, see:
(b) Lakshminarasimhan, P.; Sunoj, R. B.; Chandrasekhar, J.; Ramamurthy, V.
J. Am. Chem. Soc. 2000, 122, 4815–4816. (c) Stratakis, M.; Froudakis, G. Org.
Lett. 2000, 2, 1369–1372. (d) Yamada, S.; Morita, C. J. Am. Chem. Soc. 2002,
124, 8184–8185. (e) Yamada, S.; Saitoh, M.; Misono, T. Tetrahedron Lett. 2002,
43, 5853–5857. (f) Ramamurthy, V.; Shailaja, J.; Kaanumalle, L. S.; Sunoj, R. B.;
Chandrasekhar, J. Chem. Commun. 2003, 1987–1999. (g) Chiavarino, B.;
Crestoni, M. E.; Fornarini, S. Int. J. Mass Spectrom. 2004, 235, 145–154. (h)
Lei, Y.; Aubé, J. J. Am. Chem. Soc. 2007, 129, 2766–2767.
small N2 group. C-5 substituents in 1,3-dioxanes (the closest
model for which detailed conformational benchmarks are readily
available) have lower equatorial/axial biases than in cyclohex-
anes but they generally follow the same order. Thus, the
-∆G°25 values for Me, i-Pr, and Ph in 5-alkyl-2-tert-butyl-1,3-
dioxanes are 0.80, 0.98, and 1.03 kcal/mol, respectively.12b–d
+
1,3-Diaxial interactions between the R group and the N2
leaving group clearly play some role given the higher selectivity
of R ) i-Pr v. Me en route to 2 and 1, respectively. The fact
that the C-2 phenyl group was an outlier - giving much lower
selectivity than even the smaller methyl-containing analog -
suggested an electronic component. In particular, the transition
structure leading to the minor product series b in Scheme 2
might be stabilized by nonbonded interactions between the
(15) For literature in which cation-π interactions play a role in asymmetric
catalysis, see: (a) Birman, V. B.; Uffman, E. W.; Jiang, H.; Li, X.; Kilbane,
C. J. J. Am. Chem. Soc. 2004, 126, 12226–12227. (b) Birman, V. B.; Jiang, H.
Org. Lett. 2005, 7, 3445–3447. (c) Birman, V. B.; Jiang, H.; Li, X.; Guo, L.;
Uffman, E. W. J. Am. Chem. Soc. 2006, 128, 6536–6537. (d) Birman, V. B.; Li,
X. Org. Lett. 2006, 8, 1351–1354. (e) Birman, V. B.; Guo, L. Org. Lett. 2006,
8, 4859–4861.
(16) Cubero, E.; Luque, F. J.; Orozco, M. Proc. Natl. Acad. Sci. U. S. A.
1998, 95, 5976–5980.
3320 J. Org. Chem. Vol. 73, No. 9, 2008

Nonbonded, AttractiVe Cation-π Interactions
TABLE 1. Reactions of 2-Aryl-1,3-hydroxyalkyl Azide with 4-tert-Butylcyclohexanone
diastereomeric ratioa
hydroxyalkyl
yield
(%)
entry
azide
X
Y
lactam
a
b
1
2
3
4
5
6
4
5
6
7
8
9
OMe
OMe
H
Br
F
OMe
H
H
H
H
13
14
3
15
16
17
99
99
85
90
73
86
43
47
64
68
69
76
57
53
36
32
31
24
NO2
H
a See Supporting Information for methods used to determine structures and ratios.
FIGURE 3. Hammett plot of the stereoselectivities of the reactions from Table 1 versus σ+ parameters. The results from hydroxyalkyl azide 4 are
not included due to the unavailability of an appropriate σ+ value.
The data indicate a correlation between the electronic nature
of the 2-aryl substituent and the stereoselectivity of the ring-
expansion reaction, consistent with greater stabilization of the
intermediate leading to isomer b via nonbonded cation-π
interactions. A Hammett-like analysis of the data is shown in
Figure 3. Although its application to nonbonded interactions
constitutes a stretch of the linear free energy concept,17 the
reasonable correlation (r ) 0.91) obtained between σ+ and
the -log(a/b) values supports the hypothesis that increased
electron density in the aryl substituent assists in the formation
of isomer b.
approach was applied to the intermediates associated with the
reactions of hydroxyalkyl azides bearing phenyl substitution.
Energies were calculated using the MP2/6–311+G**//MP2/
6–31G* level of theory and focused on the N-diazonium-1,3-
oxazinane part of the intermediate C (Table 2). There are four
possible versions of the intermediate: Ca (N2+ axial/R equato-
rial), Cb (N2+ axial/R axial), Cc (N2+ equatorial/R equatorial),
+
+
and Cd (N2 equatorial/R axial). In all cases, the N2 axial
+
conformers were lower in energy than the N2 equatorial
conformers. Thus, the energy differences between Ca and Cb
were used to calculate the axial/equatorial preferences of the
substituted phenyl groups. Energies were converted to ∆G
values using standard methods and solvent was added via the
CPCM method. For comparison between the experimental data
and our calculations, the experimental ratios were converted to
∆G values using ∆G ) -RT ln K.
The calculations show a surprising degree of correlation with
the experimental results considering the small differences in the
∆G ’s between intermediates containing differently substituted
phenyl rings. Both gas phase and solvent axial/equatorial free
energy differences are provided in Table 2. Previously, the gas
phase calculations overestimated the stability of the axial/axial
conformer Cb,10 and this was again observed for the substituted
To gain greater understanding of this phenomenon, a series
of ab initio calculations was performed. In previous work,
intermediates a and b were calculated along with the transition
structures leading to iminium ether products.10 These calcula-
tions supported our overall view of the reaction mechanism and
additionally provided preliminary evidence in favor of the
+
nonbonded stabilization of the N2 leaving group by an
appropriately positioned phenyl group. In the present work, this
(17) For recent examples of linear free energy treatments of nonbonded
interactions, see: (a) Cockroft, S. L.; Hunter, C. A.; Lawson, K. R.; Perkins, J.;
Urch, C. J. J. Am. Chem. Soc. 2005, 127, 8594–8595. (b) Thomas, K. M.;
Naduthambi, D.; Zondlo, N. J. J. Am. Chem. Soc. 2006, 128, 2216–2217. (c)
Hodges, J. A.; Raines, R. T. Org. Lett. 2006, 8, 4695–4697.
J. Org. Chem. Vol. 73, No. 9, 2008 3321

Katz et al.
TABLE 2. Comparison of Calculated Free Energy Differences and Experimental Diastereoselectivitiesa
c
c
c
entry
model system
lactam
a:bb
∆Gexpt
∆Ggas
∆Gsolv
-0.5
-0.2
-0.3
-0.2
0.2
1
2
3
4
5
6
C, R ) (MeO)3C6H2
C, R ) 4-MeOC6H4
C, R ) C6H5
C, R ) 4-BrC6H4
C, R ) 4-NO2C6H4
D
13
14
3
15
16
18
43:57
47:53
64:36
68:32
76:24
5:95
-0.1
0.0
0.2
0.3
0.4
-3.5
-2.8
-2.0
-2.2
-1.8
-5.8
-1.1
-4.3 (-1.4)d
a Calculations for the solvent (CPCM, ꢀ ) 8.93) free energy differences were performed at the MP2/6–311+G**//MP2/6–31G* level of theory.
b Experimental ratios. c Differences between Cb/Ca or Db/Da. d The number in parentheses is the selectivity for equatorial versus axial attack of the
azide on the carbonyl carbon.
TABLE 3. Solvent Study with Hydroxypropyl Azide 5 and
examples. Despite this, the experimental trend was reproduced
4-tert-Butylcyclohexanone
by the theoretical data (R2 ) 0.89), aside from a slight difference
in the relationship between the unsubstituted phenyl and
bromophenyl. This is consistent with the view that the selectivi-
ties are a direct result of the axial/equatorial preference and the
strength of the cation-π interaction. Inclusion of solvent gives
free energies that are in very good agreement with the
experimental selectivities with errors of only ( 0.5 kcal/mol.
Although the CPCM method correctly reproduces the shielding
diastereomeric ratioa
effect of the solvent, the overall accuracy of these types of
entry
solvent
ꢀrb
na
2.379
4.266
7.23
yield (%)
a
b
calculations is reduced compared to gas phase. Thus, the trend
present in the experimental data and gas phase results is not as
obvious for the solvent results (R2 ) 0.65).
c
1
2
3
4
5
n-C5H12/CH2Cl2
toluene
(C5H5)2O
diglyme
78
97
98
99
99
47
47
54
72
47
53
53
47
28
53
A brief survey of solvent effects was also carried out
experimentally (Table 3). With one exception (entry 1), the
effect was relatively flat across the small number of solvents
that were examined (a greater range was not possible as the
reaction fails in highly polar solvents).5,9 In particular, neither
a dependence on dielectric constant nor a strong effect of
carrying out the reaction in an aromatic solvent could be
gleaned from these data. This suggests that the equilibrium
between intermediates A and B (Scheme 1), differing mainly
in the axial v. equatorial orientation of the aromatic group,
is not generally perturbed in the solvents examined. These
data are in agreement with previous theoretical studies that
demonstrated that going from gas phase to solvent has a large
effect on cation-π interactions, but that a change in the
dielectric of the solvent has a much smaller effect.18 It is
possible that diglyme, which is the prime outlier in this series,
is able to complex the N2+ group and thereby lead to product
that arises from a greater flux through intermediate type A
containing an equatorial phenyl group. It is also possible that
solvent effects are minimized due to the closeness of the aryl
CH2Cl2
8.83
a Determined by HPLC. b Reichhardt, C. SolVents and SolVent Effects
in Organic Chemistry, 2nd ed.; VCH: Weinhein, 1988. c A 5/2 ratio of
n-C5H12/CH2Cl2, in agreement with the experimental data (Table 4).
and cationic moieties. (We thank anonymous reviewers for
these last two suggestions.)
Conformational Effects in the Reactions of 2-Aryl-
1-Hydroxypropyl-3-Azides Containing Quaternary Stereo-
centers. It has been established that 1-methyl-1-phenylcyclo-
hexane exists as a ca. 2.5:1 mixture of conformers at -100 °C,
favoring the isomer with an axial phenyl group and correspond-
ing to an energy difference of ca. 0.32 kcal/mol at -78 °C.19a
The conformation with axial phenyl is favored because it can
avoid, through rotation, steric interactions with adjacent equato-
rial or methyl group hydrogens that are unavoidable when the
phenyl is equatorial. The situation is complicated when gem-
aryl,Me disubstitution is studied at C-5 in various 1,3 dioxanes,
where conformers favoring either equatorial or axial aromatic
rings have been observed, depending on electron donating or
withdrawing C-5 aryl substitution (an effect that is absent in
1-aryl-1-methylcyclohexanes).19b Interestingly, although tethers
containing quaternary nonstereogenic carbons are well-known
to accelerate intramolecular reactions,20 examples of stereogenic,
(18) (a) Gallivan, J. P.; Dougherty, D. A. J. Am. Chem. Soc. 2000, 122, 870–
874. (b) Biot, C.; Buisine, E.; Rooman, M. J. Am. Chem. Soc. 2003, 125, 13988–
13994.
(19) (a) Eliel, E. L.; Manoharan, M. J. Org. Chem. 1981, 46, 1959–1962.
For an excellent discussion of nonadditivity of cyclohexane-centric conforma-
tional analysis, see: (b) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry
of Organic Compounds; John Wiley & Sons: New York, 1994; pp 695–706. (c)
Cook, M. J.; Nasri, K.; Vather, S. M. Tetrahedron Lett. 1986, 27, 3853–3854.
(20) Jung, M. E.; Piizzi, G. Chem. ReV. 2005, 105, 1735–1766.
3322 J. Org. Chem. Vol. 73, No. 9, 2008

Nonbonded, AttractiVe Cation-π Interactions
SCHEME 3
FIGURE 4. Conformational expectations for Calkyl-Caryl bond rotamers
in intermediates derived from (a) 2-phenyl-1-hydroxypropyl-3-azide
and (b) 2-methyl-2-phenyl-1-hydroxypropyl-3-azide. (c) Conformational
constraint that locks the phenyl group into a bisecting conformation.
experimental and theoretical predictions are in excellent agree-
ment. We also examined the reaction of the fluoro-substituted
hydroalkyl azide 11; lactams 19a,b were obtained in a compa-
rable ratio to that for 18a,b (Scheme 3a).
These results were also examined computationally (Table 2,
entry 6; Figures 4 and 5). The most efficient cation-π
interaction between a substituted phenyl group and a cationic
species occurs when the flat face of the aromatic moiety is
directed toward the positive charge.22 In 2-monosubstituted
cases, the aromatic groups are relatively mobile thanks to the
presence of an oxygen atom in a 1,3 relationship to the
stereogenic center (Figure 4a). Thus, calculations show that
rotation about the C-Cphenyl bond in 3-axial-phenyltetrahydro-
pyran only costs ca. 1.4 kcal/mol over a range of 90°. In this
case, one expects the phenyl group to occupy a range of
conformations about the Calkyl-Caryl bond. Of course, this
ensemble is also in equilibrium with a set of conformations in
which both rings have flipped and the phenyl group occupies
an equatorial position; reactions through these conformations
lead to isomeric lactams a (an example of these alternative
conformations is given in Scheme 2). The situation changes
when a methyl group is also attached to the carbon bearing the
phenyl group (Figure 4b). As in 1-methyl-1-phenylcyclo-
hexane,19a the phenyl group is unable to bisect the heterocyclic
ring due to unfavorable steric interactions with the methyl group
(i.e., the conformation shown to the right of the arrow in Figure
4b). This analysis was supported by ab initio calculations carried
on the simplified system shown in Figure 5 (see Supporting
Information for detailed results).
quaternary carbons appearing as a stereocontrolling feature of
an intramolecular process are rare.21
Strikingly, the reaction of the geminally disubstituted hy-
droxyalkyl azide 10 was extremely selective, affording only the
single isomer 18b as determined through inspection of the crude
NMR spectrum (Scheme 3a; the stereostructure of 18b was
secured by X-ray crystallographic analysis). Conservatively
estimating the ratio at g95:5, this highly selective example
suggests that either the eq-Me/ax-Ph conformation is highly
favored in the reactive intermediate or that this conformation
is considerably more reactive than the alternative bearing an
axial Me group and equatorial phenyl (not shown). It is likely
that the eq-Me/ax-Ph preference arises from both steric and
electronic interactions. In contrast, for the systems examined
in Table 1, these two considerations work against one another.
Calculations support the greater stability of this intermediate
with the free energy preference for the axial/axial conformer
predicted to be -5.8 kcal/mol in the gas phase and -4.3 kcal/
mol in solvent. This is in contrast to the substituted aryl cases,
which showed very small to no preferences in solvent. Convert-
ing the experimental ratio to a ∆G gives a value of -1.1 kcal/
mol. This is significantly different from the calculated value of
-4.1 kcal/mol. However, the observation of stereoselectivity
also depends on equatorial attack of the azide to form the
heterosubstituted cyclohexane intermediate. Our previous cal-
culations have shown that the energy difference between
equatorial and axial attack is -1.4 kcal/mol, giving a selectivity
of 98:2. Thus, even if a large preference for the axial or
equatorial conformation of the intermediate exists, the overall
selectivity can never be greater than 98:2. Therefore, the
We wished to confirm the role of the Calkyl-Caryl bond rotation
in this reaction by synthesizing a conformationally constrained
indane-containing hydroxyalkyl azide 12. In this example, the
phenyl group is forced to bisect the plane of the perhydroöxazine
ring by connecting the geminal methyl and phenyl groups with
a methylene bridge (see Scheme 3b). Upon submitting this
(21) (a) Gunther, H. J.; Jager, V.; Skell, P. S. Tetrahedron Lett. 1977, 18,
2538–2542. (b) Trost, B. M.; Lautens, M.; Chan, C.; Jebaratnam, D. J.; Mueller,
T. J. Am. Chem. Soc. 1991, 113, 636–644.
(22) Rashkin, M. J.; Hughes, R. M.; Calloway, N. T.; Waters, M. L. J. Am.
Chem. Soc. 2004, 126, 13320–13325.
J. Org. Chem. Vol. 73, No. 9, 2008 3323

Katz et al.
SCHEME 4
(starred in Scheme 4a) and the phenyl substituent could be
responsible for this surprising result.
We considered the possibility that cation-π interactions
might also be relevant in these reactions. For example, it is
+
possible that the intermediate containing vicinal cis N2 and
phenyl groups is stabilized over the alternative trans isomer and
thereby leads to the observed product. Here, pathway divergence
only requires nitrogen inversion as opposed to the change of
ring conformations needed in the six-membered-ring containing
cases discussed above. Thus, nitrogen inversion is sufficient to
change which migratable bond is anti to the leaving group and
thus which isomer is formed. In this model, the lack of
stereoselectivity for the series in which the phenyl group is
adjacent to oxygen could be due to the lack of preference for
FIGURE 5. Possible conformers of the 3-methyl, 3-phenyl substituted
intermediate D. The structures on the left are MP2/6–31G* optimized
geometries. In the structures on the right, the phenyl ring has been
rotated by 90°; in these structures, there are greater steric interactions
with the adjacent methylenes on the six-membered ring (for Da) or
the methyl group (for Db). In the space-filled pictures, some hydrogens
have been deleted for clarity. Note that for Da, destabilizing steric
interactions exist regardless of the orientation of the phenyl ring.
+
either N2 epimer (Scheme 4b).
Ab initio calculations were first carried out for ground-state
model systems E and F in which the two migrating carbons of
the six-membered ring are replaced by methyl groups (Figure
6, Table 4). Since the five-membered ring can exist as a half-
chair or envelope, extensive searches were performed to ensure
that the lowest energy conformers were located for both cis and
trans isomers. For the 1-phenyl compound, E, a significant gas-
phase preference of 1.3 kcal/mol for the cis conformer was
calculated, while only a small preference of 0.4 kcal/mol was
observed for the 2-phenyl compound, F. In the cis epimer of
Ea, the N2+ group and the phenyl ring are aligned for a potential
cation-π interaction, whereas a similar alignment is impossible
in Eb. In F, a weaker cation-π interaction due to the increased
distance between the N2+ group and the phenyl ring is possible
analogue to typical reaction conditions, a greatly diminished
selectivity of ca. 3:1 was achieved. Furthermore, the major
isomer obtained in this way resulted from an intermediate
bearing an equatorial aryl group. The contrast between the
behavior of 10 and 11 vs 12 clearly shows that aryl group
rotation plays an important role in these reactions.
Reactions of Azidoethanol Derivatives. In the early devel-
opment of this asymmetric Schmidt reaction, we had examined
the reactions of 4-tert-butylcyclohexanone with two-carbon
hydroxyalkyl azides bearing phenyl substitution (Scheme 4).5a,7c
Thus, good selectivity was obtained for 2-azido-2-phenylethanol
21 (Scheme 4a), but the reaction of the isomer bearing a phenyl
group next to the alcohol 23 (Scheme 4b) was essentially
stereorandom. It was determined from analogy7c and calcula-
tion10 that azide addition occurred from an equatorial direction
in this system and that antiperiplanar migration relative to the
TABLE 4. Comparison of the gas-phase and solvated free energy
+
+
differences between the R/N2 cis (a) and R/N2 trans (b) models to
the experimental diastereoselectivities (kcal/mol)a
model
+
departing N2 group occurred here as well as in the azidopro-
entry
system
lactam
a:b
∆G
∆G
∆G
solv
expt
gas
panol-derived examples. Accordingly, analysis of the observed
stereoselectivity of the reaction suggested intermediacy of a
1
2
E
F
22
24
85:15
56:44
0.7
0.1
1.3
0.4
0.6
0.2
+
structure in which the N2 group and the adjacent phenyl
a Calculations for the solvent (CPCM,
differences were performed at the MP2/6–311+G**//MP2/6–31G* level
of theory.
ꢀ ) 8.93) free energy
substituent were cis to one another and the bond antiperiplanar
to the N2+ group migrates. We proposed that the minimization
of steric interactions between the migrating methylene group
3324 J. Org. Chem. Vol. 73, No. 9, 2008

Nonbonded, AttractiVe Cation-π Interactions
FIGURE 6. MP2/6–31G* optimized geometries and MP2/6–311+G**//MP2/6–31G* solvated energy differences of the cis (a) and trans (b) configured
model systems E and F. The green atom indicates the carbon that is aligned for migration.
TABLE 5. Reactions of 2-Substituted-2-azido-1-ethanols with 4-tert-Butylcyclohexanone
diastereomeric ratioa
entry
hydroxyalkyl azide
R1
R2
lactam
yield (%)
a
b
1
25
21
27
29
31
3,4,5-(OMe)3C6H2-
H
H
H
H
26
22
28
30
32
96
89
97
85
91
86
85
66
75
56
14
15
34
25
44
2b
3
C6H5-
4-(NO2)-C6H4-
C6H11
H
4
5
C6H11
a See Supporting Information for methods used to determine structures and ratios. b This literature example7c was repeated for the present work to
allow direct comparison with other cases.
in the cis conformer. However, this conformer is not the lowest
energy structure due to destabilizing steric interactions between
+
the methyl groups and the N2 group. With the addition of
solvent, a small preference is still observed for Ea over Eb (0.6
kcal/mol), but essentially no preference is seen for Fa over Fb
(0.2 kcal/mol). The solvent numbers are in agreement with the
experimental data.
FIGURE 7. MP2/6–31G* optimized geometries and MP2/6–311+G**//
The effect of substitution on the stereoselectivity of reactions
of 2-aryl-2-azidoethanol derivatives with 4-tert-butylcyclohe-
xanone was examined experimentally (Table 5). In this case,
additional electron donation gave a ratio identical to that of the
unsubstituted example (cf. entries 1 and 2, both about 5.6:1).
In contrast, the electron-poor example shown in entry 3 resulted
in a lower degree of selectivity favoring 28a (ca. 1.9:1).
However, the same direction of selectivity was obtained when
the aryl group was replaced by a cyclohexyl substituent (entry
4, 3:1 selectivity), meaning that the source of the stereoselec-
tivity in this system must have a strong steric component. In
one final example, we also examined the stereoselectivity with
1-cyclohexyl-substituted hydroxyalkyl azide 31, which veri-
fied that the nonselectivity of reactions of 1-substituted
MP2/6–31G* solvated energy differences of the cis (a) and trans (b)
conformers of model systems E and F.
azidoethanols extended to alkyl as well as phenyl substitution
(cf. Scheme 4b).
Calculations were performed on the syn/anti preferences on
models for the 2-cyclohexyl substituted intermediate G. Not
surprisingly, the anti conformer, Gb, is preferred by 1.9 kcal/
mol in the gas phase and by 2.3 kcal/mol in dichloromethane
at the MP2/6–311+G**//MP2/6–31G* level of theory (Figure
7). Because different ground-state configurations are favored
for models E (phenyl) and G (cyclohexyl), the actual migration
transition states for both the 2-phenyl and 2-cyclohexyl reactions
were also examined theoretically. The barriers for methyl
J. Org. Chem. Vol. 73, No. 9, 2008 3325

Katz et al.
TABLE 6. Comparison of the Gas-Phase and Solvated Free
Energy Barriers (kcal/mol) for Methyl Migration from the
Minimum Intermediate Structure at the MP2/6-311+G**//MP2/
6-31G* Level of Theory (CPCM, E ) 8.93)
For the 2-substituted azidopropanol derivatives, cation-π
effects are consistent with the following observations: (1) the
poor selectivity of the 2-phenyl version compared to an alkyl
control (2 v. 3; Scheme 2), (2) the dependence of selectivity on
the electronic density of a substituted aryl group (Table 1), (3)
ab initio calculations that are able to reproduce the observed
selectivities, and (4) the requirement that the phenyl group be
q
q
q
model
system
∆G
∆G
∆∆G
(b-a)
entry
environment
(a-aq)
(b-bq)
1
2
3
4
E
gas
solvent
gas
20.4
20.7
18.5
19.2
23.7
22.8
20.7
21.2
3.3
2.1
2.2
2.0
+
able to present its flat face to the leaving N2 group as
G
ascertained through geminal substitution (Scheme 3). With
respect to this last point, it must be remembered that there is a
steric component to the preference for an ax-Ph/eq-Me
conformation.
solvent
migration are given in Table 6. For both systems E and G, syn
migration is preferred over anti migration by 2.1 and 2.0 kcal/
mol, respectively. These energy differences, though, are much
larger than expected based on the experimental results (where
∆G is ca. 0.7 kcal/mol). However, the computations do predict
that the 2-phenyl and 2-cyclohexyl substituted systems should
have similar selectivities, as was observed experimentally.
These results are consistent with our original explanation for
stereoselectivity in this system.7c Thus, the migrating methyl
group and cyclohexyl group are syn to each other when reaction
takes place from the N2+/cyclohexyl anti conformer (Gb) and
this creates a large steric clash. Accordingly, the barrier to
migration from the less populated ground state Ga is lower and
leads to the major product. Although the same effect is also
operative in the phenyl-substituted cases, here the reactive
conformer and the more stable conformer are the same (Ea)
due to the ability of the phenyl group to rotate and reflecting
stabilization of the ground state by nonbonded attractive
cation-π interactions. Comparison of the gas phase and solvent
results for the phenyl vs cyclohexyl cases shows that the
difference in barrier heights is reduced by the inclusion of
solvent more significantly for E. This is because the syn
transition structure is stabilized by a cation-π interaction, which
is reduced in solvent. Solvent plays a lesser role for system G,
in which the difference in barrier heights is strictly the result
of steric effects. The selectivity differences observed for
substituted phenyl cases are consistent with a cation-π con-
tribution to the selectivity in this reaction as well, but the relative
contributions of this effect vs strict steric control cannot be
determined.
+
An interesting point concerns the presence of an axial N2
group in all of the proposed six-membered transition states.
Although this is the low energy N-configuration in this series
+
due to an electrostatic effect between the N2 group and the
C-O lone pair in the 1,3-oxazinane ring,10 this point is not
relevant to stereochemical analysis because only chairlike rings
containing an axial leaving group are capable of attaining the
necessary trans-antiperiplanar arrangement between the leaving
group and the migrating carbon. Thus, 1,3-oxazinanes containing
an equatorial N2+ group cannot contribute to product formation
at all.
The results obtained from substituted 2-azidoethanol reagents
show that steric considerations dominate these reactions. Ac-
cordingly, our hypothesis7c that minimization of nonbonded
interactions between a migrating methylene and an alkyl or aryl
group adjacent to nitrogen is still valid. Even here, however,
calculations show that the phenyl group is close enough for a
conceivable cation-π effect that may explain differences in the
selectivities noted in Table 5.
Experimental Section
General Procedure for the Synthesis of Lactams. A solution
of 4-tert-butylcyclohexanone in anhydrous CH2Cl2 (0.04 M) was
cooled to -78 °C and BF3 ·OEt2 (5.0 equiv) was added. After 30
min, a solution of hydroxyalkyl azide (1.2 equiv) in anhydrous
CH2Cl2 (0.04 M) was added to the cooled solution dropwise via a
cannula. The reaction mixture was allowed to warm to room
temperature slowly over 18–24 h at which time it was concentrated
under reduced pressure and excess 15% KOH was added slowly
to the residual oil. The reaction mixture was stirred vigorously at
room temperature for 30 min and then partitioned between CH2Cl2
and H2O. The organic layer was washed with water, brine, and
dried (Na2SO4) and concentrated to an oil.
These results are also consistent with the lack of selectivity
in reactions using 2-azido-1-phenylethanol (Scheme 4b). In this
case, little opportunity exists for steric clashes between the
migrating group and the phenyl substituent in the transition state
because the two groups are not adjacent on the five-membered
heterocyclic ring (see model structures F in Figure 6). This fact,
combined with the nearly equal stability of the two epimeric
intermediates (Table 4), leads to a nonselective outcome for
this case.
5-tert-Butyl-1-(2′-(4′′-fluorophenyl)-3-hydroxy-2-methylpropyl)-
azepan-2-one. Isolated as a mixture of diastereomers (yellow oil,
0.48 g, 93%). Rf ) 0.25 (50/50 EtOAc/hex). Compound (()-19b:
1
Mp 117–119 °C. H NMR (400 MHz, CDCl3) δ 7.28 (m, 2H),
6.99 (m, 2H), 4.94 (m, 1H), 3.85 (dd, J ) 4.7, 11.6 Hz, 1H), 3.38
(t, J ) 10.5 Hz, 1H), 2.91 (m, 2H), 2.51 (m, 2H), 2.41 (dd, J )
6.6, 15.5 Hz, 1H), 1.89 (m, 1H), 1.60 (m, 1H), 1.32 (s, 3H),
1.15–1.25 (m, 4H), 0.85 (s, 9H). 13C NMR (100.6 MHz, CDCl3) δ
178.7, 161.8 (d, J ) 241.4 Hz) 140.8, 128.2, 115.7, 115.5, 65.4,
57.2, 51.8, 51.6, 43.7, 36.4, 33.4, 28.9, 27.8, 24.5, 22.2. IR (neat)
3360, 1624 cm-1. MS (CI) m/z 336 [M+1]+; HRMS (FAB) calcd
for C20H31FNO2 [M+1]+ 336.2339, found 336.2339.
Summary
Although most stereoselective reactions that depend on the
conformation of a tether connecting two reactive groups may
be understood on the basis of size, these results demonstrate
that maximization of attractive nonbonded interactions can also
play a role. At present, it is evident that steric effects are
dominant in most of the ring-expansion reactions we have
studied. For example, previous work unveiled no unusual trends
in the reactions of 1,3-azidopropanol derivatives when a
substituent was placed on either the 1- or the 3-position.7c
However, attractive nonbonded interactions appear to be quite
significant in reactions of 2-substituted hydroxypropyl azides.
5-tert-Butyl-1-(2′-hydroxy-1′-(3′′,4′′,5′′-trimethoxyphenyl)ethyl)-
azepan-2-one. Compound (()-26a: White solid (0.61 g, 86%). Mp
1
148.2–152.1 °C. Rf ) 0.46 (10% MeOH/EtOAc). H NMR (400
MHz, CDCl3) δ 6.50 (s, 2H), 5.70 (dd, J ) 8.0, 5.8 Hz, 1H), 4.05
(m, 1H), 3.95 (m, 1H), 3.80 (s, 9H), 3.21 (m, 2H), 2.61 (m, 1H),
2.52 (m, 1H), 2.01 (m, 1H), 1.94 (m, 1H), 1.56 (m, 2H), 1.16 (m,
2H), 0.73 (s, 9H), 0.61 (m, 1H). 13C NMR (100.6 MHz, CDCl3) δ
177.5, 153.1, 137.5, 133.1, 105.4, 61.6, 60.8, 58.6, 56.2, 51.2, 44.2,
3326 J. Org. Chem. Vol. 73, No. 9, 2008

Nonbonded, AttractiVe Cation-π Interactions
36.7, 32.9, 29.1, 27.3, 24.1. IR (neat) 3400, 2980, 1600 cm-1. MS
(ES+) m/z 380.2 [M+1]+; HRMS (ES+) calcd for C21H34NO5
[M+1]+ 380.2437, found 380.2427.
1.24 (m, 6H), 1.13 (m, 7H), 0.86 (s, 10H). 13C NMR (100.6 MHz,
CDCl3) δ 177.6, 62.0, 51.7, 44.9, 36.7, 35.9, 33.0, 30.5, 30.1, 29.3,
27.5, 26.2, 25.9, 24.1. IR (neat) 3410, 2960, 1630 cm-1. MS (ES+)
m/z 296.2 [M+1]+; HRMS (ES+) calcd for C18H34NO2 [M+1]+
296.2590, found 296.2587.
Compound (()-26b: Brown oil (0.10 g, 14%). Rf ) 0.60 10%
MeOH/EtOAc). 1H NMR (400 MHz, CDCl3) δ 6.53 (s, 2H), 5.68
(dd, J ) 8.4, 5.4 Hz, 1H), 4.07 (m, 2H), 3.87 (m, 10 H), 3.66 (s,
1H), 3.25 (m, 2H), 2.67 (m, 1H), 2.56 (m, 1H), 1.99 (m, 1H), 1.60
(m, 1H), 1.21 (m, 3H), 0.77 (s, 9H). 13C NMR (100.6 MHz, CDCl3)
δ 177.7, 153.2, 137.6, 132.9, 105.5, 62.1, 60.8, 59.2, 56.2, 51.2,
5-tert-Butyl-1-(2′-cyclohexyl-2′-hydroxyethyl)azepan-2-one. Com-
pound (()-32a: White solid (0.47 g). Mp 133.0–135.8 °C. Rf )
1
0.68 (70% EtOAc/hexanes). H NMR (400 MHz, CDCl3) δ 3.70
(dd, J ) 14.0, 9.2 Hz, 1H), 3.56 (m, 2H), 3.26 (m, 2H), 2.23 (dd,
J ) 14.0, 7.6 Hz, 1H), 2.60 (dd, J ) 14.0, 7.6 Hz, 2H), 1.99 (m,
2H), 1.88 (d, J ) 12.8 Hz, 1H), 1.75 (m, 2H), 1.37 (m, 3H), 1.21
(m, 5H), 1.08 (m, 3H), 0.87 (s, 9H). 13C NMR (100.6 MHz, CDCl3)
δ 178.0, 76.1, 54.3, 51.6, 51.4, 42.5, 36.3, 33.1, 29.8, 29.1, 27.9,
44.5, 36.8, 33.0, 29.2, 27.4, 24.2. IR (neat) 3400, 2980, 1600 cm-1
.
MS (ES+) m/z 380.2 [M+1]+; HRMS (ES+) calcd for C21H34NO5
[M+1]+ 380.2437, found 380.2419.
5-tert-Butyl-1-(-2′-hydroxy-1′-(4′′-nitrophenyl)ethyl)azepan-2-
one. Compound (()-28a: White solid (0.07 g, 62%). Mp 164.4–168.5
27.6, 26.5, 26.25, 26.11, 24.0, IR (neat) 3400, 2970, 1630 cm-1
.
1
°C. Rf ) 0.10 (80% EtOAc/hex). H NMR (400 MHz, CDCl3) δ
MS (ES+) m/z 318.2 [M+Na]+; HRMS (ES+) calcd for
8.23 (d, J ) 8.8 Hz, 2H), 7.52 (d, J ) 8.5 Hz, 2H), 5.70 (t, J ) 7.2
Hz, 1H), 4.11 (m, 2H), 3.56 (br s, 1H), 3.34 (dd, J ) 15.3, 10.6
Hz, 1H), 3.18 (m, 1H), 2.69 (dd, J ) 14.5, 6.9 Hz, 1H), 2.56 (m,
1H), 2.01 (m, 1H), 1.68 (m, 1H), 1.24 (m, 2H), 0.79 (s, 9H), 0.68
(m, 1H). 13C NMR (100.6 MHz, CDCl3) δ 177.6, 147.4, 145.3,
129.1, 123.7, 61.7, 59.5, 51.2, 45.6, 36.7, 33.0, 29.3, 27.4, 24.0.
IR (neat) 3400, 2995, 1630, 1530, 1350 cm-1. MS (ES+) m/z 335.2
[M+1]+; HRMS (ES+) calcd for C18H27N2O4 [M+1]+ 335.1971,
found 335.1963.
C18H33NO2Na [M+Na]+ 318.2409, found 318.2401.
Compound (()-32b: White solid (0.461 g). Mp 110.4–111.9
°C. Rf ) 0.51 (70% EtOAc/hexanes). 1H NMR (400 MHz, CDCl3)
δ 3.80 (dd, J ) 14.4, 9.6 Hz, 1H), 3.51 (m, 3H), 3.23 (dd, J )
15.2, 6.0 Hz, 1H), 3.06 (dd, J ) 14.0, 2.0 Hz, 1H), 2.57 (dd, J )
14.4, 7.2 Hz, 1H), 2.52 (m, 1H), 1.97 (m, 2H), 1.75 (m, 5H), 1.37
(m, 1H), 1.19 (m, 5H), 1.07 (m, 3H), 0.88 (s, 9H). 13C NMR (100.6
MHz, CDCl3) δ 178.4, 75.9, 54.0, 51.6, 50.3, 42.6, 36.3, 33.2, 29.8,
28.9, 28,8, 27.9, 27.6, 26.5, 26.3, 26.1, 24.2. IR (neat) 3410, 2980,
1635 cm-1. MS (ES+) m/z 318.2 [M+Na]+; HRMS (ES+) calcd
for C18H33NO2 [M+Na]+ 318.2409, found 318.2402.
Compound (()-28b: White solid (0.04 g, 35%). Mp 160.4–167.5
°C. Rf ) 0.30 (80% EtOAc/hexanes). 1H NMR (400 MHz, CDCl3)
δ 8.21 (m, 2H), 7.52 (d, J ) 8.0 Hz, 2H), 5.87 (dd, J ) 7.6, 5.9
Hz, 1H), 4.23 (dd, J ) 11.5, 5.6 Hz, 1H), 4.08 (m, 1H), 3.15 (s,
3H), 2.70 (dd, J ) 14.1, 8.0 Hz, 1H), 2.56 (m, 1H), 2.02 (m, 1H),
1.92 (m, 1H), 1.30 (m, 3H), 0.86 (s, 9H). 13C NMR (100.6 MHz,
CDCl3) δ 177.4, 147.3, 145.5, 128.7, 123.7, 61.9, 58.5, 51.5, 45.5,
Acknowledgment. This research was supported by the
National Institute of General Medical Sciences of the National
Institutes of Health (GM-49093 to J.A.). Acknowledgment is
made to the Donors of the American Chemical Society
Petroleum Research Fund for partial support of this research
(to J.L.P.). This work was partially supported by the National
Center for Supercomputing Applications under MCB040056N
and utilized the IBM P690 and the SGI Altix. We also thank
Paul Hanson and his group for use of their gas chromatograph.
36.5, 33.0, 29.7, 27.5, 24.0 IR (neat) 3390, 2990, 1630, 1530 cm-1
.
MS (ES+) m/z 335.2 [M+1]+; HRMS (ES+) calcd for C18H27N2O4
[M+1]+ 335.1971, found 335.1962.
5-tert-Butyl-1-(1′-cyclohexyl-2′-hydroxyethyl)azepan-2-one. Com-
pound (()-30a: Pink solid (0.35 g). Mp 105.6–106.4 °C. Rf ) 0.26
1
(70% EtOAc/hexanes). H NMR (400 MHz, CDCl3) δ 3.80 (m,
4H), 3.34 (dd, J ) 15.2, 9.2 Hz, 1H), 3.30 (dd, J ) 14.4, 6.0, 1H),
2.54 (m, 2H), 1.96 (m, 2H), 1.70 (m, 6H), 1.22 (m, 6H), 0.97 (m,
2H), 0.85 (s, 9H). 13C NMR (100.6 MHz, CDCl3) δ 177.6, 62.3,
51.2, 36.9, 36.4, 33.0, 30.9, 29.8, 29.5, 27.4, 26.3, 25.9, 25.8, 24.1.
IR (neat) 3400, 2980, 1640 cm-1. MS (ES+) m/z 296.25 [M+1]+;
HRMS (ES+) calcd for C18H34NO2 [M+1]+ 296.2590, found
296.2587.
Supporting Information Available: Calculation methodo-
logy and results, experimental details (including starting material
preparations, characterization of new compounds, and stereo-
chemical determinations), copies of 1H and 13C spectra of new
compounds, and CIFs for compounds 19b, 26a, 28b, 30b, and
32a. This material is available free of charge via the Internet at
http://pubs.acs.org.
Compound (()-30b: White solid (0.08 g). Mp 126.4–129.8 °C.
1
Rf ) 0.40 (70% EtOAc/hexanes). H NMR (400 MHz, CDCl3) δ
4.14 (s, 1H), 3.83 (d, J ) 11.2 Hz, 1H), 3.63 (m, 1H), 3.33 (m,
2H), 2.79 (br s, 1H), 2.55 (m, 2H), 1.96 (m, 2H), 1.62 (m, 6H),
JO800222R
J. Org. Chem. Vol. 73, No. 9, 2008 3327